Chapter IX: The Mechanism of Acquired Immunity Against Micro-Organisms (2)

The stimulating influence of certain normal serums may be so considerable that it may prevent infection by the micro-organism, injected at the same time in a dose many times more than lethal. Wassermann[418] protected guinea-pigs by injecting into the peritoneal cavity a quantity as great as 40 times the lethal dose of typhoid bacilli, by introducing at the same time and at the same place 3 c.c. of normal rabbit’s serum, heated to 60° C. Besredka[419], who confirmed this observation, has analysed its special mechanism. He showed that the serum exercises a very marked stimulating influence on the guinea-pig’s leucocytes, which then exhibit a truly extraordinary phagocytic activity. They are seen to act in the peritoneal fluid, but they are much more active in the region of the omentum, where the leucocytes gorge themselves with micro-organisms, devouring them by dozens. The stimulating action of the heated rabbit’s serum is exercised in a similar fashion if, instead of micro-organisms, grains of carmine be injected. Very shortly after the commencement of the experiment very little carmine is found outside the cells; it is all either ingested by individual leucocytes, if the grains are small, or surrounded by numerous leucocytes when the grains are massed together; this phagocytosis is most developed in the region of the omentum, exactly as in the case of typhoid bacilli.

These facts, which so clearly demonstrate the stimulating action of the normal rabbit’s serum, prove in another way that the stimulin resists heating to 60° C., and that, in this respect, it resembles the agglutinins and fixatives. This may afford us an indication as to the nature of the stimulating substance. The possibility of obtaining an antistimulin gives us another valuable indication. Wassermann, in the work we have just cited, showed that the serum of a rabbit, previously treated with guinea-pig’s serum and injected under the same conditions as in the experiment with normal rabbit’s serum, has completely lost its protective power. The typhoid bacilli multiply freely in the peritoneal cavity and the organism of the guinea-pig is incapable of opposing a sufficient resistance. Wassermann thinks that, in this case, the disease becomes grave because of the anticytase found in the serum of rabbits treated with guinea-pig’s blood. There is no doubt that this serum is really anticytasic. But as the free cytases found in the peritoneal cavity of a guinea-pig inoculated at the moment of phagolysis, become inactive under the influence of the anticytase and play merely a minor part, it is impossible to accept the German investigator’s interpretation. Indeed, Besredka has proved that, in this case, it is the antiphagocytic or antistimulant action of the rabbit’s serum which brings about the fatal issue in the case of the typhoid inoculation.

We have laid stress on the point that an animal, whose serum is protective when introduced into another animal, may itself not be refractory against the specific micro-organism. As regards the serum of normal unvaccinated animals this has been so fully demonstrated that nowadays no one doubts it. The question is more complicated in the case of animals that have acquired immunity. As in the great majority of cases the serum of these animals is found to be endowed with a very great protective power, it has been accepted as proved that the animal which furnishes it must itself possess great immunity. The degree of protective power has even been taken as the measure of the acquired immunity. Thus, the numerous attempts to vaccinate the human subject against typhoid fever, undertaken in consequence of the researches of Pfeiffer and Kolle[420], were based on the fact that in these cases the serum of vaccinated individuals acquires a great protective power. It was argued that if this power is present it can only be due to the acquired immunity of the individuals who furnish such a serum. Undoubtedly the protective property of the fluids and the resistance are often equal; but it is none the less true that there are cases where, in spite of this property being markedly developed, the animal that furnishes the protective serum is susceptible to the action of the micro-organism and may even succumb to infection therewith.

[Sidenote: [289]]

As the hypothesis just mentioned is of importance from a general point of view it must be supported by adequate proof. It was during the course of the vaccination of rabbits against the micro-organism of the pneumo-enteritis epidemic at Gentilly that I was first able[421] to assure myself of its accuracy. I noticed that some of these rabbits, although vaccinated, ultimately succumbed to pyaemia, set up solely by this micro-organism. They were consequently not refractory against the disease, and yet their blood serum, when injected into normal rabbits along with an absolutely fatal dose of micro-organisms, was found to be highly protective. This observation drove me to the conclusion that the protective power is not a function of immunity and cannot be received as a measure of this immunity. Analogous facts have since been demonstrated in certain other cases. Thus, Pfeiffer[422] on several occasions has found that guinea-pigs, highly immunised against the cholera vibrio, have succumbed after the injection of a moderate quantity of these organisms. “On post-mortem examination of these cases living vibrios were found in the peritoneal cavity, sometimes in considerable numbers; and yet minimal doses of the heart blood given to normal guinea-pigs caused in these animals a very marked breaking down of the vibrios.” Alongside these facts may be placed others, described in the preceding chapter, of well immunised animals dying from infection, after they had been weakened by opium, cold, or other lowering agent. It is clearly seen, then, that for the manifestation of acquired immunity it is necessary that the reaction of the living cell elements should take place without let or hindrance. When this reaction fails, the possession of even great protective power is insufficient to prevent the immunised animal from contracting a fatal infection.

[Sidenote: [290]]

[Sidenote: [291]]

If, in acquired immunity against micro-organisms, it is really the cell defence which plays the most important part, we can readily imagine cases where it by itself can confer immunity without calling in the co-operation of the protective power of the fluids. When in this connection we study the resistance of an animal against various pathogenic organisms, we note, first of all, the very great variability that exists in the production of the acquired humoral properties. In certain cases, as in vaccination against vibrios or typhoid bacilli, the serum very readily becomes not only protective, but agglutinative and fixative. In other cases these properties develop with difficulty and are only manifested after a long period of vaccination. Such is the case with anthrax. After the discovery of protective serums, numerous attempts were made to obtain a serum protective against the anthrax bacillus. Several observers failed in their attempts, others were more fortunate. Sclavo[423] and Marchoux[424] were the first to succeed in obtaining a protective serum from animals hyperimmunised against anthrax. They were able to show that the serum of sheep, treated first with vaccines and then repeatedly with anthrax virus, would protect rabbits against a fatal dose of the bacillus. Marchoux even obtained, with hyperimmunised rabbits, a serum which prevented normal rabbits from contracting fatal anthrax. Sobernheim[425] was less fortunate in his first experiments. He satisfied himself that the blood serum of cattle that had recovered spontaneously from anthrax or that had been vaccinated according to Pasteur’s method, was absolutely unable to protect small animals against the anthrax bacillus, and his hypervaccinated rabbits furnished serums of doubtful activity. It was only later that he succeeded[426] in obtaining better results; especially when he used sheep. Even then he found that in the production of the anti-infective property the individuality of the immunised animals had a dominant influence. Thus, in two sheep, treated in exactly the same way, the serum of one was found to be incapable of protecting a rabbit, whilst that of the other exhibited an undoubted, although feeble, protective power.

But what is of greater interest to us, from our point of view, is that guinea-pigs which have been vaccinated against anthrax and which enjoy a considerable immunity against this disease, exhibit no protective power. In a letter from Behring I learnt that this fact had for the first time been demonstrated by Wernicke in experiments carried out in the Hygienic Institute at Marburg. After repeated and painstaking attempts this observer succeeded in vaccinating guinea-pigs against enormous doses of virulent anthrax bacilli. The serum from the animals so immunised was, however, quite incapable of protecting normal guinea-pigs against a fatal infection. This result was the more extraordinary since Wernicke’s pigeons, likewise vaccinated against anthrax, gave a serum whose protective power was quite distinct. Realising the great importance of these facts I asked M. de Nittis[427] to repeat these experiments in my laboratory. The vaccination of pigeons is an easy matter, but that of guinea-pigs presents great difficulties. He succeeded, nevertheless, in vaccinating some of these rodents very highly, and this enabled him to compare the protective power of the blood serum in the two species. That of the vaccinated pigeon was found to be endowed with this power and protected guinea-pigs and mice against virulent anthrax. The serum of the immunised guinea-pigs, on the contrary, exhibited no protective property, just as in Wernicke’s experiments. The guinea-pigs and mice, into which this serum was injected at the same time as the anthrax bacilli, died even when attenuated anthrax was used. We have, then, in this case, an example of acquired immunity, independent of any protective power of the fluids of the body.

In the course of their researches on the bacillus isolated by R. Pfeiffer from persons attacked by influenza, Delius and Kolle[428] tried to vaccinate susceptible animals (guinea-pigs) against this minute organism and to immunise animals naturally refractory (dog, sheep, goat) against fairly large doses of cultures. They succeeded in vaccinating guinea-pigs against ten times the lethal dose, but never obtained any protective serum. Nor did the other animals that were treated furnish a protective serum. “From the whole of our experiments carried on for several years”—conclude Delius and Kolle—“it is quite evident that we were unable to produce any appreciable change in the blood by the use of those methods which have produced specific immunising serums against other bacteria such as the bacilli of diphtheria, cholera, typhoid fever, and ‘blue pus’” (p. 345). Slatineano undertook a detailed study of Pfeiffer’s bacillus in my laboratory, but he found it impossible to demonstrate any unquestionable protective effect exerted by the blood serum of vaccinated guinea-pigs upon normal guinea-pigs inoculated with a fatal dose of this organism. We are not justified, therefore, in classing this bacillus with the anthrax bacillus; we may, however, cite it as an argument illustrating the difficulty that is met with, in certain examples of acquired immunity, of discovering the protective power, when feeble and masked.

The inoculation with micro-organisms of animal nature causes the development of acquired immunity, but in this case the properties of the fluids of the body are but little in evidence or they may be even _nil_. Let us return to the example of the _Trypanosoma_ of the rat which excites in vaccinated animals a protective and weakly agglutinative power of the serum. This fluid, however, is usually found to be incapable even of rendering the flagellated parasites motionless.

[Sidenote: [292]]

The question of immunity against malaria has been much discussed. It is well known that a first attack of this disease, so far from conferring any immunity of the slightest durability, leaves a certain predisposition to another attack. In spite of this the study of malaria in various countries and in individuals belonging to different races has demonstrated that there does indeed exist a certain degree of acquired immunity against this disease. During recent years Koch[429] has paid special attention to this subject and has furnished us with very valuable data, based especially on a comparative study of the blood of children and adults. The frequency of Laveran’s parasite in the former and its rarity in the latter, have led him to the conclusion that infantile malaria sets up an immunity which persists in the adult. Moreover, it has been established that in malarial countries the indigenous inhabitants exhibit an attenuated form of the disease, unaccompanied by acute attacks, but with phenomena that are chronic and very slow in development.

In spite of the existence of a certain degree of acquired immunity against malaria, all attempts to demonstrate any protective action of the serum have been fruitless. Celli[430], indeed, injected, as a preventive, the blood serum of individuals who had recovered from malaria or of others who were bled during the period of defervescence following an acute crisis of this disease, but in every instance these injections were found to be useless in preventing an attack of malaria.

We can readily understand that in a disease which is exclusively human, such as malaria, it has not been possible to perform a sufficient number of experiments to decide the question of the protective property of the blood. In this respect we shall have greater chance of obtaining satisfactory data if we direct our attention to some analogous disease attacking one of the lower animals. Such a disease we have in Texas fever, occurring in the Bovidae, as the result of the action of an animal parasite, _Piroplasma bigeminum_, which invades the red blood corpuscles much as Laveran’s parasite invades those of the human subject.

[Sidenote: [293]]

[Sidenote: [294]]

As mentioned in the preceding chapter, Smith and Kilborne and Koch have demonstrated that the Bovidae may acquire a real immunity against Texas fever. Nicolle and Adil Bey[431] at Constantinople found indigenous races that exhibited a remarkable immunity against the _Piroplasma_. Having demonstrated this fact the idea occurred to them to inoculate these refractory cattle with very large quantities of virulent blood and to make use of the serum from animals so treated for the prevention of infection in susceptible races of Bovidae. This experiment gave negative results. Lignières[432] elaborated a special method of vaccinating susceptible Bovidae and was successful in obtaining very encouraging results. A commission of veterinary surgeons from Alfort[433] appointed to verify these observations came to the conclusion that “the vaccination as carried out by Lignières was absolutely effective.”

Lignières also carried out researches on the protective power of the blood serum of his immunised cattle. In a communication to the International Congress of Medicine, held in Paris in 1900, he stated that the injection of several hundred cubic centimetres of this fluid did not protect normal animals against infection. We must conclude, therefore, that, here also, we have another example of acquired immunity unaccompanied by the presence of any protective property of the blood fluid.

These results have received confirmation from a most authoritative source. Nocard has kindly communicated to me the fact that he has tried in vain to confer immunity on normal dogs into which he has injected blood serum coming from dogs that had recovered from the disease produced by a haematozoon closely allied to that of Texas fever or serum from sheep immunised with blood from the affected dogs.

Looking at the data we have just summarised as a whole, we are compelled to recognise that, on the one hand, the protective power of the body fluids may coincide with a susceptibility to the corresponding micro-organism, and that, on the other, real acquired immunity may exist without any manifestation of this humoral property, especially as, even in immunised animals, the acquired immunity often persists longer than does this property. It must be accepted then, that, in this immunity, there exists something other than the powers of the fluids of the body, that is to say, the factor which plays the predominant part is to be sought for in the cellular elements. We need only recall the many facts collected in the preceding chapter to be convinced that in acquired immunity phagocytosis is the most constant and most general phenomenon. We find it in cases where the humoral properties are the most marked, as well as in those in which they are only slightly developed or are entirely absent. We need not again discuss Pfeiffer’s phenomenon analysed in the preceding chapter. It is sufficient to mention that this example of the extracellular destruction of micro-organisms only occurs under limited and special conditions. It is observed only in cases where the injection is made into a situation rich in leucocytes which undergo phagolysis as a result of the sudden change brought about in their conditions of existence. Further, this phenomenon is observed only in connection with micro-organisms that are slightly resistant to the influence of the microcytases. In those cases in which we meet with Pfeiffer’s phenomenon, we also meet with a widely extended phagocytic reaction.

[Sidenote: [295]]

This reaction is most pronounced where the properties of the body fluids are only slightly developed or are absent. The study of acquired immunity against anthrax provides us with a very convincing proof of this. We have already cited the example of vaccinated rabbits and rats in which phagolysis is incomparably greater than in the susceptible control animals which contract a fatal anthrax. This rule is general. It is confirmed in the vaccinated sheep and guinea-pig. The absence, or feeble development, of the protective power of the blood or of the other humoral properties in no way, then, prevents the considerable change which is set up in the phagocytes of animals that have acquired immunity against anthrax. The negative chemiotaxis of the leucocytes, so marked in susceptible animals, is modified into positive chemiotaxis as the result of vaccination. This fact, one of fundamental importance, was first demonstrated for the immunity against anthrax, later being extended to other micro-organisms. Massart[434] studied the general subject and collected a series of data which led him to say that “vaccination effects an education of the leucocytes; these latter become so adapted that they can approach the virulent micro-organisms.” The best method of forming an estimate of the change which the leucocytes undergo is by injecting subcutaneously very virulent micro-organisms capable of setting up a generalised infection. The anthrax bacillus, Gamaleia’s vibrio, the streptococci and the cocco-bacilli of swine and fowl cholera are very suitable for such study. These micro-organisms, when inoculated subcutaneously into susceptible animals, set up a very slight local reaction or none at all, in the form of an exudation of transparent fluid almost entirely without leucocytes. The micro-organisms grow freely in these exudations and soon invade the animal. In vaccinated animals the local reaction is more marked and the exudation, very rich in leucocytes, is poor in fluid; the micro-organisms remain free for a very short time, being soon ingested by the leucocytes. Their destruction, inside these cells, takes a longer or shorter time according to circumstances; but in the end it is always complete.

[Sidenote: [296]]

The difference as regards phagocytic reaction between susceptible and vaccinated animals, such as I have just described, has been generally recognised by many observers. A few opponents are still found, however, who consider that they are justified in affirming that the negative chemiotaxis of the susceptible animal does not exist and that, consequently, vaccination can in no way change it into positive chemiotaxis. Werigo made himself the spokesman of this view, which he has maintained in several papers[435]. Instead, however, of introducing the virulent micro-organisms into the subcutaneous tissue of susceptible animals he injected them directly into the veins. Using cultures of the anthrax bacillus and of the cocco-bacillus of fowl cholera he injects these into the venous system of normal rabbits. The animals soon die from general infection. If, however, these animals are killed shortly after inoculation, it is found on examination of sections that many of the micro-organisms have been ingested by the leucocytes. Werigo concludes from these facts that in the higher animals the chemiotaxis is always positive; but that it ends in the destruction of the micro-organisms in the vaccinated animals, never bringing about this result in susceptible animals. Taking all the data on this question into consideration, it is easy to convince oneself that this view cannot be accepted as correct, for not only the definite phenomena observed below the skin but also the no less demonstrative process appearing in the peritoneal cavity prove most clearly the existence of this negative chemiotaxis of the leucocytes. I need only recall Bordet’s experiment on the fate of streptococci and _Proteus vulgaris_ when injected together into the peritoneal cavity of guinea-pigs. Whilst the _Proteus_ bacilli at the end of a very short time are all ingested by the leucocytes, the streptococci remain free in the peritoneal fluid up to the death of the animal. The leucocytes which exhibit a positive chemiotaxis as regards the former, manifest a negative chemiotaxis as regards the streptococci.

[Sidenote: [297]]

In spite of the great force of these arguments, the discovery of a means of reconciling the results obtained from the inoculation of micro-organisms subcutaneously or into the peritoneal cavity, with those observed after they had been injected into the blood vessels would be of great interest, and Zilberberg and Zeliony[436] have undertaken a series of experiments with this object. Following Werigo they made use of the cocco-bacilli of fowl cholera, and found, in accordance with his observations, that the intravenous injection of these organisms, obtained from cultures in nutrient media, causes a very marked phagocytosis of the cocco-bacilli. When, however, they injected into the veins of rabbits cocco-bacilli that had been grown in the peritoneal fluid of other rabbits, they found the micro-organisms free in the blood plasma and observed only a very restricted phagocytosis in the microphages of the liver. It follows from these experiments that the ingestion of the cocco-bacilli, in Werigo’s experiments, was dependent on the presence of a large number of attenuated micro-organisms which were present in the cultures that he employed for his injections. Alongside these organisms, slightly or not virulent, were others, endowed with their normal pathogenic activity and quite numerous enough to set up a fatal infection. When Zilberberg and Zeliony replaced cultures on agar by the peritoneal exudation which contained virulent cocco-bacilli almost exclusively, the phagocytosis in rabbits, injected into the veins, was found to be almost suppressed. With the object of establishing whether the absence of the phagocytic reaction, in this case, really depended on negative chemiotaxis on the part of the leucocytes, the above cited observers performed the following experiment. They injected into the vein of a rabbit, already affected with a generalised infection by the cocco-bacillus of fowl cholera, an innocuous culture of a saprophytic staphylococcus. Post-mortem examination showed that these cocci were almost entirely ingested by the same phagocytes which refused so energetically to seize the cocco-bacilli. This experiment, analogous to that of Bordet on streptococcus and _Proteus_, compels us to reject Werigo’s conclusions as to the absence of negative chemiotaxis in the phagocytes of the higher animals. I ought to add that the work of Zilberberg and Zeliony was in part executed in my laboratory so that I was able to convince myself by ocular demonstration of the complete accuracy of their statements.

Independently of these observers and even before their work appeared, Th. Tchistovitch[437] published an interesting study on the same question. He injected very virulent streptococci into the ear vein of rabbits. These micro-organisms set up a generalised and fatal infection in which phagocytosis was completely absent or nearly so. Here again was manifested a negative chemiotaxis of the phagocytes, which, henceforth, could no longer be questioned.

[Sidenote: [298]]

In certain infective diseases terminating fatally a very marked phagocytosis is observed even in susceptible animals. The most typical example of this is furnished by swine erysipelas and mouse septicaemia. We know from the researches of Koch[438], followed by those of Loeffler[439], Schütz[440] and others, that in animals which have died from these two diseases the leucocytes are gorged with small specific bacilli. A method of vaccinating animals against the micro-organism of swine erysipelas was worked out by Pasteur and Thuillier[441] and was afterwards studied by many observers. Thanks to this method it has been possible to demonstrate the phenomena which may be observed in vaccinated animals (especially rabbits). Here also a phagocytosis takes place, even more rapid and more complete than in susceptible animals. What is more important, the intracellular digestion of the ingested bacilli is followed by the total destruction of the micro-organisms in the vaccinated animals, though in the normal animals this digestion is very imperfect.

[Sidenote: [299]]

The acquisition of immunity against micro-organisms is, therefore, due not only to the change from negative to positive chemiotaxis, but also to the perfecting of the phagocytic and digestive powers of the leucocytes—a general superactivity and adaptation of the phagocytic reaction of the immunised animal is produced. This conclusion, based upon a large number of well-established facts and in complete harmony with the whole of the data at our disposal concerning acquired immunity, has been attacked by Denys and Leclef[442] in their work on the streptococcus. They base their opposition upon experiments made _in vitro_ on the action of serums and leucocytes on this micro-organism. They have compared the bactericidal power of mixtures of the serums of normal and of vaccinated rabbits with leucocytes isolated from exudations from these two groups of animals. The leucocytes, whether derived from normal or from vaccinated rabbits, when mixed with normal serum were equally incapable of ingesting and destroying the streptococci. When mixed with blood serum from vaccinated rabbits, however, the two kinds of leucocytes exhibited a very marked phagocytic reaction. Denys and Leclef conclude from this that phagocytosis, although an important factor in immunity, plays merely a secondary part and is dependent on the humoral properties. The experiments and views of these two observers have been generally received by the partisans of the bactericidal theory of the body fluids as an actual proof of this theory. We cannot agree. Researches extending over a long period have shown us that the study of phagocytosis _in vitro_ can give only a very inexact and imperfect idea of the course of the phenomena in the living animal. Usually the leucocytes taken from the exudations, although amoeboid, no longer fulfil their phagocytic functions at a time when in the animal they would ingest micro-organisms with the greatest rapidity. As a general rule, existence outside the living body weakens them very considerably. But in some cases, rare it is true, the leucocytes although inactive in the animal exhibit intense phagocytosis when introduced into a hanging drop of fluid from an exudation or even of urine. In any case it is very hazardous to infer from phenomena which appear under these artificial conditions what takes place in the living animal. The value of the experiments of Denys and Leclef is still further marred by the fact that they mixed the leucocytes with blood serum. They appear to have lost sight of the fact that this fluid is far from corresponding to that which bathes the leucocytes in the living animal. The serums contain leucotoxin in greater or less quantity and it is not to be wondered at that the leucocytes when mixed with normal rabbit’s serum should perish very rapidly. Further, the serum of vaccinated rabbits is agglutinative (this fact, however, was not sufficiently elucidated in 1894 when the researches of Denys and Leclef were made) and the clumping of streptococci might simulate their destruction. In a word, the experiments of these observers have been carried out under such conditions that it is impossible to base upon them a refutation of data obtained in the living animal. Moreover, in the description of the phenomena which appear in the subcutaneous tissue of rabbits inoculated with the streptococcus, Denys and Leclef provide us with arguments against their own view.

[Sidenote: [300]]

These observers introduce the same quantity of streptococci below the skin of the ear of normal and of vaccinated rabbits. In the first there is soon produced a very marked oedema of the ear, in which may be seen a number of streptococci and of leucocytes that have not ingested any micro-organisms. In the second the oedema does not develop, but at the seat of invasion a number of leucocytes come up and these soon ingest the streptococci. As we see, the phenomena manifest themselves here just as they do with the anthrax bacillus and many other micro-organisms when under analogous conditions. Denys and Leclef, indeed, recognise that, below the skin of the ear of vaccinated rabbits, the small quantity of exudation fluid is not sufficient to enable us to accept it as capable of exerting any considerable influence as regards humoral properties. Nevertheless, they think that the “serum” of this fluid may exercise a certain action, but they furnish no proof of this, and seem to ignore the fact that the plasma of the subcutaneous exudation is far from being identical with blood serum obtained outside the animal. At present it is well known that this latter fluid contains cytases which are absent from the plasmas. Now, the feeble bactericidal action, if this really exists as regards the streptococcus, must be attributed to the microcytase which has escaped from the leucocytes at the time of the preparation of the serum.

To sum up, the example studied by Denys and Leclef clearly comes under the general law of phagocytic reaction in acquired immunity against micro-organisms. It is impossible to deny that the superactivity of the phagocytes which is always found in this immunity, although readily observed, cannot be demonstrated in a rigorous fashion outside the fluids which bathe the cells. There are, however, very important analogies which may be invoked in favour of this thesis. We have already cited in our fifth chapter Delezenne’s experiments on the digestion of gelatine by the leucocytes of the dog, which show in the most demonstrative fashion that these cells accustom themselves to bring about this digestion more and more quickly and this quite independently of any humoral influence.

For some time past there has been no doubt as to the fundamental fact that the phagocytes in immunised animals seize and destroy living micro-organisms. Several attempts have been made to show that such destruction of these bacteria takes place solely by the body fluids, and that the phagocytes intervene only as “scavengers” to carry off the dead bodies of the micro-organisms. The numerous observations, described in the preceding chapter, absolve us from again entering into a discussion of this question. Moreover, the majority of these opponents now recognise that micro-organisms are ingested in a living state by the phagocytes of immunised animals. Some, however, have expressed the opinion that these living micro-organisms, before becoming the prey of the phagocytes, must undergo some preliminary attenuation of virulence through the action of the body fluids. Hence the theory of the attenuating power of the fluids of the body, maintained especially by Bouchard and his pupils. During the course of our exposition of the facts concerning acquired immunity, we have several times had occasion to speak of the virulence of micro-organisms in the immunised animal. Here, therefore, we may confine ourselves to a brief summary of the observations collected on this point.

[Sidenote: [301]]

Having observed that the anthrax bacillus, when developed in the blood of immunised sheep, was incapable of giving fatal anthrax to rabbits, I expressed[443] the opinion that under these conditions its virulence had become attenuated. Later, analogous changes were shown by Charrin[444] in the _Bacillus pyocyaneus_ when cultivated in the serum of immunised animals. Bouchard[445], generalising on these data, arrived at the following theory of vaccination. “The inoculation of a strong virus into a vaccinated animal is equivalent to the inoculation of an attenuated virus. The attenuation, however, instead of being done beforehand in the laboratory, is brought about in the tissues of the vaccinated animal” (p. 18). Charrin and Roger[446] upheld this view, and the latter offered several new arguments in support of it. He observed that animals inoculated with pneumococci and streptococci grown in the blood serum of vaccinated animals, contracted a transient and benign disease merely, whilst the control animals, inoculated with the same micro-organisms, cultivated in normal serum, always died from generalised infection.

[Sidenote: [302]]

The discovery of the protective property of serums has thrown a new light upon these experiments. We must now ask ourselves: Does the innocuousness of micro-organisms depend not on the attenuation of the virus, but rather on the protective action of the serum itself? When, in the course of my researches on the Gentilly cocco-bacillus, I found that this organism, cultivated in the serum of vaccinated rabbits, became much less pathogenic than when it was grown in the serum of normal rabbits, I set myself to answer this question. Simple filtration through paper was sufficient to rid the organism of the serum in which it had grown. The inoculation of cocco-bacilli thus treated proved at once that their virulence was in no degree modified, and that it was the intervention of the serum that prevented the micro-organism from setting up the rapidly fatal disease. Issaeff[447], who, in my laboratory, carried out the investigation, was able to extend this to the pneumococcus. He obtained agglutinated cultures in the serum of vaccinated rabbits, and he compared their activity by injecting them (1) with, and (2) without their culture medium. The difference was very marked. In the first case the infection produced was much slower in its course than in the second. The virulence of the washed pneumococci was found to be the same whether they came from a culture in normal serum or from one in immunised serum. Sanarelli[448] obtained the same result with Gamaleia’s vibrio. The vibrios when grown in the serum of vaccinated guinea-pigs proved to be very virulent so soon as they were freed from the fluid in which they were grown. Later, similar demonstrations were given by Bordet[449] and Mesnil[450] with respect to streptococci and to the bacilli of swine erysipelas. We must, then, conclude that we have here to do with a general law. Some experiments made by de Nittis[451] might seem to indicate an exception to such a law. He observed that anthrax bacilli when grown in the serum of vaccinated pigeons lost a part of their virulence. It must not be forgotten, however, that he grew his cultures under special conditions; the bacillus was grown for several days at 42° C., this in itself being quite sufficient to bring about a certain attenuation of virulence.

The theory of the attenuating action of the body fluids, based on the attenuation of the virus in the serum of vaccinated animals, can no longer be maintained, as it is a well-established fact that the serum, obtained outside the body, is a fluid differing in character and properties from the plasma of the living animal. We have seen up to what point this demonstration has shaken the theory of the bactericidal action of the body fluids.

[Sidenote: [303]]

It cannot be doubted that a micro-organism may undergo a certain weakening in virulence, as well as in certain other functions, in the body of the animal that has acquired immunity. But the question must be put: Is this effect obtained as the result of humoral or of cellular action? As a general rule, exudations obtained from vaccinated animals, and containing living micro-organisms, are found to be virulent when inoculated directly into susceptible animals. This fact was established by Pasteur[452] when he first carried out his researches on acquired immunity against fowl cholera. He showed that the exudations of vaccinated fowls set up a fatal disease in normal fowls, without there being the least evidence of any attenuation of the micro-organism. The same applies to the Gentilly cocco-bacillus and to the anthrax bacillus in a very great majority of examples. De Nittis observed that the exudations of immunised pigeons produced a fatal infection in the guinea-pig and in the mouse. In the immunised guinea-pig, on the other hand, he found that the exudations soon became innocuous for these animals. This alteration, however, must be attributed not to the body fluids (which exhibit no protective or attenuating power) but to the action of the cells.

With the object of gaining some idea of the changes that the micro-organisms undergo in the immunised animal, Vallée[453] carried out a series of experiments on rabbits vaccinated against the bacillus of swine erysipelas. He enclosed these bacilli in sacs of collodion which he introduced into the peritoneal cavity of susceptible rabbits and of others that were hyperimmunised. The bacillus developed well in both cases. It gave homogeneous non-agglutinated cultures in the sacs placed in normal animals, whilst in the sacs introduced into the peritoneal cavity of hyperimmunised rabbits the bacilli grew into agglutinated filaments. This proves that the wall of the sacs permitted of the passage of the active substances elaborated in the immunised animal. Different from the point of view of agglutination, the cultures likewise exhibited a considerable difference in their pathogenic activity. The cultures developed in the sacs in hyperimmunised rabbits were found to be much more virulent than those grown in the sacs in control animals. This augmentation of virulence depends, probably, on the influence of the active substances which pass through the walls of the sacs. In any case, this experiment affords further confirmation of the impossibility of maintaining the theory of the attenuation of micro-organisms by the fluids of an animal enjoying acquired immunity.

[Sidenote: [304]]

Since the discovery of the antitoxic property of the fluids of the body, it has been accepted that its manifestation was indispensable for the acquisition of immunity. It was thought that in order to get rid of pathogenic micro-organisms the animal had first to develop the means of neutralising their toxins. These substances once prevented from exerting their toxic action, the micro-organisms were left without their weapon of attack and found themselves reduced to the condition of simple saprophytes. It was accepted, therefore, that an effective antitoxic power was always to be found in the fluids of animals that had acquired immunity. Against this explanation, however, are certain established facts. Chauveau[454] had observed that Algerian sheep, whose natural immunity was further strengthened by considerable doses of anthrax bacilli, exhibited a susceptibility to injections of anthrax blood quite as marked as that of normal sheep. The immunity against the virus, then, did not progress _pari passu_ with that against the poison. Later, Charrin and Gamaleia[455] furnished important data on this subject. They showed that animals vaccinated against the _Bacillus pyocyaneus_ and the vibrios of Koch and Gamaleia were even more susceptible to intoxication by the soluble products of these micro-organisms than were normal animals which had acquired no immunity against the corresponding bacteria. Shortly afterwards this observation was confirmed by Selander[456], in his work on hog cholera, carried out under Roux’s direction. Rabbits vaccinated against the cocco-bacillus of this disease resisted infection by the virus, but died as a result of the exhibition of the same doses of toxin that killed normal rabbits. I[457] was able not only to verify this, but to add to it the further fact that the blood serum of vaccinated rabbits, although markedly protective against infection, exercised not the slightest antitoxic action.

When, later, R. Pfeiffer set himself to study the immunity of animals against the cholera vibrio, he, along with his collaborators, was able to furnish numerous data confirming the hypothesis that animals thoroughly vaccinated against this vibrio had not thereby become more resistant to its toxin and that their anti-infective serum exhibited no antitoxic power. These results have been confirmed repeatedly and must be regarded as fully established.

[Sidenote: [305]]

Von Behring here recognised a general law which, with the aid of his collaborators, he attempted to develop. We owe to him the knowledge that the susceptibility, augmented as regards the toxins, of animals vaccinated against micro-organisms, might even serve in doubtful cases to reveal the presence of their bacterial poisons. Culture products when deprived of micro-organisms often set up no poisoning in normal animals susceptible to infection. From this fact it is generally concluded that the toxin is not present in the products in question. But animals of the same species when immunised against infection by the micro-organism, owing to their “hypersusceptibility,” react much more delicately and allow of the demonstration of the presence of bacterial poisons in fluids inactive for unvaccinated animals.

In collaboration with Kitashima[458], von Behring immunised guinea-pigs against the diphtheria bacillus, and demonstrated that two or three injections of diphtheria toxin were quite sufficient to render these animals refractory to infection by the diphtheria bacillus though they became more susceptible to intoxication. Von Behring considers that this augmentation of susceptibility to the diphtheria poison may be a means of rendering the local reaction of the living elements at the point of introduction of the bacilli more active.

In any case, it is beyond question that acquired immunity against microbial infection is quite independent of the resistance against the toxins of the corresponding micro-organism. An antitoxic manifestation of any kind, therefore, cannot be regarded as necessary for the development of immunity against the micro-organism.

Of all the humoral properties developed in acquired immunity against micro-organisms, the fixative power and the protective power are the most constant. It might naturally be suggested, as a result of this observation, that these two powers are indispensable for the manifestation of phagocytosis for the purpose of destroying and of ridding the animal of the pathogenic organisms. It is quite possible to understand how, under these conditions, the idea has been put forward that anti-infective acquired immunity is the result of two different factors: in the first place, a humoral property independent of the phagocytes and, in the second place, the phagocytes themselves. But the part played by these cells cannot be accepted as purely secondary—a view which has been advanced and defended again and again. This question is of such importance that it is reasonable to ask whence come the humoral properties, such as the fixative power and the protective power, factors of such far-reaching influence in anti-infective immunity?

[Sidenote: [306]]

Thanks to the work of several investigators this question may now be answered. Pfeiffer and Marx[459] first supplied important information concerning the origin of the protective property. Into rabbits they made subcutaneous inoculations of cholera vibrios, killed by heat (70° C.), and then examined, most minutely, the protective power of the blood and of extracts from various organs. Examining, separately, the protective power of the serum and that of the layer of leucocytes deposited in tubes, Pfeiffer and Marx were unable to find any marked difference. Nor did they ever obtain any definite effect with leucocytes collected from pleuritic exudations. From these observations they concluded that the leucocytes of the blood could not be regarded as the source of the protective substance (or “cholera antibody”). At a period when the serum as yet exhibited an insignificant protective power or none at all, the extract from the spleen often exerted an action of the most marked character. In an experiment in which the rabbit was killed 48 hours after the injection of the vibrios, 0·3 c.c. of the serum was incapable of preventing fatal infection of a guinea-pig, whereas 0·03 c.c. of an extract of the spleen exerted a marked protective effect. From this and similar experiments, Pfeiffer and Marx conclude that the spleen is the principal source of the protective substance. In order to verify this observation they injected killed cholera cultures into rabbits which had previously been deprived of their spleens, but the asplenic rabbits still produced the same amount of protective substance, and these two observers were led to conclude that the lymphatic glands and the bone-marrow might also serve as the sites of origin of this substance.

It is only during the first few days, however, that these organs exhibit a protective power greater than that of the blood. Three or four days after the injection of the vibrios the blood serum becomes richer in protective substance; the organs contain much less of it. This condition is maintained for some time, after which the blood in turn begins to get impoverished.

[Sidenote: [307]]

Pfeiffer and Marx put to themselves the question: Is the marked protective power of the spleen due to the production of preventive substance by this organ, or is it to be explained by an accumulation in the spleen of this substance manufactured elsewhere? With the object of obtaining an answer to this question they injected protective serum from other individuals into rabbits, when they found that the protective substance showed not the slightest tendency to accumulate in the spleen. These authors were compelled to conclude, therefore, that the spleen and other haematopoietic organs (lymphatic glands and bone-marrow) are the real seats of the production of the protective substance. We may add that these organs are also the phagocytic organs _par excellence_, that is to say, the centres which serve not only for the development of phagocytes but which contain a large number of the adult elements capable of exercising the phagocytic function.

Almost simultaneously with Pfeiffer and Marx, A. Wassermann[460], in collaboration with Takaki, undertook similar researches on the origin of the substance protective against the typhoid cocco-bacillus. The outcome of this work was that “it was the bone-marrow, the spleen, and the lymphatic system, including the thymus gland, which exhibited immunising power against the bacillus of typhoid fever, whilst the other organs, the blood, brain, spinal cord, muscles, liver, kidney, etc., did not at this stage show any marked specific property.”

[Sidenote: [308]]

As these observations on the production of protective substance in the phagocytic organs was one of essential importance in connection with the problem of acquired immunity, I asked M. Deutsch[461], working in my laboratory, to carry out a series of experiments on this subject. Using guinea-pigs, he injected into the peritoneal cavity cultures of the typhoid bacillus killed by heat (66° C.). A few days later the serum had become distinctly protective. At this stage, and even before the appearance of this property in the blood, Deutsch killed some of his animals and carefully measured the protective power of the extract of the various organs. He began by confirming the result obtained by Pfeiffer and Marx as to the non-production of the protective substance in the peritoneal exudation. Usually this fluid was insufficient to protect normal guinea-pigs against typhoid infection. In a few experiments only was the exudation found to be as protective as the blood serum; in most of the others, the blood serum was much more active than the fluid of the exudation. The spleen was the organ which exhibited the greatest protective power, and in nearly one-half of the cases it was more active than was the blood. The bone-marrow sometimes gave analogous though much less marked results. The spleen consequently must be looked upon as the principal seat of the production of the protective substance.

Having confirmed this observation of Pfeiffer and Marx and of Wassermann and Takaki, Deutsch tried to obtain the protective property in guinea-pigs deprived of their spleens. The experiment was quite successful, and here again his result agreed with that obtained by Pfeiffer and Marx. Guinea-pigs from which the spleen had been removed developed the protective property just as well as did the control animals; in the former the bone-marrow was found to be specially active.

When Deutsch, instead of removing the spleen from his guinea-pigs before the injection of the micro-organisms, did so some (3–5) days afterwards, there often occurred a marked diminution in the amount of the protective substance produced. We must conclude, therefore, that soon after inoculation there appears in the spleen a phenomenon which is associated with the development of the protective power. The most simple explanation of these facts is that the micro-organisms injected into the peritoneal cavity and soon afterwards seized by the phagocytes (for the most part by the microphages), are carried to the phagocytic organs, particularly the spleen, lymphatic glands, and bone-marrow. In those animals whose spleens are left intact a large number of these microphages loaded with micro-organisms make their way into this organ, a fact confirmed by direct observation. When the spleen is removed the microphages must necessarily betake themselves to other phagocytic organs. As the micro-organisms undergo intracellular digestion in the phagocytes, it is very difficult, if not impossible, to follow them for any length of time after they have been ingested, but the analogy with the phenomena of the resorption of red blood corpuscles, described in