Chapter VIII: Survey of the Facts Bearing on Acquired Immunity Against Micro-Organisms (2)

After his discovery of the granular transformation of vibrios, R. Pfeiffer, in collaboration with several of his pupils, set himself to discover how far this phenomenon was general in acquired immunity. He directed his attention to the typhoid cocco-bacillus, upon which he had already published[328] a very detailed account of work carried out in conjunction with Kolle. These observers availed themselves of the discovery made by Beumer and Peiper[329], and Chantemesse and Widal[330] and confirmed by other observers, that laboratory animals, especially mice and guinea-pigs, could be easily vaccinated against the fatal disease set up by the micro-organism of typhoid fever. As in the experimental infection of the guinea-pig by the cholera vibrio, the vaccination of the animals against the typhoid bacillus could be carried out very easily, either by using sterilised cultures or the fluids of cultures deprived of their organisms by filtration. In the small laboratory animals a most marked acquired immunity may thus be obtained, and the study of the phenomena which appear in the vaccinated organism afforded evidence of a general analogy with those which have been observed when vibrios are used. In the peritoneal cavity of the immunised guinea-pigs, Pfeiffer’s phenomenon proper does not appear, that is to say, only a few of the bacilli are transformed into granules, the large majority retaining their bacillary form; still they are evidently greatly damaged: they become motionless and agglutinate more or less completely into clumps. If, however, a few of these micro-organisms are sown on nutritive media, they multiply freely and give abundant growths. The peritoneal fluid, then, acts most unmistakably upon the typhoid bacillus, but in a much less degree than does the peritoneal exudation of guinea-pigs upon the cholera vibrio when immunised against that organism. In both cases we have a pronounced phagolysis which sets free the microcytase, whose action on the vibrio is more marked than on the bacillus of typhoid fever. This extracellular action on the typhoid bacillus in the peritoneal cavity can be easily prevented by a previous injection, twenty-four hours before, of broth, physiological salt solution, or normal serum. The suppression of phagolysis is, as in the case of vibrios and spirilla, followed by the suppression of extracellular action on the typhoid bacilli.

The same analogy is observed in the phenomena which appear beneath the skin. The bacillus of typhoid fever, when introduced into the subcutaneous tissue of vaccinated guinea-pigs, although not appreciably injured by the fluid of the exudation, undergoes some agglutination. The injurious action of the fluids of the body is here still less effective than in the peritoneal cavity. But, as in the peritoneal cavity of vaccinated guinea-pigs previously treated with broth, so in the subcutaneous exudation it is the phagocytes which destroy the micro-organisms. In both cases there is a very great afflux of leucocytes, mainly microphages. These cells ingest and digest the bacilli, which ultimately disappear. The micro-organisms ingested by the microphages, once inside these phagocytes are transformed into granules very like those observed in the cholera vibrio similarly treated. In this respect the analogy between the two micro-organisms is complete.

[Sidenote: [244]]

Oppel, working in my laboratory, has repeated Cantacuzène’s work on the retarding action of opium upon the phagocytic process. He obtained the same results: under the influence of the narcotic, the leucocytes intervened only at a late stage, with the result that the vaccinated guinea-pigs succumbed to the typhoid infection. The same conclusion must be drawn from the experiments made by A. Wassermann[331]. Guinea-pigs that had been immunised against the bacillus of typhoid fever were completely resistant to a dose that was always fatal to the control animals. When, however, along with this dose of bacilli, a certain quantity (3 c.c.) of a serum which hinders the phagocytic reaction is injected, the guinea-pigs lose their immunity and die from typhoid infection. The serum employed by Wassermann was obtained from rabbits that had been treated with the blood serum of guinea-pigs. Rabbit’s serum, thus prepared, neutralises the action of the guinea-pig’s cytase, but, as demonstrated by Besredka[332], it also exercises several other functions, one especially, that of preventing phagocytosis. In Wassermann’s experiments it was the antiphagocytic function, then, that was the important factor in the suppression of the acquired immunity of the guinea-pigs. These experiments supply a fresh proof of the great importance of the phagocytic reaction in this kind of immunity, and afford further confirmation of the analogy between the mechanism of resistance of the animal’s organism against the typhoid bacillus and that against the cholera vibrio.

In presence of this striking analogy, it is unnecessary to insist further on the details of the acquired immunity of animals against the experimental disease set up by the micro-organism of typhoid fever. It will be better to select another example from the group of bacilli. Let us first take the acquired immunity against the bacillus of blue pus (_Bacillus pyocyaneus_) which for many years has been regarded as the best example in which to study this kind of immunity. Charrin, who was the first to obtain disease with this bacillus experimentally, published several notes[333] on the acquired immunity of the rabbit against it. He demonstrated the possibility of vaccinating this animal not only with living bacilli, but also with the products of their culture; he studied the blood serum of vaccinated animals, comparing it with the serum of normal rabbits, especially as to its action on the development of the _Bacillus pyocyaneus_. Although unable to find any bactericidal power properly so called in the serum of immunised rabbits, Charrin was the first to draw attention to certain modifications undergone by the bacilli when grown in this medium. He noted that under these conditions no pyocyanin was produced, and, in collaboration with Roger, he demonstrated that, in the serum of the vaccinated rabbit, the Bacillus pyocyaneus forms packets composed of little chains of greater or less length, whilst in the serum of the normal, susceptible rabbit, it develops in the form of normal rods, the rods for the most part being isolated.

[Sidenote: [245]]

From his experiments _in vitro_ Charrin concluded that there was marked enfeeblement of the functions of the _Bacillus pyocyaneus_ when submitted to the action of the vaccinated animal organism. Bouchard[334] has gone so far as to develop a theory of acquired immunity, in which the principal part is attributed to the impossibility of the micro-organism, after it has invaded the refractory animal, secreting its fluid products; there is no vascular dilatation and diapedesis does not take place. A comparative observation of the phenomena observed in rabbits that are susceptible to the pyocyanic disease and of those met with in vaccinated rabbits, most clearly, however, demonstrates the impossibility of accepting Bouchard’s interpretation. The inoculation of the bacillus of blue pus below the skin of the ear of the normal (unvaccinated) rabbit sets up extensive inflammatory reaction with marked hyperaemia; the diapedesis of the white corpuscles takes place at a comparatively late stage of the process and phagocytosis is neither set up nor completed until very late. On the other hand, in vaccinated rabbits, infected in the same way, the hyperaemia of the ear is insignificant, but diapedesis occurs very early and phagocytosis commences at once. It is not, therefore, the impossibility for the leucocytes to traverse the vessel wall, owing to the absence of the dilatation of the veins, which prevents them from making their way rapidly to the field of battle; it is their imperfect positive sensitiveness that is accountable for the tardy and incomplete phagocytosis. This interpretation is confirmed in other cases of acquired immunity.

[Sidenote: [246]]

More recently, Paul Müller[335] has laid special stress on the part played by the bactericidal action of the serum of animals that have been vaccinated against the pyocyanic disease. For him the negative results obtained by his predecessors lose their significance, since all their experiments were carried out under conditions of aërobiosis, whilst it is only in the absence of free oxygen that this bactericidal power can be exerted at all freely. Müller, therefore, set himself to compare under anaerobic conditions the bactericidal action on the _Bacillus pyocyaneus_ of serums coming from normal and from vaccinated animals. He succeeded in demonstrating that the blood serum of vaccinated animals is more bactericidal than that of normal rabbits. Before, however, drawing any conclusion from this observation, the following question must be answered: Are the phenomena observed _in vitro_ comparable with those seen in the living animal? In preceding chapters it has been shown so often that the blood serum obtained after the separation of the extravascular clot, can in no way be identified with the plasma of the circulating blood, that it is unnecessary to argue this matter further. If we wish to gain a clear idea of the mechanism of immunity in the living animal we must observe the course of events in the vaccinated animal and not draw conclusions from observations _in vitro_ except after strict examination. All the works on pyocyanic immunity above summarised lie under the reproach that in them this rule has not been adhered to.

Since the discovery of Pfeiffer’s phenomenon in animals that have been vaccinated against the cholera vibrio, much greater care has been taken to attend to the changes that occur in the animal that enjoys acquired immunity. Wassermann[336] was the first to attempt to apply Pfeiffer’s discovery to the _Bacillus pyocyaneus_. With a race of this bacillus rendered more virulent he succeeded in producing a fatal experimental malady in the guinea-pig against which he was able by various methods to vaccinate these animals.

[Sidenote: [247]]

He thus describes the phenomena observed in the peritoneal cavity of immunised guinea-pigs. Soon after injection the bacilli of blue pus become motionless, then “the rods swell up and melt, like wax in hot water. The formation of granules, such as occur in the cholera vibrio, has been observed but rarely. The process recalls rather that which takes place in experimental typhoid fever, as described by R. Pfeiffer. In all cases the phenomenon of solution takes place entirely in the fluid of the exudation, without any co-operation on the part of the leucocytes” (p. 284). We see that we have still to do with a kind of attenuated Pfeiffer’s phenomenon, without any granular change, but with an immobilisation of the bacilli. As Wassermann has remained satisfied with the examination of the peritoneal content which, as we know, gives but an imperfect picture of acquired immunity, Gheorghiewsky[337] set himself to study the question more thoroughly under my direction. With this object he vaccinated a series of guinea-pigs with living bacilli of blue pus, a sure method of obtaining acquired immunity. On examining the peritoneal fluid (withdrawn shortly after the injection of the bacilli) of the vaccinated guinea-pigs, he found that the bacilli were motionless and had undergone a certain degree of agglutination. They were not transformed into granules but became thicker and somewhat more dumpy. These changes are observed during the period of phagolysis, when only a few scattered leucocytes are to be found in the fluid of the peritoneal cavity. About two hours after the injection of the bacilli the leucocytes begin to reappear in the peritoneal exudation, more especially the microphages, which lose no time in seizing the bacilli, some of which become transformed into granules. A few hours later the exudation, containing a multitude of leucocytes, no longer contains any free bacilli: all are found inside the microphages. Nevertheless, if a drop of the exudation now be withdrawn and kept for some time at a temperature of 37° C., it will be found that the bacilli multiply inside the dead phagocytes outside the animal. We thus obtain colonies of bacilli, a fact which clearly proves that these bacilli whilst still alive have been ingested by the leucocytes. This experiment is, therefore, very similar to the one we have described in connection with Gamaleia’s vibrio.

[Sidenote: [248]]

Even at a later period, 24 or 30 hours after the injection of the bacilli, that is to say at a period when an examination of the exudation no longer reveals the presence of bacilli, the sowing of a drop of this exudation on a nutrient medium still gives isolated colonies of the _Bacillus pyocyaneus_ capable of producing the characteristic pigments. At a still later period, when the peritoneal exudation remains sterile, a post-mortem examination of the animals enables one to recognise, beneath the peritoneal surface, small white points made up of leucocytes. The sowing of these masses almost invariably gives colonies of the _Bacillus pyocyaneus_ which form blue pigments. We see from this account that, even in the peritoneal cavity of vaccinated animals, matters by no means go on in a uniform fashion, as would appear from Wassermann’s statements. Some bactericidal action in the peritoneal fluid there certainly is, but it is quite transient, and is limited to the period of phagolysis. The majority of the bacilli resist this attack of the body fluids to continue their struggle with the phagocytes, which, however, ultimately get the upper hand. In the subcutaneous tissue the part played by this phagocytic reaction is still more general. Gheorghiewsky has studied it not only in vaccinated guinea-pigs but also in a goat which had received several large injections of the _Bacillus pyocyaneus_. He observed that shortly after the subcutaneous injection of these bacilli, the fluid which accumulates at the seat of inoculation renders them motionless and in part agglutinates them. This fluid is clear and contains a few leucocytes and a number of bacilli which still retain their usual form. Some time later the leucocytes begin to come up to the seat of inoculation and to ingest the bacilli. At the end of 10 to 15 hours all the bacteria have been seized by the microphages and we no longer find any of them free. A hanging drop of this exudation, transported to the incubator, soon swarms with bacilli which have sprung from the organisms ingested by the leucocytes.

The exudation becomes more and more abundant at the seat of inoculation and ends in the formation of an abscess, from the contents of which cultures of the _Bacillus pyocyaneus_ may be obtained for a fortnight. The bacilli, however, finally disappear, this being due to the destructive action of the phagocytes and not to that of the fluid of the exudation.

This fundamental part played by phagocytosis in acquired immunity against the _Bacillus pyocyaneus_ has been confirmed by Gheorghiewsky by experiments on guinea-pigs vaccinated and then submitted to the action of opium. As in the analogous experiments of Cantacuzène on the cholera vibrio, the opium narcosis retards diapedesis and this, for some time, increases the chances of the bacilli. A tardy diapedesis and phagocytosis, no doubt, is produced which ends in the ingestion of the bacilli, but the animal loses its acquired immunity and finally succumbs in spite of the fact that the dose of _Bacillus pyocyaneus_ was insufficient to kill a control guinea-pig vaccinated to the same degree, but not submitted to the action of opium.

[Sidenote: [249]]

The example we have just analysed relates, then, to a micro-organism which is more resistant than are the vibrios, Obermeyer’s spirilla or even the typhoid bacillus, to the action of the microcytase which has escaped from the cells during phagolysis. The _Bacillus pyocyaneus_ undergoes, in the fluids of the vaccinated animal, the action of the specific fixative and can thus be rendered motionless and become agglutinated. But this action is insufficient to ensure immunity and should phagocytosis not take place in time to ingest the bacilli, the vaccinated animal succumbs. The reaction of the phagocytes is, therefore, indispensable if the acquired immunity is to be effective. In this respect the analogy is very great between the resistance of the vaccinated animal against the various bacteria (vibrios, spirochaetes, typhoid cocco-bacilli, bacilli of blue pus) that we have so far studied in this chapter. These bacteria have, however, this in common;—they are all endowed with a considerable power of motion. Pursuing our examination of the principal data on acquired immunity against micro-organisms, we must now choose examples from the group of non-motile bacilli; amongst these we assign the first place to the micro-organism of swine erysipelas. This small bacillus has been the subject of several important researches on acquired immunity, one of which at a certain period caused quite a sensation in the bacteriological world. Emmerich[338], in an investigation carried out in collaboration with di Mattei, made an unexpected announcement. He said he believed that he was justified in affirming that the acquired immunity of rabbits against the bacillus of swine erysipelas is due to the formation, in the fluids of the body, of an antiseptic substance which very quickly destroys this organism. This substance, secreted by the cells of the vaccinated animal, was supposed to act after the fashion of a solution of bichloride of mercury and to kill a large number of bacilli, introduced subcutaneously, in from 15 to 25 minutes. This discovery was not confirmed. In a series of experiments that I carried out[339] with the object of clearing up this question, and made under conditions as favourable as possible for the demonstration of the supposed bactericidal secretion, this action was never manifested. Not only did the virulent bacilli of swine erysipelas, when injected subcutaneously into well vaccinated rabbits, remain alive in the subcutaneous exudation for hours and even days, but the attenuated bacilli of Pasteur’s vaccines likewise remained intact. These bacilli when introduced into the anterior chamber of the eye survived for even a longer period. Here, as beneath the skin, the injection of the bacilli induced an exudation rich in leucocytes, amongst which microphages predominated. These phagocytes at once began to seize the bacilli which were destroyed not in the fluid of the exudation but within the leucocytes. Long after all the bacilli had been ingested, 24 hours and more after inoculation, the sowing of the exudation frequently gave growths in appropriate media.

[Sidenote: [250]]

Emmerich[340] sought by new experiments to remove these objections, but he found that the bacilli of swine erysipelas did not disappear from the vaccinated animal until some 8 or 10 hours after they had been introduced. There is, therefore, no longer any question of a rapid bactericidal action at all comparable to that of corrosive sublimate, which would destroy the introduced bacilli in less than an hour. The limit of 8 to 10 hours, accepted by Emmerich, is still too short and is not in accordance with my experiments; but even this was quite sufficient for the appearance of a free phagocytosis, a condition that really occurs. Emmerich has not directed his researches in this direction, and his theoretical conclusions did not in the least weaken the value of my arguments drawn from the demonstration of the ingestion and intracellular destruction of the bacilli by phagocytes.

[Sidenote: [251]]

The researches on immunity against swine erysipelas then languished for some time, until the discovery of Pfeiffer’s phenomenon gave a fresh stimulus to the study of this problem. One of Pfeiffer’s pupils, Voges[341], sought to apply the results obtained in the case of the cholera vibrio to the acquired immunity against the bacillus of swine erysipelas. He studied the blood serum of animals vaccinated against this bacillus and believed himself justified in affirming the existence of an acquired bactericidal power. Under no condition, however, did he observe anything comparable to Pfeiffer’s phenomenon, and he was compelled to admit that the bactericidal action of the serum is very feeble and only takes effect on young bacilli whose membranes are as yet very delicate and not very resistant. Mesnil[342] repeated these researches in my laboratory, but his results were very different from those obtained by Voges. The blood serum of rabbits, fully vaccinated against the bacillus of swine erysipelas, proved to be a good culture medium for this bacillus, and Mesnil affirms, as the result of numerous well-established observations, that “_in vitro_, the serum of rabbits immunised against the erysipelas has no bactericidal power or a very insignificant one.” On the other hand, the same fluid had a very marked agglutinative power. The bacillus of swine erysipelas, being non-motile, does not present the abrupt change that is observed in vibrios or in the typhoid bacillus when submitted to the influence of specific serums—under which conditions these organisms at once lose their motility. But the bacilli of swine erysipelas, when introduced into the specific serum of vaccinated animals, run together into masses which become more and more voluminous and fall to the bottom of the vessel, leaving a limpid supernatant fluid. When this bacillus is sown in the serum of vaccinated animals, it is seen to develop in the form of chains, composed of a large number of segments, which fall to the bottom of the tube. These bacilli, however, whether agglutinated or developed in chains, never show any attenuation in virulence. When the serum which bathes them is got rid of by washing, they are just as virulent as are the bacilli developed in the serum of normal unvaccinated rabbits. It is important to show that this virulence is kept up in spite of the fact that the bacilli, when placed in contact with the serum of immunised animals, become permeated with the specific fixative, as shown by the experiments of Bordet and Gengou[343]. These observers, indeed, have demonstrated that the bacilli of swine erysipelas, when kept for 24 hours in the specific serum heated to 55° C., acquire the property of absorbing the cytases contained in the unheated serum of normal animals.

The study of acquired immunity against the bacillus of swine erysipelas teaches us that this immunity is not due to any extracellular destruction comparable with Pfeiffer’s phenomenon; and that this immunity causes the production of a specific fixative and of a specific agglutinative substance, whose action on the resistance of the animal, to judge from the complete virulence of the bacilli when agglutinated and impregnated by fixative, is feeble or _nil_. It is the phagocytic reaction which is dominant in the immunised animals and which brings about the intracellular destruction of the bacilli.

The history of the anthrax bacillus, another representative of the group of non-motile bacilli, is particularly interesting, the more so that for some time the researches on acquired immunity have been concentrated almost entirely on the analysis of the facts observed in animals that have been vaccinated with the two Pasteur vaccines. In this way a large number of valuable facts have been collected; of these the more important may be presented to the reader.

[Sidenote: [252]]

In my first work on this subject[344] I called attention to the fact that in the rabbit vaccinated against anthrax, the bacilli, when inoculated subcutaneously, soon become the prey of leucocytes which accumulate at the spot menaced. In the unvaccinated control rabbits, however, the anthrax bacilli remain in a free state in the fluid of the subcutaneous exudation, only a few isolated rods being found inside phagocytes. I have since been able to confirm this fact[345], which must now be regarded as fully established. In the vaccinated rabbits the leucocytes exhibit a very marked positive chemiotaxis against the anthrax bacilli, whilst in normal unvaccinated rabbits the chemiotaxis of the leucocytes in the anthrax of the subcutaneous tissue is distinctly negative. When a small quantity of anthrax culture is inoculated subcutaneously into vaccinated and into unvaccinated rabbits there may be observed, even within a few hours, a very great difference. In the former there is found at the seat of inoculation an infiltration which swarms with leucocytes in the act of devouring bacilli. In the normal, susceptible rabbit, on the other hand, the exudation produced is soft, rich in fluid, and very poor in leucocytes. The vessels in the vicinity are distended with blood, and the fact that the leucocytes do not come up to the seat of inoculation is in no way due to the absence of vascular dilatation which might prevent diapedesis. The vessels are much more dilated than in the vaccinated rabbit, and yet in the latter the emigration is incomparably greater. This essential difference must be attributed to the sensitiveness of the leucocytes, which exhibit a negative chemiotaxis in the normal rabbit but a very marked positive chemiotaxis in the vaccinated rabbit.

It has been shown repeatedly that the subcutaneous exudation, very rich in leucocytes which have had time to ingest all the bacilli, when inoculated into guinea-pigs, ensures the appearance in them of a generalised and fatal anthrax; this affords evidence that the phagocytosis is exercised against virulent and therefore living bacilli. Marchoux[346], in Roux’s laboratory, has carried out numerous experiments on the vaccination of rabbits and has observed that the inoculated anthrax bacilli cause an exudation very rich in leucocytes, and that these cells ingest and destroy the bacilli. The phagocytes easily rid the refractory animal of the bacilli in the vegetative state, but the spores are much more resistant. After being devoured by the leucocytes they may remain inside them for months without germinating. Marchoux obtained cultures of anthrax from the subcutaneous exudation taken from vaccinated rabbits 70 days after inoculation.

[Sidenote: [253]]

The fact that the bactericidal action of the blood serum on anthrax bacilli is specially well marked in the rat, suggested the idea of trying to obtain, in this rodent, an augmentation of this property as a result of vaccination. Sawtchenko[347] attempted to do this in an investigation already cited in Chapter VI, carried out in my laboratory. He succeeded in thoroughly vaccinating white rats against virulent anthrax and in showing that the blood serum of these animals rendered refractory “is bactericidal in the same degree as that of non-immunised rats.” In the vaccinated rats “the subcutaneous exudation was as free from bactericidal substances as was the lymph of the control animals.” Sawtchenko was unable to demonstrate any increase of bactericidal power except in the peritoneal exudation of rats vaccinated by injection of cultures into the peritoneal cavity.

In spite, however, of the absence of any increase in the bactericidal property of the blood serum and of the subcutaneous exudation in vaccinated rats, the cell reaction obtained in them is very different from that met with in normal, susceptible rats. In a very short time (3 to 5 hours) after the subcutaneous injection of anthrax bacilli into the control rats (susceptible), an evident oedema is produced; in the vaccinated rat there is none. The exudation, not very abundant in the latter, already contains a number of leucocytes which are actively phagocytic, whilst in the control animal, examined simultaneously, “leucocytes are rarely met with, and few of them contain bacilli.” Later, the difference becomes still more marked. Pronounced oedema occurs in the control animal, it is poor in leucocytes but rich in bacilli, which continue to multiply; but “in the immunised rat, we find not a clear exudation but a thick and purulent fluid, full of leucocytes.” These cells devour all the bacilli; not a single one remains free. “Even after 14 hours bacilli ingested by the leucocytes are present and a culture of anthrax bacilli may be obtained from fluid taken from the seat of inoculation. Further, guinea-pigs or rats, when inoculated with a drop of this exudation (which contains no anthrax spores), succumb to anthrax.”

[Sidenote: [254]]

Even before these researches on the immunity of rats had been carried out, an attempt had been made to gain some idea of the differences presented by the vaccinated fluids of animals as compared with those presented by the fluids of control animals susceptible to anthrax. In 1886 I was able to demonstrate[348] that the anthrax bacillus develops abundantly in the defibrinated blood of sheep that had acquired immunity as the result of vaccination by Pasteur’s method. When these bacilli contain spores and are inoculated into rabbits they rapidly produce a fatal anthrax; but when no spores are present the injection of bacilli does not produce a fatal disease, and such infection is well supported by the rabbits. From this I concluded at that time that the anthrax bacillus must, in the blood of the vaccinated sheep, undergo a real attenuation in virulence, an interpretation which, as will be seen in the next chapter, was found to be erroneous.

Nuttall[349] showed that the defibrinated blood of refractory sheep acted as a nutrient medium for the anthrax bacillus. Making comparative investigations, by the plate method, on the bactericidal power of the blood of vaccinated and normal sheep, he observed that, in both cases, there was, at first, a certain decrease in the number of bacilli sown, more marked in the blood of the vaccinated than in that of the control animals. Nevertheless, 8 hours after the commencement of the experiment the anthrax bacteria had produced innumerable bacilli in the blood of the refractory sheep. Nuttall satisfied himself that this feeble bactericidal power was not to be compared with the very much greater power of the blood of the rabbit, an animal specially susceptible to anthrax.

More recently the properties of the serum of sheep which have been vaccinated against anthrax have been studied very carefully by Sobernheim[350]. He also was able to show that this serum allows of an abundant development of the bacillus, and that, outside the animal, it does not exercise any more appreciable bactericidal power than does the serum of the normal sheep. The serum of the best vaccinated sheep was found to be incapable of destroying even very small quantities of anthrax bacilli. The only change that Sobernheim could make out was with regard to the thickening of the bacterial membrane. This modification, however, was not constant and could not be seen in the serum of certain vaccinated sheep.

[Sidenote: [255]]

The serum of the sheep vaccinated by Sobernheim exhibited no increase of agglutinative power as regards virulent bacilli. Gengou[351], however, made it clear that repeated injections of cultures of the first vaccine of Pasteur into dogs produced a marked augmentation of this agglutinative power; but it was only produced when the attenuated bacillus was used. The virulent anthrax bacillus, developed as isolated rods, was not affected in the least by serum that was highly agglutinative for the bacillus of the first vaccine. Gengou also made the converse experiment with the serum of a dog into which he had previously injected a number of virulent anthrax bacilli. The dog, naturally refractory to anthrax, resisted the inoculation perfectly, but its serum did not acquire any agglutinative power against the first vaccine. He concluded therefrom that “the part played by agglutinins in the defence of the animal must be regarded as extremely problematical” (p. 339). On the other hand the phagocytic reaction in the vaccinated sheep is always very pronounced and constant. Von Behring[352], in one of his most recent publications, expresses the opinion that this example of acquired immunity must be placed in the category of phagocytic immunity.

In the group of bacilli, several examples of which we have studied, the typhoid bacillus approaches still more closely to the vibrios and spirilla in its relation to humoral properties. Here may be observed a kind of attenuated Pfeiffer’s phenomenon and somewhat profound modifications taking place under the influence of the serum of vaccinated animals. The _Bacillus pyocyaneus_ is more resistant to the injurious influence of fluids taken from immunised animals. This resistance is still more marked in the bacillus of swine erysipelas and again still greater in the anthrax bacillus. Whilst, however, these properties of the fluids of the body are found to be very variable and of unequal power, the phagocytic reaction is constantly manifested and always very actively. The leucocytes which, in susceptible animals, exhibit a very marked negative chemiotaxis or only a tardy and incomplete positive chemiotaxis, have, in the vaccinated animal, this positive susceptibility developed in a very high degree.

[Sidenote: [256]]

Before quitting the group of bacteria we must cast a glance at the mechanism of acquired immunity against representatives of the group of spherical micro-organisms. Amongst the cocci the streptococci have been especially studied as regards this immunity. For long great difficulties were encountered in vaccinating animals against these chain cocci, but Roger[353], Marmorek[354], Denys and Leclef[355] overcame these obstacles and succeeded in immunising the rabbit, one of the most susceptible species, to their pathogenic action. More recently the larger mammals, notably the horse, have been successfully immunised. A certain number of important facts, the knowledge of which is useful to complete the survey of the phenomena of acquired immunity, have thus been collected.

[Sidenote: [257]]

Roger set himself to study the properties of the blood serum of rabbits vaccinated against the streptococcus, and established the fact that this fluid had not the slightest appreciable bactericidal action; the streptococcus grew in it just as well as in the serum of fresh unvaccinated rabbits. When, however, he injected cultures grown in the serum of immunised animals into rabbits, these rabbits did not die and presented only transient and insignificant lesions. From this fact Roger concluded that there must be an attenuation of the streptococcus by the immune serum, a view which was shared by several other observers. In formulating this view, however, he had not taken into account the possibility that this serum acted not upon the coccus that had developed in it but upon the organism of the animal into which it was injected. Bordet[356], indeed, was able to show that the streptococcus which grows in the serum of immunised animals is in no way weakened in virulence. When he took a race very virulent for the rabbit (Marmorek’s streptococcus) and injected a minimal dose of a culture grown in the serum of immunised animals, the rabbits died just as did the control animals, because the amount of serum introduced was too small to exert any influence. So also, when he filtered this culture and got rid of the serum bathing the streptococci, it was found to be just as virulent as that grown in the serum of susceptible unvaccinated animals.

In confirmation of the discovery made by Roger with the serum of vaccinated rabbits, Bordet showed that the blood serum of horses highly immunised against the streptococcus did not exhibit any bactericidal action. Moreover, he found that this serum caused the development of somewhat agglutinated streptococci and that it was capable of throwing streptococci grown on the ordinary media into clumps. Summing up his researches on the properties of this serum Bordet concludes that it “causes no profound change in the streptococcus. The vegetative character of the coccus is not appreciably diminished, and its morphology remains the same save for certain variations in the length of the chains. Even the agglutinative power, recognised in numerous serums by recent researches, is, in the antistreptococcic serum, developed but slightly” (p. 196).

More recently von Lingelsheim[357] has studied the properties of the serum of animals which he had thoroughly vaccinated against the streptococcus. He observed a certain slowing of the development of the coccus in this serum as compared with the growth in cultures made in the serum of normal, susceptible animals. But this retardation was slight and transient, and exhibited itself especially in serums to which von Lingelsheim, following Denys, had added leucocytes.

[Sidenote: [258]]

Von Lingelsheim also noted a certain degree of agglutination of the streptococcus by the serum of vaccinated animals, although this was much more feeble than in the case of the cholera vibrio or the typhoid bacillus, when agglutinated by their corresponding serums. Speaking generally, he regarded the direct action of the body fluids as insufficient to bring about the rapid destruction of the streptococci in the vaccinated organism. “Since the action of the bactericidal substances is limited in time, the streptococci are able to adapt themselves to these substances and recover their former energy. As the phenomena of extracellular solution, of such a form as those observed under the influence of the cholera antibodies, are absent in the case of the streptococcus and as, on the other hand, a considerable ingestion of these organisms by the leucocytes is observed ... we must seek in the activity of these cells a second important element of the defence of the animal organism” (p. 78).

[Sidenote: [259]]

To Salimbeni[358], who has carried out in my laboratory an investigation on this subject, we are indebted for the most reliable information on the phagocytic reaction in acquired immunity against the streptococcus. He studied specially the phenomena in the subcutaneous tissue of a horse, hypervaccinated against Marmorek’s streptococcus; this animal received in all, at several injections, about five litres of living culture. In spite of this refractory condition, an oedema at the point of inoculation was soon produced; in this the micro-organisms remained free and the leucocytes were sparse. But the cellular reaction, at first insignificant, developed with great rapidity and many leucocytes, amongst which the macrophages were much the more numerous, were attracted. The phagocytosis was still delayed for some time, but it continued to increase and 20 to 24 hours after the inoculation it was complete. As soon as the phagocytosis was well established the oedema began to disappear. In the thick exudation, containing a mass of leucocytes, the macrophages are filled with a very large number of streptococci packed together. These cocci develop inside the cells, cause them to burst and again become free. A fresh arrival of leucocytes, however, takes place, this time mainly microphages. These microphages seize the free streptococci that have struggled so victoriously against the macrophages; this second phagocytic phase is final. The streptococci still remain alive inside the microphages for some days, but ultimately are killed and digested by the phagocytes. At a period when, 5 or 6 days after injection, insignificant or isolated traces of streptococci are to be found in the microphages, the exudation when sown in nutritive media still gives abundant cultures. The incidents of this struggle between the streptococcus and the animal organism demonstrate the important part played by the phagocytes. The fact that the macrophages perish and allow the cocci to escape, clearly proves that these cocci have been ingested alive and virulent, and consequently that the fluid of the exudation was incapable of destroying or even of attenuating them. The macrophages, also, were powerless to bring about this result and the intervention of the microphages was necessary to cause the disappearance of the cocci. It is, however, always the phagocytes which ensure the final resistance of the animal.

In presence of these very precise results obtained from the work of Salimbeni, a work which I followed very closely, the previous researches by Denys and Leclef (_l.c._) made under less favourable conditions on vaccinated rabbits are deprived of their importance. These observers wished to get an idea of the difference between the reactions of the animal organism (_a_) after the injection of streptococci into the pleural cavity of immunised rabbits, and (_b_) after injection into that of normal susceptible rabbits. They killed the inoculated animals and found a very marked diminution of micro-organisms in the pleuritic exudation of the former. This diminution could not be attributed to a lysis of the streptococci by the body fluids, because there were never any signs of such destruction. Nor could the phagocytosis, very feeble at first, be considered as the cause of the disappearance of a large number of the streptococci. Denys and Leclef put forward a third hypothesis, which attributed this disappearance to the rapid resorption by the lymph stream of the injected fluid containing the organisms. Going over the record of their experiments it will be seen that in vaccinated rabbits the quantity of pleuritic exudation was always very much less than in normal rabbits. In presence of this feature there is reason to ask whether, in the case of the streptococci, a large number of these organisms were not fixed, along with the leucocytes, on the walls of the pleura, as in guinea-pigs that are inoculated intraperitoneally? Instead of being satisfied with merely examining the fluid exudation, the surface of the pleura should have been scraped in order to ascertain whether the phagocytic reaction was localised in this region.

In any case such incomplete results on the active immunity of rabbits in no way weaken the positive results obtained in the subcutaneous tissue of the horse, in which the phagocytic reaction plays a really preponderant part.

This example of the streptococci completes our series of bacteria in which we have studied their relations with the properties of the animal organism that has acquired immunity. We have still to see whether the acquired immunity against micro-organisms of animal origin is subject to the same law as that against bacteria.

[Sidenote: [260]]

For some years past a zealous study of the infectious diseases produced by animal micro-organisms has been carried out. Besides malaria, which occupies a most important position, attention has been directed to certain diseases in domestic animals that are set up by endoglobular haematozoa and by flagellata, and a fairly large number of accurate data have been collected with regard to Texas fever and its parasite the _Piroplasma bigeminum_, as well as upon the epizootic diseases due to _Trypanosomata_ (Tsetse fly disease or Nagana, “Dourine,” etc.).

We are indebted to Smith and Kilborne[359] for the earliest information concerning the acquired immunity of Bovidae against Texas fever. R. Koch[360] then added some very precise observations on the immunity of calves which had been inoculated with parasites attenuated in the body of the tick (_Boophilus bovis_). Lignières[361], who devoted much attention to this question in the Argentine Republic, has discovered a sure method of vaccinating the Bovidae against the “Tristeza,” the local name for Texas fever. He brought to Alfort specimens of attenuated haematozoa, and in Nocard’s presence performed successful vaccination experiments. Lignières is now engaged in devising a practical method of ensuring immunity under the special conditions found in the home of the “Tristeza.” Up to the present, however, there are no sufficient data as to the mechanism of the acquired immunity in this case. We have fuller information as to the essential phenomena observed in the organism of the rat vaccinated against _Trypanosoma lewisi_. We owe to Mme. L. Rabinowitsch and Dr Kempner[362] the first important data as to the possibility of immunising white or piebald rats against the disease produced by the flagellated infusorian. They noted that these animals when inoculated with the blood of grey rats containing _Trypanosomata_ acquire a very transitory disease which, however, confers an immunity against any subsequent infection. The flagellated organisms disappear from the blood within a few weeks, after which fresh injections of these parasites have no pathogenic effect.

[Sidenote: [261]]

Laveran and Mesnil[363] confirmed these observations, and in addition made careful observations on the mechanism of this acquired immunity. After making several inoculations with blood containing _Trypanosomata_ into white rats, they made a study of the properties of the blood serum of these immunised animals. First they established the fact that this serum exerts no microbicidal action on the _Trypanosomata_, but it agglutinates them without, however, rendering them motionless:—“The masses may be resolved into rosettes in which the _Trypanosomata_, united merely by their posterior extremities, have their flagella free and motile at the periphery.”

Laveran and Mesnil then studied the phenomena evolved in the refractory organism. When injected into the peritoneal cavity of immunised rats the _Trypanosomata_ are not acted upon injuriously by the body fluids. They are, however, devoured by the leucocytes. Laveran and Mesnil thus express themselves on this subject: “... we have demonstrated clearly and repeatedly that the _Trypanosomata_ are ingested alive, perfectly isolated and very motile, by phagocytes, and we have followed the details of this process of ingestion which recalls that of the ingestion of spirilla by the leucocytes of the guinea-pig. We consider, therefore, that the immunity is phagocytic in character.”

[Sidenote: [262]]

The main facts on acquired immunity established in connection with the most diverse micro-organisms, facts just described, may already be said to lead to certain general conclusions. They indicate in the first place that acquired immunity is accompanied by phenomena more complicated than those observed in natural immunity. In the two categories of processes observed in acquired immunity the phagocytic reaction is the only one that can be said to be constant. We find it in those examples in which the influence of the fluids of the body is most manifest, as in the experimental cholera peritonitis of the guinea-pig, as well as in those cases where the humoral action is most feeble, as in anthrax or in the _Trypanosoma_ disease of rats. We have, however, still to establish the relations that exist between phagocytosis and the part played by the fluids of the immunised animal, in order that we may, as far as possible, present a general picture of the inner mechanism of acquired immunity against micro-organisms. To attain this result we must place the reader in possession of further well-established facts, and we must postpone its discussion to the following chapter, which will be entirely devoted to the above-mentioned problem.