/ J The R61e of the Monocyte in Tuberculosis* BY R.S. CUNNINGHAM, F. R. S~BIN, 5. SUGIYAMA AND J. A. KINDWALL* From the Depbtment of Anatomy, The Johns Hopkins University INTRODUCTION T HE study of immunity may be defined as the search for the mechanism which the body itself has evolved for combating dif- ferent invading organisms. It has long been known that certain organ- isms, such as the diphtheria bacillus, liberate a toxin by some vital activity, secretion or excretion, against which the animal reacts by producing an anti- toxin. Moreover, notwithstanding the fact that neither such toxins nor anti-toxins have been analyzed chemi- cally, some of the diseases of this type have nevertheless been controlled. It is also well known that the majority of bacteria harm the body not only by the production of exo-toxins, but by other activities as well. It has been found that certain organisms produce the so-called endo-toxins, instead of exo-bxins, and to them the reac- tion of the body is much more complex, consisting in the production of agglu- tinins, precipitins, bacteriolytic sub- stances, etc. These different sub- stances have likewise not yet been analyzed chemically; nevertheless, by 1 Assisted by a grant from the Research Committee of the National Tuberculosis Association. *Mr. Kindwall's participstion in this work was made possible by a grant from the Henry Strong Denison Medical Foundation. utilizing the reactions of the animal body to such infections, certain effec- tive immune sera and effective vac- cines have already been produced. In the case of tuberculosis, it is quite clear that the human body has a marked power to produce an im- munity, since pathologists have shown such a high percentage of healed tu- berculosis, but we have as yet no di- rect control of the production of this immunity. We are thus forced to conclude that there is some factor in this particular mechanism that the body has evolved, which has so far escaped our analysis. We are now presenting as a new factor in the study of tuberculosis the concept that it is a disease which affects primarily a single strain of cells; namely, the monocytes. Sabin, Doan and Cunningham (27) showed that the epithelioid cell, the charac- teristic cell of the tubercle, is a modi- fied mono&e. We are now present- ing evidence to show that the infec- tion of tuberculosis causes an over- production of the monocyte, including all of its stages, namely the reticular cell, the typical monocyte and its two derivatives, the epithelioid cell and the giant cell; that the tubercle bacil- lus so alters the cytoplasmic activity of the monocyte that the cell becomes a suitable medium in which the bacil- 231 232 Cunningham, Sabin, Sugiyama and Kindwall lus can live and multiply, or, in other words, that the organism of tubercu- losis becomes a parasite within the cell; and that the marked overproduction of monocytes in the connective tissues is correlated, in the acute phase of the disease, with an increase in mono- cytes in the circulating blood. This new concept of the essential nature of the disease and the observa- tion that it is possible to correlate the progress of the disease with changes in the circulating blood, we believe, opens up a new experimental attack on tu- berculosis, and we think it quite prob- able that the fact that the organism of tuberculosis can live as a parasite within the monocyte may mean that the bacillus is thereby protected to some degree from the usual reactions of the body against it. Thus, our failme to obtain an effective immune serum may be the result of the intra- . cellular nature of the infection, and therefore may not indicate a fun& mental inability of the tissues to pro- duce antibodies of sufhcient power. In this study we propose, there- fore, first to present the observations that the bacillus of tuberculosis has a remarkable power of stimulating the production of new mono&es, and then of causing the differentiation of the mono&es into a specialized form, the well known epithelioid cell; and second, to show that the analysis of these activities of the tubercle bacillus forms an essential part of the analysis of the various mechanisms involved in the biological reactions of the body in this disease. In presenting this work we take great pleasure in expressing our deep appreciation of the assistance which we have received from Dr. William Charles White, Chairman of the Re- search Committee of the National Tuberculosis Association. Through- out the course of the investigation he has not only given us encourage- ment but has had a real share, through his critical judgment, in the interpre- tation of our results. In particular, it was his suggestion that the tubercle bacillus lives as a parasite within the epithelioid cell. It is likewise a pleasure to thank Mr. James Didusch, artist of the Car- negie Institute of Embryology for the drawings of the living cells. DISCRIBfINATION OF MONOCYTES FROM CLIISMATOCYTES The theory that tuberculosis pri- marily &e&s one strain of cells, namely, monocytes, is based on two concepts; first, that the mononuclear cells of the connective tissues can be separated into two distinct strains of cells, clasmatocytes and monocytes; and second, that the mono&e be- comes the epithelioid cell charac- teristic of the disease. These two concepts were presented by Sabin, Doan and Cunningham (27). These authors concluded that the phagocytic cells could be separated into two dis- tinct strains, the clasmatocytes and the monocytes, on the basis of two lines of evidence, tist, that they had a different embryological origin, and second, that they were both morpho- logically and physiologically different. They believe that the clasmatocytes are derived from endothelium; there- fore, in this concept, it is wholly ob- vious that the clasmatocyte is the cell which has been called the "endothe- lial leucocyte" by Mallory and the R6le of Monocyte in Tuberculosis 233 "endothelial phagocyte" by some of his school. It was shown by Sabin (23) that the Erst white cell to circulate in the blood, as seen in the living blastoderm of the chick, was derived from the endothe- lial wall of the blood-vessels. These endothelial derivatives appeared on the third day of incubation and she interpreted these cells as mono&es, an interpretation which we no longer believe to be correct. Of the correct- ness of the observation that the first white cell to circulate in the blood- stream is a phagocytic cell, derived from endothelium, we have no doubt, but we now believe that these freed endothelial cells do not become any of the detitive white blood-cells, or in other words, that these cells are not identical with mono&es. The same observation and the same in- terpretation, namely, that these pha, gocytic cells of endothelial derivation are transient in the blood-stream, was made by Maximow (15) on mamma- lian embryos. The early occurrence of the endothelial phagocytes in the blood-stream of the chick embryo has now been amply con&n& by one of us (Sugiyama) working in this labora- tory. These studies, in which it is possible to see a type of cell, fully iden- tified by means of the suprsrvital technique with the clasmatocyte, or endothelial phagocyte of the tissues, actually arise from the endothelial lining of a vessel, we regard as convinc- ing proof of the original production of the clssmatocyte by endothelium in the embryo. Recently, Herzog (12) has observed the same pmcess in the vessels of the tongue of the living adult frog. These observations, from studies of living tissues, combined with the work of Mallory and his school, make a large body of evidence in favor of the endothelial origin of clasmatocytes. On the other hand, Cunningham, Sabin and Doan (5) have presented evidence that the monocyte does not arise from endothelium but rather fmm the reticular cell, a primitive embryonic rest, in common with the other white blood-cells. This reticu- lar cell will be discussed later in connection with the effect of tuber- culosis on the tissues. The analysis of the morphological and physiological differences between clasmatocytes and monocytes has dt+ pended on modern methods for study- ing living cells, specifically on the use of- the so-called vital and suprrlrvital staining. The phagocytic mononu- clear cells of the connective tissues have been more generally regarded as one group, identified with the clas- matocyte, which was the tist strain to be analyzed with the newer methods. The discovery of the clasmatocyte itself came from following the reac- tions of the cells of the connective tissues to the injection of certain dyes into the blood-stream of the living animal. This so-called vital staining was inaugurated by the work of Rib- bert (21) with lithium carmine, of Rouffard (2) with isamine blue and of Goldmann (11) with pyrol blue. In the analysis of the experiments which have been made to test the reactions of living cells, it is possible, in a very general way, to classify the substances used into five groups. First, there are certain true solutions; second, colloidal suspensions of par- ticles too small to be visible with the highest powers of the micmscope, chmningham, Sabin, Sugiyama and Kindwall such as the benzidine dyes; third, insoluble substances with particles large enough to be seen with the mi- croscope, such as carbon in suspen- sion in India ink; fourth, bacteria; and fifth, red and white blood-cells in suspensions. In regard to true solutions, such as vital neutral red, we have had no evidence that any such dye in solution enters the nu- cleus of a living cell, nor are we sure that the living cytoplasm itself re- acts to such dyes; but with fixed tis- sues, certain basic dyes, for example, do enter nuclei and react with their chromatin. If such a reaction be a chemical combination, then we might define it as "true staining." The reaction of living cells to the bensi- dine dyes is quite a different phenom- enon from any such `%rue staining," because these dyes, which have been injected into the blood-stream, in colloidal suspension of ultra-micro- scopic particles, appear within the cells in quite large masses, readily visible even at a low magnification. We do not know in what form such dyes actually enter the cells, but the reaction of the cells, has been called "vital staining." In general the cells in which these dyes have been found in the form of particulate matter have been divided into two classes in accord- ance with their reactions to such dyes, first, into those that show the dye in Cne particles dispersed throughout the cytoplasm, and second, those that have the power of a non-specific agglu- tination of the substances into very large masses. The liver cell is an ex- ample of the former and the clas- matocyte is the most conspicuous example of the latter. The reaction of cells to large particles such as those of carbon in India ink is, in general, similar to vital staining, and, in the case of the clasmatocyte, practically identical. Through this character- istic reaction of clasmatocytes of agglutinating insoluble particles of dyes,~ `this strain of cells was sepa- rated out of the group of the cells of the connective tissues as a type with marked phagocytic power, and on account of this power it was called the "macrophage" (Metchnikoff). One of the most significant points so far gained in the study of living cells is a further insight into the method by which cells deal with foreign material. This point is of great importance from the standpoint, first, of the separation of the mono- cytes, from the clasmatocytes, and second, in relation to the analysis of the special relation of the tubercle bacillus to the monocyte. When vitally stained cells, that is, cells which, while still in the animal body, have stored insoluble particles of dye, are treated with supra-vital stains, we obtain evidence regarding the mechanism which the cell uses in deal@g with foreign particles. Ship ley (29) took films of fresh connective tissues containing clasmatocytes, which had been chronically stained by repeated injections of trypan blue, and treated them with neutral red. He found that the agglutinated masses of trypan blue were surrounded with fluid which took up the neutral red. It is this fluid, containing the agglu- tinated dye, which is called the "vacu- ole of phagocytosis" by some authors or the "segregation apparatus" by Evans and Scott (9). This combina- tion of the two techniques, chronic vital staining followed by the supra- R61e of Monocyte in Tuberculosis vital use of neutral red, gives the best method of analyzing these vacuoles. We have been making repeated obser- vations in this way. In studying such bits of living tissue one often has the opportunity of watching the pene- tration of the neutral red into the deeper cells of a mass and in such preparations it is quite clear that the particles of the insoluble dye have been surrounded by fluid by means of some activity of the cell, and that it is this fluid which actually stains with the neutral red. The particles of blue are thus readily seen through the red- stained fluid. This clear picture of the two different colors is not retained long, for soon the entire vacuole be- comes dark red or purple. It is not clear whether the masses of the in- soluble dye, the trypan blue or the carmine, are merely particles of dye alone or whether the particles of dye have adhered to some substance in the cell, but the presence of the fluid around them and the staining of this fluid by neutral red can be quite con- vincingly demonstrated. Into ,this fluid the neutral red penetrates and the color of the neutral red often varies quite considerably from a sal- mon red, through the scarlet of the acid reaction of the dye, to a deep maroon color. As the red color be comes more dense, the particles of phagocytized dye within the vacuole gradually become more and more obscured. It is quite clear, we think, that neither the original phagocytosis of the insoluble dye, nor the subse- quent staining of the fluid of the so- called vacuole in which the dye be- comes segregated, is a true staining of protoplasm, for neither the phago- cytized dye nor the fluid which the cell secretes around it is true proto- plasm. After such a study of clasmatocytes it becomes easy to follow the same process by staining the living cells with neutral red alone and to analyze the vacuoles that surround the debris which these cells take up during their physiological activities. Clasmato- cytes supra-vitally stained with neutral red, after they had been experimentally loaded with trypan blue or with foreign blood-cells, or physiologically loaded with debris, were shown by Sabin, Doan and Cunningham in their Plate 1 (27). As will be seen in this plate, the reactions are all the same, consist- ing of the surrounding of a foreign body by a fluid stainable with neutral red which reacts as an indicator toward the contained substances. Clasmato- cytes are characterized both by the very large size of the foreign bodies which they engulf, by their marked power of agglutinating such material, and by the fact that the engulfed material is distributed in the cytoplasm wholly without pattern. The reaction of monocytes, on the other hand, to supra-vital staining brings out a constant and character- istic pattern. In the monocytes there are certain fine bodies that stain with neutral red which occur in a rosette around the centrosphere, and the pres- ence of this rosette limits the zone for the storage of phagocytized material to the periphery of the cell. We have dealt, at some length, with the vacuoles of phagocytosis and their demonstration with neutral red in cells which were known to have phagocytized foreign particles, but it must now be made clear, for the complete study of monocytes, that the Cunningham, Sabin, Sugiyama and Kindwall reaction of living cells to neutral red is by no means confined to the stain- ing of the fluid secreted by cells for the purpose of dealing with phago- cytized material. Not every sub- stance stained by neutral red is to be considered as a vacuole of digestion. The first example of a substance which reacts to neutral red, other than the fluid of the segregation apparatus of cells, is the so-called reticular sub- stance of immature red cells. We do not regard this reaction as compar- able to the other reactions toward supra-vital staining because, as Pap- penheim pointed out, such a staining of red blood-cells is the result of a marked damage to the cell. Schilling (23) and Key (13), have demonstrated that the dye precipitates and clumps the basophilic substance, which, in the living cell, is uniformly distributed throughout the cytoplasm. Thus, this reaction is not true supra-vital stain- ing because it appears as a marked dis- tortion of the structure of the red cell, and it seems certain that no such distortion takes place in the reaction of the white cells to these dyes. The reaction of the red cell varies from the massive precipitation in the megalo- blast, through the well known stages of the reticulation, to a final stage in which only two or three droplets of substance react to the supra-vital dye (see illustration given on Plate V, Doan, Cunningham and Sabin (6)). The second reaction to supra-vital dyes is shown by the spec& granu- lations of certain cells. These granules are probably some specific type of material produced by cyto- plasmic activity and are not parts of the cytoplasm itself. Such granules are the neutrophilic, the basophilic and the eosinophilic granules of the granulocytic leucocytes. Another type of granule which reacts charac- teristically to neutral red is that found in the islet cells of the pancreas, as was discovered by Bensley (1). It is therefore evident that a very great range of substances within cells react to neutral red; it is, however, in no wise settled that all of these reactions are similar in their mechanism. If we consider first the neutrophilic granules, it is clear that they are plainly visible in the living cell without any dye. With neutral red we find that there are variations in the reaction of these granules, both in comparable cells in different animals and in the same animal at different stages in its development. The early granules of the neutrophilic myelocytes stain in- tensely in neutral red; in the human neutrophilic leucocytes the neutro- philic granules stain throughout the life of the cell, up to the time of the non-motile phase, when the granules swell, become highly refractive and entirely unstainable. In the rabbit, on the other hand, although the neu- trophilic granules stain well in the myelocytes, and in many of the leu- cocytes, occasionally there is but a slight reaction of these granules to the stain even during the most active phase of the leucocyte. The neutro- philic granules of the dog's leucocytes are very tiny and do not react to neu- tral red. All these facts show that there is some chemical evolution of the neutrophilic granules within the life of the leucocyte; nevertheless, there is a distinct substance produced by the cell which we see in the form of the neutrophilic granules and this substance is probably bound up in the R61e of Monocyte in Tuberculosis 237 specific functions of the leucocyte. stained mono&es, the actual analysis We are unable to state whether the of the substance that reacts to neutral dye actually enters the neutrophilic red has not been easy, but we think granules, as we think that it does that the study of the modified mono- enter the fluid of the so-called vacuoles, cytes of tuberculosis, the so-called or if it occupies the interphase between epithelioid `cells, has aided in this the granule and the surrounding cy- analysis. It is well known that Nae- toplasm. It is entirely clear, however, geli first analyzed the monocyte' as a that the neutral red is actually a chemi- separate group of the white cells by cal indicator in connection with the showing that the cell, which E&rlich three types of the speciEo granules of had called the transitional cell, was the leucocytes; with the neutrophilic entirely difIerent from the neutrophilic granule it gives an intermediate reac- leucocyte on the one hand, and from tion, with the eosinophilic granule large lymphocytes on the other, in the reaction is toward the alkaline re- that it contains very fhie azurophilic action of the dye, while with the baso- granules which are found in neither philic granule the reaction is definitely of the other two types of cells. In the brilliant scarlet color of the acid the supra-vital technique it was shown reaction of the dye. The neutro- by Sabm ( (24), see Fig. 4) that there phi& leucocyte is likewise a phago- are two types of substances in the cytic cell; that is to say, it has a frmc- living mono&es that react to neutral tion which may or may not be ssso- red, very tiny bodies, which in the &ted with its specific gram&&ion. living cell are arranged in a rosette This cell also reacts to phagocytized around a clear spot, the centrosphere, material by developing vacuoles and larger bodies, which are in the around the debris. In case of the neu- periphery of the rosette. The tiny trophilic leucocyte, there are many bodies have a characteristic salmon- times when the cell has no vacuoles, colored reaction when stained with and again many times when they are neutral red. We have found this present. These vacuoles show an color both constant and character- entirely different color in neutral red 1 istic. The larger bodies, on the other from the neutrophilic granules, they hand, we are quite confident, are are decidedly more scarlet, that is, true vacuoles of digestion, comparable more toward the acid reaction, are of the vacuoles of the clasmatocyte; slightly larger than the granules, and they vary markedly in size and some- vary markedly in size, so that they are what in color; though they never never likely to be confused with the show as wide a range in color, as do stained neutrophilic granules. Thus, the corresponding vacuoles of the in the neutrophilic leucocyte, there are clasmatocyte. We are also convinced two entirely different subst&Ktis which - that the fine &iticles that stain with stain with the neutral red; lirst, the neutral red in the living cell are not so-called neutrophilic granules; second, the ssme as the azurophilic granules the droplets of fluid which we call the of tied films, since the azurophilic vacuoles of digestion. granules are scattered without pattern In the case of the supra-vitally in the cytoplasm. In the living, un- Cunningham, &bin, Sugiyama and Kindwall stained monocyte, the fine particles of the rosette are just visible, because they have a very low index of refrac- tion; the larger vacuoles, when present, are plainly visible, for they have a high index of refraction. We think that the fine particles of the rosette are not seen as granules in the films of blood fixed in absolute alcohol (Wright's blood stain), but they are retained in formalin; nor are the azurophilic granules of the fixed cells visible in the living state. The most d&At point in the analysis of the monocyte is that under certain con- ditions the entire cell may be occu- pied by the larger stainable bodies. Such monocytes look like clasmato- cytes and the development of mono- cytes into this form in tissue cultures of blood has convinced Lewis, Willis and Lewis (14) that clasmatocytes and mono&es are a single strain of cells. Such a monocyte was shown by Sabin (Fig. 5, (24)) from a case of Malta fever, in which there was a very marked increase in the monocytes of the circulating blood and in which all of the monocytes became markedly vaouolated after the injection of an autovaccine. In this state the vacu- oles seemed to replace the tier par- ticles entirely. Another such mono- cyte wtw shown by Sabin, Doan and Cunningham (Fig. 19, (27)). In the former the centrosphere was still evident; in the latter the centrosphere was obscured, but the vacuoles still showed some evidence of being in a group instead of being diffusely scat- tered. In such a cell it is not clear what has become of the finer particles so characteristic of the cell; the ques- tion then arises as to whether they have disappeared or have enlarged into vacuoles. In the ma&ion of the -monocytes to tuberculosis, on the other hand, it is the fine bodies of the charaderistic rosette that increase in enormous proportions and it is this reaction which, `as we shall now dem- onstrate, indicates that the epithelioid cell and the resulting giant cell are characteristically derivatives of the monocyteo rather than of any other type of cell. METHODS General methods The purpose of the experiments which we are reporting in this paper has been to analyze the relationship which exists between the monocytes of the blood and of the tissues, and the changes which occur, especially with regard to the monocytes, in the course of acute experimental tuber- culosis. We do not feel that these experiments represent more than a very meager attempt to open up the question of the varying changes which take place in the blood cells, especially the monocytes and lymphocytes, in tuberculosis, The full exposition of this most important subject must await much more elaborate study than we have been able to carry out up to the present time. In the course of this study we have used about 75 rabbits. The.organisms which we have used were cultures obtained from the Dows laboratory of tuberculosis of the Johns Hop- kins Hospital and have been numbered Bl and H37 respectively. The organ- ism B1 was an organism of bovine tuberculosis and hasbeen used in the majority of the experiments; while H37, an organism of human tubercu- R61e of Monocyte in Tuberculosis losis obtained from the same labora- tory, has been used in only a few. The method which we have used throughout these experiments has been to remove the bacilli to a watch crystal and weigh. The weighed bacilli were transferred to a sterile mortar and ground with a little saline, more being added as the emulsion was prepared. After about 5 to 10 minutes' grind- ing the suspension was filtered through sterile cotton and then centrifuged. Samples were removed from the tubes until it was shown that practically all the masses had been thrown down. A standard loop of the mixture was then spread on a slide lover an area about 1 cm. square and the average number of bacilli per oil-immersion field determined. We are well aware that this method is not even approxi- mately exact, but it at least ensures that the eventual suspension contains no large clumps or masses of the bacilli. We have generally used an emulsion containing from 15 to 50 organisms to the oil-immersion field and the sus- pension was, in most instances, given intravenously, although a few ani- mals were inoculated intraperitoneally. Sabin (24) found that, in a case of Malta fever, there was a large increase in the number of the circulating mono- cytes, and an additional increase in their phagocytic activity as indicated by their staining with neutral red. With this concept in mind it occurred to us that perhaps B. abortus, an or- ganism closely related to the bacillus melitensis, might bring about a stimu- lation of the monocytes in our experi- mental animals and thus supply us with a mechanism for analyzing the results of infection with tuberculosis in the cese of previously stimulated animrtls. Throughout these experiments we have taken blood counts at as close intervals as was possible, many of the experimental animals having been counted daily for periods of six to seven weeks. This factor of making daily supra44tal differential counts, as well as counts of the total white blood-cells, has rendered the utiliza- tion of a larger series of animals tech- nically impossible so that, while we recognize that our series must appear small to those workers studying aller- gic, serological and immunological reactions, nevertheless, it was as large as it was possible for a small group of workers to carry through. And furthermore, our res ts have been so L striking and so easy lady-- into speciiic groups, with regard to the monocytic reactions, that it has seemed fully justifiable to consider the series quite large enough to make reliable conclusions possible. We have counted the blood of all experimental animals several times before injections and, whenever pos- sible, this period of preliminary count- ing has been extended to several weeks' duration. It is a customary opinion that the blood counts in rabbits vary much more widely than in the other animah and we were inclined in our earlier experiments to concur in this opinion, but we found, when care was taken to have such a dilatation of the vessels of the ear that the blood flowed freely, that the variations in the total counts of the blood-cells of the rabbit were reduced to within limits not greatly in excess of those which we have demonstrated to be normal in the human blood (Sabin, Cunning- ham, Doan and Kindwall (25)). Such a dilatation of the ear veins can be easily obtained if the ear is stroked 240 Cunninghi,m, Sabin, Sugiyama and Kindwall or gently tapped with the back of a knife. Throughout the study on the blood, we have been careful to take the blood for the total count at the same time at which we took the specimens for the supMGvital differentials. All of the differential counts have been made with the supra-vital technique; smears fixed in Wright's stain and also in Ziehl-Nielsen for tubercle bacilli have been made in special instances when specific observations were desired. The autopsies have been controlled by careful studies of lungs, spleen, bone-marrow, omentum and other tissue, in special cases, made upon supra-vital preparations, according to the methods described by Sabin, Doan and Cunningham (27). Sections were also prepared from tissues fixed in the routine manner and stained both by the ordinary histological stains and for tubercle bacilli. The supra-vital technique The method we have used was de- veloped by Sabin (24). The essential point in the technique is to obtain a perfectly even, thin film of a vital dye or combination of dyes on a slide, which is to be used for a preparation of fresh blood. In this way the dyes, to wh'ch the living cells react, dissolve in the normal plasma as the film is made, so that the cells are not sub- jected to any accessory fluids. For the technique it is first essential to remove all traces of grease from the slides and covers. This is done by the usual technique. They are kept in concentrated sulphuric acid to which a few crystals of potassium bichro- mate have been added for 3 to 4 days; then they axe rinsed thoroughly in running tap water, preferably hot, and transferred to distilled water and then 80 per cent alcohol. They are wiped from the alcohol with cheese- cloth and- flamed thoroughly to re- move the last traces of grease. The slides are then ready to be flooded with the stain. We have found vital neutral red and a combination of vital neutral red and vital Janus green the most use ful stains. The neutral red alone does not inhibit motility and all of the normal blood-cells react to it charac~ristically. The addition of Janus green, which stains the mite- chondria, does check motility, but is of especial. value in discriminating immature blood-cells, the cells of or- gans and the cells of the connective tissues. Therefore, for the routine blood-counts with relatively normal cells, we use the neutral red alone, but for all of the studies of abnormal blood and of the fresh tissues from the autopsies we have used the double stains. The films of stain are made as fol- lows: we keep a saturated stock solu- tion of vital neutral red in absolute alcohol; from this a dilute solution is made by adding from 20 to 30 drops of the saturated solution to 10 cc. of absolute alcohol; the strength of the stain is best judged by the color, which is a rose red; the exact strength must be tested with the material to be stained, in fact, the amount of stain must vary with the number of cells that take the dye in a given prepara- tion. Any staining of the nuclei is a sign that the stain is too strong. The double stain is made by taking 1 cc. of the dilute neutral red and Rdle of Monocyte in Tuberculosis 241 adding from 3 to 6 drops of a satu- rated solution of vital Janus green in absolute alcohol. We have found that 3 drops of Janus green per 1 cc. of dilute neutral red is the correct strength for the cells of normal blood, but for preparations from tissues more Janus green should be used, up to 6 drops. Beyond this strength the cells are killed. The slides are prepared with the dyes as follows; after they have cooled from the flaming, they are held in a horizontal position and flooded with the stain, which is quickly drained back into the bottle; the stain must neither be allowed to stand long on the slide, since the alcohol will evapo- rate, nor to touch the Angers in this process, lest a little grease be added to the solution. The slides are then placed upright until they are dry. If the flhn of stain is uneven, some of the cells will be killed, and the tech- nique is in no sense differential for dead cells. The stain can be used over and over unless it becomes greasy or full of dust. The preparations of fresh blood are made by the usual technique of ob- taining the drop on a coverslip and inverting it on the slide as soon as the blood has spread, the coverslip must be rimmed with Vaseline of a high melting point; we have used salvoline. The preparation is then placed and studied in a warm box, kept at 37oC. For preparations of the tissues the technique varies according to the or- gan to be studied. For the lung, liver and kidneys we scrape a freshly cut surface of the organ gently and mount the material as if it were a blood film. It is important to have an amount of tissue so small that it will spread out in a film practically as thin as a blood film; this is impor- tant for two reasons, f&t, because the cells are then reached by the dye, and second, because the necessity of using accessory fluids is avoided. Such preparations must also be sealed with Vaseline. For the free cells of lymph glands, spleen and bone-marrow we have found that better preparations can be made by drawing the material up into capillary pipettes from the anaesthetized animal in which the circulation is intact. In studying the cells of the diffuse connective tissues, such as subcutaneous tissue we have seldom found enough fluid present for our preparations and in this case it has proved to be better to make an artificial oedema by the injection of neutral red (1 to 10,000 in Ringer's solution) and to mount bits of the resulting gelatinous tissue. We have found that it is the cells of the circulating blood especially which are the most sensitive to accessory fluids and consider that motility of cells can never be correctly judged when they are studied in artificial solutions. From the autopsies of our animals we have made the supm-vital studies of the tissues of the lungs, liver, kidneys, spleen, lymph glands, bone-marrow and omentum and of any other tis- sues that have shown signs of tuber- culosis in the gross material. We have found these studies of the ut- most value and consider that they permit a much better diagnosis in certain particulars than can be ob- tained from fixed sections; in the first place, the supra4Gtal technique is differential for cells that cannot be discrimiiated in sections, and secondly, certain structural points such as the Cunningham, Sabin, Sugiyama and Kindwall relative independence of cells, for example, whether monocytes are struc- turally bound together in tubercles or actually free in the tissues, are more easily determined by this tech- nique than in sections of fixed tissues. EXPERIMENTAL DATA E&et of tubercubsis on the monocyte in the f&dating blood and in the tkmes The immediate effect on. the mono- cyte of the ingestion of the tubercle bacillus is an inhibition of the motility of the cell. We j'udged this because we have found, in following the blood of rabbits which had been infected with tuberculosis, that a short time after the infection there appeared in the blood monocytes which had ap parently lost their power of motility. These mono&es were quite different from the normal cells and we have &led them "modi6ed monocytes," since we are unable to say exactly in what way they have been changed. These cells were large, usually round and had apparently lost their power of motility; they stained intensely in neutral red and had the stained vacu- oles scattered throughout the periph- eral zone around the rosette. Such a cell is shown in Fig. 1, from Rabbit TB 49. This cell was somewhat ir- regular but showed no locomotion on the slide. We have demonstrated the tubercle bacilli within such mono- cytes of the circulating blood by means of the Ziehl-Nielsen technique. We consider that the cessation of motility is a sign that the cell has been dam- aged. The interpretation of the in- crease in the stainable vacuoles as the very first effect of the bacillus within the cell is an important point. It is possible to explain this change in three different ways, as evidence of increased activity on the part of the cdl, as evidence of cellular injury, or as an indication that the cell is attempt- ing to compensate by increased cyto- plasmic activity for an actual damage of its structure. In the case of the vacuoles of the clasmatocyte and of the neutrophilic leucocyte, we are confident that they are functional structures. Every re- action of the monocyte in the develop ment of the large vacuoles may not be quite so clearly functional; it is frequently true that these vacuoles take longer to stain in our preparations than the vacuoles of the clasmatocyte, but we do not believe that this change is evidence of immediate or extreme in- jury or that the monocyte is quickly killed by harboring the tubercle bacil- lus within its cytoplasm. It is, how- ever, quite clear that both by its ac- tual presence in the cell and possibly by substances which it produces, which reach the cell through the circulation, the tubercle bacillus can profoundly modify the morphological appearance and the physiological activities of the monocyte. The next stage in the effect of tu- berculosis on the monocyte we have also seen in a cell of the circulating blood, namely, the very beginning of the formation of the epithelioid cell. Such a cell is shown in Fig. 2. This cell was drawn from the blood of Rabbit TB 43, 26 days after the in- fection of the animal. We have found that the effect of the tubercle bacillus in producing the epithelioid cell is very characteristic and consists in two thingsf first, the suppression of the R61e of Monocyte in Tuberculosis 243 vacuoles which are normally present in the periphery of the cell, and second, a most characteristic and enormous multiplication of the fine particles of the rosette. A cell comparable to the one of Fig. 2, but taken from the tissues, is shown in Fig. 3, This cell was from the liver of a tuberculous rabbit (TB 36). By this multiplica- tion of the fine bodies of the rosette of the monocyte, the rosette becomes the essential characteristic of the so- called epithlioid cell. It must be brought out very clearly that the pres- ence of granules, arranged around the centrosphere is not found in monocytos alone. All young granulocytes, of course, have a centrosphere, and at a certain stage, the stage in which the cytoplasm is well filled with the spe- cific granules, these granules are ar- ranged in radiating lines, thus accentu- ating the centrosphere; this is true of the neutrophilic, the basophilic, and the eosinophilic myelocytes; in the monocyte there is likewise a spe- cific substance, in. the form of fine granules, that makes the rosette in radiating lines around the centros- sphere and, in this type of cell, in contrast to the granulocytes, there is a marked permanence of the pattern of the rosette. The cell of Fig. 3 was small as compared with the more developed epithelioid cells.' It hap- pened to have two nuclei, thus showing the same tendency toward amitosis exhibited by the normal monocyte. We have not seen division in mono&es except by amitosis. The mitochondria characteristic of the peripheral zone of monocytes were obvious in the case of Fig. 3. The most striking and characteristic change in monocytes infected with the tuber- cle bacillus consists, then, in the mul- tiplication of the fine bodies of the rosette. If the monocytes in Plate 11 of Sabin, Doan, Cunningham (27) are compared, it will be seen that there is some variation in size in the fine bodies of the rosette; for example, in their Figs. 14, 15, 16, and 18, the fine bodies of the rosette are all small, whereas in the cell of Fig. 17 they are markedly larger. All of the cells on this plate were drawn. at the same magnification, so that the size of the granules can be compared. It is interesting to note that the fine bodies of the rosette in Fig. 18 are small, and this was a cell which had been stimu- lated to marked phagocytio activity. The cell in question had engulfed a red blood-cell and several white blood- cells. In the young epithelioid cell shown in Fig. 3, which was taken from Rabbit TB 36, there had been an in- crease in the number of the fine bodies of the rosette; they were at the same time slightly larger than the fine bodies of the rosette of the average normal monocyte, such as the ones already referred to and as the mono- cyte from normal human blood shown by Sabin (Fig. 4, (24)). In the cell of Fig. 3, rabbit TB 36, the centrosphere was obvious in the center of the rosette. As will be seen in the drawing, there is a slight tone of the dye between the granules; in some instances in the epithelioid cells we have found it difficult to tell whether this tone was due to a true staining of some substance between the granules or simply to an optical effect on account of the great number of the granules. It may also be true that this staining between the granules is a diffusion of the dye from the 24.4 Cunningham, Sabin, Sugiyama and Kindwall granules due to the gradual damage to the cell. The rest of the cytoplasm of this cell w&8 practically clear, ex- cept for the mitochondria. The rab- bit from which this cell was drawn showed very many of these young epithelioid cells from the liver, together with many large epithelioid cells and giant cells; in the lung of this animal we found comparatively few of the younger epithelioid cells, but, on the other hand, we found many that were much further differentiated, together with a considerable number of clumps of the unditrerentiated reticular cells. As the monocyte becomes more and more affected by the tubercle bacillus, the rosette becomes larger and the bodies which form it become smaller Such a cell is shown in Fig. 4 (Rabbit TB 33). This cell was found in a scraping from the cut surface of the lung. In this cell the rosette was so large that it almost completely filled the cytoplasm, leaving only a small peripheral zone. Such epithelioid cells occur, but a wider peripheral zone is more frequent. In the edge of this rosette was one vacuole which stained in neutral red and a single refractive body shown in white which we in- terpreted as fat. These bodies stain with Sudan III. Most of the epi- thelioid cells from the lung of this animal showed a marked development of these droplets of fat. The most striking thing about this cell, shown in Fig. 4, was the enormous multipli- cation of the fine bodies of the rosette. These bodies were slightly smaller than those shown in Fig. 3, but still were not as fine as those shown in Figs. 6 and 8. This great increase in the a& tual number of the Cne bodies of the rosette is the characteristic mor- phological change, as seen in supra- vitally stained films, which is brought about in the monocyte by the tubercle bacillus or its products. At the same time there is a very considerable sup pression of the larger vacuoles occupy- ing the periphery of the rosette. The cytoplasm around the rosette in this cell was like ground glass and con- tained no mitochondria. The next cell of the series, shown in Fig. 5 (Rabbit TB 16) was also a typi- cal epithelioid cell. It was also taken from the lung. This cell showed the very characteristic division of the cytoplasm into two zones, the rosette and the peripheral zone. There was a greater variation in the size of the small bodies of the rosette than of the other cells, and in the lower border there were a few bodies which were decidedly larger than the rest. The wide peripheral zone of this cell was markedly granular and very charac- teristic of many of the epithelioid cells. Nothing `m this granular peripherrtl zone stained with either the neutral red or the Janus green, but in this area there were two bacilli which were very characteristic. It is in this peripheral zone that cells and other particulate material that a monocyte has phagocytized are always seen, and when such cells or debris have been phagocytized, they are always to be found within stained vacuoles of di- gestion. On this account we stress the fact that there were no stained vacu- oles in the peripheral area of this cell (Fig. 5) ; the bacilli showed not the slightest staining reaction around them and they were seen moreover to shift their position slightly in the cytoplasm, possibly through some slight movement of the latter. We have now seen bacilli R6Ie of Monocyte in Tuberculosis 245 several times in the living cells and are convinced that the monocyte does not show any of the signs toward en- gulfed tubercle bacilli which ordina- rily indicate that the material taken in is being digested. So that it seems to us as most reasonable to assume that the bacilli remain alive and capable of multiplication. This forms one factor in `the evidence that leads us to suggest that the bacilli are harbored by these modified monocytes instead of being destroyed by them. A most marked rosette with the &rest division of the granules is shown in the cell of Fig. 6, which was obtained from the lung of Rabbit TB 39. The rosette in this cell was very sharply defined and was made up almost wholly of line bodies; furthermore, in this cell the bodies reached the maximum fineness in the cells we have seen. In the cell shown in Fig. 6, there were around the rosette a few small refractive droplets which we interpreted as fat and which we think probably indicate a beginning de- generation of the cell. In the periph- eral zone of this epithelioid cell there was a small red blood-cell, which we presume had just been taken in, be- cause its color was exactly like that of the surrounding red cells. There was not a trace of neutral red about this red cell. On the other hand, many of the epithelioid cells of this rabbit showed a small amount of debris in the peripheral zone, as evidenced by stainable vacuoles. Such cells in- dicate that the power of phagocytosis is not entirely suppressed in the epi- thelioid cells. In the edge of the rosette of the cell shown in Fig. 6, there were a few highly refractive bodies; they did not stain at all in neutral red. A specimen studied in Nile Blue Sulphate did not show any staining of these droplets. In many of the epithelioid cells of this animal the entire periphery of the cytoplasm was packed with these refractive droplets, but, in frozen sections these refractive bodies stained heavily with Sudan III and hence we have concluded that they are lipoids of some type or other. We have not, as yet, studied these refractive bodies more thoroughly, although this should be done, as they represent a most obvious and impor- tant change in the cell and one which we think is probably degenerative in character. A very large epithelioid cell is shown in Fig. 7, from the lung of Rabbit TB 38. This cell showed several vacuoles in the edge of the rosette, which clearly indicated that the cell had phagocy- tized some debris; in this cell there was only one of the lipoid bodies, shown in white in the edge of the ro- sette, but in the lung of this animal the majority of the epithelioid cells showed the fat droplets either filling the entire peripheral zone, leaving the rosette intact, or else filling the entire cell, as is shown in Fig. 9. -The last phase of the epithelioid cell tends either toward a fatty de- generation or toward the formation of a giant cell, which may also pass,into the same terminal phase of fatty dt+ generation. The cell of Fig. 7, Rab- bit TB 38, shows the very beginning of the fatty degeneration. The next cell of the series (Fig. 8) was taken from the lung of Rabbit TB 36 and shows that fat droplets first fill the periphery of the cytoplasm of the epithelioid cell. These droplets of fat increase in number until they Cunningham, Sabin, Sugiyama and Kindwall occupy the entire peripheral zone of the cytoplasm. They then go on increasing in number until they en- tirely obscure both the rosette and the nucleus of the living cell. The same stages of the development of refractive droplets can be followed in the giant cell. Just why some epithelioid cells undergo this extreme degeneration before there is any nuclear division, while others go on to various stages of the giant cells, we have not been able to determine, but this condition must he associated with the extent of the injury ing.ict.4 upon the cell by the bacillus or with the general physio- logical condition of the cell. Fig. 9 is of a cell in which the en- tire cytoplasm has become 6lled with the refractive droplets referred to above; this cell, while still alive, as shown by the fact that the nucleus did not stain with neutral red, never- theless, was probably in an advanced stage of degeneration. There was only a single nucleus visible in this cell, which indicated that there had been no progression toward the giant cell type. The cell of Fig. 8 had two nuclei. This division of the nucleus without a resulting division of the cell is the method by which we believe the giant cell of tuberculosis is formed from the monocyte. Again, it seems to us that this is further evidence that there is a marked change in the cytoplasmic activities of the monocyte in animals infected with tuberculosis. Sabin, Doan and Cunningham (27) have shown first, that the monocyte has a marked tendency to divide by ami- t&s, and secondly, that amitosis is to be defined as a condition in which nuclear division precedes the division of the centrosome; complete amitosis involves three processes in definite sequence, nuclear division, division of the centrosome, and subsequent di- vision of the cell. The division of the cell seems to be dependent on the pre- vious division of the centrosome; if this be true, we have an adequate con- cept of the sequence of events that give rise to the giant cell, namely, repeated nuclear division with inhi- bition of the division of the centro- some. We, therefore, suggest that, in general, giant cells of the Lang- bans type are derived from monocytes. It is thus clear that the effect of the infection of tuberculosis on the mono- cytes causes them to increase in size and to develop a very marked dif- ferentiation of the cytoplasm into two distinct zones, the zone of the rosette and the peripheral zone. When Elms made by scraping the freshly cut sur- face of a tuberculous lung are treated as blood films and stained with the Wright's blood stain, the division of the cytoplasm into these two zones is very marked. The central zone of the rosette stains a diffuse pink in eosin; thus the fine granules of the liv- ing cell that make the rosette seem to have been dissolved in the alcohol so that they no longer appear as dis- crete particles. The peripheral zone of the cell is markedly basophilic and has the same muddy blue color as the monocyte of the circulating blood. In these cells one can count about 59 to 60 of the azurophilic granules. Thus the relationship of the epithelioid cell to the monocytes is again brought out in the presence of the azurophilic bodies characteris- tic of that cell. These observations lead us to conclude that the fine bodies R&e of Monocyte in Tuberculosis 247 of the rosette of the monocyte are substances visible in the living cell; that they are not the same as the asu- rophilic granules of the tied films which appear in the monocyte after fixation in alcohol; that the fine bodies of the rosette are retained in formalin; that when they are very markedly increased in number, as they are in the epithelioid cells, they give to the cytoplasm an acidophilic reaction inwright's blood stain. This reaction is not seen in the normal monocyte of the peripheral blood, which we inter- pret as due to the smallness in amount of the substance in the normal cell as compared with the epithelioid cell. It may be, however, that the presence of this substance in the monocyte is the factor which makes the sharp difIerentiation between the very clear blue of the cytoplasm of the lympho- cyte in Wright's blood stain, and the muddy or smoky blue of the cytoplasm of the monocyte. One of the points by which the dis- crimination between clasmatocytes and monocytes was made by Sabin, Doan and Cunningham (27) was a Merence in origin of the two types. They obtained evidence which indi- cated that the clasmatocyte originally comes from endothelium, while the monocyte arises throughout the life of the animal from an undi&entiated, embryonic type of cell, the so-called reticular cell. Thus the monocyt.43 is derived by a process of maturation just as are all other types of white blood-cells. The above named authors (5) have been able to iden- tify this primitive, embryonic rest, the so-called reticular cell, in the liv- ing connective tissues, so that the type is no longer a hypothetical progenitor for the white blood-cells, but a cell which can be readily found and identi- fied. The name `Micular cell" is not very speci6c but the cell itself can be quite clearly defined both as to its appearance and in its location. This reticular cell was identified by Doan, Cunningham and Sabin (6) in fresh Nr.ns of bone-marrow stained with suprz+vital dyes. In bone-mar- row so simplified by experimental procedures that there were no cells in the marrow except fat, endothelium and these reticular cells, this discrim- ination was easy. The reticular cell in the living state shows a complete absence of differentiated structures, both in the nuclei and in the cyto- plasm. A small clump of such cells as seen in the living state could be drawn only as a mass with a definite but common outline and with a uui- form gray tone; the nucleus may not show at all in the living cell; but if such a cell or group of cells be watched, the nuclei gradually appear, probably as the cells die. There are no discrete granules whatever to be seen in the cytoplasm which has the uni- form appearance of ground glass. The cytoplasm of the reticular cell appears to have a much more definite tone, however, than the clear part of the cytoplasm of a squamous epithelial cell, for example. In. Wright's blood stain the cytoplasm and nucleus of the reticular cells show a practically uniform- but very faint basophilic reaction. These cells can be found in great numbers in every Elm made from scrapings obtained from the cut sur- face of a lymph gland and studied either supra-vitally or stained with any methylene blue-azure mixture. The lack of a striking structure is what 248 Cunningham, Sabin, Sugiyama axid Kindwall has made the reticular cell remain a hypothetical type of cell for so long. The nucleus has little chromatin, the nuclear border is never as sharp as in a lymphocyte or in an epithelial cell; the cytoplasm has no specific structure by which it can be discriminated from other cells. There are no mitochon- dria whatever; this absence of mito- chondria sharply discriminates this cell both from the primitive white blood-cells and from the lymphocyte (see.. Cunningham ,._. ?4.@ and .Doan (5)). This reticular cell is the pro- genitor of the primitive white blood- cell which it becomes as soon as mite- chondria develop in the cytoplasm. There ia no other cell in the body with such a lack of discriminating features; that is to say, it is the most undilIer- entiated cell of the adult organism. The reticular cell is most easily formd in normal tissues in a scraping from the freshly cut surface of any lymph gland, because there are more of them in lymph glands than any- where else; it is also readily obtained from the spleen. It is much more diflicult to find these cells in normal bone-marrow because the marrow is rio crowded with myelocytes. The reticular cell can be found in very small numbers in a fresh preparation made by scraping the freshly cut sur- face of any normal lung. We have found it more easily from the septa of the lung than in subcutaneous tissue; however, if very tiny bits of fresh connective tissues are mounted on a 6hn of neutral red and Janus green, of sufficient strength, so that all of the more differentiated cells, the fibro- blasts, the clasmatocytes and the various types of the white blood- cells are well stained, these very primi- tive reticular cells, which do not react at all to either of the dyes, can be found. From the supra-vital studies of the material of tuberculous animals at autopsy, we have found that thecon- dition of the lungs has varied markedly in the relative proportions of the dif- ferent stages of the mono&es. For example, in the lung of Rabbit TB 39 (Protocol on page 255, Chart 3) the cells in the fresh scraping seemed to have. -come wholly.. from the. septa because there was none of the charao- teristic elastic tissue from around the air sacs and also no epithelium. The most striking thing about this tissue was the large masses of the primitive reticular cells which were present in the scraping; some of these masses contained only three or four cells, but others filled the whole field under the oil-immersion lens. Most of them had no granules whatever; a few had some granules, like those of Fig. 5, of Cun- ningham, Sabin and Doan (5), which did not react to Janus green. These granules may well have been the pre- cursors of mitochondria. Even when we increased the strength of the Janus green until the cells were killed, we were unable to demonstrate any mito- chondria. Besides these very large masses of the reticular cells, there were enormous numbers of epithelioid cells, some of them having rosettes that practically filled the cells, while others, like the one shown in Fig. 6, which was taken from this rabbit, had a wide peripheral zone. Only a few of the epithelioid cells of this rabbit hadmito- chondria, and these were very tiny and were in the extreme periphery of the cells. Very large numbers of the epithelioid cells of this rabbit had a few fat droplets in the periphery. Thus, in this specimen, there were cells in two different phases; first, there was a very marked production of new reticular cells, and second, the epithe- lioid cells were of the fully developed type and many of them showed the beginning of degeneration in the dii- inution of the mitochondria and in the development of fat in their cyto- plasm. In the scrapings from the lung of Rabbit TB 16 (Protocol on page 264; Chart 11A) there were, on the other hand, no reticular cells to be found, but the tissue likewise seemed to have come from the septa, for no elastic tissue was present. The scraping was practically a pure culture of modified mono&es of the type shown in Fig. 5. Films from the lung of this animal were stained for tubercle bacilli and as many as ten were found in the pe- ripheral zone of the epithelioid cells. There were very large giant cells pres- ent, some of which contained enormous numbers of fat droplets. From these two records, it can be seen that there is probably a tendency toward the de- velopment of the cells in cycles; thus, the first rabbit showed a marked wave of the production of new reticular cells, with the epithelioid cells of a preceding generation just beginning to degenerate, while the second rabbit was killed when there was compara- tively little development of new reticu- lar cells, but when there were vast numbers of the matured epithelioid cells. From the observations described h the preceding pages it seems legitimate to conclude that the effect of infection with tuberculosis is to cause an in- crease in the reticular cells of any organ which becomes infected. This local effect on the reticular cells we consider is probably a chemical one and not a direct effect due to the pres- ence of bacilli in the cells, because so far we have no evidence that the reticular cell can phagocytize the bacilli. Furthermore, the infection brings about a rapid maturation of these reticular cells into the typical mono&es and the further change of the monocytes into the epithelioid cells. We are quite sure that the mono&es and epithelioid cells take in the tubercle bacilli, but we are also sure that this multiplication of the reticular cells and their transformation into monocytes and epithelioid cells can be brought about without the im- mediate presence of the bacilli them- selves. In this report of our series of rabbits we shall show that some of the monocytes of the circulating blood be- come mod&d, and we have demon- strated the presence of tubercle bacilli within them. The fully matured epi- thelioid cell contains the bacilli, as can be proved by seeing them in the living cell and by staining them with carbol fuchsin. Moreover, it can be demonstrated that the epithelioid cell does not show a reaction toward these engulfed bacilli as demonstrated by neutral red, a reaction which we be- lieve indicates that the cell destroys and digests the material taken up. The giant cells are produced by the multiplication of the nuclei of the epithelioid cell which always takes place in the peripheral zone, leaving an undivided and central centrosome sur- rounded by an enormously developed rosette. In one of the fresh specimens, a very large double giant cell was seen, that is to say, a single cell with two R61e of Monocyte in Tuberculosis 249 250 Cunningham, Sabin, Sugiyama and Kindwall rosettes, each of which had a partial rim of nuclei. Both the epithelioid cells and the giant cells finally show the development of fat droplets, which begins near the periphery of the rosette, after which they gradually in- crease in number until they fill the entire peripheral zone of the cell and ultimately seem to replace the entire rosette. In some of the cells this increase in the amount of lipoid gran- ules is so great that there is no stain- ing to be seen at all (Fig. 9). The specihc effect of the tubercle bacil- lus on the monocyte is the enormous increase in the numbers and the de- crease in the size of the fine bodies &ining,with neutral red that charac- terize tie living monocyte. This de- velopment divides the cell into two zones, the central zone of the rosette and a peripheral zone which usually contains no stainable substance, that is, has no reaction to vital dyes but does contain the tubercle bacilli. The rosette more rarely may entirely fill the cell. The fine granules of the rosette give the zone of the rosette a pink reaction in Wright's stain, in which the typical azure granules of the mono&e can be seen. The periph- ery of the cell retains the same muddy blue reaction in Wright's blood stain as that whirjh characterizes the cytoplasm of the normal mono&e. The giant cell of tuberculosis has all of the signs of having come from a mono&e; it is a type of giant cell in which the rosette, characteristic of the mono&e, has become enormously en- larged and is centrally placed in the cell so that the multiple nuclei are confined to a peripheral zone. In this characteristic, the giant cell of tuberculosis is an entirely different type of cell from the so-called foreign- body giant cell and the osteoclaet. Thus the effect of tuberculosis is on one strain of cells, and consists in the increase of reticular cells and their maturation into monocytes and the characteristic derivatives of monocytes -the epithelioid cells, and a special type of the giant cell. Ratio of the mono&es to the lympho- cytes in the blood and the corre- lation of this ratio with the cells of the tissues. In our series of rabbits we used ten animals as controls and made repeated counts of the peripheral blood on them to obtain the normal number of the white blood-cells and the normal per- centage of the merent types of these cells. For the data concerning the normal blood of rabbits, we have also included the counts made on the rest of the animals before they were infected. From these data we have found that the average number of the white blood-cells in the normal rabbit is 11,281, taken from counts on 54 rab- bits. The average normal percentage of the monocytes proved to be 8 per cent, the actual number per cubic millimeter being 943; the correspond- ing data for the total lymphocytes is 25 per cent and 2805 cells per cubic millinxter. We are showing in Charts 1 and 2 the relative frequency of the total number of the white blood-cells and the average percentages of the monocytes in the normal rabbit to demonstrate that the range of varia- tion is not great. In Chart 2 we are giving comparative data for the mono- cytes of the normal and after infection with tuberculosis. R61e of Monocyte in Tuberculosis 251 In following the blood of rabbits ratio between mono&es and lympho- which have been infected with such cytes in the circulating blood that we massive doses of tubercle bacilli as to have been able to make a correct judg- give a comparatively acute reaction, ment concerning the condition of the we have found that there is a marked animal in the majority of experiments correlation between the progress of by following the ratio between these the infection in the tissues and the two types of cells. We have found condition of the blood. There are two that, as is well known, there is a gen- striking eeects to be seen in the blood; eral lowering of the production of both first, there is an actual and marked white and red blood-cells in tuberculo- CHART 1. CHABT SHORING TFDZ DISTEUBU- TION OR TEE AVEEAQE COUNTS OF THE TOTAL WHITE BLOOD-CJCLLB IN NOBM~L RABBIT On the abscissae are given in thousands the average number of white blood-cells per cubic millimeter. On the ordinates are given the number of animals in each group corresponding to a given number of thous- ands on the abscissae. It will be seen that from 64 rabbits, the largest group, namely 9, had a count between 9 and 10,ooO; and that the extremes were represented by very few animals. increase in the percentage of the mono- cytes which may go as high as 53 per cent; and second, the normal ratio of monocytes to lymphocytes is reversed. Thus, in a rabbit with active, acute tuberculosis the mono&es of the cir- culating blood surpass the lympho- cytes in number. so important is this CEABT 2. CEABT SHOWINQ TEE DISTBIBU- TION OF TEE A~XUIJE PEIUZENTAOES OF MON~CYTES IN TEN BLOOD OB NOBMAL RABBITB, UD IN RUBIT~ Wmcn HARE BEEN INPE~T~D WITE Twrmou~osrs On the abscissae are given the percentages in groups; on the ordinates are given the number of animals corresponding to a given range of percentagee on the abscissae. The figures on the left margin correspond to the curve from the normal rabbits, while those in the right margin correspond to that of the tubercular rabbita. It will be seen that the most frequent range in per- centagea in normal rabbits is from 7 to 9 per cent, for 21 out of a total of 54 animals were in this group, while the corresponding most frequent range in percentages for infected animals is from 10 to 12 per cent; ten out of 34 animals were in that range. 252 Cunningham, Sabin, Sugiyama and Kindwall - TB 1 15 6 998 422 2 3 24 132,373 1,288 21 6 30 142,9951,423 3, 7 29 103,5451,251 3 8 27 62,635 740 2: 10 21 71,836 654 2 11 21 52,010 433 11 13 32 93,7171$X4 1. 15 39 44,599 467 10 27 82,5841,144 1, 17 25 114,3351,734 2 18 52 94,118 712 3 19 21 92,268 870 20 25 7 4,708 1,235 21 19 82,419 1,124 22 30 133,7401,655 23 34 5 6,781 1,156 24 22 83,453 1,354 25 25 83,034 949 26 26 64,171 987 27 20 51,935 497 29 23 62,300 536 4 30 29 72,860 637 A 31 21 61,659 485 32 27 72,488 655 33 21 8 1,489 589 21 34 26 10 1,893 771 2 35 27 123,865 1,754 4 36 41 85,5761,088 2 37 17 133,1282,392 3: 38 40 203,7601,880 3. 39 33 22,838 172 2: s 59 21 71,922 844 1: 60 13 91,191 867 2: TABLE I Studies on the blood of rabbits 19 1,4021,425 6,620 5,687 Bacillus abortue, arrested tuberculosis 102,703 1,211 9,590 10,884 Miliary tuberculosis 12 3,647 1,208 10,290 10,952 Arrested tuberculosie 14 4,237 1,797 12,520 12,840 Found dead fifth day after inoculation. Slight pneumonia 20 2,443 2,048 11,680 lO,kO Moderate tuberculosis 11 1,947 949 8,600 8,331 Arrested tuberculosis 20 1,449 1,812 9,080 9,060 Moderate tuberculosie 19 2,112 3,374 11,500 18,368 Miliary tuberculoaie 10,455 Control 202,7993,469 13,455 9,720 Extreme tuberculosis 9 3,683 2,599 17,540 16,632 Bacillus abortue only 7 4,312 947 7,920 15,066 Bacillus abortus only 11,125 Control 16,760 Control 14,150 Control 19,250 Control 21,100 Cpntrol 15,700 Control 12,300 Control 14,222 Control 9,244 Control 72,700 506 8,900 6,338 Arrested tuberculosis 83,895 752 9,800 9,042 Arrested tuberculosis 7,364 Control 9,102 Control 14 1,687 1,071 7,140 6,588 Moderate tuberculosis 12 1,8111,006 7,348 7,563 Moderate tuberculosis 112,652 84113,943 7,090 Moderate tuberculosis 172,876 1,49113,600 10,038 Moderate tuberculosis 142,810 1,030 18,400 7,945 Moderate tuberculosis 202,763 1,621 9,400 8,197 Extreme tuberculosis 29 2,275 3,217 8,600 10,722 Extreme tuberculosis . 15 1,743 1,597 9,045 9,506 Moderate tuberculosis Bacillus abortue 19 2,1712,903 8,696 14,140 Arrested tuberculosis 73 76 79 65 Pl 2 3 5 6 7 8 9 10 11 12 13 14 15 16 R61e of Monocyte in Tuberculosis TABLE I--Continued 253 -e----w--- 22 112,547 1,273 16 17 1,5011,81111,580 8,309 Moderate tuberculosis Bacillus abortus 19 111,679 972 13 14 1,223 1,346 8,846 10,888 Arrested tuberculosis 32 123,247 1,180 19 102,059 1,084 9,840 10,840 Arrested tuberculosis 31 83,902 979 18 122,2641,61212,29313,668 Moderate tuberculosis 19 3 1,996 358 22 15 1,813 1,365 10,240 9,551 Moderate tuberculosis 15 81,693 865 20 162,0531,81411,36012,747 Extreme tuberculosis 18 92,332 1,231 14 14 1,600 1,673 12,960 11,606 Miliary tuberculosis 31 7 1,971 445 41 84,435 952 6,360 9,936 Extreme tuberculosis 6 4 917 564 20 172,3471,94514,12011,180 Moderate tuberc~osie 24 3 2,289 381 30 74,503 1,128 12,720 14,952 Extreme tuberculosis 19 42,126 422 27 203,2372,47610,56012,245 Miliary tuberculosis 22 101,739 875 27 112,4721,016 8,009 9,141 Arrested tuberculosis 35 104,382 1,309 34 74,162 826 13,000 11,582 Arrested tuberculosis 24 72,475 778 23 82,339 65010,360 8,307 Arrested tuberculosis 21 62,006 600 26 82,465 810 9,760 9,421 Arrested tuberculosis 21 63,001 843 21 41,492 307 13,940 7,284 Arrested tuberculosis 24 102,6671,184 24 101,878 80010,360 7,886 Arrested tuberculosis 22 82,916 987 26 82,065 745 12,240 8,593 Miliary tuberculosis 20 131,545 938 23 122,0761,075 7,680 8,839 Miliary tuberculosis -- -------- 25 82,805 943 25 142,4651,45511,281 9,978 Average.. sis as the disease becomes chronic, unless there is a secondary infection; but in the leucopenia which develops, there is an alternation in the propor- tions of the different types of white C&3. We are giving in Table I the records of 54 of our rabbits, including the controls, the animals infected with B. abortus and with tuberculosis, with the relative percentages and the actual numbers of monocytes and lympho- cytes before and after infection, cor- related with the condition found at autopsy. From these data, it becomes clear that there is an increase in the percentage and in the actual number of monocytes after infection with tu- berculosis; that the average percentage of lymphocytes remains the same, with, however, a slight decrease in their number; while there is a decrease in the actual number of the white blood-cells, i.e., a slight general leuco- penia. The effect of tuberculosis on the peripheral blood becomes more striking when the cases are analyzed with regard to the grade of tuberculo- sis found at autopsy, ae is shown in Table 2, in which it will be seen that 254 Cunningham, Sabin, Sugiyamh and Kindwall in severe infection the monocytes in- crease from an average of 8 per cent up to an average of 15 per cent. In this table it is interesting to note that the leucopenia is greater in the groups marked "moderate" and "arrested" than in the group marked "severe," which we interpret as due to the longer duration of the former experiments. When the records of our experiments were analyzed, it was found that on the basis of our studies of the blood the animals fell into three groups, TABLE II Percentage and number of monocytes and lymphocytes in the peripheral blood Of tWbbit8 14 - 25 24 25 25 - 8 2,895 94311,281 Normal before infection 15 2,7231,89011,401 Extreme and miliary tuberculok 15 2,253 1,4149,200 Moderate tuberculosis 11 2,3041,064 9,223 Arrested tuberculoeie somewhat correlated tith the clinical groups of Table II. In the first group we have included those animals in which, shortly after the infection, the monocytes increased so that the nor- mal ratio of monocytes to lympho- cytes was reversed and in which this unfavorable ratio was maintained throughout the experiment. All of these rabbits showed either miliary tuberculosis or a grade of infection even more extreme, which will be de- scribed later. These animals showed consistently a low resistance to tu- berculosis. In the second group were the animals in which the lymphocytes remained consistently above the mono- cytes, even though there was some rise in monocytes. These animals at au- topsy either showed no microscopic evidence of tuberculosis at all or a condition which we interpret as charac- teristic of arrested tuberculosis, that is, they showed a consistently high re- sistance. In the third group, which represents the largest of the three, the blood showed a repeatedly shifting ratio between monocytes and lympho- cytes and the result at autopsy cor- related with the condition of the blood which was obtained when the animal was killed. It will be noted that this grouping does not bring together all the animals that had the same grade of the disease at the time of the autopsy, for, while in Group 1 all of the cases were severe at autopsy and in Group 2 all were arrested, in Group 3, on the other hand, the results at autopsy were mixed, being extreme, moderate or arrested. The basis of our classification rather has been the nature of the reaction of the animal throughout the experiment; thus in Group 1, the entire reaction of the animal was unfavorable; in Group 2, the entire reaction was favorable, while in Group 3 the animals showed attempts to build up a resistance alter- nating with periods of low resistance; that is to say, there were alternating periods of active and arrested phases of the disease. Through this type of classification we have been able to make a better study of the ratio of monocytes to lymphocytes in tubercu- losis. High monocytes and low lym- phocytes have been found associated with active tuberculosis, while low R61e of Monocyte in Tuberculosis 255 monocytes and high lymphocytes have been found associated with arrested tuberculosis. For Group 1, representing animals with a consistently low resistance, we are giving protocols and charts (Nos. 3 to 6) of four experiments, Rabbits TB 39, P 2, P 3 and TB 13. Protocol, Rabbit TB 39 4/84/2S. Weight 3050 grams. W.B.C. 3600. For the records of the blood see Table I and Chart 3. 4/(/a6/86. Injection of 2 cc;saliie emulsion of tubercle bacilli, Bl, 50 bacilli per oil- immersion field, intravenously. d/86/% up to 6/.%9/S. Weight decreased to 2330 grams, W.B.C. as shown on Chart 3. Monocytes and lymphocytes before and after the inoculation with tuberculosis on Table I. There was no leucopenia, the average of the last four counts being 11,575. S/,93/36. Animal killed on account of the high monocytes. Autopsy: Extreme, dif- fuse tuberculosis of the lungs; supra- vital studies showed large masses of reticular cells from the septa and enor- mous numbers of free epithelioid cells of the type shown in Fig. 6. Very few of the epithelioid cells had mitochondria, but they had numerous refractive droplets of fat in the peripheral zone. Spleen sur- rounded by an enormous clot, which sug- gested an organised rupture. Mesenteric lymph glands enormously enlarged. As will be seen on Chart 3 (Rabbit, TB 39), there was a marked increase in mono&es up to 3000 per cubic mil- limeter in this animal 5 days after the intravenous injection of the tubercle bacilli, when, of the total number of the white blood-cells, 29 per cent were monocytes. From this time on, the mono&es were consistently high, reaching a maximum of 52 per cent on May 21st. On May 5th, it was first noted that some of the monocytes of the circulating blood were strikingly changed and from that time on there were marked variations in the mono- cytes. Certain of them were found to be very young forms, obviously result- ing from an increase in cell division by amitosis. That cell division was increased was also shown by the ob- servation that, on May 21st, 4 of the 52 monocyta per 100 cells were found in division. In the blood of this ani- mal occasional degenerating mono- cytes were found. The phenomenon of the modified monocytes in the circulating blood is a most interesting one. These modified monocytes were characterized by the fact that they were round, appeared to be larger than normal and showed no motility. As was described in the pre- ceding section, these modified mono- cytes of the circulatii blood contained tubercle bacilli. We believe that these monocytes in the circulating blood, in- fected with the bacilli, give the best chance to study the very first effects of the organism on the cells. The most important of these effects are a ces- sation of motility and an increase in the substances stainable with neutral red. At this stage, as is shown in Fig. 1, from Rabbit TB 49, the line of de- marcation between the fine bodies of the rosette and the vacuoles of the periphery is not sharp, since there is such a marked scattering of the vac- uoles. We are unable to say whether or not the presence of these scattered vacuoles, in the monocytes which are just beginning to show the effects of damage by the infection or more specifically by having taken in the bacilli, indicates that the immediate and normal reaction of the monocyte is an attempt to kill the invading bacil- lus. We consider, however, that the 256 Cunningham, Sabin, Sugiyama and Kindwall cessation of motility represents an im- mediate damage of the monocytes. At autopsy, Rabbit TB 39 showed a d&se, generalized tuberculosis of the hmgs. No tubercles were seen in the gross specimen; in supra4ital prepara- tions the septa of the lungs showed great masses of reticular cells and enormous numbers of scattered typi- reaction of the body in clumping the epithelioid cells into even tiny tuber- cles. Supra-vital studies from a scrap ing of such a lung show scattered epithelioid cells everywhere with no indication whatever that they had been held together by any of the usual framework of the tubercle; most of them are single, some may be in small CHAET 3. CUBVES SHOWINQ THE MONOCYTES, LYMPHOCYTES AND POLYMORPHONUCLEAR NEUTROPHILIC LEUCOCYTEE OF THE PEIUPHPXAL BLOOD FROM RABBIT TB 39 The dates are given on the abscissae and the numbers of cells on the ordinates. The curve of the total white count has not been plotted, but, in a general way, it can be judged from the other lines on the chart. The date of the inoculation of the animal with tubercle bacilli, strain B 1, is indicated on the chart. The animal belonged to our Group 1, of a severe infection without remissions. Result at autopsy, extreme tuberculosis. cal epithehoid cells like the one of clumps. Sections of such tissues show Fig. 6. Many of these epithelioid cells small clumps of monocytes, it is true, were beginning to show fatty degenera- but these clumps are not confined by tion. This lung was an example of a any framework of connective-tissue tuberculous lesion which we regard as fibres, nor are they surrounded by even more severe than the diffuse mili- lymphocytes, so that the principal and ary tuberculosis, namely, a diffuse and overwhelming characteristic of this invasive mononucleosis in which the grade of infection is an extensive inva- process is so extensive, or, as it were, sion of the tissues with single epithe- so malignant that there is not yet any lioid cells. We have other examples of Rde of Monocyte in Tuberculosis 257 this type of reaction in even more ex- Protocol, Rabbit P 8 treme form than in Rabbit TB 39. l/M/M. Weight 1930 grams. Temp. 102.6. In Charts 4 and 5 are shown the W.B.C. 11,360. For types of white blood- blood co&s from Rabbits P 2 and P 3, cells see Table I, Chart 4. and these experiments were likewise on l/16/.95. Injection of 1 cc. saline suspen- animals in which the infection was eion of tubercle bacilli, Bl, 15 organisms per oil-immersion field, intravenously. BOO0 l4000 l3000 12000 11000 10000 9000 i Sm 12 16 19 22 27Lh3 12 25 21lk2 5 3 11 12 CHART 4. CURVES SHOWINQ DATA FROM THE BLOOD OF RABBIT P 3, SIMILAR TO THOEE ON CHART 1 The animal belonged to our Group 1, of a severe infection without remissions. Result at autopsy, extreme tuberculosis. severe from the start; in neither of For data on monocytes and lymphocytes them did the increase in the monocytes see Table I. follow the injection of the bacilli as S/ 6/M. The W.B.C. were consistently quickly as in TB 39, but in both they normal in number except on this date remained COnSiS~ntlY high sfter they when they rose to 20,969 had once surpassed the lymphocytes. 8/.88/M. Lowest weight, 1695 grams. S/18/&5. Weight 2006 grams. W.B.C. 258 Cunningham, Sabin, Sugiyama and Kindwall 10,800. A&I81 killed. Autopsy: Extreme berculosis of the abdominal viscera, tuberculosis of the lungs, kidneys, peri- cardium and spleen. Supra-vital studies miliary tuberculosis of the pericardium of the lung showed large tuberoles to- and extensive tuberculosis of the lungs. gether with great masses of free epithe- The septa of the lungs showed exten- lioid cells; 8 few typical clssmstocytea, sive mononucleosis. We have not HooO , CHARTS. CUFWES SHOWINO DATASIMILAETOTBOSE ONCHART~,F~OMTHEBLOOD OF RABBIT P 3 The snhnal belonged to our Group.& of 8 severe infection without remissions. %?a&, at 8UtOpSy, milisry tuberculosie. ahOst no lymphocytes. Viscersl peri- seen such extensive masses of epithe- cardium showed miliary tuberculosis; in the kidney many tubercles were found lioid cells in the spleen of any other but no scattered epithelioid cells. The animal of our series. There were very spleen showed grester numbers of scst- few lymphocytes in these tissues. tered epithelioid cells than in any other In Rabbit P 3, there was marked &m81 of the series. A few ecattered tuberculosis of the lungs and of the epithelioid cells in the liver. kidneys. Protocol, Rabbit P S In Rabbit P 2, the organisms were given intravenously and at autopsy I/H/%6. Weight 2210 grams. Temp. 102.5. W.B.C. 12,960. For types of white blood- there was most extensive miliary tu- cells see Chart 5. R61e of Monocyte in Tuberculosis 259 1/16/H. Injection of tubercle becilli same as for Rabbit P 2. For data on mono- cytes and lymphocytes see Table I. 9/i?7/9& Highest count of the W.B.C., 17,300, with P.M.N. st 13,439. S/84/&5. Lowest weight, 1615 grams. S/U/94. Weight 1765 grams. W.B.C. 14,000. Killed. Autopsy: Lung showed very marked involvement but on the left side there were norm81 8re8s in the gross and there were zones which suggested heding. The kidneys showed numerous small tubercles. Suprs-vital studies of the lung showed tubercles ctnd many clumps of three or four epithelioid cells; it ~88 noted that they were surrounded by clumps of lymphocytes, indicating the beginning of 8 resction fsvorable to the animal. Meny free lymphocytes. The kidney showed tubercles snd an occa- sional free epithelioid cell. It was interesting in this animal, that, though the involvement was marked, there were nevertheless signs of the beginning of a reaction of in- creased resistance on the part of the animal, since we interpret an increase in lymphocytes in the tissues involved in the tuberculosis as indicative of such a reaction. The lymphocytes of the blood had, however, not yet responded to this increase in the tissues and the ratio of monocytes to lymphocytes was still unfavorable for the prognosis. Rabbits P 2 and P 3 both showed rises in leucocytes at about the same time, which involved primarily the neutro- philic leucocytes, and were thus prob- ably due to a super-added infection. The fourth example of this group is shown on Chart 6 (Rabbit TB 13). Protocol, Rabbit TB 1S 11/98/34. Weight 2470. Condition good. IS/ S/84. W.B.C.11,320. Types0fW.B.C. shown on Chart 6. f,6/f7/$4. Injection of 2 cc. saline sus- pension of tubercle bacilli having 50 organisms per oil-immersion field. i/16/26. Weight 2000. Found with par- slysed hind legs; marked leucocytosis, W.B.C. 21,360 8.m. snd 33,600 p.m.,due to an increase in the P.M.N. of 14,524 and 23,184, respectively, and in mono- cytes, 4272 snd 5630. Killed. Autopsy: Marked miliiary tuberculosis of the entire peritoneum, including the mesentery snd the sbdominal viscera. Supra-vital stud- iee showed the tubercles to be made up of typical epithelioid cells with massive rosettes. Very few young monocytes; only very slight fatty degeneration of the epithelioid cells. We did not make as many counts of the blood of this rabbit as would have been desirable, but all that were made showed that the monocytes were con- sistently higher than the lymphocytes with a marked increase in monocytes just before the animal was killed,. The slutdpsy showed the most extreme mili- ary tuberculosis of the peritoneal wall and of the abdominal viscera, together with some involvement of the lung and a zone of red hepatization in one lung which probably accounted for the rise in the leucocytes. The injection of the bacilli into this rabbit was given intra- peritoneally. It will be seen that all of the animals of this group had severe tuberculosis at the time of autopsy; two cases we have classified aa extreme and two as miliary. All of them, with the excep tion of Rabbit TB 13, (the one found paralyzed) might have lived longer, and in the case of `Rabbit P 3 it is pos- sible that a resistance might have been built' up. The records of these four experiments show that a continued severe infection was indicated in the circulating blood by an unfavorable ratio of mono&es to lymphocytes. The second group of our series repre- sents animals that were markedly re- sistant to the infection throughout the 260 Cunningham, Sabin, Sugiyama and Kindwall period of the experiment. We are giving the protocols of four animals from this group, Rabbits TB 6, TB 10, TB 29 and TB 30, as representative of markedly resistant animals. The lungs showed 8 few old ~(381s. There wss an enlsrgement of the lymph nodules of the blind pouch of the intestine, snd supra-vital studies showed one or two stimulsted mono&es from them. Other- wise no signs of tuberculosis whatever. CHART 6. CURVES SHOWINQDATAFBOMTHIPBLOOD OFR~BBITTB~~, SIMILARTO Taosn ON CHART 3 The animal belonged to our Group 1, of 8 severe infection without remissions. Result at autopsy, miliary tuberculosis. Protocol, Rabbit TB 6 18/ S/94. Weight 1899 gr8ma. W.B.C. 9800. Chart 7. 16/18/2?4. Injection of 2 cc. saline sus- pension of 50 organisms to the oil- immersion field, intraperitoneally. l/16/.&5. Weight 2030. Condition good. ,9/zo/96. W.B.C. 10,030. 9/94/M. W.B.C. 6600. a/B7/M. W.B.C. 7400. s/ .9/i%. W.B.C. 3400. S[ 4/M. W.B.C. 7120. Killed. Autop&y: In the films from the spleen there were unusually large m8sses of yellow pigment in some of the clssmstocytes, while others were filled with red blood-cells. Protocol, Rabbit TB 10 19/ g/$4. Weight 1800. W.B.C. 9299. M/17/94. Injection of 2 cc. of artline sus- pension of tubercle bacilli, 69 organisms to oil-immersion field, intraperitonedly. For the types of white blood-cells see Table I and chart. 8. RBle of Monocyte in Tuberculosis 261 l/99/96. W.B.C. 8880. Time of the leucopenia. l/17/&5. W.B.C. 5480. Time of the leu- copenia. .9/ 9/M. W.B.C. S400. 8/16/25. W.B.C. 9400. Weight 1980 grams. Animal in excellent condition. Hilled. Autopsy: A few healed calcified lesions in the omentum surrounded by lymphocytes. Lungs had a few healed lesions; no epithelioid cells found. pale, contained an average amount of air; no tubercles seen on gross inspection. Supra-vital studies disclosed a few tu- bercles in the lung made of monocytes which were full of droplets of fat. Around these tubercles were great numbers of small lymphocytes. No calcified or caseated lesions. Protocol, Rabbit TB SO 4/6/14/M. W.B.C. 10,200. For the records of the blood see Table I and Chart 10. I I h.9 h.9 it it 1217 1217 29JvC 29JvC u u 13 13 22 22 tliib9 tliib9 19 19 24 24 2tk.2 2tk.2 4 4 CE~~T 7. CURVES SHOWINQ DATA FSOY TEE BLOOD OF RABBIT TB 6, SIMILAR TO THOSE ON CHABT 3 The animal belonged to our Group 2, in which a high resistance to the infection was maintained. Result at autopsy, arrested tuberculosis. Protocol, Rabbit TB 99 4/14/~6. W.B.C. 10,SOO. For the records of the blood see Table I and Chart 9. 4/80/96. Weight 1660 grams. 4/.86/86. Injected 2 cc. of a saline suspen- sion of tubercle bacilli Bl, having 50 bacilli to an oil-immersion field, intra- venously. 6/ 6/25. W.B.C. 5600. Beginning leuco- penis. 6/ 7/86. W.B.C. 6400. 6/966/26. W.B.C. 3800. Low blood count. b/97/86. W.B.C. 2800. Lowest blood count. Weight 1750 grams. Anllal in ex- cellent condition. Killed. Autepsy:Lungs 4/80/86. Weight 1700 grams. 4/86/86. Injected 2 cc. of a saline SUB- pension of tubercle bacilli Bl, having 50 bacilli to an oil-immersion field, intra- venously. 4/87/86. W.B.C. lk,690. Highest blood count, correlated with the highest number of P.M.N., 9198. 4/$8/86. Weight 1420 grams. Looks sick. 6/ 8/.86. Weight 1570 grams. Gaining. 6/M/96. Weight 1720 grams. 6/186/26. W.B.C. 6200, P.M.N. X50-lowest blood count up to this time. Weight 1380 grams. Animal looks sick. b/87/86. W.B.C. 4SOO. P.M.N. 1988. Weight 1220 grams. 262 Cunningham, Sabiq Sugiyama and Kindwall 6/#9/#6. W.B.C.4900. P.M.N.1980. S/ l/.96. Animal found dead. Autopsy: Lungs markedly red and inflamed; leath- ery in consistence with marked bleeding, though the animal had been dead a long time. Blood watery in consistence. Ap- parently a generalized acute reaction with a healed tuberculosis in the back- ground. Rest of tissues normal. The charts (7 to 10) of all of these animals are similar; all of them record a time after the infection when the monocytes rose, showing a reaction to the tuberculosis, but in none of them 1 CHABT 8. CURVES SHOWINQ DATA FROM TEE BLOOD OF RABBIT TB 10, SIMILAR TO TEOSB ON CHABT 3 The animal belonged to our Group 2, in which a high resistance to the infection was maintained. Result at autopsy, arrested tuberculosis. did the monocytes ever go above the lymphocytes. All of them likewise showed that the period of the leuco- penis followed the period of the rise in monocytes rather than occurring syn- chronously with it. Thus, the leuco- penia is a residual effect after the animal has already controlled the in- fection. It will be noted that the greatest rise in lymphocytes was in Rabbit TB 6, shown in Chart 7, and our records show that there was an in- crease not only in their numbers but also in their motility. The records of the autopsies all indicate that these animals had no demonstrable infec- tion or were in a state in which the disease was markedly arrested. There were in some fibrosed and calcified tu- bercles; there was also a marked in- crease in the lymphocytes in the tissues which had been involved in the tu- berculosis and this increase was accompanied by an increase in lympho- cytes in the circulating blood. It will be seen in Table 1 that all of the ani- mals in this group showed either no increase or only a slight increase in the monocytes after the infection, but, on the other hand, there was a marked rise in lymphocytes. In animals TB 6 and TB 10, the increase in percentr age of the lymphocytes was not marked; but in TB 29 and TB 30 the percentages of lymphocytes rose to 42 and 43 respectively. The third group was by far the larg- est of our series and includes the ani- mals which neither succumbed at once to the infection nor showed a consistr ent resistance. We have selected 5 animals (TB 16, TB 1, TB 38, S 65 and P 1) from this group whose blood is shown on Charts 11 to 15. In general, two tendencies are to be noted in the curves on these charts; either the ratio of monocytes to lymphocytes fluc- tuated repeatedly, first one form pre- ponderating and then the other; or there were periods in which both lym- phocytes and monocytes were in- creased but tended to run along more or less parallel to each other. Some of the curves show combinations of these two types. In this group, the reaction of the animal to the disease was obviously much more complex than with the animals of our Groups 1 and 2. In general, it was possible to Rde of Monocyte in Tuberculosis 263 CHART 9. CUIWES SHOWING DATA FROM THE BLOOD OF RABBIT TB 29, SIMILAB TO TEOSB~ ON CHART 3 The animal belonged to our Group 2, in which a high resistance to the infection was maintained. Result at autopsy, arrested tuberculosis. CHABT 10. CURVES SHOWING DATA FROM THI BLOOD OB RABBIT TB 30, SIMILAE TO T~osl~ ON CHAIIT 3 The animal belonged to our Group 2, in which a high resistance to the infection was maintained. Result at autopsy; arrested tuberculosis with a superimposed acute infection. 264 Cunningham, Sabin, Sugiyama and Kindwall correlate the condition at autopsy in these animals with the ratio of mono- cytes to lymphocytes at the time of the autopsy, provided the reaction had obtained for a sufficient time. In this group when the mono&es were markedly above the lymphocytes at the time of autopsy, marked tubercu- losis was found; those in which the monocytes had been at the normal level for some time showed arrested tuberculosis, while the intermediate and the mixed types of curves corr& sponded to a moderate grade of the disease. The first example of this group, from Rabbit TB 16, (Charts 11 A and 11 B) was the most extreme example of malignant, invasive tubercular mono- nucleosis of our entire series. Protocol, Rabbit TB 16 l/.??f/96. W.B.C. 12,160. For records of the blood see Table I and Chart 11. I/G6/G6. Weight 1650 grams. i/.87/.86. W.B.C. 17,440. W.B.C. and P.M.N. rather high, as shown on Chart 11, until the time of the injection of tu- bercle bacilli. S/ 6/G6. W.B.C. 11,840. S/ 7/G6. Injection of 5 cc. of saline Bus- pension of tubercle bacilli, Bl 50 organ- isms to an oil-immersion field, intra- venously. s/ G/G6. W.B.C. S400. S/29/86. W.B.C. 6200. Lowest count. P.M.N. 2108 4/ l/$6. Animal showed marked loss in weight. Killed at the time of a rising count of monocytes. Autopsy: Lungs showed massive gelatinous pneumonia, looking as if they were a solid mass of cells. Supra-vital studies of scrapings from the freshly cut surface of the lung showed an almost pure culture of epi- thelioid cells, such as the one in Fig. 5. No elastic tissue at all in the scrapings, as if the entire tissue had come from the septa, and no epithelium from the air sacs. Epithelioid cells contained tubercle bacilli, demonstrated with Ziehl-Nielsen technique. Some enormous giant cells loaded with fat. Result: an extreme grade of invading monocytosis with little clumping of the epithelioid cells into circumscribed tubercles. It is obvious that by the term mono- nucleosis we mean the increase of the modified monocytes or epithelioid cells. It is interesting that the average num- ber of the monocytes in this animal was high before the injection of the bacilli, being 1144 as against a normal aver- age of 943 (see Table 1). After the injection of the bacilli, the average of the monocytes was 3469. We are giv- ing two charts of this animal, the first one showing the total number of the cells and the second the correspond- ing percentages. In this animal the lines of monocytcs and lymphocytes crossed repeatedly until the last few days, when the blood showed the fol- lowing astonishing numbers of mono- cytes; 4189, 4864, and 6364. At a&pay, the lungs in the gross showed a massive, diffuse gelatinous pneu- monia; they were pale and, on section, little blood ,was seen; they looked as if the entire pulmonary tissue was a mass of cells. The fresh scraping from the cut surface was practically a pure culture of infected monocytes, i.e., of epithelioid cells of the type shown in Fig. 5. Bacilli were demonstrated in these epithelioid cells. The specimens looked as if they had come from the septa entirely; for there was no elastic tissue whatever. There were some enormous giant cells full of fat drop- lets. Sections of the lung confirmed the supra-vital studies, for the septa showed great masses of epithelioid cells. Blood-vessels were seen full of these cells and some were present in R61e of Monocyte in Tuberculosis 265 CHART 11A. CURVES SHOWINQ DATA FROM THE BLOOD OF RABBIT TB 16, SIMILAR TO THOSE ON CEART 3 The animal belonged to our Group 3, in which there were alternating periods of active and inactive tuberculosis. Result at autopsy, severe tuberculosis. Rrcd 100 90 I I22 n 28h2 19 27kr.3 4 573 H 13 16 n 19 23 2s % so*1 CHART 11B. CURVES SBO~INQ TEE DATA FEOM THE BLOOD OF TEE SAME RABBIT (TB 161 AS ON CHART llA, BUT PLOTTED IN PERCENTAGES 266 Cunningham, Sabin, Sugiyama and Kindwall the exudate of the alveoli. It is obvi- ous that this animal was killed in the most active phase of the disease. The next experiment Rabbit TB 1 (Chart 12) in this group was a long one, extending from December 9th to April 7th. Protocol, Rabbit TB 1 ll/GS/G4. Weight 2350 grams. For records of the blood see Table I and Chart 12. S/ 7186. Second injection of TB-5 cc. of saline suspension of Bl-intravenous- ly. 50 bacilli to an oil-immersion field. 3/31/96. W.B.C. 4200. l/18/86. Weight 2330. l/tS/G6. Weight 2370. Condition good with actual gain in weight. 4/ 7/96. Weight 2130. Condition good. 4/ 8/.96. W.B.C. 7400. Killed. Autopsy: Peritoneal walls and viscera normal. Lungs showed a few small white masses surrounded by yellow lines suggesting CHART~~. CURVESSHOW- INQ DATA FROM THI BLOOD OF RABBIT TBl, SIYILAE TO THOSE ON CHART 3 The animal belonged to our Group 3, in which there were alternating periods of 11 (f (b IT 19 23 23 26 20 31 Apl 4 7 l.G/ G/94. W.B.C. 6440. active and arrested tubercu- losis. Result at autopsy, arrested tuberculosis. IG/lG/94. Injection of 2 cc. of saline sus- healed tuberculosis. Supra-vital studies of the lung showed an occasional mono- pension of tubercle bacilli Bl, intraperi- toneally. 50 bacilli to an oil-immersion cyte with a marked rosette, suggesting young epithelioid cells. A few epithelioid field. cells were filled with fat. The omentum 1.9/17/94. W.B.C. 5300. 1,8/GG/G4. W.B.C. 11,500. showed great masses of lymphocytes and l/.99/96. W.B.C. 4560. Lowest count. cells suggesting plasma cells. Result: healing tuberculosis. 9/ 7/96. Injected l/5 of a 24-hour culture of living B. abortus in 1 cc. of saline This picture represents the reverse of intravenously. the condition found in Rabbit TB 16, d/10/86. W.B.C. 12,200. for Rabbit TB 1 was killed when the R61e of Monocyte in Tuberculosis 267 animal was in excellent condition and the monocytes of the blood were low. In following the chart it will be noted that the lines of the monocytes and lymphocytes crossed repeatedly and that on four occasions the monocytes were markedly increased. We have no doubt that if this animalhad been killed on March 16th when the monocyets were 4929, the condition at autopsy would have been entirely different. It may be that the rise of the monocytes on April lOth, was correlated with the chondria, while others showed fatty degeneration. Thus the tubercular process was still present but markedly in abeyance. The tissues from the lung showed many lymphocytes. This experiment represents an animal with relatively high resistance. The next animal in this series, TB 38, Chart 13, is mentioned on account of an interesting correlation between the supra+ital studies of the lungs at autopsy and the chart of the peripheral blood. CHART 13. CURVES SHOWING DATA FROM TEI BLOOD OF RABBIT TB 38, SIMILAR TO THOSEI ON CHAET 3 ' The animal belonged to our Group 3, in which there were alternating periods of active and inactive tuberculosis. Result at autopsy, extreme tuberculosis. injection of B. abortus rather than Protocol, Rabbit TB 58 with an exacerbation of the tuberculo- sis. When this animal was killed, the 4/84/H. Weight 1700 grams. W.B.C. 9400. For the records of the blood see Table I mono&es had been normal in num- and Chart 13. bers for some days. The only abnor- 4/%5/N. Injection of 2 cc. of a saline aus- malities in this animal were in the pension of tubercle bacilli Bl, having 50 lungs. In the gross specimen the bacilli to an oil-immersion field. lungs showed some small white masses 6/ B/.86. Weight 1720 grams. Condition good. surrounded by yellowish borders, sug- V~l/% Weight 1650 grams. gestbg he&g tubercla. A few 6/$3/M. W.B.C.4200. Lowest Count. rather small epithelioid cells, typical 6/96/$6. Weight 1400 grams. Animal sick. of tuberculosis, were found; some of Killed. Autopsy: Lungs showed intense them were young with many mito- gelatinous pneumonia; spleen markedly enlarged. Supra-vital studies of the lung 268 Cunningham, Sabin, Sugiyama and Kindwall showed great masses of epithelioid cells like the one of Fig. 7. There were large numbers of epithelioid cells in the lung in which the rosette entirely filled the cell and again very large numbers of them in which the entire periphery of the cell was filled with fat droplets. Result: extreme tuberculosis with beginning degeneration of the epithelioid cells. The chart is quite complex. The first rise in monocytes was preceded by a marked rise `m lymphocytes, so that there was a period between May 5th and May 16th when both mono- cytes and lymphocytes were above normal. After this period there was one sharp rise in monocytes followed by a gradual fall associated with marked fluctuations in the lympho- cytes. It is obvious that such a chart indicates complex reactions on the part of the animal, on the basis of our con- cept that the condition of the blood is actually representative of specific changes in the reaction of the animal to the infection. The condition at autopsy was interesting. There was a marked gelatinous pneumonia and the scraping from the lungs showed very large masses of epithelioid cells, one of which ia shown in Fig. 7. The most striking thing about this lung was the large proportion of epithelioid cells which showed the beginning of fatty degeneration. This observation leads us to suggest that there was a cycle of young infected mono&es, that is, young epithelioid cells between the 19th and 21st, which were begin- ning to degenerate at the time of the autopsy. The first modified mono- cytes in the blood of this rabbit were recorded on May 7th, the time of the East rise in their number. In the in- terval between the 19th and the 21st the monocytes were recorded as divid- ing, modified and degenerating. On the last day of the experiment, out of 33 mono&es, 15 were modsed, while the 18 unmodified were noted as in part very young forms and in part actively motile mature monocytes. On the 23rd, degenerating monocytes were recorded in the circulating blood. We may therefore consider that the monocytes of the period from the 19th to the 25th of May represent the cycle in one .generation of mono&es from the time of the full maturity of the epi- thelioid cells of one exacerbation of the disease to the time of their beginning degeneration. We regard our observa- tions on such cycles as only suggestive, but we give them considerable weight because our entire experience indicates that there is a correlation between the condition of the monocytes of the cir- culating blood and the condition of the monocytes in the general tubercular process. It is, therefore, of great sig- nificance in following the blood of these tuberculous animals to record minutely all of the appearances of the mono- cytes. The ot&er two animals of this group Rabbits S 65 and P 1, showed similar curves (Charts 14 and 15). In both, the lymphocytes were, for a time, above the mono&es, followed by a period of repeated crossing of the two curves; the monocytes were above the lymphocytes at the time of autopsy. Protocol, Rabbit S 66 10/14/94. Weight 1605 grams. For records of the blood see Table I and Chart 14. lO/i30/.94. W.B.C. 11,920. II/ 6/%#. From this time until 12/8/24 the animal received intravenous injections of gelatin in doses of from 10 to 20 cc. of a 2 per cent solution. This was given for a possible effect on the clasmatocytes and monocytes, but without results. 15000 WOO 13000 12000 11000 iooo( 9000 8ow 7000\ 6000 5000 1 . 0s.ZOMOk8l216 27Jr6 15 19 22 23Fih7 9 10 11 12 Ij 19 24 CHART 14. CURVES SEOWINQ DATA FROM THE BLOOD OF RABBIT P 1, SIMILAR TO THOSE ON CEART 3 The animal belonged to our Group 3, in which there were alternating periods of active and inactive tuberculosis. Result at autopsy, moderate tuberculosis. Itn 12 16 l3 Itn 12 16 l3 22 27h3 22 27h3 12 13 25 27k.2 5 12 13 25 27k.2 5 3 11 12 3 11 12 CHART 15. CUBYES SHOWING DATA FROM THE BLOOD OF RABBIT 8 65, SIMILAR TO THOSIP ON CHART 3 The animal belonged to our Group 3, in which there were alternating periods of active and inactive tuberculosis. Result at autopay, moderate tuberculosis. 269 270 Cunningham, Sabin, Sugiyama and Kindwall 19/18/g4. Injection of 2 cc. of a suepen- sion of tubercle bacilli, H 37, intraperi- toneally. 50 bacilli to an oil-immersion field. 1.9/16/B4. W.B.C. 20,390. White blood cells remained high for the month of January. l/M/96. Weight 1970 grams. Condition good. l/QQ/Qb. W.B.C. 20,349. ,9/ 7/,??6. W.B.C. 14,900. Injection of l/5 of a 24-hour culture of living B. abortus, intravenously. ' Q/10/26. W.B.C. 16,000. Highest of the last period of the experiment. $/$4/%5. W.B.C. 11,520. Hilled. Autopsy: Lungs showed scars in which no epithe- lioid cells were found, but considerable numbers of lymphocytes. Liver showed a moderate number of epithelioid cells. Rem&: Moderate tuberculosis. Protocol, Rabbit P 1 l/18/86. W.B.C. 10,249. For records of the blood see Table I and Chart 15. j/16/86. Injection of 1 cc. of a saline sus- pension of tubercle bacilli Bl, 15 bacilli to an oil-immersion field, intravenously. Dose about 1.7 mgm. l/f 7/.96. Weight 2,090 grams. l/.W/Q6. W.B.C. 13,200. l/l7/lc6. W.B.C. 13,930. S/11/$6. Weight 2,000 grams. S/1.9/86. W.B.C. 3,939. Lowest count. .$/&Y/96. Weight 1,700 grams. Lowest weight. 6/18/$6. W.B.C. 16,900. Highest count. Weight 1335. Animal in good condition. Hilled. Autopsy: Lungs showed healing tuberculosis. Kidneys showed a few en- capsulated tubercles. Supra-vital stud- ies showed clumps of epithelioid cells in the omentum; in the kidneys there were some epithelioid cells both free and in small clumps. No free epithelioid cells in the lung and no reaction of lympho- cytes around the epithelioid cells. Re- sult: Moderate tuberculosis. On Chart 14 (Rabbit S 65 ) the monocytes were above the lympho- cytes on the last count, but when the chart as a whole is studied, it will be seen that the lymphocytes were above the mono&es throughout most of the experiment. In this animal there were some infected mono&es, typical epi- thelioid cells, seen in the scrapings from the liver, but not any marked tuberculosis to be made out on gross inspection. The animal whose blood is shown on Chart 15 (Rabbit P 1) showed the monocytes above the lym- phocytes more than in the preceding example. At autopsy, some tubercles were found in the omentum and in the kidneys but the process was not severe. When the records of these three groups of reactions are compared, it will be seen that there are four difIer- ent types of curves. The first two groups of animals represent experi- ments in which the reactions of the animals were so consistent that the records can be shown as composite curves. Chart 16 is a composite curve of Group 1, in which the number of different blood-counts is shown on the abscissae and the number of cells on the ordinates. This is the type of curve of a marked infection in animals with low resistance to tuberculosis, in which it will be seen that the normal position of monocytes and lympho- cytes has been reversed. The mono- cytes are Anally quite steady at the level of about 3000 per cubic milli- meter while the lymphocytes are at about the level of 1000 per cubic mil- limeter. That is to say, in `this type of curve the monocytes are above nor- mal while the lymphocytes are below normal after the infection had been established. On Chart 17, we are giving a curve of the counts of the animals in Group 3 in which there was a consistent state R61e of Monocyte in Tuberculosis 271 of marked resistance to the infection. On this chart there are two slight rises of mono&es, not in any degree, how- ever, comparable to the marked eleva- tion of the level of the lymphocytes, which were for a considerable time about 4000 cells per cubic millimeter with a singIe extreme rise over 5000 cells. The period of the very marked show that the animals had been in- fected. Thus, this type of curve is the exact reverse of the one of Chart 16, for in this one the monocytes were below normal while the lymphocytes were markedly above normal. These two charts may be considered raS repre- sentative of the two groups of animals having a low and a high resistance and COHPOSITE CUflVE CHART 16. COMPOSITE CURVE OF TEE BLOOD OF GBOUP 1 (CHARTS 3-6 INCLUSIVE), SHOW- INQ THE TYPE OF THE Cnsv~s OF TEF, MONOCYTES AND LYMPHOCZTES IN TUBERCULOSIS IN RABBITS TEAT HAD No REWSSIONS OF THE DISEASE 0 6aOO COflPOSlTE CURVE OF GROUP It. II\ /\ CHART 17. COMPOSITE CWVE OF THE BLOOD OF GROUP 2 (CHARTS 7-10 INCLUSIVE), SHORING TEE TYPE OF TEE CIJRVES OF THE MONOCYTES AND LY~~PHOCYTES IN TUBERCULOSIS IN RABBITEI WITH A CONBISTENTLY Hma RESISTANCE TO THE DISEASE rise in lymphocytes between the 8th and the 14th counts represents the time of the very favorable reaction on the part of the animals, with a subse- quent return to more usual numbers of lymphocytes and monocytes. This later period, from the 14th to the 18th counts, represents the period of leu- coDenia, so that the blood charts still may be used as typical curves for cor- responding cases. For the last group we could not plot any composite type of curve because the reactions of the animals were so varied. In general in this group there were two types of curves; 6rst, a con- dition in which the lines of lympho- cytes cross and recross each other re- 272 Cunningham, Sabin, Sugiyama and Kindwall peatedly, and second, a condition in which both lines are elevated above the normal levels, more or less parallel to each other. These two different tend-. encies in the ratio of the monocytes to lymphocytes in the blood can be found at dilIerent times in the same animal. The final outcome in this group was exceedingly varied, one of the animals showing the most extreme reaction of our series, a second showing marked signs that the infection was arrested, while the rest showed moderate tu- berculosis. From these studies we think that the autopsies of all of our animals demonstrate that in relatively acute tuberculosis. the peripheral blood can give definite information concerning the progress of the disease, which cor- relates quite accurately with the records of loss in weight and general loss of stren`gth. In our series we have found that extremely large numbers of monocytes in the blood accompanied extensive, malignant mononucleosis in the infected tissues; moderately high mono&es have accompanied miliary tuberculosis and tissue mononucleosis of a less marked intensity, while monocytes within normal limits and below normal, together with lympho- cytes markedly above normal, have been correlated with a marked arrest- ing of the disease. Such a relation between a marked increase in two of the types of blood-cells, namely mono- cytes and lymphocytes, in infected tis- sues and an increase of these types of cells in the circulating blood we con- sider a significant point both in rela- tion to tuberculosis and in connection with the relationship of the cells of the connective tissues and the cells of the blood. We have found that widely disseminated infected monocytes represent the most severe lesion of tuberculosis; on the other hand, `we have found that the surrounding of tubercles by masses of lymphocytes is evidence that some healing is taking place. The experiments with B. abortus we did not &d as helpful in relation to the study of tuberculosis as we had hoped. We used the organism alone in two animals and in connection with an infection with tuberculosis in several others. We found that, in the animals infected with B. abortus alone (see Rabbits TB 17 and TB 18, on Table l), although there was an increase in the actual numbers of the monocytes, there was nevertheless a decrease in their percentages. This shows that the increase in mono&es was due to an increase in the total number of the white blood-cells rather than a specific effect on the monocytes themselves. We did not f%rd that an infection with B. abortus increased the intensity of the infection with tubercu- losis, in spite of the added increase in monocytes, as will be seen on Chart 12, Rabbit TB 1, which showed arrested tuberculosis in spite of the secondary infection with B. abortus. DISCUSSION In these studies on blood we are offering as a new concept for an ex- perimental attack on the problem of tuberculosis the theory that the tuber- cle bacillus attacks one specific type of cell-the monocyte-and that the complete analysis of the effect of tu- berculosis on monocytes is a necessary factor for laying a new foundation for progress in the study of this disease. We have found that the effect of tu- R81e of Monocyte in Tuberculosis 273 berculosis on the whole system of the mono&es is a profound one, and con- sists primarily in a marked over-pro- duction of the diffusely scattered reticular cells which are the parent cells of monocytes. This reticular cell is also the parent form of all of the other white blood-cells, so that the effect of the tubercle bacillus on this cell is definitely to force it toward the pro- duction of monocytes. In normal connective tissues the markedly undif- ferentiated reticular cell exists only in small numbers, so that, for example, in a scraping from the freshly cut surface of a normal lung, only an occasiorial cell of this type will be found. At a certain stage in tuberculosis, repre- senting some cycle in the reaction of the body to the disease, these reticular cells may be found in an infected area, such as the septa of a tuberculous lung, in such large sheets as to cover an entire field or more, under an oil-im- mersion lens. The effect of the pro- duction of such an increase in reticular cells and their maturation into mono- cytes must be a chemical one tiecting the cells through the surrounding medium, since neither reticular cells nor very young monocytes have been found containing tubercle bacilli. There is further evidence of this fact in some experiments which we are just beginning on the effect of one of the protein fractions obtained from the tubercle bacilli with which we have injected several animals. The next stage in tuberculosis is the differentiation of vast numbers of monocytes into epithelioid cells; this effect certainly takes place in mono- cytes which have ingested the bacilli, though we have not yet analyzed the phenomenon in terms of cause and effect. It is, of course, known that tubercle bacilli are engulfed by many other cells; for example, they have been found in the Kupffer cells of the liver (Evans, Bowman and Winternits (7)). Maximow (16) has shown tuber- cle bacilli within a wide range of cells in tissue cultures, in lymphocytes as well as in mono&es and in some highly branched cells of the connective tis- sues. But it seems to us that the monocytes can be sharply discrimi- nated from all other cells by this fact that with the monocytes the major effect is on the mono&e itself rather than on the bacillus. We are unable to go so far as to state that the mono- cyte has no power whatever to destroy tubercle bacilli, but we do think the evidence points to the concept that such power, if present, is limited. Thus, histological appearance of living monocytes, in which bacilli have been seen, leads us to the conclusion that the bacilli are alive within the cell; and the numbers which can be seen within an individual monocyte point also to the idea that the bacilli multiply within the cell. The figures of cells shown by Maximow from tissue cultures in- fected with tubercle bacilli bear out this concept, for it will be noted that those cells which are definitely mono- cytes, or the giant cells derived from monocytes, have consistently the greatest number of bacilli within them; a maximum is reached in the cells of his Fig. 21, Plate 5, which have bacilli in uncountable numbers. Such num- bers of bacilli are not seen in the mono- cytes in the body of the animal but it may be that the monocytes of the tis- sue cultures have very much less resist- ance to the bacilli than ever occurs in the cells of an infected animal. We 274 Cunningham, Sabin, Sugiyama and Kindwall suggest that the tubercle bacillus lives as a parasite within monocytes; more- over, it seems to us that the concept that certain bacilli, of which the tu- bercle bacillus is one example, live as parasites within specific types of cells gives a new key for the study of im- munity. The effect of the tubercle bacillus on the monocyte is the accentuation of one of its normal structures and the suppression of another. Under the influence of this infection there is an enormous increase in the number, &c- companied by a decrease in the size, of the fine particles that make the rosette of the normal cell. This rosette then becomes the positive characteristic of the epithelioid cell; the particles that make the rosette are visible in the living cell, they are stainable in neu- tral red, soluble in alcohol but tied by form&n. In Wright's blood stain the zone of the rosette stains a diffuse pink in which the azure granules, typical of the mono&e, can be seen. The pe- ripheral zone of the epithelioid cell is the area occupied by the bacilli; it shows a marked decrease in the stain- able vacuoles; nevertheless, an occa- sional epithelioid cell can take foreign bodies such as a red blood-cell into this zone. The cytoplasm of this pe- ripheral zone shows the first signs of degeneration in the development. of droplets of fat which may ultimately fill the cell completely. The type of the giant cell which comes from the mono&es shows also the characteris- tics of the parent cell or monocyte, for the giant cell arises by an abortive amitosis when the rosette has become so enormous as to inhibit completely the division of the centrosome. There- fore, the giant cell of the Langhams type which has the nuclei in the pe- ripheral zone around the centrally placed rosette, is a derivative of a mono&e. The giant cell shows the ssme type of fatty degeneration as the epithelioid cell. `. . . It is clear that the tubercle %a& lus is not the only stimulus through which the mono&e. may become an epithelioid cell, because typical epithe- lioid cells develop in small numbers, for example, in Hodgkin's disease. We have also studied these epithelioid cells from lymph glands in Hodgkin's disease with the supra-vital technique and have found that they stain like those of tuberculosis, but in our limited experience this reaction is not so in- tense. In this connection, it is very signifi- cant that Lewis, Willis and Lewis (14) have demonstrated the production of epithelioid cells from monocytes by following drops of sterile blood in tissue cultures. They have also con- sidered these epithelioid cells as like those of tuberculous infections. It may be that the development of the epithelioid cells in the tissue cultures is an effect of a decreased nutrition on the part of the cell, due to an unfavor- able environment, which may be mter- preted as indicating that the tubercle bacillus interferes with the nutrition or some other physiological property of the normal monocyte. The effect of tuberculosis on the monocyte is so specific that it seems to us to contribute additional evidence in favor of the concept that the mono- cyte is a different strain of cells from the clasmatocyte. The monocyte is originally a cell in which cytoplasmic pattern is an especial characteristic as contrasted with the clasmatocyte and R61e of Mohocyte in Tuberculosis 275 it is this pattern which is accentuated in the infected monocyte or epithelioid cell. In the septa of sn infected lung, such as the specimen described from Rabbit TB 16, it is true that we saw no clasmatocytes, for the tissues had such an enormous over-production of the epithelioid cells that the flhns looked like pure cultures of these modified monocytes. We believe that addi- tional proof of the concept that the clasmatocytes and monocytes are two different physiological strains of cells could be obtained by taking an animal, in which such a stimulation of mono- cytes had been produced by tuberculo- sis, and giving it repeated doses of carmine or trypan blue until all of the clasmatocytes had been stained. This test we have not yet applied. In our studies on tuberculosis we have found that the effects of tubercu- losis on the monocytes are so profound as to be seen in the monocytes of the peripheral blood; in our animals in which we have induced a practically acute reaction, the monocytes of the peripheral blood became infected. In these mono&es, in which we have demonstrated bacilli, we can follow to some extent the first reaction of the cell to the disease. We were led to record the changes in these monocytes before we had found that they were infected with bacilli. These- mono- cytes have frequently made up about half of the monocytes in the ditreren- tial counts; they have occurred when both the percentages and the actual numbers of the monocytes had been markedly increased. In following the blood of the infected rabbits we have proved that the ratio of monocytes to total lymphocytes is an important law in connection with the disease. We have found that the normal ratio of monocytes to lympho- cytes is 943 to 2805, as shown in Table 2. In the first group of animals which we are showing in Charts 3 to 6, sum- marized on Chart 16, in which the reaction to the disease was consistently unfavorable to the animal, the ratio was 2129 to 1694. This is a higher ratio of monocytes to lymphocytes than is obtained when all of the severe cases, some of which had remissions of the disease, are averaged, as shown in Table 2. In Rabbit TB 16 (Chart 11A) the entire ratio after the infec- tion with tuberculosis was 3469 to 2799, while the ratio during the rmfav- orable reaction at the end was 5139 to 1694, being the average of the last three counts. On the other hand, the ratio for the group of animals shown in Charts 7 to 10, summarized in Chart 17, in which the reaction was consistently favorable to the animal, was 854 to 3047. When these figures are reduced to a ratio with the mono- cytes shown as one, the normal ratio of monocytes to lymphocytes is 1 to 2.97; the unfavorable ratio of Group 1 is 1 to 0.79, while the favorable ratio of Group 2 is 1 to 3.56. These same data are shown graphically in Charts 16 and 17. Thus a marked and continued rever- sal of the normal ratio, provided that the monocytes are markedly increased in actual numbers is, in our experi- ence, a consistent indication of a malignant stage in the reaction of the animal to the disease. The reverse condition is likewise true, for when the lymphocytes are markedly and con- sistently higher than the mono&es, even though the actual numbers of the monocytes are somewhat above 276 Cunningham, Sabin, Sugiyama tind Kindwall normal. we have found arrested tu- berculosii without exception. Our experience corroborated the well known fact that a leucopenia develops in tuberculosis. In the stage of the leucopenia we have found, in some instances, an increase in the percent- age of monocytes or lymphocytes even with an actual decrease in their num- bers, showing that the disease affects primarily mono&es And lymphocytes with a final consistent decrease in the granulocytes. We have found the leu- cope& persisting in animals in which the infection was healed or quiescent `(see Charts 9 and 10). We are thus led to suggest that the effects of tu- berculosis are primarily on the mono- cytes, with secondary effects such as a depression of the entire blood-forming tissues and a loss in weight of the animal. The majority of the rabbits of our experiments have not shown the two extremes of a consistently favorable or consistently unfavorable reaction, but rather have shown a varying reaction in which there have been repeated cros- sings of the lines of the monocytes and lymphocytes in our curves, with inter- vals in which the lymphocytes were rather steadily above the monocytes but in which both were decidedly above normal. We think that such studies give a new key for following the development of an immunity in an animal, but we are well aware that one count a day cannot be relied upon to give more than a general indication of the condition of the blood of the ani- mal, because Sabin, Cunningham, Doan and Kindwall (25) have shown a marked daily rhythm of the white blood-cells. A daily count is, how- ever, able, as the records herein re- corded demonstrate, to bring out marked variations. We therefore con- clude that the law of the ratio of mono- cytes to lymphocytes, in the circulat- ing blood, in acute tuberculosis is of fundamental importance. In these studies we have found that there is a more severe type of infection than miliary tuberculosis, namely, a con- dition of diffuse, spreading mono- nucleosis in which infected monocytes increase in enormous numbers in the tissues, with no reaction on the part of the tissues to wall them off into tubercles. This type of infection, which we have found most markedly in the lungs and in the spleen, has been correlated with the most marked increase in the percentages and in the actual numbers of the monocytes of the circulating blood. For example, our maximum figures in this regard are the ratio of monocytes to lymphocytes in the last count of Rabbit TB 16, as 6348 to 1118, the percentages being 37 to 13 per cent. In Rabbit TB 39 we had the ratio of monocytes to lym- phocytes on 5/20/25 as 5293 to 1449, with percentages 52 to 11 per cent ; and on the next day a slightly greater dis- proportion, namely 5720 to 1210, with the percentages approximately the same. .These figures are extreme ex- amples of an increase in mono&es above the total number of the lymphocytes. We have so far stressed the point of the increase in the monocytes and the reversal of the ratio of monocytes to lymphocytes because we regard that as the ,distinctively new point, but it is no less significant that an increase in lymphocytes has a relation to the healing of tuberculosis. We have con- sistently found, as has been well R61e of Monocyte in Tuberculosis ) 277 known, that lymphocytes surround healing tuberculous lesions, and that a ratio indicating healing tuberculosis is one in which the lymphocytes are above the monocytes and are increased in actual numbers above the average normal limit. We therefore consider that the study of the development of immunity to tuberculosis must include an intensive analysis of the lympho- cytes. For example, it might prove that measures for the stimulation of lymphocytes, which we have already at our disposal, as for instance, cer- tain doses of X-rays, might be used to much greater advantage when the ratio of lymphocytes to mono&es was already favorable. . We think it a significant point, with reference to the close connection be- tween the diffuse connective tissues and two of the strains of the white blood-cells namely, monocytes and lymphocytes, that there should be such a marked change in their numbers in the blood corresponding to an increase in their numbers in the connective tis- sues of an infected area. In the case of the mono&es we have given evi- dence to show that the increase in these cells in the connective tissues is due to a local differentiation of new mono- cytes, from the local overproduction of the reticular cells. From these studies we think that it is clear that the immediate steps for progress in the study of tuberculosis involve the attempt to find the sub- stances by which the change of the mono&es into the epi.thelioid cells is brought about. We believe that there is such a substance and that its dis- covery is of major importance. We think it is clear that the infection with tuberculosis brings about an increase in the number of reticular cells; it is not yet clear whether the `change of the monocyte into the epithelioid cell is initiated by some chemical sub- stance produced by the infection, which affects the cell structurally and permits bacilli to enter it and live in it more easily, or whether these changes are initiated by the presence of bacilli within the cell, or perhaps by both methods. We are, however, quite sure that the major effect of tuberculosis in the body is on one strain of cells, namely, the monocytes; we think that there is every indication that the monocyte is the cell primarily- dam- aged by the infection and that it sub- sequently becomes the host to the invading organism. In a word, tu- berculosis is a disease of the mono- cytes. The discovery of the substances which bring about the increased pro- duction of monocytes and effect their transformation into epithelioid cells, with the idea in mind that the discov- ery of the source of such a substance or substances is the first step toward producing an antibody for them, would be not only rational but a promising program. With a substance by which the overproduction of monocytes might be checked, then the problem of immunity in tuberculosis might be more analogous to, the problem in those diseases which have already been controlled. That this concept is hopeful is shown by an experiment which we have started with one of the fractions, ob- tained by chemical analysis of tubercle bacilli, given to us by the Committee for Research of the National Tubercu- losis Association from the work of Prof. Treat B. Johnson. We took a normal rabbit (TB 40) in our series, in which 278 Cunningham, Sabin, Sugiyama and Kindwall we had a long series of counts of the blood-cells, as a control and gave, daily, 10 mgm. of the fraction intra- peritoneally in 20 cc. of distilled water. After the third dose, there was an in- crease in the percentage of the mono- cytes, from an average before the experiment of 13 to 20 per cent. Twice before the injection of the frac- tion the percentage of monocytes had been high in this animal but the cells on these two occasions were small and not changed in type. On the fourth day after the first injection, when three doses had been given, we found several mono&es such as the one shown in Fig. 10, in which there had been a very marked increase in the fine bodies of the rosette. This is the exact change which we have shown is characteris- tic of the transformation of the mono- cyte into the epithelioid cell. The major problem in connection with the study of the blood at the pres- ent time is the analysis of the normal stimuli which have to do with the con- tinued maturation of the cells of the blood throughout the life of the ani- mal, because the discovery of any fat- tars involved in this normal mechan- ism may lead toward the establishing of methods by which the production or activity of a given type of cell can be controlled. It is clear that all of the experiments involving the study of the effects of various tissue extracts bear on this problem, and that the newer technique of studying living cells both supra-vitally, in cells taken directly from the animal, and in tissue culture, offers a much improved method of analyzing the results of such experiments. The next step in the study of tu- berculosis seems to us to involve an intensive study of methods for con- trolling monocytes; these methods might possibly change the rate of pro- duction of these cells or modify their specific activities. Inasmuch as, in the disease of tuberculosis, there is both an overproduction of monocytes and a marked change in them as well, which we interpret as a lessened power toward the destruction of the bacilli, methods for counteracting this over- production and for stimulating their cytoplasmic activities seem to be es- pecially indicated. Since our evidence indicates that the monocytes phagocy- tize the tubercle bacilli but then fail to destroy them, it seems particularly important to seek for methods of de- creasing the phagocytic activity of these cells or else of increasing their cytoplasmic activities so that the in- gested bacilli will be destroyed instead of being allowed to live and multiply. The approaches that seem to us feasi- ble are further studies with extracts made from tissue cultures of blood in which there had been a great overpro- duction of monocytes, according to the methods of Lewis and Lewis and of Carrel; of .tissue extracts made from the lungs of such an animal as our Rab- bit TB 38 in which the tuberculous infection had given rise to a very great overproduction of the immature forms of the mono&es; the study of the effect of the various chemical fractions from the bacilli such as we have re- ferred to above, and the study of the effect of various extraneous chemical compounds which may be found to affect monocytes. Thus through a co- operation between the modern chemi- cal studies on the nature of immunity and the modern experimental analysis of the physiology of the cells of the R61e of Monocyte in Tuberculosis 279 blood and the connective tissues there next steps in tuberculosis involve study- has developed a new method of attack ing the various chemical factors in the on the problem of tuberculosis. The disease in their effect on monocytes. REFERENCES (1) (3) (5) (f-3 (7) BENSLEY, R. R.: Studies on the pan- creas of the guinea pig. Amer. Jour. of Anat., 1911-1912, xxi, 297. BOUFFARD, G. : Injection des couleurs de benzidine aux animaux normaux. Ann. de l'lnst. Pasteur, 1966, xx, 539. CUNNINQHAY, R. 5.: On the origin of the free cells of serous exudates. Amer. Jour. of Physiol., 1922, lix, 1. CUNNINQEAM, R. S., F. R. SATIN, AND C. A. DOAN: The diiIerentiation of two distinct types of phagocytic cells in the spleen of the rabbit. Proc. Sot. Exper. Biol. and Med., 1924, xxi, 326. CUNNINQHAY, R. S., F. R. SABIN ~UJD C. A. DOAN: The development of leucocytes, lymphocytes, and mono- cytes from a specific stem cell in adult tissues. Con&i. to @nbtyoE., No. 84, Carnegie Inst. of Wash., 1925, xvi, 229. DOAN, C. A., R. 5. CUNNINGHAM AND F. R. SABIN: Experimental studies on the origin and maturation of avian and mammalian red blood-cells. Con&i. to Embryol., No. 83, Carnegie Inst. of Wash., 1925, xvi, 165. EVANS, H. M., F. B. BOWMAN AND M. C. WINTERNITZ : An experimental study of the histogenesis of the miliary tubercle in vitally stained rabbits. Jour. Exper. Med., 1914, xix, 283. (8) EVANS, H. M. AND W. SCHULEMANN: The action of vital stains belonging to the benzidine group. Science, 1914, xxxix, 443. (9) )ZVANS, H. M., AND K. J. SCOTT: On the differential reaction to vital dyes exhibited by the two great groups of connective tissue cells. Contri. to Embryol., No. 47. Car- negie Inst. of Wash., 1921, x, 3. 110) FOOT, N. C.: The endothelial phago- cyte. A critical review. The Anat. Record, 1925, xxx, 15. (11) GOLDMANN, E. E.: Die innere und iiussere Sekretion des gesunden und kranken Organismus im Lichte der vitalen FIirbung. Theil 1, Beitr. z. klin. Chir., 1969, xliv, 192. Theil 2, ibid., 1912, lxxviii, 1. (12) HE)RZOQ, F.: Ueber Capillarendothe- lien ala Wanderzelle. Klin. Wchn- schrr, 1924, iii, 91. (13) KEY, J. A.: Studies on erythrocytes, with special reference to reticulum, polychromatophilia and mitoohon- dria. Arch. Intern. Med., 1921, xxviii, 511. (14) LEWIS, M. R., H. s. WILLIS AND W. H. LEWIS: The epithelioid cells of tu- berculous lesions. "Tubercle." 1925, p. 1. (15) M~~IYOW, A.: Untersuchungen fiber Blut und Bindegewebe. I. Die friihesten Entwicklungsstadien der Blut und Bindegewebszellen beim Siiugetierembryo, bis zum Anfang der Blutbildung in der Leber. Arch. Mikro. And., 1969, lxxiii, 444. (16) MAXMOW, A.: Tuberculosis of mam- malian tissue in eitro. Jour. In- fect. Dis., 1924, xxxiv, 549. (17) MCJUNKIN, F. A.: A simple method for the demonstration of a phago- cytic mononuclear cell in peripheral blood. Arch. Int. Med., 1918, xxi, 59. (18) MCJUNKIN, F. A.: The origin of pha- gocytic mononuclear cells. Amer. Jour. Anal., 1919, xxv, 27. (19) PERMAB, H. H.: An experimental study of the mononuclear phago- cytes of the lung. Jour. Med. Re- search, 1926, xlii, 9. (26) PERMAB, H. H.: The development of the mononuclear phagocyte of the lung. Ibid., 1929, xlii, 147. (21) RIBBERT, H.: Die Abscheidung in- traven6sinjizierten gelizten Kar- mine in den Geweben. Ztschr. f. allg. Physiol., 1004, v, 201. Cunningham, Sabin, Sugiyama and Kindwall (22) SABIN, F. R.: The vitally stainable granules as a specific criterion for erythroblasts and the diflerentiation of the three strains of the white blood-cells as seen in the living chick's yolk-sac. Johns Hopkins Hosp. Bull., 1921, xxxii, 363. (23) SABIN, F. R.: Studies on the origin of blood-vessels and of red blood- corpuscles as seen in the living blas- ' toderm of chicks during the second day of infubation. Con&i. to Em- bryol., No. 36, Carnegie Inst. of Wash., 1922, ix, 215. (24) &BIN, F. R.: Studies of living human blood-cells. Johns Hopkins Hosp. Bull., 1923, xxxiv, 277. (25) SABIN, F. R., R. S. CVNNINQHAY, C. A. DOAN AND J. A. KINDWALL: The rhythm of the white blood-cells. Johns Hopkins Hosp. Bull., 1925, xxxvii, No. 1,14. (26) &BIN, F. R., C. A. DOAN AND R. S. CUNNINQHAM. : The separation of the phagocytic cells of the peritoneal exudate into two distinct types. Proc. Sot. &per. Biot. and Med., 1924, xxi, 339. (27) SABIN, F. R., C. A. DOAN AND R. S. CUNNINQHAM. : Discrimination of two types of phagocytio cells in the connective tissues by the supra- vital technique. Co&i. to Embryol., No. 82, Carnegie Inst. of Wash., 1925, xvi, 125. (28) SCHILLINQ-TOBQAU, V. : Arbeiten iiber die Erythrozyten (ii to vii). Fat. Haematol., 1912-13, xiv, 95. (29) SHIPLEY, P. G.: The physiological significance of the reaction of tissue cells to vital benzidine dyes. Amer. Jour. Phyeiol., 1919, xlix, 284. PLATE 1 FIQ. 1. Modified monocyte from the circulating blood of Rabbit TB 49. Intraperitoneal injection of tubercle bacilli on 5/B/25; cell drawn 6/23/25. Cell was supra-vitally stained with neutral red. Mitochondria are shown in black. The magnification of all the cells is indicated by the accompanying red blood-cell. FIQ. 2. Mono&e of the circulating blood of Rabbit TB 43, showing the earliest signs of the change into an epithelioid cell. Intravenous injection of tubercle bacilli on 5/29/25; cell drawn on 6/19/25. Cell was supra-vitally stained with neutral red. FIQ. 3. Epithelioid cell from a scraping of the liver of Rabbit TB 36, drawn immediately after the animal had been killed on g/4/25. Data on the blood on Table 1. The peripheral blood on that day had 40 per cent monocytes. The tissues from the liver showed many of these very young epithelioid cells. This cell had two nuclei and many mitochondria in the cytoplasm. Cell was supra-vitally stained with neutral red and Janus green. Fro. 4. Epithelioid cell from a scraping of the lung of Rabbit TB 33, drawn on 6/2/25. Data on the blood on Table 1. The peripheral blood on that day showed 17 per cent mono- &es. This was a ?very young epithelioid cell, with a single vacuole stained red in the periphery of the rosette and one refractive lipoid body, shown in white, outlined in black. Supra-vitally stained with neutral red and Janus green. FIQ. 5. Epithelioid cell from a scraping of the lung of Rabbit TB 16, drawn 4/l/25. Data on the blood in Table 1. The peripheral blood on that day showed 37 per cent mono- oytes (see Charts 1lA and 11B). This was a typical mature epithelioid cell with marked rosette of fine bodies and wide peripheral zone in which two bacilli were seen. Films from this lung stained for tubercle bacilli showed that most of them were within the epithelioid cells. Supra-vitally stained with neutral red and Janus green. BULLETIN OFTHEJOHNS HOPKINS HOSPITAL PLATE 1 `\ .I .,' . . _ `I 3 . .*. --.y /' 7 I, ` `\ ; `. .,.,, \ `i _ `. Y. `/ \ -.,. L `e b ,,;-.:- . : .: ,. ' i . . . . / BULLETIN OF THE JOHNS HOPKINS HOSPITAL PLATE 2 . _' ,' . . `I.. ____, ,d"_`-