mee Sian ne pre rer ene J +7 ; hee - oe od rte : : eg — : = - ¥ nee : A ee eae Sana Tad eS eo CT er a re es Pe ee) ee er) ee Oy ee a fT vn My Sy i) FO : d = St as SL aa aia Ta: lle A aaa teal Sa Ne oe on a a ES Td CTT a cee rere es See a ‘@Schoni of Bysene sf 4, arnt \ Tnin: eatin abs a re Secs ay THE JOURNAL OF IMMUNOLOGY VOLUME IV eal | _ Library govl of Freer ; Q sarrstiy at Zeyusht \\ BALTIMORE, MD. 1919 i . ane q, } re / F / ae , ace) os 906197 CONTENTS NoumsBer 1, January, 1919 The Study of Problems of Immunity by the Tissue Culture Method. III. A Method for Determining the Resistance of Individuals to Diphtheria Infection. Montrose T. Burrows and Yoshio Suzuki................... 1 Antibody Production in Rabbits Following Injection with Pancreatic Fer- ments. Heinosuke Wago. : berets aa eee ad) On Red Cell Globulin. C. B. Bonert sul Carl i x ‘Schnwaes Meera street: 29 NuMBER 2, Marca, 1919 Studies in Osmotic Pressure. I. The Mechanism of Boric Acid Hemolysis. MRE SITIO SAGAN Wo a otsts Sh. S ES 4 PER Satna eee ch ee ea fads He gD Studies in Osmotic Pressure. il. The Nature of Osmotic Pressure. Mit- S UNIMOG UIP Rs gey eins Ror 2 Pes 4 SN EE ae Leen face ons tap eam NuMBER 3, May, 1919 Experiments on the Effect of Agglutinins. B. Fujimoto.. 67 A Comparative Study of Hemolytic Complement and ieeibedios t in /Oralated Plasma and Serum. Susumu Watanabe. Peres tae A Method for the Production of a Homaneneans Saepenaient of Bavillists an- thracis to be used in Agglutination Reactions. Arlyle Noble. Seerees LOS On the Nature of Eclampsia. Isei Obata. res eit The Production of Anti-Human aE Suilicii, iaghea B. Wedded) a2 141 NoumBER 4, Juuy, 1919 The Bactericidal Action of the Whole Blood of Rabbits Following Inocula- tions of Pneumococcus Bacterins. George D. Heist and Solomon Solis- Cohen. RnR INE a ie sranid) Ue The Réle of Bacillus afinenaaen in wie iias “cell Ta dvense: E. M. Huntoon and S. Hannum. : 5 Akey/ Notes on the Soncteniolonen Snel ( on Gone iselective Denen of B. influenzae Pfeiffer. C. Roos.. . 189 Immunologic i iepaeibies: OF Snore andl Recetnenvel inees of ‘B. OHS. Ralph R. Mellon and Lillian M. Anderson. . 203 The Perfusion Experiment in the Study of airman Ueinee F. oem . 209 Studies in Protein Intoxication. III. Visceral Lesions in Rabbits with Chronic Protein Intoxication. T. Harris Boughton................... 213 The Mechanism of the Anaphylaxis Reaction in the Rabbit. Arthur F. Coca 219 Some Suggestive Experiments with B. influenzae; Its Toxin and Antitoxin. Ne oubornvennor tt. MM. Houghton: 4: ..0cenceMoc i ice Soh ener os ooh oo ens dam eee iii iv CONTENTS NoumBeEr 5, SEPTEMBER, 1919 A Study of the Thermolabile and Thermostabile Antilysins (Anticomplemen- tary Substances) of Human Serum. Takaatsu Kyutoku.............. 239 Experiments on the Removal of Hemagglutinin from Rabbit Anti-Human Serum. Joseph BE. Sands and Lyle Bl West..2..2..--2).20-. -.- oe eee 275 A New Method of Testing Anti-typhoid Serum. Yoshimoto Fukuhara and Masaaki Yoshioka. , . 285 A New Method of Testing eect Dy cee coe eVects Pubes hara. See Si ateia tas Pee eee . 299 Bzperimental Purpura. "Mark nd ‘Gottlieb. : ay . 309 The Antigenic Property of the Pfeiffer Baciins. as ‘Related io 2s Nees = in the Prophylaxis of Epidemic Influenza. Charles W. Duval and William H. Harris. Pe tne ee ee ee OMe 5 | The Poisons of the Teeny Baciines Shake aL) Parker’... 3) 3-202 eee 331 On the Existence of a Multiplicity of Races of B. influenzae as Determined by Agglutination and Agglutinin Absorption. Eugenia Valentine and Georgia’ M. Cooper’ :..:0. 00.52.02 2ee Race 2 =. - Wee eee Jee Studies in Protein Intoxication. IV. Histologic Lesions Produced by Injec- tions of Pepton. T. Harris Boughton. . at 2/0 Observations on the Production of an panes for ees Hates = Bacterium welchiit (Bacillus aerogenes capsulatus). William W. Ford and George Huntington Williams: 2: 3. s42m2. 222+ 22 ese oe eee el NuMBER 6, NoveMBER, 1919 The Influence of Desiccation Upon Natural Hemolysins and Hemagglutinins in-;Human Sera. John A. Kolmer .,;..0:2 2225 36. Rein cease ee eee ee The Nature of Thermolabile Hemolysins. John A. Kolmer............... 403 Complementary and Opsonic Functions in Their Relation to Immunity. A Study of the Serum of Guinea-Pigs Naturally Deficient in Complement. Hiram BD: Moore <<: 2.022. 35.6 ois > ces ease 6 oe eR Ee Ree eee ee 425 THE STUDY OF PROBLEMS OF IMMUNITY BY THE TISSUE CULTURE METHOD IiII. A METHOD FOR DETERMINING THE RESISTANCE OF INDIVIDUALS TO DIPHTHERIA INFECTION MONTROSE T. BURROWS anp YOSHIO SUZUKI The Department of Pathology, Washington University Medical School, St. Louis, Missouri Received for publication August 24, 1918 Many attempts have been made to devise a simple and accurate laboratory method for determining the presence of toxic and antitoxic substances in the blood of persons suffering from acute infections of various kinds and in the blood of nor- mal individuals. Several methods adapted for the study of certain toxins have resulted. Many of these in the hands of most workers have been difficult to manipulate. Others have failed to give the qualitative and quantitative data necessary. It has been evident since the beginning of the study of toxins and antitoxins that their presence or absence must be determined by testing their effect upon animal tissues. Other and more accurate methods will probably not be devised until the chemi- cal and physical properties of these substances are better under- stood. For a long time it has seemed apparent that the tissue culture method might aid very materially in making such determina- tions. Stimulated by the urgent need of a more careful study of many of the cases of diphtheria and other infections as they have been seen in the hospital and as they are reported from army camps we undertook recently to apply the method in this direction. As is well known the tissue cells of many animals grow readily not only in their own plasma but also in drops of plasma of many other animals. It seemed evident, therefore, that the tissue of an 1 THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 1 2 MONTROSE T. BURROWS AND YOSHIO SUZUKI animal particularly susceptible to a given toxin might be used to detect the presence of this toxin or substances which neutral- ize this toxin in the plasma of any animal in which the tissue will grow. In a previous article (1) we have already shown how this method may be thus directly applied for the study of natural immunity. The object of the experiments was to determine whether the natural immunity of certain animals is the result of a specific resistance on the part of the cells, whether it is due to the neutralizing substances in the plasma of the ani- mals, or, both. Chickens and guinea-pigs are susceptible to diphtheria toxin. Rats are resistant. The cells of the rats grow readily, not only in their own plasma but in the plasmata of chickens and guinea-pigs. In turn, tissue cells of either chickens or guinea-pigs grow actively in any one of the plasmata of the three animals. We tested the resistance of the cells of each of these animals to diphtheria toxin in the plasma of each of the animals. Through these experiments it was possible to show that the natural immunity of rats for diphtheria toxin is due both to a specific resistance of their cells and to the existence of neutraliz- ing substances in their blood which protect, also, the cells of other less resistant animal tissues. A similar study was like- wise made with tetanus toxin. In another paper (2) we have also shown how the method may be applied for the standardization of antitoxin and the study of passive immunity. For the standardization of diph- theria antitoxin the sensitive chicken cells were grown in the plasma of normal chickens to which mixtures of toxin and anti- toxin were added. Chickens were used for the study of passive immunity. It was shown that the passive immunity of a chicken previously injected with diphtheria antitoxin could be deter- mined by cultivating fragments of the heart muscle of chick- embryos in drops of the plasma of the immunized chicken to which various quantities of diphtheria toxin had been added. Chicken tissue can be readily cultivated in drops of human plasma. It became evident, therefore, that one might use this tissue to develop a method for the detection of small quanti- IMMUNITY BY TISSUE CULTURE METHOD 3 ties of diphtheria toxin and antitoxic substances in the blood of patients or normal individuals. As an introduction to this particular use of the method we have studied the neutralizing value of the blood of a number of normal adults and children. THE METHOD Fragments of the ventricular muscle of the heart of fifteen and six- teen day chick-embryo, carefully cut so that they are approximately 1 mm. in diameter, are placed on the surface of a specially cleaned cover glass and covered with a layer of medium consisting of 1 part of the plasma to be tested and 1 part of a 0.9 per cent NaCl solution (con- trol) or 1 part of the plasma to be tested and 1 part of any one of vari- ous dilutions of diphtheria toxin in 0.9 per cent NaCl solution. The medium is spread about the tissue so that it forms a layer approxi- mately 0.5mm. in thickness. A hollow ground slide previously rimmed with vaseline is then immediately inverted over the cover glass so that the tissue and the medium become sealed within the hollow chamber. When the plasma has clotted the slide is inverted, the cover is further sealed with paraffin and the slide is incubated at a temperature suitable for the chicken tissue, 39.4°C. The toxin used for all these experiments was obtained from Parke, Davis and Company. The strength of this toxin, as indicated on the label is: L + Dose = 0.29. For the study of each sample of blood at least 18 cultures have been prepared and in making these at least 8 different dilutions of the toxin have been tested. In many of the experiments we have studied as many as 12 different dilutions of toxin. Since chicken tissue can be made to grow in 100 per cent of the drops of the control medium it has been found unnecessary to prepare more than 2 cultures of each toxin dilution. The general procedure used in making the test is illustrated in table 1—Plasma no. 14 of the table below (table 4). Fresh toxin dilutions have not been made for each experiment. Fresh dilutions have been made as a routine every twenty-four or forty-eight hours. The diluted toxin is left in the ice-box except dur- ing the time when samples are taken from it. The toxin dilutions have been made by the drop method and by direct measurement. The latter has been more satisfactory. They are made in ordinary specially cleaned small glass test tubes. The highest concentration of toxin which we have used has been 10 per + MONTROSE T. BURROWS AND YOSHIO SUZUKI cent. Higher concentration may, however, be used without injuring the growth of the cells as long as the isotonicity of the solution is not too greatly disturbed and the plasma neutralizes the action of the toxin. TABLE 1 Medium: Plasma of case 14, (table IV), 1 part; various dilutions of diphtheria toxin, 1. part Control: The same plasma, 1 part; 0.9 per cent NaCl solution, 1 part Tissue: Fragments of the ventricular muscle of the heart of a fifteen-day chick- embryo GROWTH AFTER TOXIN, NUMBER OF TURE NUMBER TIMES DILUTED ABE 24 hours 48 hours 72 hours 10 1 ~ oe ee 2 a= ee 50 3 - at sas 4 - ae eee 100 5 - as aa 6 ~ - ae ++4+ 500 7 - oe ae 8 - aa a+ 1000 9 - - a aaa 10 aa a 3000 11 ~ ae ee 12 - aa mas 5000 13 - aa Hane 14 ~ aa ee 7000 15 - tit asa 16 - +++ wae Control ile — +4+ ee 18 - a= ++4+++ From the 10 per cent solution of toxin first prepared the next lower concentration is made by carefully measuring known quantities of it into a known quantity of 0.9 per cent NaCl solution and so on. The cultures are prepared as soon as the plasma is obtained and before any possible clotting has taken place. IMMUNITY BY TISSUE CULTURE METHOD 5) The preparation of plasma. Some difficulty may arise in preparing human plasma unless certain precautions in technique are especially taken. In the preparation of plasma from lower animals it has been customary here in the laboratory to draw the blood from an artery through a carefully cleaned and oiled cannula into specially cleaned glass test-tubes. A rather slender long test-tube is used. This test tube is filled half full of blood and as soon as filled it is corked and plunged into an ice water, or ice, salt and water bath so that it becomes chilled at once. Care must be taken when salt is used not to freeze the plasma. As soon as this blood is chilled it is centrifugal- ized by plunging the test-tubes containing the blood into large centri- fuge recepticles (50 cc. capacity) filled with ice, salt and water. A rapidly running electric centrifuge (International Equipment Company, size 1, style A) is suitable for this work. For preparing human plasma we have used glass test-tubes similar to those used in animal experiments. They are 12 cm. long and have an inside diameter of 7mm. They are fitted with a large cork into which a hole has been cut so that the cork fits over the end of the tube. These test-tubes are made in the laboratory out of ordinary soft glass tubing having a thick wall. They are specially cleaned to prevent chemical contamination of the medium. A cannula cannot be used in obtaining blood from man. We use a 10 cc. Luer Syringe fitted with a needle, gauge no. 19, the point of which is kept very sharp. In many cases these syringes have been carefully cleaned, coated with olive oil and sterilized in the autoclave. The needles have been similarly cleaned, coated and sterilized. A simple procedure is to place such clean needles in small test-tubes fitted with cotton stoppers. The needles are placed point down in the tube. The sharp point is held from the bottom of the tube by a wire cleaning rod having a loop end. The Syringe with two such needles in test-tubes are wrapped in gauze and canvas and sterilized. For most of this work we have simply boiled the needle and syringe in water rinsing them later in sterile isotonic NaCl solution to prevent any haemolysis. In cold weather an ice water bath is used to chill the blood. In hot weather a little salt is added to the bath. A small flask with this is taken to the ward or place where blood is obtained. The blood must be chilled as soon as taken. For the preparation of serum it is only essential to obtain the blood. For the preparation of plasma it is essential that the blood be taken as free as possible from tissue contamination and that it be agitated as 6 MONTROSE T. BURROWS AND YOSHIO SUZUKI little as possible. Chicken plasma taken free from tissue contamina- tion may be handled roughly. This is not true of human and mam- malian plasmata. Roughly handled blood will invariably clot and prevent the test. The needle must be plunged directly into the vein at the first thrust; otherwise a new needle must be taken. The blood must be allowed to flow into the syringe. Sucking the side wall of the vein into the mouth of the needle often causes clotting to take place before the plasma can be obtained and used. In obtaining plasma for these tests we have used a tourniquet, and have cleaned the skin with 80 per cent alcohol. We take about 5 ce. of blood, transfer it immedi- ately to the tubes, allowing it to run in gently, chill, centrifuge within five or ten minutes and use it, if possible, within a half hour or an hour after it is taken. Many samples of human plasma kept cold will remain unclotted for several days. Any sample of human plasma prepared with care will remain unclotted for an hour or often longer. In cold weather we allow the centrifuge to run for two or even three minutes. In hot weather two minutes is as long as any blood is centri- fugalized. The ice in the cup is melted within this time and the plasma is warm. As soon as the tubes are removed from the centrifuge they are again plunged into the ice water. From this time on the plasma is kept cold until it is placed on the cover glass. The preparation of the culture media. The medium for the culture is prepared in slender tubes. We have used tubes 33 inches long and 35 inch in diameter. These are placed in a flask containing ice and water. A pipet is placed in each tube. These pipets have long slender ends which reach the bottom of the tube; they are made of ordinary glass tubing and are fitted with a rubber bulb, which has been cleaned care- fully. Each pipet has a file mark at the neck of the slenderend. Plasma is first added to the tubes. The pipet is filled with plasma up to the file mark and emptied into the bottom of the tube. With the same pipet an equal quantity of NaCl solution 0.9 per cent or diluted toxin is taken and thoroughly mixed with the plasma. Since very small quantities of medium are necessary in preparing the cultures we use very small quantities of plasma and diluting agent. The cleaning of glassware. On account of the small quantity of me- dium used in preparing tissue cultures and the large amount of glass surfaces to which it is exposed at different times it is very essen- tial that great care be taken in cleaning the glassware if constant results are to be expected. We have used the method that was intro- duced in the laboratory by one of us a few years ago. All the glass- IMMUNITY BY TISSUE CULTURE METHOD Fi ware is treated alike, that which has been previously used as well as the new. It is boiled first in soap and water, rinsed in tap water and placed in a weak sulphuric acid solution for several hours. When removed from this last solution it is again rinsed in tap water and placed in a dish of distilled water. The final cleansing is made with steam. A jet of hot steam is allowed to play for several minutes on the inside of each tube, pipet, flask and dish. The apparatus for steaming the glassware is similar to that described by Finlay (3). An ordinary flask, Erlenmeyer or otherwise, is fitted with a tight cork through which passes a slender glass delivery tube. A funnel is placed over this tube in such a way that the tube passes through the stem of the funnel. The stem of the funnel, which is cut very short, is fitted tightly into the outer part of the cork. By this procedure the funnel is held firmly in place and its lower end is closed. It acts asa recept- acle for the condensed steam. The flasks, pipets, dishes and tubes are inverted over the delivery tube through which the steam passes. We use distilled water in the flask. After steaming, the glassware is dried and sterilized at 200°C for one hour. The prevention of infection of the cultures with bacteria from the air. During the preparation of the cultures they are exposed for some time to the outside air and in an ordinary room a large number will invari- ably become contaminated with bacteria. The cultures must be kept sterile at all times. If a small room free from dust and draughts is not available a box may be used. The one which we have used is 3 feet wide, 14 inches high and 2 feet deep. The sides are made of wood. The front, top and back are fitted with glass windows. The back reaches only to within 4 inches of the table, leaving an opening for inserting the hands and material into the box. The bottom is open. The box is placed on a black top table or a table covered with a black cloth. All cultures have been prepared in such a box. This box has been used by one of us for this purpose for several years and in many different rooms as well as in the open. It has always been found practical. The control. In making tests of this kind it is most essential that the growth of cells in the cultures is one that can be predicted and regulated. Many tests of different substances previously made with the tissue culture have been found later to be of little value on ac- count of the failure to have had a proper control. The finding of methods for controlling growth in the tissue culture formed one of the 8 MONTROSE T. BURROWS AND YOSHIO SUZUKI early problems of one of us in the study of this method. To detect peculiarities of plasma and the presence of toxic substances within a plasma one must be certain of what is to be expected under normal conditions. A statement of the theoretical deductions and facts elucidated by the development of a method for controlling growth of cer- tain tissues in culture may aid in the understanding and the application of the more minute details of the technique of pre- paring the cultures. The tissue culture cannot be compared in detail with the bac- terial culture. The tissue cells planted in plasma do not grow at the expense of the plasma. Single cells may show movement in this medium but they do not grow. After a short period of movement they show evidence of exhaustion. Growth takes place only about fragments of tissue. When fragments of 1 mm. or less in diameter are used the growth is inversely proportional to the size of the fragment. This proportion holds for a given tissue only when the fragments are of equal cellular content and arrangement. A small fragment may be made large by teasing the cells apart. The growth from such a fragment is less than that from one of the same size which is compact. The growth of cells in a tissue culture is the growth of the cells on the periphery of the fragment or cells that have migrated there. The nutrient material for the growth of these cells comes from the cells disintegrating within the fragment. The growth as it is observed in the culture is not materially different from movement or cell migration. Both are none other than a simple transfer of materials from the fragment to the media without (4). The maximum growth or migration of cells from the fragments of most tissues is seen about fragments of that tissue 1 mm. in diameter. The interchange or diffusion of sub- stances is disturbed in fragments of greater thickness. The growth may be the same but it is never greater about larger fragments (5). Any thing that again disturbs the diffusion of material from the fragments into the outer medium disturbs the growth and IMMUNITY BY TISSUE CULTURE METHOD 9 movement of cells. Again, oxygen does not diffuse readily into the plasma to a depth greater than 0.7 mm. In preparing the cultures it is essential that the cover’ glass surface and the ‘thickness of the medium layer be kept constant. Another fac- tor is the density of the clot. Dilution of the plasma with liquids leads to a greater growth of cells (6). In all cases we have di- luted the control medium exactly as in the experiment using 0.9 per cent NaCl solution. The layer of medium in each culture has been made approximately 0.5 mm. in thickness. Slight variations make very little difference. With a little practice one can obtain a suitably thick layer without difficulty. Approxi- mately 0.5 mm. is sufficiently accurate. The cleaning of cover glasses. A cover glass surface over which the medium does not spread easily is essential. All the cover glasses both new and those previously used are cleaned first by boiling in soap and water. They are then boiled three or four times in distilled water and placed in absolute alcohol for at least twenty-four hours. From this they are taken, dried and polished with an old silk or linen cloth, placed in Petri dishes and sterilized dry at 160°C. for one hour. If the sterilizer is hotter than this the surface is often changed and the drops of plasma tend to spread over them. Cover glasses cleaned in the manner just described have shown a uniform kind of surface. The drops of plasma hold up well on their surfaces. The choice of a tissue. The cells that grow most actively in the tissue culture are the connective tissue cells. Very active growth is, however, rarely seen about fragments of pure connec- tive tissue such as subcutaneous tissue. The active growths of these cells are seen about fragments of glands or skin or from other tissue containing epithelial cells. The epithelial cells disintegrate at the expense of the connective tissue and supply the nutriment. About these fragments the growth of the con- nective tissue is often excessive—far in excess of the original amount present. From fragments of pure connective tissue, for example, subcutaneous tissue, little more than a move- ment of cells is generally seen. Great difficulty has always been encountered, however, in con- trolling the growth of cells from fragments containing epithelium. 10 MONTROSE T. BURROWS AND YOSHIO SUZUKI The epithelium may grow or show movement and when it does it inhibits completely the growth of the connective tissue. Again, when it lies near the edge of the fragment the medium fre- quently undergoes liquefaction or retraction and no growth of cells results. From fragments of the more differentiated connective tissue the growth is often move active than from subcutaneous tissue. Such is the case of heart muscle. Liquefaction of the medium does not take place about these fragments. A large number of fragments of equal cellular content may be easily obtained from the ventricle of the chick-embryo. The tissue is firm and can be cut without disturbing the cellular density and arrange- ment in the fragment. Moreover, it grows readily and the ex- tent and type of the growth can be predicted if care is taken in the preparation of the cultures. The heart may be obtained easily in a sterile condition from the embryo. We open the eggs with sterile scissors, remove the embryo to a sterile dish, take out the heart with clean sterile instruments and place it in a sterile Petri dish. It is then cut into fragments with a sharp scalpel of large size. The cut- ting is done against the bottom of the dish or another clean and sterile piece of glass. By this means cleanly cut and solid fragments can always be obtained. As soon as cut they are transferred on the point of the knife or a pair of forceps to the cover glass, covered with plasma and sealed in the hollow slide chamber. Slight drying will have little effect. It is best to work rapidly, however. The tissue may be kept moist with isotonic salt solution. This is not advisable when quantitative results are expected because in such small drops the salt solu- tion clinging to the instruments and tissue causes a noticeable dilution. The age of the embryonic tissue. In the previous papers we have noted that the tissue cells of the younger embryos grow for a time in a greater amount of diphtheria toxin than those of older ones. The heart muscle cells of younger embryos may grow for a time in a concentration of toxin as great as 1/500 while those of older ones or of foetuses will not grow in a concentration greater than 1/6000. IMMUNITY BY TISSUE CULTURE METHOD 11 It is most suitable to use, however, a tissue which grows rap- idly and actively. Embryonic and foetal tissues grow more rapidly and actively than those of young animals and adults. TABLE 2 Medium: Plasma of case 5, (table IV), 1 part; various dilutions of diphtheria toxin, 1 part Control: The same plasma, 1 part; 0.9 per cent NaCl solution, 1 part Tissue: Fragments of the ventricular muscle of the heart of fifteen-day chick-embryo GROWTH AFTER es bee CULTURE NUMBER 24 hours 48 hours 72 hours 10 1 a 18 Be , = = ees 50 3 — = oa 4 = = = 100 5 _ = ee 6 — ee aa 500 @ — = = 8 = = aT 1000 9 aft. es - 10 att. — = 3000 1 — zie a 5000 13 zs ine a 14 =e aPar SParSr 7000 15 ar aineta SPA SF 16 a APSE aR aR Se Control tf + sR ara SRS SPI 18 sr ap 4 == SPar Sr 45 Again, when the younger embryonic tissue is used a sharp line of demarcation does not exist where one could say the medium is toxic or that the toxin is sufficiently neutralized. The cells from the fragments of the heart muscle of fifteen or sixteen day chick embryos grow readily and actively after twelve or twenty- four hours. They are sensitive in chicken plasma to a concen- 12 MONTROSE T. BURROWS AND YOSHIO SUZUKI tration higher than 1/6000 of the diphtheria toxin which we have used. The line of demarcation is sharp in all of our series. In many of the tests one finds the cells growing actively in one TABLE 3 Medium: Plasma of case 10 (table 4), 1 part; various dilutions of diphtheria toxin, 1 part Control: The same plasma, 1 part; 0.9 per cent NaCl solution, 1 part Tissue: Fragments of the ventricular muscle of the heart of a fifteen-day chick-embryo GROWTH AFTER TOXIN, NUMBER OF ; TIMES DILUTED CULTURE NUMBER 24 hours 48 hours 72 hours 10 i -- = = 2 — = = 50 3 + aay gearcr 4 E Seat tea 100 5 + aaaeas leat teas 6 =f apSSSE SPaeHe oF 500 7 +5 Spare Seana 8 a Shes co otaegts 1000 9 + sae ae 10 ss semech eee 3000 11 Ar Siectaats neaearse 12 4p aes ee Tete 5000 13 a. aaaear aesea se 14 a steal alae 7000 15 == = Rae spare 16 i sagt aera Control 17 — +44 +++ 18 sie SoS ee apaeaese dilution, no growth being seen in any of the media having a greater concentration of the toxin; table 2. In another there may be a small growth of cells in the higher concentration of toxin during the first twenty-four hours. These cells soon dis- integrate or the growth remains very small in amount and a sharp point is established; table 3. IMMUNITY BY TISSUE CULTURE METHOD 13 In all of the tables the relative amount of growth is indicated by the number of + marks. When no growth takes place the mark — is used. The sign + indicates the growth of a few cells. All cultures were followed for seventy-two hours. Readings were recorded every twenty-four hours. The concentration of toxin indicated in the tables is the con- centration in the salt solution. The actual amount in the medium is one-half that amount. EXPERIMENTS AND DISCUSSION The blood of 27 adults and 11 children have been studied by this method. The results of the study are given in tables 4 and 5. In most instances the Schick reaction was also performed The relative value of the Schick test as compared with our method is indicated in these tables. In tables 4 and 5 the black lines cover the dilution in which growth takes place rapidly in the cultures. In this series there are several individuals that had had diphtheria at a previous time. There are also two carriers. A few of those having had the disease; cases 10, 13, 14, 19, 26 and 27 (table IV), show a definite resistance. The carriers, also, are resistant. Individu- als, who have had diphtheria, do not however, necessarily show a high resistance; case 23 (table IV). In table 5 it is noted that the children show practically the same variations as those of adults. After two months in a num- ber of instances the blood of the same individuals studied in table IV was studied again by one of us. A new toxin was used but one having the same tested strength of the other. In a few of the eases slight differences were noted in the results. In others no such differences were seen. It is known that the plasma of an individual may be more suitable as a medium at one time than another. Plasma taken from animals that have been starved for twenty-four or forty-eight hours is invariably better than that taken from a feeding animal. Plasma taken from animals fed on a protein diet has been found to be a better medium than that taken from an animal fed on carbo- 14 MONTROSE T. BURROWS AND YOSHIO SUZUKI TABLE NUMBER 4. Previous Attack |Schicld ‘Toxin Dilutions TSE meee een ee BE aos a Shs een ep [pone [+ fo[F[ 20 [none [+ | epee [ages] + a refi as pa 20d — vyeaTts ago 25 ado xx none ; ibe Be ies s[i[ 26 | none | = | , oe = see] = BESS came oo SS So DR ee Eo DO ** Bowie ee ee ** Treated with antitoxin x Diphtheria carrier IMMUNITY BY TISSUE CULTURE METHOD 15 hydrates or fats. The differences are not striking, however. A plasma taken after a full meal is cloudy with fat. Growth is generally inhibited to a noticeable extent in such a plasma (9). To what extent variation in climate, food, ete., may effect slight differences in the relative resistance of tissues in the cul- tures is a problem of interest but one which has not been studied by us up to the present time. TABLE NUMBER 5. Tees Dilutions immunizing ose [acountizang res dose Several years ago Marks showed that diphtheria antitoxin standardization may be made by injecting mixtures of toxin and antitoxin into the subcutaneous tissue of guinea pigs. Toxin injected alone causes local oedema. This is prevented by adding sufficient quantities of antitoxin to the toxin solution. Marks claimed that this method is more sensitive than that used by Ehrlich and others for such standardization. Romer noted that necrosis followed the intracutaneous injection of diphtheria toxin in guinea pigs. He introduced a 16 MONTROSE T. BURROWS AND YOSHIO SUZUKI method of diphtheria antitoxin standardization based upon this fact. If necrosis follows the injection of toxin and antitoxin, free toxin must be present in the mixture. Completely neutral- ized mixtures cause no necrosis. This method according to Zinsser has not been found easy except in the hands of those that have had a great amount of practice and experience with it. The method has been used, however, by several investigators for the detection of toxin in the blood of diphtheria patients (7). The generally used method for determining the resistance of individuals to diphtheria is that introduced by Michiels and Schick. They made their injections directly into the skin of the patient. This method has been used in many of the large epidemics for separating those individuals that are likely to become infected from those that are not. The method, as it is used, and as we have used it in these tests consists in injecting intracutaneously 0.1 ce. of a solution containing one-fiftieth of a minimum lethal dose of diphtheria toxin. Asa control a similar amount of the same diluted toxin, heated previously for one hour at 60°C. is injected. The control is most essential because in a number of individuals a reaction is seen, generally within the first twenty-four hours, which is due undoubtedly to substances other than the toxin. This method is a quantitative test for neutralizing substances in the tissue fluids and blood of the person tested. If neutralizing substances are present the toxin is made inert. If they are not present local hyperaemia-oedema and even necrosis of a noticeable degree results from the pres- ence of the free toxin. A concentration of toxin sufficient te cause marked necrosis is never used. We have made all our readings after forty-eight and seventy-two hours and called only those positive where the hyperaemia was most marked in the area injected with unheated toxin (8). This particular time of year has not been one most suitable for obtaining cases of diphtheria for study. One of us recently has seen 3 cases and has obtained blood from them at the time of admission and before the treatment was given. The study of these cases has shown us that this method may be used for the IMMUNITY BY TISSUE CULTURE METHOD 17 quantitative estimation of toxin just as the results of the experi- ments reported in this paper have shown that the method is suitable for a quantitative estimation of the neutralizing value of the blood of normal individuals. It is evident, however, that this method finds application, with slight variation in the choice of the tissue, not only for the study of diphtheria but for the study of other toxins and anti- toxic substances. The work with bacteria in relation to pyo- genic infections has given many promising results but in spite of these results the great majority of these infections remain a menace. It has become evident that there is an urgent need for progress in new directions for the study of these diseases. While at Cornell University Medical College one of us noted a toxic substance capable of killing human tissues in the blood of a pneumonia patient. Toxic substances were also seen in individuals recovering from acute alcoholism. A more careful study of these cases may lead us to important conclusions. These latter facts have made it of greater interest to submit the report of these experiments for publication. REFERENCES (1) Suzux1, Y. 1918 A study of problems of immunity by the tissue culture method. I. A study of the cells and blood plasma of animals which are naturally resistant and others which are susceptible to diphtheria and tetanus toxins. Journ. Immunol., Vol. III, no. 3, pp. 233-246. (2) Burrows, M. T., anp Suzuki, Y. 1918 A study of problems of immunity by the tissue culture method. II. The tissue culture as a means for quantitatively estimating toxins and antitoxins and determining the distribution of antitoxin in passively immunized animals. Journ. Immunol., Vol. III, no. 3, pp. 219-232. (3) Frnuay, ALEx. 1910 Practical physical chemistry, Longmans, Green and Company, London. Fig. 51, p. 149. (4) Burrows, M. T., Burns, J. Epw., anp Suzuki, Y. 1917 Studies on the growth of cells. The cultivation of bladder and prostatic tumors outside the body. Journ. Urol., 1, no. 1, 3-15. (5) Burrows, M. T. 1914 Tissue culture in vitro. Vol. xvii, The International Cong. of Med., London, 1913, Gen. Path. and Path. Anat., Section III, Discussion no. 3, pp. 217-237. (6) Burrows, M. T. 1913 The tissue culture as a physiological method. Trans. of the Cong. of Am. Physicians and Surg., 9, 77-90. THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 1 18 MONTROSE T. BURROWS AND YOSHIO SUZUKI (7) ZinssER, Hans 1916 Infection and resistance, The Macmillan Company New York, pp. 460-463. (8) Kotmer, Jonn A. 1917 Infection, immunity and specific therapy, 2nd edi- tion, Saunders, Philadelphia and London, pp. 228-232. (9) Burrows, M. T. Unpublished notes. ANTIBODY PRODUCTION IN RABBITS FOLLOWING INJECTION WITH PANCREATIC FERMENTS HEINOSUKE WAGO From the Laboratory of Preventive Medicine, University of Chicago Received for publication September 11, 1918 Accumulated investigations give evidence that ferments when injected into suitable animals display antigenic properties in that they stimulate the production of specific antibodies. Thus: Hildebrandt (1) in 18938, produced specific antibodies to the enzyme emulsin. Morgenroth (2) in 1899 and Briot (3) in 1900, were both successful in producing a specific antibody to rennin. In 1901 Achalme (4) produced certain specific antibodies to pancreatin, while Sachs (5) a year later produced similar antibodies to pepsin. In 1903 Levene and Stooky (6) increased the antitryptic action of serum by the experi- mental injection of pancrease extract, and Halpern (7) in 1911 claimed similar results from the injection of dog pancreas into dogs. Most recently (1914) Marras (8) obtained precipitating and complement- fixing antibodies to trypsin. In the course of experiments directed toward further deter- mination of the antibodies produced to pancreatic ferments, I have made observations that possess a bearing upon the pro- duction of antibodies in general and which it is the purpose of this paper to present. The general line of experimentation was the injection of pancreatic ferments in various forms into rabbits, with a subse- quent determination of the presence or absence of specific anti- bodies in the body fluid of the recipients. For this purpose I prepared pancreatin solutions as follows. Commercial pan- creatin powder (Parke, Davis and Company) was added to physiological salt solution in the amount of 1 per cent by weight. After repeated shaking, the solution was filtered through hard- 19 20 HEINOSUKE WAGO pressed filters and finally through Berkefeld candles. The resulting filtrate was clear and sterile. Similar solutions were made of trypsin (Central Scientific Supply Company) and of amylopsin (Digestive Ferments Company). For injection, these solutions were employed intravenously, 5 cc, being introduced per kilo weight of animal. The serum of animals receiving the injections of the above materials was tested from time to time for specific antibody content. The tests were directed especially toward the detec- tion of precipitins, complement deviating antibodies, and of antibodies specifically inhibiting the ferment action of trypsin. The tests for precipitins were made against 1 cc. of the above 1 per cent ferment solutions: The complement deviation tests were made against twice the minimal activating dose of guinea pig serum for an immune serum specifically cytolytic for sheep erythrocytes: The antitrypsin tests were made against twice the amount of pancreatin or trypsin required to digest a car- mine-stained fibrin flake, according to the method of Griitzner (9). Systematic injection of animals with the pancreatin showed that only after a multiplicity of injections was there a distinctly recognizable production of antibodies of any type. After six or more injections, however, a precipitin was demonstrable which reacted with solutions of pancreatin, of trypsin, and of amy- lopsin, and with all to about an equal degree. Reacting with such solutions, the immune sera also produced complement deviation. Thus, rabbit 9 after receiving three 9 ec. injections of the 1 per cent pancreatin in six days, showed no distinct antibody content of its serum on the eighth day following the last injec- tion. But after five additional similar injections, its serum finally displayed a marked content of precipitins for pancreatin solution; 0.00025 ce. effecting a distinct reaction. Complete complement deviation was produced by 0.025 cc. On the other hand, the serum displayed no specific power of inhibiting pan- creatic digestion of fibrin. The antibodies formed therefore were clearly gained to admixed substances in the pancreatin preparation and not to the fibrinolytic ferment per se. The ANTIBODY PRODUCTION IN RABBITS 21 results cited for rabbit 9 were approximated in all of eleven rabbits similarly injected. The inability to produce antibodies in appreciable amount by a few injections of the relatively strong pancreatin solution, led to the examination of the urine as a suspected avenue of rapid excretion of the pancreatin, thus reducing its antigenic action. It was at once found that following intravenous injection, pancreatin is excreted in the urine in very large quantities. This applies not only to the ferment constituents themselves but to the admixed proteins as well. In general the urine of normal rabbits possesses no tryptic action, but within two hours after the intravenous injection of pancreatin as practiced in these experiments, the urine displayed a marked pancreatic action and continued to do so for forty-eight hours. Fifty-three rabbits were injected with the 1 per cent pancreatin solution and were killed after the duration of various time intervals the urine being at once drawn from the bladder with aseptic precautions. Forty- seven of the specimens thus obtained from animals killed in from two to forty-eight hours following the injection, displayed the power of completely digesting fibrin in vitro by the Gritzner method, most of the urines producing the digestion when diluted tenfold. They also contained proteins of native pancreatin, as shown by their reaction with specific precipitins to these substances. I next attempted to obviate this escape of the pancreatin antigens in the urine by injecting them in a less soluble form. To this end I prepared a pancreatin solution as follows: A 10 per cent solution of pancreatin was made in physiological salt solution, and after thorough shaking was filtered through hard-pressed paper. To the resulting filtrate was added 20 volumes of absolute alcohol and the mixture was allowed to stand thirty minutes with the resulting formation of a white precipitate. The mixture was then centrifugalized and the sediment was rapidly resuspended in an amount of salt solution making the concentration 2 per cent relative to the amount of pancreatin powder originally employed. The pancreatin so treated with alcohol, retained its fibrino- lytic value unmodified but when injected, it was excreted by the 22 HEINOSUKE WAGO urine much less extensively than unmodified pancreatin. The following experiment (table 1) shows that the alcohol treatment of the pancreatin did not diminish its fibrinolytic value. Although as shown above, the fibrinolytic action of pancreatin was not modified by the alcohol treatment, its susceptibility to excretion by the kidneys was much reduced as shown in the following table (table 2). The above table shows a greatly reduced excretion of the proteolytic ferment of pancreatin via the urine when the pan- creatin is injected in an alcohol modified form. There was also TABLE 1 Digestion of fibrin 1 PER CENT A. UNMODIFIED PANCREATIN Be genset aint ce. 1.0 Complete Complete 0.75 Complete Complete 0.5 Complete Complete 0.35 Complete Complete 0.25 Complete Complete 0.15 Complete Complete 0.1 Complete Complete 0.075 Complete Complete 0.05 Slight Slight 0.035 Slight Slight 0.025 0 0 0.015 0 0 a decreased excretion of the admixed pancreatin proteins as shown by the absence of a precipitin reaction with specific sera. The antigenic value of the alcohol-modified pancreatin was next tested. To this end I employed alcohol-modified pan- creatin for intravenous injection in experiments paralleled by the similar injection of unmodified pancreatin. Whereas in no instance was it possible to produce an appreciable amount of antibodies by a single injection of the unmodified pancreatin, the corresponding injection of the modified pancreatin invariably produced a very considerable amount of antibodies. Thus rab- bit 14, which received a single injection of 10 cc. of the alcohol precipitated pancreatin, solution, displayed on the eighth day ANTIBODY PRODUCTION IN RABBITS 23 such antibodies in its serum that 0.025 ce. of the serum produced a specific precipitin reaction and 0.035 cc. effected complete complement deviation. Rabbit 12, similarly injected but with unmodified pancreatin, showed no analogous antibody content of its serum. Simiar differences were observed in all experi- ments where alcohol modified pancreatin injections were opposed to native pancreatin injections. It is to be concluded, therefore, that a modification of the pan- creatin rendering it less soluble and limiting its excretion by the urine, favors its power of stimulating the production of anti- bodies to the native proteins of pancreatin. Even with the al- TABLE 2 Digestion of fibrin by twenty-four hour urine of rabbits INJECTED WITH URINE = A. Unmodified pancreatin B. Alcohol modified pancreatin cc. 0.1 Complete Complete 0.075 Complete Slight 0.05 Complete 0 0.035 Complete 0 0.025 Slight 0 0.015 0 0 cohol—modified pancreatin, however, no demonstrable antibody was formed specific in inhibiting the fibrinolytic ferment. As a further means of favorably modifying the antibody pro- duction to the proteins of pancreatin, I employed injections of sodium iodoxybenzoate. Hektoen (10) and Arkin (11) showed that when this substance is injected intravenously immediately following the introduction of erythrocytes or bacteria as antigens, the antibody production is distinctly greater than when antigen alone is injected. Applying the general technique of the above workers, I injected 5 cc. of a 7 sodium iodoxybenzoate solution into rabbits immediately following the introduction of native pancreatin as antigen. The results obtained were analogous to those recorded by Hektoen and by Arkin in that the animals receiving the iodoxybenzoate displayed a distinctly greater anti- 24 HEINOSUKE WAGO body production than the control animals not receiving the salt. The contrast was most clearly seen in animals which received but a single antigen injection. Such animals as received no iodoxybenzoate, produced no appreciable amount of antibodies, whereas all animals receiving the iodoxybenzoate produced a very distinct amount of precipitins. The following table (table 3) gives quantitative differences in the precipitin antibody con- tent of the sera of two animals receiving the same injection of pancreatin, the one with and the other without subsequent injection of the sodium iodoxybenzoate. The tests were made on the eighth day following a single injection of 10 cc. of 1 per cent pancreatin. TABLE 3 Effect of sodium iodorybenzoate upon production of precipitins to pancreatin PRECIPITIN VALUE OF SERUM FROM IMMUNE SERUM SS a | i eee Rabbit A, receiving pancreatin Rabbit B, receiving pancreatin only and iodoxybenzoate 0.1 0 Tear 0.075 .0 = 0.05 0 +++ 0.035 0 +++ 0.025 0 a+ 0.015 0 Sa 0.01 0- 0 From table 3 it is seen that whereas no specific precipitins were demonstrable in the rabbit receiving pancreatin alone, a very distinct content of precipitins was present in the serum of the rabbit receiving pancreatin and iodoxybenzoate. The experiments thus far given show that an increased pro- duction of antibodies to pancreatin proteins may be induced in two ways: First, by an alcohol modification of pancreatin, render- ing it less soluble; and secondly, by the introduction of sodium iodoxybenzoate into the recipient animal. I next sought to determine whether or not the combination of these two favorable factors would yield an antibody produc- tion greater than that induced by either singly. The experi- ANTIBODY PRODUCTION IN RABBITS 25 mental results showed that such was the case, namely, that the maximum antibody production was gained where the alcohol modified pancreatin was used as antigen and where the injection of the antigen in this form was accompanied by an injection of sodium iodoxybenzoate. The following table (table 4) presents a comparison of the amount of precipitins gained by the single TABLE 4 Relative precipitin production to pancreatin PRECIPITIN REACTION by SERUM FROM . . D. Rabbit receiv- IMMUNE SERUM] 4 Rabbit receiv- | B. Rabbit receiv- | C: Rabbit receiv- ing alcohol-modi- ing unmodified | ing alcohol-modi- earn aa epee ed _— pancreatin pancreatin fied pancreatin eonie heey AIS and. iodoxy ben- zoate ———_—_ EES oooooooqocoocoqoo 0.00035 0.00025 0.00015 0.0001 Seco ooOCCCCOCecceottttt Sooooooooom oo tf teat te ott+++ttt+++4¢4+44¢4444 injection of unmodified pancreatin, of alcohol-modified pan- creatin, of unmodified pancreatin plus iodoxybenzoate, and of alecohol-modified pancreatin plus iodoxybenzoate. The above table (table 4) shows that whereas alcohol-modified pancreatin on the one hand, and unmodified pancreatin plus iodoxybenzoate on the other, have a greater antigenic value than unmodified pancreatin alone, a far greater antibody pro- 26 HEINOSUKE WAGO duction is effected by the combined use of alcohol-modified pancreatin and iodoxybenzoate. The figures tabulated refer to differences obtained in animals receiving a single injection of antigen. Similar relations hold however in animals receiving multiple injections and by employing the modified pancreatin and iodoxybenzoate in successive injections, sera were produced of which 0.000015 cc. effected the specific precipitin reaction. No sera of comparable value were obtained by any other procedure. As the result therefore of the study here transmitted I have to conclude: 1. That the precipitins and complement deviating antibodies produced in response to pancreatin injection are distinct from such antibodies as may inhibit the proteolytic ferments of pan- creatin. The former, as in the present experiments, may be produced in large amounts in the absence of the production of the latter in any degree whatsoever. 2. That following intravenous injection of pancreatin, the proteolytic ferments and proteins of pancreatin are extensively excreted by the urine. 3. That the antigenic value of pancreatin for stimulating the production of antibodies to the contained prcteins is enhanced by an alcohol modification of the pancreatin. 4. That the production of precipitins to pancreatin as a soluble antigen is favorably influenced by the intravenous injection of sodium iodoxybenzoate. 5. That the production of precipitins to pancreatin may be enhanced to the greatest degree by employing alcohol-modified pancreatin as antigen and by accompanying its injection with that of sodium iodoxybenzoate. I wish to express my thanks to Prof. Preston Kyes for many helpful suggestions. ANTIBODY PRODUCTION IN RABBITS 26 REFERENCES (1) HinpEBRaANpDT: Virchow’s Arch., 1893, 131. (2) MorcEenrorta: Centralbl. f. Bakt., 1899, 26, Nr. 11, u. 12. (3) Brior: Thése de Paris, 1900. (4) AcHALME: Ann. de. L’Inst. Pasteur, 1901, 15 Année No. 10. (5) Sacus: Fortschritte der Medicin, 1902, 20, 425. (6) LevENE AND Srooxy: Jr. Med. Research, 1903, 10. (7) Haupern: Zeitschr. f. Immunit., Orig., 1911, 11. (8) Marras: Centralbl. f. Bakt. 1914, 1 Abt., 75, Orig. (9) Grtrzner: Pfliig. Arch., 1874., 8. (10) Hexrorn: Tr. Chicago Path. Soc., 1911, 8, no. 5. (11) Arxin: Jr. Infect. Dis., 1915, 16, no. 3. ON RED CELL GLOBULIN! C. B. BENNETT anp CARL L. A. SCHMIDT From the Department of Biochemistry and Pharmacology, University of California, Received for publication January 6, 1919 A number of workers (1) have attempted to isolate from red cells the antigen which gives rise to a specific lysin and an agglu- tinin when red cells are repeatedly injected into an animal. These attempts have confined themselves chiefly to three con- stituents of the red cell: laked blood, the insoluble stroma and extracts by various lipoid solvents. On this subject Ford and Halsey (2) report that injection of either red cells, insoluble stroma or the water-soluble portion of red cells obtained by lak- ing will give rise to a lysin and an agglutinin specific for the cell af the particular species used, these phenomena being insepara- bly connected. With the exception of hemoglobin, which is probably non-antigenic, no attempt has been made to separate from red cells a definite chemical entity which on injection will give rise to a lysin for the red cell. It may be conceived that the antigen concerned in the produc- tion of hemolysis may be one of the following possibilities: (a) the protein complex which probably exists in red cells analogous to the complex which Hardy (3) assumes exists in blood serum and to which Robertson (4) has suggested the specificity of ani- ‘mal tissues and fluids may be due; (b) a single protein constituent of the red cell which may be separated in a pure state; (¢) protein constituent occurring only in small quantities in the red cell in an analogous manner to the association of immune bodies with a particular fraction of blood serum and which cannot easily be separated; (d) lipoid constituent; (e) hemoglobin. The work 1 Aided in part by a grant from the George Williams Hooper Foundation for Medical Research. 29 30 Cc. B. BENNETT AND CARL L. A. SCHMIDT of Bradley and Sansum (5) would indicate that hemoglobin is antigenic and specific for any particular species, but when it is recalled that no attempt was made adequately to purify the hemoglobin from the protein constituents of the red cells, these results cannot be accepted without serious question. More- over, Ford and Halsey (2) were unable to produce immune bodies on repeated injections of purified hemoglobin. That the antigen is of a lipoidal nature as found by Bang and Forssman (6) and others has also been seriously questioned (7). Our work, which concerns itself with the CO.-precipitable glob- ulin of red cells, was undertaken for a two-fold purpose: (1) to study the relation of this protein to the production of a lysin for red cells, (2) the relation of this globulin to the CO.-globulin from the blood serum of the same species. The work was started by one of us (Bennett) some time ago and has since then been entirely repeated and extended. This protein is mentioned by Kiihne (8) and is probably re- lated to the paraglobulin obtained from laked red cells by Wool- ridge (9) and Halliburton and Friend (10). For its preparation we proceeded as follows: Oxalated blood from the ox was centrifuged, the serum was re- moved and the red cells were washed with normal saline nine times, the cells being thrown to the bottom by centrifuging and the supernatant fluid being withdrawn each time. The red cells were then laked with two volumes of water, and centrifuged to re- move insoluble stroma and unhemolyzed corpuscles. This fluid was then brought to a dilution of ten parts of water to one of corpuscles and saturated with CO, until the globulin separated in large flakes. It was then repeatedly washed by decanting with. distilled water until practically no further color of hemoglobin was apparent in the wash water. The suspended globulin was then centrifuged off, and the supernatant fluid removed, and to the compact mass sufficient solid NaCl was added to bring it to physiological concentration. On centrifuging off the gross par- ticles, a semi solution of the globulin was obtained which how- ever contained at this concentration some finely divided undis- solved globulin. For the purposes of injection, phenol was ON RED CELL GLOBULIN oil added to make a concentration of 0.5 per cent. Although the globulin is probably white we were unable completely to separate traces of hemoglobin, even after repeated washing. The globulin is insoluble in distilled water, but it is readily soluble in physiolog- ical salt solution. It gives a positive biuret, xanthoproteic and Millon’s reaction. On boiling, coagulation occurs, and large flakes are precipitated. On standing for a long time in the pres- ence of phenol, denaturization apparently slowly occurs. For the preparation of CO.-globulin from ox serum we used the well- known method described by Quinan (11). For the production of immune bodies two normal rabbits were each given 2 cc. doses of the concentrated globulin preparation intravenously for three successive days and after a five-day in- terval this was repeated. No symptoms were shown by the in- jected animals. Eleven days after the last injection the rabbits were bled and sera free from traces of hemoglobin were obtained. After the inactivation of the immune sera at 57°C. for one half hour to remove alexin, fixation experiments were carried out, both cell and serum globulin being used as antigen. The fol- lowing amounts were used for the hemolytic system: antigen, 0.05 ec. of a 1:10 dilution (one-quarter of the minimal inhibiting dose) ; alexin 0.15 cc. of 10 per cent guinea pig sera (1.5 units); 0.2 cc. of 1:2000 rabbit vs. sheep red cells amboceptor; 0.2 ce. of a5 per cent suspension of washed sheep cells; salt solution was added to bring the volume to 1 cc. When cell globulin was used as antigen, positive fixations were obtained for animal (1) in a dosage of 0.1 cc. of a 1:250 serum dilution and for animal (2) in a dosage of 0.2 cc. of a similar dilution. When serum globulin was used as antigen no fixations were obtained in any dilution of serum which in the absence of antigen was not inhibitory. The non-identity of the two globulins is thus apparent. This is in agreement with the work reported by Thomsen (12) who found that guinea pigs sensitized with serum alone did not react ana- phylactically to the homologous red cells and vice versa. To demonstrate the presence of red cell lysin in the sera of the immunized animals a system essentially similar to the above was used. A 5 per cent suspension of washed ox cells was used in a 32 Cc. B. BENNETT AND CARL L. A. SCHMIDT dosage of 0.2 cc., 0.15 cc. of 10 per cent alexin, immune sera in varying dilutions, and salt solution to bring the volume to 1 ce. For the serum of animal (1) lysis occurred in a dosage of 0.1 ce. of a 1:1250 dilution and for animal (2) 0.3 ec. of 1:7250 serum dilution caused lysis of the red cells. On omitting alexin, agglu- tination of red cells occurred in a serum dilution of 1: 250. Since by using the CO.-globulin of red cells as antigen we have obtained, on injection, antibodies similar to those obtained on injection of the entire constituents of the red cell, we cannot as- sume that the protein complex of the red cell is necessary for the production of immune bodies, and excluding the lipoids and hemoglobin as factors we are left with two possibilities viz., the CO.-globulin or another substance which cannot easily be separated from it. A choice of these is not yet possible. Of interest in this connection is the report of Quinan (13) that the substance occurring in serum which causes lysis of red cells is neither the soluble nor insoluble globulin nor serum albumin. He regards the substance as being probably a ferment. SUMMARY 1. It was found that the sera of animals immunized with CO.- globulin from washed red cells contained immune bodies specific for this globulin and not for the CO.-globulin from the homolo- gous serum and substances which in the presence of alexin will cause lysis of the homologous red cells and in the absence of alexin, agglutination of these cells. 2. It cannot be definitely stated whether the CO.-globulin from red cells is the antigen concerned in the production of a lysin for the homologous red cell or a substance intimately asso- ciated with it. The possibilities are discussed. REFERENCES (1) von Duncern, E.: Muench. Med. Wochschr., 1889, 46, 449. Borpet, J.: Annal. Inst. Past., 1900, 14, 257-296. Notr, P.: Annal. Inst. Past., 1900, 14, 297-330. LEVENE, P. A.: Journ. Men. Res., 1904, 12, 191-194. Srewart, G. N.; Amer. Journ. Physiol., 1904, 11, 250-281. ON RED CELL GLOBULIN oo (2) Forp, W. W., Anp Hatsey, J. T.: Journ. Med. Res., 1904, 11, 403-425. (3) Harpy, W. B.: Journ. Physiol., 1905-1906, 33, 251-337. (4) Ropertson, T. B.: Univ. of California Pub. Physiol., 1911, 4, 25-30. (5) Brapuey, H. C., anp Sansum, W. D.: Journ. Biol. Chem., 1914, 18, 497-506. Lesuanc, A.: La Cellule, 1901, 18, 336-383. (6) Bana, J., AnD Forssman, J.: Centralbl. Bakt. Orig., 1905, 40, 151-152. Bana, J., AND ForssmAn, J.: Beitr. z. chem. Physiol. and Path., 1906, 8 238-275. Davutwitz, F., AND LANDSTEINER, K.: Beit. z. chem. Physiol. and Path., 1907, 9, 431-452. Taxkakl, K.: Beitr. z. chem. Physiol. and Path., 1908, 11, 274-287. Bana, I.: Ergeb. d. Physiol., 1909, 8, 463-523. LANDSTEINER, K.: Jahresber. Ergeb. Immunitiitsfrsch., 1910, 6, 209-226. (7) von DUNGERN, AND Coca: Muench. Med. Wochschr., 1907, 54, 2321. Ritcuig, J.. anD Miuter, J.: Journ. Path. and Bact., 1912, 17, 429-431. (8) Kitunz, W.: Physiologische Chemie, Leipzig, 1868, p. 193. (9) Woouripas, L.: Archiv. Anat. and Physiol. (Physiol. Abt.), 1881, p. 387-411. (10) Hatuipurton, W. D.: Journ. Physiol., 1888, 9, 229-286. HALLisurtToN, W. D., AND Frienp, W.M.: Journ. Physiol., 1889, 10, 532-549. (11) Qurnan, C.: Univ. of California Pub. Path., 1903, 1, 1-5. (12) THompson, O.: Zeit. f. Immunititsfrsch. and exp. Therap., 1909, 3, 539-557. (13) Quinan, C.: Beitr. z. chem. Physiol. and Path., 1903-04, 5, 95-109. THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 1 STUDIES IN OSMOTIC PRESSURE I. THE MECHANISM OF BORIC ACID HEMOLYSIS MITSUJI KOSAKAI From the Department of Bacteriology of Cornell University Medical College, New York City Received for publication March 22, 1919 Of the long list of the studies of the phenomena of the he- molysis almost all have been concerned with the investigation of the properties of the various hemolytic agents, very few having pursued the question of the actual force that brings about the destruction of the corpuscles. Indeed, this ultimate force is known in but one of the many forms of the process of hemolysis that resulting from hypotonicity of the medium of suspension; that is, osmotic pressure. In the course of some experiments upon the preservation of blood corpuscles with boric acid A. F. Coca made the following observations:! boric acid in dry form or in solution may be mixed with blood in a proportion of 13 per cent, or less, of the dry substance without causing hemolysis. If blood that has been in contact with boric acid for a time is suddenly mixed with one or more volumes of 0.9 per cent NaCL solution, im- mediate complete hemolysis occurs. If the blood that has been treated with boric acid is mixed with more concentrated solutions of sodium chloride no hemolysis takes place. As the easily controlled conditions leading to this phe- nomenon appeared to be favorable for the discovery of the ulti- mate mechanism of the hemolytic process, the present further study of it was undertaken. 1 Unpublished. 35 THE JOURNAL OF IMMUNOLOGY, VOL. Iv, 2 36 MITSUJI KOSAKAI TECHNIC The corpuscle suspension used throughout this investigation was prepared from oxalated sheep’s blood by washing the lat- ter three times in physiological (0.9 per cent) NaCL solution and diluting the corpuscular sediment with the same solution up to the original volume of the blood taken. The concentration of the solutions of all of the substances employed was determined by titration so far as methods of titration for this purpose were available. The glass tubes used for the hemolysis tests were of ordinary American glass and they measured 7 cm. in length by about 1 em. in diameter. The experiments were carried out at room temperature (20° to 24°C). The influence of ordinary changes of tempera- ture upon the results was quite negligible. The direct effect of boric acid upon the corpuscles was first examined and it was found that, in the stronger concentrations, this substance is capable, after a period of time, the length of which is influenced directly by its concentration, of produc- ing complete or nearly complete laking. The protocol of an experiment showing the direct hemolytic action of boric acid is presented in table 1. That this direct hemolytic effect is not due to the acid quality of the substance; that is, to the action of hydrogen ions, seems deducible from the fact that boric acid is not dissociated in solution and also from the absence of methemoglobin formation, which is always observed after hemolysis with mineral acids and dissociable organic acids. The preceding experiment does not exclude the possibility that, although the direct boric acid hemolysis depends on the concentration of the reagent in the original mixture, the phenom- enon is influenced, also, by the quantitative relation between the corpuscles and the boric acid. This question was investi- gated by preparing two mixtures of equal volume containing an identical amount of boric acid but different amounts of corpuscles. The boric acid concentration used was such that MECHANISM OF BORIC ACID HEMOLYSIS 37 the usual amount of the corpuscle suspension (0.05 cc., tube A) was not completely hemolysed. If the degree of hemolysis were affected by the quantitative relation between the corpuscles and the reagent, then less hemolysis would be expected in the mixtures containing the larger amounts of corpuscles (that is 0.1 ec., tube B). How- ever, the degree of hemolysis was identical in the two mixtures. On the other hand, the hemolytic effect was distinctly less in a mixture in which the quantitative relationship between cor- TABLE 1 Showing the direct hemolytic action of boric acid NUMBER OF TEST TUBE I II Ill IV Vv Boric acid solution (3.5 per cent), cc.......... (O83) 205) 029551 1-95),|- 2295 Physiological salt solution (cc.)............... 0.65 | 0.45 BlOOdIsuSpensiomy(CCs))\eeae la. ce ees eee 0.05 | 0.05 | 0.05 | 0.05 | 0.05 Degree of resulting hemolysis 0 0 0 0 c EZ METIAITULL LOS fle es eae lslancscie ¢ wih wn tie ween 0 0 0 c c Pils INO soe Ws to SEE Rae ee Eo 0 tr ale c c FA) INOUE eo ace one nee eee ee ene 0 tr ale c c In all of the tables the following abbreviations are used to indicate the de- gree of the resulting hemolysis: ec = complete, alc = almost complete, v.st. = very strong, st = strong, w = weak, v.w = very weak, tr = trace, 0 = no hemoly- sis. puscles and boric acid was the same as that of mixture B but in which the concentration of the reagent was only half as great as in the latter mixture. From the results of the first experiment it is seen, furthermore, that in the lesser concentrations (tubes 1 and 2) boric acid exerts little or no hemolytic action. These concentrations were, therefore, suitable for the demonstration of the phenome- non of hemolysis observed by Coca. For this purpose a con- stant amount of corpuscular suspension was mixed with boric acid in the strongest concentration that was not directly hemo- 38 MITSUJI KOSAKAI lytic and also in three lesser concentrations. After five minutes an equal volume (1 cc.) of 0.9 per cent NaCI solution was added to each mixture and the tubes were at once vigorously shaken. The results of this experiment are presented in table 2. It is seen that complete hemolysis occurred only in the mix- ture containing the stronger concentration of the boric acid; that is, where the relative change in the concentration of that sub- stance caused by the addition of the salt solution was greatest. Furthermore, a significant characteristic of the boric acid he- molysis? is seen in the results of this experiment, in the fact that the full hemolytic action elicited by the procedure is exerted TABLE 2 NUMBER OF TEST TUBE I II lil IV Boric acid solution (3.5 per cent) (cc.)...............|-0.5 | 0.4 | 0.3 | 0:2 Bhysiologicalisalt solution (Cc.)).).0.0 7s eee 0.45 | 0.55 | 0:65 | 0.75 Bloodgsuspension: (Ce:) 2.0 5. c2.232 a ee eee 0.05 | 0.05 | 0.05 | 0.05 After five minutes, 1 cc. of physiological salt solution is added into each tube IAG ONCE: codes ctecio: dock soho ee Dee eee @ Afters ten thours: bos. cis ose ee ee c tr Fig 0 0 Ww Ww Result { within a few seconds, not the least further hemolysis occurring in any of the tubes during the succeeding ten hours. Since the cause of the hemolysis appeared to depend upon the sudden change of the concentration of the boric acid in the medium of suspension of the corpuscles it seemed possible that the addition of more than an equal volume of salt solution would cause hemolysis, also, in the mixtures containing lesser concentrations of boric acid and this was found to be true. When to the mixtures in tubes 2 and 3 in table 2 several volumes of 0.9 per cent NaCI solution were added complete hemolysis took place in both instances. *In the further discussions in this paper the term ‘‘boric acid hemolysis”’ refers to the phenomenon of hemolysis observed by Coca, not to the direct hemolysis produced by the concentrated reagent. MECHANISM OF BORIC ACID HEMOLYSIS 39 In view of these latter results it could be anticipated that the addition to the mixture in tube 1 of less than an equal volume of salt solution could result in a correspondingly lesser degree of hemolysis. That this is true is shown in the protocol pre- sented in table 3. It is seen that when a half volume of salt solution was added to the mixture only weak hemolysis occurred, the addition of a three-tenths volume causing only a trace of hemolysis. Similarly, it was anticipated that if the mixtures in the tubes 2 and 3 of table 2 were centrifuged and the supernatant fluid decanted, hemolysis could then be produced by shaking the sediment in a relatively small volume of NaCI solution—one TABLE 3 NUMBER OF TEST TUBE I Il Ill Boric acia solution (3:5 per cent). (€C.) cc. eee sn curtccedavee Ora OLS WOLs Physiological salt solution) (cc.)\\...... 0.6 63s .00- ds tock d ante 0.45 | 0.45 | 0.45 SIA MENIS EUSTON (CG:)!. ssf desc dytdun's dks Lo been clcco chldaes « 0.05 | 0.05 | 0.05 After five minutes was added to I, 1.0 cc. of physiological salt solution; to II, 0.5 cc. of physiological salt solution; to III, 0.3 ec. of physiological salt solution. tr tr Ww Wi JANG (OIE oC eae RR mR Mei Jet ae VE Nira sa nmedla Bes ua ATPeTELENWNOUTS 2 «sic ate Gehh ee beech doe en Result { which, without centrifugation, would fail to cause laking. The results of such an experiment, which are given in table 4, show this to be the case. That the hemolysis produced by the addition of salt solution to blood that has been treated with boric acid is not due to a specific action of sodium chloride is demonstrated by the fact that hemolysis occurs as in all of these circumstances if, in- stead of the solution of sodium chloride isotonic solutions of other salts or of sugar are added to the blood corpuscle and boric acid mixtures. Indeed, the same phenomenon is observed upon the addition of sheep’s serum to those mixtures. The data in hand made it appear possible that the concen- tration of the boric acid rather than its absolute quantity in 40 MITSUJI KOSAKAI the mixtures determined its hemolytic effect. That this is true, was demonstrated by repeating the experiment referred to in table 4 but using a quantity of salt solution in preparing the original mixtures such that the total volume was always 5 ec. instead of 1 cc. After these mixtures had stood over night none of the corpuscular sediments obtained by centri- fugation and decantation was hemolysed when rapidly mixed with even a large volume of 0.9 per cent NaCI solution. TABLE 4 TUBE 1 TUBE II TUBE III Boric acid (3.5 per “™ CENit) eee re 0%3 Physiological salt Same as I Same as I SOLUtIOR sri see eo 0.65 Blood suspension... 0.05 After five minutes, added 2 cc. of physiological salt solution, shaken After five minutes, ten times 0.2 ce. of physi- ological salt solution rapidly added with intervals of Result = complete he- thirty seconds, each molysis time shaken Result = no hemolysis Then this test tube cen- After five minutes centri- trifuged, into sediment fuged, into sediment added 0.5 ec. of physio- added 0.5 ce. of physio-, logical salt solution, logical salt solution, shaken rapidly Result = no hemolysis shaken rapidly Result = complete hemol- ysis The experiments thus far had elicited. the following facts: 1. The boric acid hemolysis is exerted to its full extent, in whatever degree, within a few seconds, no further hemolysis thereafter taking place. 2. The boric acid hemolysis depends on the concentration, not on the absolute amount of the reagent with reference to a constant quantity of corpuscles. MECHANISM OF BORIC ACID HEMOLYSIS 41 3. The boric acid hemolysis occurs after a sudden lowering of the concentration of the reagent in the medium of suspension of the corpuscles. The only force that is known to be developed under the fore- going conditions is that of osmotic pressure, which, of course, must be assumed to act upon a limiting corpuscular cell membrane and the further evidence, which will be presented, leaves no reasonable doubt that this force is, in fact, the ultimate cause of the phenomenon of hemolysis that we are studying. The assumption that the boric acid hemolysis is an effect of osmotic pressure necessitates the further assumption that the reagent is able to permeate the corpuscular membrane. As is well known (1) the corpuscular membrane is permeable to many substances and impermeable to others. Since boric acid had not yet been studied in this respect, it was necessary to deter- mine the question by experiment. For this purpose the method of Hedin was employed as follows: To 10 cc. of washed sheep’s blood, made up to the original volume with physiological saline solution, were added 10 ce. of physiological saline solution containing 3.5 per cent of boric acid. After 10 minutes this mixture was centrifuged and the supernatant fluid was com- pared with a mixture of 10 cc. of physiological saline solution and 10 cc. of the boric acid-saline solution as to its osmotic concentration. If the boric acid was capable of entering the corpuscles and was present there in the same concentration as in the medium of suspension, then the osmotic concentration in the fluid me- dium was, of necessity, the same as that of the mixture without the corpuscles. On the other hand, if the corpuscular mem- brane was impermeable or incompletely permeable to the boric acid, then the osmotic concentration in the medium of sus- pension must have been greater than it was in the mixture without the corpuscles. The osmotic concentration in the two fluids was determined with the eryoscopic method, with the following result: 1. Supernatant fluid of corpuscle mixture............................. 1.120 2. Control mixture without corpuscles.......................+-....... 1.122 42 MITSUJI KOSAKAI The distribution of the boric acid in the suspension of cor- puscles is, thus, uniform throughout the corpuscles and the medium of suspension. GRADUAL ADDITION OF PHYSIOLOGICAL SALINE SOLUTION TO ““BORATED’’ CORPUSCLES If the force that is operative in the boric acid hemolysis is purely that of osmotic pressure and if the corpuscles are not directly injured by the mere contact with the reagent, it should be pos- sible, by the gradual addition of a volume of salt solution that, otherwise, is hemolytic, to bring about a slow diffusion of the intracorpuscular boric acid into the surrounding medium with- out the production of hemolysis. The difference between the intracorpuscular and the extracorpuscular concentration of the boric acid should, by this procedure, easily be kept below the point at which the assumed destructive osmotic pressure is developed. The protocol of the experiment that was conducted according to this plan is presented in table 4. The result of the experiment is in harmony with the ‘‘os- motic pressure’ theory of the boric acid hemolysis and it dem- onstrates that the hemolytic effect is not caused by a direct injury of the corpuscles by the reagent. THE INHIBITION OF THE BORIC ACID HEMOLYSIS WITH CONCEN- TRATED SOLUTIONS OF ELECTROLYTES AND NON-ELECTROLYTES The already cited observation of Coca, that the boric acid hemolysis can be inhibited if, in the second step of the pro- cedure concentrated NaCl solution is substituted for the 0.9 per cent NaCI solution, is also compatible with the osmotic pressure theory. This phenomenon was subjected to a quan- titative study; the protocol of the preliminary experiments is presented in table 5. It is seen that corpuscles which have been treated with boric acid in the manner indicated can be mixed suddenly with 1 ce. of a 1.4 per cent or more concentrated solutions of sodium chlorid MECHANISM OF BORIC ACID HEMOLYSIS 43 without undergoing the least degree of hemolysis. In concentra- tions of 1 per cent or less the corpuscles are completely laked. In terms of the osmotic pressure theory it could be assumed that certain concentrations of sodium chlorid are able to offer a counter osmotic pressure upon the external aspect of the corpuscular membrane which neutralizes that developed by the boric acid within the cell. It could be assumed, further- more, that those concentrations of sodium chlorid need not prevent the diffusion of the boric acid out of the corpuscles and experimental examination of this question, the protocol of which is presented in table 6, shows that such is the case. TABLE 5 Each tube contains: Boric acid solution (3.5 per cent), 0.3 cc.; physiological salt solution, 0.65 cc.; blood suspension 0.05 cc. After five minutes, all centrifuged, and 1.0 cc. of the solutions of sodium chloride added. NUMBER OF TEST TUBE Concentration of sodium chloride NOCTACCHIN ter Se yest rp cael sO. Oe OUP at Ie 2 oS) Waa eee 1G Result PNG EOMC OM Soe ines Mics wlite c Gn levestally aw tr 0 0) 0 After two hours........... c ChooleveSunl avy tr 0 0 0 The insensitiveness of the corpuscles, after the second centri- fugation, to sudden immersion in physiological saline solution indicates that the concentration of boric acid remaining in them was no longer great enough to develop a destructive osmotic pressure. Moreover, the result of this experiment furnishes clear evidence that the boric acid hemolysis is not due to any organic change in the corpuscles caused by a chemical action of the reagent. In order to avoid confusion resulting from discrepancies that may appear to exist, with respect to hemolytic effect, in the quantitative relations in the different protocols, it may be stated that the different specimens of sheep’s corpuscles have been found to be differently susceptible to the hemolytic influence of boric acid that we are considering; that is, some speci- 44 MITSUJI KOSAKAI mens of blood could be completely hemolysed after treatment with concentrations of boric acid, which, with other speci- mens of blood, induced only a partial hemolysis. This factor has sometimes interfered with a direct comparison of the re- sults in the different protocols, but it has, in no way, detracted from the conclusions of the study, which have been based on individual experiments. If the absence of hemolysis upon the immersion of borated corpuscles in a concentrated solution of sodium chlorid is due TABLE 6 Each tube contains: Boric acid solution (8.5 per cent), 0.3 cc.; physiological salt solution, 0.65 cc.; blood suspension, 0.05 cc. After five minutes all tubes were centrifuged and the supernatant fluid in each was decanted | ] NUMBER OF TEST TUBE I II Ii IV : Physiological | 1.5 per cent | 2.0 per cent | 2.5 per cent et Ns salt solution| NaCl NaCl NaCl a na aa L 1.0 ce. | 1.0 ce. 1.0 ce. 1.0 ce. Result=. ---beaaee ¢c 0 0 0 After five minutes tubes II, III and IV again centrifuged and the sediment in each was mixed suddenly Physiological | | salt solution. .| 3 ce. 3 ce. 3 ce. ——————. qo ewe — =I.0 Sere 2s Result.......... | 0 | 0 0 | ) to a counter osmotic pressure developed against the outer sur- face of the corpuscular membrane, then a similar effect must be producible with suitably concentrated solutions of other substances capable of inducing osmotic pressure. Further- more, if the theory under consideration is correct, it must be possible to demonstrate that the minimal non-hemolytic con- centrations of all such substances for corpuscles that have been treated with a certain concentration of boric acid actually exert the same osmotic pressure. In the succeeding experi- ments these requirements are fully satisfied and the concor- MECHANISM OF BORIC ACID HEMOLYSIS 45 dant results permit the definite conclusion that the force opera- tive wn the boric acid hemolysis is, in fact, that of osmotic pressure. In these experiments the treatment of the corpuscles was always carried out in a volume of 1 ce. of the different concentra- tions of the boric acid. The concentration of the boric acid is usually indicated in the tables by the amount of a 3.5 per cent solution of that reagent that was contained in the treating mixture. For example, where the amount of boric acid is given as 0.15, this means that in making that mixture 0.15 ce. of 3.5 per cent of boric acid dissolved in physiological saline solution were mixed with 0.8 ce. of physiological saline solution and to this mixture were added 0.05 ce. of blood suspension. In order to determine the minimal non-hemolytic concen- tration of the different substances, five or six identical mix- tures of blood and boric acid were prepared and after five minutes the mixtures were centrifuged and the supernatant fluid was completely removed with a capillary pipet. With the sedi- ment of each tube was rapidly mixed (by shaking) one cubic centimeter of the different concentrations of the substance under examination. Within a few minutes these final mix- tures were centrifiged and the degree of the resulting hemolysis was noted according to the degree in which the supernatant fluid was tinged with hemoglobin. The minimal non-hemolytic concentration was taken as the lowest with which no tinging of the supernatant fluid resulted. The relative osmotic pres- sure of the various concentrations was determined with the usual cryoscopic method, the results of these examinations being recorded in the tables under the customary designation. In a preliminary experiment the relation of the osmotic pres- sure of the treating mixture to that of the minimal inhibiting concentration of pure sodium chlorid and of boric acid dis- solved in physiological sodium chlorid was studied and, as the tabulated protocols ( tables XIII and XIIIa) show, it was found that the ‘‘minimal inhibiting pressure”’ was slightly, but con- sistently lower than the ‘‘treating pressure.’’ It is impossible to interpret this difference without exact information as to the balance of osmotic pressure between the normal corpuscular 46 MITSUJI KOSAKAI contents and the “‘isotonic”’ saline solution in which the boric acid, used in the experiment, had been dissolved. A part if not all of the difference represents variation due to experimental error. It is seen, furthermore, that for corpuscles that have been similarly treated with boric acid the minimal non-hemolytic TABLE 7 Showing the method of determining the minimal concentrations of a substance (NaCl) which are capable of preventing the boric acid hemolysis NUMBER OF GROUP Amount of boric acid solution (3.5 per cent) used for treatment.............. 0.15] 0.2/0.25) 0.3/0.35] 0.4) 0.45] 0.5 Degree of resulting hemolysis Solution of NaCl, per cent == 0.9 st | c LAO w ist | ale 1b 5 I 0) Mawes lest: | sale 1e2 OF 10 wistic eS ONO 0 wistie 1.4 0 On ON lews 1 stuiice 1he6% 0 |0 | wi ale 1.6 QO | tr’) st’ |*ale eg, 0 | v.w] st 1.8 0 0 V.W 1.9 ; 0 10 2.0 0 concentration of sodium chlorid and of boric acid are found, with the method employed, to be of practically identical “osmotic concentration.” In two further series of tests, the results of which are pre- sented in tables 9 and 10, the ‘‘osmotic concentration” of the minimal non-hemolytic concentrations of four other substances was determined for differently treated corpuscles. As different MECHANISM OF BORIC ACID HEMOLYSIS 47 TABLE 8 Relation of the concentration of boric acid used in treatment to the minimal inhibiting concentrations of sodium chloride and boric acid TREATING CONCENTRATION OF BORIC ACID INHIBITING CONCENTRATION Amount of a Boric acid in physiological 3.5 per cent NaCl It sol eeeetyg | Percent | |) A pain gueinl Per cent A Per cent A PS hala kOe TT tO TS Ry AT NEL qty 0.5 hp (45) TAA: 1.85 1.095 1.54 1.07 0.45 aio 1.075 1275) 1.05 1.40 1.00 0.4 14 1.00 1.62 0.97 176s) 0.95 0.35 1.225 0.94 1.5 0.91 1.05 0.88 0.3 1.05 0.88 1.4 0.85 0.945 0.84 0.25 0.875 0.835 13} 0.83 0.77 0.80 0.2 O77 0.78 1.18 0.74 0.595 0.75 0.15 0.525 0.73 1.08 0.69 0.42 0.67 TABLE 9 Relative osmotic concentration of the minimal non-hemolytic concentrations of sodium chlorid, barium chlorid and cane sugar AMOUNT OF 3.5 NON-HEMOLYTIC CONCENTRATION OF PER CENT ba Fase Spa NaCl BaCl, Cane sugar TREATING THE SET Ser been SEH Ne a FE SR Fld COSEUECUES Per cent A Per cent A Per cent A per LE ee woh Sate Dard ot RD an) 0.2 2 0.749 Sal 0.724 11.0 0.722 0.3 1.4 0.850 4.4 0.858 13.0 0.866 0.4 1.6 0.950 4.9 0.971 14.0 0.984 0.5 1.8 1.082 5.4 1.076 15.0 W125 TABLE 10 Relative osmotic concentration of the minimal non-hemolytic concentrations of ammonium chlorid and glycerin NON-HEMOLYTIC CONCENTRATION OF AMOUNT OF 3.5 PER CENT BORIC ACID USED FOR NH:;Cl Glycerin TREATING THE CORPUSCLES Per cent A Per cent A ce. 0:2 Ne? 0.801 2.9 0.794 0.4 ey 1.010 3.8 1.041 48 MITSUJI KOSAKAI specimens of corpuscles were used in the two series of tests, the results in the two tests are not concordant. Those of table 9 agree with those of table 8, both of these disagreeing with those of table 10. However, the results included within each table are concordant among themselves and they allow no reasonable doubt that the prevention of the boric acid hemolysis with concentrated solution is due to a counter osmotic mechanism acting upon the outer surface of the corpuscles. SUMMARY After blood corpuscles have been treated for a short time with certain concentrations of boric acid that are not directly injurious to those cells, the sudden immersion of the treated corpuscles in a physiological solution of sodium chloride causes their complete hemolysis. This “boric acid hemolysis’ does not occur if the addition of the physiological saline solution is made gradually or if the corpuscles are immersed, even suddenly, in more concentrated solutions of sodium chloride or of other non-hemolytic substances. That the destructive force responsible for this form of hemoly- sis is that of ‘‘osmotic pressure’ is shown by the fact that the minimal non-hemolytic concentrations of all of the substances examined were found to be of identical ‘‘osmotic concentration.”’ REFERENCE (1) Heprn: Grundziige der physikalischen Chemie, Wiesbaden, 1915, p. 22. STUDIES IN OSMOTIC PRESSURE Il, THE NATURE OF OSMOTIC PRESSURE MITSUJI KOSAKAI From the Department of Bacteriology of Cornell University Medical College, New York City Received for publication March 22, 1919 The laws governing the development of osmotic pressure and the effects produced by that agency upon animal and vegetable cells have been exactly determined by experimental study. The nature of the force of osmotic pressure, on the other hand,-. has, of necessity, been only surmised, because of the limitations hitherto surrounding the experimental study of that force, and the ideas concerning this question are based either on theoretical grounds or on the observations upon the effects of osmotic pressure. In the writings on the subject of osmotic pressure two con- ceptions of the nature of that force are found. The first conception is drawn from the well known demon- stration by van’t Hoff of the agreement between the laws governing gas pressure and those governing the development of osmotic pressure. According to this conception, osmotic pressure, corresponding with the pressure of the gases, is ex- erted by the molecules of the dissolved substance (solute), these being thought of as continually “‘bombarding” the surface of the limiting membrane. Thus, Hedin (1) writes ‘‘da ferner der Zucker durch die Membran nicht passieren kann, so tibt der Zucker gegen die Membran einen gewissen Druck aus. Dieser wird der Os- motische Druck der eingeschlossenen Losung genannt.”’ Wells (2) says “‘since osmotic pressure, exactly like gas pres- sure, is presumably produced by the bombarding of the walls of the container by particles in solution. 49 50 MITSUJI KOSAKAI Lewis (3) writes ‘“‘It seems reasonable to suppose, therefore, that when diffusion of a solute does occur in a given direction it is due to the osmotic pressure acting as the driving force. Of course we cannot speak of the osmotic pressure of the sol- vent, but simply of the solute, since the concentration of the solute corresponds to gas concentration.” Nernst (4) writes ‘‘it must happen, of course, that the sugar will exert a pressure on the partition, which opposes its endeavor to fill the whole solution.” The further exposition by this author leaves no doubt that he looks upon osmotic pressure as being directly exerted by the solute. The second conception of the nature of the force of osmotic pressure is expressed by van’t Hoff (5) as follows: In order clearly to realize the quantity referred to as osmotic pres- sure, imagine a vessel completely full of an aqueous solution of sugar, placed in water. If it be conceived that the solid walls of this vessel are permeable to water but impermeable to the dissolved sugar, then, owing to the attraction of the solution for water, water will enter the vessel up to a certain limit, thereby increasing the pressure on the walls of the vessel [inside]. Equilibrium then ensues owing to the pressure resisting further entry of the water. This pressure we have termed osmotic pressure. | This conception ignores the dissolved substances as directly exerting the force that we are studying and looks upon the latter as merely the pressure developed by the accumulation of the water which diffuses through the semipermeable membrane into the solution containing the greater concentration of mole- cules and dissociated ions. The view just stated‘is not with- out adherents among the investigators of the subject of osmosis but it has received almost no consideration in the published treatises. The preceding study of the boric acid hemolysis (6) had revealed conditions under which the development and effect of osmotic pressure can be exactly controlled and observed under varying quantitative relationships. THE NATURE OF OSMOTIC PRESSURE | In previous investigations Eisenberg (7) had observed the laking of the corpuscles with formalin and urea under circum- stances, which, so far as they were studied by him, coincide with those which we have found to control the phenomenon of boric acid hemolysis and we had observed a similar hemo- lytic action on the part of glycerin and ammonium chloride. As will be presently set forth, a comparative study of the quantitative and time-relationships that govern the hemolytic effect of some of these reagents, has revealed facts which can be explained only with the assumption that osmotic pressure is not a direct property of a solute but is developed indirectly, as a result of a process of diffusion, by the accumulation of water on one side of a semipermeable membrane. Eisenberg observed that corpuscles which had been in con- tact with formaldehyde in a concentration below that capable of ‘‘fixing” them were immediately laked on being suddenly immersed in isotonic saline solution. This hemolytic effect did not occur if the treated corpuscles were immersed in more concentrated salt solution. Eisenberg concluded that the formalin hemolysis is not a direct effect of the reagent. He considered the phenomenon as a ‘‘water hemolysis”’ but he adduced no experimental support for such assumption. Eisenberg’s observation of the urea hemolysis (in the lesser concentration of that substance in physiological saline solution) was confined to the mere fact that that effect is produced by a sudden immersion of the treated corpuscles in isotonic salt solution. Eisenberg excluded osmotic pressure as a cause of the urea hemolysis, because of the fact that with the lesser concentration of the reagent a longer contact was followed by a stronger hemolytic effect than a shorter contact. This fact, Eisenberg thought, pointed to a direct action of the urea. The first experiments of the present study were designed for the purpose of determining whether the hemolytic action of formaldehyde and of urea were, like that of boric acid, the result of osmotic pressure and five criteria were used together in arriving at a conclusion as to that question. ‘The first of these had already been applied by Eisenberg in his observation THE JOURNAL OF IMMUNOLOGY, VOL, IV, NO. 2 52 MITSUJI KOSAKAI that the hemolytic action of both substances was developed merely by a reduction in the concentration of the substance in the medium of suspension; the second criterion was the ab- sence of hemolysis when the treated corpuscles were suddenly immersed in concentrated salt solutions; the third criterion was the disappearance of the peculiar sensitiveness of the treated corpuscles to immersion in physiological salt solution after being washed in concentrated salt solution; the fourth criterion was the absence of hemolysis when to the treated corpuscles a hemolytic volume of physiological salt solution was added not all at once, but gradually; and the fifth criterion was the demonstration of the permeability of the corpuscular membrane to both substances with the cryoscopic method of. Hedin. In all of these five respects the corpuscles treated with either formaldehyde or urea! behaved exactly like those treated with boric acid. Hence the conclusion is warranted that the hemoly- sis produced by the former two substances is, like that of boric acid, a result of osmotic pressure. In order to make a comparative study of the osmotic hemoly- sis produced by the three selected reagents the concentrations were determined in which they all produce the same degree of hemolysis upon an arbitrarily selected, uniform diminution of those concentrations. It was found that if 0.05 ec. of cor- puscular suspension were treated for ten minutes with 0.4 ce of 3.5 per cent boric acid or 0.4 ec. of 4 per cent formaldehyde or 0.8 ec. of 10 per cent urea, the total volume in each case being 1 cc., the sudden addition of 1 ec. of physiological salt solution would produce very strong hemolysis, while the addition of 2 cc. of that solution would cause complete hemolysis in all of the mixtures. Under the condition of these comparative tests the assump- tion is justified that in all three instances, where the same de- gree of hemolysis was produced, the corpuscles were being subjected to the same degree of osmotic pressure. ‘For a generous supply of urea of highest purity the author is indebted to Dr. William J. Gies of the College of Physicians and Surgeons in New York City. THE NATURE OF OSMOTIC PRESSURE 53 On the basis of these tests, the same relative concentrations were used in the subsequent comparative study. It was found that the osmotic coneentrations of the mix- tures containing 0.2 cc. of 3.5 per cent boric acid, or 0.2 ce. of 4 per cent formaldehyde or 0.4 cc. of 10 per cent urea were respectively A = 0.780, 1.091 and 1.91; and that the corre- sponding changes in concentration in the comparative hemolytic experiment were as follows: CONCENTRATION OF THE SUBSTANCE IN THE CONCENTRATION OF THE MIXTURE AFTER AD- SUBSTANCE IN THE DITION OF THE MINI- TREATING MIXTURE MAL HEMOLYTIC AMOUNT OF PHYSIO- LOGICAL SALT SOLUTION per cent per cent Borie acid, 3:5 per cent, 0.2 ce..:......-.. 0.700 0.378 Formaldehyde, 4 per cent, 0.2 cc.......... 0.800 0.267 (WreselOjpericent. O.4xcey. 2 ie est scl 4.000 Ve} It is evident that if the destructive osmotic effect that we are studying is exerted directly by the molecules of the dif- ferent substances, it should be expected that the solutions of the three substances which produce the same hemolytic effect would be found, by the eryoscopic method, to be of the same osmotic concentration; furthermore, it should be expected that a constant relation would be found between the concentration of each substance with which the corpuscles were treated and that of the respective mixture, after the minimal hemolytic quantity of physiological saline solution had been added. However, in the experiments that were undertaken to deter- mine ‘this question, neither of these two requirements was satisfied. : It is seen that although the osmotic concentration of the mixture of boric acid used in the treatment of the corpuscles is considerably less than that of the other two substances the same hemolytic effect was produced by a change in the con- centration of the reagent which was much less in the case of boric acid than it was in the other two substances; that is, exactly the reverse of what would have been expected if -the osmotic force is exerted directly by the molecules of the reagent. 54 MITSUJI KOSAKAI A comparison of the results obtained with formaldehyde and with urea shows a close correspondence in the ratios be- tween the concentration used for treating the corpuscles and that resulting upon the addition of the minimal hemolytic amount of physiological saline solution. In view of the fact that the osmotic concentration of the mixture of formaldehyde used in treating the corpuscles was only about half as great as that of the treating mixture of urea a correspondingly greater lowering of the concentration of the former reagent should have TABLE 1 Determination of the ratio of the final concentration to the original concentration of the reagent in boric acid hemolysis NUMBER OF TEST TUBE I II iil IV Borie:acid solution (3:5 percent). (ce.)..25-2..-4-.-2--| OL SO 0 oe Physiological salt solution (ce:)...-. =. oe2sseeeeeo ee 0.25 | 0.35 | 0.45 | 0.55 Bloodssuspensiont (ce). 5. 42.2. e eee Eee 0.05 | 0.05 | 0.05 | 0.05 Wirole ‘volume’ (¢e2)2 23022 ono. erin oe 4 ete OF47 1025.5) |20:6a OZ Percentage concentration of boric acid in the treat- BN PII XGUPCS seers sic tee ax hots cone Ge ae Ee 0.875} 0.7 | 0.583) 0.5 Minimal hemolytic amount of physiological salt solu- tion, added after five minutes contact without previous centrifugation (ce:)....0. ...0..<.."-25e8 oe 0.4 | 0.65|0.9 | 1.4 Final percentage concentration of boric acid......... 0.438) 0.304| 0.233) 0.167 been thought necessary to the production of an identical hemolytic effect in the two mixtures. This latter contention is verified by an examination of the protocols of three series of tests that were carried out with diminishing concentrations of boric acid, formaldehyde and urea respectively. These protocols are presented in tables L, 2 and 3. In all of these three experiments it is evident that, as the concentration of the reagent that was used for treating the corpuscles diminishes, the ratio between that concentration and the final concentration rapidly increases. THE NATURE OF OSMOTIC PRESSURE 55 TABLE 2 Determination of the ratio of the final concentration to the original concentration of the reagent in formaldehyde hemolysis NUMBER OF TEST TUBE I Formaldehyde solution (4 per cent) (cc.)............ 0.1 Ehysiological salt solution (cc.)......5.5..6....50005 0.25 loodssuspensiony (CCs)s5..2-4.4snssccess cows ne et nen 0.05 Metolewouinges (CC: 2st .see8 - oe ska ols ke 0.4 Percentage concentration of formaldehyde in the Pesca bITee ATX GUE OS eee. 's vrethalels suse 4 See deals Games ate n\e(0) 0.666) 0.572 Minimal hemolytic amount of physiological salt so- lution, added after five minutes contact without NECVIOMSUCeEM TIN gatIOMW (Ce). 1.0. ses nee ee oLES ON ole Ow ee een ee Final percentage concentration of formaldehyde..... 0.286] 0.19 | 0.143) 0.077 TABLE 3 Determination ne the ratio of the final concentration to the original concentration of the reagent in urea hemolysis NUMBER OF TEST TUBE I II Iil IV lWrearsolution, (Oper cent) (Ce.)..---cc _—_ S — —————— — Plasmas s.G et eee —— = 41 2} 43 | 44] +4] M.H Senn jc: oe eee 2 _ +1 z +? +4 +4 M.H * The respective tests were conducted after the same technic previously de- scribed. 98 HEMOLYTIC COMPLEMENT AND ANTIBODIES 99 The results are indicated in table 14 and they have shown that the plasma usually contained appreciably more hemolysin and hemagglutinin than the corresponding sera. TABLE 14 Titration of immune antisheep hemolysin in rabbit plasmas and sera | DILUTED 1:5000 NUMBER 0.05ee., 0.1 ce. 0.15 ce. 0.2 cc. | 0.25 cc. |0.3 ec. 0.35ec.|0.4 ec.0.45¢ec.0.5 ee. [SS eS eee | | Plasma |SH| MH |V.MH| CH | CH | CH|CH CH C.Hl CH Serum |S. | M.H | MH |V.M.A/V.M.H/ CHI CHI CH CH C.H Plasma |M.H) C.H | CH | CH | GH | CH ae C.H on V.H Serum |S.H|VMH| C.H | CH | CH | CH CH) CH) CH CH | | — e ES io2) oO —— TABLE 15 Titration of sheep hemagglutinin in the plasmas and sera of immunized rabbits FINAL DILUTIONS NUMBER SUBSTANCE 1:20 1:40 1: 80 1: 160 1: 320 1: 640 1:1280 14 Plasma 4 ate ote i oP = = Serum + == a = = = = 80 Plasma 4 oie a = = = = Serum a5 oF = = = = — CONCERNING THE PRESENCE OF IMMUNE TYPHOID AGGLUTININ AND COMPLEMENT FIXING ANTIBODY IN RABBIT PLASMAS AND SERA Rabbits were immunized with repeated intravenous injections of heated typhoid vaccine and bled from the carotid artery, the plasmas and sera being titrated for immune agglutinin and com- plement fixing antibody with heated and unheated plasmas and sera and with the usual controls. The results shown in tables 16 and 17 demonstrate that the plasmas always contained an equal or sometimes greater amount of agglutinin and always an equal amount of complement fix- ing antibody. THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 3 100 SUSUMU WATANABE TABLE 16 Macroscopic agglutination tests of the plasmas and sera of the immunized rabbits against B. typhosus FINAL DILUTIONS NUMBER SUBSTANCE eaflelelelele Pele | elsifeoe 2/2 /2/8/2/2/2/2/2/2/2 |= 84 { Plasma +l+]/4/4+i})—-|]-[- Serum SS Sm ee 2 { Plasma Selim (Mars apelin Mele be se se Serum “Pl seal Gabea) ect Stee Pate | aa | el) St 3 { Plasma | + | 4) ee | eee ee) ee Serum ms a Wisz ae Vc pce ace oe bs i ga a a Re TABLE 17 Thyroid complement fixation tests with the plasmas and sera of immunized rabbits DOSES OF UNDILUTED PLASMA AND SERUM NUMBER SUBSTANCE 0.0002 ce 0.0004 ce. 0.0005 ec. 0.0008 ec. Plasma eS lees Serum ee Plasma | {4 14 2/4 3]4 4/4 4) 4 4) 4 a) 4 a) Serum ae] se ffeUp 4 274 4} 4 alt a} 4 a} 4 a} 4 a} Heated 84 Unheated [ Serum —| — ary i) yy se Ber a 2 Plasma | —| + 4i43) [44/4 4/44 = Serum —| + Selietes +44 444) re Unheated Plasma | —| — + |+2 +4p444 ws Serum —-|- = 4+? + 4/4 4/4 4 pe Plasma =e + |42 44/4 4)4 4 at Serum = || = += |42 44/4 4/44 es Heated f Plasma | —| — Senate eel ctaial aia — Unheated { HEMOLYTIC COMPLEMENT AND ANTIBODIES SUMMARY 101 A brief summary of the results of this study presented in tables 1 to 17 is given in table 18 which shows in parallel columns the comparative results of tests for hemolytic complement, various natural and immune antibodies in human and rabbit plasmas In the majority of instances the various plasmas con- tained an equal or greater amount of hemolytic complement or antibodies than were present in the corresponding sera, and only exceptionally lesser amounts. and sera. TABLE 18 Summary of tables showing comparative results with plasmas and sera SOURCE 18 Human 18 Human 18 Human 18 Human 18 Human Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit SUBSTANCE Hemolytic complement Total hemolytic activity for sheep cells Thermostabile antisheep hemolysin Syphilis antibody Natural typhoid agglutinin Hemolytic complement Total hemolytic activity for sheep cells Sheep hemagglutinin Natural typhoid agglutinin Non specific typhoid complement fix- ation Non specific Wassermann reaction Immune antisheep hemolysin Immune antisheep hemolysin Immune sheep hemagglutinin Immune sheep hemagglutinin Immune typhoid agglutinin Immune typhoid agglutinin Immune typhoid agglutinin Specific typhoid complement fixation Specific typhoid complement fixation Specific typhoid complement fixation COMPARATIVE RESULTS Plasma Average unit 0.0312 ce. Equal or greater in 90 per cent Equal in about 78 per cent Positive in 12 Equal Unit 0.1 ce. 0.3 ec. cells 1:4 None (+) 0.02 ee. (+) 0.04 ce. 1: 25000 1: 50000 1: 320 1:80 1: 3200 1: 9000 1: 7200 (+) 0.002 ce. (+) 0.0008 ee. (+) 0.0008 ce. Serum Average unit 0.0412 cc. Weaker in 10 per cent Greater in 22 per cent Positive in 12 Equal Unit 0.15 ce. 0.3 ce. cells 1G None (+) 0.02 ce. (+) 0.04 ce. 1: 16000 1: 33000 1: 80 1: 40 1: 1600 1: 9000 1: 7000 (+) 0.002 ce. (+) 0.0008 ce. (+) 0.0008 ce. 102 SUSUMU WATANABE CONCLUSIONS 1. A method for the collection of plasma is described which has yielded uniform success and simplified this difficult technical problem. This technic involves the use of dried sodium oxalate after a method devised by Meeker, and a paraffined tube. 2. Sodium oxalate in the proportion of 0.001 gram per cubic centimeter of blood prevents coagulation and does not exert any injurious influence upon hemolytic complement or anti- bodies; in amounts of 0.004 gram per cubic centimeter of blood sodium oxalate may prove to be anticomplementary. 3. The oxalated plasmas of normal and syphilitic persons and normal and immunized rabbits contains hemolytic complement in the same or somewhat greater amounts than the correspond- ing sera. 4. The oxalated plasmas of syphilitic persons contain the same amounts of Wassermann antibody as the corresponding sera. 5. The oxalated plasmas of persons and rabbits contain the same or occasionally slightly greater amounts of such natural antibodies as antisheep hemolysin and typhoid agglutinin as the corresponding sera. 6. The oxalated plasmas of normal and immunized rabbits contained as much specific and non-specific complement fixing substances and specific bacterial and hemagglutinins as the cor- responding sera. 7. The general conclusion of this investigation is that hemo- lytic complement and natural and immune antibodies exist free and preformed in the circulating plasma of the blood. I beg to express my appreciation to Prof. Kolmer for direc- tions and aid in conducting this work. HEMOLYTIC COMPLEMENT AND ANTIBODIES 103 REFERENCES (1) Gencou, O.: Contribution a |’ étude de l’origine de l’alexine des sérums nor- maux. Ann. del’Inst. Past., 1901, 15, 232. (2) Herman, M.: Sur lorigine des alexines. Bull. de l’Acad. Roy. de Méd., 1904, 18, 137. (3) Gurp, F. B.: Variation in the complement content of serum and plasma. Journ. of Infect. Dis., 1912, 11, 2, 225. (4) Domeny, P.: Stammt die witksame Substanz der himolytischen Blutfliis- sigkeiten aus den mononucleiren Leukocyten? Wien. klin. Wochen- schr., 1902, 40, 105. (5) Sweet, J. E.: A study of an hemolytic complement found in the serum of the rabbit. Centralbl. f. Bakt., Orig., 1903, 33, 208. (6) Hewiert, A. W.: Ueber die Einwirkung des Peptonblutes auf Himolyse und Baktericide. Bemerkungen iiber die Gerinnung des Blutes. Arch. f. experim. Path. u. Pharm., 1903, 49, 307. (7) Lowir, M., and Scuwarz, K.: Ueber Baktericidie und Agglutination im Normalblute. Zeitschr. f. Heilkunde, 2903, 24, 205 and 301. (8) Lamporte, U.: Contribution 4 l’étude de l’origine de l’alexine bactéricide. Centralbl. f. Bakt., Orig., 1903, 34, 453. (9) Horssu1, H.: Das Verhalten der Streptokokken gegeniiber Plasma und Serum und ihre Umziichtung. Centralbl. f. Bkt., 1903, Orig., 55, 135. (10) Von Duncern: Cited from Zinsser in Infection and Resistance, 1914, 1st edition, 171. (11) Jovan, C. anv Stavs, A.: Presence de l’alexine hémolytique et bactéricide dans le plasma des oiseaux. Compt. rend. Acad. Scienc., 1910, 151, 6, 452. (12) Appis, T.: The bactericidal and hemolytic powers of paraffin plasma and serum. Journ. of Infect. Dis., 1912, 10, 2, 200. (13) FAuLoIse: Sur l’existence de l’alexine hemoletiace dans le plasme sanguin. Bull. de l’Acad. de Science. de Belg., 1903, 6, 521. (14) ScunerpEerR, R.: Ueber die Priiexistenz des Alexins in zirkulierenden Blut. Arch. f. Hyg., 1908, 65, 305. (15) Dreyer, G., AND WALKER, E.: On the difference in content of immune sub- stances in blood serum and plasma. Brit. Med. Journ., 1909, 1, 151. On the difference in content of agglutinins in blood serum and plasma. Journ. of Path. and Bact., 1909, 14, 39. (16) Cowrz, D. M.: A method for obtaining human plasma free from chemical action. Its effect on phagocytosis. Journ. med. research., 1909, 21, 2, 328. (17) Orant, M.: On the acceleration of phagocytosis in the citrated blood and citrated blood-plasma. The Kitasato Arch. Experim. Med. 1918, 2, 147. (18) Gessarp, C.: Sur l’antityrosinase. Compt. rend. Soc. Biol., 1911, 35, 591. (19) Buazrot, L.: Toxicité des extraits d’oranes. Leur neutralisation in vitro par le plasma oxalaté chauffé 4 56 degrés et recalcifié. Nécessité des sels de chaux. Rédle de la thrombocyme. Compt. rend. Soe. Biol., 1911, 34, 534. 104 SUSUMU WATANABE (20) Koitmer, J. A., AND CASSELMANN, A. J.: Concerning the Wassermann reac- tion with normal rabbit serum. Journ. med. Research, 1913, 28, 369. (21) Koumer, J. A., anp Trist, M. E.: Non-specific complement fixation by nor- mal rabbit serum. Journ. Infect. Dis., 1916, 18, 20. (22) Scumipt, A.: Weitere Beitrige zur Blutlehre. Wiesbaden, 1895, 118. (23) Kotmer, J. A.: Infection, Immunity and Specific Therapy, 1917, 2nd edition, 467. A METHOD FOR THE PRODUCTION OF A HOMO- GENEOUS SUSPENSION OF BACILLUS ANTHRACIS TO BE USED IN AGGLUTINATION REACTIONS ARLYLE NOBLE From the Research Laboratory of Parke, Davis and Company, Detroit, Michigan Received for publication April 10, 1919. Because of the fact that an accurate and reliable method for the standardization of antianthrax serum has not yet been de- veloped, while the necessity for such a method is very apparent, experiments were undertaken to make use of the agglutination _and complement fixation reactions. So far the results obtained from the animal protection tests have not proved uniform, and therefore not satisfactory, mainly because the animals on which the serum is being tested are so highly susceptible to the dis- ease, and because of the difficulty in standardizing the test culture and maintaining a standard dose, due to variations in the virulence of different strains of B. anthracis. Of the serum reactions, the complement fixation test has thus far not proven to be of value, as a stable antigen has not been produced, though further experiments are being made along this line. With regard to agglutination, the results as reported in the literature appear quite conflicting. Sobernheim in reviewing the subject states that the agglutinating action of serum on anthrax bacilli may be observed both microscopically and macro- scopically, though the immobility of the bacilli and their incli- nation to arrange themselves in clumps make the judgment of agglutination difficult; and that with many sera one always obtains distinct agglutination in strong dilution, while, on the other hand, it is often lacking in even high grade anthrax serum. He gives the statements of various writers, briefly, as follows: Sawtschenko found that horse serum agglutinated in every case, 105 106 ARLYLE NOBLE irrespective of whether it came from normal or preventatively inoculated ‘animals, while dog serum of both catagories never evidenced agglutinating power. Contrary to the report made by Sobernheim, that the specifically agglutinating power is usually lacking in the anthrax serum, Cavine states that a series of various anthrax sera proved active, in dilutions of 1—50,000 to 150,C00 even 1-500,000. Gottstein obtained completely nega- tive results in a retest of these experiments, several high grade sera from horses, cattle and sheep showing no agglutination power. Sobernheim states that the remarkable fact can be determined that a serum agglutinates, for example, the bacilli of virulent anthrax and of Pasteur’s Vaccine I, but not of Vaccine II, while another perhaps has no influence on the virulent anthrax and Vaccine II, but gives a distinct agglutination with cultures of Vaccine I. From these conflicting statements Sobernheim draws this conclusion: Although the question of agglutination for anthrax serum ‘requires further explanation, we may say, that a parallelism exists between the agglutinating and immunizing power of the serum in no ease, and the presence or absence of agglutinating action has absolutely no connection with the degree of immunity in the serum producing animal. It seems to the writer that the problem is largely, if not en- tirely, a problem of the suspension. Because of the nature of the organism, growing as it does in long chains and producing spores, it does not easily lend itself to agglutination experi- ments. However, a homogenous suspension of B. anthracis has been prepared and an increase of agglutinins demonstrated in sera from horses treated with vaccines of B. anthracis, as against sera from normal horses. The suspension for the agglutination tests was prepared as follows: The cultures employed were four strains of B. an- thracis furnished by the United States Bureau of Animal In- dustry, and designated by them “Davis,” ‘‘Chestertown,” ““N. H.,” and ‘‘6071.”’ These four strains were transplanted HOMOGENEOUS SUSPENSION OF B. ANTHRACIS 107 daily for ten days on plain agar and incubated at 42.5°C., until a sporeless and very vigorous growth was obtained. Each strain was then planted on plain agar in quart whiskey flasks and incubated for twelve hours at 42.5°C. The growths were washed off in physiologic salt solution containing 0.5 per cent formalin (about 100 cc. to a flask). The suspensions were shaken in a mechanical shaker for forty-eight hours. After standing for several days and being tested for sterility, equal parts of each suspension were mixed in a cylinder; shaken for twenty-four hours; and allowed to stand over night. The larger clumps settle out, leaving a homogeneous suspension above. This upper portion was poured off and filtered several times through four thicknesses of sterile cheesecloth. The sus- pension was then diluted with physiologic salt solution plus 0.5 per cent formalin to a density corresponding to a suspension of B. typhosus containing 2000 million bacteria per cubic centi- meter. A suspension of 5. anthracis so prepared is perfectly homogeneous, stands up for at least forty-eight hours at 37°C. and shows no spontaneous agglutination. The sera used were from thirteen horses which were treated first with vaccines of attenuated cultures and then with in- creasing doses of virulent B. anthracis. The strains were the same as used in the preparation of the suspensions. Also, sera from seven normal, untreated horses were tested. The agglutination tests. All the agglutination tests were macroscopic. In carrying out the tests, the serum dilutions were made in test-tubes with physiologic salt solution. The dilutions were never started with less than 1 ec. of undiluted serum and the volume of each dilution was always more than sufficient for the test. Special pipets, graduated to 0.5 and 1 ce. were employed throughout and a different pipet was used for each dilution. All glass-ware used in connection with the tests was clean and sterile. In the test, each small agglutination tube contained 0.5 cc. of suspension plus 0.5 cc. of diluted serum, with a control tube containing 0.5 cc. suspension plus 0.5 cc. salt solution. The tests were incubated at 37°C. for twenty-four hours. 108 ARLYLE NOBLE TABLE 1 RESULTS, JANUARY 17, 1919 Suspension — B. anthracis Sera — SERUM DILUTION Anti Nosmal ntianthrax, Horse horse 901 1057 1051 1050 1049 1047 | = 1045 1-10 Seg oR i ee Lis Urs |Paumacoe st ae ee bie |) ea a> 1-20 Semel Gees Rorccais apo |ierenonseosae | scan 140 Seat ai en ee irae Wtmace rl eae. limes ae || ans >> 1-80 ae | se clesteallstestecty ah tactics sts) ect 1-200 a a ar a SS er aU lee P| 1400 = Set We tegleahe yPiskestv che! atc sice | ee ats ate acti 1-800 = Seabee | eee (ck eo 1-1600 — “=F | sea) eR Pies | see ae 1-2000 = Bd a a (cles ol i se (i ear ee 1-3200 *E-5 | “Baka [eae ata ++) | Sera 1-6400 aad eg aieahi sf cheats +t 1-10000 ++ | + = + ts ar 1-20000 af = = = fe = 140000 - - - - — - Control - - — — - _ - +++ represents complete agglutination; ++ partial; + slight agglutination, but still with positive clumping. The results of the agglutination reactions with antianthrax sera and normal horse sera are given in tables 1 and 2. Each antiserum has been tested more than once and different bleedings fom the same horse have been tested with practi- cally no variation in the agglutination titer. The sera from five normal horses, in addition to the two given in the tables, gave agglutination titers of from 1 in 80 to 1 in 200. SUMMARY A satisfactory suspension of B. anthracis, for agglutination reactions, has been prepared by the described method. In order to be assured of a homogeneous suspension certain points must be observed. The cultures must be sporeless and must contain vigorous growths free from old organisms. The growths HOMOGENEOUS SUSPENSION OF B. ANTHRACIS 109 TABLE 2 eee RESULTS, JANUARY 18, 1919 a a ee Suspension — B. anthracis SERUM Sera — DILUTION Nosual Antianthrax, Horse horse 1665 1042 1027 1025 957 956 955 953 1-10 +++) +++] +++ ) +++] 44+ | 444+) +44] 444 1-20 +++] +++) t+4++] +++) 444) 444) 444/444 1-40 ++ | +44) t+4++] +++) +44) +44 | +44) 444 1-80 + | +++) +++) 444+] +++) 444 | 444+] 444 1-200 — | +++) +++] +++) +++] 444 | +44] +44 1400 — | +t+ |] +++) +++] 44+] +44] 444+ | +44 1-800 — | +++] 4+4+4 ] +++] +44 | 444) +44) +44 1-1600 — | t4++) +++) +++] 444+] 444) +44) 444 1-2000 — | +++] +++] 44+) 444) 444) 444+] 444 1-3200 +++] +++) +4 | +++] 444+) 444] +++ 1-6400 ++ | ++ | ++ | +4 ] + | 44 | 444 1-10000 ~ + + — | ++ | ++ 1-20000 ~ ~ - ~ - + + 140000 - -~ - - - - ~ Control - _ = = = = = = for the suspensions must be young, not more than eighteen hours old. The suspension must be thoroughly shaken; the larger clumps allowed to settle; and then carefully strained. Agglutinins have been demonstrated in the serum from horses hyper-immunized with B. anthracis. The antianthrax sera from thirteen horses have given agglutination titers of from 1 in 6400 to 1 in 20,000, as against titers of from 1 in 80 to 1 in 200 in normal horses. The agglutination tests show that certain antibodies have been produced in horses treated with B. anihracis and, in the absence of a satisfactory animal protection test or method of complement fixation the agglutination test may be used as a method for standardizing antianthrax serum. REFERENCE SoperNHEIM: Handbuch der Pathogenen Mikroorganismen, Kolle and Wasser- mann, 1904, Bd. 4, p. 800. ON THE NATURE OF ECLAMPSIA ISEI OBATA From the Forensic Institute of the Imperial University of Tokio Received for publication April 11, 1919 As is well known Dold (1) discovered, in the salt-solution extract of viscera, a poisonous property, which can be neutralized by blood serum. This so-called ‘Organgift”’ still remains a problem attracting the attention of observers. As to the nature and properties of the poison we have a number of other investigations (Dold and Ogata (2), Ascoli (3), Isar and Patane (4), Ichikawa (5), Aronson (6), Dold and Kodama (7), Yoshi- mura,! Goto! and Ishikawa),! the results of which diverge from one another. It would seem probable that the toxin above mentioned is contained also in the placenta, which may be considered as one of the viscera. The placenta is, in fact, a viscus which is present in the female sex only during a certain period of life and which is often placed anatomically in the category of tumor; it extends its villi into the maternal blood in which they are bathed so long as they exist, thus standing in a relation to the maternal body that is quite peculiar to itself. This relation- ship suggests the possibility of a resulting peculiar pathological process; it also provides an excellent material for study because of the ultimate natural separation of the organ from its site. It seemed possible, in view of the peculiar relationship referred to, that if the placenta does contain the poisonous property or substance of Dold, these could, under some unusual cir- cumstances, be responsible for the pathology and the symp- tom-complex of eclampsia. 1 Jn Japanese only. 111 1b ee ISEI OBATA The present study was pursued from this point of view. The question whether extracts of the placenta are poisonous was first investigated. PREPARATION OF THE PLACENTAL EXTRACT As soon as the placenta was expelled the umbilical cord was cut off together with that portion of the placenta which sur- rounds its attachment to it. From the so prepared placenta as much blood as possible was expressed and the decidual tis- sue was also removed. The cotyledon, preferably that por- tion most deficient in large villi was taken by weight, cut into pieces and ground in a mortar, being mixed, at last, with 0.85 per cent salt solution in a proportion of 1 to 3 parts by weight. The mixture, after being stirred, was left half an hour at room temperature before being filtered through the habutai, a fine silk. The filtrate was centrifuged and the supernatant fluid, which was designated placental extract, was used for our ex- periment. The extract was opalescent and of a pale pink color and it contained no solid particles in suspension. The placental extract employed in the present investigation was usually prepared from fresh placentae taken immediately after birth; but in some instances extracts were prepared from placentae which had been left in a refrigerator for a period that never exceeded seven hours. TOXIC SYMPTOMS IN THE EXPERIMENTAL ANIMALS The animals used in the present experiment were generally the Japanese dancing mice (a variety of Mus musculus), al- though rabbits were often employed. The following description applies to the mice, unless otherwise stated. The injections were made into the caudal veins. After the injection of a lethal dose of the extract there was an interval of ten to thirty seconds, rarely a minute before the animal became excited and fell at once in a brief clonic or, rarely, tonic convulsion, which was succeeded by a violent dyspnoea, coma and finally death within one to three minutes ON THE NATURE OF ECLAMPSIA P13 after the beginning of the convulsion. While the symptoms just described occurred in the majority of instances there were occasional exceptions in which the effects were more prolonged, death following sometimes after hours or days. Even in these cases, however, two symptoms were almost always noticeable; namely, dyspnoea and convulsions. COMPARISON, AS TO TOXICITY, OF THE NORMAL PLACENTA WITH THAT OF THE ECLAMPTIC INDIVIDUAL As the resistance exhibited by individual mice to the placental “poison”’ varied widely, it was necessary to make duplicate in- jections of each dose of the extract, in order to control this varying resistance. The lethal dose was taken as the minimal dose which killed both the individuals; in case a certain dose was able to kill only one of the two, the lethal dose is taken as an average of this dose and of that which was able to kill both animals. In table 1 are listed the cases in which the experimental animals died within ten hours after the injection; those animals which either recovered from the dyspnoea, convulsion and other symptoms caused by the injection, or died after sur- viving more than 10 hours are designated with the term ‘‘re- action.”” The animals designated ‘‘dead”’ were those which died in most cases within several minutes, though some sur- vived two or three hours and a few lived still longer, but not beyond ten hours. As shown by repeated experiments, no constant relation has been observed between the dose injected and the body weight of the mice employed in the experiments, which varied from 7 grams to 15 grams; hence the body weight is omitted in the descriptions as well as in the tables. The results presented in table 1 show that the lethal dose of the extract from normal placentae varies from 0.025 to 0.15 ec. and that from eclamptic placentae varies from 0.019 to 0.1 ec. It follows that the extracts of normal placentae can hardly be distinguished from those of eclampti¢ placentae with respect to toxic property. ISEI OBATA 114 — uol}OBaI — pop “99 900° 0 “09 GZT0'°0 PEEP uo0loead Beep —ad peop UOIJOBVOI uolpovaL pvop peop pvop peop uorpoBed u0ljoBod UOTPOVOL uo0lpovod Bee uoTovod peer prop UOTPOROI uolpOvod Peep UOTPOVOL uoTpOBoI pop uoTpovod PBep UOTPOBIA uolpOBoI peep uOTpOBOI Peep pop pop peop peop asod TVHLAT GOVUGAV Peep UNTUOM [BVULIOU JO BIUIIB[ | peop PONE is hs rer} | xy peop AL peep Ro sot) ee oe ee ee oy ee se Jip: safet| FoR f} os wooms | Pm Na pee ae VON rep fox pop Bary pvop ‘SL S er ‘0779 | SENUWINSA GaLOa’N! LOVULXA IVINGOVId AO ASOd T Wav -X@ GHL NI qaso Guam GQVINWOV Id GSOHM SNOSUMd JO SGANVN nn Ean 115 ON THE NATURE OF ECLAMPSIA Cu—,,) YSep @ poxreu O18 UOTJOVAT JNOYIIM Sosvd OY} O[TYM {SInoy yYys1d-A}10} Puokoq [BATAINS JUBOUT St ,,UOTJOVaI,, Aq ‘sInoy 4Y919-A41OJ GBI Y}vep oy} Soyrusis ,‘p ‘U peop,, fuoTJOfUT JoqjJe SINOY OATOM} UTYIIA S[VUNTUG po} UOUIIIOdx Jo YYVop SUBOUT ,,PBIP,, *370V GLE0°0 T0 T&0°0 6&0 0 Gc0'0 S00 6100 _ uolovod — pRePp — peep — uoloRad ESeP. 32) 2 uoljovod peer pgep uoTpORaL P8op. peep Been (eye uoljoRod uoTjovod Beep peep Even Ree [Se Beep [Mekyo Raee [PSY peep peep P¥ep peep pve 12) ¥ psop eal "ACL UdUIOM O19CUIBI9O JO BPUZIBT THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 3 me 0 G30 a ISEI OBATA ple) N i) | “OO TO asoa IWHLGT GADVUadaAv 116 _ — uoT}OVOI uolpovad pop Fe [278 prop UOTPOBI UOTPOVOA pvop| | pvop pwop J peop prop] | peop] —- peap| | UOT prop pvep} LVwoijovod prop prop uol}OvOd pop —- pvop -- pop Pee peep — pvop prop pvop a Beep. = peer "00 Z0 "00 €°0 Pep pBop pop pop seo peop | pvop pvop J pop peop peop \ pvop J prop Psep pop : prop prop "09 § 0 09g 0 agdLoarNnlI WOUAS AO AsOd Aouvusoid jo yyuOUT yyy Aouvusoaid jo yyuout yg NaMVL Nada GVH WOUGS HOINM LY GOoMmdd eee ua aes ir 6 “L's ea ‘OI ao | hee Rae - gpraes8-uoN ] | SONGAWNTYAd -xXq@d GH NI GaAOTd “Wa Guam vudas GASOHM SNOSuUdd JO SAWVN 3 6 WTaAVL Te ooo OOOO EEO EESSSSSSSSSSsSse < = a p¥eP. Ioqe] Jayye Avp yWS | “MW Se £0 = uoljovad pep re °'() = peop peop Ioqey 19338 Avp ys | “AW = pyep peep ; Loqe] doyye Avp yy | "OM CZ'0 —_ peep peop — peep IOgK] 10} je Aup WW9 ho N €°0 = peep ee prop IoOqe] tayjB Avp yy a ae = peop cel qe] 104J P 49 ss es os RESP, Ioqe] 1ajje Avp yyG | +7 ° és z'0 — prep : ia = peep prop Pee ote ; 4 ie = SA sai loge] 1o}}e Avp YG N ‘UY | uoeurom [eradiend Ayy]eopy S oe — peep Ioqe] 1o}jye Avp pig | -g° a ©z'0 oS peop peep]; : Pie E e) e°0 = = peop peop loqe] 1o}j8 Avp pug | “Wy : | = |geke) BBeP IOGB] LayJe ANoY prez] “ww ° fa £5 = eee Ley qe] Iopj TY Pres) “BIA LL a _ = " peep) — pwap > IOqv] 1a}jJev Noy yWOZ! “ye = SG ie peop 2 pvop Bl 19}j 4 4400}. “SM “OT = = pep} uworover pep ; aor Hy £0 a; UOTJoRad| UOTJOBAL peep SE tagged en N Ee [Behe Ioqey] tayye sanoy Zt] oN: A 0 on prop oacilt qe 1o4J Ga oN el a peep : Aoueusoid jo yyuOoUT Toy ae ~ a peut a Jo Ui} WANES Srl 23 L = a MOWjoNod ReeR Aouvusoeid Jo YyUoUT YIOT| ‘oy, “Yy : : Pp 0 —4 _ Re peop ee eee eels! “BpIAvia AY BO - | woloved pvoep| uonovor SousuBe re Je Mh Saeeue alee & es peop Aouvuderd yo yyuow wg | “N “My 920 mg peep gad u oO 0 ISEI OBATA peop peop peop pop RESP Beer Beep pop pvop pvep p¥ep peep Peep peep uoT}OBOd peep Peep peep Pvep peep peep peep peep Peep pwop pvep peop peop peop _pwop peep peep UOT}OBOL pep P¥ep peep peop peep agod TWHLOT ADVUGAV 118 GaLoarntl WOUGS FO ASOd Ioqe[ Joye skep fe youyye Surin Loge] 1ojywe ysne Loge] 1oyye Avp 46 Ioqe, toyye ABp yg yoryye Sunmna¢§ IOQ¥] oLO Jog, Ioqey] 194j8 Avp pag yow}ye suring IOQv] W1OJog Loqv] toyye Avp YT youyye Suring§ IOqB] dlOjog NGMVL NAW AVA WOUUS AHL HOIKM Ly GOIMdd panuyuog—% WTAVL SINGWIYN Gd “X@ GAHL NI oa at Gybretieyiené “UOTO o1ydure ony aqgAOTdWo aug mM vuas GSOHM SNOSUMd HO SAWVN 119 ON THE NATURE OF ECLAMPSIA aa) €'0 Z'0 a0 20 G30 Ioqe] 1ayye sABp FO I0qB] Jojye sAvp Ee yov}ye Sutrng IOQv] s1OJog Ioqv] Joyje ABp YI6 you}ye Surng Ioqev] 19}je ysne Ioqe] Joye ABp YF yovy4e suring LOB] a10Jog uso A, o19durB[oq 120 ISEI OBATA COMPARISON OF THE TOXICITY OF THE SERUM. OF NORMAL AND ECLAMPTIC INDIVIDUALS Fresh serum, when injected intravenously into the dancing mouse is toxic, causing symptoms which differ only slightly from those produced with placental extract. It was notice- able, however, that there was a longer interval between the injection and the onset of symptoms in the case of the serum than there was with the extract. In table 2 are presented the results of the comparative tests of the toxicity of fresh serum from men and from normally pregnant or puerperal women, as well as from eclamptic women. These results show that no significant difference exists in the toxicity of serum from individuals of the different groups. Furthermore, it is seen that no such difference exists between the serum obtained during the attack of eclampsia and that obtained after recovery from that condition. COMPARISON OF THE SERA OF NORMAL AND ECLAMPTIC INDIVID- UALS WITH RESPECT TO THEIR CAPACITY TO NEUTRALIZE THE TOXIC PROPERTY OT THE PLACENTAL EXTRACT The foregoing experiments had revealed no differences be- tween the materials obtained from the eclamptic individuals and those derived from normal individuals. However, as the following experiment will show, such difference was found in the capacity of the serum to neutralize the poisonous property of the placental extract. The tests of the neutralizing power of the sera were con- ducted by mixing 1 ce. of the placental extract with quantities of the fresh serum varying from 0.7 cc. to 0.025 cc., physio- logical salt solution being added to bring the total volume of the mixtures up to 2 cc. These mixtures were injected after an incubation of one hour at 37°C. The results of these tests are presented in table 3. ON THE NATURE OF ECLAMPSIA Pal It is seen that normal human serum, whether from men or from women that are pregnant or not in that condition, possesses a practically uniform power of neutralizing the poisonous property of placental extract, 0.2 to 0.3 ec. of such serum suf- ficmg to inhibit the action of 1 ce. of extract. On the other hand, this neutralizing power is considerably less in the serum of eclamptic women during the attack, the normal power being restored after the individual has recovered from the condition. As much as 0.6 cc. of the serum taken during the attack, was usually required to neutralize the toxic action of 1 cc. of placental extract, whereas after recovery 0.3 or 0.4 cc., sometimes as little as 0.2 ce. being sufficient. The normal neutralizing power of the blood was found to be restored by the fourth or fifth day of the puerperium in eclamptic women. In interpreting the diminished neutralizing power of the serum of eclamptic individuals, which we have observed, one may consider this change as either a cause or an effect of the symptoms, and the possibility that the change referred to was a result of the convulsions was experimentally investigated. To this end the sera of three rabbits were examined as to their neutralizing power before and after a toxic injection of placental extract, which produced convulsions in the three animals. The results of this experiment, which are presented in table 4, show that, in the rabbit, convulsions do not alter the neutralizing power of the serum. We have investigated the question whether a power of neu- tralizing placental poison is developed in the blood of human individuals by an immunological process. This question was studied by comparing the neutralizing power of the serum of pregnant and non-pregnant women. ‘The results of that com- parison, which are presented in table 5, show that the neutral- izing power of the blood is not increased during pregnancy. Hence, no evidence was furnished, by the study, of an immuno- logical origin of a neutralizing power of the blood. | ; psep psep Pvep Beep —) _ Peep oho peep PoP = _ pop] uorjovor pvop prop — — pvop pop] uorjoves pvop — — pvep prop pvop|'p'u peep} uorjovor — — *p'u peop pop] uoljoved|'p U peop — _ — pvop pvop pvop] uorjoved|'p'u peop — — peep RESP Peep Peep pep po ce pop] uornovod pop pvop — — — Beer BSS) Bvee BSP = = ao pop prop pvop pvop|'p'u pwop _ -— pvop pvop|'p'u pop pvop prep — -- pvop| uorjoved|'p'u pvop pop pvop pvop pvop : pvep|'p'u peep pvop pvop peop} worzoRed) UOTPOveT pvop pvop| uorovos| uoovod pvop pvop| worjoved|*p'u pvop ‘p'u pvop| uorjoved| uorpoveI pvop pvop peop pvop| uorover uoryoOveL pvop peop pvop pvep| uorjovor peep pvop Or o'r Or 01 01 $16'0 96°0 6'0 $20'0 90°0 10 TIOULNOD IIIA TIA IA —_— —— a) — OT OT oT 80 10 9°0 20 £0 F'0 A AI Ill aces ec Oi, Sete et 4 > eet ‘ = i ao Ont OT seseeeeeees* TOVULXE IVINGOVId **NOILOTIOS LIVS TVOIDOTOISAHA a0 9°0 Ul I reeeeeeeees rads IVINGNIUGIXG see eee eee ““wougds SING WNIYAdX a GHL YOd NAVE Nq@ad avi Woe -aS NOHM NOU SNOSUGd JO SANVN 122 ue | | en ee ee ee ee eee § WIV prep peep pop uorjovaL pop pop pvop ¢ pBop peop prop UuOolzpOBOL peop uOI}POBIL psap pep peop peop peep peop peop pop peop peop peop pop peop uoljOBeL peep uolpoRved uor}OBAL pop Ever, Pvep ‘p.U peop eae UOT}IVOL peep peep peop peep peop prop peep uoTJOvad pop peop peop peep Peee peop peop pvop pep uOT}OBOI peep prep BeeP prep peop peop pvop RESP pep u0T}OBOI uoTpOBaL peep Beeb Beep peep PpBep) pu peep esi rem aA aes ee Oe eae ee eee ee ree eee Sere alice rere peop [Ste 2 uoljOveI RaSe peop ‘p'u pop uOT}BOL pop peep ‘pu peop RgeP Beep UuOTJOBAL prop UOTPIBIL prep prep prop uoTJOvaL UOTPVGIL uoT}OBad peop peep peop pep pop uoloBI uorjoved uoT}OBII ‘p'u pvop BEER ‘p'u peop pop ‘p'u pvop —~ pvop| uo1lovod psep Psep — J Ioqv] 10438 Avp YIG Ioqvy] 1a4je Avp 4G IOqv] 19}jJv INoY psgz IOqv] 104 jv NOY YOST 1oqv] 19}jJ8 InoY pAg IOqeVy 19}}e Noy FT 1OqvB] Joyyy “Anoy ~P¥eP Pye Rep prop] pop peop peop peop} pop peop peop ‘pu peep pop pvop peop] peop pop pop pvop| pvop} uorovod pop pvop pop prop pop pop} pu prop ‘pru pvop pvop|'p'u pwop prop ‘pu pvop|' pu pwep pvop| prop pop prop pvop| prep prop pop pvop| pop pop prop peop} pop peop pop OT OT 01 OT OT G16°0 £60 60 = $20°0 $0°0 10 TOULNOD XI IIIA IIA iin) Bo ~- — Sy == acne ‘pu ies = SS - th Loge] 10jJe SAup TE pop prop pvop| uoTpoved —- ~ you} ye Surndg AH UOLJOBod pop pop — — — LOqv] 1oyye ysne uOTPOVIA — — = = = prop|"p'u peop 2 = _ & IOqvy] 104j8 Avp YG ‘pu peo ~- — = = = t o P i. a = = a 1Oqv] 19948 ABp YYG | "L‘IN peop prop pop peop = -- yon} ye Duran] UOTPOBOA pvop|'p'u pvop|' pu pBop — — LOG’ WLOJog $}U0T} -ed oy a = = = a 1oqey 19}ye Avp YL aed CS! ats) yeop — = = ae ‘pu Aas =a = as Loqvy] 1o}yv AVP PAG] > “WL © pvop|' pu pvap|‘pu pvop pop — — youyye Surmng pvop pop pop prop —- -- LOGB] VLOJog Bo — — a =e ne i aa ~ = Loge] 1ayye ABp WATT peep Beep pep pvep as you}ye Sung Ls Bap i) Bap yevop — IOGV] I10Jo Peet Beer peer ed GB e10}° ei 7 Baye OT Ol OT OT OT (1}e) Gem | foe LOVULXA IWLNGAOVId y 4 5 5 5 80 10 9:0 $0 70 A Nouan itee -O8 VIVA IVOINOTOISAHa | 2 = 4 g z'0 £°0 r0 G0 9°0 10 eee eee eee reeene wougg Aare IA A AL UI II J | varotaqvanawrraaxa |" 8 gHHOO moO z BA20n panuyuog—eé WTAVL pop peep prop peep pop pvep| prop pvoep| uomovor pvep| peep pop _ peop prop prop pop pvap peep peop peop pop prop pvop prop Beep pap prop Peep UOTPIBAI pvop pvop} pvop prep prop peep} pop peop pvop peop| prep peep pvop peop} pop} uolovod] UOTOROI pvep| peop} uorjovor peop pvoep|Uoloved) uoTJOROI peop RESP) ) = psep Peep Peep peep; = peep eyo Beep peep Psep jens] Eeep “‘p’u pvop UO0T}OBaI pvop prop peop peop peop pvop pvop peop peop ‘p'u pvop peep} — = — pvop|'p'u peop pop peop) p'u peap peop peop - — peop peop — — ‘pu pvap|'p'u peop — — peop} uornovor — — pvop = — = ‘p'u peep} p'u pap} p'u peop prop peop peop peop peop pvop pvep|'p'u peep — pvop _ UOLJOBA — pvop|'p'u peep|p'u peap|' pu peop ‘pu pep) pu pep) pu peop ee ee ee eee a — IOqBy] 1ayye AVp 44g peop you}}ye Suma Ioqevy] 19438 Avp yAG LOQ¥] 10} Jv SANOY YIFZ yoe}ye Suing IOqB] WOJaq 1OQv] 1a} Je SAVp 2g = Ioqvy 19};8 sXep Ze peep yov}48 Suling Loge] alojogq IOqe] 109jB AVp YIG yov}4y8 SUING IOGB] 19478 Ysne Ioqe] 19438 Ap YIFZ Ioqey] 1048 Avp PF yov}4e Sung 1OqB] o10Jaq PACA, A sliiel A A \L ) $7091} TIOULNOD uOTpOvod pop prop pvop — — pvoep| uorjovor — — pvop pop -— -- pop pvop| uornovod pop pop peop prop prop — pop prop — pop pop pop pvop pvop peop pop pop Peep Beer pop pop OT or o'r oT or $26°0 660 60 8°0 L°0 $20'0 20'0 10 20 £0 XI TIA TIA TA A =| pvop] uorjowod peop peep pvop prop peep] —_ pwep peop pvop peop ‘pu pBop peop} uorovor pvop| uorovos prop prop peop pop pvop|uoljovad 1oqey 1ozye Avp YAS Ioqv[ toyye Asp puz you} Surnqd LOGvV] oLojog ysn¢ LOGB[ 19} JB SINOY YY PP youyye Sunnd LOG¥] a1OJoq ysne* yovy}e SULING youyye duLind§ LOQv] oLOJog OT OT G10 0 G0 9°0 Ill II penuyuoo—s ALAViL £0 10 teeter e tee eceens ‘NOLO -O8 LIVS TVOIDO'TOISAHd *“sdlo Td IVINGWINAdX a INGN t NT ‘SA ns \ sn Zz ry ge ae P96 Bg Ty Oo oy SD 20% wy OS pai bh by we do T2ZER sy UdT} -vd ory -dure oy 128 ON THE NATURE OF ECLAMPSIA 129 TABLE 4 EXPERIMENTAL FLUIDS..| I It Il IV Vi VI Control ANIMALS EM- SERUM ats Alclynce ceo: 0.6 | 0.5 0.4 0.3 0.2 01 } = PLOYED IN THE ELE SLOLO GICAL: SALT 04 0.5 0 6 07 aan EXPERIMENT BOLETIONS eee . : 0.8 0.9 1.0 PLACENTAL EXTRACT....| 1.0 | 1.0 1.0 et) 1.0 1.0 [palo {| Before convul- { — | — | dead dead | dead sion — | — | dead dead ' dead Rabbit 59 ¢ : — | — | reaction | dead dead After convulsion ony ia be pee dead Anaad Before convul- { ca = dead | dead | dead dead sion --- — dead | dead | dead dead Rabbit 60 2 aes lei fj/— — | dead | dead | dead | dead pCR VUISIOR — — dead | dead | dead dead Before convul- — — dead | dead dead sion — | — dead | dead dead Rabbit 61 2 ‘ — dead | dead dead After convulsion pay = deadll dead dead a eee TABLE 5 SS eee eee ANIMALS EMPLOYED DOSE OF DATE OF INJECTION | BODY WEIGHT | 5) sceNTAL EX- RESULT IN THE OF ANIMALS |/ppacr INJECTED EXPERIMENT grams cc. 19th August 1916 1800 7.8 No reaction. ; 22d August 1916 1710 2.0 No reaction Rabbit 4. 2 %/o6th August 1916| 1700 a0) liNo reaction 30th August 1916 1800 4.0 Convulsion, dead | 19th August 1916 1190 1.2 : 22d August 1916 1220 1.3 Babbit 6. o 1) o¢th August 1916| 1230 2.0 | No reaction 30th August 1916 1230 25 Convulsion, dead 19th August 1916 1480 1.0 No reaction iss 22d August 1916 1450 12 No reaction Rabbit 10. 2 ‘| o¢th August 1916| 1310 Terri MeNouseacn aaa 30th August 1916 1220 EAE Convulsion, dead 130 ISEI OBATA COMPARISON OF THE PATHOLOGICAL ANATOMY OF ANIMALS KILLED WITH PLACENTAL EXTRACT AND THAT OT ECLAMPSIA It is important to the present thesis to show an agreement between the anatomical changes of animals dead after the in- jection of placental extract and those observed in eclampsia. Two rabbits and two dancing mice were killed with a heavy dose of the placental extract. The gross and microscopic ex- amination of the dead animals showed; a decrease in the co- agulability of the blood, hemorrhages in the lungs on both sides and often parenchymatous hemorrhage in the liver as well as the frequent formation of thromboses in both the lungs and the liver, these thrombi consisting chiefly of blood-platelets mixed with blood shadows. Both the hemorrhages and the thromboses in the animals are in agreement with the anatomical features of eclampsia, but the decreased coagulability of the blood appears to be a discordant finding. It must be borne in mind, however, that decreased coagulability of the blood is generally noted in death by acute intoxication. Hemorrhage in the viscera is also to be regarded as a usual occurrence during convulsions. The for- mation of thrombi is reported by the authors who have studied the effects of visceral extracts. Notwithstanding the parallelism just described, the ex- periment is not quite free from objection because the results spoken of above have been obtained in animals killed with a single injection. A more exact experimental counterpart of the conditions supposed, under our theory, to exist in eclampsia would be produced by exposing the animals to a more prolonged action of placental extract. Such condition was established in the following experiment. Sublethal doses of the extract were injected into rabbits three times daily over a period of seven to twelve days; the animals being finally killed with a lethal dose. In 5 of the 7 rabbits thus treated there was a marked increase in the coagu- | lability of the blood; in the other two animals the coagulability of the blood was diminished. It happens that in the latter ON THE NATURE OF ECLAMPSIA 131 two animals death was produced with several heavy doses of extract injected at intervals of two to five minutes. It is possible, therefore, that the diminished coagulability of the blood in these animals was the result of this extraordinary treatment. ; In 3 of these animals extraperitoneal hemorrhage was found along the ileopsoas muscle; in 2 of these the hemorrhage was quite recent, while in the third animal it was evidently of con- siderable duration. In 1 animal striking edema was found in the subcutaneous tissues of the whole body and in the muscles; ascites was also present. In 6 animals the alveoli of the lungs and the mucosa of the bronchioles exhibited hemorrhages; macroscopically petechiae could be seen on the surface of the lungs. Thromboses were found in the lungs, some of which were found to be layered, some undergoing a partial hyaline degeneration and some being actually organized. Fatty degeneration was demonstrable in the liver of all of the animals employed in the experiments; in those injected during three days this change was still slight, but in the animals in which the injections had been extended over seven to twelve days the degeneration was striking. Portions of the liver, as was seen in 3 of the animals, had undergone necrosis and in these Glisson’s sheath presented a perivascular infiltration. The kidney presented cloudy swelling in 5 instances, in some of which the renal cells had undergone vacuolar degeneration; others presented a slight fatty degeneration and still others showed hemorrhage in the medullary portion. In most instances, no change was found in the heart; a single case, however, presented edema and a slight fatty degeneration. In the spleen there was recognizable passive congestion, often with pigment deposits. In the literature on the pathological anatomy of eclampsia we find, first of all, the results arrived at by Schmorl (9), who gives the following features, based on a study of 73 cases. In almost all cases anaemic and ,hemorrhagic necrosis was seen in the liver and thromboses and hemorrhage were found in THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 3 132 ISEI OBATA the lungs. In the heart fatty degeneration of the cardiac muscle was frequently encountered and hemorrhage and de- struction of the muscle elements were, also, not infrequent. In the kidneys cloudy swelling of the secreting cells and fatty degeneration were usually found and necrosis and thrombosis were, also, often observed. Peterson (10) pointed out, in the liver, hemorrhage, throm- boses and intensive fatty degeneration and he detected, in the kidneys, coagulated material in the interior of the urinary tubules and hemorrhage, for the greater part, confined to the medullary portion. These pathological findings have been generally confirmed. The coagulability of the blood in eclampsia has been found to be frequently, though not always increased. From the preceding descriptions an agreement is apparent in the pathology of eclampsia and that of the animals killed with placental extract excepting the slighter alterations met with in the kidneys of the latter. I do not assume that the changes referred to are characteris- tic of intoxication with placental extract and of eclampsia only; indeed, J am aware that similar changes may be brought about, also, with extracts from other viscera. I wish only to point out the fact that pathological changes similar to those seen in eclampsia are producible in animals with placental extracts. CONCLUSIONS The present investigation has shown that in its capacity to neutralize the poisonous action of placental extract the serum of eclamptic women is much inferior to that of normal individuals, whether male or female and that the normal ca- pacity of the serum, in this respect, is restored, in eclamptic women, on the fourth or fifth day after labor. The investigation shows, also, that this abnormality of the serum in eclampsia is not brought about by the convulsion itself. Furthermore, not only has a marked resemblance been pointed out between the symptoms produced with the placental ex- PS Se ae — e ~- hates Gots Muerte ON THE NATURE OF ECLAMPSIA 133 tract and those of the eclamptic attack, but an almost perfect agreement has been found between the anatomical features of eclampsia and those of the animals which were killed by repeated injections of placental extract. From these facts we feel justified in drawing the conclusion that the true nature of eclampsia is nothing other than an intoxi- cation by the placental poison which is made possible by a weak- ening in wts normal capacity of neutralization on the part of the maternal blood. The clinical symptoms as well as the anatomical alterations mentioned above are, of course, not peculiar to intoxication by placental extract, as has been shown by several authors (Dold and Ogata (2), Kinoshita (42), Takeuchi (43). How- ever, the derivation of the injurious agents, in eclampsia, from the placenta seems indicated from the remarkable fact that the symptoms cease when the placenta has been eliminated. It is true that eclampsia does rarely occur after the discharge of the placenta, but not after a period of twenty-four hours following the placental discharge. Even for this phenomenon, however, there is an analogy in the animal experiments with the placental extracts in those instances in which the symptoms de- veloped several hours after the injection of the extract, or, as it rarely happened, when a second paroxysm occurred as late as ten hours following the injection. It may also be suggested that this later occurrence of eclampsia may be the symptom com- plex of intoxication by a poison produced by autolysis of the uterus itself as a part of the process of involution of that organ. This idea is rendered plausible by the fact, shown by Yoshi- mura (44), that uterus extracts contain a poison. While the foregoing investigation leaves little doubt as to the nature of eclampsia, one question still remains open; namely, what causes the weakening of the neutralizing capacity of eclamptic serum? Our work on this question is still in progress. 134 ISEI OBATA HISTORICAL REVIEW OF THE THEORIES OF ECLAMPSIA Eclampsia had early been looked upon as a uraemia (Lever, Frerichs, Cohnheim, Spiegelberg), but these two conditions were found to be different not only as to clinical symptoms, but also with respect to the blood findings. According to Szili (14), the freezing point of the serum of eclamptic women can not. be distinguished from that of the serum of normal individuals; furthermore, an accumulation of urinary products is not demonstrable in the blood of eclamptic women. Halbartsma and Loehlein ascribed eclamptic convulsions to a reflex effect of pressure in the pelvis of the kidney. They supposed that such pressure was produced by a compression of the ureters at the pelvic brim by the head of the foetus. This view could not be maintained because of the fact that the eclamptic convulsions take place, in most cases, during the pregnancy or at the beginning of labor; that is, at a time when the head of the foetus is not yet in position to compress the ureters. Eclampsia has been assumed to be caused by bacteria (Doleris, Poney, Blane, Mueller, Albert). Gerdes isolated from the viscera and from the placental surface of the uterus of eclamptic individuals, an organism, which he named Bacillus eclamptica. This organism, however, was identified by Hof- meister as the proteus vulgarus. Zweifel ascribes eclampsia to an intoxication by lactic acid which he found in the urine and serum of eclamptic women and he gives examples of increased lactic acid in gravidae suf- fering from nephritis as evidence that the increase of lactic acid is not an effect of the convulsions. This view seems in- compatible with the fact that an increased quantity of lactic acid is found in women suffering, not from eclampsia, but from nephritis. Similar to these views of uraemic or lactic acid intoxication in eclampsia are the views of Stumpf and of Landois according to which eclampsia is due respectively to aceton and creatinin poisoning. Both of these views are lacking in proot. ON THE NATURE OF ECLAMPSIA 355 Bouchard and his school advanced the theory that eclampsia is an autointoxication caused by metabolic products. This theory has received support from the later experiments of Chambrelent, who reported that the serum of eclamptic women is three to four times as toxic as normal serum. According to the so-called hepato-toxhaemic theory, poison- ous products of metabolism are thoroughly neutralized when the liver is normal, but autointoxication is caused by these poisonous products when the function of the liver is disturbed. The poison in such case is regarded by Massin as a leucomaine and by Ludwig and Savor as carbamic acid. This view was denied by Volhard and Schuhmacher who found that the serum of eclamptic women is not more toxic then normal serum—a conclusion that is confirmed by the present investigation. According to Kollmann and Dienst, eclampsia is a globulin intoxication resulting from imperfect function of the kidney; these authors detected an enormous increase of the fibrin con- tent of the blood of eclamptic women and they believe, this to be due to an accumulation of globulin. This excess of globu- lin the authors assume to be derived from the foetus being retained in the maternal circulation because of an imperfect function of the kidney. Dienst maintains this view in some more recent publications. According to this view, also, the serum of eclamptic women ought to be more toxic than normal serum, a requirement not met by the facts. Against this theory stand, furthermore, the instances of eclampsia in the absence of a foetus, as in two cases of the author of eclampsia with hydatidiform mole and other similar instances in the literature. Lamsbach (25) collected from the literature 68 cases in which eclampsia occurred after the death of the foetus. In 50 of these cases the foetus was found macerated; in 9 cases the foetus was well preserved, while in the remaining cases there was a hydatidiform mole. In view of these facts a theory seeking the cause of eclampsia in the foetus can not be justified. The same objection must be made to the idea of Kinoshita and his collaborators, who produced death in animals after 136 ISEI OBATA symptoms resembling those of eclampsia by injecting the al- bumin-free extract of the animal foetus. The toxic effect thus produced is ascribed by the authors to a substance named by them “‘eclampsm.”’ A number of different views assume the placenta to be the source of the condition of eclampsia. Having shown that syncytial cells actually gain access to the maternal circulation, Veit (27) carried out the following experiment: fluid expressed from the placenta was injected into the abdominal cavity of a rabbit, which then developed albuminuria. Veit observed, furthermore, that the serum of such a rabbit acquired the property of dissolving the syncytial cells and he assumed this action to be due to an immunological reaction product, which he designated “‘syncytiolysin.”” On the basis of these observations this author assumed that eclampsia is due to the rapid extrance into the maternal blood of syncytial cells in a quantity greatly in excess of the available syncytio- lysin that has developed during the pregnancy. The albu- minuria occurs, according to Veit, when the syncytial cells gain access to the maternal circulation more slowly and in more moderate excess of the available syncytiolysin. If this assumption is correct, the serum of the pregnant woman ought to show, with respect to the poisonous property of placental extract, a stronger neutralizing capacity than does that of a man or a non-pregnant woman and experimental animals ought to be made immune, by the injection of placental extract, requirements that are not fulfilled in the present investigation. Ascoli (28) agrees with Veit in the assumption of a poison- ous action of syncytial elements, but his experiment in which serum was injected into the submeningeal spaces, does not appear to be convincing. Wormser (29) repeated that experiment with negative result. Liepmann (30) looks upon eclampsia as an intoxication of placental origin, basing his belief on experiments carried out as follows: a salt solution emulsion of the powder made from the placenta was injected into the abdominal cavity of rabbits. ON THE NATURE OF ECLAMPSIA TST The emulsion from placentae of eclamptic women was _ espe- cially toxic. This result is not confirmed in the present investigation. Weichardt (31) assumes that antibodies against placental cells appear in the maternal blood-stream, when the cells gain access to it and these antibodies dissolve the cells, which then liberate a toxin called by the author ‘‘syncytiotoxin.” In the ordinary pregnancy the blood possesses, furthermore, an anti- toxic mechanism, which, according to this author, is disturbed in eclampsia. The present study has provided strong support for this theory, inasmuch as it demonstrates, for the first time, a distinct difference in the capacity of normal serum and that of eclamptic women, to neutralize the toxic property of placental extract. In a study with Pilz (32) Weichardt comes to the conclusion that placental pulp contains two different toxic principles, one of which accelerates the coagulation of the blood while the other attacks the respiratory center. According to Hofbauer (34), eclampsia is an intoxication brought about by a substance derived from the liver by a proc- ess of autolysis under the influence of a ferment produced in the placenta and discharged into the maternal blood. September, 1918 136 | From sputum—case of influenza pneumonia......... September, 1918 139 | From sputum—case of influenza pneumonia......... September, 1918 HAG AM OStOne Strate 2/6525 2 Sete od Gate tees elses. ic oe tee eee ae LARA WAT Stra 27.5.5: ones ORE = Ses crete eae 143° | Hygienic Laboratory strain 103). 25.0000. 220. .k2 144 | Hygienic Laboratory strain 159........ ............ Other strains employed 120 130 131 From tonsil; cases of influenza....................| September, 1918 132 138 PATHOLOGY OF EXPERIMENTAL INFECTIONS WITH B. INFLUENZAE It was evident from the beginning of our work that our toxic products and in the case of live bacillus infections, toxic deriva- tives had a pronounced effect upon the smaller blood vessels with most striking results as regards the respiratory tract. It is only possible to give a general description of the gross pathology of the lungs in these animals with a somewhat more detailed description of the microscopical picture in representative animals of each series. The latter descriptions were kindly made by Dr. Douglas Symmers to whom our thanks are due. The results here described are divided into three groups. 1. The effects of toxic extracts of the lungs of experimental animals. 2. The lesions produced by injections of live B. influenzae. 3. The pathology of secondary infections, spontaneous and artificially induced. ROLE OF B. INFLUENZAE IN CLINICAL INFLUENZA 183 Effects of toxic extracts of B. influenzae on the lungs of experimental animals A. Salt extract toxin (endotoxin?). In mice, guinea-pigs and rabbits during after receiving this material, the lungs invariably showed an extreme degree of congestion, with haemorrhage in spots under the pleura and sometimes a bloody exudate in the pleural cavity. On section the lungs showed patchy areas of haemorrhage and a marked universal congestion and they exuded a bloody froth. The haemorrhage was frequently seen in wedge shaped areas with the base at the pleura simulating a pulmonary infarct. Mouse. Weighing 20 grams, received 0.5 cc. of salt extract and died in eighteen hours. Microscopical examination of the lungs showed extreme congestion of the capillaries in the interalveolar walls, with patchy areas of haem- orrhage into the alveoli, associated with desquamated epithelial cells. Pig 1. Microscopical examination of the lungs showed scattered areas of haemorrhagic extravasation and intravesicular leucocytic exu- date. The intervening lung tissue was emphysematous and, in places, oedematous. Rabbit 626. Received 5 ce. of salt extract, injected intravenously and died three hours later. Microscopical examination of the lungs showed an acute congestion with a slight amount of intraalveolar haemorrhage and some atelectasis. B. Broth toxin. Generally animals dead of this poison showed in the lungs an extremely acute congestion but without the ex- tensive haemorrhage seen after injection of the salt extract. Guinea-pig 3. Received 1 ce. of broth toxin by intraperitoneal in- jection and died forty-eight hours later. Microscopical examination of the lungs showed a marked congestion with patchy areas of haemorrhage into the alveoli. Rabbit 201. Received 10 ce. of broth toxin injected intravenously and died two hours later. Microscopical examination of the lungs showed an extreme acute congestion. 184 F. M. HUNTOON AND 8S. HANNUM Streptococcus—B. influenza toxin As the pathological changes caused by this preparation were, in al’ respects, similar to those produced by the salt extract toxin no description of those changes will be made. The lesions of the lungs produced by injections of live B. influenzae into the brain This portion of the work was done in conjunction with Mr. Roos to whose paper in this issue the reader is referred for the details. Rabbit 239. Received 2 doses of live B. influenzae into the brain, dying five days after the second dose. The lungs showed acute congestion with some areas of haemorrhage. Cultures from the lungs remained sterile. Microscopical examination of the lungs showed a fairly intense in- jection of the intraalveolar capillaries with groups of atelectatic alveoli. Rabbit 232. Received 1 dose of live B. influenzae into the brain, dying three days later. The lungs showed acute congestion with haemorrhagic areas in both lower lobes. A pure culture of B. influenzae was obtained from both lungs. Microscopical examination of the lung shows an intense injection of the intraalveolar capillaries together with numerous haemorrhagic ex- travasations into the alveoli. Some of the latter areas have fused in such fashion as to produce haemorrhages of wide extent. Secondary infections, spontaneous and artificial The very frequent occurrence of secondary infections in clini- cal influenza and the prevalence of bronchopneumonias associ- ated with the mouth type of organisms led us to attempt the determination of the effect of the toxins of B. influenzae on the localization of such organisms in the lungs with their relation to bronchopneumonia. Karly in our work during an attempt to raise the virulence of a strain for mice, we had noted that four mice kept in one Jar al- though inoculated with a pure strain of B. influenzae yielded on ROLE OF B. INFLUENZAE IN CLINICAL INFLUENZA 185 death in cultures from the lungs a non-haemolyzing streptococcus, whereas other mice in the same series showed pure growths of B. influenzae. Later two guinea-pigs which had received “influenza toxin” alone but had been left in contact with a pig receiving both the toxin and a live haemolytic streptococcus died a few days later with pneumonia, both of these animals giving pure cultures of the haemolytic streptococcus from the lungs. In a third instance a rabbit was given a sublethal dose of toxin intravenously, and twenty-four hours later live B. influenzae were introduced into the nasopharynx. Death occurred in four days with both lungs yielding a pure culture of a Gram positive capsulated diplococcus, having the morphology of the pneumo- coccus. Sections of the lungs showed these organisms present in enormous numbers. The above instances are sufficient to indicate that even with experimental animals suffering from this particular intoxication spontaneous secondary infections may occur with the production of pneumonic lesions. Guinea-pig 4. Received 2 cc. of salt extract toxin by intraperito- nealinjection, waskept in contact with guinea-pig 2, which had been in- fected with streptococcus. Guinea-pig 4 died five daysafter the original inoculation and three days after the death of the streptococcus pig 2. Microscopical examination of the lung of guinea-pig 4 showed it to be the seat of a diffuse pneumonic process characterized by the pres- ence, in the alveoli, of polynuclear leucocytes together with smaller numbers of desquamated epithelial cells, an occasional red blood cor- puscle and coagulated blood serum, the whole supported in a delicate fibrinous network. The larger vessels are intensely injected. The bronchioles are filled by polynuclear leucocytes. Rabbit 492. Received 4 cc. of extract toxin intravenously. Twenty- four hours later living B. influenzae were injected into the nasopharynx. The animal died four days later. Microscopical examination of the lung showed the interalveolar cap- illaries to be universally injected, the small bronchioles containing num- bers of polynuclear leucocytes. In patches the alveoli were filled with desquamated epithelial cells and serum. A section of the lung stained 186 F. M. HUNTOON AND S. HANNUM with the Gram method showed innumerable Gram positive capsulated lance-shaped diplococci both in the alveolar walls and in the alveoli, which show the desquamation of epithelial cells. Artificial secondary infection Guinea-pig 2. Received 2 cc. of salt extract toxin intraperitoneally and on the next day a sublethal dose of haemolytic streptococcus by intraperitoneal injection. The animal died forty-eight hours after the injection of the streptococcus. A pure growth of streptococcus haemo- lyticus was obtained from lungs which showed marked congestion of the interalveolar capillaries with large areas of intravesicular haemor- rhage and patches of leucocytic exudate into the pulmonary alveoli. The control guinea-pig, which received the same dose of streptococcus as guinea-pig no. 2, survived. Guinea-pig 6. Received 1 cc. of salt extract toxin intraperitoneally. One week later the animal was forced to breathe Streptococcus haemo- lyticus, dying forty-eight hours later. Gross examination showed pneumonic consolidation of the upper lobe and portions of the lower lobe of the left lung. Microscopical examination: The sections of the lung were unsatis- factory but they showed patchy areas in which the alveoli were filled with polynuclear leucocytes. In the intervening portions the alveoli contained fibrin with a few red cells and desquamated epithelial cells. Sections of the lung stained with Gram’s method showed the pres- ence of innumerable streptococci. Guinea-pig 7. Dried and ground B. influenzae were introduced into the naso-pharynx and one hour later the animal was forced to breathe living B. influenzae. Death occurred after a further twenty-four hours. Microscopical section of the lung showed extreme congestion of the intraalveolar capillaries with a few areas of haemorrhage into the alveoli with desquamation of epithelium. Guinea-pig 8. Received 2 ce. of salt extract material intraperito- neally and at end of forty-eight hours theanimal was forced to breathe living B. influenzae. Death followed in twenty-four hours. Microscopical section of the lung showed intense congestion of the intraalveolar capillaries. Rabbits breathing toxin alone and B. influenzae alone, survived. ROLE OF B. INFLUENZAE IN CLINICAL INFLUENZA 187 SUMMARY AND CONCLUSIONS An examination of the experimental evidence of the preceding pages would appear to have established the following facts: 1. That B. influenzae is capable of producing a toxic substance. 2. That this substance, when introduced into the circulation, produces congestion of the respiratory tract with haemorrhages into the alveoli. 3. That certain conditions of symbiotic growth intensify the liberation of the toxin. 4, That as an effect of the action of the poison the lungs show a predisposition to invasion by various organisms, with the pro- duction of secondary lesions. 5. That live bacilli introduced at a remote point probably af- fect the lungs through the action of a liberated toxin. 6. That there is nothing in the serological evidence to pre- clude the consideration of this organism as an important factor in the causation of clinical influenza. - oo NOTES ON THE BACTERIOLOGY, AND ON THE SELEC- TIVE ACTION OF B. INFLUENZAE PFEIFFER C. ROOS! From the Mulford Biological Laboratories, Glenolden, Pennsylvania Received for publication May 27, 1919 The disease commonly known as influenza has been desig- nated in the past according to the clinical symptoms manifested in certain localities. In France it was known as La Grippe, from agripper, meaning to clutch, and is analogous to Blitzka- tarrh, a German word, meaning lightning catarrh. Both terms are indicative of the rapid onset of the disease, which was also commonly observed during the past pandemic. Pringle and Huxham in 17438 were the first to designate the disease by the name of influenza, from influxus, meaning influ- ence of cold or influence due to atmospheric changes. Prior to isolation and identification of B. influenzae by R. Pfeif- fer (1,2) in 1892 the infectious rdle had been assigned to vari- ous other organisms, such as the streptococci, pneumococci, Fried- ldénder’s bacillus, and even the staphylococci, apparently as one or the other of those organisms was found to be predominant, or according to the bacteriologic technic used. B. influenzae is an aerobic, hemophilic organism and, accord- ing to Pfeiffer, it prospers best upon pigeon blood agar, reaching a maximum growth in about twenty hours at 37°C. No growth occurs at temperatures below 28°C. nor above 42°C. Attempts to recover B. influenzae from specimens twenty-four hours or more old kept at room temperature, are usually unsuc- cessful, even in the fall and spring months. This has been re- peatedly shown in this laboratory by reculturing specimens of sputa or nasal secretions with negative results where B. influen- zae had been found present in large numbers in the fresh specimen. 1 Read before the Philadelphia Pathological Society, February 13, and March 27, 1919. 189 190 Cc. ROOS The usual requisite condition for the spread of the infection is close association. Since the organism is killed rapidly by ex- posure to low temperatures, infections through the various arti- cles of common use or public meeting places are of less impor- tance. There is no doubt that the spray occurring on sneezing and coughing is a potent factor in the transmission of the disease. The bacillus is very widely distributed and frequently found present in the throats, tonsillary crypts, and larynx of apparently normal individuals producing no symptoms whatever, or it may be the cause of angina, pharyngitis, and laryngitis, without excit- ing the other typical symptoms of influenza. The primary les- ions, however, are usually confined to the respiratory tract, most commonly the mucous membranes of the nasopharynx and nose, occasionally in the trachea and bronchi, less frequently the al- veoli. (Wynekoop (3, 4), Auerbach (5), Polanski (6), Scheller (7), Jochmann (8).) Luetscher (9), of the Johns Hopkins University, in 1915, in his studies on some 600 cases of non-tuberculous infections of the respiratory tract, found the pneumococcus and B. influenzae as the cause of 91 per cent of the infections of the bronchi and lungs; in these B. influenzae was alone in about 30 per cent in infections of the larynx, it was found in 75 per cent and in those of the nose, throat and sinuses, it was found in about 31 per cent. The symptom complex varies greatly, apparently due to the degree of virulence of the infectious organisms, or to symbiosis with other organisms, or to the state of resistance of the individ- ual, also to such predisposing influences as temperature and at- mospheric change. (Eade (10).) The reports as to the cause of the epidemic outbreaks of this disease as recorded by various investigators in this country and abroad have resulted generally in a repetition of inconclusive bacteriologic findings. However, it is noteworthy that the close observance of the improved bacteriologic methods has resulted in the finding of B. influenzae with great regularity. Indeed, the positive findings of the hemophilic bacilli of the B. influenzae type with improved methods of cultivation by experienced in- — eatin — Sa Oe Paty Sine ee et BACTERIOLOGY AND SELECTIVE ACTION OF B. INFLUENZAE 191 vestigators have reached almost the 100 per cent level with the typical cases of the disease. As for the negative findings in the recent pandemic at least, the author is thoroughly in agreement with Park and his able coworkers that Those of us who have, through experience, learned of the ease with which the influenza bacillus may be missed in an examination, know how little dependence can be placed upon results of negative findings unless these reports describe fully the media used, the source and dilu- tions of suspected material, the time allowed for the cultures to grow and the use of specific tests and stains, and also show that all of these factors have been handled in a satisfactory way. (Keegan (12), McIn- tosh (13), Fildes, Baker and Thompson (14), Park, Williams and asso- ciates (11), Robertson (15), Pritchett and Stillman (22).) An unusually high percentage of positive findings of B. influ- enzae has been reported from the various camps during the past year and also in the smaller epidemics that have occurred in this country and abroad since 1915. The pathologic lesions in gen- eral and those of the lungs in particular are claimed to be identical with those of the pandemic just past. (Christian (16), Thomas, (17), Abrahams, Hallows and French (18), Wolbach (23).) The work in this laboratory up to the recent pandemic, includes the study of over 100 cases beginning with the epidemic of 1915- 1916. B. influenzae was found in from 50 to 90 per cent of the cases according to the time of the year and the symptoms. ‘The pres- ence of other organisms varied accordingly and considerably. B. influenzae either in pure culture or as the predominating or- ganism, is seldom found in the specimens of nasal secretions, and then only when the symptoms are distinctly confined to the nasal cavity and characterized by copious discharge and furthermore when such specimens are collected at the very beginning of the symptoms. The author, who has been a sufferer of frequent attacks of the common cold, has been able to demonstrate this fact repeatedly on himself. Great care must be exercised to col- lect the proper specimen. The watery mucous discharge so abundant during the early stage of the disease contains very few 192 Cc. ROOS organisms of any kind. The most productive specimens are those taken from the tonsils or the posterior nasopharynx. In a series of experiments in 1916, the author was able to con- firm the observations of Jacobson (19) in regard to the symbiosis of B. influenzae with streptococci by injecting freshly isolated strains of B. influenzae into mice alone and in symbiosis with avirulent streptococci, either dead or alive. It was found that the symbio- sis of these organisms increased the virulence of B. influenzae about ten-fold. The observations of Cantani (20) showing the favorable influ- ence of other organisms upon the growth of B. influenzae in cul- tures were also confirmed and have proved of considerable value in subsequent investigations. In a series of experimental intravenous injections of rabbits with B. influenzae alone, and in symbiosis with streptococci and staphylococci, by making leucocytic counts ten minutes, six, twenty-four, and thirty-six hours after the injections, the follow- ing results were noted: B. influenzae, either alone or in symbiosis with other organisms, invariably caused a sharp drop of the leucocytes at the expense of the polymorphonuclear cells. The leucocytic count did not return to normal until about thirty-six hours after the injection. In repeated injections the same re- sults were observed. The injection of the streptococci or staphyl- ococci alone, on the other hand, caused a slight drop below nor- mal shortly after injection, soon followed by sharp rise of the leucocytes, chiefly polymorphonuclear cells, to considerably above the normal count. The blood counts during the past pandemic, as reported by numerous investigators, strikingly coincide with the above ex- perimental results in the rabbit. (Keegan (12), Eiman (21).) Beginning with September 19 and during the month of October, 1918, we had occasion to study 31 cases of clinical influenza, all characterized by the sharp onset of the disease. Seventeen of these cases were at the United States Naval Hospital, and 14 at the Presbyterian Hospital of Philadelphia. The sources of the cultures were as follows: throat, 10; pneumonic sputa, 12; pleural effusion (post mortem), 3; lung exudates (post mortem), 8; 33 specimens in all. BACTERIOLOGY AND SELECTIVE ACTION OF B. INFLUENZAE 193 All specimens were collected. with the greatest care, the throat specimens from the tonsils or posterior nasopharynx from patches showing signs of inflammation by means of sterile swabs, a tongue depressor and flash light being used. In the case of sputum the patient was asked to cough up and expectorate into a sterile receptacle. The pleural effusions and specimens of lung exudates were withdrawn by means of a sterile syringe, the chest wall being previously sterilized with tincture of iodine. These specimens were collected within an hour after death. Most of the throat and sputum specimens were collected by the author personally, and plated within three hours after the collection on freshly made horse blood agar plates, previously tested with stock strains of B. influenzae. After the plates had been sown with the infectious material, in the case of swabs, by streaking directly on the plate ; with the sputum, by selecting several pus kernels and washing them in sterile ascites fluid, then transferring to the blood agar plates and distributing by means of a sterile glass spreader, streaks were made of selected strains of B. subtilis and staphylococci known to influence the growth of B. influenzae favorably. The plates were incubated for from twenty to thirty-six hours at 37°C. and then fishings were made on blood agar slants. The delay in culturing some of the other specimens could not be avoided. The material on the swabs, or in the case of sputum a selected portion, was suspended in sterile ascites fluid and injected into white mice intraperitoneally. These were found to die in from fourteen to thirty-six hours, except a few of those injected with the throat specimens, when the animals remained alive probably on account of the very small amount of the infectious material injected. The mice were examined soon after death and cultures were made from the peritoneal fluid and heart blood. B. influenzae was recovered in most animals from the peritoneal fluid and oc- casionally from the heart blood as well. Other organisms recoy- ered from the heart blood of the mice were the pneumococci and the streptococci. The bacteriologic findings revealed the following: pneumococci 194 Cc. ROOS in 20 specimens: type I in 3, type II in 8, type III in 6, type IV in 8; B. influenzae in 27; streptococci, either of the hemolytic or viridans type, were found present in all of the throat and spu- tum specimens, absent from the three plural effusion specimens, also from 5 out of the 8 lung exudate specimens. The cultures from the throat and sputa contained a large variety of organisms besides those mentioned, such as diphtheroid bacilli, Gram nega- tive cocci of the M. catarrhalis and allied types; in a number of the sputum and throat specimens a certain type of Gram posi- tive diplococcus—an influenza like colony. From the bacteriologic findings it will be noted that B. influ- enzae and the pneumococcus were by far the most common organ- isms. Streptococci as a whole were commonly present. B. influenzae was found in all the throat and sputum specimens and in 5 of the lung exudates, not being found in the 3 specimens of the plural effusion, in which pneumococci were present in pure culture; not present in 3 specimens of the lung exudates, from which the following organisms were isolated: B. mucosus and strep- tococci In one, pneumococcus type III in another; one contained no organisms, although from a specimen of sputum taken ante- mortem B. influenzae had been isolated. With the object of establishing fixed strains of B. influenzae and increasing the virulence of that organism, intracranial injec- tions of live organisms were given to rabbits, strains isolated at different localities being used as follows: A strain from Boston, Massachusetts, a strain received from the Hygienic Laboratory, Washington, D. C.—known as the ‘“‘army strain’’—and a strain isolated in September, 1918, by the author from a fatal case of influenzal pneumonia at the United States Naval Hospital, Philadelphia. The injections were made through the petrous portion of the temporal bone by means of a Record syringe and a needle such as usod in the inoculation of the rabies virus by the Pasteur method. The cord was removed by the Oshida method. The method adopted was thought superior in that use is made of comparatively well isolated organs—the brain and spinal cord, the structure and physiologic functions of which are well de- BACTERIOLOGY AND SELECTIVE ACTION OF B. INFLUENZAE 195 fined; there is least danger from rapid invasion of microorgan- isms from other organs, and the necessary operations in regard to direct transplantation and culturing can be carried out in a sterile manner with ease. It was also considered probable that a more constant minimum lethal dose could be secured. Tests, in dextrose bouillon fermentation tubes and blood agar plates, for purity of the organisms and freedom from external con- taminations were made on all cultures and materials used for injections and were discounted from results where such were detected. Whenever animals died within a few hours after injection, death was considered to be due to ‘‘shock.”’ The cultures used were from eighteen to twenty hours old, grown on blood agar, washed off in sterile ascites fluid and injected within thirty minutes to avoid autolysis as much as possible. The dose injected varied from 0.3 to 0.5 mil. The animals used were normal, healthy rabbits weighing between 1800 and 2200 grams. The following observations were made: The minimum lethal dose of B. influenzae varies considerably for the individual strains in regard to virulence, also per gram of weight for the individual animal for any one strain. There were noticeable variations of the clinical symptoms, but these were not confined to any one strain of the organisms, nor were they proportional to the amount injected per kilo of weight. The minimum lethal dose by this method was found to be about 2500 million, 3000 million, and 4000 million per rabbit of about 2000 grams, of the Boston, Army and Philadelphia strains, respectively. It was found that the virulence of B. influenzae is not increased by intracranial passages through the rabbit; in fact a gradual attenuation takes place apparently on account of the insuscepti- bility of the rabbit to infection by this route at least. It may be noted that influenzal meningitis in human beings has been seldom observed. . Starting with a minimum lethal dose of a strain sufficient to kill the animal in about twenty hours, the spinal cord was removed immediately upon death, or the animal was chloroformed when in THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 4 196 ; Cc. ROOS the last stage. Two centimeters of the cord were cut off from the medullary end, ground up in a mortar in strictly aseptic manner, ascites fluid being used as diluent, and one-half of this amount was injected into a second rabbit. The second animal was usually found to die in from thirty-six to forty-eight hours. The third rabbit in the series injected with the same amount of material removed from the second rabbit would either survive or die from secondary infection of the lungs. The bacteriologic examinations made of the spinal cord, lungs and heart blood, and the gross pathologic lesions observed in three separate series as above described, followed a fairly definite course. ‘The first animal in the series showed large numbers of B. influenzae in the cord, the lungs and the heart blood were found sterile. Death was apparently due to acute meningitis with marked toxemia. The animal exhibiting severe meningeal symptoms with retraction of the head, appears very toxic with polypnea, a watery discharge from the nose being frequently no- ticed. Upon post mortem examination the lungs showed acute congestion, being filled with blood, and slightly emphyzematous. The meningeal symptoms in the second animal are less marked at first, with toxemia and polypnea, usually quite distinct in about five to six hours, but gradually though slowly increasing, up to about fifteen hours after the injection, when they may slowly recede. The bronchial symptoms, however, become more aggravated. The discharge through the nose, which at first is usually mucous and watery, becomes mucopurulent, occasionally tinged with blood pigment. Toxemia appears much increased. Breathing becomes labored but less rapid. The animal takes no food and grows weaker rapidly being unable to stand up and some times lying down for twenty-four hours before death. Post mor- tem examination reveals one or the other of the two more or less distinct conditions. First, little or no congestion of the brain, a considerable amount of spinal fluid, which is turbid and blood tinged. When the cord is removed B. influenzae is found present both in the cord and the spinal fluid in quite large numbers. The lungs are usually much congested and emphyzematous, ap- parently with little or no consolidation, the cultures being sterile. BACTERIOLOGY AND SELECTIVE ACTION OF B. INFLUENZAE 197 Second, much congestion and emphyzema of the lungs, and very extensive consolidation and hepatization. The cultures mostly contain the Gram negative bacilli of the B. bronchisepticus and the B. aerogenes groups, streptococci, and occasionally B. in- fluenzae as well. The heart blood is found sterile in either condi- tion, the kidneys showing acute nephritis. There is no fluid in the thoracic or pericardial cavity. The third rabbit in the series shows usually only slight meningeal symptoms which soon sub- side. The toxemia is not very marked. The bronchial symp- toms are usually more noticeable, frequently with discharge from the nose, in which case the animal is found to die from secondary infection in from three to five days. In case of death B. influ- enzae has been recovered from such animals either from the cord, heart blood, or lungs, the pathologic lesions resembling somewhat those described above. The course of disease in one rabbit seems of special interest. The animal received supposedly a minimum lethal dose of the Philadelphia stain intracranially, showing slight meningeal symp- toms and moderate toxemia in about five hours, these gradually becoming more marked but not severe. Twenty-four hours af- ter the injection the animal was considered as recovering and seemed to be getting continually better until about forty-eight hours after the injection, when severe symptoms developed and the animal died several hours later (about sixty hours after the injection). The post mortem examination revealed the follow- ing: the brain and the spinal cord were apparently little affected (no pathologic sections were made), except a moderate amount of slightly turbid cerebro-spinal fluid, which, as well as the cord, proved sterile. The one lobe of the lungs almost entirely con- solidated and having a mottled appearance, the other much congested, emphyzematous and with large hemorrhagic areas throughout. The cultural examination revealed the presence of B. influenzae in pure culture. The heart blood was found to be sterile. The kidneys showed acute nephritis. There was no fluid in the thoracic or pericardial cavity. Through the kind- ness of Dr. Huntoon, specimens of the lungs and kidney from this animal with specimens from several other animals were sub- 198 Cc. ROOS mitted to Dr. Douglas Symmers of the Bellevue and New York University Medical College, for the complete description of which the reader is referred to the article by Huntoon and Hannum appearing elsewhere in this issue of the Journal.” Immunizations of rabbits by intracranial and intraperitoneal injections were made with the three strains individually by giving sublethal doses of the live organisms twenty hours old. A remarkably high agglutinating titer, about 1:64, was ob- tained in some animals after a single intracranial injection. However, there is considerable danger of secondary bronchial infections even when a rather small initial dose is given. For the intraperitoneal injections the proper initial dose was found to be about 3000 million, the second dose about 4000 mil- lion, and so on, at five day intervals. With an initial dose of 5000 million many animals will die after the second injection if the same dose is repeated, unless sufficient time is allowed for the animal to recover before another dose is given. The cross agglutination and cross absorption tests with these , sera against the strains used in immunization, and others isolated from this pandemic, also 6 strains isolated by the author during the epidemic of 1915-1917, showed that there is a common rela- tionship among all of them. The slight differences noted seem to be in degree, not in kind. Cross protection tests in vivo on a small number of rabbits with the intracranial method showed that although the immunity conferred was not very marked and hard to obtain, yet there is a distinct cross protection against all strains. For more complete results pertaining to cross agglutination, cross absorption and especially cross protection, the reader is referred to the extensive and apparently conclusive experiments on the toxic substances of B. influenzae by Huntoon and Hannum, which have already been mentioned. 2 Page 167. BACTERIOLOGY AND SELECTIVE ACTION OF B. INFLUENZAE 199 SUMMARY AND DISCUSSION B. influenzae can be found in every case of true clinical influenza. To isolate this organism, which is most abundant in the earlier stages of the disease, it is necessary to exercise care in obtaining a suitable specimen, and since growth requirements of this or- ganism are quite rigid, special selective culture media, such as those suggested by Avery (24) and also by Fleming (25), carefully prepared and adjusted to reaction are essential for successful work. The various strains of B. influenzae apparently do not differ in kind. This is indicated by the cross agglutination, absorption, and protection tests with strains isolated at different localities during the recent pandemic as well as with those from the epi- demic of a few years ago—1915 to 1917. The toxic substances of B. influenzae show a marked action on the bronchorespiratory tract, thereby predisposing these organs to extensive invasion by the organism itself or to secondary infections. No marked increase in virulence of B. influenzae has been ob- tained by passages through laboratory animals. This, in the first place, may be due to the relative insusceptibility of these animals to the infection of this organism; secondly, to the probability that the invasive power of the organism is very limited, infection apparently taking place only when the initial toxicity is severe enough to facilitate such invasion. No bacteremia is produced by B. influenzae in laboratory ani- mals by a dose approximating a minimum lethal dose regard- less of the mode of injection chosen. Rabbits receiving intra- cranial injections will either die of acute toxemia and show no organisms in the blood stream or lungs, or where such infection passes into the sub-acute stage there is apparently a chance for a few of the organisms to get into the blood stream and to be transferred to such organs as the lungs when these have been rendered susceptible by the toxic substances of the organism. The pathologic lesions in the rabbit, gross and microscopic, in many respects resemble those of influenza in human beings as observed during the past pandemic. 200 Cc. ROOS The injection of B. influenzae into the rabbit intravenously results in a rapid and marked decrease in the polymorphonuclear cells. REFERENCES (1) Preirrer, R.: Vorliufige Mittheilung ueber die Erreger der Influenza. Deutsche med. Wochenschr., 1892, 18, 28. (2) Preirrer, R.: Die Aetiologie der Influenza. Zeitschr. f. Hyg., 1893, 13, 357. (3) Wynexkoop, F. E.: A further study of the influenza bacillus. Jour. of A. M. A., 1903, 40, 574. (4) Wynexoop, E.: Concerning the bacterioloy of influenza. Chicago Med. Recorder, 1903, 24, 343. (5) Aurrpacu, M.: Ueber den Befund von Influenzabacillen in Tonsillen und Larynx. Zeitschr. f. Hyg., 1904, 47, 259. (6) Potansx1, W.: Pathologische Verinderungen und Complicationen in Nase, Rachen, Kehlkopf und Gehérorgan bei Influenza. Gazeta lekarska, 1907, 27, 951, et seq., Jahresger. ges. Med. von Virchow und Hirsch, 1907, 2, 4. (7) ScHELLER: Ueber die Verbreitung der Influenza Bacillen. Centralblatt f. Bacteriol., Orig., 1909, 50, 503. (8) JocHMANN, G.: Influenza. Lehrbuch der Infectionskrankheiten. Berlin, 1914, 340. (9) Lurerscuer, J. A.: Bacteriological and clinical study of the non-tuberculous infections of the respiratory tract. Arch. of Int. Med., 1915, 16, 657. (10) Eanes, P.: Influenza in 1893. Lancet, 1893, 2, 798. (11) Park, W. H., Witurams, A. W., AND AsSocIATES: Studies on the etiology of the pandemic of 1918. Amer. Jour. of Public Health, 1919, 9, 45. (12) Kerean, J. J.: The Prevailing Pandemic of Influenza. Jour. of A. M. A., 1918, 71, 1051. (13) McInrosu, J.: The incidence of B. influenzae (Pfeiffer) in the present influenza epidemic. The Lancet, 1918, 2, 695. (14) Fripss, P., Baker, 8. L., anp THompson, W. R.: Provisional notes on the pathology of the present epidemic. The Lancet, 1918, 2, 697. (15) Ropertson, W. F.: Influenza: Its cause and prevention. British Med. Jour., 1918, 2, 680. (16) Curtst1an, H. A.: Incorrectness of the diagnosis of death from influenza. Jour. of A. M. A., 1918, 71, 1565. (17) THomas, H. M.: Pneumonia at Camp Meade, Maryland. Jour. of A. M. A., 1918, 71, 1307. (18) ABraHams, A., Hautows, N., AND Frencu, H.: A further investigation into influenzo-pneumococcal septicemia: epidemic influenzal ‘‘pneumonia”’ of highly fatal type and its relation to ‘purulent bronchitis.’’ The Lancet, 1919, 1, 1. (19) Jacopson, G.: Essai sur l’action pathogéne du bacille de Pfeiffer chez les animaux. Arch. de méd. expér., 1900, 13, 425. vice es SPL ee et ES eat es a aS S BACTERIOLOGY AND SELECTIVE ACTION OF B. INFLUENZAE 201 (20) Canrant, A.: Ueber das Wachstum der Influenzabacillen auf hiimoglobin- freien Naihrbéden. Zeitschr. f. Hyg., 1901, 36, 29. (21) Erman, Joun: Presbyterian Hospital, Philadelphia, (Personal communica- tion). (22) Pritcuert, I. W., anp Stituman, E. G.: The occurrence of B. influenzae in throats and saliva. Jour. of Exper. Med., 1919, 29, 259. (23) WorBacu, S. B.: Comments on the pathology and bacteriology of fatal in- fluenza cases, as observed at Camp Devens, Mass. Bull. Johns Hopkins Hospital, 1919, 30, 104. (24) Avery, O. T.: Jour. of A. M. A., 1918, 71, 2050, (25) Fuemine, A.: On some simply prepared culture media for B. influenzae. The Lancet, 1919, 1, 138. : ne nes ete) 2, IMMUNOLOGIC DISPARITIES OF SPORE AND VEGETATIVE STAGES OF B. SUBTILIS RALPH R. MELLON anp LILLIAN M. ANDERSON From the Department of Laboratories, Hahnemann Hospital, Rochester, New York Received for publication June 16, 1919 This study was undertaken in an effort to discover whether any necessary immunologic similarity existed between the two stages in the life history of spore bearing organisms. An organ- ism of this type was chosen because of the unimpeachable evi- dence that the spore is derived from the bacillus and vice versa. We have been interested particularly in the agglutinin responses of the two stages, chiefly on account of the position taken by Eberson (1) in his study on diphtheroids. In his attempt to discredit the contention that diphtheroids may under certain conditions assume a diplococcus morphology and, indeed, reproduce themselves as diplococci, he makes the statement that ‘‘it is obvious that, if one form is derived from another, there should be cross agglutination between the various forms.’’ We fully admit the logic of this position, but are some- what more interested in its truth under all conditions. Accordingly, antispore and antibacillary sera were developed in rabbits in the following manner: A four to six hour culture of the bacilli which on staining showed no spores was killed with a 2 per cent formaldehyde solution, and controlled four days. ‘ : mee ANAPHYLAXIS REACTION IN THE RABBIT » eet 4 6) Metin: Skand. Archiv f. Physiol., 1904, 15, 147. (8) Larson, W. P., anv Bet: Journ. of Inf. Dis., 1919, 24, 185. (6) Rosrosxr: Work ae phys. med. Gesellsch. zu Waurzb. 1093, N. F., 35, 15. ; (7) MicuarLis AND OPPENHEIMER: Arch. f. Anat. u. Physiol. Goat Abth. ; Suppl. 1902), p. 336. . x Coca, A. F.: Journ. of Pepmonel., 1919, 4, 209. me % She ie 114. of iy 7 ee y . aoe a ss as Be i: <0 SOME SUGGESTIVE EXPERIMENTS WITH B. INFLU- ENZAE; ITS TOXIN AND ANTITOXIN A PRELIMINARY REPORT N. 8S. FERRY anp E. M. HOUGHTON From the Research Department, Parke, Davis and Company Received for publication July 9, 1919 SUSCEPTIBILITY OF LABORATORY ANIMALS The question as to the pathogenicity of Pfeiffer’s bacillus for laboratory animals has never been given the attention it de- serves, as practically all text books dismiss the subject with a very few words, leaving an impression that it is non-pathogenic, except for the human subject and perhaps the monkey. Woolstein has called attention to the pathogenicity of this organism for the monkey, rabbit, guinea-pig and white mouse. She found that the white mouse “succumbs to intraperitoneal injections of cultures irrespective of their origin” giving rise to a peritoneal exudate containing large numbers of the influenza bacilli, as does the heart’s blood. The microorganisms were found in other organs, and the spleen was always swollen. Guinea-pigs succumb to intraperitoneal injections of one blood agar culture of all meningeal, and about one-half of the respi- ratory strains tested. The peritoneal fluid was increased at times to 8 cc. and the spleen was increased to two or three times its normal size. The kidneys were congested and the lungs showed scattered areas of congestion and inflammation. Influ- enza bacilli could be obtained in pure culture from the pleural exudate, the heart’s blood and viscera and from the surface of the pia of the brain and spinal cord. Rabbits inoculated into the ear vein succumbed in from fifteen to thirty-six hours. Small hemorrhages were found in the parietal peritoneum and within the serous coat of the intestines 233 234 N. S. FERRY AND E. M. HOUGHTON and beneath the capsule of the liver, pleura and other organs. The spleen was swollen and soft, the kidneys were much con- gested and the lungs always showed areas of hemorrhage and of inflammation. Cultures could be obtained from the heart’s blood, viscera, urine and from the surface of the brain and cord. From the congested mucous membrane of the upper nasal cavity large numbers of influenza bacilli were cultivated. The successful production of influenzal meningitis in the monkey depended upon the selection of a virulent culture and the maintenance of the pathogenicity. This disease in the monkey terminated fatally in from thirty-six hours to four days. The authors, working only with the rabbit, guinea-pig and white mouse, were able to corroborate the findings of Woolstein and have been able to show that the infection in these animals is a typical general infection or septicemia and that guinea-pigs and white mice were invariably susceptible to the strains at hand. The cultures with which we have been working were obtained through the courtesy of the Hygienic Laboratory, Washington, D. C., Dr. E. C. Rosenow, of the Mayo Foundation, Rochester, Minnesota, Cook County Hospital and the New York Board of Health Laboratory, during the month of October, 1918. These cultures have all proven invariably pathogenic for guinea-pigs and mice and occasionally for the rabbit and, by means of repeated inoculations, their virulence has been increased four- fold and over. Inoculations were made intraperitoneally and intrathoracically. An invariable picture showed a general in- fection with intense general congestion. An intrapleural inocu- lation into the guinea-pig resulted in a bloody pleural effusion, usually entirely filling both pleural cavities. After peritoneal inoculations the peritoneal fluid was not, as a rule, bloody. ~ TOXIN PRODUCTION Corroborating the work of Parker, we have been able to pro- duce soluble toxins, for all strains, fatal for rabbits in intrave- nous doses from 2 to 5 ec.; death usually resulted in about two hours. Intense prostration within about half an hour was the = lee ie Ae ap ee ge B. INFLUENZAE: ITS TOXIN AND ANTITOXIN 255 rule in all rabbits whether they died or not. In those rabbits which received nearly the fatal dose, the prostration would last several hours. The following day the rabbits gave no signs of intoxication. In producing this toxin, the technic of Parker was followed as closely as possible; the organisms were grown in veal infusion broth to which was added 10 per cent of defibri- nated rabbit’s blood (ater on horse blood was substituted for rabbit blood). This media was prepared by heating over water bath at 75°C., or until the blood coagulated and settled on stand- ing. In obtaining the toxin, the culture was incubated about eighteen to twenty-four hours and then centrifuged at high speed and filtered through the Mandler diatomaceous filter. ANTITOXIN PRODUCTION With this soluble toxic product or toxin, serum has been pro- duced by us, in a horse, which proved bactericidal as well as antitoxic for the first set of experiments. In the production of this antitoxin the horse was injected with increasing doses of the toxin, both intravenously and subcutaneously; the in- jections were given every three to seven days. The first few inoculations were made with the toxin only; for the later ones the culture was centrifuged but not filtered, hence a large number of the live organisms were being inoculated into the animal at each operation. This was for the purpose of producing an anti- bacterial as well as an antitoxic serum. Unfortunately, before enough serum was obtained for a repetition of the experiments, the horse died; however, it was felt that the results were at least suggestive enough for a preliminary report. (Death of the horse was due to thrombo-embolic colic; in no way connected with the treatment.) Experiments with anti-influenza serum Experiment 1. Test of potency of anti-influenza serum from horse 902, May 22, 1919. Serum and toxin were mixed and allowed to stand thirty minutes at room temperature. 236 N. S. FERRY AND E. M. HOUGHTON Injections made intravenously at 9.20 a.m. RABBIT NUMBER SERUM TOXIN RESULTS ce, cc 20 0.5 6 Alive 21 1.0 6 Alive 22, 2.0 6 Alive 23 Control 6 Prostrated with diarrhoea at 10.05 a.m. Dead at 10.45 a.m. Conclusion. 0.5 ce. serum protected. Experiment IT. Test of potency of anti-influenza serum from horse 902. May 22, 1919. Serum injected intravenously fifteen minutes after toxin. RABBIT NUMBER SERUM TOXIN RESULTS cc cc. 24 2 6 Dead 1 hour 25 2 6 Died 1 minute after antitoxin injection 26 Control 6 Dead 1 hour Conclusion. No protection. Experiment III. Test of protective value of anti-influenza serum (horse 902) against culture of B. influenzae. Serum injected intra- peritoneally twenty-four hours before cultures, May 21, 1919. Mimi- mum lethal dose of culture for guinea pig, 0.25 of test tube. GUINEA-PIG NUMBER SERUM CULTURE RESULTS cc, 55 1 0.75 test tube Died May 22 56 2 0.75 test tube Died May 22 57 3 0.75 test tube Died May 22 58 Control | 0.75 test tube Died May 22 Conclusion. Dose of culture too large. B. INFLUENZAE: ITS TOXIN AND ANTITOXIN Dat Experiment IV. Same test as III with smaller dose of culture. GUINEA-PIG NUMBER SERUM CULTURE RESULTS ce, 59 1 0.5 test tube Dead 5-23 60 2 0.5 test tube Alive 61 3 0.5 test tube Alive 62 Control 0.5 test tube Dead 5-23 Conclusion. 2 cc. protected. Experiment V. Test of protective value of anti-influenza serum (horse 902) against culture of B. influenzae. Serum and culture injected simultaneously, intraperitoneally, May 23, 1919. GUINEA-PIG NUMBER SERUM CULTURE RESULTS cc, 63 1 0.75 test tube Dead 5-24 64 2 0.75 test tube Dead 5-24 65 3 0.75 test tube Alive 5-24 66 4 0.75 test tube Alive 5-24 Conclusion. 3 cc. protected against large dose. DISCUSSION Irrespective of the etiological relationship of the Pfeiffer bacillus to influenza it is an interesting fact to know that a soluble toxin can be produced which will stimulate the formation of an antitoxin. It is, also, of great scientific importance to know that this antitoxin can not only neutralize the toxin in vivo as well as in vitro, but it can also protect against bacterial infection of the guinea-pig with B. influenzae. Two outstanding facts should be emphasized; namely, the extreme congestion following an injection of guinea-pigs with fs. influenzae, and the profound prostration in rabbits, due to 238 N. S. FERRY AND E. M. HOUGHTON a toxemia, following the injection of the toxin, which very closely simulates the early stages of influenza in the human subject. REFERENCES Wootstern, Martua: Am. Jour. Dis. Children, 1911, 1, 42; Jour. Exp. Med., 1911, 14, 73; Jour. Exp. Med., 1915, 22, 445. ParkER, JuLia T.: Jour. Am. Med. Asso., 1919, 72, 476. A STUDY OF THE THERMOLABIL AND THERMO- STABIL ANTILYSINS (ANTICOMPLEMENTARY SUBSTANCES) OF HUMAN SERUM! TAKAATSU KYUTOKU From the McManes Laboratory of Experimental Pathology of the University of Pennsylvania Received for publication July 3, 1919 That sera may develop antihemolytic properties is a well known phenomenon and particularly in connection with comple- ment fixation tests when sera may be found to contain substances capable of interfering with hemolysis and requiring the serum control tube for its detection. The general result of a large amount of investigation on this phenomenon has been to estab- lish the fact that these antilysins interfere with serum hemolysis by exerting some distinctive or antagonistic influence upon hem- olytic complement and for this reason the phenomenon is famil- iarly known as the ‘‘anticomplementary”’ activity of serum. Furthermore, it has been shown with human sera that heating at 56°C. for thirty minutes may remove these anticomplementary substances and for this reason they are designated as ‘‘thermo- labil’”’ while in other sera the substances are not removed by this degree of heat and these are designated as ‘‘thermostabil.”’ The sera of specimens of blood more than three days old are heated for the Wassermann test to remove the thermol- abile anticomplementary substances rather than native com- plement which in all probability has undergone spontaneous deterioration. Noguchi (1) gives an excellent review of the literature up to 1906 and in his studies with dog and sheep sera he found that the anticomplementary action of most sera developed after heating to 56°C. or higher due to the liberation of an antilytic substance, 1 Presented before the annual meeting of the American Association of Im- munologists, Atlantic City, June 16, 1919. 239 THE JOURNAL OF IMMUNOLOGY, VOL. IV, NO. 5 240 TAKAATSU KYUTOKU while heating to 90°C. reduced or removed this antilytic sub- stance. Noguchi also found that his antilysin is a lipoidal sub- stance and that it may be removed from serum by extraction with ether or, by absorption with many kinds of blood corpus- cles, which thereby acquire a greater resistance to serum hemo- lysins. Zinsser and Johnson (2) in a study of the thermolabil anticomplementary bodies in human serum found that, unlike the thermostabil substances described by Noguchi, they could not be removed by digesting serum with red blood corpuscles and that they were allied with the globulin fraction of serum rather than the lipoidal elements. Kolmer (3) in studies con- cerning the phenomenon of non-specific complement fixation by normal rabbit, dog and mule sera found that the anticomplemen- tary activity was increased by heating at 56°C. for thirty min- utes followed by a decrease when heated at 62°C. and entire re- moval by heating at 70°C. for thirty minutes. Blood corpuscles were found to absorb a portion of these antilytic substances. Both the serum lipoids and proteins (particularly the globulins) were found to be concerned in the antilytic and non-specific complement fixation reactions with normal rabbit and dog sera (4). PURPOSES OF INVESTIGATION Inasmuch as Noguchi and Kolmer worked with dog, rabbit, ox and sheep sera while Zinsser and Johnson employed human sera, the differences in the results may have been due to the fact that the phenomenon varies according to the sera of different animals; for example, human sera when heated apparently do not develop the anticomplementary properties found by Noguchi and Kolmer with rabbit, dog, mule, ox and sheep sera. At the suggestion of Professor Kolmer, I have undertaken to study the anticomplimentary substances in human sera after the following plan, in order to determine whether more light may be thrown upon the mechanism of this interesting phenomenon and a means discovered for its removal, in view of the very practical bearing that these problems have upon diagnostic complement fixation tests. ANTILYSINS OF HUMAN SERUM 241 1. The influence of heating human sera in relation to anti- complementary substances. 2. The relation of bacteria to the development of anticom- plimentary substances in serum. 3. The anticomplimentary activity of hemoglobin in human sera. 4. Changes in reaction and hydrogen ion concentration of anticomplimentary human sera. 5. The relation of the proteins of human serum to the phenom- enon of anticomplimentary activity, as determined by refrac- tometric studies. 6. The relation of ether soluble lipoids to the phenomenon of anticomplimentary activity of human serum. 7. Experiments bearing upon the removal of anticomplimen- tary substances from human serum by methods of absorption and filtration. Preliminary experiments have shown that the antilytic sub- stances in serum act upon complement, which is in entire accord with the numerous investigations of others; in my experiments these substances were found to have no direct antilytic effect upon corpuscles alone nor hemolysin alone, but have shown a direct antagonistic or neutralizing effect upon complement. My experiments also support the statements of Zinsser and Johnson that in all probability these antilytic substances do not ordinarily exist preformed as such in sera, although they may be occasionally encountered in perfectly fresh sera, but rather they are secondary products of development in sera under certain conditions. GENERAL TECHNIC The majority of the sera used in this study were secured from syphilitic individuals undergoing treatment in the clinic of Dr. Schamberg, and were employed fresh and unheated and after being heated at 56°C. for thirty minutes. Throughout this paper unless otherwise stated, “‘heated”’ refers to this degree and duration of heating in a water bath. 242 TAKAATSU KYUTOKU Tests for anticomplimentary activity were ordinarily conducted by placing in a series of chemically clean and sterile test tubes increasing amounts of serum with a constant amount of comple- ment fixed at 1 cc. of a 1:20 dilution of the sera of guinea-pigs and sufficient physiological salt solution to make the total volume about 2 cc. These mixtures were then incubated in a water bath at 38°C. for one hour and the degree of anticomplementary influence determined by adding two units of hemolysin and 1 ce. of a 2.5 per cent suspension of washed sheep cells to each tube followed by reincubation in a water bath for one hour and read- ings after the tubes had stood in the refrigerator at 0 to 2°C. overnight. The hemolysin was always titrated with each complement serum and corpuscle suspension for accurate adjustment in the hemolytic system. The usual hemolytic system and corpuscle controls were in- cluded in each experiment. PART ONE The influence of heating human sera in relation to anticomple- mentary substances While heating the sera of the dog, rabbit, ox, and sheep at temperatures between 50 and 60°C., may increase the antilytic activity as shown by Noguchi and Kolmer, similar changes do not occur with human sera. As shown in table 1, unheated human sera may be anticom- plementary while heating at 56°C. tends to remove these thermo- labile anticomplementary substances. This experiment was con- ducted with ten sera collected without special precautions and kept for five days at room temperature. All were anticomple- mentary before heating while after heating at 56°C. for thirty minutes none were anticomplementary, showing the influence of heat in removing these antilytic substances. Additional exper- iments conducted with sera heated for thirty minutes at 60°C., 70°C., 80°C. and 90°C. have shown that antilysins do not develop as described by Noguchi with the sera of the lower animals. ANTILYSINS OF HUMAN SERUM 243 Thermostabil anticomplementary substances resist heating at 50°C. for as long as two hours as shown in table 2 in an experi- ment conducted with four sera; likewise these heat resisting TABLE 1 The influence of heat on serum containing thermolabil anti-complimentary substances eae Gannon Wane LEAS ee sited PRLS Poel eked EBS. ce 1 0.2 M.H.+{| C.H Cty! (|) Co C.H C.H C.H 2 0.2 N.H S.H M.S \\; Ca C.H C.H C.H 3 0.2 N.H N.H NHS IN. St M.H C.H + 0.2 N.H N.H INE. |S HL. C.H C.H C.H 5 0.2 N.H Ne VME WC ol Cort C.H C.H 6 0.2 N.H N.H NB PS -H: M.H C.H C.H i 0.2 N.H N.H N.H.. |) VS. ME M.H C.H 8 0.2 M.H C.H CH) C2H. C.H C.H C.H 9 0.2 N.H 8.H MH M.H M.H C.H C.H 10 0.2 N.H S.H M.H C.H C.H | C.H C.H * Min. = Minute. 7 C.H. = Complete hemolysis; M.H. = Marked hemolysis; S.H. = Slight hemolysis; N.H. = No hemolysis. TABLE 2 The influence of heat on sera containing thermostabil anti-complimentary substance See NUS | AMOUNTION i GNHEATHD 5 MINUTES 10 MINUTES 15 MINUTES 20 MINUTES BER SERUM 1 0.2 N.H S.H SH MH M.H 2 0.2 S.H. M.H. M.H V.M.H V.M.H 3 0.2 N.H. S.H. M.H. M.H. M.H. 4 0.2 S.H M.H V.M.H. V.M.H. V.M.H SERUM NUM-| AMOUNT OF 25 30 45 60 90 120 BER SERUM MINUTES | MINUTES MINUTES | MINUTES | MINUTES | MINUTES OO 0.2 M.H. M.H. M.H. M.H. M.H. M.H. 0.2 V.M-H.| VoM-H. | VME VM.) VME |) Vi. 0.2 M.H. | V.M.H.| V.M.H.| V.M.H.| V.M.H.| V.M.H. 0.2 V.M.H.| V.M.H.| V.M.H.| V.M.H.| V.M.H.| V.M.H. He CO DD 244 TAKAATSU KYUTOKU antilytic substances may not be entirely removed by heating at 60°C. for thirty minutes as shown in table 3 in an experiment with twelve sera. As shown in table 1 thermolabil anticomplementary substances may be removed by heating at 56°C. for five to thirty minutes; the difference in time required for their removal is partly quanti- tative inasmuch as sera containing most antilytic substances before heating require the longer exposures. Temperatures below 50°C. appear to have but slight influence upon these anti- complementary substances while a rapid reduction occurs at temperatures between 50° and 60°C. TABLE 3 The influence of heat on sera containing thermostabil anti-complementary substances eee tr aoe. \UMeeareD| 40°C™ 45°C. 50°C. 60°C. cc, 1 0.2 N.H. N.H. N.H N.H V.S.H 2 0.2 N.H. N.H. N.H N.H V.5-H: 3 0.2 N.H Nee. N.H N.H SHE 4 0.2 N.H. N.H. N.H N.H M.H. 5 0.2 N.o. N.H N.H N.H NH: 6 0.2 N.H N.H. N.H. N.H Vi8e8 7 0.2 N.H. INSEL INGE N.H M.H. 8 0.2 N.H. N.H. N.H N.H S.H. 9 0.2 VS.) VS Eig as e S.H 5-H. 10 0.2 V.8,H. | V.S:H | V-8.2. 74V.5.a dal CAH 11 0.2 N.H. N.o. N.A: IN-E N.H. N.H. 12 0.2 INGE: N.H. N.H. S.H M.H. M.H. * In the water bath for thirty minutes. PART TWO The relation of bacteria to the development of anticomplementary substances 1. Sterile broth and egg albumin do not develop anticomplemen- tary substances. These experiments were conducted in order to determine whether plain neutral broth containing the usual amounts of protein in the form of pepton and beef extract and a ANTILYSINS OF HUMAN SERUM 245 egg white, develop anticomplementary properties when kept under sterile conditions in an incubator at 37°C. This tempera- ture was chosen as being most favorable for the occurrence of chemical changes. The results of an experiment with steril broth neutral in re- action to phenolphthalein and titrated for anticomplementary activity after incubation from one to twenty-five days, are shown in table 4, and indicate that while broth in large doses may contain thermolabil and thermostabil anticomplementary substances in mixture with 1 cc. of 1:20 pig serum, there is no increase in TABLE 4 The anti-complementary activity of plain steril neutral broth AMOUNT OF lpDay 5 DAYS 10 DAYS 25 DAYS BROTH (UNDILUTED) Unheated*| Heated | Unheated| Heated | Unheated| Heated | Unheated| Heated is} is) 0.05 Co Wa Org al C.H (OA s ey fg Or 5 | C.H CH C.H 0.1 C.H C.H C.H C.H C.H C.H CH C.H 0.2 C.H C.H C.H C.H C.H C.H C.H C.H 0.3 C.H C.H C.H C.H C.H C.H C.H C.H 0.5 CABe peOoH «|. C.H.. jp CH. se ClBRN Cure. te CuEb vel. E 1.0 Con. (a Grr aaa CH.” ieC Eee eB ie stinne |, CUE... |). Cab. foes) VME? eV Mic VME) ©oiir? Vi IMer. | Cee.) °C C.H. 2.0 VEE iene MOE C sb vine ee sbte |. Can tC ie 3. NAB) ONE), SS. |) MGR ViVi MeEL.! Voir. OMS Control | C.H. : C.H. | CH. | CH. | CH. | CH. | CH. | CH. *C.H. = complete hemolysis; V.M.H. = very marked hemolysis; M.H. = marked hemolysis; S. H. = slight hemolysis; N.H. = no hemolysis. this antilytic activity over a period of twenty-five days. As a general rule, heating broth at 56°C. for thirty minutes reduced the antilytic activity. The results of an experiment with steril egg albumin shown in table 5, indicates that this substance rich in proteins, does not become anticomplementary under steril conditions. 2, Steril human sera develop thermolabil anticomplementary substances. In the majority of instances fresh steril human sera are not anticomplementary as tested with a satisfactory comple- ment of the sera of guinea-pigs; as shown in these experiments, 246 TAKAATSU KYUTOKU however, steril human sera develop thermolabil but not thermosta- bil anticomplementary substances when kept sealed in ampules over a period of time. As shown in tables 6,7, and 8 steril sera develop thermolabil anticomplementary substances in from three to seven days when kept hermetically sealed in ampules at 37°C. TABLE 5 The anti-complimentary activity of egg albumin RESULTS AMOUNT OF EGG WHITE (1; 5) First day After 7 days ce. 0.2 CaEk* C.H 0.4 C.H C.H 0.6 C.H C.H 0.8 C.H C.H 1.0 C.H. C.H. 2.0 Care C.H. Control CSE C.H. *C.H. =complete hemolysis. TABLE 6 The anti-complementary activity of steril sera kept in an incubator AMOUNT OF FIRST DAY AFTER 4 DAYS AFTER 7 DAYS AFTER 17 DAYS SERUM (1: 10) Unheated| Heated | Unheated| Heated | Unheated| Heated | Unheated| Heated but thermostabil substances do not develop during periods as long as seventeen days. At room temperature thermolabil substances develop in about seven days (table 9) but in the refrigerator (temperature 0-2°C.) these did not develop over a period of thirty-one days (table 10). Thermostabil anticomplementary substances did not develop in any specimen. aS Sten = > sents aes ANTILYSINS OF HUMAN SERUM 247 TABLE 7 The anti-complementary activity of steril sera kept in an incubator AMOUNT OF FIRST DAY AFTER 4 DAYS AFTER 7 DAYS AFTER 17 Days C—O ee (1: 10) Unheated} Heated | Unheated| Heated | Unheated| Heated Unheated| Heated ce. 0.1 C.H Ci C.H C.H C.H C.H (Gyel C.H 0.5 C.H C.H C.H C.H Cf C.By |} iVIM-H.| Ch 1.0 C.H C.H C.H C.H C.H C.H M.H C.H 1.5 C.H C.H C.H CE. Vi. MEHih CET M.H C.H 2.0 C.H C.H @:E €.H.. |\ VM.) Ci M.H C.H 3.0 Cai. OMe Ol a (Gilg le Jn fied & MA A nes Gon & Un MPH iC. A. Control | C.H (Orsi a ls C.H C.H. C.H. C.H. Care C.H TABLE 8 The anti-complementary activity of steril sera kept in an incubator AMOUNT OF FIRST DAY AFTER 3 DAYS AFTER 7 DAYS AFTER 10 Days pe a a a a (1: 10) Unheated| Heated | Unheated| Heated | Unheated| Heated Unheated| Heated rrr =e et ee 0.1 C.H. Cia C.H. (Ons C.H. C.H. C.H. C lH 0.5 Ce (Gsleb Cre Ca. | VeMe Eh, |) ‘nae (Cel, C.H. 1.0 C.H. CBS VV M.H.| (CH i) ViSae Com: SiH C.H. Ths CAE C.H. S.H. C.H. Visits) Core S.H. (Ons & 240 V@aak C.H. eH C.H. N.H. Ce: N.H. C.H. 3. CABO CORLEY SOE. | Cue “ene Herm Green) Nee Se ae ontrel)) CH. | C-He i C.E:, | Cao eG neh ead) Clmbi Gum: TABLE 9 The anti-complementary activity of human sera kept at room temperature 1 Day 7 DAYS 31 Days AMOUNT OF Se ROMS § | aeetoah pee a tas aa Unheated Heated Unheated Heated Unheated Heated ee el a S| eee ae 0.01 C.H C.H ClHe C.H C.H C.H 0.05 C.H C.H C.H. C.H S.H C.H 0.1 C.H C.H C.H. C.H S.H C.H 0.2 Car: CE V.M.H. Cire N.H. rH: 0.3 C He C.H. V.M.H. C.H. N.H. C.H. Control (Gplal C.H. CAE (Onlel, CEE C.H 248 TAKAATSU KYUTOKU TABLE 10 The anti-complementary activity of steril sera kept in a refrigerator (Oto 2°C.) 1DayY 7 DAYS 31 Days AMOUNT OF aed | | === ee | SE ee Unheated Heated Unheated Heated Unheated Heated eres ts eS Ee 0.01 C.H C.H C.H. C.H S.H C.H 0.05 C.H C.H Care C.H N.H C.H Ost C.H C.H (skit C.H N.H C.H 0.2 C.H. C.H. (OME C.H. N.H. C:He 0.3 C.H. CH: C iH. CH N H. C.H. Control C.H. C.H. Cie C.H. C.H. C.H. 3. Serum dried in filter paper does not become anticomplementary. Ordinary fresh human serum collected under usual conditions, with no special aseptic technic and containing bacteria upon culture, was quickly dried by fanning, in Schleich and Schull’s paper no. 597 and kept in tightly stoppered bottles at room tem- perature. Anticomplementary tests with these fresh sera before and after heating and in amounts up to 0.3 cc. showed the com- plete absence of anticomplementary substances. Subsequent tests with the dried paper at intervals up to twenty days as shown in table 11 indicated thermolabil and thermosta- bil anticomplementary substances do not develop under these conditions. 4. Sera containing bacteria develop thermolabil and_ thermo- stabil anticomplementary substances. In striking contrast to these results with steril sera are the results of experiments with sera contaminated with various microorganisms and particularly staphylococci; contaminated sera invariably develop thermolabil and thermostabil anticomplementary properties as described by Craig (5). In table 12 are shown the results of anticomplementary tests with plain neutral broth kept in an incubator for three days; one portion was kept steril and a second portion was inoculated with Staphylococcus albus. Titrations show that the latter became highly anticomplementary and particularly when used unheated. 249 ANTILYSINS OF HUMAN SERUM pO) STON | ele GOM els hOll 18 (Oi as (0) |pashfO)| (salty) dally) AshOl lp sekOr Is hO)| alah istOl is eO jes br@) 0 [or} UO ye DTON| | eyo) Pas Oe) s Ol el byOh iam s iO). ols OH fois l{O) al sts O)y/ el {O)4) 1s (tO) Pals tO) fle lrOy || ie lr) || eis br@y a8 40) 0'€ 00 X ¢ SLT) ALS EGON | 218 OO) heel 8 IO Sele WOH} 18 fO) (ees Ores GO)|/es MON; IslOn| IshO)y] MshO) Piso mislO)ehOMieIshioyp ita v6 OST x¢ ETO) feel (gO )4| ele GO Nes 3 ber Wy kad sO el a0) || eg) a hrO)y (ale GO) Pas MOMs Oh | lelO) tse) | elsliOr isi) | sl) Sat OOT x WS Ur OM | 1 (70) fale Oy a) 3 by Py cal s Oa peal s EXO) (ll BOs ieee Oy et sOrles ls (Opals Oddieale brOyt| 18 0) |pas ei) |S hiOy lle iO) aT 08 xg ASTON) 18 Ur Bi ioa) & OO alias Gta els EO Pas We a ie ON ea oI @ kia a ltO ois rO) || lsO) as iO) MM ShO} ash) asco) del) 6°0 09 xg “1EUTO)| (e815) pia s (Oa fits Fer i aS (2 Ja ad 2 1 @ Dees 3 i @ YF tO ellen SEO py (als Os KO) pies iO} | elrOl ds tiO)4| isl Op] 18h tO) 9°0 nx gs DET OV ol UFO ap 8 Os ds OM 23 8 Gt lea 2 BO) fam Sc @ js taal Cc Ya led c EO) feels IO) aed LO) als MO pele lrOr| ale 10) jf eli) hale ltO) £0 0c XG “LeEfOy| 1a TOM) 18 RON ls OMe a GO), [tS VOM) aidan SO MELON altO) | as MOr SiO M/s Ma) || Seb Oilers ht) jp S10 (0) Orx¢ ‘09 “UU peyeoyL ae p9789R is i peeve eee peyeoq aanie peyBoAL tae pe7yBvoy ane pe7yeaA Lie pezeay eae Ge D ee a ae kaa Lk eae al Ae ae. Ga | ake La ETC Co Lee lites =. Wwouwas uddvd UWINIOS BATPOBUT | UWINIIS dATJOW | WINES oATJOVUT | WINIIS9ATJOW | WINIJOSaATJOVUT | WNIesoATOW | WNIesoATpOVUT | UINIOsSeAIWOy [TO ENOOWV) aad ; : : ONIGNOds 40 GZI8 -auu00 a0 Siva 0% alo Siva (I a1O siva ¢ aio x4Va [ dadvd wajpyf uo parip wndas fo fhzvayon hupywawadwoo-yup ay 7, Tl WIGVL 250 TAKAATSU KYUTOKU TABLE 12 The anti-complementary activity of steril and staphylococcus broth ee Oe st STERIL BROTH STAPHYLOCOCCUS BROTH (UNDILUTED) aa a Unheated Heated Unheated Heated cc. 0.1 C.H Ci: C.H C.H 0.2 C.H C.H. C.H C.H 0.4 C.H C.H. V.M.H C.H 0.6 C.H GC: V.M.H C.H 0.8 O- C.H. M.H C.H 1.0 CH CH. S.H. C.H. 2.0 C.H. CH S.H. V.M.H. Control C.H. C.H. C.H. C.H. In table 13 are shown results of tests with the two sterile human sera kept at room temperature sealed in ampules for a period of two weeks; a portion of each serum was inoculated with Staphylococcus albus. As shown in this table the steril portions of each serum developed thermolabil anticomplementary prop- erties as previously described while those portions containing staphylococci developed more of this variety of anticomplemen- tary substance and thermostabil antilytic substances in addition. TABLE 13 Anti-complementary activity of steril human serum and serum containing staphylococci * SERUM I SERUM II AMOUNT OF i SERUM Steril _ Staphylococci Steril _ Staphylococci (1:10) inoculated serum inoculated serum Unheated| Heated | Unheated| Heated | Unheated| Heated | Unheated| Heated 30 |NH.| cH. | No. |vmMuH| No_| CH | OUNE. | ONE Control | GHG H.|.C- | @m.| peeelcHn |-c.a 4 een * These steril and staphylococcus inoculated sera were kept at room tem- perature for two weeks. Te a A ANTILYSINS OF HUMAN SERUM 251 While staphylococci occur most frequently among contaminated microorganisms found in sera as ordinarily collected, and produce thermostabil and thermolabil antilytic substances, other micro- organisms are capable of producing the same changes. In table 14 are shown the results of an experiment conducted by dividing a steril human serum into five parts and inoculating four parts with Staphylococcus albus, B. coli, B. typhosus and B. subtilis re- spectively reserving the fifth part as steril serum. All were sealed in ampules and placed in an incubator for ten days when anticomplementary tests were conducted with each portion TABLE 14 The réle of various bacteria in producing anti-complementary substances in human sera* AMOUNT STERIL STAPHYLOCOCCUS B. TYPHOSUS B. COLI B, SUBTILIS | | | (1:10) ws Heated! Unheated| Heated | Unheated} Heated ma Heated | Unheated! Heated SS Oe eG | CH. Wl) Co: €.H. Cie |) CB? iG@tHe |G. CH. (iG:H. ee ©. Veo .| V.M.H.|) C.Ae | Coe. GiB: | (C.B | VME G:H. Ones. | C.Aoit IN Be | NE: | V.MiB,) CB. Sse ) VM. Be oS: Vi M_ ONG LH: | C.H: |) N.H; | .N.H. SH +). C20: | NE Beles: B N.H. | V.M.H. Beene. | CH} Nome | Ni. °| NBs 4 ClB | NGR |e Sib N.H. S.H. Con- 5 trol] C.H. | C.H.| C.H. GE Cho) Coe 4 He iC. C.H. Cyr * Each serum was inoculated with an equal amount of culture of each germ. before and after heating at 56°C. for thirty minutes. As shown in this table the steril portion did not become anticomplemen- tary whereas the four portions containing the bacteria developed thermolabil and thermostabil anticomplementary substances and especially those contaminated with staphylococci and B. subtilis. A question of considerable interest in this connection is whether the cocci themselves or their products act as antilytic substances. To answer this question an experiment was conducted by dividing a steril rabbit serum into three parts as follows: One part was sealed in ampules and kept steril;a second portion was inoculated with staphylococci and incubated for five days. At the end of this time the number of cocci per cubic centimeter was counted 252, TAKAATSU KYUTOKU and an equal number of washed cocci from agar cultures were added to the third portion of steril serum carried over in sealed ampules in an incubator. The results of anticomplementary tests upon the three portions before and after heating are shown in table 15; the steril portion had not become anticomplemen- tary whereas both portions containing staphylococci developed thermolabil and thermostabil anticomplementary substances. Inasmuch as the addition of washed cocci to sterile serum ren- dered the latter anticomplementary it is logical to conclude that TABLE 15 The anti-complementary activity of steril and contaminated rabbit sera AMOUNT OF B SERUM* C SERUM D SERUM BeROM. ja ee Eee (1: 10) Unheated Heated Unheated Heated Unheated Heated cc 7S iia eee ee eee 0.1 CH CREE Cli C.H Cm C.H 0.2 C.H C.H C.H C.H C.H Carl 0.4 CH C.H C.H C.H (@AE| CH 0.6 Car C.H CH: CH C.Hi CaF 0.8 C.H CH C.H. C.H. (Gas! C.Hi 1.0 C.H C.H M.H. C.H Ar C.H 2.0 C.H C.H N.H. M.H V.M.H V.M.H 3.0 C.H C.H N.H. S.H N.H M.H Control C.H CoE C.H C.H CE C.H *b = steril serum kept in an incubator for five days; c = serum inoculated with staphylococci and kept in an incubator for five days; d = steril serum to which washed staphylococci were added. the cocci themselves irrespective of their products, are antilytic, which is in accord with the well known anticomplementary activity of bacterial-antigens in general. Similar results were observed in an experiment in which plain steril neutral broth was employed. One portion was placed in an incubator for five days; a second portion was inoculated with Staphylococcus albus and the number of cocci per cubic centimeter was counted at the end of five days. To the third portion, carried over in the incubator for five days, was added a corre- sponding number of washed cocci from agar cultures and anti- complementary tests were conducted with the three portions ANTILYSINS OF HUMAN SERUM 253 before and after heating. As shown in table 16 the steril portion remained free of antilysins whereas both portions containing cocci developed thermolabil and thermostabil anticomplemen- tary substances. As will be described later, the removal of these cocci from serum and broth by porcelain filtration removed the anticomplementary activities of both. TABLE 16 The anti-complementary activity of steril and contaminated broth AMOUNT OF B BROTH* C BROTH D BROTH BROTH CONDING- |p oma cae i | fai OP EL LL | = TED) Unheated Heated Unheated Heated Unheated Heated cc 0.1 C.H. CoH. C.H. (Gals li C.H. C.H. 0.2 (Grlsle aH C.H. (Onlal CH Ce 0.4 C.H. CH: C.H. C.H. (Gplel Ci: 0.6 C.H. (Gah C.H. aE. @xHE C.H. 0.8 C.H. CH: V.M.H. Cyr CHE CHE 1.0 Cire C.H. V.M.H. (GAEL. M.H. V.M.H. 2.0 CH. Cle: M.H. M.H. N.H. N.H. 3.0 (OH si C.H. M.H. Sis ly N.H. INGE Control C.H. CE. C.H. (Gals. C.H. C.H. *b = steril broth kept in an incubator for five days; ¢ = broth inoculated with staphylococci and kept in an incubator for five days; d = steril broth to which washed staphylococci had been added. PART THREE The anticomplementary activity of hemoglobin in human serum Since the sera of old specimens of blood collected under usual conditions, deeply tinged with hemoglobin and usually contami- nated with bacteria, frequently prove anticomplementary both before and after heating, experiments have been conducted for the purpose of determining whether the products of disintegration of erythrocytes alone may contribute the antilytic substances. A solution of fresh sheep cells washed four times was prepared in 100 ce. of steril distilled water by dissolving as much blood as possible with gentle shaking; this solution was filtered through paper and rendered isotonic with sodium chlorid. Anticomple- 254 TAKAATSU KYUTOKU mentary tests conducted with the fresh solution before and after heating showed the material to be highly anticomplementary inasmuch as amounts as low as 0.1 ce. contained sufficient ther- molabil and thermostabil anticomplementary substances com- pletely to inhibit hemolysis as shown in table 17. Additional experiments were conducted by collecting human blood under sterile conditions into test tubes and conducting tests with a portion of the sera for anticomplementary activity before and after heating. In one experiment the remainder of the serum was left on the clot in an incubator for eight days, TABLE 17 The titration of anti-complementary activity of sheepcell hemoglobin AMOUNT OF FIRST DAY AFTER 10 DAYS HEMOGLOBIN SOLUTION i (UNDILUTED) Unheated Heated Unheated Heated ma 0.1 N.H. N.H. N.H. N.H. 0.2 N-H. NER: N.H. N.E- 0.4 N.H. INGER: N.H. iIN-H 0.6 N.H. N.H. N.H. N.H. 0.8 N.H. INCE: N.H. N.H. 1.0 N.H. N.H. N.H. INGE 2.0 N.H. N.H. INSEL: N.H. Control (ai 5 C-:H. C.H. C.H. at which time it was discolored a deep red with liberated hemo- globin but steril when cultured; anticomplementary tests with this serum before and after heating showed that it was highly antilytic, especially before heating (table 18). In a second experiment conducted in the same manner with two sera but kept at room temperature and titrated for anti- complementary activity at intervals, showed the development of thermolabil anticomplementary substances in about ten days, at which time the sera contained from 20 to 25 per cent of hemo- globin (table 19). ANTILYSINS OF HUMAN SERUM 255 TABLE 18 Anti-complementary properties of hemoglobin in steril serum Ragen cneen ci FIRST DAY AFTER 8 Days* , oe Unheated Heated Unheated Heated ce: 0.1 C.H CoH C.H C.H 0.5 C.H C.H M.H C.H 1.0 C.H CE S.H C.H eS C.H C.H. S.H M.H 2.0 C.H. (Onlel N.H. S.H. 3.0 C.H. (On18 | N.H. S.H. Control Orr CH: C.H Ci * This serum was steril upon culture. TABLE 19 The anti-complementary activity of hemoglobin in steril serum 1 DAY 4 DAYS AMOUNT OF SERUM No. 1 No. 2 No. 1 No. 2 (UNDILUTED) Unheated} Heated | Unheated| Heated | Unheated| Heated | Unheated| Heated Maia Crema) ca. | @au! | em) em): cme) | Ga: Control |. CH. | CH. | CH. | cH. | cH. | GH | CH. | CH 7 DAYS 10 Days LSC SERUM No. 1 No. 2 No. 1* No. 2 (UNDILUTED) Unheated} Heated | Unheated| Heated | Unheated| Heated | Unheated| Heated 03 | CcH.| cH.| cH. | cH | cH. | CH lvMH! CH. Control | C.H. Crt (OME [ @hHe (Oplai (@eE @aEl CEH * Serum 1 was steril upon culture at end of ten days; but serum 2 was con- taminated. These sera developed a deep red color at the end of seven to ten days. THE JOURNAL CF IMMUNOLOGY, VOL. Iv, NO. 5 256 TAKAATSU KYUTOKU PART FOUR Changes in reaction of anticomplementary sera In view of the fact that minute traces of inorganic acids and alkalies have been found highly anticomplementary, (6) experi- ments have been conducted to determine whether changes in the reaction of steril and contaminated sera may be detected by colorimetric methods as these sera develop thermolabil and thermostabil antilysins. In studying changes in the reaction of the sera a new indicator was employed described by Bronfenbrenner (7) and composed of equal parts of a one-half per cent watery solution of China blue with a 1 per cent solution of rosolic acid in 95 per cent alcohol designated as C-R. In preliminary experiments I have found that this indicator may be prepared free of bacteria and as tested with dilutions of normal sodium hydroxid and normal hydrochloric acid is at least four to six times more delicate.an in- dicator than phenolphthalein; anticomplementary titrations of serum alone and serum plus ten drops of indicator per cubic centimeter have shown no increase of antilytic activity of the latter. The addition of one drop of C-R to each cubic centimeter of fresh, clear and almost colorless human serum imparts a purplish red color indicating neutrality; incubation of clear steril serum plus the indicator sealed in ampules over a period of three weeks, usually caused a slightly deeper red color, indicating a change of reaction in the direction of alkalinity and titrations at this time showed the presence of small amounts of thermolabil anticompli- mentary substances. Similar experiments conducted with almost colorless sera con- taminated with Staphylococcus albus showed that production of alkaline changes as indicated by the development of redder tints over a period of ten days, after which no further changes in color could be detected; at the end of twenty days these sera were highly anticomplementary before and after heating. Studies bearing upon changes in the hydrogen ion concentration of steril and contaminated sera during the time anticomple- CSF ae: ANTILYSINS OF HUMAN SERUM 257 mentary substances were being formed, were attempted by means of a colorimetric method. The results of these experiments indicate that no definite information was to be obtained by this method. Specimens of human sera were collected under rigid aseptic conditions and in glassware prepared with particular care to insure chemical cleanliness. Anticomplementary tests and hydrogen ion determinations were conducted with each of the fresh sera, following which each specimen was divided into two portions and treated as follows: the first portion was kept steril and the second inoculated with Staphylococcus albus, both being placed in prepared ampules hermetically sealed in an incubator for seven days. Fresh steril serum showed a Ph of 7.6 to 7.8 and was free of anticomplementary activity before and after heating; after seven days in an incubator steril sera showed a Ph of 7.8 to 8.0 and thermolabil anticomplementary substances, while the contaminated sera showed a Ph of 7.4 to 7.6 and both thermo- labil and thermostabil antilysins. PART FIVE Refractometric studies of anticomplementary sera As previously stated, Zinsser and Johnson have found that the thermolabil anticomplementary substances of human serum are removed by precipitating the globulins with ammonium sulphate; Kolmer found a reduction in the antilytic properties of dog, rabbit and mule sera with similar methods. These investigations indicate therefore that the anticomplementary substances of serum are, at least in part, associated with the protein constituents and especially the globulins. In this investigation studies were conducted after the refrac- tometric method of Robertson (8); the results of a series of experiments are shown in table 20. In conducting these experiments steril and contaminated human sera were employed; anticomplimentary and refracto- metric tests were conducted with each serum while fresh and then repeated seven to twenty days later; the sera were kept in an incubator sealed in ampules to prevent evaporation. TABLE 20 Protein changes in anti-complementary sera as determined by the refractometric method ANTI-COMPLE- RESULTS OF REFRACTOMETRIC MENTARY TESTS* DETERMINATIONS Bes CONDITION OF SERA oe 42 b 5 2 b Sal ee | ae | Boe lee wail ge (32) ee A eoiees ia) & al < io) a Diesar cc. cc. per cent| per cent| per cent| per cent 1 Ares ee. secs hey oe OF 0 8-3) |) 5: Galea 7 days; contaminated.......... 0205 7)'0.05 "|" 9.8" |] 9°39) “O25 1 eee 9 Wes: dota 8 he 6 on cpael ete cypereierec tere 0 0 S7-4|) a0) ale eal 7 days; contaminated.......... 070531'0205)| 10:0 | °7.9"| 2o1peates 3 BTS :o38 02.704 Ses hee aaa te 0 0 C20 | BET | (4ASOR BS 20 days; contaminated......... 0.05:).0:3 | 10.2.4 -.9:2.) 1cO8\eetes A Breshis:) detncok ere ee eee 0 0 129°| 428. |, SUE sae 20 days; contaminated......... 0.05 | 0.05 | 11.2 | 10.3] 0.9} 1.8 5 Bresh:,;.. ans Sacer oe eheee eee 0 0 Sof ed Bole 12 alee ZOTOAYSs SbETH 2250 . Seis vamos 0.15 | 0 9.0 |.°6.0.| 3:04 gies 6 Wresh sore ysci ke cee els oe OE 0 0 Wis2 EOI, 245 nl eeles. 20 days; sterile s2.c..2- eens 0.15 | 0 9.0): 4.9 ))" 455) pas OSE tence 'c Wena re kc emperor 0 0 7.2.1 6.2 | T.00\" aa6 7 Cidayss Steril wc sch wereceortaee Os i0 8-9.1 (6:44). 62-5 ed 7 days; contaminated.......... 0.05 | 0.05 | 8.6] 6.6] 2.0}; 1.0 ORY Re aey. ison icis( ares eee 0 0 8.5.) S07 5) O:Salaaies 8 Madaye sterile, ators ek ee Are Ose 110 83515-0508) aes 7 days; contaminated.......... 0205710205 ||. “O27 1) V8 7 1299) ee 9 INot hiltored « oc\. <2. 06,2. seeeeee 0-1 4) 0.3 S27 |6.6 | -2 51 aes IBN ORe Gi erin enc tiscceccitels cemteee 0 0 422)" 470 | Of2a alee 10 INotsiitered..22.65 os scsi ts eee 0.05 | 0 TGA G4 MW) de eee Pulpered er. eee ti ccc cars s ae 0 0 3.4) od.211 | 0222) 2220 u Unheatede. 3292 5) oct Re 0.06 S20 7 57) SOrS | aoe Fleated Gycree sen... ises ete e 0.2 8.2)) 7:4)| 0:85) y2z# e | TOrnhies red Sein esa is oe Ore ea 0.02 8.6] 7.0] 1.6] 2.4 || “Hteated!t renee. 25129 2 REE OL2 8-18) "6/8 )| Le eas * Smallest amounts of undiluted serum proving anti-complementary are given in this table. +0 = Not anti-complementary in dose 0.3 ec. undiluted serum which was the maximum dose tested. t Filtered through Kitasato candle filter. § Heated in a water bath at 56°C. for thirty minutes. 258 ANTILYSINS OF HUMAN SERUM 259 Owing to the individual variations in the amounts of albumin and globulin in the different sera, the results of these studies will not bear too close analysis, but a general analysis warrants the following conclusions: 1. Contaminating sera proving highly anticomplementary before and after heating have shown an increase of the total protein and especially of the albumin fraction; it is reasonable to infer that these changes are due to the presence of bacteria. 2. Sterile sera containing thermolabil anticomplementary sub- stances showed a slight increase of total protein and especially of the globulin fraction. 3. Sera strongly anticomplementary before and after heating and filtered through Kitasato filters, which removed both the thermolabil and thermostabil antilysins, showed a decrease in total proteins and especially of the globulin fraction. 4. Refractometric determinations of anticomplementary sera before and after heating showed practically no changes in the protein constituents even though there was a marked reduction but not a complete removal, of the anticomplementary substances. PART SIX The relation of ether soluble lipoids to the anticomplementary activity of human serum As previously mentioned, Noguchi has found that the antilysins liberated in dog, ox, and sheep serum as a result of heating at 50 to 60°C. may be removed by heating the serum with ether and the extract, freed from lecithin and certain related bodies, contains the antilysin in a concentrated but not in a pure form, which can now be taken up in a saline solution in which it dis- solves. Kolmer found that the dog and rabbit sera extracted with ether were more antilytic than untreated portions of the same sera until they were heated at 56°C. when the extracted portions were less antilytic; that is, extraction of those sera with ether increased the thermolabil anticomplementary activity but reduced the thermostabil substances. In my experiments conducted with anticomplementary human sera, extraction with ether did not remove the thermolabil or 260 TAKAATSU KYUTOKU thermostabil anticomplementary substances, but rather brought about such changes as to increase the antilytic activity of the sera. In these experiments sera were titrated for antilytic activity before and after heating; 3 cc. of each were then extracted with 10 cc. of ether and the latter was carefully separated in the centrifuge and decanted into an evaporating dish. The residue complementary tests were then repeated with the sera before and TABLE 21 The extraction of anticomplementary sera with ether ANTICOMPLEMENTARY TITRATIONS No. SERA Unheated serum Heated serum 0.05 0.2 Plain serum......... M.H C.H. Plain serum after 1 extraction......... M.H S.H. Ether residue in sa- linens 5.£e 2b A ee C.H. Cane Plain serum......... Cale C:He Plain serum after 2 extraction......... S.H. N.H. Ether residue in sa- MING ee ecctssc cane Care CE Plammsserulieni:s. .2/.o-. S.H. N.H. Plain serum after 3 @xXtraction......3-.. S.H. N.H. Ether residue in sa- linvetrerdne kek. 2 Crna Care *N.H. = No hemolysis (strongly anticomplementary); S.H. = slight hemoly- sis; M.H. = marked hemolysis; C.H. = complete hemolysis. of serum was then extracted once more with the same volume of ether and the latter was decanted into the same dish. Anti- after heating and with the same complement serum; the ether was evaporated and the residue was taken up in 3 cc. of physio- logical saline solution and tested for anticomplementary activity before and after heating in the same manner as the sera. The results of an experiment with three sera are shown in table 21; ANTILYSINS OF HUMAN SERUM 261 sera 1 contained large amounts of thermolabil and small amounts of thermostabil antilytic substances; serum 2 contained thermo- labil antilysins only while serum 3 contained large amounts of both. As shown in this experiment extraction with ether did not remove the thermolabil antilysins except to a slight extent in serum 3, and saline solutions of the ethereal residues did not con- tain antilytic substances; as a general rule extraction with ether increased the amounts of thermostabil antilysins inasmuch as the antilytie activity of all sera heated after extraction with ether was higher than that with the plain sera after heating. PART SEVEN The removal of anticomplementary substances from human serum by methods of absorption and filtrations 1. The influence of absorption with barium sulphate. Wechsel- mann and Lange have shown that absorption of syphilitic sera with barium sulphate increases the delicacy of the Wassermann test probably by removal of natural antisheep hemolysin as shown by Noguchi and Bronfenbrenner; experiments have been con- ducted with anticomplementary sera to determine if barium sul- phate, kaolin, bone ash and other substances were capable of removing the antilysins. In one set of experiments sera were diluted 1:10 and titrated for antilytic activity before and after heating and portions of 2 cc. were treated with 1.1 cc., 5.5 cc. and 11 ce. of a 7 per cent sus- pension of barium sulphate in physiological saline solution; these mixtures were incubated for an hour and the barium was removed by centrifugation. The supernatant sera were collected and further diluted with saline solution until the final dilution was 1:10 when the antilytic tests were repeated before and after heat- ing; the results are shown in tables 22 and 23. As shown in these tables barium sulphate is capable of absorb- ing a portion of the antilysins of human sera and especially thermolabil antilysins; the larger amounts of barium removed more antilysin than the smaller amounts. 262 TAKAATSU KYUTOKU TABLE 22 The absorption of thermostabil and thermolabil antilysins by barium sulphate AMOUNT OF] UNTREATED SERUM no. 1* No. 2 no. 3 SERUM (1: 10) Unheated| Heated. | Unheated| Heated | Unheated| Heated | Unheated| Heated cc, 0.1 C.B,..| C2.) | vCB | Csey | A@eBe hy Geely) (CARE: aeaete 0.2 C.H..,| C.Hi i) (CH. 6) eC 2E a COE), WC) |) © aera 0.3 G:H..|| Cale | (CA) Cabin wean. 1 Cubl. ) (Cu. eae 0.4 CaS "|, OR | (CAH AGATE aa et oa te Seb | (Et | eel 0.5 CAH.) | (CB |) (CH. Coe eC sy) °C. iG Gare 0.6 C.H: |. C:H.4\| -C.Be >| GABE Al WGuk) |) (Ca. |), KC SE Gabe 0.7 CAB, | ‘(CHG |) (Ce ais Cah VG Seen il) OAH) MG BEE AGEs 0.8 V.M.H:| (C.H.. | (CEE | Cone Cae) CoB 7) (Can ean 0.9 V.M.H.| V.MCB)) (Cane (Cee) SCAB) (CA, 2) AC. eae 1.0 M:H. | V-MUH.| UCI. st (C21) SCs a) Cae. |, (CE a eal 2.0 N-H: | N B.S 3B SME IN2 | WOME | VE MORE Ss Ee S.H. 3.0 N-H.) NGS =| ONSET ONG Ne) SNC NEE NS Ete * No. 1 = 2 ce. serum + 1.1 ce. of 7 per cent BaSQ.; No. 2 = 2 cc. serum-+ 5.5 ec. of 7 per cent BaSO,; No. 3 = 2 cc. serum + 11 ce. of 7 per cent BaS O, TABLE 23 The absorption of thermolabil anticomplementary properties from human serum by barium sulphate AMOUNT OF | UNTREATED SERUM no. 1* No.2 no.3 SERUM RP A Ae ea es A Re ks A eee (1:10) Unheated| Heated | Unheated| Heated | Unheated| Heated | Unheated| Heated cc. 0.1 GrH:.,| “C.H. | (C.Hi.| WC Hos) CH. |) ACEH ts) SCS: teases 0.2 C.H.. | ‘C.H: |, C:H. | CoB C.B.. | (© A) (@. SG re 0.3 CB. CH. | -C.H. \) @3HE” | Cin (Cea) CHa, | Gane 0.4 IVS, 8} COE.) . CA) | ACOH: | (CAR a OCs © Ee ae ere 0.5 MiG | C.H.. | CH. «| \CiHe a) CYS pl s@ariey |) (CJ mete 0.6 MGH. | C.H. | CoH. | ClH | (CHE iG@AE sel) Cary || Gare 0.7 Sele & C.H.. | “CuH-). Cat.” | (Clean a: | CE 2 |) seat 0.8 S.H. C.H. | CiH. {) (Co | (CARES eC. H."’| (C. He | Gare 0.9 ME Cl. ») CoH: Cas | W@aERayCsBL, |) SC Hey a@ams 1.0 M.-H C.B.. | VMs.) (Ciae |) (CAE ae CE.) CES Gaeta 2.0 V.M-HAGC:.H. | V.M.H.| C21. 9 «Cag CH. | Ca easels 3.0 M-BoweC!H. |) MoH. 1) S@JHe | Cnaet | CoH. | CARY Teese *No. 1 = 2 cc. serum + 1.1 cc. of 7 per cent BaCO,.; no. 2 = 2 cc. serum+ 5.5 ec. of 7 per cent BaSOu; no. 3 = 2 cc. serum + 11 ce. of 7 per cent BaSOx. ANTILYSINS OF HUMAN SERUM 263 In a second set of experiments sera were diluted 1:10 with saline and titrated for antilysins before and after heating; 2 ce. of each serum was then diluted 1:10 with increasing strengths of barium sulphate as 7, 10 and 20 per cent suspensions in saline ; the mixtures were incubated for an hour followed by removal of the barium and antilytic tests with the supernatant, diluted serum before and after heating. The results of experiments conducted in this manner are shown TABLE 24 The absorption of antilysins from human sera by barium sulphate AMOUNT OF | UNTREATED SERUM No. 1* No. 2 NO.3 SERUM eer ee (1:10) Unheated| Heated | Unheated| Heated | Unheated| Heated Unheated| Heated cc. 0.1 C.H GHal CH (Galsl C.H C.H C.H C.H 0.2 C.H C.H C.H (Cle C.H C.H (Ona C.H 0.3 C.H C.H C.H @sHe C.H C.H (Opts! C.H 0.4 C.H C.H C.H C.H C.H C.H C.H CoE 0.5 C.H C.H C.H C.H. C.H C.H C.H C.H 0.6 VAI Get C.H Cart C.H C.H C.H C.H 0.7 S.H VeMCE.|.V7M.H: |. C2: Gal C.H C.H C.H 0.8 N.H M.H M.H. | V.M.H.| V.M.H.| C.H (Oval C.H 0.9 N.H S.H S.H M.H. S.H. | V.M.H.| V.M.H.| V.M.H 1.0 N.H S.N S.H V.M.H.| S.H. | V.M.H.| V.M.H.| V.M.H 2.0 N.H N.H N.H N.H. N.H. N.H N.H N.H 3.0 N.H N.H N.H N.H. N.H N.H N.H N.H * No. 1 = 2 ce. serum + 18 ce. of 7 per cent BaSO,; no. 2 = 2 ce. serum +18 cc. of 10 per cent BaSO,; no. 3 = 2 ec. serum + 18 ce. of 20 per cent BaSOx,. in tables 24 and 25 and they indicate that barium may remove small amounts of both thermolabil and thermostabil antilysins. 2. The influence of absorptions with kaolin, charcoal and other substances. In these experiments sera were diluted 1:10 and titrated for antilytic activity before and after heating; portions of 20 cc. were then treated with 0.5 gram of the following: Kaolin, silicon, charcoal (wood) and bone ash, previously steri- lized; mixtures were made in mortars and after incubation the diluted sera were recovered by centrifugation and filtration through paper and antilytic tests were conducted with each before and after heating. 264 TAKAATSU KYUTOKU The results of an experiment of this kind are shown in table 26; kaolin and bone ash usually removed slight amounts of both thermolabil and thermostabil antilysins but not to the same extent as barium sulphate. In additional experiments anticomplementary sera were ag vided into five portions of 2 cc. each and kaolin added to the first four in increasing amounts as 0.1, 0.2, 0.3, and 0.4 gram; after thorough mixing, these and the fifth portion (control) were heated at 56°C. for thirty minutes; the kaolin was removed, each serum TABLE 25 The absorption of antilysins from human luetic sera by barium sulphate Pare nen UNTREATED SERUM No. 1* No. 2 No. 3 tat | |_——a SS (1: 10) Unheated| Heated | Unheated| Heated | Unheated| Heated | Unheated| Heated ce. 0.1 Oe af C.H C.H C.H C.H C.H C.H C.H 0.2 C.H Co C.H C.H C.H C.H CoH C.H 0.4 C.H C.H C.H CH C.H C.H CH C.H 0.6 C.H C.H C.H CE: Cur CoH C.H C.H 0.8 V.M.H.| C.H C.H (Onlat. C.H C.H C.H C.H 1.0 MH | VeM JE) “CH C.H: C.H C.H C.H C.H 20 SoHe | VeMeHe | IC are C.H. CsH: C.H. (HEL (OnE 3.0 N.H. SiH. Cal: Opleb C.H. Olek Cari: C.H. Control | C.H. (Cals Cro: CoH CAH: CEA C.H. CHE * No. 1 = 2 cc. serum + 18 cc. of 7 per cent BaSO,; no. 2 = 2 ce. serum + 18 ec. of 10 per cent BaSO,; no. 3 = 2 ce serum + 18 ce. of 20 per cent BaSO,. was diluted 1:10 and this was followed by titrations for thermo- stabil antilysins. The results of an experiment shown in table 27 indicate that kaolin is capable of removing small amounts of antilysin; with contaminated sera it is highly probable that a part of this result is due to the removal of bacteria by the processes of absorption, centrifugation and filtration. 3. The influence of absorption with washed red blood corpuscles. Since Noguchi has found that red blood corpuscles may absorb thermostabil antilysins from dog, ox and sheep sera and thereby acquire an increased resistance to serum hemolysis, similar experi- ments have been conducted with anticomplementary human sera. 265 ANTILYSINS OF HUMAN SERUM ‘HO ‘H'O ‘HO ‘HO ‘HO ‘HO ‘HO ‘HO ‘HO ‘HO ‘HS “AN ‘HS ‘HS ‘HS EGS ‘HW ‘H’S ‘HS “HN ES ‘HS ‘HW ‘HS ‘H'S ‘HS ‘HW ‘HS ‘A'S ‘HS “HW ia Be: “HW ‘HS ‘HW ‘HS “HW ‘HW ‘HW ‘HW ‘HO “HW “HWA ‘HW “HW ‘HW ‘HWA | ‘HWA “AW ‘HW HO HO WO AWA | CHICA | CHIWA WO HO WA WIA elt] ‘HO ‘HO ‘HO zl 8 IO) eG) oH OD. ‘HO ‘HO “HWA pezeeyy pezyeequy poyeozy pozyeoquy pozwoyzy poyeoquy, pozeeyy poyeoquy pozeoyy pezyeaquy HIIM naar eae HIM Seeker HIIM er merce HIIM page auer ae RQ RS DY ST saoup}sqns snoipa fiq wnsas upuny wolf sarunisqns fiupjuawmajdwoo-Yyun fo JoAOWAL OY], 9% AIAVL Joryu0g 0's (OT: 1) wauds 40 INQOWY 266 TAKAATSU KYUTOKU Each serum was divided into portions of 0.5 cc. and diluted with 4.5 ec. of a 2.5 per cent suspension of washed sheep cells (1:10) after heating for thirty minutes at 56°C., 60°C., 70°C. and unheated. These mixtures were incubated at 37°C. for three hours and centrifuged; the supernatant fluids (sera diluted 1:10) were titrated for antilysins as was, likewise, a portion of the serum diluted 1:10 and treated as above but without corpuscles, as a control. The results of the antilytic tests showed that the red blood corpuscles had not removed the antilysins; occasionally with contaminated sera the thermostabil antilysins were removed to a TABLE 27 The removal of anticomplementary substances from human sera with kaolin NUMBER OF SERUM AMOUNT OF SERUM (1: 10) No.1 No. 2 No. 3 No. 4 Control ce. ene ay 0.1 C.H. CHA CoH: (GAEL CHEE 0.5 Care Cie. CAE. ‘Cualy CB: 1.0 Car: V.M.H. C.H. Cie: V.M.H. 15 S.H. M.H. V.M.H. C.H. S.H. 2.0 S.H. S.H. S.H. S.H. INZHE 3.0 N.H. INGHE N.H. N.H. N.H. slight extent but no more than could be accounted for by the removal of a portion of the bacteria during the process of centrifuging. The corpuscles recovered from the sera were re-suspended in steril saline solution and their resistance to serum hemolysin was determined with decreasing amounts of complement. Equal sus- pensions of untreated cells were titrated at the same time as controls. The results of these titrations showed no differences in the resistance of the corpuscles. According to these results red blood corpuscles do not absorb the thermolabil or thermostabil antilysins of human sera nor acquire increased resistance to serum hemolysis by contact with them. ANTILYSINS OF HUMAN SERUM 267 4. The influence of neutralization. As previously stated, steril and contaminated sera containing thermolabil antilysin alone or in conjunction with thermostabil antilysins, gradually became more alkaline in reaction to a point where further changes apparently cease. In a series of experiments large volumes of anticomplementary sera were titrated for the degree of alkalinity and hydrochloric acid added to the neutral point. Titrations of such sera heated and unheated before and after neutralization, showed no differences in antilytic titers. In additional experiments anticomplementary sera were treated with varying amounts of a 4 per cent solution of boric acid in physiological saline solution, but without influencing the antilytic titers of the sera either before or after heating. 5. The influence of filtration. As shown by Muir and Brown- ing (9), filtration of active serum through porcelain removes hemolytic complement; we have been able to corroborate this observation; we have found that active guinea-pig serum passed through a small chemically clean and sterile Kitasato filter is ren- dered inactive by removal of all traces of hemolytic complement. Experiments conducted with human sera containing thermolabil and thermostabil antilysins have shown that filtration of a 1:10 dilution of serum through these small earthen filters effectually removes all of the antilysins In conducting these experiments human sera were diluted 1:10 with physiological saline solution and a portion titrated for antilysins before and after heating; 10 cc. of the remaining por- tions were passed through the filters by suction and the tests were repeated before and after heating. The results of one experiment with four sera are shown in tables 28 and 29; in table 28 are shown the results with the sera before filtration and in table 29 the results after filtration. As shown in these tables filtration through chemically clean and steril Kitasato filters removes thermostabil and thermolabil. antilysins from sera diluted 1:10; with filters used more than twice removal is incomplete as likewise with undiluted sera, 268 TAKAATSU KYUTOKU owing to the small size of the filters and progressive deterioration following their frequent use. With sera containing bacteria and thermostabil antilysins filtration of diluted sera through paper and prolonged centri- fuging also tends to remove small amounts of antilysin, probably TABLE 28 The anticomplementary substances in human serum before filtration No. I No. II No. III No. IV AMOUNT OF SERUM (1:10) Unheated| Heated | Unheated| Heated | Unheated| Heated | Unheated| Heated 300 | Naa) wee wee wee! | Sn | ene ee eee Geol} GH. | SCR. || Cab ol 9G. Hes | @ad WuGlEL a |) Glee eae TABLE 29 The removal of anticomplementary substances by filtration through Kitasato filters No. I No. II No. III No. IV AMOUNT OF SERUM (1:10) | Unheated| Heated | Unheated| Heated Unhcsted| Heated |\Unbeated| “Heated 8 ol ae alae onounr 3.0 Control by the removal of a portion of the bacteria, as shown in table 30; filtration through the Kitasato filters, however, effectually re- moves all antilysin under the conditions mentioned above. Of further interest in this connection are the results of addi- tional experiments with emulsions of Staphylococcus aureas, which Sas ee SS gt See = PIS ANTILYSINS OF HUMAN SERUM 269 have shown that filtration removes in large part or entirely the antilytic substances (table 31); however, alcoholic extracts of tissues reinforced with cholesterin used as antigens in the Wasser- mann test, diluted with saline solution and filtered show no re- duction in the antilytic titers (table 32) although the antigenic properties are entirely removed (table 33). TABLE 30 The removal of antilysins by filtration through Kitasato filters, centrifuging and paper filtration PORCELAIN FILTERED| CENTRIFUGALIZED PAPER FILTERED amount or | ONFUTERED pene SERUM SERUM SERUM SERUM se Rh ek 8 | ee (1: 10) Unheated | Heated |Unheated | Heated |Unheated | Heated |Unheated | Heated TABLE 31 The influence of filtration wpon the antilytic activity of a broth culture of staphylococcus aureus AMOUNT OF BROTH (UNDILUTED) BROTH SE 0.1 0.2 0.4 0.6 0.8 1.0 2.0 3.0 Control cc cc ce ce ce. ce ce cc Unfiltered Broth... ... V.M.H.| M.H.| M.H.| M.H.| M.H.| M.H.| S8.H. foals Eee Ord a i Filtered broth............ C.H.. |. CH... | CH. |: CAA Wesel Vi M.A) Vi. MoE @uet Filtration of syphilitic sera effects the complete removal of thermolabil and thermostabil antilysins with practically no in- fluence upon the syphilitic antibody concerned in the Wassermann test, if account is taken of the effect upon the test following the removal of the antilysins (tables 34 and 35). TABLE 32 The influence of filtration on the antilysin in an alcoholic extract of beef heart reenforced with cholesterin aateceh tes UNFILTERED FILTERED ANTIGEN ———S a ee (1:5) Unheated Heated Unheated Heated ce. 0.1 Cl: C.H @sHi C.H 0.2 C.H. CH C.H (Gulal 0.4 Coe C.H C.H C.H 0.6 C.H. CEL C.H C.H 0.8 C:H. (Ofle le C.H C.H 1.0 CoH: C.H. (ORIEL C.H 2.0 V.M.H V.M.H. CoH V.M.H. 3. N.H. N.H. V.M.H. V.M.H. Control C.H. Ce (UE: CoH: TABLE 33 The influence of filtration upon the antilytic sensitiveness of an alcoholic extract of beef heart reenforced with cholesterin UNFILTERED FILTERED ANTIGEN (Cc. B.H.) 1:10 ; Unheated Heated Unheated Heated cc, S| Et Catre rns ~ eee 0.05 +2 +2 — — 0.1 +3 +4 _ —- 0.15 +4 +4 - — 0.2 +4 +4 _- — 0.25 +4 +4 —_ _ 0.3 +4 +4 _ - Serum/controly.....0025e2-n- _ _ — Hemolytic control........... _ - — - TABLE 34 The influence of filtration on the anticomplementary substances in human serum UNFILTERED FILTERED are aaa SERUM L ‘ : Unheated Heated Unheated Heated cc. hla... ..latccdwcha an. ctl ubieiieiieie Gl sei Opal C3 C.H C.H C.H 0.5 C.H CH. C.H C.H 1.0 V.M.H V.M.H. C.H C.H 2.0 M.H. M.H. CH. @sEF 3.0 N.H. N.H. CoH. CH Control C.H. Cane (Gps CG lHe ANTILYSINS OF HUMAN SERUM pat i} Additional experiments with dog sera have shown that filtra- tion removes the substance that is responsible for the non-specific complement fixation reactions described by Kolmer and his associates (10). Heating dog serum at 56°C. for thirty minutes tends to increase its antilytic titer and the power of fixing com- plement in the presence of lipoidal and bacterial antigens; as TABLE 35 The influence of filtration upon the Wassermann antibody in human serum UNFILTERED FILTERED AMOUNT OF SERUM (1:10) a Unheated Heated Unheated Heated cc. 0.1 El — - _ 0.2 +3 +2 — _ 0.4 -P4: oie +1 _ 0.6 +4 +4 +1 +1 0.8 “ind 4: 4 +2 0 =-4 Ses +4 +2 2.0 Se +4 +4 +3 Serum control,.2.0.:....1... +3 +3 - — Antigen control............. - — = _ Hemolytic control........... _ _ — — es ee I Tene bl TABLE 36 The influence of filtration on non-specific complement fixation by heated dog serum se. Ue bog serum | aNricnN COMMEMEN? | aeaorvenn ||| coneqaeles ira (1: 10) (1: 10) ; UNIT 2.5 PER CENT.) Unfiltered Filtered ce. ce, ce, cc, 0.1 0.1 1 2 1 G-H: CoH: 0.2 0.1 1 2 1 C.H. Cali; 0.4 0.1 1 2 1 CAE C.H. 0.6 0.1 1 2 1 Os i C.H. 0.8 0.1 1 2 if V.M.H. CH. 1.0 0.1 1 2 1 M.H. CH. 2.0 0.1 1 2 1 M.H. CH. S.c.* 2.0 0 1 2 C.H C.H. A.C. 0.1 if 2 C-H, C.H. EC; 0 1 2 | 1 On cE C:H. ee * S.C. = serum control; A.C. = antigen control; H.C. = hemolytic control. THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 5 DAS fe) TAKAATSU KYUTOKU shown in table 36, filtration removes.the substance from dog sera that is capable of yielding non-specific complement fixation reactions with bacterial antigens. SUMMARY AND CONCLUSIONS 1. Anticomplementary substances (antilysins) in human sera may be divided into two kinds: (a) those removed by heating at 56°C. (thermolabil antilysin) and (b) those not removed by heating (thermostabil antilysin). 2. Human sera do not develop antilysins as a result of heating, as may occur with rabbit, dog and mule sera. 3. Steril sera develop thermolabil but not thermostabil anti- lysins. Steril sera kept at 37°C. may develop these thermolabil antilysins in from three to seven days; at room and lower tem- peratures (0 to 2°C.), longer periods are required. 4. Fresh sera rapidly dried in filter paper do not develop anticomplementary properties. 5. Human sera containing various bacteria and particularly staphylococci rapidly develop thermolabil and thermostabil antilysins. 6. Staphylococci alone rather than their products elaborated during cultivation in broth and serum, produce the antilytic effects of contaminated serum. 7. Large amounts of hemoglobin in salt solution and serum, steril and contaminated, exert anticomplementary activities before and after heating. Traces of hemoglobin are usually free of these antilytic effects. 8. Steril and contaminated sera containing antilysins gradually become alkaline in reaction; neutralization of this alkali with hydrochloric acid does not remove the antilytic activities of the sera. 9. Definite changes in the hydrogen ion concentration of steril and contaminated human serum before and after the develop- ment of antilysins, could not be determined with colorimetric methods. 10. Steril human sera containing thermolabil antilysin showed a slight increase of total protein and especially of the globulin ANTILYSINS OF HUMAN SERUM 273 fraction, as determined by the refractometric method of Robert- son; contaminated sera containing thermolabil and thermo- stabil antilysins showed a marked increase of total protein and especially of the albumin fraction, probably due in part to the presence of bacteria. Anticomplementary sera passed through a Kitasato filter which removes the antilysins, showed a reduction in total protein and especially of the globulin fraction. Heating anticomplementary sera at 56°C. for thirty minutes, which re- duced the content of antilysins, had no appreciable influence upon the protein constituents. 11. The removal of ether soluble lipoids from anticomple- mentary human sera did not remove the thermolabil and thermo- stabil antilysins. 12. The antilysins of human sera are closely allied with the protein constituents and especially the globulin fraction. 13. Absorption of human sera with barium sulphate tends to remove a portion of thermolabil and thermostabil antilysins; kaolin, bone ash and charcoal also remove antilysins, but to a lesser degree. 14. Absorption of human sera with washed erythrocytes does not remove thermolabil or thermostabil antilysins as may occur with the thermostabil antilysins of dog serum; erythrocytes ex- posed to the antilysins of human serum do not acquire an in- creased resistance to serum hemolysis. 15. Filtration of diluted human sera through new, chemically clean and steril Katasato filters removes all thermolabil and thermostabil antilysin; likewise the filtration of diluted heated dog serum removes the substances responsible for non-specific complement fixation reactions. Filtration of the sera of luetic persons has practically no influence upon the antibody concerned in the Wassermann reaction. 16. Filtration of freshly prepared bacterial antigens removes the antilysins; filtrations of antigens for the Wassermann test as alcoholic extracts of beef heart reinforced with cholesterin diluted 1:20 with saline solution, does not remove the antilysin but com- pletely removes the antigenic substance. 274 TAKAATSU KYUTOKU I beg to express my appreciation of the kindness of Professor Kolmer for outlining the experiments and technic of this inves- tigation and for his supervision of the work; also to Dr. Charles Weiss, for aid in conducting the hydrogen ion determinations, and to Dr. Hatai and Dr. Toyama for aid in the refractometric work. REFERENCES (1) Noaucut, H.: The thermostable anticomplementary constituents of the blood. Jour. Exper. Med., 1906, 8, 726-749. (2) Zinsser, H., anp Jonson, W. C.: On heat sensitive anticomplementary bodies in human blood serum. Jour. Exper. Med., 1911, 13, 31-42. (3) Koutmer, J. A.: The effect of heat on normal rabbit and dog sera in relation to antilytic and non specific complement fixation reactions. Jour. Infect. Dis., 1916, 18, 64-87. (4) Koumer, J. A.: The relation of serum lipoids and proteins to non specific complement fixation with normal rabbit and dog sera. Jour. Infect. Dis., 1916, 18, 46-63. (5) Crata, C. F.: The relation of certain bacteria to non specific fixation with the complement fixation test for lues. Jour. Exper. Med., 1911, 13, 521. (6) Brown, C. P., anp Koutmer, J. A.: The preparation of glassware and saline solution for the Wassermann reaction. The influence of acids, alkali and other factors. Amer. Jour. Syph., 1919, 3, No. 1. (7) BRONFENBRENNER, J.: A new indicator for direct reading of hydrogen ion concentration in growing bacterial cultures. Jour. Med. Research, 1918, 39, 25-33. (8) Rosrertson, T. B.: A micro-refractometric method of determining the per- centages of globulin and albumin in very small quantities of blood serum. Jour. Biol. Chem., 1915, 22, 233. (9) Murr, R., AND Brownina, C. H.: On the filtration of serum complement. Jour. Path. and Bacteriology, 1909, 13, 262. (10) Koutmer, J. A., Trest, M., anp Heist, G. D.: Non specific complement fixation by normal dog serum. Jour. Infect. Dis., 1916, 18, 27-31. EXPERIMENTS ON THE REMOVAL OF HEMAG- GLUTININ FROM RABBIT ANTIHUMAN SERUM JOSEPH E. SANDS anp LYLE B. WEST From the laboratory of experimental pathology of the University of Pennsylvania Received for publication June 16, 1919 In conducting the complement fixation test for syphilis with an antihuman hemolytic system, agglutination of the erythro- eytes not infrequently interferes with hemolysis and the fre- quency and intensity of agglutination constitutes a drawback in the usefulness of the antihuman system. One of the sources of this agglutinin is the immune rabbit hemolytic serum and at the suggestion of Professor Kolmer, we have conducted a series of experiments with rabbit immune sera with the object of determining whether or not agglutinin for human cells may be removed without disturbing the hemol- ysin content. Professor Kolmer has been of the opinion that drying antihuman rabbit serum on paper after the method of Noguchi, resulted not only in facilitating the preservation and manipulation of the hemolysin but also in the deterioration of agglutinin without commensurate destruction of hemolysin and that this constituted an important reason for the use of paper amboceptor in complement fixation tests when an antihuman hemolytic system is employed. Our experiments have supported this view and in addition they have shown that filtration also tends to remove the agglutinin from a hemolytic serum consti- tuting an observation of considerable theoretical and possibly of some practical value. TECHNIC In conducting hemolysin titrations one cubic centimeter of dilutions of heated rabbit hemolytic serum varying from 1: 10 to 1: 120 were placed in a series of test tubes; 0.1 ce. of a 1:10 275 276 JOSEPH E. SANDS AND LYLE B. WEST dilution of guinea-pig complement, 0.5 cc. of a 1 per cent suspension of washed human cells and 0.4 cc. of salt solution were added and incubation was conducted in a water bath at 38°C. for one hour, the readings being made after the mixtures had stood over night in a refrigerator. The agglutination tests were conducted by placing in a series of test tubes 1 cc. of dilutions of heated rabbit hemolytic serum varying from 1:10 to 1: 960 together with one cubic centimeter of a one per cent suspension of washed human cells; incubation took place in a water bath at 38°C. for one hour and the readings were made the following morning after the tubes had stood in a refrigerator. In presenting the results the titer of each serum is given as the actual amount of serum as such or as calculated with a solution of dried serum, producing agglutination or hemolysis -in a constant volume of 2 cc. and in the time specified above. A number of different rabbit hemolytic sera were prepared by immunizing with washed human cells after various methods’ (1) and employed in conducting these experiments. THE INFLUENCE OF DRYING UPON HEMAGGLUTININ AND HEMOLYSIN These experiments were carried out with two methods: (a) Measured amounts of each immune serum were rapidly dried by fanning with cold air in a measured amount of Schleicher and Schull’s paper no. 597 and the amount of serum per square millimeter of paper was estimated; hemolysin and agglutinin titrations were then conducted with increasing amounts of paper and the results were compared with tests conducted with vary- ing amounts of fluid serum. (b) In the second method, measured amounts of serum were rapidly dried by fanning with cold air in weighed dishes and the amount of dried product was calculated per cubic centimeter of serum. This amount was then weighed and dissolved in 10 cc. of physiological salt solution yielding opalescent solutions of dried serum equal, or very nearly so, to a 1:10 dilution of HEMAGGLUTININ FROM ANTIHUMAN RABBIT SERUM 277 fluid serum secured by diluting 1 cc. of serum with 9 ce. of physi- ological salt solution. Comparative hemolysin and agglutination tests were then conducted at the same time and with the same corpuscles and complement. The results of a series of these experiments are summarized in tables 1 and 2. TABLE 1 The influence of drying rabbit antihuman serum in filter paper upon agglutinin and hemolysin INFLUENCE ON AGGLUTININ INFLUENCE ON HEMOLYSIN BERA : 3 Titer of serum after Titer of serum after Titer of serum Titer of serum drying drying 1 0.0013 0.0015 0.016 0.015 2 0.0033 0.0144 0.04 0.01 3 0.0041 0.006 0.06 0.015 4 0.00125 0.003 0.0166 0.013 5 0.00083 0.001 0.01 0.008 6 0.008 0.036 0.016 0.009 7 0.002 0.007 0.01 0.01 8 0.0125 0.02 0.01 0.009 TABLE 2 The influence of drying rabbit antihuman serum in dishes upon agglutinin and hemolysin INFLUENCE ON AGGLUTININ INFLUENCE ON HEMOLYSIN SERA EES ee eS as ee eee eee ES Titer of serum Titer after drying Titer of sernm Titer after drying 1 0.0013 0.02 0.016 0.016 2 0.0033 0.015 0.04 0.04 3 0.0041 9.035 0.06 0 4 0.00125 0.006 0.0166 0.016 5 0.00083 0.003 0.01 0.01 6 0.008 0.05 0.016 0.01 7 0.002 0.003 0.01 0.009 8 0.0125 0.01 0.01 0.0085 As shown in tables 1 and 2, drying sera in filter paper and in dishes usually resulted in a well defined reduction of the agglu- tinins; the hemolysins showed no appreciable deterioration but rather an enhanced hemolytic activity, which we ascribed to the removal of the inhibiting influence of the agglutinins. 278 JOSEPH E. SANDS AND LYLE B. WEST THE INFLUENCE OF FILTRATION UPON HEMAGGLUTININ AND HEMOLYSIN These experiments were conducted by filtering 5 to 10 ce. of each immune serum undiluted and diluted 1: 10 with physio- logical salt solution through Kitasato and Chamberland candle filters and comparing the agglutinin and hemolysin content with unfiltered serum tested at the same time with the same indicator antigen and complements. The results of a number of experiments are given in table 3. As shown in table 3, filtration through Kitasato and Chamber- land candles usually removed a large amount of hemagglutinin TABLE 3 The influence of filtration of rabbit antihuman sera wpon agglutinin and hemolysin INFLUENCE ON AGGLUTININ | INFLUENCE ON HEMOLYSIN SERA FILTER SS a ee nn Before After Before After filtration filtration filtration filtration 1 (1: 10) Kitasato 0.0013 0.006 0.016 0.016 De Gtr=10)) Kitasato 0.0033 0.01 0.04 0.03 SECU 10) Kitasato 0.0041 0.0125 0.06 0.025 4 (1: 10) Kitasato 0.00125 0.033 0.0166 0.008 5 (1240) Kitasato 0.00083 0.0125 0.01 0.03 6 (1: 10) Kitasato 0.002 0.006 0.01 0.008 7 (undiluted) Kitasato 0.002 0.003 0.01 0.01 8 (1: 10) Chamberland 0.0125 0.005 0.01 0.008 vithout appreciable influence upon the hemolysins. Best results were observed with sera diluted 1:10 with physiological salt solution and with new filters cleansed with sterile distilled water and sterilized. Filters used more than three times with proper cleansing and burning in a blast lamp between filtrations became so porous that slight or no influence was exerted upon the filtered sera. Not infrequently the hemolytic titer of a serum was increased as a result of filtration and we have ascribed this to the removal of the inhibiting influence of agglutinin. The first portion of the filtrate always showed a greater reduc- tion of agglutinin than following portions; serum 5 for example, HEMAGGLUTININ FROM ANTIHUMAN RABBIT SERUM 279 showed an agglutinin titer of 1: 1200 before filtration, 1: 160 with the first 5 ce. of filtered serum and 1: 400 with the second 5 ec. When 10 ce. of salt solution were then drawn through the empty filter some of the agglutinin was withdrawn, inasmuch as this salt solution showed an agglutinin titer of 1: 200 and a trace of hemolysin. THE INFLUENCE OF BARIUM SULPHATE ABSORPTION UPON HEMAGGLUTININ AND HEMOLYSIN Wechselmann and Lange having shown that absorption of syphilitic sera with barium sulphate tends to remove certain constituents and increasing the degree of complement fixation in the Wassermann reaction, we have conducted additional ex- periments upon the influence of barium sulphate upon rabbit antihuman sera; Noguchi and Bronfenbrenner (2) have shown that barium sulphate tends to remove natural antisheep hemol- ysin from human sera and may in this manner increase the delicacy of the Wassermann reaction as shown by Wechselmann and Lange. Our experiments were conducted by adding to 1 ec. of immune serum increasing amounts of a seven per cent suspension of barium sulphate and incubating for one hour with frequent shakings. Physiological salt solution was then added to make the final dilution of serum 1:20 and each mixture centrifuged and filtered through paper to remove the barium. The exact proportions were as follows: lec. serum plus 1.1 cc. of 7 per cent barium plus 17.9 cc. salt solution lee. serum plus 5.5 ce. of 7 per cent barium plus 13.5 cc. salt solution 1 ce. serum plus 11.0 ee. of 7 per cent barium plus 8.0 ce. salt solution 1 ce. serum plus 19.0 ce. of 7 per cent barium Agglutinin and hemolysin titrations were then conducted with several immune sera before and after absorption with barium sulphate; the results are given in table 4. As shown in table 4, barium sulphate removes from rabbit antihuman serum varying amounts of hemagglutinin; 1.1 cc. of 280 JOSEPH E. SANDS AND LYLE B. WEST a 7 per cent suspension.of this substance removes from 1 ce. immune serum almost as much agglutinin as the larger amounts. With two sera the hemolytic titers were appreciably increased as the result of absorption with barium; with a third serum (no. 6) a decrease in hemolytic activity was observed due pre- sumably to removal of hemolysin. We are unable to explain the increase of hemolytic activity of sera 2 and 3; the suspension of barium itself was not hemolytic as tested by centrifuging and filtering a portion of the 7 per cent suspension used and testing the filtrate with the same suspensions of washed human cells employed in these tests. TABLE 4 The influence of barium sulphate absorption of antihuman rabbit serum upon hemagglutinin and hemolysin | UNTREATED 1.1 cc. OF BARIUM | 5.5 CC. OF BARIUM | 11 Cc. OF BARIUM |19 CC. OF BARIUM Aggluti- | Hemoly-| Aggluti- | Hemoly-| Aggluti- | Hemoly-| Aggluti- |Hemo- |Agglu- |Hemo- nin sin nin sin nin sin nin lysin | tinin | lysin DP SOOn 25 a) 180) E20) | el SOM eic4 Olas ett SO Np elie SO iat Ss 240 216) P80) | 50 SO 50) GOs She 5On alr Ge eON 60) W605) 2a ol GO mle | lero ay alc 2On ae THE INFLUENCE OF ABSORPTION BY HUMAN CELLS UPON HEMOLYSIN AND AGGLUTININ ~ It is well known that human cells added to antihuman immune serum will absorb the specific hemolysin and agglutinin, but inasmuch as agglutination usually appears earlier than hemol- ysis in complement fixation tests conducted with an antihuman hemolytic system, we have conducted several experiments by adding to 1 cc. of heated immune serum 1 cc. of washed packed human cells and centrifuging after incubation in a water bath at 55°C. for fifteen minutes to determine what proportion of hemagglutinin and hemolysin were removed from the serum under these conditions. The results of experiments with three sera are shown in table 5. HEMAGGLUTININ FROM ANTIHUMAN RABBIT SERUM 281 As shown in table 5 absorption with human cells removed agglutinin and hemolysin but by adding a large volume of cells and removing them after a short exposure in the serum as con- ducted in our experiments, a proportionately larger amount of agglutinin was apparently removed. TABLE 5 The influence of absorption with human erythrocytes of rabbit antihuman sera upon hemagglutinin and hemolysin BEFORE ABSORPTION AFTER ABSORPTION SERUM Agglutination Hemolysis Agglutination Hemolysis iets 1: 60 1: 50 1:10 2 1: 300 1: 25 45) 15 3 1: 240 220 1:25 1: 20 THE INFLUENCE OF HYPERTONIC SOLUTION OF SODIUM CHLORID UPON THE HEMAGGLUTININS AND HEMOLYSINS IN RABBIT ANTIHUMAN SERUM Owing to the well known influence exerted by varying con- centrations of sodium chlorid upon the agglutination of bacteria and also upon opsonins and the phenomenon of phagocytosis, we have also conducted a series of experiments with three im- mune sera for the purpose of determining the influence of hyper- tonic solutions of sodium chlorid upon the hemagglutinins and hemolysins. Solutions of chemically pure solium chlorid in distilled water were prepared in concentrations varying from 0.8 to 5 per cent and 10 ce. of each placed in a series of test tubes; to each tube was added 0.1 cc. of washed packed human cells giving approxi- mately a one per cent suspension in each. Agglutination tests were conducted by placing 1 ce. of each suspension in corresponding test tubes and adding 1 cc. of a dilution of immune serum in distilled water known to contain two units of agglutinin, which doubled each dilution of sodium chlorid. After mixing and incubating for one hour in a water 282 JOSEPH E. SANDS AND LYLE B. WEST bath at 38°C. the results were read after standing in a refrigerator over night. Hemolysin tests were conducted by placing 1 ec. of each corpuscle suspension in corresponding tubes and adding 0.5 ce. of a dilution of immune serum in distilled water known to contain two units of hemolysin, 0.1 cc. of a 1:10 dilution of guinea-pig serum complement in water and 0.4 ce. of distilled water, which resulted in doubling each solution of sodium chlorid. The results were read after incubation in a water bath for one hour followed by refrigeration over night. The results of several experiments are shown in table 6. TABLE 6 The influence of hypertonic solutions of sodium chlorid upon hemagglutinins RESULTS OF AGGLUTINATION TESTS IN THE PRESENCE OF INCREASING PERCENTAGES OF w SODIUM CHLORID AMOUNT OF SERUM USED UNIT OF SERUM* e | SIS ESSE IS ESI ESS ES =) pss Vie ised cd inset Peco) C= Sa Ve Sa) bp SI ft Cos pee! |S m | mm mL LN ORIN NIN 1 | dec. of 1:750 | 1 cc. of 12350 [fH e 4 lel2|2 |e 2 | Lee. of 1:300 | 1c. of 1:150 [+ |+)4]=|+|=|=|+/=|£/=|+|+]=/= 3 | 1 ee. of 1: 240 lec. of 1:120 J Jp] pelle lel] ep el eye f=] =]— * Titrated in the presence of 0.85 per cent sodium chlorid. 7 + = agglutination, = partial agglutination. As shown in ‘table 6, concentrations of sodium chlorid in a final dilution of 1.6 per cent and higher tend to protect human erythrocytes against hemagglutinin, but do not entirely prevent agglutination in solutions of as high as 2.5 per cent unless the amount of agglutinin in the serum is relatively small, as in serum 3. As shown in table 7 hemolysis is interfered with by final dilutions of sodium chlorid ranging from 1.7 or 2 per cent and higher; for this reason hypertonic solutions of sodium chlorid are of slight practical value for reducing the influence of hemag- glutinins in complement fixation tests in which an antihuman hemolytic system is employed. yO ORE OE GaSe ~g HEMAGGLUTININ FROM ANTIHUMAN RABBIT SERUM 283 TABLE 7 The influence of hypertonic solutions of sodium chlorid wpon hemolysis ayes RESULTS OF HEMOLYSIN TESTS IN THE PRESENCE OF INCREASING oO aaa aan Pain co PERCENTAGES OF SODIUM CHLORID OF SERUM pe 1% 1.1% 1.2% 1.3% 1.4% 1.5% 1.6% 1.7% Tee ce. cc. evn a Taran ected ; 1 0.016 0.03 (Opel OAel Oe (Ost Osi Cael |) Omen | Wis 2 0.04 0.08 Cab Oa shal Ose Cael |(CRst Char (Cust | (Mat 3 0.04 0.08 CoH |\C Et | R@ EE | iCsEe | i@ae @sEs | k@zHe |p Men RESULTS OF HEMOLYSIN TESTS IN THE PRESENCE OF INCREASING en UNIT Si cea PERCENTAGES OF SODIUM CHLORID OF SERUM* TaD 1.8% 1.9% 2% 2.1% 2.2% 2.3% 2.5% a ce, cc, 5 aia ae (a ALA AOR (| eed | gl 1 0.016 0.03 See NG EE ENE NEED ENED SNE NEL 2 0.04 0.08 Misael Wiel IN Els IN Sl Waal | Nfs 3 0.04 0.08 SH jeNEe | INGER | NEE NEES ISNeEO | NeGEL * Titrated in the presence of 0.85 per cent sodium chlorid. 7 C.H = complete hemolysis; M.H = marked hemolysis; S.H = slight hemolysis; N.H = no hemolysis. CONCLUSIONS 1. Owing to the tendency for retardation or prevention of hemolysis of human erythrocytes by rabbit antihuman serum, due to the presence of hemagglutinin for these cells, experiments were conducted for the purpose of determining whether practical methods for the removal of the hemagglutinin could be devised. 2. Drying serum upon filter paper by the method of Noguchi, or in evaporating dishes was found to result in distinct destruction of hemagglutinin with slight or no destruction of specific hemol- ysin. The use of paper amboceptor therefore in complement fixation tests employing the antihuman hemolytic system is advisable for this reason in addition to being a satisfactory method for the preservation and manipulation of this constituent. 3. Filtration of immune sera and especially 1:10 dilutions through satisfactory Kitasato and Chamberland filters removes a large amount of hemagglutinin with slight or no depreciation in hemolytic activity; hemolytic activity may be increased, due presumably, to the removal of the hemagglutinins. 284 JOSEPH E. SANDS AND LYLE B. WEST 4, Absorption of rabbit antihuman serum with barium sul- phate removed large amounts of hemagglutinin; the removal of specific hemolysin was irregular and occasionally the hemolytic activity of the treated serum was increased. 5. Absorption of rabbit antihuman serum with a large volume of washed human erythrocytes for a short period tends to remove a proportionately larger amount of hemagglutinin than hemolysin. 6. Solutions of sodium chlorid in final dilution of 1.5 per cent and higher tend to prevent hemagglutination; dilutions of 1.7 or 2 per cent and higher tend to interfere with hemolysis so that the use of hypertonic solutions of sodium chlorid are of no practical value in preventing hemagglutination in comple- ment fixation tests in which an antihuman hemolytic system is employed. We wish to express our sincere thanks to Professor Kolmer for his advice and directions and to Miss Anna Rule for aid in conducting some of the experiments. REFERENCES (1) Koumemr, J. A., anp Ruts, A.: A study of methods for the preparation and preservation of hemolysins. Studies in the standardization of the Wassermann reaction, VIII. Amer. Jour. Syphilis (in press). (2) Noeucut, H., aNnp BRONFENBRENNER, J.: Barium sulphate absorption and the serum diagnosis of syphilis. Jour. Exper. Med., 1911, 18, 217-228. A NEW METHOD OF TESTING ANTITYPHOID SERUM YOSHIMOTO FUKUHARA anp MASAAKI YOSHIOKA From the Pathological-Bacteriological Institute in Osaka, Japan (Director: Professor A. Sata) Received for publication June 28, 1919 It is well known that no antibodies have heretofore been dis- covered in antibacterial (antiinfectious) sera, whose quantitative determination could be used as a measure of therapeutic value. Nevertheless, practical experience has shown that there are a number of antiinfectious sera which, notwithstanding their variable content in one or another of the known antibodies, are practically useful and can be quantitatively titrated with the use of animals. This titration can be carried out by the prelim- inary protective injection of the serum followed immediately or later by the injection of living bacteria. However, since the time of Pfeiffer it has been the custom to inject the serum in varying quantities combined with ten lethal doses of a living virulent culture of the bacteria. The injections are made into the peritoneal cavity. The quantity of serum which was just able to protect the animal against ten lethal doses of bacteria was designated by Pfeiffer as the titer or immunity unit of the serum. For this test in typhoid it has been customary to use strains of such virulence that a fifth to a tenth of one oese could kill a guinea-pig in twenty-four hours after peritoneal injection. The varying virulence of the cultures offered a source of difficulty in making the test. As to the cause of this variability we shall speak later. The protocols of an experiment presented in tables la, 1b and ic demonstrate the disadvantages of the usual method of titrat- ing an antibacterial serum. 285 286 YOSHIMOTO FUKUHARA AND MASAAKT YOSHIOKA In carrying out this test we chose the procedure given by Pieiffer for testing anti-cholera serum. One cubic centimeter of the diluted immune serum of goat A was stirred in a test-tube with ten lethal doses of three different living virulent strains of typhoid organisms (Monoyama 2, Monoyama 3 and Takayama PI) from an eighteen to twenty hour agar culture and this suspen- sion was injected intraperitoneally into a guinea-pig of about 250 grams weight. TABLE 1a The testing of goat’s serum A with strain Momoyama 2; lethal dose, 0.125 oese SERUM DILUTION GUINEA-PIG WEIGHT RESULTS grams 230 Lived ; 230 Lived Bie 260 Lived 245 Died after 15 hours 1:22 230 Died after 15 hours ; 230 Lived 1:24 250 Died after 15 hours i 240 Died after 15 hours 1:27 235 Died after 17 hours ; \ 230 Died after 17 hours 1:32 230 Died after 17 hours : 230 Died after 17 hours It is seen that the titer of the goat’s serum is about 1-20 with strain Monoyama 2; 1-25 with the strain Takayama PI and 1-55 with the strain Monoyama 3. In other words, 0.04 ce. of the immune serum protects against ten lethal doses of strain Takayama PI; 0.018 ec. protects against ten lethal doses of strain Monoyama 3 and 0.05 ec. protects against ten doses of Monoyama 2. The method of testing antityphoid sera that is illustrated in the foregoing table is obviously unreliable, since the result of the determination of protective value depends, here, on the strain of the bacteria employed in the test. TESTING ANTITYPHOID SERUM 287 The experiment demonstrates clearly the need of a constant standard of protective value, which could only be established with the use of a constant unit of the infectious material; that is, of the typhoid culture. It occurred to us that the principle TABLE Is The testing of goat’s serum A with strain Takayama PI; lethal dose, 0.11 oese SERUM DILUTION GUINEA-PIG WEIGHT RESULTS grams 230 Lived 230 Lived ee 235 Lived 235 Lived 240 Died after 18 hours i 250 Lived weal 235 ie 280 Lived 250 Lived 1:28 235 Lived 245 Died after 18 hours 260 Lived 1:30 255 Lived 270 Lived 260 Died after 18 hours 240 Lived Riss 250 Lived 220 Died after 18 hours 1:36 255 Lived : 230 Died after 17 hours 1:37 230 Lived ; 230 Died after 18 hours employed in the testing of antitoxic sera might be applicable, also, to the testing of antibacterial sera and, as will be presently shown, our anticipations in this regard have been realized. Our first step in applying this principle was the selection of a standard protective unit, which, as in the case of the standard THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 5 288 YOSHIMOTO FUKUHARA AND MASAAKI YOSHIOKA units of antitoxin, is necessarily an arbitrary one. The unit was arrived at in the following manner. Varying amounts of antityphoid serum 180 were mixed with ten lethal doses of an agar culture of the typhoid strain Takayama and these mixtures TABLE Ic The testing of goat’s serum A with strain Momoyama 3; lethal dose, 0.1 oese SERUM DILUTION GUINEA PIG WEIGHT RESULTS grams 1:45 240 Died after 18 hours : 240 Lived 1: 50 230 Lived 230 Lived pee 240 Lived 1:58 230 Died after 18 hours ; 245 Lived 1:60 225 Died after 18 hours ‘ 225 Died after 18 hours 240 Lived 230 Lived ies 230 aed 245 Lived 245 Lived is 235 Lived 230 Died after 18 hours 1:65 235 Died after 18 hours 230 Lived 1:66 245 Lived ‘ 250 Died after 18 hours were injected intraperitoneally into guinea-pigs. The amount of serum which just sufficed to protect the animal from death was taken as the unit and it was found, as a result of repeated tests, that 1 gram of the dry serum contained 4410 such pro- tective units. The protocols of these tests are presented in TESTING ANTITYPHOID SERUM 289 TABLE 2 Testing of the original standard serum 180 PS RESULTS SERUM DILUTION GUINEA PIG WEIGHT ie ey ee CSS grams vee 230 Lived 1:5 255 Lived 1: 1000 230 Lived 1: 2000 230 Lived 250 Lived 1: 4000 230 Died after 49 hours 240 Lived 1: 4200 250 Lived 1: 4300 250 iiveg 250 Lived 1: 4400 250 Lived 250 Lived 1: 4410 230 Lived 1: 4415 230 Died after 20 hours 230 Died after 21 hours 1: 4420 230 Died after 2 days 230 Lived 1: 4440 250 Died after 46 hours 240 Died after 21 hours 1: 4460 240 Lived 230° Lived 1: 4480 240 Died after 27 hours 240 Died after 20 hours 1: 4500 235 Died after 20 hours 230 Died after 20 hours 1: 50000 230 Died after 20 hours 290 YOSHIMOTO FUKUHARA AND MASAAKI YOSHIOKA table 2. The minimal lethal dose of the strain employed was about 2 oese. It was found advisable, in order to exclude indi- vidual differences in the experimental animals, to test the critical amounts on several animals. In measuring the amount of the bacteria we have used the bacteriometer of Rosenthal, which we have somewhat modified (bacteriometer of Fukuhara). The number of the bacteria is calculated from the graduations on the capillary tube containing the bacterial sediment. The standard antityphoid protective unit having been thus established, it was necessary to see whether, in the standard- ization of the antitoxic sera, a constant measure of protective power could be determined in the various typhoid cultures, with the use of the arbitrary standard protective unit. In investi- gating this question, we have adopted the conception of L; and the L, dose. The ly) dose we have taken to be the largest amount of the living bacteria which when mixed with the pro- tective unit and intraperitoneally injected did not cause the death of a guinea-pig of 250 to 300 grams weight; the L, dose has been considered to be the smallest amount of typhoid bacteria which when mixed with the protective unit of antiserum and injected intraperitoneally into a guinea-pig of 250 to 300 grams weight caused the death of the animal in the course of twenty-four hours. The protocols of the determination of these two limits in three different strains of typhoid bacteria are presented in table 3. It is seen that both of these values differ in the different strains as follows: STRAIN Ly Lo graduations graduations IM OMG YANN 2 ro Seg iots breccias ON ooe Sere eee ri 4.3 AKA VaIT Gd GPs eee tel ox knnlo dios bt vs, ket eon sic ae 7.2 5.0 PRA A VEINS. Gets sepia Os. oes Lose eakte eee eae 12.0 9.7 The L, and L, doses having been determined for the strain of bacteria to be employed, it remained to be seen whether these doses can be used, in turn, to determine the protective TESTING ANTITYPHOID SERUM 291 unit of an antityphoid serum. This question was pursued in a manner analogous to the method employed in testing antitoxic sera. The L, dose (or the L, dose) was mixed with different amounts of the serum to be tested and the mixtures were in- - TABLE 3 Determination of the L4 and Lo dose AMOUNT OF BACTERIA MY YAMA MOMOYAMA 2 TAKAYAMA PI IN GRADUATIONS ao Fy Fy 14.0 13.0 12.0 11.5 11.0 10.0 9.7 (SSNS @W@)s ||} Stee] | CO9000O00C0OHH Ss ae lel Iie Om sOrr SaaS ao ese) | Sexe) ©'O OS Orsi isi elee akelied o) ae | OOO eo OS" OrO Ore = Sl ey SS 00 00 exe) SPSO99990900C0O0 FAOHAH COHKAHAH eX) DP PESO PEO ES RSE I IO Or rica Gea cen. RONGSOHNWANOOABONE OHHOOOCCOHANH HAHAHAHA KAA AD The minimal lethal dose of the respective strains were: Takayama, 1.2 gradua- tions; Momoyama 2, 1.7 graduations; Takayama PI, 1.1 graduations. T indicates the death of the animal within twenty-four hours. O indicates that the animal survived the iniection. 292 YOSHIMOTO FUKUHARA AND MASAAKI YOSHIOKA jected intraperitoneally into guinea-pigs. The protocols of such a test are presented in tables 4a and 4b. It is seen that the method possesses considerable reliability. Notwithstanding the wide variation in the L, dose of the dif- ferent strains of bacteria, the determined protective value of the antisera was practically the same whichever strain was used in the test. As a result of numerous experiments we have come to the conclusion that the use of the Ly dose is not practical. We prefer, therefore, to test the antisera with the L, dose. TABLE 44 The testing of typhoid antiserum 180B with the use of the Ls dose SERUM DILUTION TAKAYAMA P MOMOYAMA 2 637 — (JX) 3 tae tae Fy A ol eles lalor Fl lloe Plier P| wo S OHO HAHAHAHA ee ee feo ie tm Nile ilar [Le ee) nel lr (2) ©) Fa ee snuatie aan weaits Legmine -aalewe Wal sas wainGeuccnaaes bo S C10 OFF Stal SS Cate Hiei eo fre iehLeel leel Tac hherlilicel Jerltssh = Teel lse [ee MOO ae lil hile WO thee Vile |) o) 2) The L+ dose of the three strains was: Takayama P, 7.3 graduations; Momo- yama 2, 7.6 graduations; 637, 55 graduations. To recapitulate, our method is as follows: with the use of an arbitrarily selected protective unit of an antityphoid serum, which is preserved in the dry state as a standard antiserum, the L, dose of a typhoid culture is determined; with this L, dose is then determined the protective unit of the serum to be tested; that is, the largest amount of the serum which, when mixed with the L+ dose of the bacteria will not prevent the death of the test animal within 24 hours. If, for instance, this amount should TESTING ANTITYPHOID SERUM 293 be found to be 0.001 ec. then, naturally, 1 cc. of the serum con- tains 1000 protective units. Although the use of the Ly) dose in testing antityphoid serum was found to be unserviceable, nevertheless, this dose can be used and the same result can be obtained with it as with the L, dose. This is shown by an experiment, the results of which are pre- sented in table 5. TABLE 48 The testing of typhoid antiserum 80B with the use of the Ly dose 8ERUM DILUTION MOMOYAMA 2 TAKAYAMA P 637 1: 18000 ADM 1: 12900 ANSAL 1: 12700 Av Ab ALO) Av ab 1: 12500 Aver ie AP a 1: 12800 AVAL fed AVM 1: 12100 Te ALeaR Anat 1: 12000 AVAL Av “ar Aa 1: 11980 fT © TO) 1: 11960 IO) T@ lee Ti 1: 11940 eS TO LO 1: 11920 TO TO ol a 1: 11900 TO TO 1: 11700 ARO) ome) 1: 11500 ATs ALO) Tae le 1: 113800 IPG 1: 11100 AUO) TO TO 1: 11000 Ata e 1: 10500 OO 1: 10000 AIO) 1: 7000 OO OO TO The L; dose of the three strains were: Momoyama 2, 3.98 graduations; Taka- yama P, 4.42 graduations; 637, 54 graduations. The standard serum was preserved in the dry state in Ehrlich tubes. When a new standard serum is to be tested, a tube of the old standard serum is opened and the contents are diluted with glycerin-water so that 1 ce. of the fluid contains exactly 10 protective units. With this diluted serum the L, dose of a typhoid bacterial culture is determined and this dose is, in turn, employed to estimate the protective value of the new standard serum. 294 YOSHIMOTO FUKUHARA AND MASAAKI YOSHIOKA In order to aid in the general adoption of a uniform standard protective unit, the advantages of which are the same as those of the antitoxin units, our institute will be glad to supply standard serum on request. In the Li dose of typhoid bacteria we possess a uniform criterion for the determination of the protective value of an antityphoid serum. Naturally, any strain of typhoid bacteria - could be used for testing. The virulence control is not necessary in every test. It would be also unnecessary to increase the virulence of the strain by animal passage, if this were possible. Even the strain 637, 4 oese of which could not always kill a guinea- TABLE 5 The testing of antityphoid serum 180B with the Lio dose of bacteria SERUM DILUTION TAKAYAMA P MOMOYAMA 2 sO Tt: eee ke) rials ag) ena or arms ear rr No) Se- tous) Sei! Cre Fira oo Sy ie) belt) i The Lo dose of the two strains was: Takayama P, 4.9 graduations; Momo- yama 2, 4.3 graduations. pig, was found to be as available for the testing of antityphoid sera as other virulent strains. This fact in itself speaks for the practical value of our method of testing antityphoid sera. It would be practically advantageous to use a culture which would maintain its properties for a considerable time. However, the typhoid cultures constantly diminish in virulence. This we have found to be especially true if the cultures are kept in the ice-box without being transplanted. The diminution of the virulence under such conditions is shown in tables 6a and 6b. Even here, however, the toxicity of the culture remains fairly constant for one to two months. In order to prevent consid- erable variations in the test doses, it seems advisable to transplant TESTING ANTITYPHOID SERUM 295 the cultures every three to six weeks and to preserve the cultures at low temperature. We have found that the falling off of the virulence of typhoid cultures, which occurs even when the cultures are regularly trans- planted, cannot be prevented by animal passage. The con- stancy of the L, dose cannot be maintained by this means. This fact is shown in a series of determinations made over a period of a year and a half, the results of which are presented in table 7. TABLE 6a Variations in the lethal and Li dose of the strain Momoyama 2 AFTER DAYS LETHAL DOSE L4 pose graduations graduations 21 1.07 1.80 40 1.00 1.80 60 1.00 2.00 100 1.22 2.60 130 1.20 About 3.00 150 1270 7.60 TABLE 68 Variations in the lethal and Li dose of the strain Takayama P AFTER DAYS LETHAL DOSE L4 DosE graduations graduations 21 0.53 3.10 40 0.57 3.30 60 0.58 3.30 100 0.68 4,20 130 0.70 4.40 Pfeiffer and his pupils and also Strong came to the conclusion, as a result of their studies, that the binding power, virulence and immunizing power of cholera cultures are parallel properties. Wassermann, Hetsch and Kutscher, Petterson, Meinicke, Jaffe and Flemming, on the other hand, were unable to find such a parallelism in cultures of typhoid and cholera. Friedberger and Moreschi and other authors are of the opinion that the partial receptors of different typhoid strains vary considerably from one 296 YOSHIMOTO FUKUHARA AND MASAAKI YOSHIOKA another and they believe that they have shown that the titer of a bactericidal serum, as determined with one typhoid strain, cannot be considered an absolute measure of its value. That this opinion is incorrect is shown by the results which we have obtained in testing sera 180B and 80B with typhoid strains of different virulence (see tables 4a and 4b). TABLE 7 Variations in the lethal and L dose of strain Momoyama 2 (effect of animal passage) DATE OF TEST LETHAL DOSE L4 pose 1917 graduations graduations November d'.¢ S228 Ga.0e0. beeen 1.07 1.80 1918 MAMUAEY MORN A see ceccrcrociersiete serteeke 1.00 2.00 MURTY gO) 7e x ute nose = aia ati ceva 1.22 2.60 (passage) IMarCni 20: oo tae Sis. nda adeisite eg ee 0.98 3.90 (passage) IV ALGEs eh Le 2h ltrs choo uae ee 0.95 (passage) lyon eras oe tee See 1.25 4.30 (passage ) PAN DUIS LO Sorte ices oon cs ears ftereaes 1.30 5.00 (passage) DEDEDE NE evakc be cnu seaceee ee 1.60 5.60 (passage) WCC DER ZO et sere eater. ties emer a at 1.00 5.50 (passage) INO VEIMDEEEAO jt te aeins.s:s.scice see « 0.96 4.90 IDecembersO eee ee oan sk oi 0 ate 2s 5.30 1919 March AG ere titie sccm oc.aces ete 1.05 (passage) March cle eect tn tea tist we dsctat 0.98 IVY 1G 8 raph Eee rte Beastie lo ceoisicios DE 0.95 The results presented in table 4 show that the simultaneous testing of the same serum.with different strains of typhoid bacilli leads to the same value. Hence, we must assume that the receptors of different typhoid strains possess no specific differences with respect to the protective antibodies and that each kind of partial antibody corresponds with a common hap- tophore group of the bacterial protoplasm. TESTING ANTITYPHOID SERUM 297 However, in the experiments described above we have shown certain differences in the various typhoid strains; namely, with respect to the L; dose. How are these differences to be ex- plained? We must assume with Pfeiffer that the highly com- plicated molecule of the bacterial protoplasm possesses receptors which seize the corresponding antibodies. In virulent bacteria we must assume an increase in both the number and the avidity of the receptors. The differences in the binding power; that is, the differences in the size of the lethal and L, doses are not explicable as due to a dissimilarity of the receptors but by differences in their number and avidity with respect to the antibodies. If the receptor apparatus of one strain of the bacteria is uni- form in it, it should be expected that the L, dose, as determined TABLE 8 Summary of the results of the tests of two antityphoid sera with the three strains of typhoid bacteria SERUM WITH TAKAYAMA P WITH MOMOYAMA 2 WITH STRAIN 637 units units units 80B 12,000 11,980 12,000 180B 17 18 lye with any standard serum would be available for the testing of all other antityphoid immune sera. That this assumption is not a mistaken one is shown in table 8, which is to be compared with table 5. It is seen that the values obtained in the two sera with the different strains present no noteworthy differences. The small variations probably represent the inevitable experimental error. SUMMARY 1. Our experiments have shown that the virulence titer of the living typhoid bacilli is very variable according to the strain and that no quantitative relationship exists between the viru- lence titer and the binding power with relation to the protective unit of antiserum. For this reason the lethal dose of the bacteria 298 YOSHIMOTO FUKUHARA AND MASAAKI YOSHIOKA is not a reliable measure of a protective unit of an antityphoid serum. A reliable test of a bacterial antiserum can be made only through the use of a standard serum. 2. Our new method of testing typhoid antisera is as follows: with the use of an arbitrary protective unit of antiserum (stand- ard serum) the L, dose of a typhoid culture is determined; with this L, dose the serum to be tested is mixed in varying quantities; the largest amount of the serum which will just permit the death of a guinea-pig of 250 grams weight within twenty-four hours after the mixture has been injected into the peritoneal cavity is taken as the protective unit. The value of an antityphoid serum is expressed by the number of such pro- tective units contained in 1 ce. of the serum. 3. Any strain of typhoid bacteria can be used for the test. The virulence control is not important. It is also unnecessary to attempt to increase the virulence of the cultures. Efforts in this direction have failed in our hands. 4. The general adoption of our method and particularly of our protective unit will naturally permit, for the first time, a quan- titative comparison of the antityphoid sera prepared with dif- ferent strains and in various laboratories. A NEW METHOD OF TESTING ANTITOXIC DYSENTERY SERUM YOSHIMOTO FUKUHARA From the Pathological-Bacteriological Institute in Osaka, Japan (Director: Professor A. Sata) Received for publication June 28, 1919 The testing of antitoxie dysentery serum can be carried out by injecting mixtures of the toxin and antitoxin and also by injecting these two reagents separately. Kraus and Doerr made use of the latter method, while Todd, Villiard and Dopter, Kolle and his co-workers, Schottelius and also Inomata used the former method with sufficiently reliable results. Kraus and Doerr carried out the test of the serum upon rab- bits of 800 to 1000 grams weight. They injected four lethal doses of the toxin into the marginal vein of one ear and, at the same time, varying quantities of the serum to be tested into the marginal vein of the opposite ear. On the basis of their ex- periments these authors thought that only such sera should be used in the treatment of human dysentery that could protect in a quantity of 0.1 ce. However, according to our experiences the separate injection method produces variable results and it is, therefore, unreliable. Inomata, working in the sero-therapeutic institute in Osaka, recommended the following mixture method. Thirty lethal doses of the fluid dysentery toxin were mixed with diminishing amounts of serum and after being kept an hour at 37°C. the mixtures were injected intravenously into rabbits. The pro- tective unit of the serum was taken as the smallest amount which would just prevent the death of the animal. My own comparative experiments have convinced me, also, that the determination of the immunizing and curative value of anti- dysenteric sera is best accomplished with the mixture method. 299 300 YOSHIMOTO FUKUHARA An inevitable requirement of a method of testing the anti- toxic power of a serum is the establishing of a uniform standard of measurement. The previous authors have used the lethal dose of the toxin as the toxic unit. However, the fluid dysentery | toxin loses its toxicity gradually, even when it is carefully pro- tected against injurious influences. Furthermore, it was not known whether the toxicity and binding power of dysentery toxin are parallel functions. However, we have found that dysentery toxin can lose its toxicity while retaining its binding power. This instability of the fluid dysentery toxin led us to adopt, as a standard of measurement, an arbitrary unit of the stable antitoxin as was done by Ehrlich in the testing of diphtheria antitoxin. This necessitated the preparation of a ‘‘standard serum,’ which should be used not only in the testing of the new antitoxic sera but also in the study of the constitution of the dysentery toxin itself. PREPARATION OF THE DYSENTERY TOXIN Not every dysentery strain is suitable for the preparation of soluble toxin. We have examined a series of 10 strains with respect to toxin production and among these we have found only one satisfactory toxin producer (strain Fujimoto) for which I am indebted to Dr. Inomata. As a culture medium we have used a mixture of a solution of pepton prepared from pigs’ stomachs with meat infusion. The meat infusion was prepared by boiling the chopped meat with acetic acid according to Hida, in order to destroy the muscle sugar. The extract was obtained, as usual, by filtration. A reaction of 5 per cent of normal NaOh beyond the neutral point with phenolphthalein gave the best results. The strongest toxin that we obtained killed 1500 gram rabbits within five days after the intravenous injection of 0.01 ce. or occasionally as little as 0.005 ec. The first toxin used in our experiments was obtained by filtration of the three weeks broth culture. The clear fluid thus obtained was preserved with toluol. The direct toxicity was measured by the amount that could kill a rabbit of 1500 grams weight by intravenous injection in four TESTING ANTITOXIC DYSENTERY SERUM 301 or five days. Larger animals require correspondingly larger amounts of the toxin. Smaller amounts than the selected lethal dose may cause the death of the animals in from eight to eleven days or they may fail to kill but produce paralysis. The use of a number of animals is advisable as this permits a more exact estimation of the toxicity. A similar advantage is secured by the use of larger series of animals in the testing of the anti- toxic sera. We shall revert to this point later. THE SELECTION OF THE STANDARD UNIT OF ANTITOXIN AND THE CORRESPONDING TOXIN UNITS; NAMELY, THE Lo AND THE Li; DOSES The standard antitoxin unit which we adopted was, naturally, an arbitrary one. We took it as the smallest amount of the serum which could just prevent the death of the animal when injected intravenously in combination with 100 minimal lethal doses (2 ec.) of the toxin that happened to be at our disposal. The protocol of the experiment in which this quantity was arrived at is shown in table 1. From the results of this experiment it was calculated that 1 gram of the dried serum contained 42 antitoxin units. The determination of the L, and L, dose had, naturally, to be carried out with the use of the above selected standard serum. The interval between these two doses is very variable. By repeated tests with the L, dose we have found that this unit of the toxin is not practically useful in the testing of antitoxic sera. The L, dose was, therefore, preferred for that purpose. The criterion in the use of the latter unit is the death of a rabbit of 1500 to 2000 grams weight within four to five days. In all of these tests, whether for the determination of the L, dose or for the testing of an antitoxic serum, it is advisable to employ a number of animals in order to avoid the difficulty of individual variations in the resistance of the animals. If an antidysentery serum is to be tested as to its antitoxic value the L.. dose of the toxin that is to be used in the test must first be determined. Table 2 presents some examples of tests conducted to this end. 302 YOSHIMOTO FUKUHARA TABLE 1 The testing of the original ‘‘standard serum’”’ SERUM DILUTION WEIGHT RESULTS grams 1: 140 1100 Paralysis; dead after 2 days 1: 100 1200 Dead after 13 days 1:80 1320 Lived ; 1780 | Paralysis; dead after 23 days 1:75 1870 Paralysis; dead after 23 days ‘ 1200 Paralysis; dead after 7 days 1:70 2100 Lived : 1350 Dead after 9 days 1:65 1730 Lived P 1350 Paralysis; dead after 6 days 1:60 1350 Paralysis; recovered ; 1640 Lived 1:55 1230 Lived ; 1620 Paralysis; dead after 4 days 1:50 1400 Lived ; 1400 Paralysis; dead after 16 days 1:45 1580 Dead after 2} days : 1680 Paralysis; dead after 9 days 1: 43 1580 Paralysis; dead after 3 days , 1470 Lived oe { 1270 1: 41 1610 Lived ; 1700 Lived ce: 1400 - | Lived 1:35 1500 Lived ; 1440 Dead after 13 days TESTING ANTITOXIC DYSENTERY SERUM Determination of the Ly and Ly. dose of dysentery toxin 9 TABLE 2 303 oo essssssssSsS— ONE UNIT OF ANTITOXIN PLUS THE AMOUNTS OF TOXIN MENTIONED BELOW ce, 0.8 0.9 1.0 Et 1.4 1.5 1 ri 1.8- 1.9 2.0 2.1 2.2 2.3 2.4 RESULTS Lived Lived Lived Lived Lived Lived Paralysis; recovered Lived Dead after 2} days Paralysis; recovered Lived Lived Dead after 24 days Paralysis; recovered Paralysis; recovered Paralysis; recovered Paralysis; dead after 5 days Paralysis; recovered Dead after 2 days Paralysis; recovered Paralysis; dead after 23 days Paralysis; dead after 7 days Paralysis; dead after 14 days Paralysis; recovered Dead after 14 days Paralysis; dead after 14 days Paralysis; dead after 3 days SSS ES NE THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 5 304 YOSHIMOTO FUKUHARA It is seen, here, that the injection of one antitoxin unit mixed with 1 cc. of toxin caused symptoms but no paralysis. With 1.01 ec. of the toxin, paralysis resulted; consequently, the I, dose was taken aslee. With 2.1 cc., 2.2 cc. and 2.3 cc. the toxic TABLE 3 The testing of anti-dysentery serum 237 with dried dysentery toxin B (L dose 0.39) SERUM DILUTED WEIGHT RESULTS grams 1: 125 1550 Lived 1: 150 1720 Lived 5 1600 Lived 1: 200 1470 Lived 1:225 1440 Paralysis; dead after 23 days ; 1540 Lived : 1390 Lived ; oo 1350 Dead after 24 days 1:275 1720 Paralysis; dead after 5 days ; 1430 Lived eee 1810 Lived 1790 Dead after 3} days 1: 300 1560 Dead after 8} days ; 1850 Paralysis; recovered 1: 305 2020 Dead after 13 days ; 1750 Paralysis; dead after 3} days 1: 310 1290 Paralysis; dead after 3 days ; 1780 Dead after 3 days effect becomes greater. Finally, all the animals that received mixtures containing 2.4 ce. regularly die after two to three days. This amount, therefore, was taken as the L, dose. The fluid toxin preserved under toluol becomes constantly weaker. For this reason the fluid toxin must be tested from time TESTING ANTITOXIC DYSENTERY SERUM 305 to time with the standard serum with respect to the L, dose. The minimal lethal dose of the toxin must, also, be determined. If it is desired to guarantee the constancy of the L; dose in order to avoid its repeated determination, we must use a dry toxin obtained by precipitation with ammonium sulphate. Such dried toxin preserved in Ehrlich vacuum tubes can maintain its properties for years. The testing of antidysentery serum is conducted with the use of the L, dose of the toxin in the manner similar to the method of testing diphtheria antitoxin. The mixtures, after being kept for an hour at 37°C. are injected intravenously into rabbits of 1500 to 2000 grams weight. The largest amount of the serum which when mixed with the L, dose of toxin and injected intra- venously will permit the death of the animal in four to five days is taken as the antitoxic unit for that serum. The results of such a test are shown in table 3. It is seen that in this instance 1 cc. of the serum 237 contains 305 antitoxin units. Our standard serum is preserved, as usual, in the dry state in vacuum tubes. Our institute will be glad to supply tested standard serum upon request. THE BINDING RELATION OF THE DYSENTERIC TOXIN AND ANTITOXIN According to Todd, the union between dysentery toxin and its antitoxin takes place considerably more quickly at warmer temperatures than it does in the cold. At 37°C. five minutes were sufficient, whereas two hours were required for complete neutralization at 0°C. Our experiments, also, indicate that differences in temperature exert a noteworthy influence on the course of the toxin-antitoxin reaction. The protocol presented in table 4 shows that complete union takes place after one hour at body temperature. The union thus takes place rather slowly. However, one can be sure of complete union within one hour at 37°C. If the dysentery toxin behaved like diphtheria toxin, it should be expected that the minimal lethal dose would increase with time, while the binding power would remain unchanged; that 306 is, the L, dose would not increase. YOSHIMOTO FUKUHARA However, the results of tests presented in table 5 show that this is not the case. Showing the influence of time on the reaction between dysentery toxin and antitoxin we Re orcs After 30 minutes at | After one hour at jena 37°C. 37°C. “at 90°C me ie eee 1.5 Lived 137 Lived D, 24 1.8 Par., Rec. Par: + Rec, 19 Par., D4 Lived D, 93 Par., Rec. oct Par., D 23 Par: DS: 2.0 Pars. 2 Lived D 13 D2 ; Lived Par., Rec. 21 Par., Rec. Par. D5% Par., D 23 Par., D 8} : D2 Par., Rec. Par:,, DAs 22 Par., Rec. Lived Par., D4 Par., D 5} " D 20 hours Lived Par. D7 Par., Rec. 23 D 13 Par., D 23 D 13 Par., Rec. ‘ D2 Par., Ree. Par., D5 24 D 13 Par., D 23 Par., D 14 ; Par., D3 Par., D3 25 Par., D 24 D 13 D4 . D2 Dif 2.6 Du TABLE 4 INJECTED AFTER BEING MiXED WITH ONE UNIT OF THE ANTITOXIN Par. = paralysis; Rec. = recovered; D = died (the numerals indicate the number of days that the animal survived the injection). It is seen that in this case as the lethal dose increased from 0.009 ec. to 0.08 cc. the Ls. dose rose from 1 ce. to 7.5 ec. This experiment makes it probable that the changes in the dysentery TESTING ANTITOXIC DYSENTERY SERUM 307 toxin represents an alteration into non-toxic substances which possess no avidity for the antitoxin. Expressed in the termi- nology of Ehrlich, a destruction of the toxin had taken place which affected not only the so-called toxophore group but also the haptophore group. The changes are, thus, different from those which take place in diphtheria toxin. TABLE 5 Influence of age on the toxicity and the binding power of dysentery toxin 9 MINIMAL LETHAL DATE OF TEST Ont L+ DosE BPE Ayes) UL ieee icy ciaiesecsite avec! ola esaes, s/o winy'a wilese: e'alone 0.009 1.00 PETITE OMG cn iret ifs nates eh akc ayalerale readers 0.012 1.20 I COMA CL ON SOLO! sas ara aveie wld aieieys cba a eialedisie's 0.014 1.23 Nearer ber TAWA ON Ase erate = abcess « ies alleys, cfavaye 0.080 _7.50(?) PRE AMUESIV VERE Strats op2)s< 1s o\s) ste 1s aio! s)2! sissies 6, 19/09 e.a%< 0.170 bi * The L, dose, at this date, was too large to be estimated. SUMMARY 1. In order to determine the antitoxic value of anti-dysentery serum it is necessary to select a standard serum which can be preserved in a dry state in the vacuum tubes of Ehrlich. 2. The standard antitoxin unit adopted by us was the amount that neutralized 100 minimal lethal doses of the toxin which at that time was at our disposal. 3. For the testing of other anti-dysentery sera the so-called L, dose of dysentery toxin was adopted as the smallest amount which when mixed with the antitoxin unit and injected intra- venously into a rabbit of 1500 to 2000 grams weight caused the death of the animal within from four to five days. The L, dose of the toxin is employed, in the usual manner, to determine the relative antitoxic value of newly prepared antisera. 4. The deterioration of aged dysentery toxin is referable to a change of the toxin molecule into a non-toxic modification which possesses no avidity for the antitoxin. The toxin molecules lose their toxicity without the formation of toxoids; in this re- pect the dysentery toxin is unlike diphtheria toxin. EXPERIMENTAL PURPURA MARK J. GOTTLIEB From the Laboratory for Clinical Research, New York, N.Y. Contribution No. 1 Received for publication July 24, 1919 Although purpura has been known for a long time as a disease in which hemorrhages occur spontaneously or from slight prov- ocations, it has only been of late that we have recognized the fact that a condition may exist in the individual wherein after the tissues have been cut or bruised, they will bleed for an abnor- mally long time. Neither the person so afflicted nor his medical attendant is cognizant of his condition because these patients do not show outward signs of the disease except after sustain- ing an accidental or surgical wound. Associated with this pro- longed bleeding is a diminution in the number of blood platelets and the red blood cells are easily laked by hypotonic or hyper- tonic salt solutions. Hess (1), who has studied this condition carefully, suggests that in the blood stream of these individuals there exists a toxic substance which has the power of dissolving the blood platelets: and of rendering the erythrocytes easily laked by heterotonic: salt solution and that this substance causes a disturbance in the nutrition of the lining of the blood vessels. He believes that the reduction in the number of the blood platelets is more ap- parent than real and comes to this conclusion after an experi- mentation by which he finds that although the number of visible blood platelets be reduced, the remaining portion exists in solution in the blood stream. Lee and Robertson (2) have been able to produce experimental purpura in guinea-pigs. ‘They proceeded as follows. The washed blood platelets of two guinea-pigs were injected intra- venously into a rabbit and this was repeated every seven days for four 309 310 ; MARK J. GOTTLIEB successive times; on the seventh to the tenth day after the last admin- istration of guinea-pigs’ blood platelets, either the rabbit was bled to death or blood was withdrawn from the heart and the serum thereof was injected intraperitoneally into guinea-pigs in varying doses; as a result of this procedure a condition was produced in these animals wherein the bleeding time was very much prolonged, while the coagula- tion time remained normal; the blood platelets were reduced in number. Those animals which received a sufficiently large dose of the serum usu- ally died and on post-mortem examination there were found blood clots and blood stained fluid in the pleural and peritoneal cavities, hemor- rhages into the walls of the intestines, heart muscle and liver. Before death there was a continuous ooze of blood from small cuts which were made in the ears for the purpose of procuring blood for the various tests. This anti-platelet serum not only had the property of producing this condition in the living animal but it also dissolved blood platelets in vitro. We have repeated this experiment in the following way: one or two guinea-pigs, depending upon size, were exsanguinated by severing the vessels of the neck and allowing the blood to flow into a flask containing citrate solution; the blood thus obtained was centrifuged at high speed for fifteen minutes, and after the supernatant fluid was removed, the white coat which was found on the surface of the sediment was pipetted off; normal saline solution was added to the collected blood platelets and the mixture was thoroughly agitated; this was centrifuged again for fifteen minutes; the supernatant fluid was removed and the layer of blood platelets was pipetted off and again washed as before; the washed platelets were then suspended in saline solu- tion and injected intravenously into a rabbit; although we have washed blood platelets many times, we have never been able to procure them entirely free from red blood cells, probably because the quantity of blood used was small and in our endeavor to procure the maximum amount of platelets some red cells were naturally taken up with them; the rabbits were subjected to weekly intravenous injections of the blood platelets thus pro- cured according to the succeeding tabulation. EXPERIMENTAL PURPURA Sila White rabbit no. 1; weight, 2006 grams December 30, 1918. Blood platelets of 1 large guinea-pig, intrave- nously. January 6, 1919. Blood platelets of 2 large guinea-pigs, intrave- nously. January 14, 1919. Blood platelets of 2 medium sized guinea-pigs, intravenously. January 21, 1919. Blood platelets of 1 large guinea-pig, intrave- nously. January 31, 1919. Blood platelets of 2 small guinea-pigs, intrave- nously. February 13, 1919. Etherized and exsanguinated and the blood col- lected under as steril conditions as possible. TIME ANGORA GUINEA-PIG, 501 GRAMS BLACK GUINEA-PIG, 490 GRAMS February 11, 1919: Coagulation time] 7 minutes 5 minutes Bleeding time ...| 9 minutes 9 minutes Blood platelets.. .| 484,000 per cubic centimeter} 649,000 per cubic centimeter SG piles... es: 2 cc. normal rabbit’s serum} 1.6 cc. serum from rabbit given intraperitoneally number 1 (anti-platelet) intraperitoneally February 12,1919, 10 a.m: Coagulation time| 6 minutes 6 minutes Bleeding time. ...| 10 minutes Continuous Blood platelets. ..| 250,000 333,000 EINES eiehvers\<.3S:a as Normal Bloody and contained mod- erate number of red blood cells and a few hyalin casts. February 12, 1919, 4.30 p.m: Coagulation time} 3 minutes 2 minutes Bleeding time ...| 10 minutes Continuously bleeding Blood platelets . .| 485,000 171,000 February 13, 1919; 9 a.m. Alive and well Found dead.* * Autopsynotes. Bloody discoloration of peritoneum. Small quantity of blood stained fluid and clots in peritoneal cavity. Hemorrhagic area in the wall of the right auricle. Blood stained fluid in pleural cavity. Right lung intensely con- gested but not consolidated. ole MARK J. GOTTLIEB Two guinea-pigs of approximately equal weight were selected and action of the serum obtained from white rabbit number 1 was compared with the action of normal rabbit’s serum. It will be noticed in the above protocol that 2 ec. of normal rabbit serum were given to the angora pig while 1.6 cc. of anti- platelet serum were given to the black pig. There was a tem- porary reduction of blood platelets in the blood of the pig that received normal rabbit serum while the reduction in the number of blood platelets in the pig that received anti-platelet serum was progressive. The activity of the serum of rabbit number | (anti-platelet) against guinea-pig’s blood platelets as compared with the action of normal rabbit serum was also determined. In this connection it might be well to add that Lee and Robertson found that com- plement is essential in the reaction and we accordingly used the rabbit serum on the same day that it was obtained, in order that the complement that is normally present in sera would be opera- tive in the tests without the addition of extraneous complement. In the following protocol 0.5 cc. of diluted serum was mixed with 0.05 cc. of a 5 per cent suspension of guinea-pig’s blood platelets in 0.9 per cent salt solution. After the mixtures were made, they were incubated for one hour in the water bath at 37°C. and the results were then read. DILUTION RABBIT 1 NORMAL RABBIT 1-1 Complete lysis Complete lysis 1-2 Complete lysis Complete lysis 14 Complete lysis Complete lysis 1-6 Complete lysls Complete lysis 1-8 Complete lysis Partial lysis 1-10 Complete lysis No lysis 1-12 Complete lysis No lysis 1-15 Complete lysis No lysis 1-20 Complete lysis No lysis 1-25 Complete lysis No lysis 1-30 Complete lysis No lysis 140 Complete lysis No lysis 1-50 Complete lysis No lysis 1-60 Complete lysis No lysis 1-80 Partial lysis No lysis EXPERIMENTAL PURPURA ole From the foregoing we learn that normal rabbit serum is lytic to guinea-pigs’ blood platelets in dilutions of the serum not higher than one to eight, whereas the serum of the rabbit that was treated with guinea-pig’s blood platelets dissolved them in dilu- tions far greater. Because of the fact that we were never able to obtain blood platelets entirely free from red blood cells, it is quite natural that we should endeavor to determine how much hemolytic anti- body was produced in the rabbit in conjunction with the platelet anti-body (3). A rabbit was prepared as in the case of rabbit 1 and its serum was used on the same day that the blood was re- moved from its heart. This serum was used in the following experiment. In a set of test tubes 0.5 ce. of diluted serum and 0.05 ce. of a 5 per cent suspension of guinea pigs’ red blood cells were mixed. In another set 0.5 ec. of diluted serum and 0.05 ee. of a 5 per cent suspension of guinea-pigs’ blood platelets were mixed. ‘The tests were incubated for one hour at 37°C. and then read. Rabbit serum 2 DILUTION PLATELETS ERYTHROCYTES 14 Complete lysis Complete lysis 1-8 Complete lysis Complete lysis 1-12 Complete lysis Complete lysis 1-16 Complete lysis Complete lysis 1-24 Complete lysis Complete lysis 1-32 Complete lysis Partial lysis 1448 Complete lysis Partial lysis 1-64 Complete lysis Partial lysis 1-96 Partial lysis No lysis 1-128 Nearly complete lysis No lysis 1-192 No lysis No lysis 1-256 No lysis No lysis 1-384 No lysis No lysls Here we find that a hemolytic antibody was produced in con- junction with the platelet antibody but, apparently, not in as great an amount. The question must necessarily arise in one’s mind: is this hemolytic antibody also a factor in the production 314 MARK J. GOTTLIEB of so-called experimental purpura or does it exert the preponder- ating influence in its production? This matter will be reported in a succeeding contribution (4). Black rabbit 3 was injected intraperitoneally with the blood serum of rabbit 2 for the purpose of determining whether the purpuric serum of one rabbit would produce in a normal rabbit protective substances against the action of this purpuric serum or whether the transferred purpuric serum would confer upon the serum of the recipient rabbit the property of producing purpuric symptoms. The following protocol indicates how the serum was administered. April 25, 1919. 2 cc. of rabbit serum 2 injected intraperitoneally into rabbit 3. April 26, 1919. 2 cc. of rabbit serum 2 injected intraperitoneally into rabbit 3. April 28, 1919. 2 cc. of rabbit serum 2 injected intraperitoneally into rabbit 3. April 30, 1919. 2 ec. of rabbit serum 2 injected intraperitoneally into rabbit 3. May 2, 1919. 2 cc. of rabbit serum 2 injected intraperitoneally into rabbit 3. May 5, 1919. 2 cc. of rabbit serum 2 injected intraperitoneally into rabbit 3. May 7, 1919. 2 cc. of rabbit serum 2 injected intraperitoneally into rabbit 3. May 9, 1919. 2 cc. of rabbit serum 2 injected intraperitoneally into rabbit 3. May 19, 1919. Paracentesis cardia was performed on rabbit 3 and the serum obtained was used in the following experiment. The lethal dose of the serum of rabbit 2 (anti-platelet) was determined to be about 1.2 cc. This dose was sufficient to cause the death of guinea-pigs whose average weight was 500 grams, within forty-eight hours. The dose used in this last experiment was 1.5 ce. sees om tg Rn sae a a pe ee Meaae Tae eee a i EXPERIMENTAL PURPURA 315 GUINEA-PIG No. 1 No. 2 No.3 POESEREN GS 55 oie dn tre 473 grams 478 grams 463 grams May 21, 1919: Bleeding time.| 15 minutes 15 minutes 33 minutes Platelets...... 281,000 266,000 510,000 BSE2 pom. o 3.2 3 1.5 ec. serum rabbit| 1.5 cc. serum rabbit} 1.5 cc. serum rab- 2 (antiplatelet) 2 (antiplatelet) bit 2 (antiplate- intraperitoneally intraperitoneally let) intraperi- toneally May 22, 1919: 10.30 a.m. Bleeding time.} 1 hour 5 minutes Continuous bleed-| Continuous ing bleeding Platelets......| 77,000 57,000 128,000 10.45 a.m. .... 2 cc. serum rabbits} 2 cc. normal rab- 3 intraperiton- bit serum in- eally traperitoneally AA p.m... «5. Platelets...... 52,000 278,000 325,000 Bleeding time.| Bleeding continu- | Bleeding continu-| Bleeding contin- ously ously uously Bpms ss... 2 cc. serum rabbit 3] 2 cc. norma’ rab- subcutaneously bit serum sub- cutaneously May 23, 1919: TPAD coe oh Found dead Still alive Still alive May 24, 1919: 240 a ee Found dead Still alive Autopsy....... Hemorrhages into} Hemorrhages into subcutaneous tis-| intestines, stom- sues, intestines,| ach, liver and stomach, and heart heart muscle It will be noticed in the above cited experiment that in the case of guinea-pig 3 normal rabbit’s serum apparently had the effect of counteracting the action of the anti-platelet serum. The serum of rabbit 3 did not seem to be as effective as normal rabbit’s serum; the life of guinea-pig 2 was prolonged but its death was not prevented. In this instance we may infer that the treatment of rabbit 3 with the serum of rabbit 2, partly nullified the protective qualities of the serum of rabbit 3. 316 MARK J. GOTTLIEB Whether the serum of rabbit 2 remained operative as such in rabbit 3 or whether it produced some substance in rabbit 3 which offset those protective agents, remains open for further investigation. REFERENCES (1) Hess, Aurrep E.: Proc. of the Soc. for Exper. Biol. and Med., 1917, 14, No. 3. (2) Lez AnD RosperTsoN: Journal of Medical Research, 1916, 28, 323. (3) LepincHam, J. C. G.: Lancet, 1914, p. 1673. (4) LepINGHAM AND Brepson: Lancet, 1915, p. 311. THE ANTIGENIC PROPERTY OF THE PFEIFFER BACILLUS AS RELATED TO ITS VALUE IN THE PROPHYLAXIS OF EPIDEMIC INFLUENZA CHARLES W. DUVAL anp WILLIAM H HARRIS From the Department of Pathology and Bacteriology, Tulane University of Louisiana Received for publication July 28, 1919 During the height of the epidemic of influenza which occurred in New Orleans in the fall of 1918, when the Health authorities were recording between two and three thousand new cases daily, we undertook to determine the value, if any, of specific bacterial protein in the prevention of the infection. Prior to the systematic vaccination of a relatively large number of persons, we had satisfied ourselves (1) that the Pfeiffer bacillus played an im- portant réle in the clinical disease known as epidemic influenza. Furthermore, while this work was in progress we noted that B. influenzae possessed distinct antigenic properties, which indicated that some degree of protection against the disease might be afforded in man through its employment as a vaccin. The futility of all other methods to check the overwhelming spread of influenza suggested efforts along prophylactic lines already established and well recognized for various other epidemic diseases. Inasmuch as previous visitations of epidemic influenza extended over a period of only a few weeks, the possibility of establishing protective sensibilizators sufficient for such a dura- . tion, seemed to justify attempts at transient immunization with specific bacterial proteins. Again, if the epidemic continued in violence for a longer period revaccination could be carried out. With these basic principles at hand, we instituted to as great an extent as possible this method of defense after proper represen- tation to those concerned of its hypothetical limitations. We further realized that at the time of the previous great epidemic in 1890-92 the influenza bacillus was discovered late in the epidemic 317 318 Cc. W. DUVAL AND W. H. HARRIS period (2) and bacterial vaccin immunization had not yet formed a part of our protective armamentarium. ‘Therefore, it was apparent that only a thorough trial of vaccin in the present epidemic could demonstrate its degree of efficiency or possible worthlessness; in either instance no harm could result in conse- quence of its administration. The successful accomplishment with other vaccins such as small-pox, typhoid, and more recently the meningococcus, encouraged the hope that perhaps at least something might be achieved in the use of a vaccin of the Pfeiffer bacillus or that knowledge bearing upon its status as a preventive method might be gained for consideration of its employment in future epidemics. Vaccin and its preparation. As it was our express purpose to attempt the protection against a specific infection caused by one and not several bacterial agents, we employed as a vaccin the protein of killed cultures of B. influenzae. We did not consider that there was anything to be gained for prophylaxis against Pfeiffer infection through the use of mixed vaccin (3), or one that contained, with B. influenzae, the pneumococcus, streptococcus, staphylococcus, etc., since these latter species, in our opinion, play only a secondary réle in the epidemic disease. While the use of a heterologous vaccin could do no harm there seemed no occasion to attempt the immunization against possible second- ary invaders, particularly as their activity followed the in- fection of the host by some other excitant. Even though the disease is not caused alone by the Pfeiffer bacillus, it was con- sidered logical to attempt the prevention of the infection by exciting, in the prospective host, immune substances specific for . this microorganism. If the primary infection could be pre- vented, dependent secondary infections in consequence would not arise. To accomplish this protection the Pfeiffer antigen alone was employed to stimulate the maximum amount of anti- body production without simultaneous interference with the mechanism by other antigens. The cultures included several isolations from recent cases of the epidemic disease and a strain of B. influenzae (Wollstein) obtained from the Rockefeller Institute which has been in ANTIGENIC PROPERTY OF PFEIFFER BACILLUS 319 our possession for a number of years. While at first it was thought advisable to use a number of influenzal isolations, having in mind the possible existence of types it was later found unneces- sary as the agglutination tests carried out upon patients’ sera did not suggest ‘‘variants” of B. influenzae, the absence of which seemed already determined by Wollstein (4) in her study of a large number of isolations from widely separated localities and from various pathological conditions. In regard to the antigenic property of various influenzal cultures it was found that the one from the Rockefeller Institute possessed this to a high degree, and perhaps greater than other cultures employed by us. Because of its power constantly to produce specific antibodies and to give rise to a marked local and constitutional reaction in the vaccinated host this culture was finally employed to the exclusion of all others. The influenza bacilli for vacecin were grown upon the surface of a solid nutrient agar medium, 0.6 per cent acid against phe- nolphthalein, to which was added before allowing to solidify, one per cent steril, saline washed human erythrocytes (complement free). These were thoroughly mixed with the liquid agar medium in large Erlenmeyer flasks and the admixture was cooled rapidly to ensure against settling out of the erythrocytes away from the surface on which growth was to be encouraged. Where freshly drawn blood was used, the defibrinated sero-cellular part was heated at 56°C. to destroy the complement before adding it to the nutrient medium. The solidified culture medium was seeded by flooding the surface with a fresh saline suspension of B. influenzae, introduced by means of a sterile pipet. Incubation at 37°C. was carried over a period of 48 hours, and the growth was then washed off with steril saline solution; care was taken not to disturb the culture medium. The collected suspension was now devitalized by saturation with chemically pure chlo- roform (Merck’s) which was allowed to remain for ten to fifteen minutes in the admixture. The mixture was agitated from time to time to insure uniform diffusion of the germicidal agent. Repeated tests showed that the chloroform killed B. injlwenzae in a few seconds, and apparently without injury to the antigenic THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 5 320 Cc. W. DUVAL AND W. H. HARRIS property of the bacterial protein even though this chemical remained in contact with the culture for a period of days. After the culture was devitalized the chloroform was allowed to settle out which in greater part occurred within 4 few minutes. The supernatant suspension of killed bacilli was now removed to steril flasks. These were lightly plugged with cotton and placed at 37°C. for a time sufficient to remove all traces of chloroform, which usually required less than one hour. The chloroform-free culture was then standardized and tested for viability after which it was ready for use. The vaccin prepared in this manner gives a suspension of bacterial protein free from all other ex- traneous irritants and unappreciably altered in its antigenic power. - For inoculation purposes we employed only the freshly pre- pared vaccin (not more than a week old). The vaccinations were all made by subcutaneous injection into the outer aspect of the arm below the insertion of the deltoid muscle. The dose em- ployed for adults was approximately one billion killed B. in- fluenzae for the first injection, one-half this number for the second, and one billion for the third injection, allowing an interval of three days between the inoculations. Arbitrarily it was decided that three doses constituted a complete treatment, following in this respect the ordinary method of vaccination against typhoid fever. We did not consider lipovaccin under the existing con- ditions of the epidemic for the reason that its action might be slower, and we desired the maximum amount of host response in the shortest possible period. OBSERVATIONS UPON THE EFFECTS OF VACCINATION Approximately five thousand persons were vaccinated by sub- cutaneous injection with influenzal protein which had been specially prepared by devitalizing the culture with chloroform. The majority of those vaccinated were employees in the large commercial houses, banks and factories of the City of New Orleans. Because of the interest and cooperation taken by the various heads of these establishments it was possible systemati- ANTIGENIC PROPERTY OF PFEIFFER BACILLUS 321 cally to observe the effects of the vaccin upon selected groups for a period of months after the injections. In addition to these groups we vaccinated several thousand persons widely distributed over the city. In each instance the vaccin was administered to those who at the time were apparently well, and stated that, so far, they had escaped influenza infection. For our group series the vaccinations were completed within a period of two weeks during which time the epidemic had reached its maximum intensity (see chart). However, before the epi- demic showed any signs of recession our series in greater part had received the complete vaccin treatment, namely three doses given at intervals of three days between each injection. CHART— SHOWING NUMBER OF CASES PER WEEK FOR CITY OF NEW ORLEANS. OCTOBER OCTOBER OCTOBER Coreaey ZENER 82 To {5° 155 TO 227% 22270 2T= nT oe ao. ae VACCINATION CARRIED CUT GETWEEN OCTOBER 182 — 23% WHICH CORRESPONDS THE WEEK OF GREATEST INTERSITY OF THE EPIDEMIC. Hight weeks after the subsidence of the epidemic, a second visitation (recrudescence) occurred, which afforded an unusual opportunity to determine in those previously vaccinated the duration of protection occasioned by the specific protein of B. influenzae. Out of a total number of our group series, 2608 persons who received the complete treatment of vaccin (three injections) 98.3 per cent did not contract the disease. Of the total number of 346 who received only two injections 92 per cent did not contract influenza during the first epidemic. Seventy-six per cent of the 118 receiving but one injection did not develop the a2 Cc. W. DUVAL AND W. H. HARRIS disease. ‘This latter group shows a marked difference in the per- centage of infection compared with those receiving the full treatment. It is noteworthy that in those vaccinated that developed influenza the character of the infection was mild and without pneumonic complications. There were 866 individuals, forming a part of the forces vaccinated, who refused injection; 375 of these or 41 per cent developed influenza as contrasted with only 3.3 per cent occurrence of the disease in those vacci- nated. These persons afforded a means of control inasmuch as their occupation and daily environment were identical with those that were vaccinated. While the incidence of infection is higher in the controls than that shown for the general city sta- tistics, the difference is explained by the fact that these indi- viduals were employed in large establishments located in the congested business sections. Tables 1 and 2 indicate the results with individual vaccinated groups. Table 3 presents a general summary with controls of the results obtained in all groups. While we realize that the number of persons vaccinated is too small to draw any sweeping conclusions relative to the percentage of absolute protection afforded and the duration of the immunity we believe, however, that even in these relatively few cases, the results indicate specificity of the Pfeiffer bacillus and efficacy of the vaccin as a prophylaxis in the infection. The clinical reaction. In the majority of cases a reaction occurred at the site of inoculation, usually in the form of a localized erythema. This was anything from a mild circum- scribed redness, 4 to 5 em. in diameter, to a markedly swollen and reddened skin and subcutaneous tissue involving the whole arm and the greater part of the forearm. ‘This local inflam- matory reaction gradually subsided and the redness faded out in three to five days. In addition to the local effects there occurred, in the majority of the cases, a well defined constitutional reaction. This host response was frequently so striking in many of its clinical aspects as to simulate the early stages of influenza. In some instances the reaction was severe enough to confine the individual to bed for a period of eight to ten hours, such usually PA ESR AY 1 NN Oe SS mE ners ANTIGENIC PROPERTY OF PFEIFFER BACILLUS 323 TABLE 1 Showing results of vaccination for individual group A as paliicieeers NUMBER VACCINATED NUMBER REFUSING VACCINATION EMPLOYEES A 1000 364 398 238 (100%) (36.4%) (39.8%) (23.8%) : Number not : Number not Number developing aocoloni Number developing : . ping : developing influenza aioane influenza Aineace 5 393 93 145 (1.2%) (98.8%) (39%) (61%) No pneumonia 27 pneumonia TOTAL NUMBER DEVELOPED INFLUENZA VACCINATED PERSONS Q Number of GROUP A persons Per cent Protected per cent Wine injection... ..5..9...... 26 4 15 85 Two injections.............. 30 1 3.3 96.7 Three injections:........:... 342 0 None 100 TABLE 2 Showing results of vaccination for individual group B roraL | NUMBER NUMBER NUMBER VACCINATED NUMBER REFUSING VACCINATION EMPLOYEES|2!CE AT TIME 1200 517 583 100 (100%) (43%) (48%) (8%) * Number not : Number not Number developing develoni Number developing : : ping : developing influenza aicease influenza disease 22 561 By/ 63 No pneumonia 14 pneumonia TOTAL NUMBER DEVELOPED INFLUENZA VACCINATED PERSONS, GROUP B ee Per cent Protected per cent Meae injection. .........2- <0: 40 14 35 65 mee miyections.............. 191 8 41.8 58.2 Three injections............. 452 0 None 100 324 Cc. W. DUVAL AND W. H. HARRIS appearing as early as six to eight hours after the first injection of the vaccin. We wish to mention in this connection that the severity of the reaction accounted for a number of. our cases receiving only one injection, the individuals stating they did not care to submit to the inconvenience of a second inocula- tion. In our experience the second injection of vaccin rarely gave rise to more than a mild constitutional reaction, and the third to little if any at all. In general some form of constitutional response to the vaccin TABLE 3 Summary of the results of collective groups VACCINATED DEVELOPING CONSTITUTIONAL NUMBER OF INFLUENZA REACTION PERSONS VACCINATED Num Per | Pro-| | Severe| Mild | None ber cent | tected One INJECION «,. ¢)s 1.2 oo. - 118 29 | 24 76 Mwo injections..:). 020. .3 6.0. 346 28 | 8 92 30 60 10 Three injections............. 2608 45 | 1.7 | 98.3 POG es Sistas ies cloereeds 3072 102 3.3 | 96.7 pee Gee Ee a INCIDENCE OF DISEASE Num- Per Pro-_| Vacci- | Unvac-| Differ- ber cent | tected | nated | cinated| ence per cent|per cent|per cent|per cent Wontrolseayeee coe sicc soo cles 866 Persons not vaccinated..... 375 | 41.6 | 58.4 | 3.3 | 41.6 | 38.3 occurred in 90 per cent of the cases. The severe reaction was noted in 30 per cent. Only in 10 per cent did an appreciable reaction fail to develop (see table 3). In persons reacting the constitutional symptoms ranged from a slight headache, mild pains over the body, lassitude and a half to one degree of temperature, to severe frontal and occipital headache, neuralgic pains over the body, not infrequently ushered in with chill and nausea and followed by a temperature of 101° to 102°. Itis noteworthy that even with persons that responded more violently to the vaccin, the reaction invariably subsided ANTIGENIC PROPERTY OF PFEIFFER BACILLUS 325 in from six to eight hours after the onset, leaving the individual perfectly normal to all intent and purpose. As the epidemic was raging at the time vaccination was insti- tuted we expected to have a number of cases develop the disease shortly following the first injection of vaccin. We reasoned that a number, though apparently well at the time, would in reality be in the later stage of the incubation period, and in consequence not to be benefited by vaccin therapy. Under these circumstances it was surprising how few developed the infection during the process of the vaccination period. However, our tabulated results do not exclude any case that contracted influenza after the first injection of vaccin. Such results would indicate that the negative phase, considered by some as a contraindication for vaccination during the epidemic, does not exist or is a negli- gible factor. Blood findings. With a view of determining whether the leucocytic blood picture is altered or influenced by vaccin in- jection and to compare it with that noted in influenza infection, total and differential counts were made upon selected groups of cases before and after vaccin administration. In the vacci- nated instances the leucocytic count was, as a rule, somewhat higher than normal, the rise being due to the neutrophiles (5). Leucopenia was not noted in any of the vaccinated persons, but on the contrary the leucocytic count ranged from 9000 to 12,000. A diminution in the normal number of leucocytes which is a characteristic feature in influenza infection was hardly to be expected in the vaccinated case where the antigen is introduced subcutaneously. In this connection it may be mentioned that typhoid antigen injected subcutaneously excites an increase rather than a decrease of the total leucocytic count (6), whereas in the natural disease the organism produces a leucopenia. Duration of the protection. ‘This could only be approximated by the length of time antibodies for the Pfeiffer antigen remained in demonstrable quantities in the blood of those vaccinated. On the average these substances were found persisting in the sera for a period of two months though often in greatly diminished quantity as compared with that demonstrable ten days to two 326 Cc. W. DUVAL AND W. H. HARRIS weeks after vaccination. However, the second epidemic which occurred about two months after the vaccination, afforded in those again exposed further data bearing upon the duration of the protection. Of those vaccinated at the time of the first visitation of the disease and who successfully passed through this epidemic, 8 per cent contracted influenza during the recru- descence. Of the recorded group controls or those who refused vaccination in October and did not develop the disease during the first epidemic, 56 per cent contracted it in the period of the second epidemic. . TABLE 4 Showing probable duration of protection, group C PERSONS VACCINATED IN DEVELOPING INFLUENZA IN OCTOBER, 1918 (PERIOD OF FIRST} JANUARY, 1919 (PERIOD OF RE- PERIOD OF PROTECTION EPIDEMIC) CRUDESCENCE) : 272 21 Pevacninated in J 1919 Developing influenzain January, evaccinated in January, 1919 3 months 154 1 Total number persons not Developing influenza during Developing influenza during vaccinated in October, 1919 first epidemic recrudescence 238 93 82 A survey of the results for our series of cases, taking into account only those who did not contract the disease during the second epidemic, indicate two to three months as the average duration of protection afforded by the injection of the chloroform prepared Pfeiffer vaccin. The antigenic specificity. Inasmuch as the sera from infected individuals with B. influenzae yield .quite constantly specific agglutinins and to some extent lysins of the complement fixing group, it seemed consistent to expect that normal persons injected with the protein of the Pfeiffer bacillus would likewise contain in the blood specific antibodies. Agglutination and complement fix- ation tests were therefore carried out upon the sera of a number of those vaccinated in order to obtain an index of the production of such substances. Agglutinins were constantly found present ANTIGENIC PROPERTY OF PFEIFFER BACILLLS aan in variable amounts; the titer, however, was not, as a rule, as great in the sera of the vaccinated individuals as in the cases of the actual disease. Dilutions of 1:80 gave good reactions while in higher dilutions only exceptionally was the clumping clear cut. The complement fixation tests were not as constantly positive in the vaccinated cases as the agglutinin reaction; the former occurred only in sixty per cent of the cases. It was evident, however, that at least a certain amount of antisubstances or sensibilizators occurred, which justified the employment of vaccin as a prophylactic. It was considered that even though the protective factors stimulated were of low titer and transient in duration, they were of value in the defense against the fulmi- nating epidemic disease. DISCUSSION AND SUMMARY During the height of the epidemic of influenza in New Orleans more than five thousand persons were vaccinated by us with a specially prepared protein suspension of the Pfeiffer bacillus. Of this number approximately 90 per cent did not contract influ- enza, either during the period of the epidemic or its recrudescence which occurred two months afterwards. This percentage com- pared with that of our group controls (those refusing vaccin) and the city statistics indicate that a considerable degree of protection was established in all those who were completely vaccinated and it is reasonable to suppose that the protection was the result of the inoculation with B. influenzae antigen. The culture used in the preparation of the vaccin was grown upon the surface of hemoglobin nutrient agar at a temperature of 37°C. for thirty-six to forty-eight hours, when the growth was washed off and suspended in normal saline solution and immedi- ately devitalized by being saturated with chemically pure chlo- roform. In our experience the highest potency vaccin was obtained from the influenza culture, which was killed with chlo- roform. With this agent the viability of the bacilli was destroyed in a few seconds without apparently causing any deleterious eilect upon the protein immunizing substance. While it is 328 Cc. W. DUVAL AND W. H. HARRIS recognized that chloroform in the presence of water and under the influence of light (actinic rays) in the course of time decom- poses and give rise to COCI, (carbonyl! chloride) and HCl (hydro- chloric acid), which are poisonous substances, it can be stated that these compounds are not evolved in the chloroform vaccin prepared in the manner herein described. In this connection it may be said that the devitalizing agent is left in contact with the saline suspension of bacilli for a short period of time and is then completely removed from the mixture. In regard to the behavior of chloroform upon the vaccin we believe that it devitalizes the bacilli by rapidly absorbing the water and in consequence increasing the permeability which re- sults first in plasmolysis then rupture of the bacterial cell, lib- erating without effecting in any manner its toxic moiety. There is nothing to show that chloroform per se has any direct chemical action upon bacterial protein whether alone or in combination with water and sodium chloride. On the other hand, this cannot be said for heat and the phenol derivatives when used as vaccine devitalizing agents. Here the destruction of bacterial life is due to a direct action upon the protoplasm, which is coagulated to a greater or lesser extent and in consequence there is destroyed at least a part of the antigenic property of a protein (7, 8). Therefore, in our opinion chloroform as a germicide in the prep- aration of vaccin has distinct advantages over all other agents commonly employed. The use of heat, tricresol, etc., for killing the culture, while effective from a germicidal standpoint, injures to some degree the immunizing property of the protein and particularly is this the case with B. influenzae. Influenza vaccin, freshly prepared with strict attention to pre- serving its maximum amount of antigenic value, will excite in the human host the production of specific immune bodies, the degree depending not so much upon the quantity as the quality of the antigen injected. Therefore, the freshness of the vaccin and particularly the method of its preparation are factors of paramount importance if results are to be obtained in the pro- phylaxis against epidemic influenza. We believe that failure to obtain sufficient protection is, to a large extent, the fault of the ANTIGENIC PROPERTY OF PFEIFFER BACILLUS 329 vaccin employed. A contrast study of the results obtained with influenzal vaccins devitalized by heat, tricresol and chloroform will show a wide variation in antigenic effectiveness. The culture killed by heat at 56°C. is practically worthless as an antigen while that prepared by the phenol derivatives even in 0.25 per cent strength is altered in its power to produce antibodies. While our total number of vaccinated persons is too few to permit of the statement that B. influenzae vaccin is a constant specific in the prevention of the disease clinically known as epidemic influenza, it can be said, however, that the results are interesting and significant from the standpoint of prophylaxis. While the duration of specific antibodies in the blood of the host varies within wide limits for different individuals, it may be said that these substances remain in the circulation for a period of ten weeks on the average, as shown by the agglutination and complement fixation tests. In many isolated cases, however, these tests are positive six months after vaccination. There occurred in all of the individuals vaccinated a local reaction at the site of inoculation. Usually this was simply a small sharply circumscribed area of erythema; however, in fully 30 per cent of individuals the reaction involved a greater area and there was considerable oedema of the subcutaneous tissues in the vicinity of the inoculation site. Even in the more severe type of local reaction there was a complete subsidence in three to five days. Constitutional effects following the administration of the vaccin were noted in 90 per cent of the individuals receiving it. In 30 per cent of the cases this host response simulated in symptom-complex the early toxemia of true influenzal infection; however, the reaction subsided in six to twelve hours after the onset. This type of reaction was so striking in its analogy to the early stage of the natural infection that it strongly suggested proof of the relationship between the clinical diseases known as influenza and the artificial toxemia produced by the injection of the Pfeiffer bacillus vaccin. From the observations herein reported for our series of vacci- nations we believe there is every indication that the influenza 330 Cc. W. DUVAL AND W. H. HARRIS protein employed gives rise to the production of protective substances and therefore justifies its use in the prophylaxis of epidemic influenza. REFERENCES (1) Duvau, C. W., AnD Harris, W. H.: Jour. Infect. Diseases, (in Press). (2) Pretrrer, R.: Ztschr. f. Hyg., u. Inf., Krakh. 1893, 13, 357. (3) MinaxkeEr, A. J., AND IRvINE, R. S.: J. Am. Med. Assn., 72, 12, 847. (4) WouusteEIn, M,: Jour. Exper. Med. 1915, 22, 445. ‘ (5) McConnetu, G.: Jour. Am. Med. Assn., 72, 20, 1457. (6) Smiru, J. W., Jour. Am. Med. Assn., 72, 4, 257. (7) Repoport, F. H.: Jour. Am. Med. Assn., 72, 9, 633. (8) Perry, M. W., anp Koumer, J.: Am. Jour. Immunol., 1918, 3, 258. THE POISONS OF THE INFLUENZA BACILLUS JULIA T. PARKER From the Depariment of Bacteriology, College of Physicians and Surgeons, Columbia University, New York Received for publication July 29, 1919 It was shown in a preliminary report (1) that the filtered blood broth cultures of B. influenzae are toxic for rabbits. It was stated that the serums of rabbits which had been immunized to the poison possess neutralizing power for the poison. Whether this poison belonged to any known class of poisons was not established. It is the purpose of this paper to give in greater detail the work done to date on this poison of the influenza bacillus with especial regard to its possible classification. EXPERIMENTAL Preparation of media The poison is produced by the growth of B. influenzae on several different mediums, in fact, on any medium which con- tained rabbit hemoglobin in sufficient quantities. ‘The medium, however, from which the poison for this work was obtained, was made as follows: Veal infusion broth was prepared in the usual way with the exception that an extra amount (2 per cent) of peptone was used; it was found that the Parke, Davis and Company peptone gave the best results; the required amounts of broth were measured into Erlenmeyer flasks and sterilized by the fractional method; the final hydrogen ion concentration should have a pH of from 8 to 8.2; to these flasks enough steril defibrinated rabbit’s blood was added to make it 5 per cent of the whole; the flasks were then heated in a water bath at 75°C. until the blood coagulated on 331 332 JULIA T. PARKER standing; this requires from three to five minutes after this tem- perature (75°C.) had been reached. This medium is called for convenience the dark medium. The dark medium and the greater part of the others enumerated below were prepared with the idea of getting rid of the bacteri- cidal action of the rabbit serum either by discarding the serum or by heating it. Most of the following mediums were not used in this work, not because very toxic products could not be pro- duced from them, but because of the much larger amounts of hemoglobin they contained after filtration. Although the amount of hemoglobin contained in the original inoculation was not toxic per se, it was possible that it might be where larger doses were given in immunizing the rabbits and where multiple lethal doses of the poison were tried out. Another reason for using the dark medium was that when filtered it could even be boiled without producing a coagulum, a necessary factor in some of the experiments performed below. Other methods used 1. Washed rabbit red cells (hemolyzed with distilled water 1-3), 1 to 5 parts of broth. 2. Same as 1 except that the red cells were not washed. 3. Washed red cells hemolyzed with ascitic fluid and broth. 4. Rabbit whole defibrinated blood (not hemolyzed) with broth 1-5. 5. Rabbit red cells (not hemolyzed) with broth. Very slight if any poison could be produced from the three following mediums—which we believe is explained by the fact that the organisms grew very poorly, or not at all. 6. Same as 1 except that the broth contained 1 per cent glucose. 7. Same as 2 except that the rabbit cells were replaced by sheep cells. 8. Broth and ascitic fluid (no hemoglobin). It may be noted that the rabbits which were immunized with poisons produced on the mediums 1, 2 and 3 were resistant to the poisons produced on the dark medium. POISONS OF THE INFLUENZA BACILLUS 333 Strains used The cultures used were of very small Gram negative pleomorphic bacilli which only grew in the presence of hemoglobin. Although we obtained poisons from seven different strains of B. influenzae, most of this work was carried out with only two strains, 9 and E. We employed these two strains especially because we had found that they produced regularly the most toxic filtrates. It was found, moreover, that an immunity against a poison produced from any one strain will protect against a poison from any other strain. It was also found that an im- mune serum produced by inoculating rabbits with the poison from any two strains will neutralize in vitro at least one lethal dose of the poison from any other strain. An illustration of this latter point is given in table 1. TABLE 1 Showing the protective action of an antiserum produced by inoculation with two strains against any other strain of B. influenzae RABBIT WEIGHT POISONS SERUM REMARES MM rch iss. GW chan, 1 1600 5 ec. 1 1 ec. 340 Lived 2 1580 5 ee. 1 1 ce. 340 Lived 3 1600 5 ee. 1 Dead 1 hour 50 minutes 4 1580 5 ce. 2 1 ec. 340 Lived 5 1595 5 ec. 2 1 ec. 340 Lived 6 1610 5 ec. 2 Dead 1 hour 50 minutes Poisons | and 2 were poisons produced from two strains which Dr. Zinsser brought back from France. Serum 340 was from a rabbit which had nine injections of 9 and E, its last three injec- tions being of 15 cc. each. The animal had been bled six weeks previous to this experiment. The mixtures were incubated at 37°C. for one-half hour before injection. 334 JULIA T. PARKER INOCULATION OF MEDIA Incubation and filtration after inoculation Erlenmeyer flasks containing relatively small amounts of the dark medium, which had previously been tested for sterility, were inoculated with one slant of B. influenzae for every 50 ce. of the medium. The flasks were incubated at 37°C. for from six to twenty-four hours. It is advisable to shake the flasks several times during this time. After incubation the final pH should be between 7.2 and 7.6. If the poison has a greater acidity than this, it will only have a slight toxicity. After incubation the cultures were centrifugalized and the super- natant fluid filtered through a Berkefeld candle. The filters used were found impervious to the influenza bacilli and remained so during this work. Symptoms The incubation period in rabbits inoculated intravenously with a lethal dose of the poison is invariably thirty to forty minutes. At this time the breathing rate increases, followed usually by very severe dyspnoea. There may or may not be great weakness. An hour after the inoculation the animal is extremely sick. The temperature has risen a little at this time but not higher than 104°F. There are usually muscular twitchings in the neck. The animal may be flattened out on his abdomen either from weakness or from pain. This pain may or may not be followed by diarrhoea. The animal gets progressively worse; he is weaker; the temperature falls below normal and he usually dies within three hours after the injection with or without convulsions. If a sublethal dose is given, the incubation period is longer, the symptoms are milder and the temperature usually rises to 106° to 107°F. in one to three hours. There is almost invariably a great loss of weight in these animals, this loss being propor- tional to the dose and sickness of the animal. If the injections are given intraperitoneally the rabbit takes two to three hours to sicken and the symptoms are very light POISONS OF THE INFLUENZA BACILLUS 30D even when a larger dose of the poison is given. Subcutaneously injected, the poison is not toxic. At the point of inoculation there may be a sloughing off of the skin one to two weeks later. The poison is not toxic for guinea-pigs when injected intra- venously in 0.5 ec. doses and it does not kill white mice when 2 cc. are given either subcutaneously or intraperitoneally. PROPERTIES OF THE POISON The original toxicity of the poison may be preserved for two days if kept frozen solid, otherwise it deteriorates very quickly, even after two or three hours standing. For this reason the toxicity of the poison must be tested and the experiments made on the same day. If 2 ce. of a poison killed a 1400-gram rabbit within three hours we considered it a good one. In most of the experiments given below, the poison was tested in this way before it was used. To safeguard against the possible variation in the resistance of rabbits and against possible deterioration of the poison, two lethal doses were usually given. A few experiments were done on the effect of heat on the poison. They were not convincing. It is possible that there are two poisons concerned, one thermolabil, destroyed when heated to 75°C. for one-half hour, the other thermostabil, not destroyed even when it is boiled. This thermostabil toxicity may be due to extracted bacterial substances which all these filtrates must contain. More will be said of this later. The experiments in table 2 were directed to this point. The poison is precipitated with six volumes of absolute alcohol and appears to retain its original toxicity. This was demonstrated in the following way: six volumes of absolute alcohol were added to 6 cc. of a poison whose lethal dose was 5 cc. The resulting brownish precipitate was centrifu- galized off and dried. It was then taken up with 6 cc. of isotonic salt solution in which it completely dissolved. A rabbit, weighing 1030 grams, was inoculated with this solu- tion and died with typical symptoms in one hour and twenty min- utes. The experiment was repeated with another poison with the same result. THE JOURNAL OF IMMUNOLOGY, VOL. IV, NO. 5 336 JULIA T. PARKER The dark medium is not toxic in itself. This is shown in the following experiment: The medium was prepared as usual and incubated for twenty- four hours to insure sterility. It was then separated into two flasks A and B. Flask A was inoculated; flask B was not inocu- TABLE 2 Showing the effect of heat on the poison of B. influenzae RABBIT werca POISON REMARKS Experiment 1 grams 7 1450 | 5 cc. A, unheated Very sick 35 minutes after inoculation. Dead 2 hours, 10 minutes 8 1350 | 5 ec. A, heated at 48°C. | Sick 1 hour after inoculation. Recov- 3 hour ered 9 1310 | 5 ec. A, heated at 55°C. | Slightly sick 1 hour after inoculation. 3 hour Recovered Experiment 2 10 1620 | 3 cc. B, unheated Very sick 40 minutes after inoculation. Dead in 53 hours 11 1900 | 5cc. B, heated at 56°C. | Very sick 40 minutes after inoculation. 4 hour Recovered 12 1930 | 10 cc. B, heated at 56°C. | Very sick 30 minutes after inoculation. 4 hour Recovered 13 1860 | 5 cc. B, heated at 70°C. | Slightly sick 1 hour 15 minutes after 3 hour inoculation. Slightly sick all the afternoon 14 1950 | 10 cc. B, heated at 70°C. | Slightly sick 1 hour 15 minutes after 3 hour inoculation. Slightly sick all the afternoon 15 1850 | 10 cc. B, boiled 5 minutes | Slight labored breathing 1 hour 30 min- utes. Slightly sick all the afternoon lated. Both flasks were incubated for eighteen hours at 37°C., and then centrifugalized and filtered. The toxicity of the two filtrates was tested with rabbits. The results of these tests are presented in table 3. This protocol shows clearly that the filtrate of the medium is not toxic in itself. POISONS OF THE INFLUENZA BACILLUS aon TABLE 3 Showing that the toxicity of the filtrate from cultures of B. influenzae is not a property of the culture medium nanos | wstonr | MATERIAL [wntour arrun] ourvanewce 2 | gescxas mr ae amr | ema [Ye 16 1250 2cc. A Dead in 2 hours 17 1250 2 cc. B 1310 60 grams gain} Not sick 18 1475 10 cc. B 1520 45 grams gain} Not sick 19 1775 15 ce. B 1755 20 grams loss | Not sick EXPERIMENTS WITH ANTISERUM TO THE POISON The serum was produced by injecting rabbits with increasing doses of the poison. The serum of rabbits that had withstood at least four to five lethal doses (15 to 20 cc.) of the poison was used in the following tests. The injections were given five to eight days apart and the amounts were usually increased as follows: ce cc. EES EH AMTECEION. & og) apie.e sue <,0'e:0.0 2 Kifthanyjection=sessce sees 12 Second injection.............. 5 Sixthenjectioneeces sce 15 Ehird myjection. ...... 0.2.4. 66< 8 Seventh injection............. 18 Fourth injection.............. 10 Kighth injection.............. 20 The rabbits were bled five to ten days after the last injection. A horse also was immunized. The experiments were carried out by three different methods. A. By mixing one to two lethal doses of the poison in vitro with the serum before injection. B. By giving the serum fifteen minutes before or fifteen minutes after the injection of the poison. C. By mixing multiple lethal doses of the poison with serum before injection. A considerable number of tests were made with the use of the first procedure; namely, that of mixing one to two lethal doses of the poison with serum and incubating the mixtures at 37°C. for one-half hour before inoculation. Five of these protocols are given in table 4. 338 JULIA T. PARKER In this experiment 1 cc. and 2 ec. of immun serum 308 detox- icated a poison that killed the four control rabbits in two hours or less. Normal serum is seen to have no effect on the poison. Three out of four different immun serums neutralized a poison that killed the two controls in eight to twenty hours. The protected rabbit that died (rabbit 27) survived the controls seven days. The experiment shows that immun serum 531 protected in 1.5 cc., 1 cc. and 0.5 cc. amounts, but not in 0.25 cc. The controls died in one and a half hours and five and a half hours. Experiment 4 is similar to experiment 1, table 4, except that the poison was not as toxic and that horse instead of rabbit immun serum was used. It is seen that normal horse serum also has no detoxicating effect on the poison. Here the horse immun serum neutralized the poison which killed the two controls in three hours and forty-five minutes, and three hours and thirty-five minutes in 0.25 ec. amounts. In each case the control animals getting the poison alone and poison and normal serum were injected last in order to give the tubes containing these mixtures the benefit of a few minutes extra standing. In the third and fourth experiment in table 4 the loss of weight in the animals that survived was a good indication of their sickness. It is seen that the animals that received the greatest amount of immun serum lost the least weight and vice versa. This is almost always the rule when the rabbits are of approximately the same weight when injected. The weights in experiment 5, table 4, are a rare contradiction to this rule. Experiment 4, table 4, is given as it shows that the horse serum had some protective action even though the horse at this time had only had two injections of the poison. The horse had had 13 injections when the serums used in experiment 5, table 4, was obtained. It is seen from the foregoing protocols that the immun serum from both rabbits and from a horse neutralized in vitro at least one lethal dose of the poison. It is also seen that normal serum has no such effect. POISONS OF THE INFLUENZA BACILLUS 339 TABLE 4a Showing the neutralizing power of antisera prepared against the poison of B. influenzae Experiment 1 RABBIT WEIGHT POISON SERUM RESULTS RY | 0 TMA | any Were ty | 20 1450 5 ec. A |2 ce. 308 Lived 21 1530 5 cc. A | 1 ce. 308 Lived 22 1500 5 ec. A |2 ec. Normal} Dead 1 hour 15 minutes 23 1530 5 ec. A |1cee. Normal|] Dead 1 hour 45 minutes 24 1500 5 ec. A Dead 1 hour 20 minutes 25 1550 5 ee. A Dead 2 hours TABLE 48 Experiment 2 RABBIT WEIGHT POISON SERUM RESULT MMs el So Ue ha i Leen 26 1650 4.5 ce. B 1 ce. 308 Lived 27 1650 4.5 ec. B 1 cc. 368 Dead 8 days 28 1660 4.5 ec. B 1 ce. 369 Lived 29 1670 4.5 cc. B 1 ec. 324 Lived 30 1680 4.5 cc. B 1 ec. Normal Dead 8 hours 31 1710 4.5 ec. B Dead 20 hours UII ee ——————————e TABLE 4c WEIGHT RABBIT WEIGHT POISON SERUM RESULTS AFTER LOSS 24 HOURS grams 32 1540 | 5ee. C 0.25 ce. 531 Dead 1 hour 33 1540 | 5ee.C 0.5 ec. 531 Lived 34 1520 | 5ece. C 1.0 cc. 531 Lived 35 1520 | 5ec.C 1.5 cc. 531 Lived 36 1580 | 5ec. C Dead 53 hours 37 1600 | 5cc. C Dead 13 hours 340 JULIA T. PARKER TABLE 4p Experiment 4 RABBIT | WEIGHT POISON SERUM RESULTS 72 LOss HOURS grams grams grams 88 | 1560} 5ce. D | 2cc. Immun horse! Lived 1540 20 (3/17) 39 | 1680 | 5ec.D | 1cc.Immun horse | Lived 1330 | 350 40 | 1680 | 5ce. D | 2 cc. Normal horse | Dead during night 41 | 1725 | 5ce. D | 1 ce. Normal horse | Dead during night AD it, 00 srorces Dead 1 hour 40 minutes 43 | 1810 | 5cc. D Dead 7 days 1400 | 410 TABLE 45 Experiment 5 WEIGHT RABBIT|WEIGHT| POISON SERUM RESULTS | L088 HOURS grams grams | grams 44 | 2340} 5ecce.E | 0.25 cc. Immun | Lived 2280 60 horse (5/3) 45 | 2050} 5cc.E | 0.25 cc. Immun | Lived 1960 90 horse 46 | 2100 | 5cc. E 0.5 cc. Immun | Lived 1920 | 180 horse 47 | 2245| 5cc. E 0.5 ce. Immun | Lived 2135.|) 16 horse 48 | 2255; 5cc. E 1.0 cc. Immun | Lived 1745 | 210 horse 49 | 1880 | 5cc. E 1.0 cc. Immun | Lived 1640 | 190 horse 50 | 2855 | 5 cc. E Dead 3 hours 45 minutes 51 | 2510} 5cc. EB Dead 3 hours 35 minutes Although comparatively little work has been carried out with the serum used prophylactically or therapeutically, still there seems to be little doubt that it is effective here also. Five of the protocols are given in tables 5a, 5b, 5c, 5d and 5e. In this experiment one of the two rabbits was saved by being POISONS OF THE INFLUENZA BACILLUS 341 TABLE 5a Showing the prophylactic and therapeutic action of the antiserum upon the injection of the poison of B. influenzae Experiment 1 RABBIT WEIGHT POISON SERUM RESULTS oy oe marae Pe ot 52 1360 5 ce. A 5 ec. 307 Dead 1 hour 20 minutes 53 1380 5 ec. A 10 ec. 307 Lived 54 1400 5 ce. A 5 ec. 308 Lived 55 1410 5 ce. A 10 ce. 308 Lived 56 1400 5 ce. A 5 ec. Normal | Dead 1 hour 20 minutes 57 1410 5 ec. A 10 ec. Normal |} Dead 23 hours 58 1400 5 ec. A Dead 1 hour 55 minutes 59 1440 5 ee. A Dead 2 hours 7 minutes TABLE 58 Experiment 2 RABBIT WEIGHT POISON SERUM RESULTS rota UU ae 60 1475 5 ce. B 10 ec. 410 Lived 61 1450 5 ce. B 10 ce. 410 Dead 3 hours 45 minutes 62 1515 5 ce. B 10 cc. Normal | Dead 1 hour 35 minutes 63 1475 5 cc. B Dead 1 hour 25 minutes 64 1555 5 ce. B Dead 2 hours TABLE 5c Experiment 3 RABBIT WEIGHT POISON SERUM RESULTS TE en (a ein De aT 65 1490 5ec.C | 10 ce. 311 Dead 3 days 66 1520 5 ee. C 10 ce. Mixed immune se-| Dead 24 hours rum 67 1510 5 ec. C | 10 cece. Normal |} Dead 1 hour 15 minutes 68 1560 5 ce. C | 10 ce. Normal | Dead 1 hour 25 minutes 69 1590 5 ec. C Dead 41 minutes 342 JULIA T. PARKER TABLE 5p Experiment 4 RABBIT WEIGHT POISON SERUM RESULTS EM eran | ee eran) cea. em 70 1500 4cc. D | 10 ce. 345 Lived 71 1520 4ce.D | 10 cc. 345 Lived 72 1530 4cc.D | 10 cc. Normal | Dead 14 days 73 1546 4ece. D | 10 ec. Normal | Dead 1 hour 20 minutes 74 1535 4 ec. D Dead 45 minutes 75 1545 4cce. D Dead 1 hour 25 minutes TABLE 5r Experiment 5 RABBIT WEIGHT POISON SERUM RESULTS ip Se a em (pm pT 76 1540 5 ec. F 9 ec. 410 Lived 77 1600 5 ec. F 9 ec. 343 Lived 78 1600 5 ec. F_ | 9 ce. Normal | Dead 6 days 79 1680 5 ec. F 9 ec. Normal | Dead 1 hour 5 minutes 80 1620 5 ec. F Dead 7 days 81 1700 5 cc. F Dead 6 days injected with 5 cc. of immun serum fifteen minutes before the injection of the poison. Both rabbits getting 10 cc. of immun serum fifteen minutes after the injection of the poison survived. The four controls died; one of these rabbits (no. 57), which received 10 cc. of normal serum fifteen minutes after the inocu- lation of the poison, lived twenty-one hours longer than the other three controls. This result shows that in this case the normal serum may have conferred a slight amount of protection. The poison in this experiment was a'very toxic one. One out of two of the treated animals lived; the other died in three hours and forty-five minutes, surviving the three controls by about one hour. The normal serum did not have any effect on the poison. This is also a very violent poison. All of the five rabbits succumbed; however, the two treated animals lived a relatively much longer time than the controls. The normal serum is seen to confer no protection. POISONS OF THE INFLUENZA BACILLUS 343 In this experiment the two treated animals survived. Three of the four controls died in less than one hour and a half. The other one (no. 72), which got the normal serum, lived for fourteen days. Here again the normal serum seems to have protected the rabbit to a certain degree. In this experiment 9 ce. of the two different immun serums saved the two rabbits, while the four controls died. In this case normal serum did not show any nonspecific protective action. In this series all of the rabbits except numbers 52, 54, 56, were inoculated with the specified serum fifteen minutes after the in- jection of the poison.. The latter three rabbits were given the serum fifteen minutes before the injection of the poison. Table 5 shows that where relatively large amounts of antiserum are give either fifteen minutes before or fifteen minutes after, at least one lethal dose of the poison, the animal is usually saved. Although normal serum exerts occasionally a slight curative power, this effect, can, in no way, be compared with that of the immun serum. EXPERIMENTS WITH MULTIPLE LETHAL DOSES OF THE POISON Several attempts were made to neutralize multiple lethal doses of the poison with immun serum, but with little success. Several protocols are give below where relatively large amounts of immun serum have been mixed with multiple lethal doses of the poison. It will be seen that less than 50 per cent of the animals that received such mixtures survived. Five cubic centimeters of the poison in this experiment was probably more than one minimal lethal dose. Although both rabbits died, the rabbit getting the 15 cc. of the poison and the immun serum survived the control by twenty hours, showing some neutralization. In this experiment one of the two rabbits getting 3 minimal lethal doses of the poison and 6 ec. of the immun serum lived, while the two controls died. This experiment is similar to the preceding one. However, the immun serum showed protection even in rabbit 88 as this 344 JULIA T. PARKER TABLE 64 Showing the failure of the antiserum to neutralize multiple lethal doses of the poison of B. influenzae Experiment 1 RABBIT WEIGHT POISON SERUM RESULTS AGT i itpranse Nee, Wem be nh) 82 1450 15 cc. A 6 ec. 531 Dead 22 hours 83 1400 5 cc. A Dead 1 hour 30 minutes TABLE 638 Experiment 2 -RABBIT WEIGHT POISON SERUM RESULTS MRT |) iprame dr a Or ee 84 1340 15 cc. B 6 ec. 523 Dead during night 85 1400 15 ce. B 6 cc. 523 Lived 86 1440 5 ce. B Dead 3 hours 87 1410 5 ce. B Dead during night TABLE 6c Experiment 3 RABBIT WEIGHT POISON SERUM RESULTS grams 88 1620 6 ce. 528 Dead 7 days 89 1600 8 ce. 523 Lived 90 1680 Died during night 91 1620 Dead 1 hour 25 minutes animal survived the controls by at least six days. The mixtures were incubated for 30 minutes at 37°C. before injection. It is well known that animals of the same species may differ in their ability of producing antibodies against the same antigen. This was also the case with our rabbit immun serums which we tested in this respect in several instances. An extreme example of this point is given below. The mixtures were incubated as usual before inoculation. It will be seen that serum 536, which was obtained from a rabbit that had had seven injections of poison (the last two being 18 cc. and POISONS OF THE INFLUENZA BACILLUS 345 20 ce. respectively), had no neutralizing action, while serum 510, obtained from a rabbit which had seven injections (the last two being 15 cc. each) was relatively active in this respect. This led to a test of the agglutinins, precipitins, and complement fixing antibodies of these two serums to discover whether there was any relation between these antibodies and the neutralizing activities of a serum. Serum 536 was found to agglutinate completely B. influenzae 9 and E +++ ina dilution of 1-40; serum 510 was +++ in a dilution of 1-100. No precipitins could be demonstrated in TABLE 7 Showing a difference between two rabbits in the production of the power of neutralizing the poison of B. influenzae RABBIT WEIGHT POISON SERUM RESULTS grams cc 92 1200 5.0 1.0 cc. 5386 Dead during night 93 1200 5.0 0.5 ec. 536 Dead during night 94 1320 5.0 0.25 cc. 536 Dead during night 95 1200 5.0 1.0 cc. 510 Not sick. Lived 96 1280 5.0 0.5 cc. 510 Not sick. Lived 97 1280 5.0 0.25 ce. 510 Sick. Lived 98 1350 5.0 Dead during night 99 1320 5.0 Very sick. Lived either serum. Serum 536 (no. 10 in table 8) fixed complement completely, +++-+, in 0.01 cc. amounts, while serum 510 (no. 9 in table 8) fixed completely in 0.001 cc. amounts. In these two instances, therefore, the complement fixing antibodies ran parallel with the power of neutralization. This was also the case with no. 4, table 8 (compare with neutralizing properties, table 4, experiment 5) and with no. 12, table 8, which we found neutralized easily in 0.25 cc. doses, at least one lethal dose of the poison. A number of recent and old immun serums were then tested for their complement fixing power. Some of the results obtained are given in table 8. The antigen was made of cultures 9 and E and one-third of the anticomplementary dose was used in the tests in table 8. The animals were bled on the dates noted in table 8. 346 JULIA T. PARKER It is seen from this table that, with few exceptions, the oldest serums have the least power of fixing complement. In only one case did we try the neutralizing action of anoldserum. Although this serum had previously neutralized a lethal dose of the poison TABLE 8 Showing the complement-fixing power of various antisera produced against the poison of B. influenzae AMOUNT OF SERUM S E 3 § S s E REMARKS Horse Sages Ese Ee 0 SON oleate te | oe got late stacteada| sale its 0 0 0} 0.1 and 0.01 slightly anti- complemen- tary NORE eo oe ae el ones ee ore RL Leen Ue ey 0 0} 0.1 and 0.01 slightly anti- complemen- tary Bs Altes ale tea tc ata oh teal ects ate teeta 0 0) 0.1 and 0.01 slightly anti- complemen- tary Rabbit 636, 5/3") SIFE--F=F] | + =e 0 0 308, 12/6 | 6) + 0 0 0 306, 12/21] 7) + 0 0 STD ited NET Mrs) ar a ec 0.1 slightly an- ticomplemen- tary S10) S27 ao eee teat oa a ae) tet a ia 0 0 OS0, (oi20 WO ao tedodnt-ot ta) 533, (4/2419 |Past} 0 0 0 0 572, 6/11 [12\++++|+++4+]+++4]++4+4/44+4+4/44++4|4 in 1 cc. amounts, when tested again four months later, it had lost all this power. These facts are noted here because they seem to indicate the possibility that the neutralizing power of a serum runs parallel with its complement fixing power. More light is thrown on this phenomenon by experiments cited below. POISONS OF THE INFLUENZA BACILLUS 347 Although it seemed very improbable that any other antiserum produced by inoculating some other organism into rabbits could protect against the influenza poison, still in order to set aside any possible nonspecific action of an immun serum produced from another organism, 1 and 2 cc. of a high titer typhoid rabbit serum were incubated with a lethal dose of the poison of BP. influenzae. There was no neutralization. THE POISONS IN THE VACCIN OF B. INFLUENZAE The question as to whether our poison could be produced from dead bacteria, fresh or autolyzed, whole or filtered was next investigated. It seemed possible that the poison of B. influ- enzae might be related closely to those produced from dysentery (Shiga Kruse), typhoid and cholera cultures by Kraus and Doerr (2), Kraus and Stenitzer (3), and Kraus (4), respectively. These observers found that their poisons were contained in bac- terial extracts as well as in broth filtrates and that antiserum produced by immunizing either with the extracts or with the filtrates produced equally good serums against the poison. Pfeiffer (5) in his original work on B. influenzae found that the same symptoms were caused in rabbits by the inoculation of dead or living cultures and that the same amount of culture was necessary to kill in either case. It was, therefore, thought advisable to test out the toxicity of vaccins of B. influenzae. EXPERIMENTS WITH VACCIN OF B. INFLUENZAE The vaccin was prepared as follows. The heavy eighteen hour growth of B. influenzae (9 and E) on coagulated rabbit blood slants was washed off with isotonic salt solution, 2 cc. to each tube. 0.1 cc. of normal NaOH was added to each 100 ce. of vaccin. After heating for one-half hour at 58°C. the bottle was shaken in the shaking machine for one and one-half hours. Microscopic examination of such material at the end of this time showed the bacilli to be well broken up and staining poorly. The autolyzed vaccin (B) was prepared in the same © way, but was incubated at 37°C. for twenty hours in addition. 348 JULIA T. PARKER Both vaccins were very opaque and sedimented extremely slowly. EXPERIMENTS WITH FRESH VACCINS It will be noticed in table 9 that antigen A (fresh vaccin) is very little toxic. The difference in resistance of rabbits to the bacterial antigen is well illustrated in this experiment by the fact that rabbit 102, which had received an injection of only 0.5 ec., was much sicker and lost a correspondingly greater amount of weight than did the other rabbits. This variability in rabbits was shown in almost all our experiments with the vaccin and was in marked contrast to the experiments with the broth poison, where the resistance was fairly even. TABLE 9 Determination of the toxicity of a vaccin of B. influenzae WEIGHT RABBIT | WEIGHT VACCIN RESULT eae LOSS 100 | 1430 | 2.0 cc. A | Sick 1 hour after injection; quick breath-|} 1250 | 180 ing. Better after 2 hours A | Slightly sick 1 hour after injection 1235) 165 102 | 13880 | 0.5 cc. A | Very sick 1 hour after injection lying on} 1105 | 275 abdomen. Diarrhoea. Sick all the afternoon 103 | 1380 | 0.25 ec. A | Very little sick 1235 | 155 EXPERIMENT WITH AUTOLYZED VACCIN The vaccin B (autolyzed vaccin) was possibly slightly more toxic than vaccin A. The symptoms of the rabbits were the same in either case. It was our intention to learn the lethal dose of this antigen and see whether a good immun serum to the broth poison would protect against it. This was tried twice with- out showing any neutralizing activity of the serum. One protocol is given below, table 10. One cubic centimeter of the horse serum which protected in 0.25 cc. amounts (see table 3, experi- ment 5) was used in the test. The mixtures were incubated at 37°C. for one hour. POISONS OF THE INFLUENZA BACILLUS 349 EXPERIMENTS WITH VACCIN FILTRATE The Berkefeld filtrate appeared to have about the same toxicity as the whole vaccin. A dose of 3.6 of vaccin A filtrate was given to a rabbit weigh- ing 1150 grams. He was sick all day and died during the night. A smaller dose (2 cc.) was given to a 900-gram rabbit. This animal survived. The toxic properties of the vaccin are therefore filtrable, a fact which, of course, was to be expected, but which we thought best to prove by experiment before going on. TABLE 10 Showing the failure of a broth poison antiserum to protect against the toxic action of a vaccin of B. influenzae RABBIT |WEIGHT| ANTIGEN SERUM RESULTS 104 | 1660 | 3cc. B | 1 ce. Immune horse | Very sick but recovered 105 | 1800 | 3 cc. B | 1 ce. Immune horse | Very sick. Dead during night 106 | 1870 | 3cc. B Very sick. Diarrhoea. Dead 3 days B 107 | 1810 | 3ce. Sick, but recovered EXPERIMENT WITH BOILED FILTRATES OF THE ANTIGEN As the broth poison was still toxic even when boiled it was thought advisable to test the vaccin filtrate under the same conditions. In this experiment the vaccin was prepared as be- fore with the exception that the bacilli were killed by shaking with toluol instead of by heat. It was found that the boiled vaccin filtrate was slightly more toxic than the unboiled. An example of this is given in table 11. Five cubic centimeters of the filtrate corresponded to one slant of culture. From the foregoing experiments with bacterial vaccin, it does not seem likely that the bacterial extractive substances as ob- tained above possess more than slight toxic properties and these would appear in the broth filtrates together with the true toxins. 350 JULIA T. PARKER It seemed possible that the toxicity remaining after heating to 75°C. or over might be due to these substances. With the hope of throwing more light upon this bossialates the following experiments were carried out (tables 12 and 18). The broth poison was heated to 75°C. for one-half hour before the serum was added. The mixtures and poisons alone were incubated as usual at 37°C. for one-half hour before injection. An explanation is required before discussing this experiment, table 12. On reviewing this work, there appear to be two pos- sibilities in regard to the poison of B. influenzae. TABLE 11 Showing the effect of boiling on the toxicity of the filtrate from the vaccin of B. influenzae WEIGHT RABBIT WEIGHT VACCINE FILTRATE RESULTS are ie Loss HOURS grams grams grams 108 1200 BrCG: Very sick. Recovered 1180 20 109 1190 5 ec. Boiled | Verysick. Dead4days} 1120 70 1). That the filtrate contains only one poison; that heating it merely destroys part of its toxicity and that the remaining toxicity is the same as its original toxicity except that it is attenuated. 2). That this poison of B. influenzae contains two poisons: one, a true toxin, which is only produced during the growth of the bacilli on blood broth, which is filtrable, and thermolabil, and which may be neutralized in multiple lethal doses by the immun serum; the other, also filtrable, but thermostabil and not neutralized by the immun serum. Considering table 12 from this point of view it is seen, in the first place, that except for a slight amount of sickness and con- siderable loss of weight (220 grams) 5 ec. of serum 572 was able to neutralize three lethal doses of the unheated broth poison (see rabbit 110). On the other hand, although 15 cc. of the heated broth poison is only one lethal dose (see rabbit 111), rab- bit 112, which was inoculated with 15 cc. of the heated poison POISONS OF THE INFLUENZA BACILLUS 351 (one lethal dose) and 5 cc. of serum 572, was, if anything, sicker than rabbit 110. If the immun serum had the same effect on the heated poison as on the unheated, one would have expected the heated poison given rabbit 112 to have been entirely neutralized. But rabbit 110 was, if anything, less sick and lost less weight than rabbit 112. Hence, it seemed most likely that the immun serum TABLE 12 Showing the neutralizing effect of antiserum 572 on the poisonous action of the heated and unheated broth poison of B. influenzae WEIGHT RABBIT | WEIGHT POISON SERUM sora LOSS RESULTS HOURS grams grams grams 110 | 1600 | 15 ce. 5 ec. 572} 13880 | 220 | After 1 hour diarrhoea slightly sick all the afternoon. Lived 111 | 1610 | 15 ce. heated 1170 | 440 | After 45 minutes sick, dysp- noea. Sicker than rabbit 108. Dead 6 days 112 | 1620 | 15 cc. heated |5 cc. 572| 1350 | 270 | After 55 minutes sick. Di- arrhoea. As sick as rabbit 110. Dead 6 days 113 | 1640 | 5 ce. After 40 minutes sick. Di- arrhoea. Lying on side. Very sick all the afternoon. Dead 21 hours 114 | 1660 | 5 ce. After 40 minutes sick. Dysp- noea. Very sick all the afternoon. Dead _ during | the night. neutralized a poison that the heated broth did not contain. The fact that rabbit 111 was sicker and lost more weight than rabbit 112 seems to indicate that the serum was in a slight way pro- tective in this case also. It was thought that this might be a nonspecific action aud that normal serum would possess ‘this same power, just as it may have helped in a slight degree when the normal serum was used curatively (see table 5). This was investigated further in table 13. THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 5 352 JULIA T. PARKER The following points are brought out in table 13. It is seen that 5 ce. is the lethal dose of the poison and 20 cc. of the heated poison. The only rabbits that survived were nos. 115 and 118, the former getting four lethal doses of the unheated poison and 6 ce. of serum 572 and the latter one lethal dose of the heated poison and the same amount of serum. Rabbit 118 was sicker than rabbit 115, which corroborates the findings in table 12; namely that the immun serum has not the same effect on the unheated as the heated poison. No other deduction can be made from this experiment. Normal serum had some detoxicating effect on the unheated poison (see rabbit 116) while it had the reverse — effect in the case of the heated poison (see rabbit 119) in which ease the result was probably due to unusual susceptibility of this rabbit to the heated poison. TABLE 13 Comparison of the neutralizing effect of normal and immun serum upon the poisonous action of the heated and unheated broth poison of B. influenzae WEIGHT AFTER RABBIT | WEIGHT POISON SERUM RESULTS 22 Loss HOURS grams grams grams 115 | 1500 | 20 ce. 6 ec. 572 Lived 1380 | 120 116 | 1500 | 20 ce. 6 ec. Normal) Dead 4 days 1310 | 190 117 | 1550 | 20 cc: heated Dead 5 days 1450 | 100 118 | 1510 | 20 cc. heated! 6 ee. 572 Lived 1350 160 119 | 1515 | 20 cc. heated} 6 ec. Normal) Dead 1 hour 20 minutes 120 | 1560 | 5 ce. Dead 1 hour 15 minutes 121 | 1580 | 5 ce. Dead 1 day 1400 | 180 EXPERIMENTS WITH ANTISERUM TO THE VACCIN These experiments were undertaken to discover whether a serum obtained by immunizing rabbits with the vaccin could detoxicate the broth poison in vitro. If this could not be done it would be an added point in favor of the theory that the toxicity of the vaccin and broth filtrates were different. It was difficult to immunize rabbits against the vaccin. Four cubic centimeters of the vaccin (the bacilli from two slants) POISONS OF THE INFLUENZA BACILLUS 353 was the largest quantity injected. However it was compara- tively easy to obtain a high titer serum. ‘The serum used in these tests fixed complement in a quantity of 0.00025 to 0.00001 ec. Our best antiserums to the broth poison fixed in 0.0005 ce. to 0.00025 ec. and detoxicated at least one lethal dose of the broth poison in a quantity of 0.25 cc. Two protocols of this work are given in tables 14 and 15. TABLE 14 Showing that the antiserum to the vaccin of B. influenzae lacks power to detoxicate the broth poison WEIGHT RABBIT WEIGHT POISON SERUM RESULTS 24 LOSS HOURS grams cc grams | grams 122 1410 5 1 ce. 567 Very sick. Recovered 1230 | 180 123 1510 5 0.5 ec. 567| Dead 45 minutes 124 1380 5 Dead 55 minutes 125 1420 5 Very sick. Recovered 1325 95 TABLE 15 Showing that the antiserum to the vaccin of B. influenzae lacks power to detoxicate the broth poison RABBIT oa WEIGHT POISON SERUM RESULTS Ne en | a ee 126 1350 5 1 ec. 548 Dead 5 days 127 1280 5 1 ec. Normal Dead 1 day 128 1300 5 Dead 4 days 129 1229 5 Dead 2 days Serum 567 obtained from a rabbit that had had three injections of vaccin and whose serum fixed complement completely in a quantity of 0.00025 cc. and partially in a dose of 0.0001 ce. The broth poison was made, as usual, on the dark medium. In this experiment (table 15) one control and one treated animal died. No protection can be made out here, for of the two animals that survived the one getting the 1 cc. of serum 567 lost more weight than the control. Serum 548 fixed complement completely in a dose of 0.00025 354 JULIA T. PARKER ec. Although the poison used in the above experiment was a weak one, serum 548 did not save rabbit 126. From the foregoing experiments it is seen that the power to fix complement bears no relation to the power to neutralize the broth poison. Thus, it appears that in antiserums produced by immunizing with this poison, the power to fix complement and to neutralize the poison run parallel, not because these anti- bodies are in any way related to each other in the detoxication of the poison, but because rabbits that can produce a serum of high titer for fixing complement are also the best for the pro- duction of antitoxin. The last point to be considered was whether the filtrate of the dark medium after incubation with dead B. influenzae would be poisonous. The poisons in this case would be in the nature of proteotoxins. This was tried out three times and in one case the dead bac- terial filtrate made a rabbit sick. ‘This protocol is given below. Experiment. Two hundred cubic centimeters of the dark medium was prepared as usual and was separated into two flasks A and B, 100 ec. to each flask. A was inoculated with 1 slant each of E and 9 alive; B was inoculated with 1 slant each of E and 9, which had been heated to 58°C. for one-half hour and proved steril. After eighteen hours incubation A and B were centrifugalized and filtered. Rabbit 130, weight 1690 grams, was injected with 4 cc. of A. Forty minutes after the injection the animal was very sick with severe dyspnea. Later it became weak and died during night. Rabbit 131, weight 1760 grams, was injected with 4 cc. of B. One hour and twenty-five minutes after injection the animal exhibited weak, rapid breathing. Recovered. In the other two experiments there was slight if any sickness in the animals given the B filtrate even when twice the amount given the control was injected. We also tried incubating the flasks with dead bacilli for forty-eight hours without, however, observing any increase in toxicity. It is possible that rabbit 131 was especially susceptible to the bacterial extracts which, of course, are contained in small amount in these filtrates and to ey POISONS OF THE INFLUENZA BACILLUS 355 which, as previously shown (table 9), rabbits vary greatly in their resistance. We hope to clear up this point later. DISCUSSION In this work an effort has been made to classify the broth poison of B. influenzae. ‘To date this has not been possible. It was of first importance to ascertain whether this poison is a true soluble toxin. Although it possesses some of the attributes of such toxins, it lacks others. The fact that the antiserum will detox- icate, in vitro, three minimal lethal doses of the poison in only about 50 per cent of the cases, even when an excess of serum is used, speaks against its being a true toxin. On the other hand there is a possibility that the poison is a combination of two poisons; the first a true soluble toxin, which is produced during the growth of B. influenzae and can be neutralized by the antiserum in multiple lethal doses; the second a thermostabil filtrable poison, which is found also in bacterial extracts and against which the antiserum has no powers. Ac- cording to this conception, the animal receiving multiple lethal doses of the broth poison and the immun serum would not die from the toxin, but from a lethal dose of the thermostabil poison present in the filtrate. Being unable to prove that the broth poison is a soluble toxin, we next tried to determine whether it could be similar to the dysen- tery, typhoid, and cholera poisons. In this we also failed, for these latter poisons could be produced as well from dead bacterial extracts as from broth filtrates, which was not possible with our poison. Not enough work has been done on the possible proteotoxin formation of the Pfeiffer bacillus on the blood broth mediums. The fact, however, that our poison can be neutralized in vitro with relatively small amounts and in vivo with larger amounts of antiserum, speaks against its being a proteotoxin. It must also be remembered that rabbits can be immunized with great ease to four or five lethal doses of the poison, which has never been done with anaphylatoxin, produced by whatever method. 356 JULIA T. PARKER CONCLUSIONS 1. Bacillus influenzae produces a filtrable poison which is lethal to rabbits when given intravenously. 2. The poison is only partly destroyed when heated to 55°C. for one-half hour. When heated to 75°C. for one-half hour or boiled for five minutes, over two-thirds of its toxicity has been lost. 3. Rabbits can be immunized to at least four or five minimal lethal doses of this poison. 4, One-quarter to 1 cc. of the immun serum can neutralize in vitro one to two lethal doses of the poison. 5. Five to 10 cc. of the immun serum, when given intrave- nously fifteen minutes before or fifteen minutes after the injec- tion of one to two lethal doses of the poison, will usually save the rabbit. 6. Five to 8 ce. of immun serum, when mixed in vitro with at least three minimal lethal doses of the poison, will save about 50 per cent of the rabbits. Influenza bacterial extracts, fresh or autolyzed, are poisonous to rabbits in relatively large amounts. The symptoms are the same as with a sublethal dose of the broth filtrate. A dose corresponding to one and one-half heavily grown slants of B. influenzae will kill 50 per cent of the rabbits injected. 7. The Berkfeld filtrate of the bacterial extracts is nearly as toxic as the extracts themselves. Boiling this filtrate does not destroy its toxicity. 8. The immun serum has no effect in vitro even in large amounts in detoxicating the bacterial extracts. 9. Antiserums produced by immunizing with vaccins of B. influenzae do not neutralize in vitro a lethal dose of the broth poison. 10. While the evidence is by no means conclusive, it seems probable that the poison of B. influenzae contains two poisons; the first, the more important one, a true soluble toxin, filtrable, ther- molabile, against which antitoxins can be produced; the second, present also in the vaccin of B. influenzae, also filtrable, but POISONS OF THE INFLUENZA BACILLUS 357 differing from the first poison in its thermostability, and in the fact that it is not detoxicated by the antitoxin. I am much indebted to Dr. James G. Dwyer and Dr. Stewart Polsom for assistance in the inoculations of the horse and to Miss Lucy Pope for technical help with the cultures. I am also much Indebted to Mr. James J. May, who has been of the greatest assistance in all the work connected with these experiments. REFERENCES (1) Parker, Juuia T.: J. A. M. A., 1919, 72, 476, (2) Kraus anp Donrr: Zeit. fur Hygiene 1906, 55, 1. (3) Kraus anp Srenirzer: Men. klin. Woch. 1907, 20, 344. (4) Kraus: Men. klin. Woch. 1907, 20, 1280. (5) Preirrer: Zeit. f. Hygiene, 1893, 13, 357. ON THE EXISTENCE OF A MULTIPLICITY OF RACES OF B. INFLUENZAE AS DETERMINED BY AGGLUTINATION AND AGGLUTININ ABSORPTION EUGENIA VALENTINE anp GEORGIA M. COOPER From the Bureau of Laboratories, Department of Health, New York City Received for publication July 31, 1919 In the general investigation of the bacteriology of the recent epidemic of influenza, carried out under the supervision of Dr. Wm. H. Park and Dr. Anna W. Williams,! it was early seen that one of the most important questions was the identity or non- identity of the strains of B. influenzae from epidemic cases. Although the practically uniform presence of B. influenzae in actual cases suggested its etiological importance, the further demonstration was needed that the strains encountered were identical before a conclusion was possible. It would be difficult to conceive of a pandemic spreading from country to country, due to different races of a microédrganism. Preliminary observations indicated that the relationship of the strains isolated could be determined most easily by agglutina- tion and agglutinin absorption. It was found that agglutinating sera, active in dilutions of 1-800 to 1-1000, could be produced without great difficulty by intravenous injection of rabbits, starting with live or killed bacili. The following schedule will serve as a general outline of the method we have employed. Increasing amounts, first of heated suspensions, later of live suspensions were injected intravenously into rabbits. The doses were increased from an amount equivalent to the growth on + of a slant to that on 1 to 2 slants. The organisms were culti- vated on “‘chocolate” agar? in 6 by 2 inch tubes with about a 1 We are indebted to Dr. Charles Krumwiede for direct technical supervision. 2 Glycerine veal agar, 10 cc.; citrated horse blood, 0.5 cc.; add blood to the boiling hot agar; mix thoroughly and slant. 309 360 EUGENIA VALENTINE AND GEORGIA M. COOPER 13 inch length of slant surface. Injections were made on three consecutive days, and then a rest period of four days intervened. After two series of injections, on the fourth day following the last injection of the second series, a trial bleeding was tested. If the titer of the serum was satisfactory, the rabbit was etherized and bled to death. If it fell short, the injections were con- tinued as outlined above and trial bleedings were tested on the fourth day after the last injection of each series until the serum was found sufficiently potent. With the majority of strains, the immunization required about four weeks. The agglutination antigen was prepared by scraping into 0.8 per cent salt solution the growth of a twenty-four-hour culture on ‘‘chocolate’’ medium. Vigorous shaking by hand was suf- ficient to give an even suspension of the bacilli. Some strains encountered had a tendency to clump spontaneously, but these clumps were extremely unstable, and on slight shaking, could be easily distinguished from serum agglutination. The suspension made on different days but from the same cultures, showed only a slight to moderate degree of variation in regard to their agglu- tinin ability. In spite of such variations, the general results, as shown in the tables, have been relatively uniform and repeated tests have been suffic ent in number to exclude any possibility of error due to this factor. The technique for the agglutination tests was as follows: The sera were diluted to 0.1 of the final dilutions given in the tables; 0.1 of serum dilution and 0.9 of the bacterial suspension were pipetted into tubes and thoroughly mixed by shaking. Very thin suspensions, that is, suspensions showing only moderate cloudiness, were found to give the sharpest readings and were, therefore, employed. Uniform density was approximated by reading type through the suspension in a test tube 1 inch in diameter. Two cultures failed to agglutinate on the first trial. However, repetition gave sharp results. Differences so marked were most probably due to error n selecting cultures for sus- pension (confusion of names and numbers). The tests were in- cubated in the water bath at 45°C. for two hours. Readings oe MULTIPLICITY OF RACES OF B. INFLUENZAE 361 were made at the end of this time; these readings were checked by a second one made after the tests had been in a cold room over night. These readings differed to a negligible degree; some strains showed better agglutination at the first reading and others at the second. The second readings were on the whole more difficult to make than the first, because of the tendency of the organisms to settle to a much larger degree even in the absence of actual agglutination. The earliest readings are re- ported in the tables. The symbols +, +, X and — were adopted to show the grades of agglutination. By +, a complete agg!u- tination is indicated. The clumps varied from large to medium in size with different strains, but the supernatant fluid cleared as the clumps settled. By +, a less complete reaction is indi- cated; the clumps were of smaller size and the supernatant fluid, after the settling of the clumps, showed slight opalescence. A X is used to indicate still less marked reactions than the preceding, as slight clumping or a settling with no coherent clumps; in short, any trace of reaction occurring beyond that found in the control tubes. The first strains selected for study, with one exception, were isolated at post mortem from lung or larynx. If there were an epidemic strain it seems that the strains having the ability to extend from the nasopharynx and invade the lung would most likely be the actual epidemic variety rather than an accidentally present parasitic type. As has been shown, the influenza bacillus is very frequently encountered in the nasopharynx of apparently normal individuals as well as in the throats of persons suffering from diseases due to other agents. The preliminary strains were nine in number, and an agglutinating serum was prepared for each. For comparison, a culture of B. influenzae (B. I.) isolated in 1914 from the lung at autopsy in a case of pneumonia and a serum prepared against the culture were included. The results of the direct agglutinations are shown in table 1. The general absence of cross agglutination, or the slight degree when present, as shown in this table, indicates that even among these ten strains there exist dissimilar varieties. To exclude the possibility that the cross agglutination obtained =|+|+)+ = =| | ==] =e EI] —| ||) eK ee =|—|—|—|— ||| Kel el —|—|=|=|—|-|-|-|- —|—|—|—|=|—|-14-]4]4]+4]4+!-1-|=|-|=!-|=]-|-|-|-I-1l-|-}- |" | Fl} | ]F III] Ir rire} —|—|X}=|+]+|-]—|-|}<|=|-|-|-|-]X|-|-]-]-|- =|+|+|+/+}—|-|-|-|-|-|-|- ee (lial felis +\=|=|=|=|— =|=|=|=|—|=|=|—|— —|—|—|—|—|=|=|—]—]—|=|-|=|=|=|—|-|4 44 III =|]=]-|-|-|-} |<)" IF |- IF IFIFll |e] ir} Sa eee fel eee etal ae | | | ee | i Peal tell eee omen lb ied ed ee = sl er Fel el fn ees ee! exec P| ame Pelee celle] cal fc —|=|—|—|-|=|-|-|-|-|=|=|=|-|-|-|-|-|-l-l-l-l-1-]-]-]-]-]-]"}-}- 1-17 —|—|-+]+)+)+)+|-|-|-|-|-|-|-|-|-|- SS) SE el rel Sr eee hol cel lef oe fe fs fce Sle Sls SSS Sl Foe] Kamer al el naff al aL f(a —|—|x|=|+/+]—|—|x|*|*/-|—|—|x]#]—-|—|]#]#]-]-]<||]41=|-|-1<|=]-|-|-|-]X[-]- ilps peel ee SISISISISISlSlSslSlSlelselselSlSlelelelelslsisisisisislsisieisisisislisisislslslSlsliSisisisiSlSlSlslSlsls Os Poms cL |e VL | SAU et Sa ea I Lc Sol hee | el eee g 5 4(¥) ‘1a urdureyy AdIJpor Sd}VSVIT SUIBIT[TM uvyely Ipsl9AVeT sayAg aa, uosueg Ps cc a 8S ee Se “UOTPVUIJNIISB 99B1} 10 FBIM “UOLJVUTYNI SSB SUOI4S UOTZBUIYN[ SSB oyopdui0od = ‘PIGE poyepost “T “a + ‘survays Asdoyny , Vugas ONILVNILOISDV suoynuyn bBo 799LUq «Tl GIAVL I dl sae *KaIJpor) "S99 BSBA — Sure tM: “+ aByBA, SNIVULS 362 MULTIPLICITY OF RACES OF B. INFLUENZAE 363 was any indication of close relationship or identity, agglutinin absorption was resorted to. The method of absorption was as follows. The growth from twenty-four-hour cultures on “‘choec- olate’’ medium was scraped into 0.8 per cent NaCl solution and centrifuged until the organisms were sedimented and tightly packed in the tip of the centrifuge tube. Sufficient serum was added so that the volume of the bacilli to the total volume of serum dilution and bacilli would be as 1 is to about 6 to 10. Depending on the volume of bacilli, the amount of serum needed was cal- culated and sufficient saline was added to give a final dilution of 1-15. To compensate for the water in the packed mass of organisms, the volume was considered as one-half water. This in actuality is an over calculation so that the subsequent dilu- tions are.actually lower than stated. We intentionally adopted this procedure as it gave no advantage to the test, therefore adding to the conclusiveness of the findings. The organisms were thoroughly distributed in the diluted serum; first by breaking up the clump at the tip of the tube with a platinum loop, then by shaking until an even suspension was secured. The mixtures were incubated in a water bath at 45°C. for three hours and in order to keep the organisms con- stantly in contact with the serum, the tubes were shaken every fifteen minutes during this period. The mixtures were then stored in the ice-box over night. On the following morning, they were centrifuged and the supernatant fluids were used for agglutination in the requisite dilutions. Control experiments have shown that the absorption mass was three to four times the amount necessary to obtain complete absorption of a serum when the homologous strains were used. It seemed preferable to employ a large absorption dose to bring out any common although quantitatively different absorptive capacity, even though this introduced a possible source of error due to non- specific absorption. Further control experiments since carried out, have shown that the latter does not occur with doses of this size. Thus 2 cc. of serum diluted 1-20 was absorbed with masses of culture ranging from 0.1 ce. to 1.75 ce. of packed culture. The heterologous strain used for these absorptions 364 EUGENIA VALENTINE AND GEORGIA M. COOPER gave considerable direct cross agglutination with the serum em- ployed. In spite of this, even the largest dose of this strain did not appreciably diminish the agglutinins for the homologous serum strain. The results of the absorptions with the strains and sera already described are given in table 2. The results given in table 2 indicate that the ten strains studied are different one from the other, or in other words from nine cases of influenza were isolated nine distinct races of the influenza bacilli. This result was so surprising that it seemed advisable to extend the study to another series of cultures, especially those isolated in the early stages of the disease; also from cases during convalescence, as well as strains isolated from apparently normal throats. An additional series of cultures obtained post mortem from lung, larynx, or trachea also, six strains isolated by mouse passage, from sputa of suspected influenza pneumonia, were included. Two antisera prepared with nasopharyngeal strains isolated early in the disease were employed in addition to the nine sera employed in the previous series. The results of these aggluti- nations as well as the strains employed are given in tables 3 and 3a. The results in these tables again indicate relatively little relationship between the strains employed and the strains used for the antiserum. In some instances, a considerable degree of cross agglutination was noted. Te determine again whether this indicated a close relationship or identity, absorption was resorted to. The results presented in table 4 show that, in two instances, identity in absorptive capacity were encountered. The Sykes post mortem strain was identical with the L.4. nasopharyngeal strain isolated during convalescence. The Laverdi post mortem strain was identical with the mouse 2 strain obtained from sputum by mouse passage. Possible contact between the individuals from whom the first two strains were isolated can be definitely excluded. The other two strains were isolated from sailors admitted to the Willard Parker Hospital from a United States MULTIPLICITY OF RACES OF B. INFLUENZAE 365 Receiving Ship within eight days of each other. In this instance, there is a probability of contact. Conclusive data is not available. Although these two instances of identity were encountered, the general absence of relationship of the strains from this series TABLE 2* Agglutinin absorptions AGGLUTINATION AFTER ABSORPTION AGGLUTINATION BEFORE ABSORPTION Of absorbing | Of homologous SERUM ABSORBED BY strain strain e Strains | o|S/S/ 8/8} 2 o/S/S!S!/Slo/s/sisis 6) a) A] a] | O SSS SS Sle Sse Benson +/+/+/+/+]/—| Benson |—/—/]—/—/—/—/-—|— ee | ae Benson..... Lee +\/+)/#/-|-|-| Lee |J—j—|—|-|-— ++} } +e] Godfrey |+|+|+|X|—|—| Godfrey |—|—|—/—|/— j++} |} Benson |+/+/+/X/-—|/—| Benson |—|/—/—/—|— fe} || WG sess’ Lee +H/+/+/+/+/-—| Lee 8 j—j—J—/—|—/—|-—/—/-J-— Godfrey |+|+/+|+|X/—| Godfrey |—|—|—|—|-— +i+/+]4+]x Benson +!+/X/—|—|—| Benson |—/—|—/—|— Se ee ae Se Wirahani....<: fy Bea P e 8 ag8 NT FE Benson |+/+/+/—|/—|—| Benson |—|—|—/—|— feat ate | atte Williams....4} Williams |+/+/+|+/+/—| Williams |—|—|—/—|—|—/—|—|—/— Godfrey |+|+|X|—|—|—| Godfrey |—|—/—|—|— ee |e ere ee Benson +/=/X/—|—/—| Benson |—/—|—/—|— te eee Laverdi |=/=|X|—|=-|—| Laverdi |—|—/—|—|— Se a Pe eee Masates /+/+/+|/+/+|/—| Masates |—|}—/]—]—|—|]—|—]—|-—|-— Godfrey |+|+/+|X|—|—| Godfrey |—|—|—|—|— BE ee) eee Benson +/<|/—|—|—|—| Benson |—|—/—|—|— i+] +e} Godfrey.....4| Williams |x|—|—|—|—|-— Williams |—|—|—|—|—|+|+]+/+|]= Godfrey |+|/+|+/+)+|—| Godfrey |—|—|—|—|—|—|/—|—|-—|- Benson +/+/+/<|/—/—| Benson = ||} — Peele eK Rampin.....)| Williams |+|X/X|—|—|—| Williams |—|—|—|—|—|+/+/+/+|x \) Rampin § [-F]-+/+/+/+|—| Rampin” |—|—|—|—|=|—|—|=—|-|— * Autopsy strains. Direct agglutinations in table 1. 366 EUGENIA VALENTINE AND GEORGIA M. COOPER STRAINS AGGLUTINATIN Be LL; M II At Benson Group names | Condition| — Period of disease IS Oana ae sane Influenza? {911 == olla allies nae ia 1 Ae: ESO a ach ate Infiuenza | Convalescent, 6 weeks ra ead fe bol a 1 Rese SUM a aeaS Influenza | Convalescent, 5 weeks oll aloes 1,7.....2+-+++-.| Influenza | 4 days Ta: Base ered kare sets Influenza | Convalescent, 6 weeks | |—|—|—|—|—|—|—|—|— Sleds ends Influenza |5days 90 J JJ Sf FS lle Sik iec sameeree Influenza | 1 day Sis eee ee Influenza |5days J J—J — J J — J J - I] | i -I- Pi a ee ake SOE Oe Influenza | 3 days Sid aeatke ieee Influenza | 4 days DiGi -sacenee Influenza | 5 days Civ Beaanorenas ds Influenza | 5 days Siti ts Seeregetene an ace Influenza | 4 days MTG iatciaw ere the Influenza | 3 days AE I eee oar Influenza | 3 days MAG Sich acess oS Influenza | 3 days 1 8 2 Dee pea ee Jie Influenza | 3 days MELD. acters Influenza | 3 days MGa Gace eee Influenza | 3 days Saree dates oe Influenza | 3 days —|=|-Fl=|X|-!— |= |= |xl— G2 says eee Influenza | 1 day Sle lS lS 1 | |SIS 1S = Sf i | Hud............| Influenza? | Convalescent, 4 weeks |—|—|—|—]—|—]—|—]—|—|-—|— Rech i024. 255 Influenza | 3 days Recs lessees Influenza | 4 days J Eb iobeade Sane Influenza | 3 days Grahsssoesno.6e Influenza | 3 days 20 bth eaeeeae Influenza | Autopsy trach'; s32.<0- Influenza | Autopsy Wielunpre. act ce Influenza | Autopsy (LEACH ae ee Influenza | Autopsy Calungie, feet Influenza | Autopsy trache-b.. 5. Influenza | Autopsy IB Atbach) a eet: Influenza | Autopsy ID} bi pey ee ane Influenza | Autopsy IPAtraca teens. Influenza | Autopsy Mouse Tia...) 3 Inf. pneu. | 7 days MOUBE:2 4 ...j5\5-5 Influenza | 6 days Mousses: 22-20. Influenza | 4 days Mouse:4..5....- Influenza | 4 days Mouse 55-2. -* Inf. pneu. | 7 days , Mouse 6....... Influenza | 4 days x =|= ++ /+/=)<]-[+ + =|—[+/+}+}]=|=|4}4 * General series. 7 H II = Hebrew orphan strain—see table 5 a. M Il a = Marine strain—see table 5. tI Mouse 1-6 Recovered from mouse after noculation of sputum. MULTIPLICITY OF RACES OF B. INFLUENZAE vations 5 USED Laverdi Trahan jp —| — | —| — | — | — | — —|—|—]| — | — | — Williams Masates Godfrey Rampin Not tested —|-|-|- Not tested =|=—|=|— —| Not tested —|-—|-|-|-— Not tested =|—|—|— Not tested —|-|-|— Not tested oN Not tested —|=|-|-— Not tested —|-|-|-— Not tested —|-|-|- Not tested +/X|—|— Not tested —|-—|-|— Not tested —|-|-|-— Not tested —|-|-|- Not tested =|-|-|- Not tested +|+|=|X Not tested =| —|—|— Not tested —|-|-|-— Not tested =|/—|—|— Not tested —|-—!|-|- elt +/+] t+] =[4+]+]+|+[+]+ THE JOURNAL OF IMMUNOLOGY, VOL. lv, NO. 5 | 100 | 200 | 400 | 700 | 100 | 200 | 400 | 700 | 1000 I | x | | | | | | CONTROLS 367 368 EUGENIA VALENTINE AND GEORGIA M. COOPER Direct aggl u STRAINS AGGLUTIN. Group names and numbers Condition ONL shit BASRA ER ees: OS ae eR Normal Om tase asa et tees aoe eee Normal Oke Be eeerisas Horcicsn tao tact concmccarc Normal | |X/—/]J-—|—|-] |-—|-—J-|-I-| |-—|—/J-—j-|— OPE SER AER A A atae Leen Normal | j=|=|=|=|—| -SGeel es Olas i aakladette oe Slee atts cee aes oes Normal | |X/—|—/J-—|-| |-—/—|-—l—|-—|$]-—l-—|-i-i— ONG ccc5 Seach Oil de Teale eee Normal 5 3} TBR ENO DOD OnE OR COR Sen CODON Once Normal j OO ada ts RR Oe San eee Normal { - ONO is cs.kn.s eee ee eee ee Normal i | | Odi bei rhe Normal -: Oats Sea ree Normal 1 013 . - | O 14 : Opler Normal i | 0 16 iM (07 1 Paneer ae Hane carota mErnsatacr Normal it OWS Pe iinGnne sees sas seta eee Normal L| PA es cas ges n eee mae Lote tees tence Norma] ! PAD! ws © nesters Page Aten Sens fe Ae tS Normal a | 122 3 et RRR AN SACP ea Ton OerS ae ocr Normal 1220 Bee em OR anc cacti tts Ae acer Normal 1 Re SETAE Sor ae ection ipcteee Influenzat 1 A Neto SEE Oen ar aca cck isco riccetare Normal saa SIA Siecle che «cae see Oe eae en ee as a nie Normal * General series continued. 7 Convalescent, 4 weeks 369 MULTIPLICITY OF RACES OF B. INFLUENZAE nued 4 2) 8 & 8 x SC | mm | ee |e me —|-—)|/— |= —|—|-|x —|-|-|x Godfrey Xx —|+ —| P< SS | | El | = Masates x Williams esse = | eS Ee Se Pett tlt itll itil tit itll titi tlle] |t}t}t]tt]=|+]+]+]+|+|—]+4]+4]+]41x 370 EUGENIA VALENTINE AND GEORGIA M. COOPER of cases to the strains used to produce the antisera was strongly suggestive of multiplicity. A possible explanation would be that not one of the serum strains used was an actual primary TABLE 4* AGGLUTINATION AFTER ABSORPTION AGGLUTINATION BEFORE ABSORPTION ae Of absorbing | Of homologous SERUM ABSORBED BY strain strain fe Strains Paes ee be o;o; colo oo} oc; oo} & |=) R|S/=/5 S| =| 8/S/8|8] 5/8] S/= Hl i BU a Pp ea Pps) |) 1 (at ee ear 82 +i+/=IX|X/—-| S2 J—-j—-|-—|j-—|-— + )4)+] +} Mouse 5 |+/+|)X|—|—|—| Mouse 5 |—|—|—/—|— +)+]+]+|+ a Sykes +)+)+\/+/=|—-| Sykes = |—|—|—|-—|-—|-/-—|/-|-|- ohyiiges Ge ere Te tale) 2S Mita “j4l4l4l4l2lewaia | [=| El Siaaa 1 G1 GA eae ML 2 +/+)=/—|—|-| ML2 = |—|—|—|-—|— +/+/+]=|« ONG +/+\/4)/+|X|-| O7 J—|-—|-j-|- +HI+I+]=|— Laverdi |+/+/+/+/=/—| Laverdi |—|—|—|—|—|—|—|—|—|= Tacerdi Mouse 2 |=/=|=/=/+=/—|} Mouse2 |—|—|]—/—|—|—|—/-—|-—|— +S 6 C. trach. |-+/+/+/xX]—|—| C. trach. |—|—]—|]—/—/4+/4+/+]=|x C. lung |+/+|X/—|-—|—| C. lung |—|—|—|-|- ea BB Godfrey ae +/+|+=/+|-—| Godfrey |—|—|/—|—|—|-—|-—|-—|-|-— Goatrey.....{ Ae Ae a ee ar Ee Rube ps | Masates |+|+|+|+|+|—| Masates |—|]—|—|—|—/—|]-—|-—|-|- Masates.....4] Borham |+/+/+/+/+/—| Borham j|—|—|—|—|-— +)+/+/+}+ Sareska |+/+|/+|X|—|—| Sareska |—|—|—/—|— Be ee ere a. Tee Lee HI+|+i+/=l—| Lee 9 = |S/—)—|-|- | - Salee 1; Sareska j|+/+/=/x + ze + = ALAR des Influenza Early* - = — = = - 1c it Gee eree Influenza 3 days ~ - _ = = = EI GH ee is -.- Influenza Early* = _ — = = = 136 ites seer Influenza 5 days — - — = = =e 15 les fo eee Influenza Early* - — — = = = EE IQ) ee ss Influenza Early* — = — = = zee 1c 7 Rees Bee Influenza Early* — _ — = a = 18 Ged SO ae Influenza Karly* — - - — = = 18 Pie aM Influenza Karly* - — - = = LS H 231f......| Influenza Autopsy — — = = = Bs PAG ae er Influenza Autopsy _ — _ - _ _ H 25u.......| Influenza Autopsy - _ — = = = Hi 257:..:...| Influenza Autopsy — _ — = = ae H 271.......| Influenza Autopsy - _ — = = aes H 277.......| Influenza Autopsy - _ — = = = * Detailed data not available, early cases 2 to 4 days of disease. | L = Lung; T = Trachea. MULTIPLICITY OF RACES OF B. INFLUENZAE ano TABLE 6* AGGLUTINATION AFTER ABSORPTION AGGLUTINATION BEFORE ABSORPTION Of absorbing | Of homologous SERUM ABSORBED BY strain strain Strains $ sl/slgleisié s|£|8/8/8|s/s|s/3/8 M3a [+/+/4+/=|xX/-| M3a — |—|—/—/—|—|—|-—/-|-|- M3s- |+/+/4|/+/=/-| M3s_— |—|—|—/—/-|—/|-/-|-|- MPSA... See PE PS wise SSS sees M 738 =X i=|=|—|» Mis | el ll ele pe[=||== M13 |+/+/=/=| a = a A o oa) ca) 376 ‘sorties ATIUIVA y . —|<|=|+|+/+/+)-|-|-|-|-|-|=|-|-|- |i | 4 elle al le eel ¢ jezuenguy | 1BsoB)) —|—| ><} >< =e |e |e EEE ]-E]-e] te] | |e] EE |X| 2 |] Ede] | EEF] ] =| I] g jeauenguy jo" oueqowry —|-|-|-|-|- x¢|—|—|—|=|—l>< |< e141 1+ 1-1 =—|— 1] =| +-1-|-|—]< |= |= |—|—| |] |] ><] =|] F YE] |p fee ||] g fezuenguy jo PPRVYOLNL pa i a —|—|—|—| —|} =} —} — | — elit ot ezuonyuy |** °° ** *OZUaLO'TT shop slolalalelclalelalelelelelelalslslelelclsisizlelslelsisislelglelaisigisiz fom ke SIisisiS(SlSelSlelelslsieisis[ejseisisl(sisljelsisigicisisisisisic alesse = SH Oa con ici SSS SS = sie\sis\e = A KS asBasip | jo | uolipu0g Soule N ° poedg Is, IBsoBy ouByoBry) AI’ Bpasuy [PBA ozu910'T : SS SS SS ee dasa SNOILATIA GNV VUGS ONIGYNILATODY SNIVULS en ee eS eee See «lL HIAVL MULTIPLICITY OF RACES OF B. INFLUENZAE 377 strains in the throat of M. 7 is not necessarily an indication of primary and secondary infection by two strains of the same bacterium. The individual may have possessed. the two strains originally, or the second strain may have been implanted by contact. TABLE 8* SS TET AGGLUTINATION AFTER ABSORPTION AGGLUTINATION BEFORE ABSORPTION ABSORBED Of absorbing Of homologous SERUM BY strain strain STRAINS = 3 . ie! Strains old = o;o!lice!icio — 3 — 3s o;coice o;joio;ol;olc | | Al = =| O SISiR/ SRS) 3B) S/R|s ———— Michael |-+/+-|+/+|+||—| Michael |—|—|—|—|—|-—|-—|-—|—|-|-|— Michael...4| Gaetano|+|+|+|]+|X|—|—| Gaetano|—|—|—|—|—|— +/+|+]+|x|— Caesar |+|+/+|xX|—|—|—| Caesar |—|—|—|—|—|-— 4+)4|4]+1<|— Michael |+|+|X|—|—|—|—| Michael |—|—|—|—|—|— +} ]+-|+-]-+]= Angela |+|+/+/+|+|=|—| Angela |—|—|—|—|~[—|—|—|7|7|7|7 Angela.... Mees 2 No 0 aa | ua al Gaetano|+|+|+|+|X|—|—| Gaetano|—|—|—|—|—|— +\+|+]+|+|+ Michael |-+|-+|+|+|+||—| Michael |—|—|—|—|—|— +I+)/+)=|X|— Mar Angela |+|+|+|+|=|X|— Angelo |—|—|—|—|—|-— +)+}+}=/xX|— Ye--*-)) Mary ce ee ee edt DUES 209) tf foe | al be Ses Se Gaetano|+|+/+|+|+|x|—| Gaetano|—|—|—|—|—|— +)+)+)+/x|- | Michael |-+|-+|=||—|—|—| Michael |—|—|—|—|—|—|4|+|+|+|+|= Gaetano...,| Angela |+|+|+|X|—|—|—| Angela —|—|—|—|—|—|4+ | + }+]+|4+]= Gaetano|-+|+|+/+|+|+|—| Gaetano|—|—|—|—|—|-|—|-|7/-]7|— Michael |-+|+|+|+|—|—|—| Michael |—|—|—|—|—|—|+|+|+|1+|*|* Gaesae Angela +|+|x|x|—|—|—| Angela —|—|—|—|—|—| +} +/4+)4+)+|x ne Gretanol = |=(<|=|sl}— 1p Gaetanol—|— |) |— | siaiety eas @aesar (EEE |= @aeser | ala | |e el an * Family series. Direct agglutinations in table 7. An opportunity arose to study the strains isolated from the members of a single family all ill with influenza. The onset of the disease in the different members was very close together. Antisera were prepared with the strain isolated from each individual. The results of the direct agglutination and of the agglutinin absorption are given in tables 7 and 8. No identities 378 EUGENIA VALENTINE AND GEORGIA M. COOPER were encountered. It would seem that in such a series infected from a common source or one from the other, identities should be encountered, if B. influenzae were the primary infectious agent. TABLE 9 Summary of Investigation S| Sab |S g SERIES =) s & Ree an a FINDINGS Sea 240| 284 = aH 5 < & P O< AULODSY <) ws... /.%. 10 10 7 | All strains found to be distinct races. One autopsy strain identical with a miscellaneous strain. Another au-. topsy strain identical with another miscellaneous strain. Miscellaneous..... 73 0 18 | No two strains were found to be identi- cal. For identities see Autopsy Series above. IMBTINGS. 2.0.5 sjs<00 54 4 14 | Two strains from different individuals were found identical. Of isolations from the same individual. 4th day identical with 6th day Ath day identical with 8th day iden- tical with 11th day 2nd day identical with 8th day 4th day identical with 8th day iden- tical with 15th day 3rd day identical with 7th day, but 14th day was different : Hebrew orphans...| 28 5* | 16 | Two strains were found identical. Hamatllivne sepereicia 6 6 10 | All strains were found to be distinct Taces. * The results with one serum is given in table 5a, but the strains were tested with four other sera. No additional identities were discovered. For economy of tabulations and to avoid confusion, a con- siderable amount of agglutination results as well as absorptions are not recorded in the previous tables. Both the Hebrew Infant Asylum and the marine culture series were tested with nearly all the antisera. No identities were discovered in these MULTIPLICITY OF RACES OF B. INFLUENZAE 379 tests. Likewise, a marine antiserum other than the one given in tables 3 and 3a was utilized to test nearly all the strains. Here, again, no identities were encountered. Finally, all the strains still available were tested with anti- sera, Sykes, Laverdi, Benson, marine no. 4 and Hebrew Infant Asylum no. 11. These were the sera with which the identities were found as noted in the previous tables. Wherever cross- agglutinations were encountered, the results were again checked by agglutinin absorption. No further identities were found. A summary of the results given in the preceding tables as well as the additional agglutinations above described, is presented in table 9. DISCUSSION The results recorded indicate that, under the term B. influenzae, we are dealing with a group of organisms which, for practical purposes, is heterogeneous in character as determined by immu- nological reactions. We have evidences of the existence of small sub-groups. The actual frequency with which identical strains would be encountered could only be determined by the use of many more sera than we have employed. Our own results, however, indicate at least that identical strains would be infrequently encountered. The existence of a multiplicity of races is advanced as evidence that B. influenzae is not the primary etiological agent in epidemic influenza. To controvert this evidence, it would be necessary to prove either that the agglutinin absorption capacity of B. influenzae was a characteristic susceptible to rapid change or that, in no instance, have we utilized an actual primary infecting strain for the production of our serums. By analogy with the constancy of the agglutinin absorptive capacity shown by other varieties of bacteria, the former assumption is extremely improb- able. The latter assumption that we have failed to employ an actual epidemic strain is rendered very remote especially by the employment of strains isolated early in the disease from groups of cases in close contact. a 5 ie = at a ee ial” P< a = - 9 E STUDIES IN PROTEIN INTOXICATION IV. HISTOLOGIC LESIONS PRODUCED BY INJECTIONS OF PEPTON T. HARRIS BOUGHTON From the Depariment of Pathology of the University of Illinois, College of Medicine, Chicago Received for publication September 10, 1919 It has long been known that the parenteral injection of albu- moses or “pepton” into animals produces marked effects on blood pressure and on the coagulability of the blood. The earliest important work was reported by Schmidt-Muelheim in 1880 (1) and by Fano in 1881 (2). The similarity of this ‘‘ pep- ton poisoning” to anaphylactic shock was noted by de Waele in 1907 (3), by Biedl and Kraus in 1909 (4), and by Arthus (5). Since these original observations were recorded an enormous amount of study has been devoted to this topic with the purpose of analysing the mechanism of this toxie action, and possibly of throwing some light on the mechanism of anaphylaxis. Hardly any observations, however, have been recorded on the histologic changes resulting from pepton poisoning. Simonds (6) describes marked passive hyperemia occurring in the livers of dogs dying from acute pepton shock, and thrombi in the sinusoids of the livers after several injections of peptone. Nolf (7) mentions damage to the endothelium of hepatic capil- laries. A few writers have described the histologic changes occurring in anaphylaxis, but no attempt has been made to com- pare the effects on the organs and tissues produced by anaphy- laxis and by pepton poisoning. The purpose of the present investigation is to make such a comparison. The work on the physiological effects produced by pepton and by purified albu- moses when introduced parenterally into various animals has 381 * ‘ 382 T,. HARRIS BOUGHTON shown beyond question the very close parallelism between this and the intoxication of sensitized animals by native proteins. DeKruif and Eggerth (8), however, believe that the two phe- nomena are not identical since they state that the toxic prin- ciple of Witte’s pepton easily passes membranes that hold back all of the anaphylotoxin. The histologic method of study while not so direct as the physiologic or chemical may yet elicit information not otherwise obtainable. MATERIAL AND METHODS In the present study fourteen guinea-pigs were used, varying in weight from 200 to 400 grams. They were injected with Witte’s pepton intraperitoneally. The initial dose was usually 0.6 gram per kilo and this amount was increased to the limit of tolerance, which was usually found to be about 2 grams per kilo when the doses were given a few days apart; however, some of the animals resisted several doses of this size. The maximum number of doses received by any animal was eight and the average was five. Death usually occurred several hours after the last in- jection, but sometimes not until a few days later. The animals were under observation up to seven weeks, and the average length | of the observation time was three weeks. RESULTS The symptoms usually shown consisted of ruffing of the hair of the head and neck with respiratory distress (frequently hic- cough) and prostration. Most commonly if an animal does not succumb to the first injection, a later injection of the same size — is without evident effect, but a larger dose will bring on the — symptoms noted. On post mortem examination the lungs were found to be fully expanded, and hyperemia of the organs was marked. In a few cases subserous hemorrhages were observed, but these were not marked nor extensive. On microscopic examination the kidneys showed the mods marked change. ‘The epithelium was affected in every case. in most of the sections there was marked and extensive swelling STUDIES IN PROTEIN INTOXICATION 383 and vacuolation of the epithelium of the convoluted tubules and of the ascending limb of Henle’s loop. In some eases the process went on to complete disruption and necrosis of some of the epithelial cells, and was followed by regeneration. In a few cases vacuolation was not so prominent, and the process pro- ceeded from parenchymatous degeneration to necrosis. The epithelial changes were most marked in those cases in which death occurred soon after the last injection. In cases where several days elapsed before the animal’s death the changes noted were somewhat less marked. Hyperemia was present and marked in every case. Small hemorrhages were observed in about one-half of the cases. The vessels of the kidney were almost unaffected: in two cases the walls of the smaller arteries showed slight swelling. There were no areas of fibrosis or infil- tration observable. In the livers the epithelium was affected in all but one case, and this was an animal that had received but a single injection. Parenchymatous degeneration was observed in nearly all cases, and was frequently severe. Fatty degeneration of mild grade was also observed. Small areas of focal necrosis were seen in half of the series, and there was some evidence of epithelial regeneration. Leukocytic infiltration was frequently observed, especially about the vessels. The vessels of the liver were markedly altered. Swelling of the walls of the smaller arteries was present in nearly every case. This swelling involved both intima and media, frequently rendering the internal elastic lamina unduly prominent, and occasionally showing fissuring of the media. In most of the sections proliferation of the endo- thelial cells of the intima could be seen, as evidenced by the increased number of nuclei present. Fibrosis was not observed. In the hearts the only pathological change noted was round celled infiltration which, however, was present in practically all cases, and was marked in a few. The gross appearance of the organs post mortem, and the lesions found microscopically in this series show a striking simi- larity to those found in guinea-pigs with anaphylaxis (9). In general the lesions produced are somewhat less severe in the THE JOURNAL OF IMMUNOLOGY, VOL. IV, NO. 5 384 T. HARRIS BOUGHTON pepton animals. This may be due to the fact that the pepton animals were under observation for a much shorter period than the albumen animals, and received a smaller number of injections. The marked resemblance between the gross and microscopic changes produced by injections of pepton and by injections of native proteins (especially when considered in connection with the physiological observations recorded by many authors) fur- nishes another reason for considering these two phenomena, as closely related. CONCLUSIONS 1. Intraperitoneal injections of Witte’s pepton into guinea- pigs produce lesions of the liver, heart, and kidney. 2. These lesions consist of degeneration and necrosis of epithe- lium, followed by regeneration; of edema of the walls of the smaller arteries, with endothelial proliferation; of perivascular infiltration; and of hyperemia. 3. On post mortem examination the lungs are expanded and the organs hyperemic. 4. Both gross and microscopic lesions are very similar to those produced in guinea-pigs by injections of native proteins. REFERENCES (1) Scumipt-MvuetHerm: Archiv. f. Physiol., 1880, 30. (2) Fano: Archiv. f. Physiol., 1881, 277. (3) pE Waz.eE: Bul. de 1’Acad. roy. de med. de Belg., 1907, 21, 715. (4) Brept aND Kraus: Wien. klin. Wehnshrft, 1909, No. 11. (5) Arruus, C.: R. de l’Acad des Sci., 148. (6) Stmonps, J. P.: Journ. Inf. Dis., 1919, 24, 297. (7) Nour: Archiv. Internat. de Physiol., 1904, 1, 242. (8) Dz Kruir, P. H., ann Eccrrts, A. H.: Journ. Inf. Dis., 1919, 24, 505. (9) Boueuton, T. H.: Journ. Immun., 1916, 1, 105. Sh cat A OBSERVATIONS ON THE PRODUCTION OF AN ANTI- HAEMOTOXIN FOR THE HAEMOTOXIN OF BACTERI- UM WELCHII (BACILLUS AEROGENES CAPSULATUS) WILLIAM W. FORD anv GEORGE HUNTINGTON WILLIAMS From the Department of Bacteriology, School of Hygiene and Public Health, Johns Hopkins University Received for publication July 22, 1919 It has previously been shown by Ford and Lawrence (1) that the hemolytic property of milk cultures of Bacterium welchii is to be attributed to the production of a true bacterial haemolysin or haemotoxin by this organism. The opinion previously held that the destruction of blood corpuscles in cultures of this species is due to the presence of acids (butyric and lactic) is not justified, since milk cultures retain their blood-laking powers after com- plete neutralisation. In tests with the neutralised whey the discoloration of the corpuscles due to the production of met- haemoglobin from acids is not seen and the haemolysed blood takes on the brilliant red color usual with true bacterial haemo- toxins. Furthermore, cultures of the gas bacillus lose their haemolytic activity on being heated to 62-63°C., while solutions of both butyric and lactic acids ean be boiled some time with- out losing their power to destroy blood corpuscles. Finally, the haemolysin of B. welchiz is digested by pepsin-HCl and by pancreatin, and is precipitable with ethy! alcohol, properties which also tend to place this substance in the group of bacterial secretory products. During the course of the past few months the study of the haemotoxin of the gas bacillus has been con- tinued and our previous observations have been confirmed. In addition several new facts have been brought out which it seems desirable to report at the present time, 385 386 WwW. W. FORD AND G. H. WILLIAMS SOURCE OF THE CULTURES The strains of B. welchii which have been employed for the study of the haemolysins found in milk cultures have been obtained in all instances from samples of Baltimore milk. Quantities varying from 500 to 1000 cc. have been heated to temperatures ranging from 80° to 85°C. for twenty to th rty minutes and incubated at 37°C. for twenty-four to forty-eight hours. In the majority of such samples the characteristic reaction of stormy fermen- tation appears after the lapse of this time. Such milk cultures contain aerobic bacteria and a number of other anaerobes in addition to the gas bacillus. Various methods of obtaining pure culture have been tried out. In the early part of the work rabbits were inoculated intravenously with small quantities (1 to 2 ec.), killed in five minutes and kept in a warm place for eighteen to twenty hours after which milk and agar cultures were made from the blood and organs. Frequently other or- ganisms survived in the animal body and appeared in the cultures so that it was necessary to resort to the use of blood-agar plates from which the typical haemolytic colonies could be fished. As the work progressed it was found that this plating in agar could be dispensed with and pure cultures could be obtained from the animals by the simple method of ‘‘continuous transfer” of fairly large quantites of culture (1 to 2 cc.) from one milk tube to another. The rapidity with which B. welchii develops in milk at 37°C. enables it to overgrow nearly all other species and if transfers are made at the end of every twelve hours some half a dozen transfers suffice to give an uncontaminated strain. Cul- tures obtained with this apparently crude method have been tested repeatedly. They contain nothing but non-motile en- capsulated organisms—Gram-positive with an admixture of Gram-negative forms—which give characteristic reactions in culture media and produce the typical lesions on animal inocu- lation. Blood-agar plates made from such cultures reveal nothing but haemolytic colonies and aerobic cultures remain steril. We feel confident, therefore, that pure cultures can be obtained easily with this method. The same method of con- PRODUCTION OF AN ANTI-HAEMOTOXIN 387 tinuous transfer can be applied to the original milk flasks which have shown the stormy fermentation and in about two-thirds of the samples a pure culture of the gas bacillus can be obtained. In a certain number however the gas bacillus is overgrown by certain acid-resistant streptococci which may happen to survive the original heating of the milk. Animal passage, therefore, is usually advisable. It may be noted that the gas bacillus de- velops in milk tubes without incubation under anaerobic con- ditions provided cream is left in the milk to form a surface layer. While this cream layer may be more pervious to oxygen than was previously supposed and does not provide anaerobic conditions in the depths of the medium its presence in the milk has proved a distinctly favorable factor for the cultivation of the gas bacillus. Finally, the difficulty of keeping milk cultures alive may be lessened by the addition of a few particles of powdered calcium carbonate to the milk tubes. The acidity of the cultures is much lessened by this procedure and transfers can be obtained from such tubes twelve to fourteen days old. In general the gas bacillus dies out in milk in three to four days. PREPARATION OF THE HAEMOTOXIN The haemotoxin of Bacterium welchii can be obtained from massive milk cultures of the organism with the following method. Large flasks of steril milk containing 800 to 1000 cc. are inoculated by pouring into them the contents of a twenty-four to forty- eight hour milk culture (10 to 12 ec.). These flasks are incubated at 37°C. and by the end of twenty-four to forty-eight hours the characteristic reaction usually appears. This material is now filtered through coarse filter paper to remove the particles of curd and the filtrate as it appears is immediately neutralised by caustic soda or caustic potash. ‘The filtration is usually slow, requiring eighteen to twenty hours, during which time there is a further development of organisms in the filtrate with some ‘in- crease of acidity. The material is again completely neutralised, upon which a thick gelatinous precipitate appears. ‘This settles to the bottom of the flasks and is easily removed by passing the 388 W. W. FORD AND G. H. WILLIAMS fluid through filter paper. If this product is not entirely free from bacteria it can now be filtered through a Berkefeld candle. The final solution is clear, straw-colored, somewhat viscid and foams easily on shaking. About 400 cc. of this product can be obtained from a liter culture of the gas bacillus. STRENGTH OF THE HAEMOTOXIN Haemolytic solutions prepared by the method here described have a fairly uniform strength. When titrated against rabbits’ corpuscles 1 cc. of a 5 per cent suspension of blood cells is usually dissolved completely by 0.05 cc. representing a dilution of 1—20 TABLE 1 Determination of the strengih of haemotoxin from neutralised milk cultures of Bacterium welchit WHEY RABBIT BLOOD, 5 PER CENT HEMOLYSIS cc. ce. 1.0 1 Complete 0.75 1 Complete 0.5 1 Complete 0.25 1 Compl.te 0.1% 1 Complete 0.075 1 Complete 0.05 1 | Complete 0.025 1 Complete 0.01 1 Partial Control 1 ce. NaCl 0.75 per cent 1 Negative Strength, 0.025; dilution, 1-40. of the original material. The reaction takes place somewhat slowly, requiring at least four hours at 37°C. In some instances the haemolysin is stronger than this, the haemolytic unit being 0.025 cc., a dilution of 1-40, but this is somewhat uncommon. Very rarely a more powerful haemotoxin is found with a hae- molytic unit of 0.01 cc., a dilution of 1-100. STABILITY OF THE HAEMOTOXIN The haemotoxin of the gas bacillus is a relatively stabile sub- stance. When kept in the dark at a low temperature the strength of the solution remains fairly constant. Thus a prep- PRODUCTION OF AN ANTI-HAEMOTOXIN 389 aration was made about June 15 of the past year which had a strength of 0.025 ec. (1-40). This was used repeatedly for testing sera for about four weeks during which time it slowly deteriorated until its strength on July 10 was 0.05, a dilution of 1-20. This preparation was then put aside and not tested again till October 26 when it had a strength of 0.075, a dilution of about 1-13. A similar preparation with a strength of 0.2 ce. (1-5) on December 14, 1918, was tested repeatedly for about five months and on May 7, 1919, its strength was the same. Curve of union of haemotoxin ana entihaemotoxin 16 units § units 4 units 2 wits L mit -0008 ec. -004 ec. +008 cc. C% Cc. 002 oc. Fie. 1 PRODUCTION OF THE ANTIHAEMOTOXIN The attempt was now made to immunize rabbits with the haemotoxin of the gas bacillus produced by the method described. Six rabbits were given increasing doses of haemotoxin subcu- taneously, starting with small quantities representing low mul- tiples of the haemolytic unit. No definite effect upon the animal’s health could be noted. The weight remained stationary and there were no changes at the point of inoculation. As the doses became larger the health of the animals was not appreciably affected but a good deal of subcutaneous oedema and induration developed at the site of inoculation. This oedema and swelling were transient however and they usually disappeared in three or 390 W. W. FORD AND G. H. WILLIAMS four days. Of the animals originally selected one died of a laboratory infection and one developed subcutaneous abscesses necessitating its destruction. The other rabbits remained under treatment four months and showed the presence in the blood stream of antihaemolytic substances which could be demon- strated in high dilutions of the serum. STRENGTH OF THE ANTIHAEMOTOXIN The serum of rabbits treated with increasing doses of the haemotoxin of the gas bacillus contains an antihaemotoxin for this substance. If the serum be allowed to come in contact with the haemotoxin it neutralises it completely so that when blood corpuscles are subsequently added no solution takes place. The usual strength of the antihaemotoxin in animals under treatment was about 1—1000 when tested against freshly prepared hae- molysins with the use of the haemolytic unit above described as the index. This is shown in the following table. The immune sera were always tested in quantities of 0.1 cc. or less. Normal rabbits’ serum in larger amounts, 1 cc., 0.5 ee. and 0.125 cc., protects the corpuscles against several multiples of a haemolytic unit, but this reaction appears to depend upon the bathing of the corpuscles in a protective medium rather than upon a true neutralisation of the haemotoxin. The protection afforded by normal serum never appears beyond a dilution of 1-10. In the serum outlined in table 2 the corpuscles were com- pletely protected against the haemotoxin by a quantity of 0.0008 cc. which represents a dilution of the serum of about 1—1250. The usual strength of the anti-serum was somewhat less, about 1—1000. METHOD OF COMBINATION While the titration of the anti-serum against minimal doses of the haemotoxin offers a simple method of determining its strength this method is subject to certain possible errors. The length of time required for solution of the corpuscles by the haemotoxin and for estimating the strength of the antihaemotoxin is too long PRODUCTION OF AN ANTI-HAEMOTOXIN 391 and in addition the deterioration of the haemotoxin may be such as to make the quantity regarded as the unit incapable of caus- ing a complete solution of the cells. Later, therefore, the sera were always titrated against several multiples of a haemolytic unit, usually against multiples representing 2, 4, 8 and 16 un.ts. Such a series can be read at the end of four hours and in addition it gives some information as to the mode of union of TABLE 2 Strength of anti-haemotoxin Serum from rabbit I tested vs. an haemotoxin of unit strength 0.1 HEMOTOXIN SERUM PROTECTION cc. 0.1 0.1 +44 0.1 0.08 +4 0.1 0.06 +44 0.1 0.04 aol ok 0.1 0.02 ee 0.1 0.01 +++ 0.1 0.008 ++4+ 0.1 0.006 +++ 0.1 0.004 +++ 0.1 0.002 +44 0.1 0.001 ++ 0.1 0.0008 +++ 0.1 0.0006 ++ 0.1 0.0004 + 0.1 0.0002 0 0.1 0.0001 0 0.1 1 cc. NaCl 0.75 per cent 0 Complete protection is indicated by +++; partial by ++; slight by +; none by 0. the antihaemotoxin with the haemotoxin. With fresh prepa- rations the substances combine apparently according to the laws of multiple proportions. This is shown in table 3. It is interesting to note in this connection that the antihae- molysin produced in animals by the immunization with the haemolysin in fungi also combines with the haemolysin according to the laws of simple multiples, as was shown by Ford and Rockwood (2). 392 W. W. FORD AND G. H. WILLIAMS COMBINATION OF ANTISERUM WITH OLD HAEMOTOXIN As the haemotoxin deteriorates it loses its strength and its method of combination with the antihaemotoxin does not appear to be as simple as when fresh. Larger quantities of anti-serum are required to neutralise given quantities of haemotoxin and the line of union with multiples of a haemolytic unit is no longer a straight line but a rapidly ascending curve. The laws governing this union have not been worked out completely but it apparently follows the same rule as other toxins and antitoxins and may be explained by the theory that toxoids are formed as the haemotoxic solutions age. ‘This will be taken up in a subsequent publication. TABLE 3 Strength of antihaemotoxin from rabbit I determined by titrating the serum against multiples of a haemotoxic unit HEMATOXIN QUANTITY OF SERUM PROTECTING COMPLETELY UNITS WHEY cc. ce. 16 1.6 0.02 8 0.8 0.008 4 0.4 0.004 2 0.2 0.002 il 0.1 0.0008 If this series is put in the form of a curve with abscissas and ordinates it will be seen that the union is represented by practically a straight line (see chart 1) CONCLUSION By the immunisation of animals with the haemolytic or haemotoxic substance produced in milk by Bacterium welchir an anti-haemolysin or anti-haemotoxin of a strength of 1-1000 to 1-1250 has now been produced. The demonstration that this substance can act as an antigen offers the final proof that it belongs to the group of true bacterial haemolysins or true haemotoxins. REFERENCES (1) Forp, Witui1am W., anp Lawrence, JosepH H.: Johns Hopkins Hospital Bull,, 1917, 28, 245. (2) Forp, Witut1am W., anp Rockwoop, Erxet M.: Jour. of Pharm, & Exp. Ther., 1913, 4, 235. THE INFLUENCE OF DESICCATION UPON NATURAL HEMOLYSINS AND HEMAGGLUTININS IN HUMAN SERA JOHN A. KOLMER From the Dermatological Research Laboratories of Philadelphia and the McManes Laboratory of Experimental Pathology of the University of Pennsylvania Received for publication October 24, 1919 As part of a series of studies upon natural hemolysins and hemagelutinins in human sera (1), the influence of desiccation upon these substances has been investigated and the results have proven of particular interest in reference to the method of San- ford (2) of using normal human sera dried on cover glasses for the purpose of conducting microscopical agglutination test: in the classification or typing of blood; these studies have proven especially interesting and of practical importance since Sands and West (3) have found in these laboratories, that drying rab- bit antihuman serum in paper and dishes results in the destruc- tion or inactivation of a portion of the immune hemagglutinins and suggesting that the practice of using dried human sera for agglutination tests after the methods of Sanford (2) and Hart- man (4), may be open to error due to deterioration of the normal isoagglutinins. The influence of desiccation upon normal isohemolysins in human sera has also been the subject of investigation; in the studies previously referred to (1) sera were occasionally en- countered containing natural hemolysins but no agglutinins for the corresponding corpuscles and I am of the opinion that best results in the matching of bloods for transfusion are obtained by conducting both agglutination and hemolysins tests. It is pos- sible that these hemolysins may produce reactions following the transfusion of blood compatible for the recipient insofar as the agglutinins are concerned. 393 THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 6 394 JOHN A. KOLMER INFLUENCE OF DESICCATION UPON NATURAL HEMAGGLUTININS Fresh unheated human sera were examined for the presence of agglutinins by a macroscopic test consisting in mixing in small test tubes 0.1 ce. of each serum with 1 cc. of a 1 per cent sus- pension of washed corpuscles, incubating at 38°C. in a water bath for an hour and reading the results after the mixtures have stood in a refrigerator over night. Sera showing the presence of hemagglutinins were then dried by distributing 1 cc. of each evenly and carefully over accurately measured squares (usually 60 by 60 mm.) of paper (W. and R. Balston No. 2) and drying them at room temperature; each square was then ruled and cut into ten parts, each part carrying the equivalent of 0.1 ce of serum. Macroscopical agglutination tests were then repeated with each fluid and dried serum at the same time, with the same corpuscles and under identical conditions. In some experiments the dried sera were kept at room tempera- ture and in a refrigerator while the corresponding fluid sera were kept in a refrigerator over a period of two weeks for the purpose of studying the deterioration of the hemagglutinins under these conditions; since a variation in results could occur with differ- ent lots of corpuscles, the same corpuscles were kept in 1: 800 formalin in physiological saline (5) over the entire period of observation. The results of several experiments in which this macroscopic technic was employed are shown in tables 1, 2 and 3; it is to be emphasized that these observations were made macroscopically inasmuch as comparative microscopical and macroscopical tests have shown that the former may show some agglutination which cannot be detected by the naked eye in test tubes. By using 0.1 cc. serum with 1 cc. of 1 per cent suspensions of corpuscles the macroscopical tests were, however, usually clear and decisive and the presence or absence of agglutination was readily deter- mined by comparison with the corpuscle controls. The reactions shown in table 1 were conducted with the cor- puscles of six different persons and the dried sera were employed about three days after preparation; as shown in this table not INFLUENCE OF DESICCATION UPON NATURAL HEMOLYSINS 395 all of the isohemagglutinins in human sera are destroyed by drying, but some deterioration usually occurs. A few tests of this kind with typed bloods in which type II and type III sera dried in paper were employed, have shown that the resistance of these isohemagglutinins to drying are about the same and that the disappearance of hemagglutinin for one lot of corpuscles and the persistence of a second is due largely to quantitative relations, titration tests having shown that the latter is present TABLE 1 The influence of drying serum in paper upon human isoagglutinins RESULTS WITH SERA BEFORE DRYING RESULTS WITH SERA AFTER DRYING* Sees) i) ie. ha ee em eed (nee SECMM ee S| eee: Sols jek eeuieae |e" ig ee ele] el ele] Ble] | | or ON oof om owe oo om oN od om os oo 6) 6) 6) 6) 6) 6) 6) 6) 6) o) 6) e) 2 —t}| - == = ae + a = = = = — 3 Sc ed a RN De a eee eat i on ea 4 Se fe cma (ear wai (ig cn fc (Ue eee) (ea) te 8 2 Sell a da Ih oo eae 9 +); —-/] -|] - —{/ —/} —| —| =} —-|] -| - 10 —;} —| -/] - —{/ —}; —| —|] =} =] —-]| = 11 +} —-/} —| —-}| —}| —}| —-}| —-| -| -}] -|{ - 12 Ly esi lige amir PA PRR ie ll sss! [ear | eengeild ikea ga (gh 13 +}; —-/|/ —}| —-| —}| —| —-}| —-{| -|] -| -|{ - 15 LE Sa eee (mea Vc SPN aa Fe 19 —/}| —} —}| +) -—-| +) —-] -—-] -] -] - | + 20 Se | oe Aare, aio ead Silene * Tested seventy-two hours after drying; kept at room temperature. +t — = no agglutination; + = partial agglutination; + = agglutination. in the original serum in larger quantity and consequently present in sufficient amounts in the dried state to produce agglutination despite deterioration of a portion. Human sera generally contain relatively large amounts of agglutinin for rabbit corpuscles as may be determined by titra- tions; consequently drying and testing 0.1 ec. of serum may not show much depreciation in agglutinin as shown in table 2; a serum containing a small amount of agglutinin in the fresh state may, however, fail to agglutinate after drying. 396 JOHN A. KOLMER Human sera usually contain smaller amounts of agglutinin for rat corpuscles and consequently the influence of drying sera upon these agglutinins is quite apparent; as shown in table 3 considerable deterioration generally occurs. TABLE 2 The influence of drying human serum in paper upon natural antirabbit agglutinins RESULTS WITH SERA BEFORE DRYING RESULTS WiTH SERA AFTER DRYING* Beau Se Beg ee ee | ee) a so) = S = = = = S = =) =) =] e ep a e me 2 S et Ee e e ie o- ON ooo om ow oo on ON os8 ox“ os oo oO .@) oO iS) 1S) iS) ie) .@) 1) 1S) ie) o 2 SiMe inchveariosce tt orelltmriie oe = ee 3 Sona erie aoe ai lacaedt eee aime as i ST 6 Stil atau bee cE Ne ie =) | =a 9 Sianll Sia oN pee fee 8) = elf af — | = 10 -- - _ _ - - — -- -- -- — - 11 SPU karts Gall are ie Se i 5 WI ae = = | =) 12 See itor tees) Momo me Merely Grealcr seb Se 17 Se alee eles | acelear kar Se = * Tested ninety-six hours after drying; kept at room temperature. + + = agglutination; — = no agglutination. TABLE 3 The influence of drying human serum in paper upon natural antirat agglutinin RESULTS WITH SERA BEFORE DRYING RESULTS WITH SERA AFTER DRYING* S g g S g g g 8 3 & 3 5 2 5 S = = 3 r=} 5 = = = 3 ao a Q & [on a a a a a a i= Sm | 6N | Seo | Sw | SO | SO |] Sa | SN | So | SH | GO | GO 6) e) 6) e) e) 6) 6) 6) e) 6) 6) 6) 3 5 il (ite i all el lec ed eee 8 Se) AR Os SR el es 15 =e he ea se ee eee 16 figs she Sel SS a ee 20 che perce toe | oR) ee aie) ely alee rete * Tested seventy-two hours after drying; kept in a refrigerator. 7 — = no agglutination; + = agglutination. Tables 2 and 3 also show the presence of group hemagglutinins in human sera for rabbit and rat corpuscles similar to the group for human corpuscles, previously described (1). Microscopical tests were conducted by drying sera upon clean cover glasses (two loopfuls of serum placed in the center of a INFLUENCE OF DESICCATION UPON NATURAL HEMOLYSINS 397 glass) and allowing them to dry at room temperature; when used, a loopful of a 1 per cent suspension of washed corpuscles was added to the dried serum and thoroughly rubbed up with it and the mixture was suspended in a vaselined hollow ground slide followed by microscopical examination after standing about fifteen minutes at room temperature. The results of two experiments with anithuman and antirabbit agglutinins in human sera are shown in tables 4 and 5; numerous tests were made over a period of fourteen days to determine whether the agglutinins in dried sera tend to further deterioration. TABLE 4 The influence of drying human sera on cover glasses upon normal isoagglutinins RESULTS WITH RESULTS WITH DRIED SERA* SERA CORPUSCLES FRESH FLUID Sb 1 day 3 days 7 days 14 days Type II es Lype TIT ++t + =e = a Type Il Type III a ++ |} +4+ | +4 ] +4 Type II Type III = pa sae + “S - Type II Type Il =a ++ | ++ - ~ Type II Type I aaa + + + te Type III Type II aPSe ++ | ++ | ++ + Type III Type II + — = =e = Type III Type II aoe a _ - = Type lit: | Type Il Seats Sco i) ran + ~ @ype tlt’) | Type l ASE 55 + + - * Kept in a refrigerator. +++ = strong agglutination; + = weak agglutination; — — no aggluti- nation. As shown in table 4 if a serum contains a large amount of agglutinin capable of producing agglutination of practically all cells in the preparation (++), it may show a weaker reaction (+) after drying, some clumps of cells being seen but also large numbers of non-agglutinated corpuscles; a serum yielding a weak reaction in the wet state has been observed to fail to agglutinate after drying. Not infrequently however, sera yield- ing strong reactions in the fresh state yielded well marked agglu- tination over a period of two weeks, the dried sera on cover glasses being kept sealed in papers in a refrigerator as described by San- ford. As shown in table 4, deterioration of agglutinins was 398 JOHN A. KOLMER found within three or four days after drying; agglutinins escap- ing destruction in this time were usually preserved in a refrig- erator over the two weeks’ period of observation. Similar results were observed with the agglutinins for rabbit corpuscles in human sera (table 5); owing to the presence of these agglutinins in large amounts in human sera, the majority of dried sera showed no appreciable differences in agglutination from the corresponding fluid sera, inasmuch as sufficient agglu- tinin remained after drying to mask deterioration of a portion of the agglutinins. TABLE 5 The influence of drying human serum on cover glasses upon natural antirabbit agglutinin SERA AFTER DRYING* SERA BEFORE su DRYING 1 day 3 days 7 days 14 days 21 days 2 +44 a a a ++ ++ 3 + - - -- - - 4 ++ a + + + ~ 6 ++ at a ++ a+ a 7 ++ + - - - + 10 ++ | ++ nae nae ++ oe * Dried sera kept in a refrigerator. j++ = strong agglutination; + = weak agglutination. THE INFLUENCE OF DESICCATION UPON NATURAL HEMOLYSINS The natural hemolysins in human sera vary in resistance to the deterioration of desiccation; antisheep and antiox hemolysins and especially the former, are most resistant while antihuman, anti- guinea-pig and other hemolysins are guite susceptible. With the exception of the natural hemolysin for sheep corpuscles it would appear that the balance of hemolysins present in human sera for human corpuscles and the corpuscles of the lower ani- mals, are more susceptible to deterioration by drying than the corresponding agglutinins. Experiments with the hemolysins were conducted in the same manner as the macroscopic agglutination tests except that sera about three days old and practically free of complement were INFLUENCE OF DESICCATION UPON NATURAL HEMOLYSINS 399 employed in the fluid state and after being dried in paper, com- plement being furnished to the fluid and dried sera by adding 0.2 ce. of 1:10 dilutions of the mixed sera of guinea-pigs pre- viously absorbed at a low temperature with the corpuscles being used to remove the corresponding natural hemolysin if present. The results of an experiment with antihuman hemolysins are shown in table 6; the dried sera were employed twenty-four hours after preparation and the results of this and similar experiments show the marked deterioration of these isohemolysins following desiccation of the sera. TABLE 6 The influence of drying human serum in paper upon normal isohemolysins RESULTS WITH SERA BEFORE DRYING RESULTS WITH SERA AFTER DRYING* Srpoeen fice {oa toes er Wea leek ean eee Me Malia or ae oo ost a se Sr ga eee ar oe ne 1 N7| S N S C C N N N N N N 4 S S N N NS) N N N N N N N 9 NS) S N N N N) N N N N N N 12 N S N N Nie HP IME {Pe N N N N N 20 N N N N Ss N N N N N N N * Texted three days after drying; kept at room temperature. 7 N = no hemolysis; S = slight hemolysis; M = marked hemolysis; C = com- plete hemolysis. The results of a similar experiment with sheep cells are shown in table 7; marked deterioration of the natural antisheep hemo- lysins also occurred but owing to the presence of relatively large amounts of antisheep hemolysin in human sera in addition to a greater resistance of these hemolysins to deterioration (4), many of the dried sera were hemolytic and remained so over a period of at least three weeks. Additional experiments were conducted by exposing washed sheep corpuscles to equal amounts of the same serum before and after drying at a low temperature followed by removal of the cells and two washings with saline solution and titration of each suspension for the degree of sensitization by distributing the sus- 400 JOHN A. KOLMER pensions in doses varying from 0.1 to 1 cc. in test tubes and fur- nishing complement by adding 1 cc. of a 1: 20 dilution of hemo- lysin free guinea-pig serum. After one hour in a water bath at 38°C. the “hemolytic. index” of each serum was read off, that is, TABLE 7 The influence of drying human serum in paper upon natural antisheep hemolysin RESULTS WITH SERA BEFORE DRYING RESULTS WITH SERA AFTER DRYING* 3 3 3 3 3 3 3 3 3 3 3 3 pe B Pa Ee 2 zy Sy ee re ey ee on ON ooo ow ow oo on ON ose ot on oo Oo 1) iS) eo) ie) 1S) ie) 1é) is) 1o) ie) 16) 2 CT Ae Cc C Cc C NIN | N | N ipa 4 M; M| C Mj| C Met ON IN} NO IN| 1) SING 5 C C C C C C NighoNcts3S N |. Nai 7 M; M| C NS) Cc Ss N | N- | N.| No) Nee 9 C Cc Cc C C C Nye} oS N } Nay 10 C Cc Cc C Cc C 1 a Ves Ch We MA a BP M 12 C Cc C Cc Cc C No aN: os N | Moi as 13 C C C C Cc Cc N | ON | ON) N [NG 17 Cc C C C C C Nas Ss Ss Nae 18 C C C Cc C C M; M/| C Ss Cc; M * Tested two days after drying; kept at room temperature. + C = complete hemolysis; M = marked hemolysis; S = slight hemolysis; N = no hemolysis. TABLE 8 The influence of drying human serum in paper upon natural antisheep hemolysin HEMOLYTIC INDICES OF SHEEP CELLS SERA Sensitized in serum before drying} Sensitized in serum after drying cc. cc. a 0.1 0.5 2 0.1 0.5 3 0.05 None 4. On None the largest amount of corpuscle suspension showing complete hemolysis, and the method permitted an accurate measure of the degree of deterioration of the hemolysins consequent to drying as measured by the degree of sensitization of the same amounts of corpuscles in the fluid and dried portions of each serum. INFLUENCE OF DESICCATION UPON NATURAL HEMOLYSINS 401 The results of experiments of this kind with natural antisheep hemolysins are shown in tables 8 and 9 and with antiguinea-pig hemolysin in table 10. Invariably deterioration of these hemo- lysins took place as a result of drying which sometimes progressed to a slight extent over a period of three weeks. TABLE 9 The influence of drying human serum in paper upon natural antisheep hemolysin HEMOLYTIC INDICES SERA Before drying Dried 1 day Dried 2 days cc cc. 1 0.6 0.3 2 0.2 None 3 0.5 0.1 4 0.4 0.1 5 0.3 None 6 0.4 None i 0.3 None TABLE 10 The influence of drying human serum in paper upon natural antiguinea-pig hemolysin HEMOLYTIC INDICES SERA | Before drying Dried 1 day Dried 2 days cc. cc. 1 0.2 0.1 None 2 : 0.1 None None 3 0.1 0.5 None CONCLUSIONS 1. Drying normal human sera upon cover glasses and in paper at ordinary room temperatures frequently results in marked or complete deterioration of the normal isohemagglutinins. 2. Deterioration of these normal isohemagglutinins is espe- cially evident within the first to fourth days after the sera have been dried. 3. Similar results were observed with hemagglutinins in nor- mal human sera for the corpuscles of the lower animals. 402 JOHN A. KOLMER 4, Human sera containing large amounts of normal hemag- glutinins when dried under ordinary conditions and properly kept in a refrigerator may prove satisfactory for microscopical tests for at least two weeks, due to the presence of sufficient agglutinins escaping destruction. Only such sera should be used for drying and tests should be made at the end of the first week to determine if agglutinins are present before the cover glasses are used for the typing of bloods. 5. The hemolysins found in normal human sera for the cor- puscles of persons and the lower animals also deteriorate upon desiccation under ordinary conditions and are somewhat more susceptible than the hemagglutinins. 6. For the grouping of blood, sera should be kept in a fluid state sealed in ampoules at a low temperature, both hemagglutinins and hemolysins in normal human sera being highly susceptible to heat. REFERENCES (1) Kotmer, J. A., Trist, M. E., anp Fuicx, A. M. A study of the natural ; thermolabile and thermostabile hemolysins and hemagglutinins in human serum in relation to the Wassermann reaction. Amer. Jour. Syphilis (in press). (2) Sanrorp, A. H.: A modification of the Moss method of determining iso- hemagglutination groups. Jour. Amer. Med. Assoc., 1918, 70, 122. (3) Sanps, J., anp West, L.: Experiments upon the removal of hemagglutinins from rabbit antihuman sera. Jour. Immunology (in press). (4) Harrman, F. W.: New methods for blood transfusion and serum therapy. Jour. Amer. Med. Assoc., 1918, 71, 1658-1659. (5) Koumer, J. A., anp Brown, G. P.: The red corpuscle suspension for the Wassermann reaction. The preservation of red blood cells. Amer. Jour. Syphilis, 1919, 3, no. 2. THE NATURE OF THERMOLABILE HEMOLYSINS! JOHN A. KOLMER From the Dermatological Research Laboratories of Philadelphia Received for publication, November 14, 1919 Studies bearing upon the presence of natural hemolysins in human serum for the erythrocytes of the lower animals have been largely conducted with heated sera and principally for antisheep hemolysin, owing to the widespread use of the sheep hemolytic system in complement fixation tests; for this reason the natural hemolysins are commonly regarded as thermostabile or heat resistant, as is generally true of the immune hemolysins. Thiele and Embleton (1) have shown that a portion of the natural antisheep hemolysins present in active human sera and of the immune hemolysins in rabbit sera are thermolabile or heat sensitive, being destroyed or inactivated by heating serum at 56°C. for thirty minutes. These investigators have sought to prove that natural hemolysins and the first production of immune hemolysins are thermolabile and of the nature of differ- entiated complements, which may be absorbed by the corre- sponding corpuscles at a low temperature; later in the course of immunization, the immune hemolysins become more and more differentiated from complement becoming thermostabile in nature and removable by absorption with corpuscles without absorption of complement. Sherman (2), however, claims that all hemolysins, both nat- ural and immune, are thermostabile and that the reduction in hemolytic activity of a serum as a result of heating is due to “masking” of the hemolysin rather than an actual destruction 1 Presented before the meeting of American Association of Immunologists, June 16, 1919. This investigation has been conducted through the use of funds accruing from the distribution of arsphenamine. 403 404 JOHN A. KOLMER or inactivation. Sherman designates as “obvious” hemolysis all that are active in heated serum and as ‘“‘masked”’ hemoly- sins, those proving active in unheated but not in heated serum, but the presence of which may be shown by absorption tests, the corpuscles becoming sensitized in heated serum to the same de- gree as in unheated or active serum. In other words, according to this view, certain hemolysins may become ‘‘masked” as a result of heating the sera and incapable of sensitizing erythro- cytes in the presence of complement and producing hemolysins during the usual period of incubation of an hour at 38°C. but are TABLE 1 Summary showing the percentage of human sera in amount of 0.1 cc. containing hemolysins for the erythrocytes of various animals HEMOLYSINS ERYTHROCYTES nn Thermolabile Thermostabile per cent per cent AETURTVNRTN EY Seta gc, 3 cite eCate- eae oeTe eereere 4to 44 Oto 4 CEM nie Ce rr tase k RR Oe Cne ee ete 85 to 95 80 to 95 (Oe Bent sas Meee eee ete Reye eaht aes Le A Tiee ah ee 72 to 88 5 to 30 Guimesapigs: 2 F405 ti <5 5 aisle ee eyes See eee WORLUG 20 to 25 TRIES 0] ohh a, ie hee ae ER PR i ale le ROLE a ree 94 to 100 Oto 2 TRIG rg ek oN, ee Ad RAM a Nes ea me A AIMS Di] nia 4to 32 OLtorZ PO Fe oreo a es es aie Ses ocr PaO Ae one SRE e ee | OOO EEUU 0 to 15 DO PEPE oA ee tere te Ae kee Teas PD Un eta oes About 98 About 4 CWhickenttae ys byt sts. fee) BUN Ee eee ee About 62 0 not actually destroyed being capable of sensitizing corpuscles in mixtures of heated serum and cells, the latter undergoing hemoly- sis when removed from the serum and mixed with complement. In a study of the natural hemolysins and hemagglutinins in active and heated human sera in relation to the Wassermann test (3) we have found that a large percentage of active or un- heated human sera contain a wide variety of hemolysins for the corpuscles of the lower animals and after heating the sera at 56°C. for thirty minutes the majority of these hemolysins are destroyed or rendered inactive; indeed, the natural hemolysis in human sera for human corpuscles and for those of the guinea- pig, rabbit, rat, dog, hog and chicken are very largely of this NATURE OF THERMOLABILE HEMOLYSINS 405 thermolabile variety as shown in a summary given in table I. This study also showed that the hemolysins and hemagglutinins in human sera for the erythrocytes of the lower animals occur in groups similar to the four groups of isohemagglutinins and that figures based upon work conducted with the corpuscles of any one animal of a species are only approximately correct. The percentages shown in table 1 were secured by testing each serum with corpuscles from at least six different animals of each species and showed the marked variation in the hemolysin content of sera for the corpuscles of any given species. PURPOSES OF INVESTIGATION The purposes of this investigation were mainly twofold, namely, to study the relation of the natural hemolysins in human serum to human complement and of antisheep immune hemolysins in rabbit serum to rabbit complement, to determine whether these hemolysins are differentiated complements according to the theory of Thiele and Embleton; secondly, to study the fate of the natural hemolysins in human serum as a result of heating. There can be no doubt that heating human serum at 56°C. for thirty minutes either destroys outright or inactivates certain hemolysins, notably that for guinea-pig cells; it is commonly believed that they are actually destroyed and I have sought to determine whether this occurs or whether the hemolysins become masked as described by Sherman, without undergoing complete destruction. PART ONE Thermolabile and thermostabile hemolysins In differentiating between the natural thermolabile and ther- mostabile hemolysins in human sera shown in table 1, all sera were tested fresh and active and again after being heated in a water bath at 56°C. for thirty minutes; with active sera the com- plement of each serum was utilized while with heated sera com- plement was furnished by the mixed sera of guinea-pig previously absorbed at a low temperature with the corpuscles in order to remove any traces of hemolysin and hemagglutinin. 406 JOHN A. KOLMER This degree of heat is purely arbitrary; deterioration of natural hemolysins is not marked until a serum has been exposed to a temperature of 52° to 56°C. for fifteen to thirty minutes (tables 2 and 3), but at 62°C. deterioration occurs quite rapidly and is generally completed after thirty to forty-five minutes (table 4). In my experience natural antisheep hemolysin is more resistant to heat than any of those studied (tables 3 and 4) and anti- TABLE 2 The effect of heating human sera at 56°C. upon natural antiguinea-pig hemolysin ACTIVE 5 MINUTES 15 MINUTES 20 MINUTES 30 MINUTES SERA 1 N# Co) Co Ne SMC Ne GS sei ON cS" 1 Nanas 2 NC. (Cu NM ING INS SEN IN| ING) INGot Inia ea 3 N | C | Cop Nel MC.) NN oS Nal ON HIN Nine t N/S |}/C UN | NS | NN NENG NO) ONO ANGTSNaen 5 NS | CUNT NON ENGIN TENN SN Ne aN ai *C = complete hemolysis; M = marked hemolysis; S = slight hemolysis; N = No hemolysis. TABLE 3 The effect of heating human sera at 56°C. upon natural antisheep hemolysin ACTIVE 5 MINUTES 15 MINUTES 30 MINUTES 60 MINUTES SERA Se ee eed 0.01} 0.1 | 0.2 | 0.01) 0.1 | 0.2 | 0.01) 0.1 | 0.2 | 0.01} 0.1 | 0.2 | 0.01) 0.1 | 0.2 1 STC Te (N | Sores N iS | Cal NES so Naiomee 2 Cre Ee. | C | CVG.) Ce C.-C Wt SAC 44Ca tS 2 tea ae 3 S16€)/C 18S 1|C1e]S |M-C | SM i Mailtss ieee 4 M/C |}C |S |C|/C/S (Mi © |S |S.) @)) Ssissaieme *C = complete hemolysis; M = marked hemolysis; S = slight hemolysis; N = no hemolysis. guinea-pig hemolysin one of the most susceptible (table 2), but all natural hemolysins are much more susceptible to heat than immune hemolysins and in this regard bear a resemblance to the thermolabile nature of complement. If a temperature of 62°C. and an exposure of one hour were adopted for the differentiation of thermolabile and thermostabile hemolysins, all natural hemoly- sins in human sera would be classed as thermolabile, but 56°C. NATURE OF THERMOLABILE HEMOLYSINS 407 for thirty minutes has been generally adopted and with this exposure all natural hemolysins may be divided into the two classes. The experiments shown in tables 2, 3 and 4 were conducted with fresh sera; heating was conducted in a water bath. Washed guinea-pig and sheep corpuscles were employed in dose of 1 ce. of 1 per cent suspension with the graded amounts of each serum. Complement was furnished by the mixed hemolysin-free sera of guinea-pigs in dose of 0.2 ce. of 1:10 dilution; the results were read after water bath incubation for one hour. TABLE 4 The effect of heating human sera at 62°C. upon natural antisheep hemolysin ACTIVE 5 MINUTES 15 MINUTES 30 MINUTES 45 MINUTES SERA 1 See iec. | MiC | CoeNetS” | MLN iS lias chen | Nets 2 CALC) es) |) M |) CoN Se oS IN: PANWIRSESTEN: | NG Ten 3 SaeCr Peaks tM Cor NerS | M | NaN sN EN EN’ ON 4 MCE as |S MUN | NS TaN TEN aS: IN| es aN *C = complete hemolysis; M = marked hemolysis; S = slight hemolysis; N = no hemolysis. The relative susceptibility to heat of natural hemolysins and complements in human sera Bearing directly upon the question of the relation of natural hemolysins to complement raised by the work of Thiele and Embleton, is the relative susceptibility of both to heat. As pre- viously stated most of the natural hemolysins are quite suscep- tible and the thermolability of complement is well known; for these reasons experiments designed to bring out differences in resistance if they exist, required frequent repetition and the closest attention to technical details. Inasmuch as natural antiguinea-pig hemolysin appears to be very thermolabile, most of my experiments were conducted with pig cells and attempts made to study the comparative resistance of this particular hemolysin and the complement of human sera. 408 JOHN A. KOLMER These experiments were conducted by two methods; in the first method fresh human sera were titrated in varying amounts with 1 ce. of a 1 per cent suspension of washed pig cells to determine the smallest amount of serum proving completely hemolytic and also titrated with antihuman hemolysin and 0.1 ce. of a 5 per cent suspension of washed human corpuscles for the unit of com- plement. The sera were then heated at 56°C. and the titra- tions were repeated at intervals to determine the relative rates of deterioration of the hemolysin and complement. The results observed with one serum are shown in table 5 and they indicate that the resistance of natural antiguinea-pig hemolysin to heat is but slightly greater than the complement of human serum. TABLE 5 The influence of heat upon the complement and natural antiguinea-pig hemolysin of human serum HEMOLYTIC ACTIVITY SA sig GS COMPLEMENT ACTIVITY{ SERUM 0.02 | 0.05} 0.1 | 0.2 | 0.02} 0.04} 0.06} 0.08) 0.1 | 0.2 Active—unheated...........0:2+<...2| Mil_C.) Ce | NS aes Cc Heateditfor 5S aminutes. . 22-25-82. Ni N 1S Go) NEN. EN OES) EMae Heated for 10 minutes............ NININYUS EN UN UN Ne Nee Heated for 15 minutes............ NIN | NIN | NIN IN} NOUNS * Titrated with 1 cc. of a 1 per cent suspension of washed pig cells. + Titrated with 0.1 ce. of 5 per cent human cells and antihuman hemolysin. tC = complete hemolysis; M = marked hemolysis; S = slight hemolysis; N = no hemolysis. Similar experiments with natural antisheep hemolysin have, however, usually shown a much greater resistance of the hemolysin as compared with the complement of each serum. In the second method the hemolytic activity of fresh sera for guinea-pig cells was determined in titrations employing varying amounts of serum with 1 cc. of 1 per cent suspensions of washed pig corpuscles; the sera were then placed in water at 56°C. and titrated at varying intervals being reactivated by the addition of 0.2 ce. of 1: 10 dilutions of guinea-pig serum complement free of isohemolysin for the particular cells being employed. Numerous experiments have shown that human serum complement is very NATURE OF THERMOLABILE HEMOLYSINS 409 sensitive to heat and that an exposure of five to ten minutes at a temperature of 56°C. results in marked or complete deterioration. The results of one experiment are shown in table 6; as a general rule an exposure of five minutes completely removed the hemoly- tic activity of fresh sera for guinea-pig cells, but that this was due to the destruction of the complement rather than of the hemolysin is indicated by the fact that with reactivation of the serum hemolytic activity was not completely lost until the expo- sure was prolonged to twenty minutes. TABLE 6 The influence of heat upon thermolabile antiguinea-pig hemolysin TITRATIONS WITH 1 cc. OF 1 PER SERA CENT OF PIG CELLS NET e bine ISGRUNS a1. 3 ice pn he Ane delae es hao ec IN |e Meee? C1 mena heated. fOr 5 MINULES.:..) ce lee ede INP NG ENGIN Nea aN Serum heated 5 minutes + complement............|N|N|N|S |M]|C Serum heated 10 minutes + complement............ NUGNE Nes; ies i) Mi Serum heated 15 minutes + complement.............N|N|N{[N {S/S Serum heated 20 minutes + complement............ NENG ONIN, WEN is Serum heated 30 minutes + complement.............N|N|N|N/]|NJIN *N =no hemolysis; S = slight hemolysis; M = marked hemolysis; C = complete hemolysis. { Complement furnished as 0.2 cc. of 1: 10 pig serum. Experiments conducted after these methods indicate therefore, that the natural hemolysins are more resistant to heat than the complements in’ fresh human sera. The relative susceptibility to"age of natural hemolysins and comple- ments.in human sera Experiments conducted?in exactly the same manner as de- seribed above have shown that the natural antiguinea-pig and antisheep hemolysins in human sera are more resistant to the deterioration of age than the complements of the same sera. In conducting these experiments sterile sera only were employed THE JOURNAL OF IMMUNOLOGY, VOL. IV, NO. 6 410 JOHN A. KOLMER in order to prevent the development of anticomplementary sub- stances and the titrations were made at intervals of one, three, five, seven and ten days with sera kept at ordinary room tem- perature and at 7° to 9°C. in a refrigerator. While the com- plements had usually completely disappeared within three days in sera kept at room temperature and in five to seven days in sera kept in the refrigerator, the natural hemolysins had under- gone only partial deterioration, the sera being reactivated with hemolysin free guinea-pig serum complement as described above. TABLE7 The relation of the complement activity of human serum to its hemolytic activity for guinea-pig erythrocytes SERUM TITRATION OF COMPLEMENT* TITRATION OF HEMOLYTIC ACTIVITY ON Oa a a eee er eee eee BER /0..01/0..02/0..03/0..04/0..05|0..06 ad ea 0.1 0.01/0..02|9..03)0.04 0..05/0..0610 .07|0. 08/0..09 0.1 1 INT NIMIC |}CIC)/ClLC|/CiCININIS | MIC] Ci} Cy Clee 2N | NINIS | CGE Pe1 Ce | NasSeMsy Cri iC ere 3 -IN |NIS |MIM/ C/G]. €]Cl-C| NS Ml CCC Ce Cike 4 IN|IN(|S|M/C)C|C|C|C/CININ|N|S | C7 CC} Cl eere SIN NI NIN) SC) C PE1C UC {NIN SC) Cire) | Carian 6 IN |NIS | C/] C1) CC] € Cre NIN S fe | C1CVene Tete * With 0.1 ce. of 5 per cent human cells and two units of antihuman hemolysin. + N = no hemolysis; S = slight hemolysis; M = marked hemolysis; C = complete hemolysis. The relation of the complement activity of human serum to natural hemolysins For the purpose of further study of the possible relationship of the natural hemolysins to the complements, fresh human sera were titrated for complement activity in an antihuman hemolytic system and for natural antiguinea-pig hemolysin by using vary- ing amounts of serum with 1 cc. of 1 per cent suspension of guinea-pig cells. The purpose of these experiments was to de- termine whether sera particularly rich in complement for sensi- tized human corpuscles were likewise markedly hemolytic for guinea-pig ceils or whether sera poor i complement were likewise poor in hemolytic activity for the guinea-pig cells. NATURE OF THERMOLABILE HEMOLYSINS Ait The results of experiments with six sera shown in table 7 indicate that the two properties are not absolutely parallel. Sera 3 and 5, for example, were slightly below the average in complement activity while their hemolytic activity for guinea-pig cells was quite marked; serum 1, on the contrary, was quite rich in complement and below the average in hemolysin content. The differences, however, were never marked and were brought out only by strict attention to technical details but I believe the results warrant the conclusion on the basis of the technic employed, that complements and natural hemolysins in human sera are not necessarily present in parallel amounts. The influence upon the complement sera of the removal by absorption with erythrocytes Thiele and Embleton have based part of their claim for the existence of thermolabile hemolysins and their relation to com- plement upon the observation that the addition of complement increases the hemolytic power of an active serum for sheep ery- throcytes. This is undoubtedly true but the deductions are not necessarily so; as shown in tables 8 and 9 the addition of hemolysin-free guinea-pig complement to active human and rabbit immune sera increased the hemolytic activity of these sera, but inasmuch as the guinea-pig complements were pre- viously absorbed in the cold with sheep corpuscles and proven free of hemolysin for these cells in doses as large as 0.5 ce., I ascribe the results to the mechanism of the well known phenome- non that an excess of complement will, to a certain extent, in- crease hemolysis in any hemolytic system although, in view of our ignorance regarding the true nature of serum complement and the mechanism of its activity, the explanation of why this occurs is only a conjecture. Another of the arguments given by Thiele and Embleton as support for their theory that hemolysins are differentiated com- plements and that the immune hemolysins are products of evo- lution from complement, is that absorption of sera at a low temperature with erythrocytes results not only in the removal 412 JOHN A. KOLMER of hemolysin but of complement as well and that the washed sensitized corpuscles will undergo some hemolysis in the test tube due to complement and hemolysin they have absorbed. In their experiments it was found that the absorption of normal sera at a low temperature with sheep erythrocytes removed at TABLE 8 The influence of complement upon the hemolytic activity of human serum FOR 5 PER CENT WASHED SHEEP CELLS FOR 5 PER CENT WASHED PIG CELLS SERUM 0.1 | 0.2/0.3 0.4/0.5 0.6 | 0.7| 0.8|0.9| 1.0} 0.1] 0.2|0.3| 0.4} 0.5] 0.6| 0.7/0.8! 0.9| 1.0 Active serum 0.1 COS Ree: CTICICIM/IS;|S/IS/IS;/|S;/IS/|CIC|IMIM|S|S/S|IS{S1]{S Active serum 0.1 ec. + comple- MEGA ses ¢ C{|CICIC;C;CIiCl/C|C;C;C|;|C;C!1C|C|iMIM;IM|IM|M Heated serum 0.1 cc. + comple- MeTIGAS eee. 3: C HCH auc7e MIS{|S/S{|S|N|IN|IN|IN|IN|IN|ININININ *0.5 ec. of 1:10 dilution pig serum (hemolysin free). + C = complete hemolysis; M = marked hemolysis; S = slight hemolysis. TABLE 9 The influence of complement upon the hemolytic activity of rabbit antisheep serum 10 PER CENT SUSPENSION SHEEP CELLS SERUM Unheated serum 0.1 cc........... Cae WC NE | My eM Msgs s|S Unheated serum 0.1 cc. + com- plement*). 6 s¢sdee3. 0... 0:es0) Ca: | C. | Caves e: Cre a eee Heated serum 0.1 cc. + comple- THON Glee eee ee eect eiae an ve eles CijIMIM|IM;|IM;{S/S{S/S1{S *0.5 cc. of 1: 10 hemolysin free pig serum. the same time all or a part of the complement, indicating that “‘the hemolytic substance appears to be complement so modified that it can combine directly with the red cells in the cold and directly cause hemolysis.’”? With the sera of rabbits immunized with human corpuscles, these investigators found that absorption in the cold removed the immune or thermostabile hemolysin NATURE OF THERMOLABILE HEMOLYSINS 413 and less and less complement as immunization proceeded or, in other words, that the first production of immune hemolysin in rabbits is thermolabile and allied to the complement, absorption removing not only the hemolysin but a part of all of the com- plement as well, whereas later, when the immune hemolysin became thermostabile, absorption resulted in the removal of the hemolysin but little or none of complement. In my experiments fresh human and rabbit immune sera were chilled, treated with chilled washed erythrocytes and the mixtures were maintained just above the freezing point for two TABLE 10 The influence on the complement of human serum of the removal of natural antisheep hemolysins COMPLEMENT ACTIVITY* HEMOLYTIC ACTIVITY FOR SHEEP CELLS SERUM 0..01/0..02/0..03/0.04/0. 05}0..06)0..07|0..08)0..09] 0.1 |0 .01)0..02/0..03/0..04/0..05)0..06|0..07|0..08/0..09] 0.1 No. 1. Untreated EGU S, sraevaro-e « NS NM MLC | Cre re (Gre | Ni NPNSe Cre | Cle rere No. 1. Treated RETIN oss ios)s cia NINIS |M|IM;/C/IC\|CICiICIN|INININININININININ No. 2. Untreated BETUIN s . =:.- se tee Nii NiIS|S/|M{|M/C;}C]Cy{C ihalberedsserumee. .5..:!.o.- 20 cee oe NINN] Ne] NOON? NY Nae Filtered serum + complement*...|S |M}]C|C|C}C|C|C}]Cj{C *9.1 cc. of 1: 20 guinea-pig serum. +N =no hemolysis; S = slight hemolysis; M = marked hemolysis; C = complete hemolysis. The results of an experiment in which fresh human serum and guinea-pig corpuscles were employed is shown in table 17; this serum proved completely lytic for 1 cc. of a 1 per cent suspen- sion of guinea-pig corpuscles in dose of 0.07 cc. After filtration its hemolytic activity was lost due to the removal of the com- plement by the filter; the addition of a small dose of guinea-pig complement previously absorbed with the same corpuscles to remove any isohemolysin which may have been present, restored hemolytic activity beyond that of the unfiltered serum. NATURE OF THERMOLABILE HEMOLYSINS 419 PART TWO The absorption of natural hemolysins from active and heated human sera Numerous experiments have been conducted to determine whether certain natural hemolysins in fresh human sera are actually destroyed by heating or simple inactivated or masked as described by Sherman; these experiments were conducted as follows with each serum after preliminary tests had shown the presence of hemolysin: 1. With each fresh unheated serum 0.1 cc. was placed in a series of ten test tubes; increasing amounts of 5 per cent suspen- sion of washed corpuscles varying from 0.1 to 1 ec. were added and the volume of each tube was equalized with salt solution. After water bath incubation for one hour the results were read and the largest dose of corpuscles completely hemolysed was desig- nated as the hemolytic index of the serum, hemolysis being due to the activity of natural hemolysin and native complement. These tests were then repeated with sera heated at 56° and 62°C. for thirty minutes in exactly the same manner except that hemolysin-free complement was furnished in dose of 1 ce. of 1:20 dilution of guinea-pig serum absorbed by the corpuscles at a low temperature over night; the results showed the degree of inactivation or destruction of hemolysin as the result of heating. 2. One cubic centimeter of each active or unheated serum was now chilled and treated with 4 ce. of 5 per cent suspension of cells (also chilled) and the mixture kept at about the freezing point over night; the next morning the cells were removed by centrifuging and washed twice with iced salt solution. After the last washing the cells were suspended in 4 ce. salt solution and distributed in doses ranging from 0.1 to 1 cc. in a series of ten test tubes; complement previously absorbed with corpuscles for the natural hemolysin under study, was added in constant dose of 1 cc. of 1:20 dilution. After one hour in a water bath the results were read and the largest dose of corpuscles showing complete hemolysis was designated as the hemolytic index of cor- 420 JOHN A. KOLMER puscles and giving an exact measure of the degree of sensitiza- tion. These tests were repeated in exactly the same manner with portions of each serum heated at 56° and 62°C. for thirty minutes. With this technic it was possible to study the influ- ence of heat upon natural hemolysins as determined in the usual TABLE 18 The absorption of natural antisheep hemolysin from active and heated serum HEMOLYTIC ACTIVITY OF 5 PER CENT SUSPEN- | HEMOLYTIC ACTIVITY OF 5 PER CENT SUSPEN- SERUM SIONS OF CORPUSCLES AFTER EXPOSURETO SIONS OF CORPUSCLES AFTER EXPOSURE TO NUM- ACTIVE SERA HEATED SERA BER 0.1/0.2/)0.3|0.4)0.5|0.6)0.7/0.8/0.9}1.0)0.1/0.2/0.3|0.4)0.5/0.6/0.7/0.8/0.9) 1.0 1 |C¥;/CICICIM|IM|IMIMIMIM|MIS/S/Si/S/S/S{S/S1/S8S 2 ICIMISISISISISISISIS|IN|IN|INININININIS|S]S 8 IC |CI|CICICIM|M|MIMIM;|CIC|M|M|M/S/S/S|S|{S 4 IMIMIS{SISISISISIS|IS|NININ|INININ/ININ|N[S 5 IC |C|IM|IM|IM|MIM|M|MIM|IM|iMIMI|M/S/S/S/S/S]{S 6 IC |CICIMIMIS/IS/ISIS/IS/CIMIS/IS|IS/S/IS/SjiS1/S 7 IC|C{|CICIMIM|M|IMIMIM|M|S/S/S/S/{S/S{S}S]/S 8 IC |C|IMIMIMIM|M|IMIM|[MIS/S/S/S/S/S/S|{S/S]/S8S *C = complete hemolysis. TABLE 19 The absorption of natural antisheep hemolysin from unheated and heated human sera . oe, Le ae Ee HEMOLYTIC INDICES OF SERA SERA Unheated 56°C.* 62°C:* Unheated 56°C.* 62°C.* 1 0.5 0.5 0 O49 Ch PADS? 0 2 0.5 0.5 0.2 0.5 0.3 li, Ome 3 0.5 0.5 0.2 0.5 0.2 0.1 4 0.5 0.5 0 0.4 0.2 0 5 0.5 0.5 0 0.5 0.1 0 6 0.2 0 0 0.3 0 0 * Sera heated at 56°C. and 62°C. for thirty minutes. manner by direct titration and also by determining the degree of sensitization occurring in unheated and heated sera under identical conditions. These experiments were conducted with ten to twenty-five human sera for natural antisheep, antiox, antiguinea-pig, anti- NATURE OF THERMOLABILE HEMOLYSINS 421 rat, antidog and antirabbit hemolysins and the results observed with three to nine sera are shown in tables 18 to 24 as examples of each series; these results may be summarized as follows: 1. A general statement on the fate of natural hemolysins as a result of heating the sera cannot be made, as the various hemolysins behave somewhat differently by reason of variation in resistance to heat. 2. Antisheep hemolysin is most resistant to heat; after an ex- posure of serum to 56°C. for thirty minutes the hemolytic index is reduced although corpuscles are usually sensitized to the same degree as in unheated serum. After exposure to 62°C. TABLE 20 The absorption of natural antiox hemolysin from unheated and heated human sera HEMOLYTIC INDICES OE CORPUSCLES AFTER ABSORPTION HEMOLYTIC INDICES OF SERA Unheated 50°C. 62°C. Unheated 56°C.* 62°C.* 1 Onl 0.1 0 0.1 0 0 2 0.2 0.1 0 0.2 0.1 0 3 0.5 0.3 0 0.5 0.2 0 4 0.4 0.2 0 0.3 0.1 0 5 0.1 0.1 0 0.1 0 0 6 0.3 0.1 0) 0.2 0.1 0 * Sera heated at 56°C. and 62°C. for thirty minutes. for thirty minutes the hemolytic activity is lost or greatly re- duced and absorption removes a proportionately smaller amount of hemolysin or none at all (tables 18 and 19). 3. Antiox hemolysin behaves in a similar manner; the hemo- lytic indices of sera heated at 56°C. for thirty minutes are lower than unheated sera and totally lost after heating at 62°C. Absorption of sera heated at 56°C. shows the same or slightly less sensitization than occurs in unheated serum; absorption of sera heated at 62°C. showed no sensitization (table 20). 4, Antiguinea-pig hemolysin is usually destroyed by heating sera at 56°C. for thirty minutes and always after heating at 62°C. for the same period. Corpuscles are sensitized very slowly in unheated sera at low temperatures and agglutination 422 JOHN A. KOLMER frequently occurs; under identical conditions, I have only occa- sionally found corpuscles sensitized in sera heated at 56°C. and never after heating at 62°C. (table 21). 5. With antirat, antidog and antirabbit hemolysins heating sera at 56°C. for thirty minutes generally reduced or totally removed hemolytic activity although the corpuscles were sen- sitized to the same or slightly less degree than occurred in un- heated sera; after heating at 62°C. for thirty minutes hemolytic activity of the sera was totally destroyed and the sera failed to sensitize corpuscles (tables 22, 23 and 24). TABLE 21 The absorption of natural antiguinea-pig hemolysin from unheated and heated human sera HEMOLYTIC INDICES OF CORPUSCLES AFTER ABSORPTION HEMOLYTIC INDICES OF SERA SERA Unheated 56°C. 62°C. Unheated 56°C. 62°C. 1 0.1 0 0 0.1 0 0 2 0.1 0 0 0.2 0 0 3 0.1 0 0 0.3 0 0 4 0.2 0.1 0 0.3 0.1 0 5. 0 0 0 0.1 0 0 6 0.1 0 0 0.3 0 0 7 0.1 0 0 OL 0 0 8 0.2 0 0 0.3 0.1 0 9 0.2 0 6) 0.3 0 0 The general results of these experiments were to show that apparently a portion of the natural hemolysin in human sera for the corpuscles of the sheep, ox, guinea-pig, rat, dog and rab- bit may become masked or inactivated as described by Sherman, that is, while the hemolytic activity of the serum is reduced in a direct titration, the corpuscles become sensitized to the same degree as in unheated serum; but a portion of some hemolysins are actually destroyed by heating at 56°C. for thirty minutes as shown by the fact that sensitization is less or nil in heated serum as compared with unheated serum and this is particularly true of such hemolysins as those for guinea-pig cells which are more sensitive to heat than those for sheep and ox corpuscles. NATURE OF THERMOLABILE HEMOLYSINS 423 As is well known, sensitized corpuscles are more vulnerable to complement and hemolysis than plain cells mixed with the same amount of hemolysin and complement and “masking” of hemoly- sins by heat may be due to the production of amboceptoids TABLE 22 ‘The absorption of natural antirat hemolysin from unheated and heated human sera HEMOLYTIC INDICES OF CORPUSCLES atid AFTER ABSORPTION HEMOLYTIC INDICES OF SERA Unheated 56°C. 62°C. Unheated. 56°C. 62°C. 1 0.1 0.1 0 0.1 0 0 2 0.1 0.1 0 0.1 0.1 0 33 0.1 0.1 0 0.1 0) 0 TABLE 23 The absorption of natural antidog hemolysis from unheated and heated human sera HEMOLYTIC INDICES OF CORPUSCLES AFTER ABSORPTION HEMOLYTIC INDICES OF SERA SERA Unheated 56°C 62°C. Unheated 56°C 62°C j 0.2 0.2 0 0.2 Ont 0 2 0.1 Trace 0 0.2 0 0 3} Trace Trace 0 0.1 0 0 TABLE 24 The absorption of natural antirabbit hemolysin from unhealed and heated human sera my HEMOLYTIC INDICES OF CORPUSCLES AFTER ABSORPTION HEMOLYTIC INDICES OF SERA SERA Unheated 56°C 62°C Unheated 56°C 62°C ik 0.1 0) 0 0.2 0 0 2 0.2 0.1 0 0.3 0.1 0 3 0.1 0 0 0.2 0.1 0 which block unchanged hemolysins in the short exposure of one hour whereas in prolonged absorption of heated sera the cells may become sensitized with sufficient unchanged amboceptor to render them as vulnerable to complement as cells exposed to unheated serum; according to my experiments heating sera at 56 and 62°C. results in an actual destruction as well as inactiva- 424 JOHN A. KOLMER tion or masking of natural hemolysins, the different hemolysins found in human sera varying considerable in resistance to these changes. SUMMARY 1. Natural hemolysins in human sera are more resistant to heat and age than complement. 2. There is no relation between the complement content and natural hemolysins in human sera. 3. Absorption of active human sera with corpuscles removes hemolysin but not complement. 4. Washed sensitized cells have not absorbed complement and do not undergo hemolysis unless complement is furnished. 5. Filtration may remove complement from a serum without any or but slight removal of a natural hemolysin. 6. Heating serum sera at 56°C. for thirty minutes results in the partial destruction and inactivation (‘‘masking’’) of natural hemolysins, the different hemolysins varying in their resistance; heating at 62°C. results in the destruction of natural hemolysins. CONCLUSIONS 1. These experiments indicate that the natural hemolysins in human sera are distinct substances and not differentiated complements. 2. Natural hemolysins are susceptible to heat being inacti- vated (masked) or destroyed when sera are heated at 56°C. and totally destroyed by heating at 62°C. The natural hemolysins in human sera vary in resistance to heat, antisheep hemolysin being most resistant (thermostabile) and antiguinea-pig hemoly- sin being most susceptible (thermolabile). REFERENCES (1) Turetz, F. H., anp Empieton, D.: The evolution of antibody. Ztschr. f. Immunitiatsf. u. exper. Ther., 1914, 20, 1-51. (2) SHerman, H.: Thermostabile and the socalled thermolabile hemolysins. Jour. Infect. Dis., 1918, 22, 534-541. (3) Koumer, J. A., Trist, M. E., anp Frick, A. M.: A study of the natural ther- molabile and thermostabile hemolysins and hemagglutinins in human serum in relation to the Wassermann reaction. Amer. Jour. Syph. (in press). COMPLEMENTARY AND OPSONIC FUNCTIONS IN THEIR RELATION TO IMMUNITY A STUDY OF THE SERUM OF GUINEA-PIGS NATURALLY DEFICIENT IN COMPLEMENT HIRAM D. MOORE From the Veterinary Laboratory of the Vermont State Agricultural Experiment Station, Burlington, Vermont Received for publication December 19, 1919 SOME TITRATIONS OF GUINEA-PIG COMPLEMENT Several years ago, it was discovered in the Veterinary Depart- ment of the Vermont State Agricultural Experiment Station, in connection with the work on the complement fixation test, that the blood serum of some guinea-pigs is very deficient in comple- ment. The late Ramon C. Downing found that in some in- stances, as much as 1 cc. of guinea-pig serum, which he tested, produced no hemolysis of sensitized corpuscles. In the course of his work, Downing used two different hemolytic systems, one with sheep erythrocytes, the other with horse erythrocytes. His work is unpublished. At the time that this deficiency of complement in the serum of some guinea-pigs was discovered, Dr. F. A. Rich, head of the Veterinary Department of the Vermont State Agricultural Ex- periment Station, and Ramon C. Downing began a breeding ex- periment with such animals. The results of this experiment have not yet been published; I may say, however, that this deficiency in complement was found to be a heritable condition. As far as I am aware, the complement-deficient guinea-pigs at this Experi- ment Station are the only ones of the kind in existence; we have been able to multiply them by careful breeding and we have hundreds of them at present. ; The writer’s tests with the hemolytic systems used by Downing have confirmed this remarkable deficiency in complement. The 425 THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 6 426 HIRAM D. MOORE sera of several hundred guinea-pigs have been tested, which showed practically no hemolysis with 1 cc. of undiluted serum. Since most of the sera of the complement-deficient guinea-pigs, which we have, show practically no hemolysis in a quantity of 1 cc., we usually made the rough test shown in table 1 in order to determine whether the sera were or were not deficient in complement. In a few instances the serum of guinea-pigs that had been found deficient in complement when tested with sensitized horse’s erythrocytes, were also tested with sensitized human blood cor- puscles. Most of the sera so examined were found to be more TABLE 1 Rough test of the sera of complement deficient guinea-pigs. 0.5 cc. of 1 per cent washed horse erythrocytes and three units of amboceptor in each tube GUINEA-PIG NUMBER 1.0 cc. 0.3 cc. 0.4 cc. ooo+o oooqoce Se orororS 1 2 3 4 5 6 (normal) ++ ++ ++ ++ = Complete hemolysis; + = partial hemolysis; 0 = no hemolysis. * Over one thousand guinea-pigs, whose undiluted sera showed practically no hemolysis in a quantity of 1 cc. when titrated in this manner, have bee tested by the writer. . deficient in complement for the former system than for the latter. The result of this experiment, the protocol of which is presented in tables 2 and 3, seems to offer further evidence of a multiplicity of complement. In a further experiment a mixture of the sera of seven guinea- pigs, all of which had been found entirely wanting in comple- mentary power with sensitized horse’s corpuscles, was tested with 1 cc. of washed sensitized human corpuscles. With this hemo- lytic system, also, a deficiency of complementary action was demonstrable; 0.1 cc. was found to be the minimal complemen- tary dose as compared with the normal dose of 0.02 ce. COMPLEMENTARY AND OPSONIC FUNCTIONS 427 On the other hand the complementary action of these sera was certainly quantitatively greater toward human cells than it was toward the horse corpuscles, and the difference was possibly a qualitative one. No hemolysis whatever was produced in 1 ce. of 1 per cent sensitized horse corpuscles by 1 cc. of any of the undiluted sera. TABLE 2 Rough test of the sera of complement deficient guinea-pigs. 0.5 cc. of 1 per cent washed horse erythrocytes and three units of ambocepter in each tube DEGREE OF HEMOLYSI AUSE GuIENRCrIG 8 CAUSED BY THE UNDILUTED SERA NUMBER 1.0 ce. 0.3 ce. 0.04 ce. 0.02 cc. | 0.01 ce. 0 .1cc.* 688 0 0 0 0 0 0 31G 0 0 0 0 0 0 98Z 0 0 0 0 0 0 722 teeta 0 0 0 0 0 494 Sar 0 0 0 0 0 716 aeSF 0 0 0 0 0 68C =e4 ++ 0 0 0 0 73G (normal) Secs ara se ++ + 0 * 0.1 cc. of normal serum and 0.5 ce. of 1 per cent washed unsensitized horse erythrocytes in this tube. TABLE 3 Rough test of the sera of complement deficient guinea-pigs. 1 cc. of 1 per cent washed human erythrocytes and two units of amboceptor in each tube DEGREE OF HEMOLYSIS CAUSED BY THE UNDILUTED SERA GUINEA-PIG NUMBER 1.0 cc 0.3 ce 0.04 ce 0.02 ce 688 ara “EF 0 0 31G FESR i 0 0 98Z aEae staat 0 0 722 seni siecle 0 0 494 staat ae 0 0 716 apar a 0 0 68C Sar alain 0 0 73G (normal) ++ =P45 aes tale The question suggested itself whether the complement defi- cient guinea-pigs differ from the normal animals in their resist- ance to bacterial infection, and this question was experimentally examined. In one parallel series of experiments, no difference 428 HIRAM D. MOORE was found between the complement deficient and the normal euinea-pigs in their ability to produce complement fixing and agglutinative antibodies upon the repeated injection of killed bac- teria. The appropriate examinations of the sera of the comple- ment deficient animals, after the immunizing injections had been made, showed no increase in the complement content of the blood. This result confirms the original observations of Von Dungern, that the complement content of the blood is not affected by the process of antibody formation. On the tenth day after the last injection of the killed organism all of the animals of this series were given a small dose; (one-tenth of a twenty-four hour agar slant culture) of the live bacteria. As a result of this inoculation, one complement normal pig died, whereas all of those deficient in complement survived. On April 14 another inoculation was given, this time one-half agar slant culture of living bacilli and three-fifths agar slant culture of living cocci. As a result of this inoculation three complement- deficient animals and four complement-normal animals died. This experiment seemed to show quite conclusively, that the de- ficiency in complement did not interfere with immunity acquired through a systematic immunization. Subsequent results have confirmed this conclusion. In a second experiment, a comparative test of the resistance to infection of non-immunized guinea-pigs of the two kinds (com- plement-deficient and complement-normal) was made. In this test live cultures of Bacillus cholerae suis were used. Of 100 complement-deficient guinea-pigs 77 succumbed to the inocula- tion of the live bacteria, whereas only 20 of the 100 complement- normal guinea-pigs that were similarly inoculated died. This result seems to indicate that with the noted deficiency of comple- ment, there is an associated deficiency in natural resistance to artificial bacterial infection. While it may be thought that this deficiency in natural resistance to bacterial infection is directly due to the deficiency in complement, it is conceivable that other factors that may be concerned in the mechanism of natural im- munity, are lacking in these complement-deficient animals, and that the observed deficiency in natural immunity is due to the lack of these other factors. COMPLEMENTARY AND OPSONIC FUNCTIONS 429 In view of this possibility and in view of the known presence in normal blood of opsonins, to which have been ascribed some property of resistance to bacterial invasion, the present compara- tive study of the opsonin content of the blood of the complement- deficient guinea-pigs was undertaken. In this study the comple- ment-deficient sera employed were such as caused no hemolysis of 0.5 ce. of 1 per cent doubly sensitized horse corpuscles in a quantity of 1 cc. of the undiluted serum, whereas the normal sera caused complete hemolysis of the same corpuscular unit in a quantity of 0.06 cc. of a 20 per cent dilution. The technic of the opsonin determinations followed through- out the study was practically identical with that of Wright with some minor differences. The bacterial emulsion was made from cultures of the typhoid bacillus. In the earlier experiments horse leucocytes washed four times in 0.9 per cent saline solution, were used. For the further determinations leucocytes from the comple- ment-deficient guinea-pigs were used with complement-deficient serum, while leucocytes from normal guinea-pigs were used with complement-normal serum. The sera to be examined were prepared from blood that had been secured by heart puncture with hypodermic needle; the sera were employed undiluted. They were drawn into a Wright pipet with equal volumes of the washed leucocytes suspension and the bacterial emulsion and then well mixed on a glass slide in the ‘usual manner. The mixture, after being drawn into the pipet, the end of which was then sealed in a flame, was incubated for exactly fifteen minutes at 37°C. At the end of this period the seal was broken and the contents of the pipet was thoroughly mixed again and evenly spread on an absolutely clean glass slide, with the end of another glass slide in the usual manner of spread- ing a blood film. The smear was air-dried and then stained with Jenner’s stain. In the determinations all of the bacteria in 100 leucocytes were counted in every smear. The results of these determinations are presented in table 4. 430 HIRAM D. MOORE TABLE 4 Showing the opsonic index of complement-deficient guinea-pigs GUINEA-PIG Normalee.. Complement- deficient ....: Complement- deficient..... INormalae seer Complement- deficient..... 18, 8, 0, 4, 6, 1, 6, 5, 8, 3, 18, 2, 0, 0, 15, 8, 0, 6, 4; 6,.9, 10; 10, 5, 6, 7,.9,14 9, 8,41, 7, 6.6, bo 10" 7,10, 3949) 12 1a toe 12,9) 6:7; 9,3, °8, 15, 19; 6, 610, OGG, 8,7, 14:12, 19.15.90. 25 10: 105).8 AT 14, 4, 16, 8, 2, 12, 4, 12, 9, 9, 14, 16, 8, 6, 5, 10, 6, 18, 18, 9,7, 12, 15, 5, 20, 10, 18, 10... 0:4 1311 80) 2 6.846.450 0; 0,.5, 0:0; 2:-648006,°9-10) 24-358, 0) 1216, 2.00 24 Best 8 2) 4d 4g BS 2,0, 6, 2, 2,2, 4,4, 4, 15, 2, 5,1 , 2, 4, 18, 0,354, 2, 3.00, 0.0, B93 ’ ~~ 9, 10, 6, 5, 5, 8, 7, 6, 8, 11, 10, 6, 0, 9, 9, 10, 7112, 6, 18,0, 40, 12,8149, 12, 10,"8 2820) 8, 9,14, 10, 9, 13, 12, 10, 12, 15, 9, 11, 10, 10, 5, 12, 11; 12°10, 8, 10,°10) 0, 12, tie: 14, 11, 11, 6, 9, 14, 8, 0, 6, 11, 9, 9, 6, 0, 6, 519) 10; 11, 9,9, 13, 4,7, 0,4.5,a3011) a, 9, 10, 11:.8, 0; 9,11 10/ 1, 16 ios, 11. 5, 4, 6, 3, 5, 0, 9, 12, 10, 3, 5, 3, 0, 5, 3, 2, 6, 8:3, 1, 0;5,. 43.6.3, 0, 6roee al29: 10, 7:18, 6, 1, 0:72 0F 16, 9s 12aueee ts -7,-4, 45,8, 4:0, 6, 656, 8 Seta; 6°26, 5, 5, 6, 0, 0, 0, 0, 6, 4, 3, 5, 4, 3, 4, 5, 10, 1, 0,33, 6: 6.0.4, 415.0925: 7, 53,84 TOTAL INDEX 743 |) 0.323 240 877 0.4207 369 856 0.5408 463 431 COMPLEMENTARY AND OPSONIC FUNCTIONS TABLE 4—Continued I TOTAL GUINEA-PIG cow ots SOQ aws BSSSSS it 1D oo Say od is + Sen est pees TN orsth, © Pi Gre EG) = Ca ie ey ys a ns BAe © = es Se ee Ses SoS oc Bon ee Gas oe OMN ~ntos 1D AS Co) (Sh Gh See rao SS Lach _—) FSS) x a aaa Cae pares. SnCu eace 3 RS NS ae = ms Seo: 19 Ss se ms ~ ~ Oi HO GO: De SLO ee ~ON 5 S'S =: ~ 1S ~ .S are Be eet st oe Shes ew ae SO Corres ee en Nig ~ oO 7 tis ek See Oa WSS Fee a Se Sn Say = tren st | 00 - Sie Awad NA 7 Ope eS rat ae Aa : ieee a Se) Hs Gh ete NONE ee ae On ee eo eens ays). ors LS eae ics ‘eeoN Seeaass - ga. ONS ae ~ AO oD = 4) « ada o esa ~ on a lipfes: rie aw CaS oo = 69) ~ ~ 4 SS a 1a a eae eC See eh a ee ey Serene eS tt SS nee ee See Suns SH re reas NL nN ey ao fae en z ~ Oey = 2 By Wael oe OP es ec a SeaN Oo 7 nw NONNON aS eel ee Ses Fe 28 SP CIS Ree pha ee athe a - ee e e Sn ie eas a OREES ge =e Hie, OD) rt) ie, CN ites onMON ~ 26 suee: 3 wD ob S) Geaigel 2s + mS) pap ~©O BES OS) art ce Not elen min gre OO ae is Se) cy 2c aS MOS yn ese) OT es oD PEO E ie es mooon Ree Sate Aon Posen or Kol Sie set na An ee | Teton ae Ne n . Cet ye re n Ses: Gelert ee molt ase . . 10 SVS ee SS) Een ee SSS 5 Ms yo S) OW ~ wn ~ ns Aer erie) apace | bap den ke mene a co Ss CQ) ian Si ce SLs 1 or) SH MOO 7 DW OnNnnoooce oe a ou on nS) oo os cot Ces RUM NC eiiess Resiiten, pay Nes, en how tent nein SiS ES se) « S SHEEN GS MHON CAMH AWAAtMAOMT Soon oOoCO aN a 68 FON ao! Wig n o 4 oo Oo a md = S NA i : : he : © 1 ‘ 1 > id i] ~ coe ¥ ds =a : 5+ : oS ¥ i a a : a+ : a § . a 8 - oe gq 8 ee On = O-m os] =a 0 i= Oem 2) ost = Oo g forms] 3 = 0 5 ae E as E as E Ba Sol a iol as o} ot H ie) ° o) ° fe) Pa O [o} fe) Z S Z O Ae Z 0 nt —————__—___.,_—__——_—_———" Re) Ne) ~ NDEX * Counted as a check on number 5 by the serologist of the State Laboratory. 432 HIRAM D. MOORE TABLE 4—Continued GUINEA-PIG TOTAL! INDEX SS eed 50; 41 S88S. de Oso 1) 11S, 0k Baer, Normal.......}| 25, 14, 12, 29, 15, 20, 19, 23, 4, 6, 3, 4, 20, 25, 22,2, 15, 42, 6, 32, 4, 30, 4, 2, 0, 8 8, lh 1), 1556.12709 0199 12).40). ih eh 65s 8 (| 4, Oy OF 010-49; 0, 3,735130; 10,1 0/39 2s 0; '2) 0.5421 Complement- 1A, 32:40\'0;/ 25; By KBR ALO, Ta 55, 12) deficient.....|| 0, 4, 7, 50, 0, 12, 7, 0, 2, 20, 1, 4, 0, 0, 0, 0, BM: 5p. Ag tO, a eiiee Me, bra a ee 354 (| 25, 1, 6,32, 6, 28, 5, 16,30) 6, 4, 24, 12;.7, 9, 4° 10/18, 12/376 TO nish 20 Nak 7 OS 38N A; 5, 6R43; 45205, A p40 ee 10 10,10, 14, Normal....... 5:0) 0,,0912, 2.045) OH wea, te 14, 9, 7, 0, 0, 0, 7, 4, 1,0, 0,8, 30, ce 6,'5,.0, 8)-1,:5) 28, 6) 0,/15715,.1, 2,0, 19, 12: 6; 05 36,8) IANO ITPGRSl eae 902 |) 0 0 4, 7, 0,3, 8, ty; t,.2, 0; ? 0.4955 Complement- } deficient. .... 40, Normal.. 10 (| 56, 0, 12, 0, 29, 49, 34, 0, 0, 0, 53, 6, 36, 0, 6, 26, 11, 17, 0, 11, 10, 3, 19, 2, 0, 3, 22 2, 28, 5, 0, 2, 0, 10, 3 4, 0, 6, 6, 0, 0, 6, 1, 4, = AU: 6 0, 6, Complement- || 0,5, | deficient....5| 3, 0, 1 0,3 , 0, 0, 0, 52, , COMPLEMENTARY AND OPSONIC FUNCTIONS 433 TABLE 4—Continued Se ee eee | GUINEA-PIG TOTAL INDEX 55, 7, 45, 10, 10, 0, 2 1342) 13.0).4.,7, 1 15, 0, 12, 60, 0, 5, 0, 3, 0, 9, 10, 0,5 TeOn2e2,Otsge a: J BASS 118).5,-0; 4.01 5, 3,3, 82) 10 Normal..... 7 , 40, 0, 11* 50, 0, 24, 0, 0, 0, 0, 2, 2, 4, 42, 10, 2, 20, 5, oun 25, 8, 3, 6, 0, 2, 34, 6, 10, 1, 2, 0, 2, 18, 0, 0) 2, 2.0, 2,2, 4,30, 2) 2° 0, 32.790, 32. Complement- deficient... Ny ay 2050,-0;,0) 0,3; 2 2 "18, 14, 0, 0, 0, 0, 0, 0, 0, 0, ee 05:0), 2.5, o9 oy 11, Normale. .: 0; 14; 0, 0, "40, Wena 7, 0, 26, 0, 0, 25, 0, 3 0, 0, 0, 0, 5 12 Complement- 0 0,9 00; 0.4761 deficient..... 6 0 0 21, 0, 12, 15, 6, 0, 12, 8, 0, 18, 0, 0, 28, 19, 15, 6, 18, 27, 7, 6, 17, 7,9, 0, 5, 0, 14, 3, 25, 0, 25, 20, 48, 20, 10, 0, 13, 6, 0, 0, 22, 20, 6,38; 2, 6,.0; 0, 7, (6, 2, 0, 23,10) 16,2, 5, 2, 23, 18, 16, 28, 19, 39, 0, 0, 17, 10, 4, 1, 5,42) 0. 21,315, 0; 12:7 8, 0, 3.5, 0, 6, 18, 11, 6, 18, 0, 17, 24, 27, 13, 8, 25....| 1082 0 3; 2 1 —--—-. a ro OO OO ooo oO —_ seek Roe eh Soa oo! econ moO oO Worn So SUS xs 13 es, = S = bo ~~ ; Te 722.0% 31 bn S40; 0,0, 0 24.0, 0; 0: 080,18, 23, 2,10) 7, 31, 6, 0, 0 16,20: 2: §0;0,1027..9-45).0;0;, 0; 23.2 0, 12: 5,0; 2, 3: 0; 4, 0;:0,080.,1 0, 45, 0, 0, 4, 0, 0, 0, 0, 16, 5, 0, ~ Complement- deficient .... : 0.6663 d 0 3 1s 1918, 22. -0;.0; 0, ~ 5, 1, 8, 0, 22, 0, 19, 0, 0, 9, 24, 0, 0, 14, 18, 18, 4, 0, 2 0, 14, 13. Bain seo Be) mea 0 6, 0, MOORE HIRAM D. 434 TABLE 4—Continued Som I iS iS Ns 8 z S % oO oO Oo —_——-—--_-r-— er e——————————— ree ee q io) Nn a) sH B o 83 iS 5 3 A ~ Yes ror) oS & eA Mo aie OG ar Bae a le TOS OO ON On) a (<=>] Chas jee ye hee ee es are eT es ee tN tte Te Te ell =) re ri ee ae Reason eo eo is tear i i Sas tae Sa OS) S an CN is Ae) ey) cl Wet es foo! oo a© SC en NTT ,T~ wo COk Ste a: el Se ee ee ele Se Telarc a S| A EAE) aera aco IDO os ~a© axe a SS cs (ore) oo = 00 t=) 5 cy 0 mS sO © oe oO ns (on) aH AQ «~ 2-0 4 Orpen ee 5 N 7.210 . ~10 FH ACO MO ios) ox EES Srl ao Ss ain Bs) S00 OS moos .Q oe (ar), WR CDie aes Gots - «19D on oe Onn Se eee lel vere We eer te Se etal Go NS oy nS OY Bie persevere es ONO - --OOnS ~ a AOR = a art © > . SR CO ee cee oNmR ool See SoG eR) Orr wate tninn =i) CONG a a) Om ~DO +N ee Complement- deficient. . Complement- N i (Veta Ce et Te) ee a N deficient.... Complement- * Counted as a check on number 17 by the agricultural bacteriologist. kh MOORE HIRAM D. : | : TABLE 4—Continued GUINEA-PIG Normal.... Complement- deficient Complement- deficient.... Normal... ge Pe M Je) e a z ia] —) (=) OW —-_ ~~ SS nN 4 oa) = 3 fen 13 PE St nN oD 5 Yo) a 6S © i is) row oNOS : Socios & Sonate : CODON D Soa ado SHaSSOS OISS Se coneow Oot ON toi Ot So oa Aid k Oo A SM aS a ~ aA ers ted Ce ps aes ary es ee Ss oe Sees ri a * cas oo cone d ~ oO ol anaowo So Or ee RaW ap c=) SS: aS Beat ga watincia) So ION ces io ee Se ee eae <7 nigortinan NS consi ee SVEN = cere o Rod) SO ay 50 «SH tie wen ~~ Rofo) ~ ~aSSO «a» oy 0 oO i ie aa os Se ON et = os iS ono ee aay epee 1 ON of = 60 es ~ O00 ow SE — ST — I ee ea) Sean Nv a ane Sr ae i=l aw Oo Py le is AT CN) ceed me Oo Ir wt 0 - (Os AimVcepl SS) si ae CN om SN it SON OD en may ne cs So) ees eae OO i sata Drees ao tees Cea ets Be Aon . . (I bo) OANA YT OH ~O YN es = ASA ois Ge) a Sey an ee ig epee eee mt = = S Ses Sy Des eee eee acta cm SOSS MQ HiGglonneo soa os 5 SOR AN S OD eg OO a> Dooce O as mo Rent he. Seta S On) =F} syicall rs “A Ye) COR ty a Pees Sreyye Ce) DAL [oS 2. SON Ose isd an ON 219 Ne) ets > SCoMen oOo. © i 5c CNN Sepedigees: ard = ated oS ee = fae oOoeP LNA ~ OD A ee ma Sop Re. me aw oO .-0C «a ae eS Seki (api lees woe cents Sonmoood S50 oe ee aS ~AS 8 ea SSM SN = nN ae UE * ~ eee act nin ex oe gat) SP ere pes aps > si) wan os) es aya ere S erin. on Pel Sw o ie Qo fe = wa No ee ee SH xs (57 =) ie eer = | a RICO) ie PRA. | - oO fan) ye) = oe owt me ey ener a Sie - ON EO NNT SCOD oO sl alee pe SEOIGN im en ey lee Ne ae) ce Ho. gon oO Nivel OO... ~O 09 I MO AO WA ta 8 -~ CRE cles ~O MD . .10 © Oo a re ONb- a) (er (mim cn GEN eee ee NHN UNO at Lone a Jif pee Ea IP SE SEREGES ~MOnMNO Nr OON © [or] ey Complement- deficient... ~~ A Os os oR o Gr) Sas OB be oe i) aa oO woo Ce ee Nee Se et ’ , 6, 30, 3, 4, 18, 3, 13 4 (S617 s11,48 5365 2 5] 1 ’ 0 1 0 0 Diners ’ ’ 0 ? 4 ? SS eee Mee here o SSN ees ae & Ss at OS of oa couse Serato. Sa ass SaNSsgaesc Owfod= Tow ise Ovo cot Soo sGRdd SATS N N Complement- deficient ' ° Oo eo o — o q ° ca) Q mo) o ~ n o an S| co HH o ~ o 3 Q Lal ° he o 2 q =} q o q ~ nm ~ q o n o HH Q o mH HY o Q g =} =| | 12) ios} co fa & e} Z o ~ ral ° on = Q 3 ~ o qd ~ ut ° os (=| o o qd ~ we fo] q o A &0 a a o = aq z ra o ne) a o oO ins} mH o > is} o qd r=) Leal fo} q = ~ 3 g | ~ n o o am ~ oS o ~ oe g 5 ica) eo o n ° ry il 3 A, o mH a =) rei ol o a q = =| Seal S) n 5 ° Oo o a ~ COMPLEMENTARY AND OPSONIC FUNCTIONS 439 TABLE 4—Concluded Summary of all the counts excepting number 10 and its check number 11 COMPLE- MENT- NORMAL DEFICIENT 240 743 369 877 463 856 180 279 361 734 229 387 164 402 354 653 447 902 350 735 721 1,082 298 635 532 785 374 953 215 571 97 288 212 426 155 592 441 704 434 628 177 607 165 739 272 547 360 553 207 736 Tat Tk eee eae rns GBS oe BABS Ce cite BORO RE GUAT gis rar cas cic too 7,817 | 16,464 Average phagocytic index... .....----s sees eee cere eees 3.1268 | 6.5856 Average opsonic index of complement-deficient serum...... 0.4748 eee The foregoing determinations demonstrate a distinctly subnor- mal opsonic power in all of the sera of the complement-deficient guinea-pigs. While this result might be thought to indicate some relationship between the opsonic and the complementary functions of serum as has been assumed by some, it would seem to contradict an assumption of identity of those two func- tions; because in these sera some opsonic property, nearly half, is retained, whereas complementary activity seems to be com- pletely absent (less than 1 per cent of the normal activity). 440 HIRIAM D. MOORE SUMMARY The great deficiency noted by Ramon C. Downing in the com- plementary power of the blood serum of some guinea-pigs, has been confirmed by the writer’s experiments. Complement titrations made by the writer, in which positive and negative control was employed, seemed to indicate that the apparent lack of complement in the blood serum of some guinea- pigs is not due to the presence therein of anything interfering with the action of amboceptor. The biological tests applied to the blood serum of complement- deficient guinea-pigs which had been systematically immunized by repeated inoculations, gave positive evidence of the presence of both bacteriolytic and agglutinating antibodies; moreover, when inoculated with large doses of live virulent culture, such guinea-pigs manifested more or less active immunity to the bac- teria injected. Titration of the blood serum of complement-deficient guinea- pigs, before and after the production of active immunity, gave practically the same complement titer. The results of experimental inoculation of one hundred non- immunized complement-deficient guinea-pigs and a like number of complement-normal guinea-pigs, seem to indicate that com- plement bears some relation to natural immunity. The results of counting the bacteria in 4800 leucocytes of complement-deficient and complement-normal guinea-pigs show that the phagocytic index of the serum of complement-deficient guinea-pigs is about one half that of the complement-normal animals. CONCLUSIONS The conclusions drawn from the foregoing study are as follows: 1. That the apparent lack of complement in the blood of the guinea-pigs used in this work was not caused by inhibition of the amboceptor. 2. That the lack of complement is not necessarily inimical to the production of acquired immunity. COMPLEMENTARY AND OPSONIC FUNCTIONS 441 3. That the animal tissues do not make up any deficiency in the complement content of the blood during active immunization. 4, That the complement titer varies with the opsonic-index and in the same direction. 5. That the opsonic determination is too tedious to lend itself readily to practical use. 6. That complement and opsonins bear a definite relation to each other. THE JOURNAL OF IMMUNOLOGY, VOL. Iv, NO. 6 INDEX TO VOLUME IV Acid hemolysis, The mechanism of boric. Pace . 85 Agglutination and agglutinin absorption, On the exiciomes of. a | multiplicity of races of B. influenzae as determined by.......... .. 359 reactions, A method for the production of a iHasnaeaorta): pucpension of Bacillus sie acis to be used in. ae . 105 Agglutinin absorption, agglutination onl On the. enetenee nO? a Sroluiiers of races of B. influenzae as eioenned Dect ratane “se sox ga Oe Agglutinins, Experiments on the effect of. SOR ri tne Ors Rae nT Anaphylaxis reaction in the rabbit, The sehen ae BOR Mee A) ——, The perfusion experiment in ine study of.. seh 4 S . 209 Rederan) Lillian M., and Mellon, Ralph R. ieareneterie cree tied of spore and vegetative stages of B. subtilis . er: . 203 Antibodies in uxalated tid and serum, A Coenpeeative study, oF Homolyue complement and........... 51 UC Antibody production in eb urts Follomine prreccion soil nancreanta Seaton 19 Antihaemotoxin, Observations on the production of an, for the haemotoxin of Bacterium welchii (Bacillus aerogenes capsulatus). ..........+.. 24.2 885 Anvinuman haemolysin: The production’ of 60.0... 6.2.0. s see ete ween 141 — serum, rabbit, Experiments on the removal of hemagglutinin from..... 275 Antilysins (anticomplementary substances of human eae A study of the thermolabile and thermostabile . Mis ote SEN ate sei De ein eR oOe Antitoxic dysentery serum, A new meted of joacnioe ae San ahs ee Renmei gins) O40;5 Antityphoid serum, A new method of testing. . : BP cette Meo, Bacillus anthracis, fa method for the production one a omopeneons suspension of, to be used in agglutination reactions. Se RAT CEES HES Chad —— influenzae in clinical influenza, The rdéle one CA ; Stoo LOW Bacterins, pneumococcus, The Beene aeiion ia he a Hole blood of mipbits followine*the inoculations Oe io.) 0. Gs ise eos slaw ac hee ee cot 147 Bacterium welchii (Bacillus aerogenes capsulatus), Observations on the pro- duction of an antihaemotoxin for the haemotoxin of. Pasar. .. 380 Bactericidal action of the whole blood of rabbits followin Gnoculnvicuel of pneumococcus bacterins, The. a aia sch enon ae, a C. B., and Sehenide, Carl L. mM On me cell plopalen aE Ue eed 29 : ee iicnzac. On the existence of a multiplicity of races of, as determined by agglutination and agglutinin absorption........ . 359 —— —— Pfeiffer, Notes on the bacteriology of, era on nthe peletiive eae of 189 —— ——, Some suggestive experiments with; its toxin and antitoxin......... 233 Boushian T. Harris. Studies in protein intoxication. III. Visceral lesions in rabbits with chronic protein intoxication . é . 213 ——. Studies in protein intoxication. IV. Feceolesty lesinas produced ae AEC COM SROTATE TOE ONI:, <1 255 'de Lt) ol Hater oete ou ho iaia' a's: «io0Giar sighs Gia sels «ton OO 443 444 INDEX B. subtilis, Immunologic disparities of spore and vegetative stages of........ 203 Burrows, Montrose T., and Suzuki, Yoshio. The study of problems of im- munity by the tissue culture method. III. A method for determining the resistance of individuals to diphtheria infection.................. 1 Cell globulin, On red... Ee ai yh is We ate Ean Sea es SN ne Clinical influenza, The role of Bactitie Mfieieoe in. 3 arn ist aa Coca, Arthur F. The mechanism of the cess! Seen in ae rabbit 219 ——. The perfusion experiment in the study of anaphylaxis................ 209 Complement, A study of the serum of guinea-pigs naturally deficient in.... 425 Complementary and opsonic functions in their relation to immunity....... 425 Cooper, Georgia M., and Valentine, Eugenia. On the existence of a multi- plicity of races of B. influenzae as determined by agglutination and agelutinm absorption’: 252.202. 10c .4 siehecensn ose eae oe oe ee 309 Desiccation, The influence of, upon natural hemolysins and hemagglutinins in ae sera. pee .. 393 Determining the fesistante of Gadvaunle te ‘diphtheria ateeHiont ‘é jnothed {00-25 bide lah el iala, hoa) «/eienevacél