CELLULAR IMMUNITY SANFORD S. ELBERG. if its generation time is so long that it causes too great a tax on host cell culture and maintenance:

Transcription

1 CELLULAR IMMUNITY SANFORD S. ELBERG University of California, Berkeley, California CONTENTS I. Introduction II. Response of Phagocytic Cells to Infection.. 67 A. Interaction between Parasite and Phagocytic Cell III. Response of Mononuclear Phagocytes to Mycobacterium tuberculosis and Brucella melitensis. 71 A. Description of Methods B. Responses of Monocytes Parasitized with M. tuberculosis Specificity of serum components Specificity of "immune" monocyte Some additional properties of serum component Effect of immune serum Time needed for serum component's action Modification of degenerative effect of tubercle bacilli by combined action of immune cells and immune serum IV. Reactivity of Immune Monocyte as Reflection of Response of Intact Animal V. Breakdown of Microorganisms within Phagocytic Cells A. Phagocytic Destruction of B. melitensis.. 87 VI. Summary and Conclusions VII. References I. INTRODUCTION The term, cellular immunity, has traditionally implied that the parasite is ultimately destroyed in or by one of a series of host cells. In support of this concept, there has accumulated a massive body of circumstantial evidence on the fate of microorganisms in the intact animal. There is, however, also a body of contradictory data on the survival and multiplication of microorganisms in cells maintained in cell-culture environments, much of the contradictoriness being a result of varying experimental conditions. On the other hand, it is difficult to accept such variation as wholly responsible for the existing differences in response to the question: does the large mononuclear cell of animal peritoneal exudates suppress the growth of certain parasites? If we are to examine this question and evaluate the pertinent evidence in some detail, certain criteria are necessary. These are: (a) enumeration of microbial populations should reflect the intraphagocytic census and not be complicated by I Grateful acknowledgement is made to the Microbiology Branch, Office of Naval Research, and to National Institutes of Health for grants in aid of the research described, in part, in this review.' including extracellular microbes or their descendants; (b) the methods to measure intracellular activity should include growth of the parasite if its generation time is so long that it causes too great a tax on host cell culture and maintenance: lengthening of cells, cording, and other morphological changes indicative of viability may be utilized; (c) in addition to morphological observations, present day studies should include a demonstration of persistence, latency, or "unit growth" such as protoplast-like formation and its perpetuation; (d) precautions should be taken that aggregation of bacterial units on release from host cells does not complicate the interpretation of viable cell counts; (e) attention should be paid to homogeneity of the parasite population in terms of virulence and colonial or antigenic smoothness; (f) there should be an awareness that the utilization of autologous sera from animals, in different stages of the infection under study, imposes on the phagocyte cell system unknown complications arising from the variable degree of bactericidal action of the serum itself. II. RESPONSE OF PHAGOCYTIC CELLS TO INFECTION It is advisable at this point to examine the present status of work on the responses of phago- 67

2 68 SANFORD S. ELBERG [VOL. 24 cytic cells to infection by some viral and bacterial agents. Though the survey cannot be complete, the examples chosen give a cross-section of our information and the trend of present studies. In the case of viral diseases, the role of blood leucocytes, macrophages, and other cells of the reticulo-endothelial system has not been extensively studied. In cellular immunity studies with viral as well as bacterial infections, many investigators have disregarded or evaluated as unimportant to their experimental designs the nonsterile type of immunity or latent infection. Although it is experimentally proved, in the case of miyagawanella and brucella infections, that birds as well :as mammals with clinical and occasionally with anatomical microbial latency can resist massive superinfections, this host-parasite equilibrium is frequently overcome. Striking differences between the opsonic action of heated and unheated immune serum have been reported to occur when the phagocytizing cell employed was the mononuclear type but not the polymorphonuclear, in the case of miyagawanella (Meyer (68)) and vaccinia viruses (Beard and Rous (7)). In the former case, the conclusion was drawn that "the ingested viral particles were greatly, if not entirely, suppressed in their multiplication in the presence of immune serum and exudate cells secured from immune guinea pigs when the infectiousness was tested intraperitoneally on mice." Normal serum and cells derived from immune animals also destroyed the miyagawanella virus and, furthermore, destruction was accelerated in the presence of phagocytes resembling the epithelioid cell (Meyer (68)). The picture with fibroma virus seems to be different. The principal cellular contribution appears to be an enhancement of the virus-inhibiting capacity of specific immune serum without much advantage in favor of cells from immune hosts. But Ginder (32) has clearly shown the ability of the leucocyte and macrophage from fibroma-immune animals to enhance the neutralizing capacity of immune serum alone, although even in these experiments, Ginder observed occasionally that cells from normal animals also enhanced the serum phenomenon and not because of any "antibody" contained in the cells. Benedict (9) has studied the responses of rabbit and guinea pig monocytes to meningopneumonitis and psittacosis viruses, utilizing a system in which the host cell populations were carefully controlled. Several points in these experiments are of interest in the context of this review. Neither virus was cytopathogenic and the monocytes of normal animals supported growth of the viral particles; no latent or noninfectious phase was observed, in contrast to the results obtained in the chick embryo fibroblast-meningopneumonitis system. Monocytes derived from immune guinea pigs suppressed psittacosis virus proliferation. Whereas infected guinea pigs in three cases yielded suppressing-type monocytes, this suppression was only partially seen in the case of two similarly infected animals. These findings suggest great sensitivity for the monocyte test or else heterogeneity in monocyte response reflecting similar qualities in the stock animals. Comparisons between the effectiveness of monocytes derived from normal and immunized animals in the case of mycobacterial infections have yielded certain contradictory observations which apparently result from differences in experimental methods. Lurie (57, 58) made the basic observation that parasitized phagocytes from immune rabbit donors, when placed in the anterior chamber of the eye of a normal rabbit, inhibited growth of their contained tubercle bacilli, independently of the body fluids. This observation was confirmed by Suter (97, 98) and Abe (1), who showed in guinea pigs and rabbits that phagocytes from animals vaccinated with BCG suppressed the multiplication of tubercle bacilli. Suter further reported that attenuated strains (BCG and R1Rv) multiplied in monocytes from normal animals, whereas the avirulent type of strain did not. Mackaness (61) confirmed the events in the normal phagocytes described by Suter for the virulent strain but found no difference in the rates of multiplication or length of lag periods of the Branch and R1Rv strains either in immune or normal monocytes. However, examination of the data in Mackaness' paper reveals that the average number of bacilli per infected monocyte was less in the immune cells for each time interval. Part of the explanation for the diversity of results may reside in the absence in the cell culture environment of immune serum, which will be shown below to exert an important influence on the maintenance of monocyte structural integrity. Raffel (79) states that he could find no inhibition by normal macrophages of the avirulent strain H37Ra and the attenuated strain (BCG).

3 19601 CELLULAR IMMUNITY 69 In the experiment just cited, the differences may well reflect variable destruction of the host cells with subsequent release of viable bacilli. The age of the culture used to infect the monocytes could also have affected the subsequent growth rate. A culture 1 week old, for example, of strain RlRv increased 300 per cent in 8 days in the monocytes whereas a 2-week-old culture increased 60 per cent and 3-week-old cells failed to multiply. Marked differences in the growth rates and length of lag period of strains of different virulence will be observed if attention is paid not only to the age of the culture at the time monocytes are parasitized (62) but also to the cytotoxic action of the strain. If unrecognized cytotoxicity is proceeding independently in normal monocytes suspended in normal donor serum, loss of monocytes and subsequent rephagocytosis will affect the observations. The extremely difficult task of evaluating results obtained in vitro with cell cultures and relating these to events in vivo was put into clear perspective by the work of Brieger (13). For example, true tubercles did not develop in infected spleen explants cultivated in vitro although they appeared in the spleen left in vivo. Acid-fast bacilli were hard to find in spleens left in the animal for as long as 3 weeks after infection, whereas explants of spleens removed at various periods after intravenous infection became densely infiltrated after 10 days in culture, although only a few infected macrophages and no free bacilli were present in the spleens when first removed for explantation. This finding suggests that living cells in vitro promote the growth of tubercle bacilli although inhibiting it in vivo (Brieger (13)). Excellent resolution of the numerous technical difficulties in this type of experiment has been achieved by Stinebring and Kessel (96) for the Brucella abortus-monocyte system, by Howward (43) and Larsh and Shepard (54) for the Histoplasma capsulatum-monocyte system, and by Furness (30, 31) and Oakberg (75) for Salmonella typhimurium. The E. coli-monocyte system was studied by Rowley (85). An attempt to understand the important part played by the mononuclear cell in infection and immunity is greatly aided by the findings of Cavanaugh and Randall (16), who have studied the properties of Pasteurella pestis not only in the vertebrate host but also in the vector, Xenopsylla cheopis. Cavanaugh and Randall showed that the virulent bacilli from the blocked proventriculus are actually phagocytosis-sensitive and nonencapsulated, properties normally associated with avirulence in this species. Yet the bite of the blocked flea is classically the path leading to bubonic plague. The events following infection by flea bite, according to these workers, consist of regular phagocytosis by polymorphonuclear neutrophiles and presumably destruction of most of the parasites. There is, however, an occasional organism which escapes this host cell and is ingested by a mononuclear cell, which it multiplies in and eventually destroys. Cavanaugh and Randall found that populations of P. pestis emerging from monocytes are phagocytosis-resistant and encapsulated, and thereby enabled to establish the infection. Such a system should provide a "field day" for those interested in mechanisms of antigen biosynthesis, in view of the knowledge already available on the virulence antigens of P. pestis. The significance of these findings is increased when viewed in relation to the observations on the pathogenesis of Bacillus anthracis infection by respiratory introduction of spores. Ross (84) has conclusively shown by the most painstaking histopathological analysis that anthrax infection is established after an occasional monocyte has ingested a spore and transferred it to a lymph node. Only then do germination, growth, multiplication, and establishment of the infection occur Ċhance, independent action, transference, and mediation of change of bacterial properties are recurring themes in which the mononuclear cell is intimately involved. It is essential and timely that a rigorous comparison of the biological activities of mononuclear cells derived by intraperitoneal injection of starch, glycogen, caseinate, broth, oil, and other irritants into mice, guinea pigs, rabbits, and rats be made before the area of investigation becomes as confused as the Tower of Babel. Despite the difficulties involved in this type of experiment, results of a more consistent nature have been obtained in the case of brucella infections. Virulent strains of Brucella grow in monocytes derived from normal animals (rabbits, mice, rats, guinea pigs); they are inhibited in monocytes derived from infected animals (and presumably in the nonsterile phase of immunity

4 70 SANFORD S. ELBERG [VOL. 24 at the time of the experiment); and are not affected by the presence of "immune serum" in the cell culture (Elberg et al. (21); Holland and Pickett (42)). The latter workers have pointed to the bimodal distribution of inhibiting powers of normal and immune monocytes, which suggests a clonal type of distribution of the ability to respond to Brucella parasitization. They suggest that a selection of monocytes with inhibiting properties occurs under the influence of an infection, since normal monocytes also at times show an "immune" inhibiting response (12, 29, 79). Certainly this suggestion has important consequences for future experimental designs in the light of the clonal theory of antibody formation recently put forth (73). The theory of independent parasite action, however, also stimulates thought as to the explanation of monocyte behavior in terms of variants in the bacterial population and the graded abilities of such variants to multiply in a host cell population (Meynell and Stocker (67)). The specific activity of different organs must be taken into account in any analysis which relates data in vitro to conditions in vivo. For example, the carefully documented and extensive studies of McCune and Tompsett (64) and Mc- Cune et al. (65) re-emphasized the differences in response of mice to tubercle bacilli, depending on whether one's attention is focused on the events in the lung or spleen. Less bacterial activity observed in the spleen, in terms of rise and fall of bacillary populations, was accompanied by a lesser response of these bacterial populations to antimicrobial drugs than was the case in the lung. More important, however, to our concept of the validity of tissue response as a guideline to bacterial activity, was the finding that although the lungs from many of the experimental groups were indistinguishable from each other in terms of extent or composition of lesion, there were within these same organs definite population differences of a uniform and predictable nature. As McCune and Tompsett (64) point out, "once a population of tubercle bacilli attained a sufficiently high level, the resulting pulmonary lesions showed a steady progression thereafter even though the size of the population remained stable." This hint of a "critical mass" is repeated in a similar but not identical situation in anthrax infection with respect to the bacillary numbers in the blood (Smith and Keppie (92)). The phenomenon of latency provides provocative data for our understanding of the serial events in intracytoplasmic parasitism. The presence of viable tubercle bacilli in quiescent areas for long periods, the latent streptococcus that results from penicillin treatment of pharyngitis, and the presence of brucellae in animal tissues in the absence of any detectable ancillary signs are examples of the danger of overemphasizing the ability of cultured phagocytic cells to prevent the multiplication of contained bacilli as the criterion par excellence of cellular immunity (cf. Rogers and Tompsett (83); Tompsett (100)). Egestion of altered parasites, increased generation time, and spheroplast or protoplast-like development are some of the many alternatives available to phagocytes as their individual contribution to the cellular community. Perhaps equally important is maintenance of the phagocyte's own vital structure in the early stages of the parasitism. This aspect will be discussed subsequently. A. Interaction between Parasite and Phagocytic Cell The title of this section of the Symposium, "Cellular Immunity," has been interpreted in terms of very limited scope because of the extremely broad coverage of the Symposium as a whole, many of the participants of which have valid reasons for considering their own presentations as descriptive of the cellular aspects of nonspecific resistance and immunity. In actual fact one could claim that this section has been pushed to the wall and therefore one has taken this literally, for provocative reasons, to think of a part of our experimental data in terms of maintaining the monocyte wall. A second part of the experimental data has been directed to a consideration of the other wall involved, that of the parasite. The result of this orientation may well be a discussion of enzymatic "wall-to-wall carpeting" and the effect of the two systems, host and parasite, in weakening each other's wall structure, whether this be by the general class of N-acetyl- 1,4-muraminidases or by esterases directed toward bacterial lipoprotein wall components on the one hand or, on the other hand, by as yet undefined systems of the parasite which chelate essential binding materials or hydrolyze components of the host cell wall (Repaske (82); Colobert (17)).

5 19601 CELLULAR IMMUNITY 71 One concept which has been especially rewarding in an attempt to define part of the natural history of cellular immunity has been that of A. A. Miles, namely, that the events occurring in the early stages direct the outcome of the response of the host to the parasite and that a critical period ensues immediately in this interaction in which maneuverability of the response is experimentally possible, but after which this reversibility gives way to irreversible events. III. RESPONSE OF MONONUCLEAR PHAGOCYTES TO MYCOBACTERIUM TUBERCULOSIS AND BR UCELLA MELITENSIS If one examines the response of mononuclear cells, derived from the inflamed rabbit's peritoneum, to either Mycobacterium tuberculosis or Brucella melitensis, interacting in a Mackanesstype chamber, so as to follow the fate of each of the monocytes in the chamber after parasitization, one may interpret the results, described subsequently in detail, as indicating that preservation of the structural integrity of the monocyte is an active response. Thereafter, whether the parasitized bacilli multiply or not is more or less irreversible, depending upon many conditions such as initial and subsequent parasite to monocyte ratios, surface and other modes of phagocytosis. Additional mediating influences are discussed by others in this Symposium. A. Description of Methods An experimental design, familiar in some respects to that of other investigators, yet differing in certain essentials, will illustrate the means which have been used to study the mononuclear response. This summary of the methods is taken from the detailed reports (Fong et al. (26-28); Elberg et al. (21)). Monocytes from adult rabbits, both normal and BCG-vaccinated, which had been injected 5 days earlier with "Klearol" intraperitoneally, were harvested in Tyrode's solution. The cell suspension was filtered through gauze, sedimented, and the cells were redispersed in trypsin in Tyrode's solution. After 25 min at 25 C, the cells were washed 3 times with chilled Tyrode's solution. After a final centrifugation the packed cells were dispersed in fresh normal or immune rabbit serum. Cell suspensions obtained 5 days after administration of oil (arbitrarily designated as monocyte suspensions) generally contained 85 per cent or more large mononuclear cells, although other cell types such as polymorphonuclear leucocytes, small mononuclear cells, and a few fibroblasts were also present. The H37Rv strain of M. tuberculosis, recently passaged in guinea pigs and maintained on Trudeau medium, was grown for 1 week in Tweenalbumin medium and the sedimented cells were washed twice in fresh medium. They were resuspended in a small volume of medium and centrifuged to remove larger aggregates. The supernatant fluid, which largely consisted of bacilli occurring singly, was used for parasitization of monocytes after microscopic enumeration of the bacteria. The bacilli and normal or immune monocytes suspended in fresh normal or immune rabbit serum were mixed in a ratio of 10 bacilli per monocyte in most experiments. The mixtures were placed in paraffin-lined bottles, centrifuged, and refrigerated at 4 C for 1 hr. At the end of this time the supernatant fluid was removed as completely as possible, the sediment was resuspended in a small volume of medium, and the suspension was diluted to contain approximately 15 monocytes per mm.3 This diluted suspension of infected monocytes was used for culture in vitro. The standard nutrient medium for cell maintenance in these studies consisted of 40 per cent aged, Selas-filtered rabbit serum (normal or immune) in Tyrode's solution. The ph of the medium was adjusted to 7.4 by gassing with 5 per cent CO2 in air. Maintenance of monocytes in vitro was also examined in Hanks' solution, lactalbumin-yeast extract medium, tissue culture No. 199 medium (Difco), and TACPI medium (Tyrode-amino acids-cocarboxylase-p-aminobenzoic acid-insulin) as described by Trowell (101) for cultivation of lymph node cells. Immune serum was derived from rabbits given 1 to 3 intravenous injections of BCG 30 days apart and bled 4 to 5 weeks after the last injection. Only positive reactors to 0.2 ml of 1:100 Old Tuberculin (O.T.) were used. The same animals which were used as sources of normal and immune sera served as donors of normal and immune monocytes. The Mackaness chamber was assembled as follows: the bottom cover slip of the chamber was sealed into position and approximately 0.05 ml of a monocyte suspension containing 15 cells per mm3 was introduced into the space enclosed

6 72 SANFORD S. ELBERG [VOL. 24 by the small "perspex" ring. This yielded a total of about 500 to 1000 monocytes in the center well of the chamber. The top cover slip was inserted and the chamber left at room temperature for 10 to 15 min to allow settling and adherence of cells to the bottom cover slip. Sufficient medium to fill approximately two thirds of the culture chamber was introduced via one of the lateral drill holes. The entire chamber was sealed by insertion of the stainless steel pins and incubated at 37 C. The number of cells in the central well of the culture chamber was determined at the start of an experiment and at certain intervals thereafter with a phase contrast microscope at 100- fold magnification. Enumeration of cells was facilitated by a special square constructed in the ocular of the microscope. This square was subdivided into 9 smaller squares by cross hairs. The large square was correlated with horizontal and vertical line markings on a specially constructed mechanical stage whereby each marking on the stage corresponded to the area encompassed by the large square in the ocular. The entire area in which the monocytes were confined was covered in the following manner: the small "perspex" ring was brought into the field of vision of the objective, and the outer margins of the ring at its widest points in the horizontal and vertical axes were located. This served to delineate the area of examination as an imaginary square within which was located the "perspex" ring. By starting at one corner of this imaginary square and moving the stage one marking at a time in a horizontal direction, the monocytes along the entire width of the imaginary square were counted. The stage was then moved one marking in a vertical direction and the horizontal movements repeated. These successive movements were continued until the entire area of the imaginary square was included. Morphologically intact cells only were counted and no attempt was made to distinguish different cell types. Counts made by different experienced individuals did not vary more than 5 per cent, and replicate counts by the same worker seldom varied by more than 2 per cent. When it was desired to establish the proportion of infected monocytes following parasitization, a cover slip was prepared as described previously and the parasitized cell suspension was introduced into the space confined by the "perspex" ring. Following adherence of cells to the cover slip, the fluid was removed with a capillary pipette and the specimen was allowed to air-dry. The specimen was fixed with heat, passed through successive changes of xylol and alcohol, and stained by the Ziehl-Neelsen method. The proportion of infected cells after parasitization was determined by counting a total of 200 stained cells. Antisalmonella serum represented the aged, pooled sera of rabbits immunized with Salmonella rutgers (65a). For absorption of antisera, 3 successive portions of homologous absorbing antigen were employed to remove agglutinating and nonagglutinating antibody. Dialysis of anti-bcg serum was carried out in cellophane bags against normal serum medium for 7 days with daily changes of medium. Normal serum medium was used rather than Tyrode's solution alone, since dialysis of serum against the balanced salt solution yielded sera which failed to maintain cultures of monocytes. Serum globulin was precipitated twice, redissolved in Tyrode's solution, and dialyzed against Tyrode's solution for 4 to 7 days. The globulin preparation was not utilized until preliminary tests had shown that the preparation was no longer toxic for cultures of rabbit monocytes. Before use, enough Tyrode's solution was added to the dialyzed globulin preparation to restore its volume to that of the original serum used in its preparation. In experiments designed to test the ability of globulin to protect immune monocytes against virulent tubercle bacilli, the solution of normal or immune globulin was used in place of Tyrode's solution. Enough normal serum was added to yield a medium consisting of 40 per cent serum in globulin solution. In experiments involving determination of viable bacteria in monocyte cultures, the infected monocytes were washed 3 times with Tyrode's solution to remove extracellular bacteria and then suspended in maintenance medium. Approximately 20,000 to 40,000 infected cells were introduced into Carrel flasks. The number of bacteria present in the flasks was determined at the start of the experiment and after varying periods of incubation at 37 C, utilizing the contents of 3 Carrel flasks as a unit. Lysis of the monocytes was achieved by adding saponin to the flasks to a final concentration of 2 per cent.

7 19601 CELLULAR IMMUNITY 73 This concentration was adequate to lyse monocytes and effect release of intracellular bacteria without causing inactivation of the bacteria. The entire content of the Carrel flask was subjected to repeated pipetting to effect disaggregation of bacterial clumps as well as to facilitate lysis of cells. The sample was then examined microscopically for evidence of aggregation. Most samples treated in this manner consisted of single bacterial cells. Dilutions of the sample were made in Tween-albumin liquid medium and each dilution was added in triplicate to the surface of each of 4 glycerol-blood agar plates by the drop method. The plates were incubated for 12 to 14 days at 37 C, and those plates which contained suitable numbers of colonies were counted and the average of all 4 plates calculated. Parallel cultures were generally set up in the Mackaness-type culture chambers to allow correlation of cellular degeneration with bacterial multiplication. Attention to certain details revealed the means whereby difficulties encountered by earlier workers could be overcome. For example, degeneration of the cell cultures during the first 24 hr of incubation, when uncontrolled and unrecognized, invalidates estimates of monocyte response to bacterial growth; exposure of the cell suspensions from the exudate to 0.25 per cent tryp- TABLE 1 Effect of trypsin on enhancement of monocyte survival time Avg Per Cent Degeneration Time of No. of after Incubationb Trypsinization Monocyte at 25 C Donorsa 24 hr 48 hr 72 hr miss a Indicates number of rabbits used as sources of monocytes; the monocytes of each rabbit were cultivated separately and in duplicate or triplicate. brefers to average per cent degeneration (from initial count) of all cultures in a given series; thus, if 4 rabbits were used and each cell suspension was prepared in duplicate or triplicate, the figures would represent the average per cent degeneration of 8 to 12 cultures (from Fong et al. (26)). sin in Tyrode's solution provided a selective process which at least rendered the cell suspension amenable to the cell culture environment. The influence of trypsin is clearly indicated in table 1. Nontrypsinized cells possessed less survival potential than similar cells preliminarily subjected to the enzyme. The different capacities of these 2 categories of cells to survive during cultivation in vitro were most manifest in the first 24 hr and somewhat less apparent after 48 and 72 hr of incubation. Trypsinization of cells for 20 to 35 min at room temperature permitted better maintenance of cells than that afforded by a 5-min period of exposure. Data in table 2 show that trypsinized cells cultivated in 40 per cent normal rabbit serum in Tyrode's solution exhibited only minor changes in cell population over a period of approximately 72 hr. Since nontrypsinized cells cultivated in 40 per cent normal rabbit serum in Tyrode's solution exhibited gross degeneration during the first 24 hr of incubation, the possibility existed that the nutrient medium used might have been inadequate; moreover, since others have employed Hanks' solution in their studies, it seemed advisable to compare the relative survival of cells grown in rabbit serum in the presence of these 2 balanced salt solutions. In addition, 3 other nutrient media, TC No. 199, lactalbumin-yeast, and the TACPI medium of Trowell (101) were studied. In the experiments shown in table 2, TC, lactalbumin-yeast, and TACPI media failed to yield adequate maintenance of cells. The amount of degeneration in these media, as opposed to that occurring in 40 per cent normal rabbit serum in Tyrode's solution, was approximately 12 to 18 times greater. In view of these results, no counts were made after 24 hr except for one experiment with lactalbumin-yeast medium in which a slight further degeneration of cells was noted after incubation for 48 hr. Comparison of the results obtained in 40 per cent normal rabbit serum in Hanks' solution with those in Tyrode's solution indicated a better survival of cells in the latter. The advantage of Tyrode's solution over Hanks' solution was most noticeable at 24 hr but a significant difference was still apparent at the 48-hr interval. The average per cent degeneration of cells in Ty-

8 74 SANFORD S. ELBERG [VO L. 24 TABLE 2 Survival of monocytes in different media Composition of Mediuma Avg Per Cent Degeneration after Incubation Ratio of Degeneration in Each Medium to Degeneration in 40% RS-Tyrode after Incubation 24 hr 48 hr 72 hr 24 hr 48 hr 72 hr 40% RS-Hanks 15.3 (12)b 18.6 (6) 6.3c (3) 6.9:1 3.9:1 2.5:1 40% RS-TC (no. 199) 25.7 (3) _d 11.7:1-40% RS-LY 27.7 (6) 43.5 (2) 12.6:1 9.1:1-35% RS-TACPI 40.8 (8) :1-40% RS-Tyrode 2.2 (67) 4.8 (37) 2.5c (10) 1:1 1:1 1:1 20% RS-Tyrode 35.0 (3) : % RS-Trode 45.0 (3) 20.4:1 a RS refers to normal rabbit serum; LY indicates lactalbumin-veast medium; TACPI refers to Trowell's medium (101). I The numbers in parentheses indicate the total number of culture chambers counted in the course of several experiments. c A number of the cultures counted showed an increase in cell population at 72 hr; these chambers were arbitrarily recorded as 0 per cent degeneration in calculating the average per cent degeneration (from Fong et al. (26)). d Samples not counted. rode's solution throughout the experimental period of 72 hr was within the range of the limits of error of the method of enumeration; hence, the degree of degeneration recorded may not necessarily represent loss of cells during this period. (An elegant analysis of the sources of error inherent in techniques attempting to follow the fate of ingested bacilli has been presented recently by Hanks (36).) Studies on the influence of serum concentration upon ceil survival, carried out in a limited number of chambers, indicated a more beneficial effect of high serum concentration. The amount of cellular degeneration with 5 to 20 per cent serum was 15 to 20 times greater than that with 40 per cent serum. B. Responses of Monocytes Parasitized with M. tuberculosis Although cultivation of uninfected, trypsinized cells in 40 per cent normal rabbit serum in Tyrode's solution provided a consistent procedure for maintenance of cells in vitro, cells which had been parasitized with the virulent H37Rv strain of M. tuberculosis and cultivated under similar conditions exhibited various degrees of degeneration. This degeneration was reflected in a disappearance of monocytes and some accumulation of granular debris. "Shadow" forms (nonrefractile cells with indefinite boundaries and absence of internal structures) invariably preceded a decreased total monocyte count. The degenerative effect of virulent tubercle bacilli on monocytes is shown in table 3. The experiments recorded in this table were made with the trypsinized cells of normal rabbits and rabbits rendered tuberculin-positive by 1 to 3 injections of BCG. It may be seen that uninfected normal or immune cells (those obtained from rabbits injected with BCG) cultivated in the presence of normal rabbit serum showed little change in cell population during the experimental period. Very marked degeneration of both normal and immune cells occurred, however, upon parasitization of cells with virulent tubercle bacilli. The average per cent degeneration of normal or immune cells after exposure to H37Rv was between 26 to 35 per cent after 24 hr of incubation. Immune cells derived from animals receiving 3 injections of BCG were no more resistant to the degenerative action of tubercle bacilli than were cells from animals given 1 or 2 injections of BCG. It is of interest that the average per cent degeneration of parasitized cells was less at 72 hr than at either of the 2 preceding intervals. One inference which may be drawn is that some degree of cell proliferation occurred despite the presence

9 19601 CELLULAR IMMUNITY 75 of virulent tubercle bacilli. The number of experiments shown in table 3, in which parasitized cells exhibited an increased cell count, is small, but various other experiments (not shown) have tended to confirm this observation. The serum of BCG-vaccinated rabbits exerted a noticeably favorable effect upon survival of cells following their parasitization with the virulent H37Rv strain. Infected normal cells cultivated in normal serum exhibited marked degrees of cell degeneration in 24 hr; similar cells maintained in immune serum showed no significant change in cell population during the same interval. The protection of normal cells by immune serum, however, seemed to be merely a delaying action, for the degree of cell degeneration in immune serum after 48 hr approximated that of normal cells in normal serum. A more persistent inhibition of cellular degeneration induced by virulent tubercle bacilli was observed when immune cells were cultivated in immune serum. This prolonged protection of immune cells by immune serum is of particular interest; although the results presented in table 3, when taken alone, indicated no significant difference in the behavior of normal or immune cells, the results shown in table 4 imply a basic dissimilarity in immune and normal cells. In order to define some of the parameters of the degenerating action of the tubercle bacillus, the activity of 3 strains of different virulence as Treatment of Cellsa well as the effect of some bacterial products were studied. Monocytes of tuberculin-negative rabbits infected with H37Rv, BCG, or H37Ra and cultivated in the presence of normal serum revealed that the 3 strains of tubercle bacilli differed in their ability to effect degeneration of monocytes. Marked degeneration of monocytes infected with the virulent H37Rv strain was apparent after only 10 hr of incubation; by 48 hr, almost one half of the original cell population had been lost. Induction of monocytic degeneration by the BCG strain of tubercle bacillus was apparent at the 24- hr interval, but the degree of degeneration was slight. The amount of monocytic degeneration doubled during the next 24 hr of incubation, but it was still considerably less than that caused by the H37Rv strain. Infection of monocytes with the avirulent H37Ra strain resulted in no apparent effect upon the cells during the entire period of observation. Induction of cell degeneration by virulent tubercle bacilli bore little relationship to the cytotoxicity of tuberculin. Experiments tested the ability of normal immune cells suspended in normal or immune serum to survive treatment with H37Rv, O.T., P.P.D. (purified protein derivative), and a filtrate of a week-old culture of H37Rv in Tween-albumin medium. Comparison of the results following exposure of monocytes to H37Rv with those effected by O.T. revealed a number of interesting differences. TABLE 3 Survival of monocytes after infection with tubercle bacilli Avg Per Cent Degeneration after Incubation 24 hr 48 hr 72 hr Uninfected normal cells (13) b 5.1 (11) 5.5 (2) Infected normal cells (15) 35.2 (13) 12.5 (2) Uninfected immune cells (10) 2.0 (9) 0 (2) Infected immune cells (13) 36.8 (13) 10.3 (3) Uninfected immune cellsd... 0 (4) 0 (2) Infected immune cellsd (6) 39.0 (3) - The proportion of infected monocytes in the parasitized cell suspensions of tables 3 and 4 varied in different experiments, but was generally between 20 and 30 per cent. a All cultures were made in 40 per cent normal rabbit serum in Tyrode's solution. b Numbers in parentheses refer to total numbers of cultures counted. c Refers to cells from rabbits given 1 or 2 injections of BCG. d Refers to cells from rabbits given 3 injections of BCG. (From Fong et al. (26).)

10 76 SANFORD S. ELBERG [VOL. 24 TABLE 4 Effect of Serum of BCG-vaccinated animals upon survival of infected monocyte cultures Nature of Sample Testeda Avg Per Cent Degeneration after Incubation Ratio of Degeneration in Each Sample to Degeneration of Immune Cells in Immune Serumb after Incubation 24 hr 48 hr 72 hr 24 hr 48 br 72 hr Infected normal cells in normal serum (10) 34.1 (9) 12.5 (2) 9.9:1 6.2:1 4.1:1 Infected normal cells in immune serum (11) 26.4 (11) 5.5 (2) 1.5:1 4.8:1 1.8:1 Infected immune cells in normal serum (14) 33.2 (14) 10.3 (3) 9.4:1 6.0:1 3.4:1 Infected immune cells in immune serum (13) 5.5 (13) 3.0 (2) 1:1 1:1 1:1 a Control (uninfected) cultures not shown since results on controls resembled data given in tables 2 and 3. b The terms "immune cells" and "immune serum" are used qualifiedly (to designate a difference between normal and BCG-vaccinated animals). (From Fong et al. (26).) Normal monocytes cultivated in the presence of normal serum were affected by exposure to both agents. Contrasted with the results obtained with H37Rv, degeneration of monocytes by O.T. appeared sooner (4 hr), was less pronounced (10 per cent as against 20 to 47 per cent for H37Rv), and showed no progression with time; these same differences between the effects of O.T. and H37Rv may be observed both for normal cells cultivated in the presence of immune serum and for immune cells in normal serum. It is apparent, however, that immune cells were more susceptible to the effects of O.T. than were normal cells (24 per cent as against 10 per cent for normal cells). The most significant difference between the effects of O.T. and H37Rv was exhibited by immune cells cultivated in the presence of immune serum; under these conditions, the immune cell was protected against the degenerative effects of H37Rv (the average per cent degeneration of 5 shown for the 44-hr interval was within the limits of error in the method of counting monocytes), but not against the cytotoxic action of O.T. (as evidenced by the early occurrence and the high percentage of degenerated cells). It may also be significant that 3 of the 4 cultures treated with O.T. were stimulated to proliferation earlier than any of the cultures infected with H37Rv. The effects of P.P.D. upon normal and immune monocytes cultivated in the presence of normal or immune serum are shown for the 20-hr period. A comparison of these results with those recorded for H37Rv and O.T. at the same time interval discloses a similarity in action of P.P.D. and O.T. Both substances caused marked degeneration of immune cells, and the action of both, uilike that of H37Rv, was unaffected by the type of serum used. Since P.P.D. exhibited little cytotoxic effect on normal cells, it seems that the low level of cell degeneration evoked by O.T. in the cultures of normal monocytes may be referable to certain impurities in the latter reagent. When the results of tests employing a culture filtrate of H37Rv are compared with those given by the viable, intact bacterial cell, a definite parallelism in action becomes apparent. Both agents caused marked degeneration of normal and immune monocytes when these were cultivated in the presence of normal serum. Similarly, cultivation of immune cells in an immune serum resulted in protection of these monocytes against the degenerative effects of both agents. Although the potency of culture filtrates may vary from one experiment to another, a basic similarity in action of H37Rv and its culture filtrate is strongly suggested. These findings indicate that the effects of H37Rv upon monocytes differed from that of O.T. and P.P.D. but were indistinguishable from that of a culture filtrate of the virulent bacilli. 1. Specificity of serum components. The non-

11 1960] CELLULAR IMMUNITY 77 specificity of the serum components of rabbits immunized with BCG was established in a number of similar experiments on the nature of this serum activity. The results shown in table 6 compared the survival of normal and immune monocytes infected with virulent H37Rv and cultivated in the presence of normal serum, homologous antiserum, or heterologous antiserum. Examination of the data in the upper half of table 6 reveals a protection of normal monocytes by antiserum from rabbits immunized with BCG, Salmonella rutgers, or crystalline egg albumin. This was evidenced by an absence of monocyte degeneration as reflected in a constant average per cent degeneration over a 24-hr period. No such protection of normal monocytes was evident with normal rabbit serum (average per cent degeneration of 38 over the 24-hr period). As might have been expected from the results obtained for normal monocytes, a similar protection was afforded immune cells by either homologous or heterologous antisera but not by normal serum (no degeneration in immune sera as compared with 11 per cent in normal serum after 24 hr). There was very little difference between the behavior of normal and immune monocytes over the 24-hr period of observation in this experiment. Various other experiments have demonstrated that immune (anti-bcg) serum only delayed the onset of degeneration in infected normal mono- TABLE 5 Effect of HS7Rv, O.T., P.P.D., and culture filtrate upon survival of monocytes in vitro Type of Monocyte& Monocytes Exposed to Type of Serumc Avg Per Cent Degenerationd after Incubation 4 hr 10 hr 20 hr 44 hr Normal Tween-albumin Normal X Normal Tween-albumin Immune X Immune Tween-albumin Normal X Immune Tween-albumin Immune X Normal H37Rv Normal Normal H37Rv Immune Immune H37Rv Normal Immune H37Rv Immune Normal O.T. Normal X Normal O.T. Immune X Immune O.T. Normal Immune O.T. Immune X Normal P.P.D. Normal _ 0 Normal P.P.D. Immune 1- Immune P.P.D. Normal Immune P.P.D. Immune _ 27 Normal Culture filtrate Normal 31 Normal Culture filtrate Immune - 7 Immune Culture filtrate Normal - 32 Immune Culture filtrate Immune - 0 anormal designates monocytes of tuberculin-negative rabbits; immune refers to monocytes of rabbits rendered tuberculin-positive by BCG immunization. b Ratio of H37Rv to monocyte was 10:1; O.T. (4 times concentrated) was diluted 1:120 before use; the solution of P.P.D. contained 50 Ag per ml.; the culture filtrate was from a week-old culture of H37Rv in Tween-albumin medium. c Refers to the type of serum (normal or anti-bcg) in which monocytes were suspended at time of parasitization or exposure to O.T., P.P.D., or culture filtrate. The same type of serum was used for subsequent cultivation of monocytes. d Refers to average per cent degeneration (from original count of approximately 500 to 1000 monocytes) of 2 to 3 cultures. X indicates increase in number of monocytes; - indicates sample not counted. (From Fong et al. (27).)

12 78 SANFORD S. ELBERG [VOL. 24 TABLE 6 Effect of ivarious sera upon survival of monocytes infected with H37Rv Avg Per Cent Type oserof y Per Cent Degenerationd Monocytea' Type of Srm of Infected Cells0 10 hr 24 hr Normal Normal Normal Anti-BCG Normal Antisalmonella Normal Anti-ovalbumin _e - 0 Immune Normal Immune Anti-BCG Immune Antisalmonella Immune Anti-ovalbumin Control (uninfected) monocytes suspended in the different types of serum showed no degeneration. a Normal refers to monocytes from tuberculinnegative rabbits; immune refers to monocytes from rabbits vaccinated with BCG. I Normal designates serum of tuberculin-negative rabbit; anti-bcg refers to serum of rabbit immunized with BCG; antisalmonella refers to serum of rabbit immunized with Salmonella rutgers; anti-ovalbumin refers to serum of rabbit immunized with a solution of crystalline egg albumin. c Obtained by examination of 22)0 stained cells. d Average per cent degeneration (from initial count) of 2 to 3 cultures. e Sample not counted. (From Fong et al. (27).) cyte cultures. Whether or not anti-ovalbumin and antisalmonella sera behaved identically in this respect was not established. The participation of a serum component of the cell culture system (Shepard (88-90)) was revealed most clearly by the ability of some sera to support phagocytosis of staphylococci in inverse relation to their ability to support growth of the microorganism within the phagocyte. Absorption studies on such sera implicated an antibody-like substance with no specificity as far as bacterial activity was concerned. The results presented in table 6 reveal an element of nonspecificity in the action of serum components, for sera against S. rutgers and ovalbumin proved equally effective in protecting immune monocytes against the necrotizing action of virulent tubercle bacilli. 2. Specificity of "immune" monocyte. The results of experiments designed to test the specificity of monocytes in resistance against tubercle bacilli are presented in tables 7 and 8. The results shown in the upper half of table 7 indicate a similarity in the behavior of normal monocytes and monocytes derived from an animal immunized with S. rutgers, i.e., both showed early (10 hr) and marked degeneration (30 to 35 per cent in 10 hr and 46 to 50 per cent in 48 hr) when they were infected with tubercle bacilli and cultivated in normal serum. Moreover, the degeneration of both types of monocytes was delayed (a lower level of monocytic degeneration in 24 hr as opposed to the marked degeneration usually encountered when normal serum was used) but not suppressed by immune (anti-bcg) serum. It is in connection with cultivation in immune serum that the monocytes of the normal animal and the TABLE 7 Specificity of?uonocytes in resistance to infection with H37Rv Source of Monocytea Normal Normal Immune Immune Salmonella Salmonella Normal Immune Salmonella Type of Serumb Normal Anti-BCG Normal Anti-BCG Normal Anti-BCG Normal Anti-BCG Antisalmonella Per Cent Infected Monocytesc Avg Per Cent Degeneration after Incubationd 10 hr 24 hr48 hr The different types of control (uninfected monocytes suspended in the various sera) showed no change in cell numbers during the experimental period. a Monocytes derived from normal (tuberculinnegative) and immune (tuberculin-positive) rabbits, and from rabbits immunized against Salmronella rutgers, respectively. I Normal refers to serum of tuberculin-negative rabbits; anti-bcg refers to serum of rabbits immunized with BCG; antisalmonella refers to serum of rabbits immunized with S. rutgers. c Obtained by examination of 200 stained cells. Ratio of H37Rv to monocyte was 10:1. d Average per cent degeneration count) of 2 to 3 cultures. esample not counted. (From Fong et al. (27).) (from initial

13 19601 CELLULAR IMMUNITY 79 TABLE 8 Demonstration of cross immunity between tuberculosis and brucellosis at the cellular level Source of Cells Source of Serum Cells Exposed to Avg Per Cent Degeneration after Incubation 24 hr 48 hr TB. immunea Normaib H37Rv TB. immunea TB. immune H37Rv 5 10 TB. immunea Br. immune H37Rv 1 25 TB. immunea Normal Br. mel TB. immunea TB. immune Br. mel TB. immunea Br. immune Br. mel Br. immunec Normal H37Rv Br. immunec TB. immune H37Rv 1 2 Br. immunec Br. immune H37Rv 2 27 Br. immunee Normal Br. mel Br. immunec TB. immune Br. mel Br. immunee Br. immune Br. mel TB. immune Normal Dubos medium - TB. immune TB. immune Dubos medium 0 5 TB. immune Br. immune Dubos medium 0 7 Br. immune Normal Dubos medium 0 0 Br. immune TB. immune Dubos medium 0 1 Br. immune Br. immune Dubos medium 1 2 a "TB. immune" indicates that the serum or cells were obtained from rabbits immunized with BCG (tuberculin-positive). b "Normal" indicates that the serum or cells were obtained from normal rabbits which were tuberculin-negative and had no titer to the standard Brucella-agglutinating antigen kindly supplied by Dr. C. Mingle, Agriculture Research Service, U. S. Department of Agriculture. c "Br. immune" indicates that the serum or cells were obtained from rabbits immunized with a nondependent mutant from a streptomycin-dependent mutant strain of Brucella melitensis, Rev Is (as distinguished from merely being infected animals as other investigators have mistakenly confused the terms and methodology). (From Elberg et al. (21).) animal immunized with S. rutgers differed sharply from the homologously immune monocyte (derived from an animal immunized with BCG); the immune monocyte maintained in the presence of immune serum was completely refractory to the degenerative effects of virulent tubercle bacilli. It might be reasoned that the monocytes of an animal immunized with S. rutgers would exhibit their full potentialities only when suspended in homologous antiserum. This premise was tested as shown in the lower part of table 7. It is evident that after infection with H37Rv, cultivation of the monocytes of an animal immunized with S. rutgers in homologous antiserum failed to confer on such cells a state of resistance equivalent to that of the homologously immune monocyte. The behavior of the monocyte of the salmonella-immunized rabbit in antisalmonella serum resembled its behavior when cultivated in the presence of heterologous (anti-bcg) antiseruan (a delayed degeneration which appeared at the 24-hr period and not at the 10-hr interval). The data in table 8 show that immunization with BCG and Brucella melitensis vaccine conferred parallel protection on the monocyte against both homologous and heterologous parasitization. The data also suggest that the immunity conferred by the brucella vaccine was slightly more effective against both homologous and heterologous challenges than that conferred by BCG under the conditions of the experiment. However, the effects of both vaccines on the resistance of the monocyte to destruction by either organism is clearly indicated. 3. Some additional properties of serum component. Since either homologous or heterologous

14 80 SANFORD S. ELBERG [VOL. 24i TABLE 9 Survival of infected immune monocytes in the presence of absorbed immune sera Per Cent Per Cent Degenera- Type of Monocyte Serum Used in Cultivation of Monocytes Infected tiona (48 hr after Monocytes incubation) Immuneb Normal rabbit serum Immuneb Anti-BCG serum 14 1 Immuneb Absorbed anti-bcg serumc 14 3 Immuneb Antisalmonella serum 14 2 Immuneb Absorbed antisalmonella serumd 14 0 a Represents average per cent degeneration (from initial count) in 2 replicate cultures. b Monocytes obtained from tuberculin-positive rabbits immunized with the BCG strain 30 or more days previously. The immune monocytes were parasitized in the presence of the same serum as that subsequently used for their cultivation. c Absorbed with washed, whole, heat-killed H37Rv strain. d Absorbed with washed, whole, formalin-killed Salmonella rutgers. (From Fong et al. (28).) TABLE 10 Survival of infected immune monocytes in immune globulin medium Per Cent Degenera- Type of Medium Usedlin-Cultiva- Infected tion after Monocyte tion of Monocytes Monocytes Incubatio 24 hr 48 hr Immuneb Normal serum Immune Immune serum Immune Normal globulin Immune Immune globulinc arepresents average per cent degeneration in 2 replicate cultures. b Immune monocytes were parasitized in the presence of the same medium as that subsequently used for their cultivation. c These media contain 40 per cent normal rabbit serum in globulin solution. (From Fong et al. (28).) immune sera was effective, the action of the serum factor appeared to be nonspecific. Results obtained with absorbed sera and the globulin fraction of immune serum (table 9) confirm the earlier observations concerning the protective effects of homologous and heterologous immune sera. Data shown in table 10 indicate that there was no close association of protective serum factor and antibody globulin. There was some delay in the onset of degeneration in infected monocyte cultures in the presence of immune globulin; the less than 5 per cent degeneration of infected immune monocytes in immune globulin during the first 24 hr of monocyte-bacterium interaction compared favorably with control cultures of infected immune monocytes cultivated in immune serum. Immune globulin, however, permitted an average degeneration of 30 per cent in 48 hr. Although infected cells cultivated in immune serum showed a slight loss of cells after 48 hr (9 per cent), there can be little doubt that whole immune serum proved much more effective in protecting immune cells against virulent tubercle bacilli than immune globulin. It may be noted that the percentage of infected monocytes was fairly high (35 per cent), which might account for the slight degeneration of immune monocytes cultivated in immune serum; in other experiments in which the percentage of infected monocytes was lower (15 to 20 per cent), degeneration occurred in cells which were cultivated in immune globulin but not in whole immune serum. The delayed degeneration of infected immune monocytes in immune globulin was not apparent when the infected immune cells were cultivated in a normal globulin medium as evidenced by the 32 per cent degeneration of these cells after 24 hr of monocyte-bacterium interaction. The relation of the serum factor involved in these studies to the phagocytosis-promoting factors previously described (Tullis and Surgenor (102); Pollack and Victor (77)) has not been examined. Heating of immune serum at 60 to 70 C failed to destroy the protective serum factor. Dialysis of immune serum did not remove the protective factor, for infected immune monocytes main-

15 1960] CELLULAR IMMUNITY 81 tained in a nutrient medium consisting of dialyzed immune serum and Tyrode's solution likewise showed less than 5 per cent loss. Intradermal injection of rabbits with a viable BCG strain or a heat-killed BCG strain of tubercle bacillus stimulated the appearance of a protective factor in their sera within 5 days. This was evidenced in the average per cent degeneration of monocytes; infected immune cells cultivated in each of the 5-day sera exhibited less than 5 per cent degeneration at both 24 and 48 hr after infection. The 5-day monocytes of animals injected with a viable strain of BCG possessed a level of cellular resistance which protected them against the virulent H37Rv strain. In contrast to the less than 5 per cent degeneration of these infected monocytes, it may be observed in table 11 that 31 per cent of the monocytes derived from animals injected with heat-killed BCG had been destroyed 48 hr following infection. It is therefore suggested that evocation of the protective serum factor can be achieved with relative facility and that its appearance is not necessarily correlated with the existence of a high level of cellular resistance. It seems, moreover, that evolution of tuberculin hypersensitivity, as manifested by a positive skin reaction, was not an essential condition for either development of cellular resistance or production of protective serum factor. 4. Effect of immune serum. The resistance exhibited against virulent tubercle bacilli by immune cells in the presence of immune serum may involve: (a) a direct action of immune serum upon virulent tubercle bacilli with consequent modification of their behavior toward the cell, (b) an indirect cellular alteration permitting survival of bacteria in cells without destruction of the cells, and (c) modification of the bacterial cells resulting from the combined action of serum and cell. Two consequences of direct action of serum upon bacteria may be modification of the bacterial potential for induction of monocytic degeneration and for proliferation within monocytes. When normal or immune serum-treated bacteria were used to parasitize immune monocytes, which were then washed twice in Tyrode's solution, resuspended in normal serum medium to contain 15 cells per mm3, and cultured in a Mackaness-type culture chamber, little change in the ability of the bacteria to cause monocytic degeneration was observed. Approximately 35 and 37 per cent, respectively, of the immune monocytes infected with either normal or immune serum-treated bacteria and cultivated in normal serum medium were destroyed during a 48-hr period. Downloaded from TABLE 11 Relationship of protective serum factor to cellular resistance and tuberculin sensitivity Degeneratioflc after Tuberculin b ~~~~~~~~~~~Infected Incubation Reactiona ~~~~~~~~~~~~~~cytes 24 hr 48 hr % % % Immunizing Agent Amount Injected Reactionn Test Systemb Mono- at: Live BCG 2.0 X 107 Negative 5-day serum + immune cells + H37Rv Killed BCGd 40.0 X 107 Negative 5-day serum + immune cells + H37Rv Live BCG 2.0 X107 Negative 5-day cells + immune serum + H37Rv Killed BCGd 40.0 X 107 Negative 5-day cells + immune serum + H37Rv Live BCG 2.0 X 107 Positive Immune cells + normal serum + H37Rv Live BCG 2.0 X 107 Positive Immune cells + immune serum + H37Rv atuberculin reaction at time of removal of serum and monocytes. b The 5-day serum was incorporated into the nutrient medium used for cultivation of monocytes; the immune cells and immune sera were obtained from tuberculin-positive rabbits injected with live BCG 30 days earlier. c Represents average per cent degeneration (from initial count of approximately 500 to 1000 monocytes per culture chamber) in 2 to 3 replicate cultures. d Killed by heating at 60 C for 30 min. (From Fong et al. (28).) on March 17, 2019 by guest

16 82 SANFORD S. ELBERG [VOL. 24 When immune monocytes were infected and maintained in normal serum medium, the increase in bacterial concentrations 48 hr later was 6-fold for bacteria previously exposed to normal serum and 5.7-fold for bacteria treated with immune serum. The increase in bacterial concentration in immune monocytes cultivated in immune serum medium, in which there was no degeneration of cells, was 5.1-fold. Thus, increase in bacterial cell numbers within infected monocytes does not necessarily result in destruction of the monocytes. A consequence of degeneration of infected immune monocytes in normal serum medium is accumulation of extracellular bacteria. The fact that the bacterial increase under these circumstances was nearly identical with that of monocyte cultures in immune serum (in which bacterial multiplication would primarily be intracellular since there was no degeneration of cells) suggests that the bacterial increase in these immune monocyte cultures was mainly intracellular. Inoculation of either normal serum medium or immune serum medium (without monocytes) with 1 X 105 bacteria per ml did not result in any increase in bacterial count after 48 hr of incubation, so that the medium's ability to support an increase in bacterial population does not approximate that which takes place when monocytes are present. These results indicated an inability on the part of immune serum to affect directly either the capacity of virulent tubercle bacilli to induce degeneration of monocytes or their ability to multiply intracellularly. To test whether immune serum increases mononuclear cell resistance to necrotization by virulent bacilli, monocytes (normal or immune) were cultivated in the presence of serum (normal or immune) for a period of 24 hr, then treated in the same manner as monocytes derived directly from animals (i.e., trypsinization, washing, and resuspension in the desired serum), and infected with virulent tubercle bacilli. Cultivation of normal monocytes in immune serum failed to render these cells resistant to virulent tubercle bacilli; the per cent degeneration of these monocytes after 48 hr of incubation in the presence of immune serum following infection was 33, a value similar to that observed for normal monocytes (28 per cent) previously cultivated in normal serum medium and maintained in immune serum medium after infection. The per cent of infected monocytes in these 2 test systems was not markedly different and could not reasonably account for the absence of resistance in the test system consisting of normal monocytes previously cultivated in immune serum. When immune monocytes were cultivated in normal serum medium for 24 hr, infected with virulent tubercle bacilli and then transplanted to immune serum medium, there was little degeneration (3 per cent) after 48 hr of incubation; the behavior of these cells resembled that of immune monocytes grown for 24 hr in immune serum medium before infection. Prior cultivation of normal monocytes in immune serum medium or of immune monocytes in normal serum medium failed to alter the inherent properties of these cells with respect to their resistance to necrotization by virulent tubercle bacilli. 5. Time needed for serum component's action. Since immune serum failed to alter the capacity of virulent tubercle bacilli to induce cellular degeneration and to proliferate intracellularly, and failed to modify the susceptibility of cells, it was necessary to determine the stage of tubercle bacillus-monocyte interaction at which participation of serum factors became critical. The two stages of bacilli-monocyte interaction studied were parasitization (carried out at 4 C) and postparasitization (incubation at 37 C). For the parasitization studies, immune monocytes were infected with virulent tubercle bacilli in the presence of fresh normal serum, fresh immune serum, or heated (56 C for 30 min) immune serum. In the postparasitization experiments, the infected immune monocytes were cultivated in the presence of aged normal serum or aged immune serum. The data shown in the first two rows of table 11 reveal that parasitization of immune monocytes in normal serum resulted in degeneration of the monocytes even though the infected immune monocytes were subsequently cultivated in the presence of immune serum medium. The necessity for serum factors during the postparasitization period of infected immune monocytes is shown in the central portion of table 11. When immune monocytes were infected in the presence of fresh immune serum and cultivated in the presence of normal serum medium, degeneration occurred both at 24 and 48 hr (25 and 32 per cent, respectively) as contrasted with the

17 1960] CELLULAR IMMUNITY 83 less than 5 per cent degeneration occurring when similarly infected monocytes were cultivated in the presence of immune serum medium. Certain reservations about the degree of specificity of the serum component are still to be answered. The need for the presence of the serum component throughout the entire period of parasitization and subsequent culture of the monocytes suggests that the sera may actually be "cross-reactive," as glycogen is cross-reactive with certain pneumococcal polysaccharides. Rather large amounts of glycogen are needed to force the reaction with antipneumococcal sera to precipitation and there is a distinct probability that the cross-reaction would have been missed if smaller amounts of glycogen had beem employed in the quantitative precipitin test (M. Heidelberger, personal communication). The active material may exist in the fast-moving electrophoretic component of the serum. 6. Modification of degenerative effect of tubercle bacilli by combined action of immune cells and immune serum. The third hypothesis, that modification of the bacterial cell, if it occurred at all, would most likely result from the combined ac - tion of immune cells and immune serum, was tested by studying the ability of intracellularly passaged bacilli to induce monocytic degeneration. Six- to 7-day-old cultures of the H37Rv strain of tubercle bacillus were derived from (a) Tweenalbumin medium, (b) infected normal cells cultivated in normal serum medium, and (c) infected immune cells cultivated in immune serum medium. The intracellularly passaged bacteria were released by addition of saponin, concentrated by centrifugation, washed several times in Tweenalbumin medium and finally resuspended in a small volume of the medium. The various bacterial suspensions were used to infect monocytes. Infection of normal monocytes with the various bacterial suspensions followed by cultivation of the infected monocytes in immune serum medium revealed a difference in the capacity of these bacilli to induce degeneration of monocytes. Examination of table 12 (rows 4 to 6) shows that bacteria grown in vitro and bacteria passaged in a normal system (normal cells and normal serum) caused considerable degeneration of monocytes (42 and 37 per cent, respectively, after 48 hr of incubation), whereas bacteria passaged in an immune system (immune monocytes and immune serum) exhibited a decreased potential for destruction of monocytes (8 per cent after 48 hr of incubation) after a period of intracellular existence. Although the 8 per cent loss in cell population in this experiment suggested partial retention of the bacterial capacity for destruction of monocytes, other experiments have demonstrated that the bacteria recovered from an immune TABLE 12 Capacity of intraceltularly passaged bacilli to cause monocytic degeneration Source Bacilli for Infection of Cultivation Serum for Per Cent Per Cent Type TypeofMonocytes; Sourceofgacil~fol: Mooye Infected netdmncts Monocytes \Monocytes Infected Degeneration" after 48 hr Incubation Normal Normal systemb Normal Normal Immune systemb Normal Normal Tween-albumin Normal Normal Normal system Immune Normal Immune system Immune 16 8 Normal Tween-albumin Immune Immune Normal system Normal Immune Immune system Normal Immune Tween-albumin Normal Immune Normal system Immune 15 3 Immune Immune system Immune 17 2 Immune Tween-albumin Immune 19 0 arepresents average per cent degeneration in 2 or more culture chambers. bnormal system indicates bacteria derived from infected normal monocytes cultivated in normal serum medium for 6 days at 37 C; immune system refers to bacteria from infected immune monocytes cultivated in immune serum medium. (From Fong et al. (28).)

18 84 SANFORD S. ELBERG [VOL. 24 system caused no degeneration of monocytes when recovered after 48 hr of bacilli-monocyte interaction. The results of the experiments described above demonstrate that the mononuclear cell of peritoneal exudates obtained from animals regarded as immune, when in contact with variously derived antisera, constitutes a host system highly immune to the normally destructive action of certain microorganisms. The evidence indicates that the reactivity of the immune monocyte is more specific than the reactivity of the serum component of the immune animal. The limited specificity of the immune monocyte was reflected by the cross-reactivity between infections by M. tuberculosis and B. melitensis. Other data suggest in addition that the acquired resistance of the monocyte is directed by the two above-cited organisms toward the class of parasites generally regarded as intracellular. IV. REACTIVITY OF IMMUNE MONOCYTE AS REFLECTION OF RESPONSE OF INTACT ANIMAL Suggestions that the reactivity of the immune monocyte is a true reflection of the response of the "whole animal" come from the studies of Pullinger (78) and Nyka (74), who demonstrated similar relations between M. tuberculosis and Brucella abortus infections in guinea pigs. In fact, guinea pigs have responded to this type of experimentation in a positive fashion in mixed infections involving Coxiella burnetii and Brucella species (Mika et al. (69)), Brucella suis and M. tuberculosis (Henderson et al. (38) and Henderson (this Symposium)). The cell with which we have been concerned is part of a system of which the histiocyte is the primordial unit, a cell which takes its form in part from its environment. The peripheral section of its cytoplasm is hyaline and without granules, whereas the perinuclear part is dark, granular, and contains a basophilic cytoplasm with rod-shaped mitochondria and various types of granules or vacuoles. The cytoplasm is finely reticulated with varying sized vacuoles. The scattered and abundant phagocytic vacuoles, the size of the cell, the abundance and smallness of its mitochondria, and its pattern of mobility are aids in differentiating the monocyte from lymphocytes and plasma cells. The view has been expressed by Tompkins (99) that "monocytic morphology represents a specific physiologic status in the life of a cell which is normally destined to become a macrophage, i.e., a young cell on its way to growing up, just beginning to be actively phagocytic, and not yet capable of taking vital dyes, let alone taking the mass of tissue debris and foreign substance which is the function of the macrophage" (cf. Policard (76) and Bessis (10) for review of morphology and physiology of the histiocyte and leucocytes in general). Much of the discrepancy in the literature that is concerned with the fate of ingested bacilli (vide infra) could stem from the differences in physiological state of the cell' in cultuire as a consequence of the differences in inflammatory incitants, culture media, gaseous tensions, and inorganic environment, leading to populations of monocytes responding quite differently depending on the initial parasite to monocyte ratios. Studies concerned with events accompanying or following phagocytosis are increasingly aimed at a biochemical basis of explanation. It has been reported on several occasions that the oxygen uptake of peritoneal exudate leucocytes of the polymorphonuclear type is very low and that the uptake markedly increases during phagocytosis, suggesting a significant expenditure of energy required by the cell (De Gregorio (18), Delaunay et al. (19), Stahelin et al. (93-95), Harris and Barclay (37)). The oxygen uptake of peritoneal exudate leucocytes is less than that of blood leucocytes, suggesting that the process of exudation is in itself somewhat damaging (De Gregorio (18)). The refined experiments carried out by Sbarra and Karnovsky (86) are an important beginning in the understanding of the intermediary metabolism concerned in phagocytosis. The respiratory activity of phagocytizing cells increases within the first hour after mixing the cells and particles and then declines; respiration is proportional to particle concentration to a maximal point, after which a more or less steady level is maintained. The increment of oxygen uptake is proportional to the serum concentration, being maximal at 30 per cent, as well as to the number of particles ingested. Glucose consumption and lactic acid production respond to the act of phagocytosis to a greater extent under anaerobic conditions. The CO2 production from glucose induced by phagocytosis was predominantly via direct oxidative pathways, the CO2 derived from

19 1960] CELLULAR IMMUNITY 85 the C-1 of glucose undergoing a 7-fold increase as against a 2.5-fold increase for the C-6 -* CO2 reaction. Under anaerobic conditions, the C-1-* CO2 production doubled under the impact of the phagocytic reaction. Experiments with dinitrophenol and potassium cyanide indicated that phosphorylation and cytochrome systems are not involved in the glycolytic pathway during phagocytosis. Lactate production was increased during the first 30 min of phagocytosis, suggesting that with the energy demand, glycolysis proceeds rapidly. With lactic acid accumulation, and as the phagocytic reaction slows down or ceases, the glycolytic reaction reverses itself (Sbarra and Karnovsky (86); cf. Fisher and Ginsberg (23) who utilized similar methods to analyze the inhibition of phagocytosis by influenza virus, and Alonso and Nungester (2) on the inhibitory effects of pneumococcal polysaccharide on cell metabolism). The increased respiration attending phagocytosis of strains of M. tuberculosis follows proportionally the virulence of the strain. Recent reviews on the general biochemistry of polymorphonuclear neutrophiles (Bazin (6) and Delaunay et al. (19)) and on the biochemistry and enzymatic activities of leucocytes with special reference to states of health and disease (Valentine (103)) are of special interest. The richness in enzymatic systems of cultivated macrophages and monocytes has been attributed to the activation of resting histiocytic cells and their transformation under the stimulus of explantation (Gropp and Hupe (35)). During the differentiation of monocytes in culture, at least 24 hr were required for the rapid accumulation of acid phosphatase, polysaccharide, ribonucleic acid, and protein by macrophages and giant cells (Goldstein and McCormick (33)). It has been proposed that the acid phosphatase is the most important single enzyme in the monocytes of animals resistant to M. tuberculosis infection, that the presence of this enzyme is a specific response to the mycobacterial phosphatide and that the enzyme is a prominent feature in necrotic material of lesions in susceptible but not resistant animals from degenerating polymorphonuclear cells (Grogg and Pearce (34)). However, a more general occurrence of the enzyme was demonstrated by Goldstein and McCormick (33) in various nonphagocytic cells of the body, which casts doubt on the specificity of this enzyme for the phagocytic activity of the macrophage. The presence of acid phosphatase suggests, on the other hand, that the transformation of the monocytes is not a degenerative process. The presence of acid phosphatase in macrophages where none was demonstrable in monocytes in culture probably represents only a change from a low to a high level of activity of phagocytosis which endures through giant cell formation (Weiss and Fawcett (106)).. BREAKDOWN OF MICROORGANISMS WITHIN PHAGOCYTIC CELLS Parallel with the great number of studies on the enzymatic complement of phagocytic cells and the biochemical description of the early stages of phagocytosis are the investigations of the actual steps in the intracellular breakdown of the parasite. Pathways to this latter objective are via the analysis of protoplast formation, a study of the compounds released from bacteria by serum components, and a study of intramonocytic lysis of the bacterium and the substances which affect this reaction. The rigidity of the surface "envelope" of gramnegative bacteria is a function of the lipoproteinamino sugar complex in the walls. Action by lysozyme, deprivation of diaminopimelic acid (DAP), or penicillin action weakens the wall and causes spherical shapes to appear (Weidel and Primosigh (105)). Lysozyme can be replaced by the properdin component of serum if antibody, complement, Mg++, and sucrose are available (Muschel et al. (70, 71)). In the presence of complement, sensitized cell suspensions of Vibrio comma and Vibrio metschnikovii evidence spherical cell transformation and at least half the cell-wall DAP is released (Shafa and Salton (87)). This allows a reasonable conclusion that "Pfeiffer's phenomenon is explicable in terms of an enzymatic release of the cell-wall peptide which in the normal cell provides the wall with a rigid structural framework" (87). Amano et al. (3-5), studying a similar system, had observed that lysozyme can act after the antibody-complement system to accelerate lysis. Presumably peptide disintegration weakens the wall structure and amino sugar hydrolysis completes the reaction. No appreciable lysis occurred during DAP release under antibodycomplement action (87) although such lysis was inferred (3-5). The. release of bactericidal and antibacterial substances from the polymorphonuclear type of

20 86 SANFORD S. ELBERG [VOL. 24 cell has been studied far more than from the mononuclear cell. For example, phagocytin from rabbit leucocytes (Hirsch (40, 41)) is bactericidal but not bacteriolytic in vitro, although it might serve as the first reacting component in vivo where bacteriolysis was occurring. It will be of great interest to know whether phagocytin in its action on gram-negative bacteria acts on the muramic acid component or on the peptide structure. Considering its nonlytic activity in vitro on gram-negative bacteria, it may be presumed that phagocytin has for its "substrate" the peptide component. The fact that activity of phagocytin is enhanced at acid reactions implies that intracellular conditions might be quite appropriate for its activity under the impact of the inflammatory reaction and the consequent production of lactic and other organic acids. Antibacterial substances isolated from leucocytes and active against gram-positive organisms have been characterized. One, "leukin," has been carefully restudied by Skarnes and Watson (91) after it had been relatively ignored since its earlier discovery. The substance appears to be derived from the nuclear fraction and has the properties of a protamine. It has been postulated that one role for the action of the "leukin" type of substance is its release from the leucocytes into the inflammatory environment and its extracellular or preparatory action on the parasite prior to ingestion. This would, as Skarnes and Watson (91) suggest, provide an additional substance to account for the histopathological picture in early anthrax lesions. The other two antibacterial substances active on gram-positive organisms are the mitochondrial fraction (Fishman and Silverman (24)) and the lysozyme-like substances of polymorphonuclear leucocytes and bovine spleen (Amano et al. (3-5)). Little is known of their mode of action, chemical nature, or site of action. Since release of lysozymelike enzymes is one response to injury of leucocytes by bacterial products and can serve an an index of leucocyte damage (Kerby (47)), it is difficult to evaluate whether a given enzyme is produced as a secretory product upon the stimulus of inflammation or actually reacts on its bacterial substrate intraleucocytically. The known kinetics of action of lysozyme and the requirement that the susceptible organism should be in an actively growing and actively metabolizing state (Becker and Hartsell (8)) suggest that one of the most significant inhibitors to effective phagocytic action intracellularly would be the ability of the parasite to slow its own biosynthetic mechanism in the face of the burst of respiratory activity attending the act of phagocytosis. By a feedback arrangement the parasite might control the available amount of leucocytic enzyme substrate produced in its cell wall. Alternative to this inhibitory action on the part of the virulent parasite (expressed as an ability to enter a "latent" metabolic state) would be the ability to respond to the intraleucocytic environment by a process of adaptation of cell-wall constituents of slightly altered configuration, and hence insusceptible to phagocytic action. Evidence for this exists in an "experiment of nature" in which certain streptococci, for which the correlation between MI substance and capsule formation with virulence and phagocytability does not hold, can be rendered phagocyte-sensitive by treatment with trypsin (Wiley and Wilson (108)). Trypsin alters the bacterial surface by its action on the peptide area of the cell wall. The early work of Lamar (50-53) on the interaction of certain fatty acids and bacteria was predicated on the view that recovery from local bacterial infections also involves the effects of definite chemical substances always present in a focus containing disintegrating organisms and tissues. Noguchi (72) had shown that bactericidal and hemolytic activity of leucocyte extracts was dependent upon their content of certain higher unsaturated fatty acids or their alkaline soaps. Under the influence of the lipase which is abundant in monocytes the fatty acids were liberated. Lamar showed that oleate enhanced serumlysis of pneumococci, which tended to be incomplete with normal serum but complete with immune serum. This reaction was operative in vivo as well as in vitro, in mice as well as in monkeys. The possible site and mode of action of lipotropic macromolecules in tuberculosis was suggested by Mackaness' (60) study on the suppression of growth of tubercle bacilli in phagocytic cells. Polyoxyethylene ethers derived from a high molecular weight condensate of octyl alcohol and formaldehyde were active in vivo but not in vitro against tubercle bacilli. It was suggested that these substances became concentrated in monocytes where they acted to solubilize membrane lipids of ingested parasites through removal or displacement of hydrophobic groups. It was pro-

21 19601 CELLULAR IMMUNITY posed that analogous infection-enhancing or "pro-tuberculous" agents act by displacing the more hydrophilic lipids, rendering the lipid layer of the bacterium even more resistant than it is normally (Lovelock and Rees (56)). It may be suggested that this type of nonspecific alteration of phagocytic internal environment is responsible for the action of certain bacterial constituents (endotoxins) in increasing mouse resistance to various types of heterologous infections, since the activity of the endotoxin appears to be associated with "Lipid A" component and is associated also with the sequence of hydrophilic and hydrophobic groups of this lipid (Boehme and Dubos (11), Evans and Perkins (22)). Endotoxin, however, has additional effects such as increased phagocytosis, activation of properdin, activation of stressor organs, and stimulation of leucocyte migration. An alternative explanation for the nonspecific enhancement of resistance by substances such as endotoxin is that of a reflex mechanism. First contact irritates physiologic and nervous receptors and by direct action on the central nervous system prevents reflexes normally leading to lethality, thereby resulting in increased survival (Krylov (49), Kreschko and Guobis (48), Karaev et al. (45)). Concerning events inside the polymorphonuclear phagocyte which are part of nonspecific cellular immunity, it has been suggested for example that the protection of pathogenic staphylococci from the intracellular milieu is associated with the action of coagulase. Rogers and Tompsett (83) and Tompsett (100) demonstrated that pathogenic staphylococci survived in such situations longer than saprophytic variants. According to these authors the exotoxins and leucocidins are more responsible for the pathogenic variant's staying power, perhaps enhanced by coagulase. Other reports suggest more importance for the leucocytic lysozyme and ribonuclease, the pathogenic strain resisting this complex whereas the nonpathogenic strains are lysed. According to this view, the saprophytic strains possessing the lysozyme substrate in their walls are attacked by the two enzymes. Development of penicillin resistance and concomitant resistance to intraleucocytic action are attributed to loss of receptor sites for the lysozyme component of the host cell and consequent in vivo survival (Mesrobeanu et. al. (66) and Mai et al. (63)). The "critical time" concept has arisen even in consideration of intraphagocytic digestion, for it has been shown that division of streptococci egested from leucocytes or escaping after leucocytic destruction by leucotoxin depends upon how long they had been in the cell before liberation. If egestion occurred within 15 minutes following phagocytosis, proliferative power was maintained, whereas 30 min inside the leucocyte caused an irreversible event (Wilson et al. (109); Fleck (25); cf. events in phagocytosis of virulent and avirulent P. pestis reported by Burrows and Bacon (14)). A. Phagocytic Destruction of B. melitensis The results of studies carried out in the writer's laboratory by D. J. Ralston and B. Baer (unpublished data) with the vaccine strain of B. melitensis, Rev Is (Herzberg and Elberg (39), Elberg and Faunce (20)), summarized partly in tables 13-17, suggest that death of these bacteria may result when the mucopolysaccharide wall substrate has been exposed by an agent acting to cause preliminary damage. In the infected monocyte both in vivo and in vitro, glycine is capable of rendering brucellae susceptible to lysozyme-like action (cf. Welsch (107)). In studies on the survival and multiplication of smooth brucella cells in infected monocytes of normal rabbits, glycine added to the Tyrodeserum tissue culture medium significantly depressed the bacterial growth and, as seen in table 13, in certain experiments brought about marked intracellular bacterial destruction. The concentrations used, in the order of 1000 to 2000 ;g per ml, were not harmful to the monocytes themselves. Extracellular growth of brucellae was prevented by the presence in the extracellular fluid of 50 to 200 pg per ml of streptomycin. Subsequent analyses in vitro of the intramonocytic TABLE 13 Effect of glycine on intracellular brucellae Viable Brucellae Recovered per ml from Washed and Lysed Monocytes after Incubation for 1 day Control preparation 1.5 X 105 Glycine, 0.03 M, added 2.6 X 104 after phagocytosis 5 days X X 102

22 88 SANFORD S. ELBERG [VOL. 24 TABLE 14 Effect of monocyte extract on growth of glycine-treated brucellae Treatment of Bacterial Cells Viable Cells per ml after Incubation at Ohr 24 hr 96 hr Control (serum-tyrode) X X X 108 Glycine, 0.03 M, alone.1.8 X X X 109 Monocyte extract.1.8 X X X 107 Glycine plus extract. 1.8 X X 101 <105 TABLE 15 Action of lysozyme and monocyte extract on glycinetreated rough and smooth cells of Brucella melitensis strain Rev Is Sample Treatment in Tyrode's Buffer Solution at 37 C Per Cent Turbidity Decrease Rough Smooth Rev Is Rev Is 40 min 40 min 1. Monocyte extract, 1/5 dilution Glycine, 0.3 M Extract 1/5 + glycine 0.3 M min 0 min 4. Lysozyme, 5,.g/ml Glycine, 0.3 M Lysozyme + glycine, 0.3 M Method: Viable brucella cultures were harvested into saline after 18-hr growth on Albimi agar at 37 C and suspended to give a final concentration equivalent to 53 Klett units (by direct count this is equivalent to approximately 2 X 109 cells per ml.). The monocyte extract was prepared by a freeze and thaw procedure. The monocytes were suspended in Tyrode's solution, frozen and thawed 6X. The debris was spun out and the supernatant was stored at 4 C until used. The active material was quite stable at refrigerator temperatures. For the lysis test, mixtures containing 0.3 M glycine, 5.0 jig lysozyme, or a 1/5 dilution of monocyte extract were made in a solution containing 20 per cent Tyrode's and 80 per cent 0.01 M phosphate buffer, ph 6.2. The total volume was 5 ml in a Klett-size test tube. Rough and smooth cells were added and the tubes were incubated stationary at 37 C. Readings were made at intervals in a Klett photoelectric calorimeter equipped with a red filter. effects suggest that a lethal physiological environment may be induced by the combined use of glycine and a lysozyme-like agent present in the cells. This agent, which is not identical to lysozyme, may be extracted from the monocytes of both normal and immune rabbits by a variety of procedures whereby the cells are ruptured in saline buffer solutions at ph 3.5 to 7.5. By itself the enzyme has no killing action on the brucellae, although it is capable of agglutinating rough cells and may delay the growth of very young smooth cells. However, when combined with sufficient amounts of glycine, death and lysis occur (tables 14 and 15). This effect when studied extracellularly occurs more rapidly with fast-growing younger brucellae than with stationary phase cells. Intracellularly, rough cells are more susceptible than smooth. Preliminary experiments indicate that the rate of lysis is dependent upon both the glycine concentration and amount of monocytic enzyme. The surviving brucella cells appear to be capable of multiplying in the presence of appreciable levels of these substances but may again pass into a susceptible stage. In the presence of glycine, crystalline lysozyme substitutes for the action of monocyte enzyme (table 15). Both enzymes act on heated and chloroform-acetone-treated or butanol-treated heated cells, causing the release of a viscous material identifiable as bacterial deoxyribonucleic acid. Under these conditions a decrease in bacterial turbidity does not always occur. Cells which have been killed and treated with lysozyme or the monocyte enzyme are rendered more susceptible to the action of pancreatic lipase, trypsin, and chymotrypsin as measured by drop in turbidity and the release of additional viscous material. With living brucellae in a nutrient medium such as Albimi broth, glycine and either crystalline lysozyme or monocyte enzyme produce lysis and death. This response is not brought about by the following enzymes combined with glycine: trypsin, chymotrypsin, papain, lipase, deoxyribonuclease, ribonuclease, or D-aminO acid oxidase. The proteolytic agents, however, exert a slight

23 19601 CELLULAR IMMUNITY 89 lethal action. With living cells, the concentration of glycine required to render brucella cell walls susceptible to lysozyme ranges from 0.1 to 0.3 M, but this is dependent upon the age of the bacterial cell and other as yet undefined cultural conditions. The appearance of bacterial deoxyribonuclease in the culture medium suggests that death may also depend upon the loss of this material following the action of lysozyme on the cell wall. Under certain conditions, such as insufficient enzyme or TABLE 16 Effects of antiserum and monocyte extract on growth in vitro of brucellae Medium Increased Turbidity (Klett Units) at 4½jhr 18 hr Broth alone Broth M Glycine Glycine + extract Glycine + extract M sucrose Antiserum Antiserum + extract Glycine + antiserum + extract 0 9 glycine, or rapid oxygenation, apparently many brucellae are capable of repairing the damage, since they may grow in the presence of large amounts of these substances. The role of glycine may be similar to that of Versene (ethylenediaminetetraacetic acid) in chelating certain essential metal ions that are necessary for the maintenance of a healthy cell wall, as demonstrated with certain gram-negative bacteria. It may also act by causing an unbalanced wall synthesis, perhaps producing protoplast-like forms already reported for brucellae by Gerhardt and also indicated by the effect of sucrose shown in table 16. These findings suggest that intramonocytic death of these bacteria begins as a result of damage to the cell wall. There may be a number of other amino acids, peptides, products of inflammation, and perhaps specific antibodies that act similarly. The survival of brucellae might then depend on the levels of both the initial walldisturbing agent and of intracellular "lysozyme," the immune monocyte facilitating more rapid uptake of the wall-disturbing agent and faster synthesis of intracellular enzyme. Participation of immune serum in the monocytic digestion process is illustrated in tables 16 and 17, the data for death as well as lysis suggesting that the antibody is a preparative reagent for Downloaded from TABLE 17 Effect of glycine, immune serum, and their combination on intracellular Brucella melitensis, strain Rev Is, in normal rabbit monocytes No. Intracellular Organ- pg/ml Test Condition isms per ml after 72 hr 37 Viable Monocytes per Ml after Incubation at C in Presence of 200 Streptomycin 0 hr 24 hr 72 hr 1. Control in Tyrode's solution 6.6 X X X X Glycine, 0.03 M, added at 7 hr 8.4 X X X X 105 :3. Rabbit antiserum, 1/100l 7.4 X X X X Glycine, 0.03 M, and antiserum, <103 1 X X X 105 1/100 Method: Monocytes were harvested into Tyrode's solution on the 5th day after injection of 30 ml sterile Klearol into 2 female rabbits. The yield was centrifuged lightly and resuspended into approximately 20 ml Tyrode's solution, washed once, and suspended in a Tyrode's solution-serum medium to contain 1 X106 monocytes per ml. The monocytes were mixed with an overnight growth of Rev Is suspended in Tyrode's solution in an initial infection ratio of 100:1 (1 X 101 bacteria to 1 X 106 monocytes). Samples were placed in small 12-ml centrifuge tubes, 2 ml per tube, slanted, and phagocytosis was allowed to occur at 37 C. After 7 hr the monocytes had settled to the glass walls, the residual, unphagocytized bacteria were decanted; streptomycin was added to a final concentration of 200,.g per ml; 1 hr later the infected monocytes were treated with glycine and/or immune serum. a Prepared by injection of rabbits with a suspension of B. melitensis, Rev Is, inactivated by ultraviolet light and suspended in "Arlacel" and "Klearol" (D. J. Ralston, and B. Baer, 1959, unpublished data). on March 17, 2019 by guest

24 90 SANFORD S. ELBERG [VOL. 24 intramonocytic digestion. This reaction may be part of a general phenomenon of absorption acting to render cells sensitive to lysis, as suggested by its similarity to absorption of specific bacteriophage to prepare protoplasts of Escherichia coli (Carey et at. (15)), absorption by staphylococci of specific bacteriophage rendering the cells sensitive to virolysin (Ralston et al. (80, 81)), and absorption of antibody rendering cells sensitive to the enzymatic action of complement (Shafa and Salton (87)). VI. SUMMARY AND CONCLUSIONS The traditional dichotomy of cellular immunity into its specific and nonspecific aspects appears to be a nonproductive concept at the enzymesubstrate level of intracellular management of a parasite. The preparation by antibody of an area on the parasite wall for intraphagocytic action, as so clearly indicated in recent work, seems to lead to an explanation in precise chemical terms of the meaning classically attributed to the term "opsonin" by the earlier students of immunity. Whether the enzymatic attack on the parasite is intra- or extracellular is, at this level, quite beside the point, since it is a matter of evolutionary development of a particular parasitism. In this respect, it is submitted that "cellular immunity," as distinguished traditionally from noncellular, or, more correctly "extracellular" immunity, simply reflects one aspect of the ecological niche in which the parasite involved is to be found. But what appears to be more cogent, the general outlines of its destruction are the same wherever it is found. To return to a more biological and perhaps evolutionary mode of analysis, data have been described in this review which show that: (a) normal monocytes are "lysed" by certain ingested virulent bacteria but not by avirulent ones and that "immune" monocytes are not "lysed" by homologous parasites; (b) "immune" monocytes can support growth of the homologous parasite temporarily but it then decreases and stops; (c) the immunity in its quantitative expression, from data on persistence of the parasite in a latent state as well as from "growth curve" experiments, suggests that the parasite does actually persist as an elementary particle; (d) the immunity is group specific, judged by studies on mixed infections in the whole animal as well as in cell culture situations; (e) situations can be experimentally created where ingestion occurs but not intraphagocytic multiplication and where, by a change in the identity of the serum component of the cell culture system, intraphagocytic multiplication can be induced. This finding may be interpreted as follows: in the presence of one serum, the ingested parasite is converted to an elementary unit of persisting quality; in another serum this conversion does not occur and the parasite multiplies; (f) the stubborn persistence of the concept of "infection immunity" in the case of tuberculosis, brucellosis, and other infectious diseases is strongly supported by data from so-called "superinfection type" experiments. The infection-immunity phase depends on the presence of the infection and is lessened quantitatively by the evolution of the sterile phase of immunity (Kelly et al. (46)). The resemblance of the six points just cited to the cellular immunity aspects of lysogeny (Lwoff (59)) is rather striking: (a) lysogenic bacteria can produce phage; (b) they can adsorb homologous phage, whose genetic material probably persists in the bacterial protoplasm, unable to replicate in an immune bacterium; (c) the immunity of the lysogenic bacterium is bound to the persistence of the prophage and when this develops into phage the immunity to a second infecting phage is thereby lost; (d) the immunity seems to be groupspecific, and is a corollary of prophage's inability to develop; (e) one of the strangest phenomena in this connection is the existence of a prophage in the guise of a phage whereby lysogeny can be transduced. A point of convergence upon development of the ideas of cellular immunity (cf. effect of serum in item (e)) appears in the observation by Lennox (55) and Jacob (44) that the cell receiving the "transducing prophage" can produce phage and lyse, or lysis may be prevented (i.e., production of phage is prevented but prophage is maintained) by lowering the temperature of the system. As recently stated by another, "a reviewer ought not to be tempted by his subject-matter, even if it owes a very great deal of its attraction to aptness for constructive speculation, to indulge in probably quite unwarrantable hypothesismaking of his own. He retreats with apologies" (Weidel (104)). It is submitted that these considerations of certain aspects of cellular immunity are not only of interest to problems of immunity but are part and parcel of an evolutionary ap-

25 19601 CELLULAR IMMUNITY 91 proach to parasitism. The changes we see in the responsiveness of parasites to host-cell environment and vice versa must be but reflections of the events occurring as a parasite develops to the stage where, for example, it no longer requires a maintenance host and its arthropod vector. The existence of pneumonic plague, the domiciliation of parasites to the stage of maternal-to-fetal transmission in utero as with Trypanosoma cruzi, show that, while we are still deeply involved in the adjustment reactions of which "cellular immunity" is but one, we are also concerned with the terminal stages of ecological complexes in nature where lysogeny, masking phenomena, and latency appear to serve as some examples. VII. REFERENCES 1. ABE, S Studies on the intraperitoneal cells of a guinea pig and the multiplication of tubercle bacilli in the intraperitoneal mononuclear cells cultured in vitro. Sci. Repts. Research Insts. Tohoku Univ., 8, ALONSO, D. AND NUNGESTER, W. J Comparative study of host resistance of guinea pigs and rats. V. The effect of pneumococcal products on glycolysis and oxygen uptake by polymorphonuclear leucocytes. J. Infectious Diseases, 99, AMANO, T., NISHEMOTO, M., MAI, S., SEKI,. Y., KASHIBA, S., FUJIKAWA, K., AND ORIHARA, M Studies on immune bacteriolysis. V. Fractionation of active substances from acidic extract of bovine spleen. Med. J. Osaka Univ., 5, AMANO, T., MORIOKA, T., SEKI, Y., KASHIBA, S., FUJIKAWA, K., AND ICHIKAWA, S Studies on the immune bacteriolysis. VIII. The mechanisms of the immune bacteriolysis. Med. J. Osaka Univ., 6, AMANO, T., MORIOKA, T., SEKI, Y., AND KASHIBA, S Studies on the immune bacteriolysis. IX. The effect of immune bacteriolytic system and lysozyme on the Med. J. Osaka Univ., cell wall preparation. 6, BAZIN, S Biochemistry of polymorphonuclear leukocytes. Rev. fran. dlin. biol., 1, BEARD, J. W. AND Rous, P The fate of vaccinia virus on cultivation in vitro with Kupffer cells (reticulo-endothelial cells). J. Exptl. Med., 67, BECKER, M. E. AND HARTSELL, S. E The synergistic action of lysozyme and trypsin in bacteriolysis. Arch. Biochem. Biophys. 55, BENEDICT, A Growth of meningopneumonitis virus in normal and immune guinea pig monocytes. Nature, 181, BESSIS, M Cellular structures discovered by the electron microscope in leukocytes. Rev. hematol., 11, BOEHME, D. AND DuBos, R. J The effect of bacterial constituents on the resistance of mice to heterologous infection and on the activity of their reticulo-endothelial system. J. Exptl. Med., 107, BRAUN, W., POMALEs-LEBR6N, A., AND STINEBRING, W. R Interactions between mononuclear phagocytes and Brucella abortus strains of different virulence. Proc. Soc. Exptl. Biol. Med., 97, BRIEGER, E New approaches to the study of experimental infection. Advances in Tuberculosis Research, 5, BURROWS, T. W. AND BACON, G. A The basis of virulence in Pasteurella pestis: The development of resistance to phagocytosis in vitro. Brit. J. Exptl. Pathol., 37, CAREY, W. F., SPILMAN, W. M., AND BARON, L. S Protoplast formations by mass adsorption of inactive bacteriophage. J. Bacteriol., 74, CAVANAUGH, D. C. AND RANDALL, R Role of mononuclear phagocytes in the pathogenesis of flea-borne plague. Federation Proc., 18, COLOBERT, L Study of lysis of pathogenic salmonellae by lysozyme after partial delipidation. Ann. inst. Pasteur, 95, DE GREGORIO, P Oxygen consumption of leukocytes and phagocytosis in different experimental conditions. Boll. soc. ital. biol. sper., 32, DELAUNAY, A., PAGES, J., AND MAURIN, M Le metabolisme respiratoire des leucocytes. Ann. inst. Pasteur, 72, ELBERG, S. S. AND FAUNCE, K., JR Immunization against brucella infection. VI. Immunity conferred on goats by a nondependent mutant from a streptomycindependent mutant strain of Brucella melitensis. J. Bacteriol., 73, ELBERG, S. S., SCHNEIDER, P., AND FONG, J Cross-immunity between Brucella melitensis and Mycobacterium tuberculosis.

26 92 SANFORD S. ELBERG [VOL. 24 Intracellular behavior of Brucella melitensis in monocytes from vaccinated animals. J. Exptl. Med., 106, EVANS, D. G. AND PERKINS, F. T The ability of pertussis vaccine to produce in mice specific immunity of a type not associated with antibody production. Brit. J. Exptl. Pathol., 35, FISHER, T. N. AND GINSBURG, H. S The reaction of influenza viruses with guinea pig polymorphonuclear leucocytes. III. Studies on the mechanism by which influenza viruses inhibit phagocytosis. Virology, 2, FISHMAN, M. AND SILVERMAN, M. S Bactericidal activity of rat leucocytic extracts. I. Antibacterial spectrum and the subcellular localization of the bactericidal activity. J. Exptl. Med., 105, FLECK, L Le leukergie et son role dans l'immunit6. Ann. inst. Pasteur, 92, FONG, J., SCHNEIDER, P., AND ELBERG, S. S Studies on tubercle bacillus-monocyte relationship. I. Quantitative analysis of effect of serum of animals vaccinated with BCG upon bacterium-monocyte system. J. Exptl. Med., 104, FONG, J., SCHNEIDER, P., AND ELBERG, S. S Studies on tubercle bacillus-monocyte relationship. II. Induction of monocyte degeneration by bacteria and culture filtrate: Specificity of serum and monocyte effects on resistance to degeneration. J. Exptl. Med., 105, FONG, J., CHIN, D., AKIYAMA, J., AND EL- BERG, S. S Studies on tubercle bacillus-monocyte relationship. III. Conditions affecting the action of serum and cells. Modification of bacilli in an immune system. J. Exptl. Med., 109, FREEMAN, B. A. AND VANA, L. A Host-parasite relations in brucellosis. I. Infection of normal guinea pig macrophages in tissue culture. J. Infectious Diseases, 102, FURNESS, G Interaction between Salmonella typhimurium and phagocytic cells in cell culture. J. Infectious Diseases, 103, FURNESS, G. AND ROWLEY, D Transduction of virulence within the species Salmonella typhimurium. J. Gen. Microbiol., 15, GINDER, D. R Resistance to fibroma virus infection: The role of immune leukocytes and immune macrophages. J. Exptl. Med., 101, GOLDSTEIN, M. N. AND MCCORMICK, T Cytochemical studies during the differentiation of normal human monocytes in vitro. Am. J. Pathol., 33, GROGG, E. AND PEARSE, A. G. E The enzymic and lipid histochemistry of experimental tuberculosis. Brit. J. Exptl. Pathol., 33, GROPP, A. AND HUPE, K Demonstration of enzymes in macrophages cultivated in vitro and in epitheloid cells. Arch. pathol. Anat. u. Physiol. Virchow's, 331, HANKS, J Assay of the fate of mycobacteria in cell and tissue cultures. Am. Rev. Tuberc. Pulmonary Diseases, 77, HARRIS, H. AND BARCLAY, W. R A method for measuring the respiration of animal cells in vitro, with some observations on the macrophages of the rabbit. Brit. J. Exptl. Pathol., 36, HENDERSON, D. W., LANCASTER, M. C., PACKMAN, L., AND PEACOCK, S The influence of a pre-existing respiratory infection on the course of another superimposed by the respiratory route. Brit. J. Exptl. Pathol., 37, HERZBERG, M. AND ELBERG, S. S Immunization against brucella infection. III. Response of mice and guinea pigs to injection of viable and non-viable suspensions of a streptomycin-dependent mutant of Brucella melitensis. J. Bacteriol., 69, HIRSCH, J. G Phagocytin: A bactericidal substance from polymorphonuclear leucocytes. J. Exptl. Med., 103, HIRSCH, J. G Studies of the bactericidal action of phagocytin. J. Exptl. Med., 103, HOLLAND, J. AND PICKETT, M A cellular basis of immunity in experimental brucella infection. J. Exptl. Med., 108, HOWARD, D. H Observations on tissue cultures of mouse peritoneal exudates inoculated with Histoplasma capsulatum. J. Bacteriol., 78, JACOB, F Transduction of lysogeny in Escherichia coli. Virology, 1, KARAEV, A. I., KAFAROVA, R. Z., AND RUSTA- MOVA, S. H Effect of stimulation of chemoreceptors of the liver and of mechanoreceptors of the colon on phagocytic activity of leukocytes. Biull. eksptl. biol. med., 41, KELLY, E. H., GORELICK, H. N., SILVERMAN,

27 1960] CELLULAR IMMUNITY 93 S., AND BRAUN, W Further studies on superinfection in experimental bovine brucellosis. J. Infectious Diseases, 93, KERBY, G. P Release of enzyme from human leucocytes on damage by bacterial derivatives. Proc. Soc. Exptl. Biol. Med., 81, KRESCHKO, L. AND GuoBIs, G On the connection of the cellular immunity reaction and the general course of experimental staph intoxication with the typological characteristics of the nervous system in dogs. Zhur. Mikrobiol. Epidemiol. Immunobiol., 29, KRYLOV, V. N Analysis of physiological basis of so-called depression immunity. I. Effect of homo- and heterological antigenic irritations and of non-specific stimuli on manifestation of depression immunity. Zhur. Mikrobiol. Epidemiol. Immunobiol., 29, LAMAR, R. V Chemo-immunological studies on localized infections. I. Action on the pneumococcus and its experimental infections of combined sodium oleate and antipneumococcus serum. J. Exptl. Med., 13, LAMAR, R. V Chemo-immunological studies on localized infections. II. Lysis of the pneumococcus and hemolysis by certain fatty acids and their alkali soaps. J. Exptl. Med., 13, LAMAR, R. V Chemo-immunological studies on localized infections. III. Some further observations upon the action of certain soaps on the pneumococcus and its experimental infections. J. Exptl. Med., 14, LAMAR, R. V Chemo-immunological studies on localized infections. IV. Experimental pneumococcic meningitis and its specific treatment. J. Exptl. Med., 16, LARSH, H. AND SHEPARD, C. C Hela cells and Histoplasma capsulatum. Phagocytosis and subsequent intracellular growth. J. Bacteriol., 76, LENNOX, E. S Transduction of linked genetic characters of the host by bacteriophage P1. Virology, 1, LoVELOCK, J. E. AND REES, J. W Possible site and mode of action of certain lipotropic macromolecules in tuberculosis. Nature, 175, LURIE, M. B Studies on the mechanism of immunity in tuberculosis. The fate of tubercle bacilli ingested by mononuclear phagocytes derived from normal and immunized animals. J. Exptl. Med., 75, LURIE, M. B Native and acquired resistance to tuberculosis. Am. J. Med.. 9, LWOFF, A Lysogeny. Bacteriol. Revs., 17, MACKANESS, G. B The action of drugs on intracellular tubercle bacilli. J. Pathol. Bacteriol., 64, MACKANESS, G. B Growth of tubercle bacilli in monocytes from normal and vaccinated rabbits. Am. Rev. Tuberc., 69, MACKANESS, G. B., SMITH, N., AND WELLS, A. Q Growth of tubercle bacilli (intracellular) in relation to virulence. Am. Rev. Tuberc., 69, MAI, S., SEKI, Y., KASHIBA, S., INOVE, K., AND NISHIMOTO, M Studies on immune bacteriolysis. IV. The effect of the acidic extract of leucocytes on immune bacteriolysis. Med. J. Osaka Univ., 5, MCCUNE, R. M. AND TOMPSETT, R Fate of Mycobacterium tuberculosis in mouse tissues as determined by the microbial enumeration technique. I. The persistence of drug-susceptible tubercle bacilli in the tissues despite prolonged antimicrobial therapy. J. Exptl. Med., 104, MCCUNE, R. M., TOMPSETT, R., AND MCDER- MOTT, W The fate of Mycobacterium tuberculosis in mouse tissues as determined by the microbial enumeration technique. II. The conversion of tuberculous infection to the latent state by administration of pyrazinamide and a companion drug. J. Exptl. Med., 104, a. MCWHORTER, A. C. AND EDWARDS, P. R A new Salmonella serotype (Salmonella rutgers). Cornell Vet., 46, MESROBEANU, I., MESROBEANU, L., MITRICA, N. T., AND PAPAZIAN, E The leucocytic enzymes and problem of survival of M. pyogenes in the leucocytes after phagocytosis. Minerva med., 48, MEYNELL, G. G. AND STOCKER, B. A. D Some hypotheses on the aetiology of fatal infections in partially resistant hosts and their application to mice challenged with Salmonella paratyphi-b or Salmonella typhimurium by intraperitoneal injection. J. Gen. Microbiol., 16, MEYER, K. F Immunity and mononuclear response in psittacosis. Schweiz. med. Wochschr., 12,

28 94 SANFORD S. ELBERG [VOL MIKA, L., GOODLOW, R. J., VICTOR, J., AND BRAUN, W Studies on mixed infection. I. Brucellosis and Q fever. Proc. Soc. Exptl. Biol. Med., 87, MUSCHEL, L. H. AND MUTO, T Bactericidal reaction of mouse serum. Science, 123, MUSCHEL, L. H., CAREY, W. F., AND BARON. L. S Formation of bacterial protoplasts by serum components. J. Immunol., 82, NOGUCHI, H fuber gewisse chemische Komplementsubstanzen. Biochem. Z., 6, NOSSAL, G. J. V. AND LEDERBERG, J Antibody production by single cells. Nature, 181, NYKA, W Enhancement of resistance to tuberculosis in mice experimentally infected with B. abortus. Am. Rev. Tuberc. Pulmonary Diseases, 73, OAKBERG, E. F Constitution of liver and spleen as a physical basis for genetic resistance to mouse typhoid. J. Infectious Diseases, 78, POLICARD, A Morphology and physiology of the reticulo-histiocytic cell. In Physiopathology of the R.E. system. Blackwell, Oxford. 77. POLLACK, A. D. AND VICTOR, J Phagocytosis-augmenting factor produced by leucocytes. Proc. Soc. Exptl. Biol. Med., 89, PULLINGER, E. J The influence of tuberculosis upon the development of B. abortus infection. J. Hyg., 36, RAFFEL, S Mechanism involved in acquired immunity to tuberculosis. Ciba Foundation Symposium. Exptl. Tuberc. Bacillus and Host, 1955, RALSTON, D. J., BAER, B., LIEBERMAN, M., AND KRUEGER, A. P Staphylococcal virolysin, a phage-induced lysin. Its differentiation from the autolysin of normal cells. J. Gen. Physiol., 40, RALSTON, D. J., BAER, B., LIEBERMAN, M., AND KRUEGER, A. P Lysis from without of S. aureus K1 by the combined action of phage and virolysin. J. Gen. Physiol., 41, REPASKE, R Lysis of gram-negative organisms and the role of Versene. Biochim. et Biophys. Acta, 30, ROGERS, D. E. AND TOMPSETT, R The survival of staphylococci within human leukocytes. J. Exptl. Med., 95, Ross, J The pathogenesis of anthrax following the administration of spores by the respiratory route. J. Pathol. Bacteriol., 73, ROWLEY, D Bactericidal activity of macrophages in vitro against Escherichia coli. Nature, 181, SBARRA, J. A. AND KARNOVSKY, M. L The biochemical basis of phagocytosis. I. Metabolic changes during the ingestion of particles by polymorphonuclear leukocytes. J. Biol. Chem., 234, SHAFA, F. AND SALTON, M. R. J. Nature, Personal communication. 88. SHEPARD, C. C Phagocytosis by HeLa cells and their susceptibility to infection by human tubercle bacilli. Proc. Soc. Exptl. Biol. Med., 90, SHEPARD, C. C Comparison of the growth of selected mycobacteria in HeLa, monkey kidney, and human amnion cells in tissue culture. J. Exptl. Med., 107, SHEPARD, C. C Staphylococcus a'ureus in human cells in tissue culture. Influence of serum factors. Federation Proc., 18, SKARNES, R. C. AND WATSON, D. W Characterization of leukin: An antibacterial factor from leucocytes active against gram-positive pathogens. J. Exptl. Med.. 104, SMITH, H., AND KEPPIE, J Mechanisms of microbial pathogenicity. Symposium Soc. Gen. Microbiol., 5, STXHELIN, H., KARNOVSKY, M., AND SUTER, E Studies on interaction between phagocytes and tubercle bacilli. J. Exptl. 94. Med., 104, STXHELIN, H., SUTER, E., AND KARNOVSKY, M Studies on the interaction between phagocytes and tubercle bacilli. I. Observations on the metabolism of guinea pig leucocytes and influence of phagocytosis. 95. J. Exptl. Med., 104, STXHELIN, H., SUTER, E., AND KARNOVSKY, M Studies on interaction between phagocytes and tubercle bacilli. II. Action of phagocytes upon C14-labelled tubercle bacilli. J. Exptl. Med., 104, STINEBRING, W. R. AND KESSEL, R Continuous growth of Brucella abortus in mononuclear phagocytes of rats and guinea pigs. Proc. Soc. Exptl. Biol. Med., 101, SUTER, E Multiplication of tubercle bacilli within mononuclear phagocytes in