ON THE MECHANISM OF OPSONIN AND BACTERIOTROPIN ACTION : VI. AGGLUTINATION AND TROPIN ACTION BY PRECIPITIN SERA. CHARACTERIZATION OF THE SENSITIZED SURFACE

THE EFFECT OF COMBINATION WITH DIAZO COMPOUNDS ON THE IMMUNOLOGICAL REACTIVITY OF ANTIBODIES

Sufficient coupling with any of five different diazo compounds eventually destroyed the reactivity of all the antisera here studied. The rates of inactivation varied considerably among the several antisera. By stopping the reaction at intervals, it was possible to prepare partially inactivated antibodies of peculiarly modified reactivity. Thus, the flocculating activity of diphtheria antitoxin with toxin was completely destroyed long before there was any demonstrable impairment of its protective titer in vivo. The first change induced in antipneumococcus horse sera was the apparently complete loss of reactivity with the capsular carbohydrate at a time when the agglutinating, animal-protecting and complement-fixing activity of the sera were only slightly affected. On further coupling, the sera no longer caused visible agglutination; but aggregation of the serum-treated bacteria could be induced by centrifugation. Still further coupling destroyed all antibody activity. Rabbit antisera to egg albumen and horse serum no longer precipitated the homologous antigen after treatment with diazo compounds, probably due to their failure to combine with the antigen. The hemolytic, complement-fixing and lipoid-flocculating activity of coupled rabbit antisera to sheep red blood cells fell off in parallel; the hemagglutinin seemed somewhat more resistant. The reagin of syphilitic serum was destroyed almost instantaneously by comparatively small amounts of diazo compounds. Finally, in the case of antityphoid agglutinin, the isoelectric point of the coupled antibody, measured on the surface of specifically sensitized bacteria, was found to shift from an original value of pH 4.7 to one of less than pH 2.7 as progressively more sulfanilic acid radicals added on to the antibody molecule. The groups in protein which participate in its reaction with diazo compounds probably include aliphatic amines, the imidazole ring of histidine, the indole group of tryptophane, the NH of proline and hydroxyproline and the phenyl group of tyrosine. Although it has been possible to modify antibodies chemically so that they combine with the corresponding antigens without causing their aggregation, the experiments here described furnish no indication as to which of these groups in antibody protein are primarily concerned in the antigen-antibody reaction, and which are responsible for the secondary flocculation. Such localization awaits the development of a technic for attacking individual groups in the protein molecule.

THE DEFORMABILITY AND THE WETTING PROPERTIES OF LEUCOCYTES AND ERYTHROCYTES

The resistance to deformation of polymorphonuclear neutrophile leucocytes under the conditions of our observations has been shown to be on the average considerably less than the resistance to deformation of large mononuclear leucocytes. It is recognized of course that the viscosity of leucocytes, as of other cells, may be markedly influenced by osmotic conditions (17), by the reaction of the suspending medium (18, 19), by temperature, or by injury (20, 21). Although the conditions of our observations were quite different from those of the body, they were nevertheless closely similar to those of simultaneous phagocytosis experiments in which the cells functioned exceedingly well (3). Moreover E. R. and E. L. Clark (22) have noted that polymorphonuclear leucocytes in the tails of living tadpoles were more fluid than the macrophages. And Goss (23) in microdissecting human polymorphonuclear neutrophiles reports that they are more fluid than the clasmatocytes and monocytes studied by Chambers and Borquist (24). Other types of leucocytes have in our experience seemed to fall between the large mononuclear and the polymorphonuclear leucocytes in their average resistance to the interfacial tensions. The leucocyte of each type studied is surrounded by an exceedingly delicate membrane. This membrane appears under the dark-field microscope as a pale, silvery line not distinguishable by inspection alone from a simple phase boundary between two immiscible liquids. That this is a membrane, however, and not a mere interface between immiscible phases, seems certain. In the first place the cell cytoplasm and the suspending medium are not immiscible. When the cell organization is broken down by the interfacial tension the greater part of the cell contents is immediately dissolved or dispersed. Goss (23) has noted that when the membrane is torn with a microdissection needle disintegration at once spreads over the membrane and the cytoplasm undergoes profound change. Moreover it is improbable that a simple phase boundary could exist in the presence of so much protein, lipoid, and other surface active materials as are present in protoplasm; the tendency of these substances to lower the free interfacial energy must necessarily tend to their adsorption in the interface until, if sufficient material is available at the interface, an adsorption film or membrane may be formed. Kite (25), in a pioneer microdissection study, described the polymorphonuclear leucocyte as "naked" protoplasm. The contradiction between this statement and those just made is more apparent than real. For the capacity swiftly to form a limiting membrane between itself and other liquids is an attribute of "naked" protoplasm, as has been shown by the beautiful experiments of Chambers (20). The present study of the wetting properties of leucocytes shows that their external membranes are hydrophilic, a character suggesting a surface in which proteins, probably bound water and salts (27), possibly the polar radicles of soaps or fatty acids, rather than non-polar lipoid groupings, are predominantly exposed. This makes it the more remarkable that a cell of such fluidity as for instance the polymorphonuclear leucocyte, composed largely of water and of water-soluble materials, should maintain its integrity in an aqueous medium with the aid of a membrane so delicate and so mobile. The mobility of the membrane, frequently extended in forming new pseudopodia or spreading over the surface of particles being ingested, must require constant entrance into and exit from the membrane of component materials, and their constant reorganization there. The limiting factors in the reformation of such a membrane would be the amounts of adsorbable materials available and their rates of movement up to the surface rather than the time required for orientation there, since the latter phenomenon is exceedingly rapid. Harkins (29), for instance has calculated that at a water-water vapor interface at 20 degrees C., from the area occupied by one molecule of water, a molecule would jump out into the vapor and a vapor molecule would fall into this area of the surface 7,000,000 times in one second; the time of orientation of the water molecule he estimates to be of the order of 1/100,000,000 second or less. The mammalian erythrocyte possesses a surface membrane capable of being folded and of withstanding tension in the interface. This has also been stretched by microdissection needles (21). The surface of the erythrocyte, as evidenced by its wetting properties, is relatively hydrophobic, relatively non-polar in character, as compared with the leucocyte. Evidence indicating that the erythrocyte surface contains both lipoid and protein components has been summarized in earlier papers (8, 30). We have little to add here other than to point out that the wetting properties of the chicken erythrocyte surface are similar to those fully described for the mammal. A serious source of error in certain isoelectric point determinations is discussed.

CARBOHYDRATES ADSORBED ON COLLOIDS AS ANTIGENS

Evidence is here given that polysaccharides can be rendered antigenic by haptogenic adsorption upon a colloid carrier. The polysaccharides studied were those of B. anthracis, the meningococcus, Streptococcus viridans (Bargen), B. proteus, S. morgani, B. dysenteriae, both the Shiga and Hiss types, and the pneumococci. With the polysaccharide of Type III pneumococci, we have been unable in 6 weeks to produce any detectable protective antibodies, but we were able to produce anticarbohydrate antibodies. All the bacterial carbohydrates were non-antigenic alone when used in the doses indicated, though containing some nitrogen. Dextran, which was free from nitrogen was also rendered antigenic by the adsorption method.

THE PROCESS OF PHAGOCYTOSIS : THE AGREEMENT BETWEEN DIRECT OBSERVATION AND DEDUCTIONS FROM THEORY

The phagocyte, then, is a complex system delicately responsive to internal and external influences. Interfacial tensions, and under certain conditions viscosity, are critical factors in determining the ingestion of particles with which the phagocyte has come into contact. Deductions from the formulation of these factors by Fenn and Ponder are in agreement with observation and with experimental analysis. However, other and still unformulated forces also enter into the behavior of these remarkable cells.

CHANGES IN BACTERIAL VOLUME AS THE RESULT OF SPECIFIC AGGLUTINATION

Measurements indicate that bacterial antigens increase in volume as the result of specific agglutination. There is a general parallelism between the increase in antigenic volume and the concentration of the immune serum. The phenomenon is specific. There is no increase with normal serum; with absorbed serum the increase is slight and it can be correlated with the presence of unabsorbed antibody. The effect is enduring as shown by volumetric determinations upon repeatedly washed, agglutinated bacteria.

AN ANALYSIS OF THE OPSONIC AND TROPIC ACTION OF NORMAL AND IMMUNE SERA BASED ON EXPERIMENTS WITH THE PNEUMOCOCCUS

1. In normal unheated human serum, virulent pneumococci may be prepared for phagocytosis by two separate antibodies, acting in conjunction with complement. One of these is the type-specific anticarbohydrate antibody reacting with the carbohydrate fraction of the pneumococcus. The other is probably also a type-specific antibody, but quite distinct from the former, and therefore must react with a different antigenic constituent of the bacterium. 2. In the normal human serum heated to 56 degrees C., these two antibodies may, after prolonged contact with the organism, promote phagocytosis of the pneumococcus without the adjuvant action of complement. 3. Although these two antibodies are equally effective in the phagocytosis of 24 hour culture organisms by normal blood, the anticarbohydrate antibody tends to become the predominant factor as the pneumococci approach the state in which they exist in the animal body. 4. In so far as we have been able to show, the anticarbohydrate antibody is the only antibody in immune serum which can induce phagocytosis. This substance by itself is active in a phagocytic system, but just as in the normal serum, complement enhances its effect. The failure to demonstrate the presence in the immune serum of an antibody, distinct from the anticarbohydrate antibody, analogous to that found in the normal serum, may be due to the experimental difficulty of removing all the anticarbohydrate antibody from a concentrated immune serum. 5. Thus it is seen that a single well defined antibody (the anticarbohydrate antibody) may be responsible for the phagocytic action of normal unheated serum, normal heated serum, inactivated immune serum, and immune serum activated by complement. These facts appear to us to invalidate Neufeld's division of the phagocytic antibodies into (a) bacteriotropins (antibodies, the phagocytic titre of which is not raised by the addition of complement); (b) opsonic antibodies (antibodies, comparable to the lysins, which are only active in the presence of complement). 6. Complement alone is incapable of inducing phagocytosis of the pneumococcus. In the phagocytic process, it appears simply to increase the speed at which the reaction takes place. Its role may be compared to that of a catalyst in a chemical reaction. 7. On the basis of these findings, it is proposed that the term "tropin" be discarded as misleading and unnecessary, and that the term "opsonin" be retained to denote any heat-stable antibody which prepares bacteria for phagocytosis. Contrary to current usage, it would not suggest a combination of antibody with complement.

THE MODIFICATION OF ANTIBODIES BY FORMALDEHYDE

Certain strains of bacteria which have only minimal zeta potentials over a wide range of pH, and upon which surface deposits can be formed, afford a favorable means of studying certain chemical and physical properties of the surface deposits. Films of specific antibody-globulin upon these bacteria possess basic groups which can combine with formaldehyde. Combination of these groups with HCHO under the conditions of the present experiments shifts the isoelectric point of the sensitizing film toward the acid side by about 0.6 to 0.8 pH unit, and reduces the agglutinating tendency of the sensitizing film. Antibodies may be formalinized before combination with antigen without marked change in their specific combining affinities. The properties of the sensitizing films are similar whether formol treatment occurs before or after the antigen-antibody combination. The nature of the basic groups has been discussed.

THE IDENTITY OF THE MECHANISMS OF TYPE-SPECIFIC AGGLUTININ AND PRECIPITIN REACTIONS WITH PNEUMOCOCCUS

The experimental results which have been described demonstrate the following facts: 1. In the type-specific agglutination reaction, when the organisms are not present in sufficient numbers to absorb completely all the antibodies from the serum, more antibody is bound by cellular S than is required for the process of agglutination. 2. The excess of antibody thus bound can then unite with additional amounts of the specific substance when this is added in soluble form to the agglutinated material. 3. If an excess of the free S is added to an agglutinated mass of antibody and bacteria, the organisms are redispersed and in the suspended state are again specifically agglutinable. 4. When a solution of the specific polysaccharide is added in excess to an homologous immune serum, a prozone is created in which precipitation is inhibited; moreover, if, at this point, type-specific pneumococci are added to the mixture, inhibition of agglutination also occurs. 5. The reactive substance in the type-specific agglutination and precipitation reactions is the same, i.e., the capsular polysaccharide. In the former instance it is combined in the bacterial cell; in the latter, it is in a soluble, chemically purified state and entirely separate from the body of the cell.

CONDITIONS INFLUENCING THE DISAPPEARANCE OF LIVING BACTERIA FROM THE BLOOD STREAM

1. The simultaneous intravenous injection into normal and actively immunized rabbits of equal quantities of living staphylococci or paratyphoid bacilli is followed by a distinctly accelerated rate of removal of the bacteria from the blood streams of the immune animals. 2. This altered reactivity is due essentially to specific active immunization. 3. The bacteria pass rapidly through the capillary bed of the lungs, extracellularly and dispersed for the most part, and become generalized through the blood stream. 4. The bacteria are quickly removed from the circulating blood in the immune animals and less rapidly in the normal ones, by various organs, particularly the liver and spleen, where they accumulate in enormous numbers, become adherent to the lining membrane of the sinusoids of the liver and apparently to the macrophages of the spleen and are phagocytosed by the macrophages and leucocytes in these organs. 5. Associated with this effect are morphological changes in the bacteria as shown by swelling, loss of staining power and evidences of increased cohesiveness and decreased viscosity, these changes being apparent as early as 2 minutes after their intravenous injection. 6. Inasmuch as these changes are not seen to a marked degree within the lungs or other organs, they are probably the result of a local antigen-antibody reaction of a bacteriotropic type in the two organs generally considered to be most actively concerned with the production of immune bodies. 7. By means of this accelerated bacteriotropic effect in the actively immunized animals, phagocytosis is facilitated and intracellular digestion of the bacteria is enhanced.

TOXIN-ANTITOXIN REACTION WITHOUT NEUTRALIZATION

1. Collodion particles adsorb diphtheria or tetanus or botulinus toxins. These toxins are retained on the particles when washed but are at least in part released in the animal. 2. The adsorbed toxins are neutralized by adsorption of the corresponding antitoxins but are unaffected by other serums. 3. When collodion particles are treated first with tetanus antitoxin, then with diphtheria toxin, they are not toxic, but they become toxic when they are treated first with diphtheria antitoxin, then with the diphtheria toxin. Similarly when collodion particles are treated first with diphtheria antitoxin and then with tetanus toxin, they do not become toxic, but they become toxic when they are treated with tetanus antitoxin and tetanus toxin.