CHAPTER XII

ARTIFICIAL IMMUNITY AGAINST TOXINS

Adaptation to poisons.—Artificial immunity against bacterial and vegetable toxins and against snake venom.—Principal methods of immunisation.—Immunisation by toxins and toxoids.—Inoculation against diphtheria toxin.—Phenomena produced in the course of vaccination against toxins.—Rise of temperature.—Leucocytosis.—Development of antitoxic power.—Properties of antitoxins.—Mode of action of antitoxins.—Action of antitoxins _in vitro_.—Their action in the organism.—Influence of living elements on the combination of antitoxin with toxin.—Antitoxic action of non-specific serums, of normal serums and of broth.—Immunity against toxins is not in direct ratio to the amount of antitoxins in the body fluids.—Hypersensitiveness of an animal treated with toxin.—Diminution of the susceptibility of the organism immunised against toxins.

Hypotheses as to the nature and origin of antitoxins.—Hypothesis of the transformation of toxins into antitoxins.—Hypothesis of receptors detached from cells as the source of antitoxins.—Hypothesis of the nervous origin of tetanus antitoxin.—Fixation of tetanus toxin by the substance of the nerve centres.—The relations between saponin and cholesterin.—Anti-arsenic serum.—Part played by phagocytes in the struggle of the animal against poisons.—Probable part played by phagocytes in the production of antitoxins.

[Sidenote: [359]]

Although scientific men succeeded only a little more than ten years ago in vaccinating against poisons by artificial methods, savage races and ancient peoples at a very remote period undoubtedly possessed methods of counteracting the effects of certain venomous substances. The frequent observation of cases in which doses of poisons, insufficient to cause death, brought about a more or less durable resistant condition, must result in the elaboration of artificial means of preventing the intoxications.

Von Behring[521] points out that analogous facts must have been known to the physicians of ancient times; and it is in such knowledge that we must look for the source of the dogma put forward by Hippocrates, that the factor which produces a disease is also capable of curing it.

[Sidenote: [360]]

To Pliny we are indebted for the now well-known story, that Mithridates of Pontus possessed the means of protecting himself against various poisons by a process of adaptation, and, amongst others, by the use of the blood of Pontine ducks to which he had given poisons by the mouth.

The adaptation of horses and of the highlanders of Styria to arsenic, as well as that of the many morphinomaniacs to morphia, is known to everybody. A man, habituated to morphia, is able to consume daily a dose several times the fatal one; indeed, cases have been known of people acquiring the power of consuming two, and even three, grammes of morphia per diem. Man may acquire an adaptation to toxic substances of the most diverse character, such as arsenic, alcohol, morphia, nicotine, etc.

[Sidenote: [361]]

Even when we had obtained much information concerning acquired immunity against micro-organisms we still knew nothing of the mechanism of such adaptation, or as to the possibility of acquiring a special immunity against bacterial poisons. Charrin and Gamaleia’s discovery that animals vaccinated against a micro-organism are just as susceptible to its toxic products as normal animals, led Bouchard[522], in whose laboratory it was made, to say that the idea of the adaptation of cells to bacterial poisons must be dropped. He developed this thesis at the International Congress at Berlin in 1890, and formulated it as follows: “When we inject a healthy animal and a vaccinated one with the soluble products of the micro-organism which has been used for the vaccination, the dose required to kill each animal is exactly the same. Let us not speak, then, of the training of the leucocytes, and of the adaptation of the nerve cells to bacterial poisons: it is pure rhetoric.” At this time we had only just commenced to acquire exact knowledge concerning the toxins of micro-organisms. For a considerable period they were sought for amongst the ptomains, very stable substances allied to the alkaloids; here, however, we were working in a wrong direction. It was not until the classic researches of Roux and Yersin[523] on diphtheria toxin, published in 1888 and 1889, that the true nature of bacterial poisons was revealed. It was found that we were not dealing with ptomains, but with soluble ferments, substances of indeterminate chemical composition, allied to the albuminoids, and, like them, unstable. The methods adopted by Roux and Yersin in their study of diphtheria toxin enabled other investigators to discover the analogous toxins of several other bacteria. Knud Faber[524] and Brieger and Fränkel[525] soon succeeded in separating the toxin from the tetanus bacillus, a toxin capable of producing in animals tetanic contractions as typical as those obtained with cultures of the tetanus bacillus.

These investigations inaugurated a new era in microbiology and enabled us to attack the problem of acquired immunity against bacterial toxins scientifically. Within a few months of the declaration made by Bouchard at the Berlin Congress, there appeared, almost simultaneously, the earliest publications on the possibility of vaccinating laboratory animals against the toxins of diphtheria and tetanus by artificial methods. Immediately after the discovery of these poisons, the attempt was made to immunise various species of animals against them, but here very great difficulties were met with; the animals, after receiving increasing doses of toxin, became thin and ultimately died. It occurred to Fränkel[526] that the toxic action of the diphtheria poison might be weakened by subjecting it to a temperature of 60° C. Independently, von Behring and Kitasato[527] used chemical substances, especially iodine trichloride, to attenuate the action of the tetanus and diphtheria toxins. The animals which resisted these modified poisons were found to be capable of tolerating gradually increasing doses of unaltered and very active toxins. By the use of these methods it was found possible to obtain a definite and lasting immunity against these microbial products.

[Sidenote: [362]]

The discovery of the possibility of vaccinating against bacterial toxins was soon followed by the demonstration of the antitoxic power of the blood of animals that had acquired such artificial immunity against these poisons. Everyone knows of and appreciates von Behring and Kitasato’s great discovery. It opened up a new and fruitful field of research from most diverse points of view. Ehrlich[528] was able to apply it to the vaccination of animals against the vegetable poisons ricin, abrin and robin, and thus to establish rigorous methods of immunisation and to obtain very important results concerning immunity against toxins in general. He also succeeded in demonstrating that animals vaccinated against these vegetable poisons, which, by their nature, approximate to the microbial toxins, develop in their blood a most marked antitoxic property.

Some years later, the discovery of antitoxins was extended to snake venoms, poisons of animal origin which, like the vegetable poisons studied by Ehrlich, present a chemical composition analogous to that of the microbial toxins. Phisalix and Bertrand[529] and Calmette[530], working independently, discovered methods of vaccination against snake venom and were able to demonstrate the existence of an antitoxic power of the blood in immunised animals.

The works above briefly referred to gave us the fundamental basis of our present knowledge on acquired immunity against toxins.

It would be very interesting to be able to determine whether the lower animals can be vaccinated against the toxic substances to which they are susceptible. Unfortunately in the study of this problem we encounter very great difficulties. Making use of various methods I have often tried to solve it. The crayfish is susceptible to snake venom and to the ichthyotoxin of eel’s serum, and I have tried at various times to vaccinate it against these poisons. The results, however, were so inconstant and even contradictory that I was unable to draw any definite conclusion from them.

[Sidenote: [363]]

It is, indeed, very difficult to vaccinate the lower vertebrata against poisons. Several attempts have been made in my laboratory to immunise frogs against tetanus toxin, but without success. Calmette and Deléarde[531] obtained the best results with abrin. They succeeded in vaccinating frogs—which are not very susceptible to this vegetable toxin, though they are far from presenting a real natural immunity—against doses which are absolutely fatal for the control animals. These observers, however, had to proceed very cautiously, and they allowed a very long interval between each injection of abrin. The blood of their vaccinated frogs not only did not prove to be antitoxic against abrin, when injected into mice, but for long retained sufficient of this toxin to poison normal mice. This experiment certainly tells against the hypothesis that the acquired immunity of frogs is due to the development of a specific antitoxic power in their body fluids, but it does not settle the question definitely since it may be objected that the blood, whilst toxic for mice, might, still, be antitoxic for the frog. The antitoxin of this blood might merely be incapable of neutralising all the abrin present. Fresh investigations, then, are necessary.

Even in the higher vertebrata, it is often very difficult to obtain a real vaccination against the various toxins. In the small mammals, which exhibit a great susceptibility to these poisons, it is specially difficult to obtain an artificial immunity. As Vaillard and von Behring have demonstrated, it is possible to vaccinate such animals by means of gradually increasing doses of unmodified toxins, but this method demands much time, is often dangerous, and hence is not very practical. Poisons that act through the alimentary canal are the most serviceable for vaccination, as has been demonstrated by Ehrlich. This investigator had to abandon the vaccination of mice by means of subcutaneous injections of ricin on account of the sloughing set up at the point of inoculation. He then had recourse to vaccination by way of the mouth, which gave very good results, not only with ricin but also with abrin. This mode of vaccination, however, is applicable to a small number of poisons only.

We can also vaccinate mammals, even laboratory rodents, such as rabbits and guinea-pigs, by means of unmodified snake venom, but this method is a very delicate one and must be carefully watched. It is necessary to begin with very small doses of venom, continue them for some time, and increase the amount of venom injected very slowly. Calmette[532] modified this method by inserting, below the skin and leaving it there, a piece of chalk impregnated with small quantities of venom and surrounded by collodion through which the venom diffuses very slowly and continuously.

[Sidenote: [364]]

Large mammals, sheep, oxen and horses, can be more easily vaccinated by means of unmodified toxins, but they also require to be treated with very special precaution. Salomonsen and Madsen[533] have given the history of their horse, immunised with diphtheria toxin. Into a mare weighing 665 kilos they were able to inject at the commencement only 1 c.c. of this toxin, and the dose had to be increased very carefully.

In the presence of all these difficulties in the use of unmodified toxins for vaccination, a different method is now generally adopted in the immunisation of animals, small or large, for the purpose of scientific research or for the preparation of toxins on a commercial scale. Vaccination is commenced with toxins modified by heat or by chemical substances. The diphtheria and tetanus toxins, those most employed in the serotherapeutic industry, are subjected to various degrees of heat. Fränkel[534] was the first to make use of this method for vaccination against diphtheria, and Vaillard[535] for vaccination against tetanus. It consists in introducing large doses of filtered cultures, heated to progressively lower degrees of temperature, 60°, 55°, 50° C., and then giving gradually increasing quantities of filtered cultures whose toxicity is unaltered. This method is very convenient for small animals, but for large mammals it is greatly simplified by injecting for a certain period toxins heated to 60° C., and, later, replacing these by unmodified toxin.

[Sidenote: [365]]

Phisalix and Bertrand[536] applied an analogous method to the vaccination of the guinea-pig against the venom of the viper. This poison, which resists much higher temperatures than do the tetanus and diphtheria toxins, received a preliminary heating to 80° C. in order that it might be inoculated without danger into small animals. Under these conditions it confers a certain immunity, but even when heated to 80° C. it, in many cases, still remains sufficiently active to produce fatal results. For this reason, in the vaccination of animals for the preparation of antivenomous serum on a large scale, Calmette had recourse to another method, that of attenuating the venom by means of chemical substances.

Von Behring and Kitasato[537] were the first to make use of iodine trichloride in the vaccination of animals against the toxins of tetanus and diphtheria. In their early experiments this substance was injected before the toxins were introduced. Later, the mixture was made _in vitro_ and then injected into the animals. Roux devised another method which had the advantage of being simple, certain, and easily employed, for which reason it was soon introduced into commercial and scientific practice. It consists in the injection of mixtures of the tetanus or diphtheria toxins with Lugol’s iodo-ioduretted solution. The iodine, in small doses, instantly neutralises or modifies these poisons and is itself borne well, even by small animals. By employing progressively increasing doses of these mixtures, in which the amount of iodised solution becomes smaller and smaller compared with that of the toxin, we are able, without difficulty, to vaccinate the most susceptible animals and enable them to withstand considerable doses of the pure toxin. By this method it is possible to immunise guinea-pigs against the most active tetanus toxin. The method serves equally well for the preparation of horses for injections of unmodified toxins. For a longer or shorter time (according to the susceptibility of the horse) toxins which are mixed with Lugol’s iodised water are injected. Having made sure of the resistance of the horse, larger and larger quantities of pure, unmodified toxin may be introduced with impunity.

For the immunisation of mammals of all sizes (guinea-pigs, rabbits, dogs, horses) against snake venom, Calmette, in his work at Lille, also makes use of venom modified by chemical substances, but his method differs from those we have just described. During several weeks he injects increasing quantities of venom, mixed with decreasing quantities of a solution of 1:60 of hypochlorite of lime. After this treatment the animals become capable of tolerating fatal doses of unmodified venom and can be injected with larger and larger doses.

[Sidenote: [366]]

In recent years a method of vaccinating horses against certain microbial toxins, and especially against the diphtheria toxin, by means of mixtures of toxin and antitoxic serum, or with these two products successively, has been introduced. Babes[538] was the first to extol this method as the best for obtaining a high and durable immunisation. Afterwards, several other observers, amongst whom I may cite Pawlowsky and Maksutow[539], Palmirsky, and especially Nikanoroff[540], took up this question, and communicated very encouraging results. Von Behring[541] also found it very useful in certain cases. Thus, for the vaccination of guinea-pigs against tetanus toxin, he recommends the injection of a mixture containing antitoxin and an unneutralised excess of toxin. Under these conditions he easily succeeds in immunising these small animals in cases where all other methods fail. As a general method of vaccination against toxins, however, this method has not fulfilled its promise, and Roux, who tried it several times, was not at all satisfied with it.

[Sidenote: [367]]

This method of immunisation by mixtures of toxin and antitoxin is often spoken of as the method of vaccination by _toxones_. This name, “toxone,” was first applied by Ehrlich[542] to a product developed by the diphtheria bacillus in culture media, a product less and differently toxic than is the true diphtheria toxin, yet capable of neutralising antitoxin. The idea of toxones presented itself to Ehrlich in connection with a fundamental fact noted by him, namely, that when to a non-toxic mixture of diphtheria toxin and antitoxin there is added one and even several lethal doses of the former, the animal is not affected. To make it succumb to intoxication it is sometimes necessary to add more than 20 lethal doses of toxin. To explain this paradoxical result, Ehrlich formulated the hypothesis that, in the soluble products of the diphtheria bacillus there exist two poisons: (1) the true toxin which exhibits a very strong affinity for antitoxin, and (2) the toxone which possesses less avidity for this antibody. When to an inactive mixture of the products of diphtheria bacilli and of antitoxin, there is added a fresh quantity of these same products, the added toxin, owing to its greater affinity, replaces the toxone of the previous combination. In the mixture to which is added one or several lethal doses of diphtheria poison, the toxone only is found free, all the toxin being combined with the antitoxin, and, as the toxone is only feebly toxic, the animal resists without suffering any serious illness.

Madsen[543] adopted the theory of the diphtheria toxone, and affirmed that this substance poisons but slowly, produces neither early nervous symptoms nor loss of hair, but excites slight oedema at the point of inoculation and late paralyses. Susceptible animals may die from toxones, but very much later than as the result of poisoning by the toxins.

Ehrlich’s pupils have extended the theory of toxones to other bacterial poisons. Thus Madsen[544] has described a similar toxone in tetanus poison—the tetanolysin of Ehrlich—which dissolves the red blood corpuscles, and Neisser and Wechsberg[545] refer to a toxone in the poison produced by the staphylococcus.

Ehrlich also describes _toxoids_ as occurring in diphtheria poison. The toxone, he maintains, is a product of the diphtheria bacillus itself, but the toxoids (protoxoids and syntoxoids) represent the toxin modified without further aid from the bacillus. The toxoids, though not toxic, retain all their avidity for antitoxin. According to Ehrlich’s conception, the molecule of toxin, under the influence of various factors, readily loses its toxic or _toxophore_ group, capable of poisoning the animal, whilst still retaining its _haptophore_ group, the group that combines with the antitoxin. The toxoids then would represent this haptophore group of the diphtheria toxin. Without being injurious to animals, the toxoids are capable of neutralising the antitoxin and of setting up in the animal the formation of this antibody. In the experiments carried out by the method of Babes and of the Russian authors we have just mentioned, there would be, according to the view held by Ehrlich and his school, an immunisation by the toxoids.

[Sidenote: [368]]

The toxones, however, are also capable of vaccinating against the toxin and the toxone and of giving rise to the production of a diphtheria antitoxin, active against these two poisons. This is what is affirmed by Madsen[546] and by Dreyer[547], according to a communication made by the latter to the International Congress of Medicine held at Paris.

[Sidenote: [369]]

By means of the various methods briefly described above, is obtained a real acquired immunity against the various bacterial and vegetable poisons and the venoms. On the other hand, with the methods of vaccination mentioned in the eighth chapter, which confer a substantial immunity against micro-organisms, we cannot demonstrate, in the vaccinated animals, a resistance against the corresponding toxins greater than in the unvaccinated control animals. The animals, so thoroughly vaccinated against certain micro-organisms that they withstood enormous doses of culture, did not become capable of resisting the minimal lethal dose of the poison. We are led to conclude, therefore, that immunity can only be obtained against certain of the toxins. For this reason we must regard the attempt made by von Behring to obtain a real immunisation against the toxin of cholera as an important forward step. Before von Behring’s attempt, various species of animals had been frequently and very substantially vaccinated against the cholera vibrio, but these animals, even when most thoroughly vaccinated, were completely non-resistant to the cholera toxin. Von Behring suggested to his pupil Ransom[548] the idea of immunising guinea-pigs, not with microbial cultures living or dead, as had usually been done previously, but exclusively with the fluids of the cultures, deprived of the vibrios by filtration. In order, however, to attain the desired object, it was necessary to prepare fluids sufficiently active to poison the unvaccinated control guinea-pigs with certainty. The results of these investigations confirmed his anticipation, and Ransom soon found himself in possession of guinea-pigs well vaccinated against the cholera poison. He was mistaken, however, in supposing that, in all cases of immunity acquired against Koch’s vibrio, we have to do, in the main, with a purely antitoxic immunity. An investigation carried out in the Pasteur Institute[549], whilst confirming the facts discovered by Ransom, lead to different results as regards their interpretation. It was demonstrated that the immunity against the vibrio is in no way founded on a resistance against its toxin and that we have to do with two very different acquired immunities. The vaccination obtained with the bodies of the micro-organisms induced a refractory condition against infection by the living vibrio, but not the slightest resistance against the toxin. The immunity, on the other hand, which is conferred by the injection of soluble products, deprived of the micro-organisms, is effective not only against the toxin of cholera, but also against infection by the vibrio. When an animal is vaccinated with cultures, or even with the bodies only of the vibrios, cholera toxin is introduced, but the toxin, under these conditions, is incapable of setting up antitoxic immunity. It would appear that the presence of the vibrios may constitute some obstacle to the production of this immunity.

Soon afterwards, Wassermann[550] pointed out that the same rule applies in the case of the _Bacillus pyocyaneus_. With whole cultures of this bacillus he obtained in guinea-pigs an immunity exclusively against infection, whilst with cultures in a fluid medium, deprived of the bacilli, he was able to vaccinate his animals both against the pyocyanic toxin and against the infective peritonitis produced by the living micro-organism. The same double immunity could also be obtained in laboratory animals against the typhoid bacillus and several other bacteria.

When animals were subjected to different methods of vaccination against toxins, the manifestation of certain phenomena more or less constant was observed; amongst these must be pointed out especially the rise of temperature, a local reaction and certain modifications in the body fluids.

Fever is a very general symptom in the course of the vaccination of mammals. A rise of temperature is almost always observed as a result of the injection of toxins. It is very variable, both as regards duration and intensity, and cannot serve as an indicator of the result of the vaccination. In this respect, such great differences have been observed that the attempt to establish any general laws has had to be abandoned.

[Sidenote: [370]]

[Sidenote: [371]]

Local reaction is also a phenomenon which is very frequently observed during vaccination; to this von Behring[551] paid great attention. He and his collaborators found that normal horses when injected subcutaneously with small or large doses of tetanus toxin did not present any exudation at the seat of inoculation. The horses which died as the result of a tetanus intoxication and those which got better behaved from this point of view in much the same fashion. In horses, however, which are being vaccinated and which are periodically subjected to gradually increasing doses of toxin, tumefaction at the seat of injection is never absent. Von Behring attributes this difference to the primordial insusceptibility of the living elements which govern exudation in the subcutaneous tissue to tetanus poison. It is only during the process of vaccination that these cells become susceptible and capable of manifesting a visible reaction. I consider that this difference is due more probably to a change in the chemiotaxis of the various elements which contribute to the inflammatory exudation reaction, from a negative to positive type. The cells do not react at the commencement, not because they are not susceptible to the toxin, but rather because their susceptibility is too great. During the course of vaccination they become sufficiently adapted to the poison to be able to manifest their normal inflammatory reaction. This explanation certainly harmonises with the fact that during the period of vaccinations in general and of vaccination against toxins in particular, the blood usually presents a more or less distinct hyperleucocytosis. Now, as is well known, this phenomenon of hyperleucocytosis is one of the most striking manifestations of a positive chemiotaxis in white corpuscles. It is true that, as to this reaction during the course of vaccination, the views of observers are not unanimous. Besredka[552], as the outcome of his work on this subject, expresses himself very distinctly. “During the course of an immunisation against diphtheria toxin,” he writes, “one always observes a marked reaction in the goat, either at the beginning or at an advanced stage of the period of injections and especially in the first few hours after injection” (p. 322). Nicolas and Courmont[553] in their first memoir maintain that hyperleucocytosis “is not necessary for immunisation.” Nevertheless, in the description of their experiments, which were performed on horses vaccinated against diphtheria, it is clear that the number of white corpuscles is often markedly increased. Further, in several cases they describe the formation of tumours at the point of inoculation, some of which end in suppuration. Under these conditions, it is not possible to deny a vaccinal reaction on the part of the leucocytes. Later, Nicolas, Courmont and Prat[554] published a second memoir on the same subject, in which they seek to confirm their view of the uselessness of hyperleucocytosis in vaccination against the poison of diphtheria. They give details of experiments on several species of animals and insist specially on the conditions in which they have not observed hyperleucocytosis. “The doses from the first have always been extremely weak and with the addition of Lugol’s solution to attenuate them; only very gradually have we reached stronger doses, as _that is one of the indispensable conditions for the avoidance of leucocytic variations_, whilst obtaining a good and rapid immunisation” (p. 974). These special precautions to avoid hyperleucocytosis demonstrate clearly that this phenomenon is usually produced during the course of vaccination. It is quite natural that we should, by proceeding very slowly and with small doses of toxin, succeed in diminishing or even suppressing the afflux of leucocytes; but this fact cannot in any way minimise the importance of the leucocytic reaction in vaccination. In these particular cases, this reaction may take place without the number of leucocytes in the blood being noticeably increased. In reading the details of the experiments made by the Lyons observers, it will be seen that, in spite of all their precautions, they were unable to prevent the production of hyperleucocytosis. In all their cases, where they took the precaution to count the leucocytes several times a day, there was an undoubted increase of these cells. We may here recall Salomonsen and Madsen’s account of the immunisation of a horse against diphtheria toxin, in which they point out the frequency of tumefactions and even of abscesses. In most cases the pus was sterile, which renders it probable that the white corpuscles had accumulated at the seat of inoculation as the result of some influence exerted by the diphtheria toxin.

[Sidenote: [372]]

By far the most important and remarkable change met with in animals vaccinated against toxins and venoms, consists in the appearance of antitoxic power in their blood and fluids in general. This fact was, as already mentioned, first demonstrated by von Behring and Kitasato[555] in the blood of rabbits immunised against tetanus. The blood itself, or the blood serum, mixed with a quantity of tetanus toxin more than sufficient to cause fatal poisoning, sets up no disease when injected into animals. In their earliest researches, von Behring and Kitasato kept the mixtures in contact _in vitro_ for 24 hours, before injecting them into test animals. Later, they found that this prolonged contact outside the body was unnecessary and that they could obtain successful results by injecting the serum of vaccinated animals and the toxin simultaneously, even at different points of the body. This discovery was immediately afterwards applied by its authors to diphtheria and, in the case of both intoxications, confirmed by numerous observers.

[Sidenote: [373]]

For some time we were satisfied with vaccinating small laboratory animals and establishing the antitoxic power of their blood serum; later, the vaccination of large animals, especially horses, was commenced with the object of obtaining large quantities of antitetanus and antidiphtheria serum for medical use. During the course of these experiments the principal characters of the antitoxic fluids were established. It was deemed desirable to isolate the antitoxic substance from the blood serum in order to get rid of every unnecessary and inactive admixture, so that the antitoxin might be used in as pure a form as possible. This idea of isolating the antitoxic substance had, however, soon to be abandoned as impossible of realisation. Antitoxin is a non-crystallisable substance, of unknown chemical composition, which adheres firmly to the albuminoid substances of the serum. It is usually regarded as belonging to the same albuminoid group of substances, though it is not possible to prove this satisfactorily. Von Behring[556], however, who studied this question in collaboration with Knorr, denies the albuminoid nature of tetanus antitoxin. After demonstrating that this antitoxin, when the antitetanus serum is submitted to dialysis, passes through the dialysing membrane, these observers found that they could not obtain the characteristic reactions of albuminoids in the dialysed fluid. It must be admitted, however, that this negative result is not sufficient to justify a denial of the albuminoid nature of antitoxin. When Nencki and Mme Sieber[557] sought to produce the reactions of albuminoid substances with the digestive juice of _Nepenthes_ (the well-known insectivorous plant) they got no result; but after the concentration of the juice _in vacuo_, it at once gave the characteristic reaction with nitric acid, and also with acetic acid, potassium ferrocyanide and Millon’s reagent.

The antitoxins may be precipitated along with the globulins and are distinguished, in general, by a fairly great resistance against physical and chemical influences. In this respect they are allied to the agglutinins, the fixatives and the precipitins, considered elsewhere, and are sharply distinguished from the cytases. The antitoxins resist temperatures which destroy the cytases and remain unaltered to beyond 60°–65° C. They are more stable than the delicate toxins of tetanus and diphtheria, but they are more easily altered than the toxins of cholera, of _Bacillus pyocyaneus_ and the venoms. When stored in a dry state in the residue of evaporated serums and protected from light and air, the antitoxins will keep for a very long time without showing any notable attenuation. This property is very important in practice.

[Sidenote: [374]]

The antitoxins, in this respect also resembling the fixatives and the agglutinins, are humoral substances in the strictest sense of the term. They are found not only in prepared serums but abound also in the plasma of the circulating blood, and in the plasmas of the lymph and of exudations. Vaillard and Roux[558] have shown that the clear acellular serous fluid of the oedema produced by the slowing of the circulation in rabbits vaccinated against tetanus toxin, is as antitoxic as the blood itself. Even the aqueous humour of a strongly immunised animal is antitoxic, though to a less degree. On the other hand, the saliva and urine exhibit very little antitoxic power, even when they are derived from animals hyperimmunised against tetanus toxin. Milk, as first demonstrated by Ehrlich[559], is fairly rich in antitoxin, although much less so than the blood. According to the estimation of Ehrlich and Wassermann[560], in the same immunised animal, milk contains one-fifteenth to one-thirtieth of the amount of diphtheria or tetanus antitoxin contained in the blood. Pus is always less antitoxic than blood or blood serum. According to Roux and Vaillard (_l. c._, p. 82), the pus of their rabbits vaccinated against tetanus toxin was only one-sixth or one-eighth as antitoxic as the serum of the blood. In Salomonsen and Madsen’s[561] antidiphtheritic horse the cellular sediment of the pus was about one-half as antitoxic as the blood.

For the development of the antitoxic property in the fluids of the body, it is not essential that animals should belong to species susceptible to the corresponding toxin. Animals naturally most refractory against the poisons of diphtheria and tetanus are also capable of producing antitoxins. Vaillard[562] demonstrated this fact in the fowl. This bird, which is naturally refractory against tetanus, usually acquires a very marked antitetanic power in its blood after one or more injections of tetanus toxin. He observed, however, that, in fowls thus treated, at a stage when the fluids of the body are antitoxic, the albumen of the egg is not so. The antitoxin, therefore, does not pass into this nutritive secretion, as it does into the milk of mammals. On the other hand, as has been demonstrated by F. Klemperer[563], the vitellus of the eggs of fowls treated with tetanus toxin in time acquires an antitoxic property of the most marked character.

[Sidenote: [375]]

[Sidenote: [376]]

The antitoxins, found especially in the fluids of the body but only scantily in the cells, exert some action on the toxins. What is the nature of this action? This question, much studied and discussed, is one of very great importance in connection with the general problem of acquired immunity against toxins. In his first memoir, written in collaboration with Kitasato, von Behring (_Deutsche med. Wchnschr._, Leipzig, 1890, S. 1113) formulates his first thesis as follows: “the blood of a rabbit immunised against tetanus possesses the property of destroying tetanus toxin.” This idea of destruction, which would remove all toxic power from the poison, would naturally present itself to the mind and was at once accepted by a great many observers, but the numerous facts now accumulated on the subject will not allow us to accept a real destruction of toxins by antitoxins. Tizzoni[564] was one of the first to point out certain contradictions between the theory of destruction and the phenomena produced in animals injected with tetanus toxin and antitoxin. Buchner[565] also brought forward new facts which led him to conclude that antitoxin, instead of acting directly on the toxin, exerts its influence exclusively on the living elements, thus protecting the animal against intoxication. Amongst the arguments advanced by the Munich observer, the principal one is drawn from the different action of mixtures of tetanus toxin and antitetanus serum on various species of animals. It has been clearly shown that the guinea-pig is more susceptible to tetanus than is the mouse. In poisoning with tetanus toxin it requires an absolutely larger quantity of toxin to kill the guinea-pig than to kill the mouse. But if we take into account the weight of these animals, the conditions change completely. Thus, to cause a fatal tetanus in a guinea-pig, which weighs twenty times more than a mouse, we need only inject into the former a dose at most ten times greater than that necessary to produce fatal intoxication in the mouse. Buchner prepared a mixture of tetanus toxin and antitetanus serum which, in the mouse, produces no tetanic phenomenon or only sets up feeble and transient symptoms. According to the theory of direct action, we must assume that in this mixture the toxin is completely or almost completely neutralised by the antitoxin of the serum. But when Buchner injected the same quantity of mixture into guinea-pigs, without increasing it in proportion to the greater weight of these animals, he produced a tetanus of the most marked character. There has, consequently, remained in the mixture a sufficient amount of free toxin, whose tetanigenic action is manifested in the guinea-pig, an animal, as we have seen, more susceptible than the mouse. Buchner’s experiment has been verified by several observers. Roux and Vaillard[566] carried out others which afford similar evidence. The same mixture of tetanus toxin and specific serum which is borne without the least difficulty by normal guinea-pigs, causes typical tetanus in other guinea-pigs of the same weight, and apparently in the best of health, but which have been immunised some time before against the Massowah vibrio. In another series of experiments, Roux and Vaillard injected into guinea-pigs a very large amount of antitetanus serum “capable of immunising them thousands of times,” and, shortly afterwards, a lethal dose of tetanus toxin. The normal guinea-pigs were thoroughly resistant to this test, whilst several guinea-pigs into which were also injected the products of other micro-organisms, acquired tetanus. Analogous results were obtained with mixtures of diphtheria toxin and antidiphtheria serum. Roux concludes from these facts “that the antitoxins act on the cells.” Against the theory of the destruction of toxins by antitoxins, he invokes the influence of heat on mixtures of these two substances. Calmette[567], under Roux’s inspiration and in his laboratory, carried out various experiments on antivenomous serum. A mixture of this with snake venom, in such proportion that the poison became inactive, regained its toxicity after being heated for five minutes at 68° C. A normal animal, injected with this mixture, succumbed as if it had received pure venom. On being heated at 68° C. the serum lost all its antitoxic power over the venom, and the latter, which only becomes modified at a much higher temperature, remained intact. Later, a similar result was obtained by Wassermann[568] in his experiments with pyocyanic toxin. This poison is resistant at even higher temperatures than is snake venom, whilst the antitoxin of the serum is destroyed under the same conditions as are the other antitoxins. Taking advantage of these peculiarities, Wassermann boiled the mixture of pyocyanic toxin and antitoxin serum, being careful to dilute it with two volumes of distilled water before doing so. This mixture which, before it was heated, was quite innocuous for guinea-pigs, again became a fatal poison after the destruction of the antitoxin.

[Sidenote: [377]]

These experiments prove clearly that, in the action of the antitoxin on the toxin, there can no longer be any question of an actual destruction of the latter, a view which has been accepted by both von Behring and Ehrlich. But, as pointed out by Roux at the International Congress at Budapest in 1894, the manifestation of the toxic action of the venom after it has been heated along with antitoxin, may be reconciled with the view that the combination between the two substances, if such take place, must be very unstable. This same remark may be applied to Wassermann’s experiment. Therefore the great majority of observers, if not all, admit that the antitoxin combines with the toxin to form an innocuous and unstable substance which can be decomposed by heat and by other agents. The researches on the action of antitoxins _in vitro_ have had a powerful influence in determining this view.

[Sidenote: [378]]

We have already in Denys and van de Velde’s[569] experiments an indication of the direct action of certain antitoxins. These observers showed that the serum of animals vaccinated against a Staphylococcus is capable of neutralising _in vitro_ a particular toxin to which van de Velde gave the name of _leucocidin_. When it was added to a drop of the exudation from a rabbit, this leucocidin in a very short time destroyed the white corpuscles, by dissolving the cell content but leaving the nucleus untouched. When Denys and van de Velde prepared mixtures of leucocytes, leucocidin and antileucocidic serum _in vitro_, the white corpuscles retained their normal condition for a very long time. The leucocidin was, therefore, rendered inactive by the direct influence of the corresponding antitoxin. These facts have been confirmed by Bail[570] and other observers and even extended to certain other microbial toxins. Thus, the _Bacillus pyocyaneus_ produces a leucocidin which kills the white corpuscles and dissolves their contents[571]. With the object of facilitating experiments with these leucocytic poisons and the corresponding antitoxic serums, Neisser and Wechsberg[572], of the Institute of Experimental Therapeutics at Frankfort, invented a method which allows us to observe the phenomena of the destruction of the leucocytes and of the antitoxic power in test tubes, without having recourse to a microscopical examination. They applied the fact, discovered by Ehrlich, that living formed elements reduce methylene blue and, depriving it of its oxygen, decolorise it. Leucocytes from aseptic exudations are introduced into tubes and a weak solution (2%) of methylene blue is poured on them. To prevent the re-oxidation of this colouringmatter by the oxygen of the air, the surface of the fluid is covered with a layer of liquid paraffin. If the leucocytes are living, the lower blue layer becomes decolorised in a short time (in about two hours); when the corpuscles are dead, decoloration does not take place. By adding to the mixture of leucocytes and colouring matter some leucocidin, alone or along with antileucocidic serum, it is possible not only to observe with the naked eye the phenomena which take place in these cases, but also to estimate to some extent the proportions of poison and counterpoison.

All these researches make it clear that the antitoxin acts directly on the leucocidin. Similar facts have been noted as regards certain other organic poisons and their antitoxins. Shortly after the discovery of antileucocidin by Denys and van de Velde, Kanthack made a communication to the Physiological Society in 1896[573], exhibiting tubes in which the coagulating action of Cobra venom on the blood had been prevented by the addition of antivenomous serum. Of all the experiments, however, made to prove the direct action of antitoxin on toxin, Ehrlich’s[574] have played the most important part in the study of this question. Ehrlich directed his attention to ricin which, as Kobert demonstrated, has the property of agglutinating the red corpuscles of defibrinated blood. This phenomenon can be easily observed _in vitro_. In tubes containing red blood corpuscles, the addition of ricin causes these corpuscles to agglutinate into clumps and to fall to the bottom of the tube, leaving a clear supernatant fluid. After adding progressively increasing quantities of antiricic serum to the tubes containing fluid blood and ricin, Ehrlich was able to demonstrate that small quantities of antiricin merely retarded the precipitation of the red corpuscles, whilst larger doses completely prevented it. Having studied the proportions of ricin and its antidote, necessary to retard and prevent the fatal poisoning of animals, Ehrlich was struck by the parallelism which is exhibited between the action of the antitoxin in the living animal and that in the test tubes.

[Sidenote: [379]]

The study of anticytotoxins, discussed in the fifth chapter, has furnished another opportunity of observing the action of antitoxins _in vitro_. Camus and Gley and H. Kossel were the first to observe the action _in vitro_ of antitoxic serum against the ichthyotoxin of eel’s serum. Since this observation, this phenomenon has been repeatedly studied in the antihaemolysins and antispermotoxins. The antidiastatic serums also act _in vitro_ and, as their effect can be demonstrated on soluble ferments placed in contact with unorganised bodies, such as gelatine and casein, the purely chemical character of the reaction is all the more strikingly shown. We are indebted to von Dungern, Briot and Morgenroth for accurate observations on this subject.

Martin and Cherry[575] made use of a different method to demonstrate the direct action of antitoxins on toxins which exhibit their toxic power on the animal organism. They chose snake venom mixed with antivenomous serum. The mixtures were filtered under great pressure [50 atmospheres] through a film of gelatine, under the idea that, if the venom and antitoxin were not chemically combined, the former alone, owing to its much smaller molecules as compared with those of the antivenom, would pass into the filtered fluid. This fluid should, under these conditions, possess a toxic power for animals, when the mixture, used for filtration, was deprived of the larger molecules. Martin and Cherry left the venom and the antitoxic serum in contact for periods of varying length, before filtering the mixtures. As the result of a series of such experiments carried out according to this scheme, they found that the product of the filtration made after some minutes’ contact between the two substances, was distinctly toxic; whilst the filtrate obtained after a contact of half-an-hour was absolutely innocuous. From their observations these authors conclude that the antitoxin enters into chemical combination with the venom, but that the combination does not take place instantaneously, a certain amount of time being necessary for its accomplishment.

[Sidenote: [380]]

In addition to the time factor others have an influence on the combination between toxins and antitoxins, as is seen from Ehrlich’s[576] and Knorr’s[577] investigations. Both observers have shown that antitoxin neutralises the toxin more slowly in dilute solutions than in more concentrated form. For this reason, when animals are injected with very weak solutions, the toxin may manifest its action before it can be neutralised by the antitoxin; this may lead to erroneous conclusions. On the other hand, according to data furnished by these authors, temperature also exerts an influence on the combination. Lowering the temperature retards, whilst raising it accelerates the neutralisation of the toxins by the antitoxins. Insisting on the purely chemical character of the combination between these two substances, Ehrlich and Knorr adduce the fact that this combination, in cases where we have a complete neutralisation of the toxin, follows, most rigorously, the law of multiple doses, that is to say, in order to render innocuous a hundred doses of toxin we have only to take a hundred times the quantity of antitoxin.

The series of facts summarised above demonstrate distinctly that antitoxins act directly on toxins. But how can this result be reconciled with the observations given above according to which must be admitted the no less real influence of the organism of the living animal on intoxication by mixtures of antitoxin with toxin? Knorr[578] sought at first to minimise the importance of the facts brought forward by Buchner and Roux. He failed to corroborate Buchner’s results and found that the injection of mixtures, made with very large doses of tetanus toxin (20,000 times the minimal lethal dose) and corresponding quantities of antitetanus serum, brought about the same effect in guinea-pigs and mice. By modifying the quantity of antitoxin, he rendered the mixture equally innocuous or equally toxic for these two species. But the data given by Knorr are quite sufficient to prevent us from accepting his conclusion. In his experiments, as in those of Buchner, the guinea-pigs manifested a greater susceptibility and died from mixtures which, in mice, caused merely a tetanus of medium intensity.

[Sidenote: [381]]

Some have sought to explain Buchner’s experiment by assuming that the mixtures, lethal for the guinea-pig and innocuous for the mouse, owed their toxic action to the presence of the _tetanus toxone_ and not of the true tetanus poison, the _tetanospasmin_. This hypothesis of toxones, as stated above, was put forward by Ehrlich as the outcome of his ingenious researches on the constitution of the diphtheria poison. As, however, the toxones must act differently from the toxins, we can only attribute to their action the results in those cases where the guinea-pigs die without presenting typical symptoms of true tetanus, that is to say without spasms. Now, in Buchner’s experiments, a much larger proportion of these animals, injected with the same mixtures as the mice, succumbed and exhibited the characteristic tetanic convulsions. Even in those cases, however, where the death of the guinea-pigs might be attributable to an intoxication by the toxone, the general result could not be altered. The toxones are, according to Ehrlich, manufactured by the micro-organisms in the culture media and form an integral part of the natural microbial poisons. Again, they are neutralised by antitoxic serums. If, therefore, in spite of there being the same quantity of toxones and of antitoxin in the mixtures, these mixtures become more toxic for the guinea-pig than for the mouse, we have an indication that some special change must take place in the animal to upset the conditions of toxicity.

Weigert[579] accepts the accuracy of Buchner’s experiment, which, indeed, can no longer be denied, but explains it on the hypothesis that there is some substance in the animal possessing a very great affinity for the toxin. This substance is supposed to be capable of decomposing the innocuous combination of the antitoxin with the toxin, just as heat does in Calmette’s and Wassermann’s experiments, described above. In both cases the toxin would be set free to exert its noxious action. Such a hypothesis is very probable, because it agrees with direct observation, but it compels us to accept some new phenomenon which is produced not _in vitro_, but in the living animal, and which carries on its work in a very different fashion in the guinea-pig and in the mouse.

In the present imperfect state of our knowledge it is very difficult to form any idea of the precise conditions which must intervene in the organism of the guinea-pig to cause the tetanus toxin to act in a mixture with antitoxin which is much more innocuous for the mouse. In order, however, to satisfy those who seek to understand these complex phenomena, it may be useful to cite another example of antitoxic action in which certain factors are distinguished by their simplicity.

[Sidenote: [382]]

Lang, Heymans and Masoin[580] have demonstrated that hyposulphite of soda prevents poisoning by prussic acid. This terrible poison becomes innocuous if we take care to introduce into the animal by any channel whatever (subcutaneously, intravenously, or by the stomach) a sufficient quantity of hyposulphite of soda. Under these conditions the sulphite is substituted for the hydrogen of the prussic acid, transforming the poison into sulphocyanic acid, which has no action on the organism. The hyposulphite of soda, then, acts as the antitoxin of the prussic acid, thanks to a chemical reaction of substitution between bodies of simple composition. We have never yet succeeded in reproducing this reaction _in vitro_, whilst in the animal body it is effected with very great ease. Consequently, we are quite justified in invoking special conditions in the body of the living animal; this, however, does not preclude the possibility of a transformation of the toxic substance into an innocuous substance through a chemical reaction. It is probable that analogous phenomena may also be met with in the action of true antitoxins on the microbial toxins or allied substances (venoms, vegetable toxalbumins).

[Sidenote: [383]]

The case of the destruction of micro-organisms, which is now more easily studied because it is possible to observe with the eye the fate of these organisms in the animal, is a further source of valuable information. The direct action of cytases on certain bacteria, such as the cholera vibrio, can be just as easily demonstrated _in vitro_ as can the action of antiricin on ricin. If we proceeded to argue from this, a perfectly accurate observation, that the living animal plays no part in the destruction of the micro-organisms and that this destruction takes place always in a fashion analogous to Pfeiffer’s phenomenon _in vitro_, we should undoubtedly arrive at an erroneous conclusion. We know already, as has been indicated in previous chapters, that the granular transformation of vibrios is only part of a whole series of phenomena of destruction of micro-organisms, the great majority of which phenomena require more or less active intervention of the animal organism. In reality, matters usually go on in a very complicated fashion, in which direct and indirect actions are blended in varied proportions. In the examples described elsewhere, we see, alongside the granular transformation, an agglutination into clumps and immobilisation, and an ingestion and intracellular destruction of micro-organisms. The final phase, no doubt, is always a chemical or physico-chemical action, exerted against the micro-organism, but how varied are the means used to bring about this result! We may surely be allowed to suppose that analogous phenomena may take place in the action of antitoxins on the toxins.

[Sidenote: [384]]

Just as, in the analysis of the influence of serums on the micro-organisms, it was found useful to study the action of certain fluids less complicated than the anti-infective specific serums, so we may utilise information furnished by the antitoxic action of fluids other than the true antitoxins. Cases are by no means rare in which normal serums exert a certain influence on toxins. Thus, Pfeiffer[581] noted that the normal blood serum of the goat has the power to prevent fatal poisoning by the cholera toxin. Freund, Grosz and Jelinek[582] observed an analogous action of solutions of nucleohiston on diphtheria intoxication and Kondratieff[583] demonstrated the same action of an extract of the spleen on the tetanus poison. Calmette[584], in collaboration with Deléarde, studied the influence of a whole series of fluids on abrin intoxication. Whilst physiological saline solution was absolutely incapable of preventing the death of animals, fresh broth exerted an undoubted antitoxic power. Amongst normal serums, ox serum exhibited a certain antirabic property. More, however, than the serums of normal animals, have those of animals immunised against various toxins other than abrin (antitetanus, antidiphtheria, antivenomous serums, &c.) been found to possess the power of preventing intoxication by abrin. These facts are connected with others of analogous nature, previously demonstrated by Calmette[585], of which I may cite the following: the serum of animals vaccinated against tetanus toxin is active, though to a less degree, against snake venom; the serum of rabbits vaccinated against rabies, a serum powerless to protect against this disease, is, however, very markedly effective against the same venom; the serum of animals immunised against snake venom is also antitoxic against scorpion venom (I have myself had the opportunity of confirming this fact on several occasions). In all these examples, the serums have proved to be less efficacious against poisons other than the toxin with which the animals that furnished the blood had been treated. Ehrlich[586], too, has demonstrated that animals vaccinated against robin (toxalbumin of _Robinia pseudacacia_) produce a serum, antitoxic not only against this poison but also against ricin. It need scarcely be added that in all these cases of non-specific action of serums derived from vaccinated animals, no question of any antitoxic effect of normal serums can enter. In all the experiments just summarised, the serums of normal animals, used as controls, were found to be inefficacious.

[Sidenote: [385]]

If, in the case of the non-specific action of serums, it were allowable to advance the hypothesis of a direct influence of these fluids on the toxins, it would still be impossible to sustain this view where broth fulfils the antitoxic _rôle_. This fluid, much simpler in composition than any serum, is an excellent culture medium for micro-organisms and one in which the toxins develop well and can be kept for a fairly long period. There is, therefore, not the slightest ground for assigning to it any direct antitoxic action, on the contrary, everything leads us to regard it as an indirect agent, which acts by stimulating the reaction of the animal organism. Here, then, the case would be quite analogous to that of the action of broth as a protective agent against certain bacterial injections, a subject already discussed in the tenth chapter. In this same category of indirect influences also, must be ranked the example of the antitoxic action of the blood of the crayfish against scorpion venom. I have demonstrated in a series of experiments that the fresh blood of the crayfish has the power to prevent fatal intoxication of mice by scorpion venom. Injected in a dose of from 1 to 1·25 c.c., several minutes or an hour before the injection of the rapidly fatal dose of scorpion venom, the crayfish’s blood exerts a very distinct preventive action. It might be supposed from this that the crayfish belongs to the group of animals insusceptible to scorpion venom. This, however, is not the case. The crayfish is very susceptible to this poison and succumbs to a quarter the dose necessary to kill a mouse. The blood of the crayfish is, therefore, completely ineffective as a protective to the crayfish itself, and only exerts its action when introduced into the body of the mouse. It might be concluded that it is only after it has been drawn from the crayfish that the blood acquires its antitoxic power. Experiment contradicts this supposition. Crayfish blood, when injected into another crayfish, in equal or greater amount than is necessary to protect a mouse, is incapable of preventing fatal intoxication by scorpion venom, although, here again, the crayfish received only one-quarter of the dose of venom used for the mice.

[Sidenote: [386]]

We are, therefore, compelled to believe that the crayfish’s blood is antitoxic for the mouse, not in virtue of its direct neutralising action on the venom, but owing to some indirect influence on the organism of the mouse. It is impossible to define, exactly, the mechanism of this action. We may suppose that the blood of the crayfish contains some substance which, by itself, is insufficient to prevent the intoxication, but which becomes active in the presence of some other substance, also inefficacious by itself, met with in the organism of the mouse. Here we should have something analogous to what is met with in immunity against micro-organisms where both fixatives and cytases intervene to bring about the destruction of micro-organisms. By making researches _in vitro_ on the action of the fluids on bacteria, we may easily observe certain phenomena which appear to indicate their direct influence. Take the case of the fluid of an oedema from an animal vaccinated against the cholera vibrio which renders this micro-organism motionless and agglutinates it _in vitro_; the oedema of an unvaccinated animal produces no such effect. If, however, we were to conclude from this fact that, in the oedema of the living animal or in its subcutaneous tissue, everything goes on as in the test-tube and that no other phenomenon of reaction against the vibrios is produced, we should fall into a grave error. It is extremely probable that, in the resistance of the living animal against the toxins, the phenomena are more complicated than are those observed _in vitro_. The example of the blood of the crayfish which prevents the poisoning of the mouse, without having any influence on that of the crayfish itself, may here serve as a guide to us. It is possible that, as in the struggle against the micro-organisms, we have here a co-operation of two substances, each one of which, by itself, is inactive. One of these substances would be found pre-existent in the blood of the crayfish, the other forming part of the organism of the mouse. Perhaps the action of this blood is even more complicated and only becomes active through the mediation of some constituent of the living cell.

Our study of immunity against toxins long ago revealed cases in which this resistance cannot be attributed simply to the antitoxic action of the body fluids. Animals vaccinated against living micro-organisms may succumb to infection in spite of the presence of a strong anti-infective power of the body fluids; similarly animals immunised against toxins may die from intoxication in spite of the antitoxins contained in their fluids. Facts of this order are not rare. Roux and Vaillard[587] on several occasions observed animals which died from tetanus although they had a large supply of antitoxin in their blood. Von Behring[588] and his collaborators, Knorr, Ransom, and Kitashima, also collected a large number of analogous facts. They showed that horses that have been treated for a long time with tetanus toxin and whose blood serum is very antitoxic, still experience marked illness after fresh injections of toxin and may even succumb, in spite of the presence of a large amount of antitoxin in their blood. In these cases the morbid phenomena are undoubtedly different from those typical of tetanus. Instead of the muscular contractions which characterise this disease, the above observers noted disturbance in the regulation of the body temperature, exudative inflammation around the point of inoculation, impairment of appetite and fall of body weight. Sometimes they observed muscular tremors and marked feebleness in the movements. These symptoms differing from those of typical tetanus, it may be asked whether this poisoning is not due to special substances other than tetanus toxin in the fluids injected. Von Behring does not think that this is the case, for he found that by adding antitetanus serum the formation of exudations at the seat of inoculation was suppressed. These exudations, then, must be attributed to the tetanus toxin.

[Sidenote: [387]]

In the cases where animals immunised against diphtheria toxin fall ill and even die as the result of fresh injections of toxin, in spite of the presence of a large quantity of antitoxin in their blood, we might also cast doubts on the diphtheritic character of the poisoning, because the clinical picture of this poisoning is not a very typical one. At the Pasteur Institute, where a large supply of antidiphtheria serum is prepared, we see, from time to time, horses, which have long been undergoing the process of immunisation and are furnishing a very good serum, suddenly fall ill and die from intoxication, without presenting any symptom of infective disease. On one occasion, there was actually quite a small epidemic of fatal poisonings as the result of the injection of a quantity of diphtheria toxin not exceeding the doses which had been well borne previously. Amongst the horses, inoculated with the same toxin, five of the best furnishers of serum died. The others, some of which were producing only a weak serum, remained unaffected.

Von Behring and Kitashima[589] have given a detailed history of a young horse which had become very susceptible as the result of vaccination with diphtheria toxin. It finally succumbed to the intoxication in spite of the presence of diphtheria antitoxin in its blood.

If, in these examples, we have any reason to doubt the specific nature of the intoxication, all doubt must give way before the case described by Brieger[590]. One of his goats, well immunised with tetanus toxin, which, for months, had furnished a good serum and even an antitetanus milk, after an injection, stronger than the preceding ones, was seized with tetanic contractions. These, becoming general, brought about the death of the animal with the symptoms of classic tetanus. The blood, drawn off after death, exhibited strong antitoxic power.

[Sidenote: [388]]

As the result of these observations von Behring formulated the theory of a hypersusceptibility acquired during immunisation. “Paradoxical as it may appear,” he writes[591], “there can no longer exist any doubt that horses which have acquired a high immunity as the result of treatment with tetanus toxin, present a histogenic hypersusceptibility of the organs which react against the tetanus toxin.” In support of this thesis von Behring compares the effect produced by this toxin on horses immunised with this same poison and on normal horses treated with antitoxic serum from other horses. The former, in spite of the fact that they contain in their blood 1,500 times more antitoxin than do the latter, are, nevertheless, less refractory to tetanus toxin. This feeble resistance is due, in von Behring’s opinion, to the much greater susceptibility of the living elements in the horses treated with repeated doses of the poison.

Von Behring’s theory of this form of acquired specific hypersusceptibility has been confirmed by several well-observed facts. These show that, in the animal subjected to treatment by toxins, phenomena of very diverse order are evolved simultaneously: on the one hand, cell reactions which bring about the production of antitoxins; on the other, an increase in the susceptibility of some of the living elements to the specific poison. We are, however, justified in asking if the great difference between the immunity of animals treated with toxin, and that of others treated with antitoxic serum, can be altogether attributed to this hypersusceptibility?

Let us examine in a little more detail some examples of this hypersusceptibility. We know that the guinea-pig is characterised by its great natural susceptibility to the toxins of tetanus and diphtheria. Small doses of these poisons are quite sufficient to produce in it a fatal intoxication. But it is possible to diminish greatly this feeble resistance of the guinea-pig by frequent injections of very small quantities of toxin. Knorr[592] increased their susceptibility to tetanus toxin by daily injections of one-tenth of a minimal lethal dose. The animals died before they had received the ten tenths of this dose. The hypersusceptibility produced under these conditions might be so great that one-fiftieth of the minimal lethal dose was capable of causing death. From these facts we can understand the great difficulty experienced in the earlier attempts to vaccinate guinea-pigs by means of unmodified toxin.

Von Behring and Kitashima[593] made analogous researches on the susceptibility of guinea-pigs to diphtheria toxin. By frequent injections of very small doses of this poison they succeeded in killing these animals with ¹⁄₄₀₀ of the minimal lethal dose _distributed over several injections_. They never succeeded in vaccinating guinea-pigs with increasing doses of pure diphtheria toxin. Their animals died even when they commenced with one-millionth of the minimal lethal dose.

[Sidenote: [389]]

Here, then, we have examples of the greatest hypersusceptibility that it is possible to observe. When we compare it with the changes in the antitoxic power of the blood, we find that these are even more marked. Thus, Salomonsen and Madsen’s horse, to which we have already referred, presented extraordinary oscillations in this power. After receiving, during the course of immunisation, a fresh dose of diphtheria toxin, the antitoxic value of its blood suddenly fell more than one-third (35%). In order to neutralise, completely, this dose of toxin, when injected into a normal animal mixed with antitoxic serum from this same horse, a very small quantity of the blood of the latter would have been sufficient. The injection into the immunised horse should have passed unperceived, as this animal contained in its body more than 50 litres of strongly antitoxic blood. Nevertheless the antitoxic power of this blood fell 12,000 times more than it ought to have fallen according to the calculation made upon the data just indicated. This fall is incomparably greater than the increase of susceptibility to toxin in the most significant examples reproduced above.

[Sidenote: [390]]

As the fact above cited is not at all unique, it is probable that the phenomena which appear in the animal subjected to vaccination by toxins, must be much more complicated than is usually supposed. If the fresh injections of these poisons bring about a specific hypersensitiveness on the one hand, and on the other a great fall in antitoxic power, followed by its still more notable augmentation, it is evident that the introduction of toxins must give rise to a great perturbation in the cell functions. The general analogy between acquired immunity against micro-organisms and against toxins probably rests on similar bases. Kretz[594] has already advanced the hypothesis that, in antitoxic action, two factors, comparable to the cytases and fixatives in the antimicrobial action, co-operate. In the absence of one of these elements we can understand that the one which remains may be incapable of bringing about the neutralisation of the toxin. For this reason the antitoxic serum may act very differently in the organism of the animal which produces it and in that of a normal animal which receives it. An explanation which is adequate for the antitoxic action of the blood of the crayfish injected into mice serves equally well in the case of the antitoxic influence of the serums of animals which themselves succumb to intoxication.

Wassermann’s[595] experiments on the anticytase serums might appear to supply an argument against the hypothesis we are defending. Having shown that animals injected with antityphoid serum die of intoxication when serum which prevents the action of the cytases is introduced simultaneously, Wassermann put the question: May not the action of the antitoxins be prevented by this same anticytase serum? To solve this point he injected into guinea-pigs a mixture of antidiphtheria serum with toxin in excess and a fairly strong dose (3 c.c.) of anticytase serum, upon which we have already spoken (see Chapter VII). The animals, so treated, behaved exactly as did the animals used for control which received the same quantities of antitoxin and toxin but without the addition of anticytase serum. Wassermann concludes from these experiments that the exclusion of the cytase, contrary to what takes place with antimicrobial serums, in no way impedes the action of the antitoxins. This conclusion, which appears at first sight to be justified, cannot, however, be accepted, as the two examples chosen by Wassermann, typhoid infection and diphtheria intoxication, differ very profoundly from each other. In the former, we have an experimental typhoid peritonitis which kills the control animals in less than 24 hours, whilst the second is diphtheria in which the controls do not succumb until the sixth day after injection. The effect of the anticytase serum being only very transitory, it is quite natural that this should manifest itself in an infection of short duration and should not do so in a slow intoxication. Besides, Wassermann himself has shown that in several other cases of immunity against micro-organisms (the bacilli of influenza and of leprosy) the injection of his anticytase serums does not interfere with the perfect resistance of the animals. But even were it demonstrated that the cytases really play no part in immunity against toxins, the intervention of some other similar factor could always be evoked.

[Sidenote: [391]]

The analogy between immunity against micro-organisms and that against toxins may facilitate the study of the relations between the latter and the antitoxic power of the body fluids. In the preceding chapters we have described examples in which animals possess a protective power in their blood but are not refractory to the corresponding infection; on the other hand, we have cited cases in which acquired antimicrobial immunity exists without the blood presenting any appreciable protective power. The idea of measuring acquired immunity against micro-organisms by the measurement of the protective or agglutinative power of the blood must therefore be abandoned, and it is impossible to regard immunity against toxins as a function of the antitoxic property of the body fluids. As we have seen, animals completely refractory to tetanus, such as the cayman, whose immunity does not depend on the antitetanic power of the blood, develop antitoxin after the injection of toxin. A similar state of affairs, but less pronounced, has been demonstrated by Vaillard as occurring in the fowl. The fowl, in spite of its very marked natural immunity against tetanus, produces antitetanin as the result of the introduction into its body of tetanus toxin; the rabbit, on the other hand, a susceptible animal, may acquire a real immunity without the development of any antitoxic power in its fluids. An additional fact was noted by Vaillard[596]. He showed that the repeated inoculation of tetanus spores along with a small quantity of lactic acid, made below the skin of the tail of rabbits procured for them an immunity against tetanus toxin, although no antitoxic property appeared in their blood. In his experiments, one hundred volumes of blood serum were found to be incapable of neutralising a single minimal lethal dose of the toxin. The rabbit, however, still remains quite capable of developing antitetanic power in its fluids. All that is necessary is to inject into it some tetanus toxin heated to 60° C. or treated with Lugol’s iodo-ioduretted solution. As the outcome of his researches Vaillard concludes that the antitoxic property of the body fluids “is not sufficient ... for the general interpretation of acquired immunity, as it cannot be demonstrated in all animals which have become refractory.”

[Sidenote: [392]]

[Sidenote: [393]]

The facts I have just mentioned were demonstrated early in our study of the antitoxic power of the animal organism. Since then a large number of analogous data have been collected. Recently, von Behring and Kitashima[597] have had to abandon the immunisation of monkeys against diphtheria toxin because of the poor yield in antitoxin which they obtained. The blood of one of their monkeys that had acquired a resisting power against very large doses of diphtheria toxin showed only a very moderate antitoxic power. In establishments where antitoxic serums are prepared on a large scale the workers have become convinced that the yield of antitoxin has no direct constant ratio to the immunity of the animal. This has been demonstrated repeatedly at the stables of the Pasteur Institute. Thus, of two horses, treated at the same time and in exactly the same way with diphtheria toxin, one furnished a very good antitoxic serum which was maintained at 200 units Ehrlich, rising up to 400 units, whilst the other never reached 150 units[598]. And yet both these animals possessed the same immunity against diphtheria toxin. They tolerate considerable doses of toxin and react merely by a slight or insignificant rise in temperature. In another series of horses, which have been immunised for nearly seven years, one remained capable of yielding a large quantity of antitoxin, seeing that the value of its serum oscillated between 200 and 300 units. After five years of this state of things the antitoxic power began to fall considerably, without, however, any corresponding loss of immunity. Indeed, an injection of 250 c.c. of toxin (of which 0·002 c.c. was sufficient to kill a guinea-pig) began, at the commencement of the present year, to be borne without the least febrile reaction. An attempt was made to raise the antitoxic power of the blood by making intravenous injections of toxin and of diphtheria culture, but in vain. The yield of antitoxin continued to fall and it became necessary to employ this horse for another purpose than the preparation of antidiphtheria serum. This is by no means an isolated example. Of a large number of treated horses it frequently happens that certain individuals, without being particularly susceptible to a given toxin, are found to be incapable of producing any corresponding antitoxin[599].

In presence of the fact that animals very resistant to toxins may possess no, or only an insignificant antitoxic power in their fluids, and that, on the other hand, animals in which this property is highly developed may succumb to intoxication, it may be readily understood that immunity against toxins and the antitoxic power of the body fluids may be two distinct conditions. Von Behring has clearly demonstrated the fact of the cellular hypersensitiveness of the animal immunised against the corresponding toxin and has laid great stress upon this fact. He came[600] to the conclusion that “the immunity of the tissues and the production of antitoxin follow a parallel course in their development so slightly that, in spite of an abundant accumulation of antitoxin, the susceptibility of the elements of the tissues may increase in an extraordinary fashion.” If, during the course of immunisation, this susceptibility can increase so greatly, it is probable _à priori_ that under certain circumstances it might also diminish notably. After demonstrating “that in time the antitoxin disappears from the blood of animals immunised with toxins without any consequent disappearance of immunity,” von Behring formulated the conclusion that in these animals “the living elements of the animal, which were previously susceptible to the poisons, have acquired an insusceptibility towards the same substances.” This result fully accords with the facts of the change of the negative chemiotaxis of phagocytes into positive chemiotaxis for micro-organisms during the acquisition of anti-infective immunity.

[Sidenote: [394]]

Later, von Behring[601] changed his opinion. Whilst still accepting the change of cellular susceptibility in the direction of hypersensitiveness in animals immunised against toxins, he refused to admit the change in the opposite direction. The cells, according to him, never lose any of their susceptibility, so that acquired immunity against toxins cannot be obtained otherwise than by means of antitoxins capable of neutralising the poison in a susceptible or hypersusceptible animal. This new theory von Behring upheld in several papers and it is met with in his most recent publications. Nevertheless, certain well-established facts compel us to accept an immunity against toxins as coming about as the result of a diminution of the susceptibility of the vaccinated animal. Parallel with his researches on the increase of the susceptibility of guinea-pigs to tetanus toxin, researches discussed above, Knorr[602] describes analogous experiments on rabbits. When these animals are injected with fractions of the minimal lethal dose, frequently repeated, the rabbit not only does not become hypersusceptible to tetanus but exhibits a greater and greater insusceptibility. Whilst guinea-pigs, treated according to this method, die from tetanus before they have reached the minimal lethal dose, rabbits, as the result of frequent injections of small quantities of tetanus toxin, become capable of resisting five times the lethal dose (for normal rabbits) without exhibiting the slightest symptom of illness. Against the attribution of this result to the acquired insusceptibility of the living animals it might be objected that the immunity, in this case, may depend on the antitoxic power of the fluids of the body, developed with great rapidity. Such an objection cannot be raised in the case of horses which become insusceptible to toxins after a long period of vaccination. The horse whose history was given above, when discussing the diminution of antitoxic power, may serve as an example. At the commencement of its vaccinal period, in 1894, it reacted to the injection of 10 c.c. of diphtheria toxin by a rise of temperature of 1° C. Four years later, when its blood had become very antitoxic (350 units per c.c.), it was necessary to inject 350 c.c. of toxin to obtain the same rise of temperature. Quite recently, having now lost the greater part of its humoral antitoxic power, this horse exhibited no rise of temperature after an injection of 250 c.c. of strong diphtheria toxin. The diminution of the specific susceptibility is produced in this case in a most marked fashion; it is not therefore to the antitoxic property of the body fluids that this case of immunity must be attributed.

[Sidenote: [395]]

The insusceptibility acquired against poisons of different kinds is observed also in cases where the adaptation is not accompanied by the production of humoral antitoxic properties, as in the immunity of frogs against abrin. This form of immunity may be traced through the organic series down to such lowly developed organisms as the plasmodium of the Myxomycetes, which as we have seen readily becomes adapted to different poisons (see Chapter II).

It can be clearly seen, then, that immunity against toxic substances is a very complex phenomenon which it is impossible to reduce simply to an antitoxic function of the fluids of the body. For this reason we cannot accept a theory which would confine this kind of immunity within the narrow limits of a simple reaction between two substances, a reaction quite comparable to that observed in a test-tube. Attempts have been made to determine with almost mathematical precision the conditions under which it is possible to communicate to the animal a resistance against microbial toxins and formulae have been constructed to define these conditions. But the application of these formulae has been found to be a much more difficult matter. In Prussia, with the sanction of the Government, regulations have been enacted as to the procedure to be followed in the testing of antitoxic serums, and a paragraph has been added which requires a post-mortem examination of the guinea-pigs employed for this purpose in the case of diphtheria antitoxin. “The dead animals,” says this instruction, “must be submitted to a post-mortem examination, and special attention must be directed to the presence of any pre-existing diseases (tuberculosis, pseudotuberculosis, pneumonia) which may have induced hypersusceptibility in the animals under experiment.” Do we not see in this a proof of the important intervention of the organism of the living animal which may modify the results of calculations based upon too rigorous formulae? It must not be forgotten, too, that in addition to the three diseases named in the instructions, we have a number of other factors which may influence the receptivity and the resistance of animals. We have already cited Roux and Vaillard’s experiments which demonstrated that even animals which have been previously subjected to vaccinal inoculations against certain micro-organisms, exhibit a hypersusceptibility to mixtures of toxins with antitoxins.

In view, then, of this complexity of the phenomena of acquired immunity against toxins, it would be very important could we learn something of the nature and origin of antitoxins. Unfortunately, as we shall see, these questions are, as yet, far from having received a satisfactory solution.

[Sidenote: [396]]

Struck by the fact that antitoxins exert a specific action on the toxin which has been employed in the treatment of the animals that produce the serum, certain observers have sought an explanation on the hypothesis of a transformation of toxin into antitoxin. We have already seen that antitoxic action is not always absolutely specific; we have serums which prevent intoxication by various kinds of poisons, e.g. antitetanus serum, which is active against both tetanus toxin and snake venom. There is, however, a great quantitative difference between the influence of the antitoxin on the toxin with which the animals have been prepared and on a different poison. Thus, in the example just cited, in order to neutralise snake venom it is necessary to use a much larger quantity of antitetanus serum than against the toxin of tetanus. The classical example of the specific influence of antitoxins is the absolute inactivity of antidiphtheria serum against tetanus and the same non-effect of antitetanus serum against diphtheria intoxication. The most simple explanation of this specificity of action appeared to be the supposition that each antitoxin contains a part of the corresponding toxin, modified by the organism of the animal. H. Buchner[603] advocates this hypothesis. I myself[604] said “that it is probable that antitoxins, at least in great part, represent a modification of the toxins prepared by certain cells in the animal body; this product is then poured into the blood.” This view was stated as a “probability” and consequently contains no affirmation in the least definitive. I was, therefore, quite prepared to give it up under the weight of the crushing criticism formulated by several very distinguished observers. It was objected; first, that antitoxin is produced by animals in very great disproportion to the quantity of toxin they have received; secondly, that the animals which receive an injection of antitoxin eliminate it from their body much more rapidly than do those which prepare it in their own body; thirdly, that antitoxins are sometimes found in the blood of healthy animals, who have had no attack of the disease nor any injection of the specific toxin. Let us examine these objections more closely, objections all based on well-established facts.

[Sidenote: [397]]

It has been shown that the antitoxin produced by the animal is sufficient to neutralise a quantity of toxin much greater than that which was injected into the animals supplying the antitoxic serum. Knorr[605], from his experiments, calculated that a horse reacts to one unit of toxin by the production of 100,000 units of antitoxin. This statement certainly does not allow us to affirm that all the antitoxin corresponds to toxin, but it does not eliminate the possibility that toxin, subjected to the influence of the cells of the animal body, may be found, in a modified form, in the product of these elements. This hypothesis would be quite sufficient to explain the very remarkable specificity of antitoxins.

[Sidenote: [398]]

If the toxin, in order to be modified by the living cells, must be subjected to some special action on the part of the latter, we can readily understand that this process must demand a greater or less length of time; this would lead to a much slower elimination of the antitoxin than in the case where it had been injected, ready prepared, into a normal animal. From the commencement of his researches on immunity against poisons, Ehrlich[606] distinguishes two kinds of this immunity, an _active immunity_ which is obtained as the result of the introduction of toxins into the animal, and a _passive immunity_, another form of the refractory condition which is set up by the injection of antitoxic serum formed in the actively immunised animal. Von Behring[607] applies the term _isopathic immunity_ to active immunity, and to passive immunity that of _antitoxic immunity_. It is generally admitted that the first kind of immunity is more slowly acquired, but that it persists for a much longer period than the second (passive or antitoxic immunity) which is acquired immediately after the introduction of the antitoxin, but which, on the other hand, lasts for a short time only. This view is supported by numerous observations on the very rapid disappearance of the refractory condition. According to von Behring the great difference in the duration of the isopathic and antitoxic immunities is only an apparent one. It is due to the fact that antitoxins are usually introduced along with the serum of different species which sets up a strong reaction and is rapidly eliminated from the animal. Thus the antitoxic serum of the horse is usually injected into small animals such as guinea-pigs, rabbits, and mice. When, however, von Behring injected horses with antitoxic serums from other horses, the antitoxic immunity lasted almost as long as in animals vaccinated with toxins. Ransom[608] has developed this thesis in a work carried out in von Behring’s Institute at Marburg, and supports it by comparative researches which demonstrate the more rapid disappearance of the antitoxin when introduced with the serum of a different species than when introduced with that of the same species.

Even should we accept the current view on the greater duration of the antitoxic power of the blood in isopathic immunity, the hypothesis of the transformation of toxin by the cells of the animal is not necessarily invalidated. If a part of the toxin introduced into the animal remains stored for some time in an organ it is evident that only gradually can it be subjected to the transforming action of the cells. It is impossible, in the present state of our knowledge, to demonstrate this proposition, but we may invoke in its favour the prolonged persistence of red blood corpuscles when introduced into the body of a different species of animal (see Chapter IV). These corpuscles are in the end always completely digested but the process is of long duration.

[Sidenote: [399]]

The same hypothesis will also explain a fact, first demonstrated by Roux and Vaillard[609]. They have shown that after repeated bleedings of rabbits immunised against tetanus, the antitoxic property of the blood was soon raised to almost the same value as before. Salomonsen and Madsen[610] have confirmed the fact of the regeneration of antitoxin after the bleeding of their animals (horses and goats) immunised against diphtheria. Those authors who do not accept the possibility of the transformation of toxins in the production of antitoxins, regard these facts as absolutely incompatible with the hypothesis which they attack. Thus, Weigert[611] considers that the regeneration of antitoxin after bleeding can only be understood by accepting that antitoxin, like the blood, may be reproduced in the actively immunised animal without any fresh introduction of toxin. It is, however, just as simple, we think, to explain the fact in question by the hypothesis of a provision of toxin stored up in certain cells. This also is sufficient explanation of another observation made by Salomonsen and Madsen[612], who showed that pilocarpin is capable of augmenting the production of antitoxin. Since it is the living cells which transform the toxin and excrete the antitoxin, it is quite natural to suppose that every factor which stimulates cell function may be capable of causing an increase of the product transformed by the cells.

The third argument invoked against the possibility of the transformation of toxins into antitoxins is based on the fact that the serum of normal horses has sometimes a certain degree of antitoxic power against diphtheria toxin. The horses have never suffered from diphtheria, therefore the antidiphtherin of their blood has nothing to do with diphtheria toxin. It is not known why the blood serum of certain untreated horses is from the first active against diphtheria toxin, whilst that of others exerts absolutely no action on the same poison. We know only that this property is far from being constant in the equine species. Perhaps it is acquired as the result of the penetration into the animal of some pseudo-diphtheria bacillus, whose frequency and number are very great. In order that the microbial products may give rise to the formation of antibodies, it is not at all necessary that the micro-organisms should produce an evident disease. Thus, to cite one example only, Foerster[613] observed a considerable agglutinative power against the typhoid cocco-bacillus in the serum of a child which was found living among a family of typhoid patients but which, itself, presented no morbid symptom.

The criticism, directed against the hypothesis that modified toxin enters into the production of antitoxin, may not be sufficient to show the incorrectness of this view; it does not follow, however, that the view is right. In the present state of our knowledge it is impossible to solve the problem definitely, and as the hypothesis of transformation gives us the best idea of the specificity of the action of antitoxins, it has a right to be taken into consideration as much as any other.

[Sidenote: [400]]

Ehrlich[614] has formulated another hypothesis to explain not only this specificity but the origin of antitoxins in general. This is the ingenious hypothesis of side-chains or of receptors, which has already been considered in other chapters of this work. It is now for the first time brought forward in relation to the antitoxins properly so-called, that is to say substances capable of preventing intoxication by microbial toxins. In order to make his hypothesis as clear as possible Ehrlich begins by explaining its bearing on the concrete example of tetanus antitoxin. “When we introduce into an animal a small quantity of tetanus toxin, it is easy to obtain exact proof that it is quickly fixed by the central nervous system, probably by the motor cells of the ganglia; that the central nervous system more than any other organ attracts the tetanus toxin and retains its toxic molecules very firmly.” There we have the side-chains of the protoplasm fulfilling this rôle and subjecting the living protoplasm to the prolonged action of the poison. Once it is combined, the side-chain becomes incapable of fulfilling its normal function, and there is induced on the part of the living elements the production of new chains of a similar character. Following the law that the reaction is stronger than the action, there is an over-production of these side-chains which finally so embarrass the cell which has developed them that they are excreted by it into the blood plasma. Once expelled into this plasma, they continue to manifest their affinity for the tetanus toxin, an affinity which must be even greater in the case where the chains are found in the blood than when they were connected with the cell. Owing to this affinity, these chains, now in the blood, fix the tetanus poison introduced into the animal and prevent it from reaching the susceptible nerve elements. Antitoxins, according to this hypothesis are, therefore, nothing but overplus side-chains poured into the body fluids. Ehrlich extends his theory to a whole series of bodies capable of causing the formation of antitoxins and antidiastases. “It is probable,” he says, “that all analogous bodies can only become toxic to the animal on condition that the animal is capable of fixing their toxophore groups in certain of the organs that are important for its life” (p. 17).

[Sidenote: [401]]

According to this theory tetanus antitoxin must pre-exist in the central nervous system of the normal animal. In the immunised animal, the side-chains must be reproduced in very great quantity in the nerve cells and pass thence into the circulation. Indeed, Wassermann, a supporter of this theory, made a search for tetanus antitoxin in the nerve centres of normal animals. In collaboration with Takaki[615] he made the important discovery that the brain and spinal cord of small mammals (guinea-pigs and rabbits) when triturated with tetanus toxin prevent the manifestation of its toxic action in animals most susceptible to tetanus. The brain was always found to be more active than the spinal cord. The property of neutralising the toxin of tetanus belongs to the solid parts of the nerve centres; the fluid of the cerebral emulsion is incapable of exercising this action.

This discovery was soon confirmed. Ransom[616] demonstrated it almost at the same time, and independently of Wassermann and Takaki; and the fact is indisputable. It remains to be seen whether the “antitoxin” of the nerve centres of normal animals is really the same as that which is found in the fluids of animals immunised against tetanus toxin, as is accepted by Wassermann and the other partisans of the side-chain theory. The former is characterised by a very local reaction; it is incapable of being dissolved and distributed through the body of the animal. This is shown by Marie’s[617] experiments, and my own[618], all carried out in my laboratory. All that is necessary is to introduce, beneath the dorsal surface of the thigh of a guinea-pig, a quantity of the cerebral substance sufficient to neutralise several times the lethal dose of toxin, and below the skin of the ventral aspect of the same thigh, a lethal dose of this toxin, when it will be found that the guinea-pig contracts a fatal tetanus. The antitoxic action of the nerve substance extends, therefore, for a short distance only; it is strictly local.

[Sidenote: [402]]

The view that the action of the substance of the pounded nerve centres is different from the neutralisation of the toxin by the antitoxin of the body fluids is further confirmed by the fact that the fixation of the tetanus poison by the cerebral substance is very transient. We have shown that a mixture of toxin and pounded cerebral substance, that does not produce any tetanic symptom when injected into the peritoneal cavity of guinea-pigs, sets up a grave tetanus when it is injected subcutaneously into the thigh. In the latter case the toxin becomes separated from the particles of the cerebral substance that had fixed it. Danysz[619] convinced himself that the mixture of pounded brain with tetanus toxin when it is left in physiological saline solution, in distilled water, or in a 10% solution of sea salt, allows the tetanus toxin to pass into the macerating fluid. The fixation of the toxin to the cerebral substance is, therefore, more comparable to the mordanting of colouring matters by the tissues than to a real combination.

Observers who have repeated the experiments of Wassermann and Takaki have been greatly struck by the difference between the action of the pounded cerebral substance and that of the living brain upon the tetanus toxin. Whereas the former, taken from the guinea-pig, an animal very susceptible to tetanus, prevented intoxication when employed in minimal dose, the living brain of the same species was found to be incapable of neutralising the most minute quantities of toxin. On the other hand, Roux and Borrel[620] have shown that the brain of rabbits, whether untreated or vaccinated against tetanus, was very susceptible to the action of the tetanus toxin. This toxin, injected directly into the brain, set up in both groups of rabbits a special and characteristic cerebral tetanus. On the other hand, when a little of the cerebral substance of the rabbits, mixed _in vitro_ with tetanus toxin, was injected into other susceptible animals, these remained unaffected.

This great difference between the antitoxic action of the living brain and that of the pounded cerebral matter, on the one hand, and the rigorous localisation of the antitetanic influence of this cerebral substance, on the other, have suggested to several observers the idea that the brain cannot be regarded as the organ of formation of the true antitoxin, such as is found in the fluids of immunised animals. This view has been expressed by Roux and Borrel, Marie and ourselves. Knorr[621] also shares this view, being struck by the fact that rabbits attacked by tetanus remain for weeks with contractions, but are incapable of producing in their nerve-cells sufficient antitoxin to disintoxicate them, although their blood is already loaded with dissolved antitoxin.

[Sidenote: [403]]

At this period it was generally supposed that, in accordance with Ehrlich’s theory, the hypothetical side-chains were capable, under certain conditions, not only of fixing the tetanus toxin, but also of neutralising it. It was said, therefore, that these chains, reproduced in large quantities in the cerebral cells, must exercise their neutralising action in the brain itself. Consequently, when it was seen that, in Roux and Borrel’s experiments on vaccinated rabbits, this organ was itself affected, it was concluded that the brain must not be regarded as the producer of the antitoxin.

Later, Ehrlich and his supporters, amongst whom I will name especially Weigert, have developed the theory of side-chains in a much more detailed fashion, leading to a different interpretation of several facts previously established. Ehrlich distinguishes in the toxin molecule a _haptophore group_ which combines with the side-chain or the corresponding receptor of the living elements, and a _toxophore group_ which produces the poisoning of the protoplasm. The side-chains, inactive for the toxophore group, neutralise only the haptophore group. Consequently, when these side-chains are numerous in the nerve elements which produce them, they may be a source of great danger to this living element, by attracting the toxic molecules. In this case, these chains, or receptors, serve to attract the poison, just as the badly adjusted lightning-conductor attracts lightning. For this reason rabbits vaccinated against tetanus become tetanic when the toxin is injected directly into the brain. It is only at a distance from the nerve centres that the receptors, excreted into the body fluids, fulfil their rôle of true antitoxins. There they combine with the haptophore group of the toxic molecule, leaving the toxophore group intact; this latter group, however, diverted from the nerve-cells, is incapable of exercising an injurious action.

[Sidenote: [404]]

From this point of view not only the cerebral tetanus of vaccinated rabbits, but also the hypersusceptibility of immunised animals, upon which von Behring has so strongly insisted, may be explained. The argument, drawn from these facts, against the nervous origin of tetanus antitoxin, loses, therefore, much of its weight. If we confront this hypothesis with the other data collected on the question, the solution of the problem becomes beset with great difficulties. Previous to the discovery made by Wassermann and Takaki, I attempted to solve the problem by removing from fowls portions of the brain and spinal cord, proposing to take advantage of the fact that birds, which are capable of producing antitoxins, withstand these operations fairly well. My hopes were not fulfilled; I could never keep my fowls alive long enough to complete the experiment. We must, therefore, for the present, be content with indirect arguments. If the nerve centres do really produce the tetanus antitoxin and excrete it into the blood, we ought at a given moment to find in these organs a greater quantity of this substance than in the blood and the other organs. The reader will recall the researches of Pfeiffer and Marx, and of Deutsch, who demonstrated the possession of a greater richness in protective substance by the phagocytic organs of animals, treated with micro-organisms, than by the blood serum. The same result might be obtained by a comparative investigation of the tetanus antitoxin in the nerve centres and the blood of animals immunised against tetanus. My experiments directed to this point have not been favourable to the hypothesis of the nervous origin of tetanus antitoxin.

In fowls, killed as soon as tetanus antitoxin began to appear in the blood, the brain and spinal cord did not exhibit the slightest antitoxic power[622]. We might be tempted to explain this result as due to an accumulation of toxin in the nerve centres which would prevent the manifestation of the antitoxin. For this reason, in my later researches[623], I made use of animals that had been long immunised, but whose blood was still antitoxic. I killed a fowl which had not received any toxin for about eight months, and a guinea-pig into which the last toxic injection had been made almost two years before the date of this experiment. After removing a portion of the brain the blood of these two animals was found to be more antitoxic than before the operation, which indicated that the source of the antitoxin was as yet uninjured. To ascertain whether this source was to be found in the nerve centres I made a comparative determination of the antitoxic power of the brain, of the spinal cord and also of several other organs, of the blood and of the exudations. The result was still negative. The nerve centres were found to be less antitoxic than the blood and other fluids of the body, and even less active than such organs as the liver and kidneys.

[Sidenote: [405]]

[Sidenote: [406]]

In support of the hypothesis of the nervous origin of tetanus antitoxin there remains, then, only the fact of the retarding action of the cerebral substance upon tetanus. In the absence of other arguments this assumes a preponderating importance. We have seen that this action is based on a fleeting and not very firm fixation of the toxin by certain parts of the brain and the cord. Are we justified in regarding this as comparable to the more stable fixation observed in living animals susceptible to tetanus intoxication? Soon after Wassermann and Takaki’s discovery I pointed out that the pounded brain of frogs mixed with tetanus toxin does not prevent animals, into which this mixture is injected, from contracting fatal tetanus. This observation was confirmed by Courmont and Doyon[624], in several series of experiments carried out under various conditions. They found that “the brain of the frog, heated or unheated, when mixed with tetanus toxin even for several hours, at the temperature of the laboratory or at 38° C., even in considerable doses, does not possess any neutralising property.” This fact would not be in any way wonderful if we had to do with an animal insusceptible to tetanus; but in the frog, as we have said in the preceding chapter, this is far from being the case. In the cold it does not readily become tetanic, but above 25°–30° C. it becomes very susceptible. The tortoise, which is very refractory to this intoxication, has a brain which, when pounded and mixed with tetanus toxin, exerts a certain preventive power over susceptible animals. Nevertheless, the brain of the living frog, as demonstrated by Morgenroth, absorbs this toxin. There is, therefore, a difference between the absorption of the tetanus poison by the living elements and by the pounded cerebral substance. A similar result is obtained with several other toxins. Diphtheria poison is very toxic when injected directly into the brain of the guinea-pig or rabbit. Even the rat, as demonstrated by Roux and Borrel[625], is readily affected by this toxin under these conditions. Doses which when inoculated subcutaneously are well borne by the rat, when introduced into the brain set up a fatal intoxication in this animal. And yet the brain, when pounded and mixed with diphtheria toxin, can never protect susceptible animals from intoxication. Numerous attempts to reproduce Wassermann and Takaki’s experiment with the diphtheria poison have always been unsuccessful. Attempts to obtain the same result with snake venom have also given negative results. Calmette[626] made several experiments with emulsions of rabbit’s brain and snake venom with the object of ascertaining whether the elements of the nervous system possess against venom the same properties as against tetanus toxin. “None of these emulsions”—concludes Calmette—“exhibited either the slightest protective or antitoxic power _in vitro_. There is, therefore, no analogy of action between what takes place in the nerve elements against tetanus toxin and against venom.” Nevertheless venom, like diphtheria toxin and tetanus toxin in the frog, exerts an undoubted action on the nerve centres.

Again, the protective fixation of poisons to the cerebral substance is not the exclusive privilege of tetanus toxin. Kempner and Schepilewsky[627] obtained the same result with the toxin of botulism (produced by van Ermenghem’s anaerobic micro-organism which sets up intoxication of intestinal origin in certain cases of poisoning by food). The brain and spinal cord of the guinea-pig, when triturated with physiological salt solution and mixed with botulinic toxin, prevents intoxication in susceptible animals, exactly as in Wassermann and Takaki’s experiments with tetanus.

When Kempner and Schepilewsky wished to obtain some idea as to the substance or substances in the nerve centres which fix the toxin of botulism and thus prevent poisoning, they found that lecithin and cholesterin, mixed with this toxin or injected separately and simultaneously, protected mice just as completely as did the cerebral substance. On the other hand, they found a difference as regards the two substances when injected before the toxin was introduced; they were then unable to prevent poisoning, though the cerebral substance exerted an undoubted protective influence. Kempner and Schepilewsky also showed that heating altered the preventive action of lecithin and cholesterin less than it did that of cerebral emulsion.

[Sidenote: [407]]

These observers extended their researches to the protective action of fats and demonstrated that olive oil when emulsified and neutralised with soda and mixed with twice and even four times the lethal dose of botulinic toxin, prevented the contraction of a fatal poisoning by mice. Tyrosin also protected mice against this intoxication, not only when injected simultaneously with the poison, but even when introduced into the animal 24 hours before the poison was administered. Kempner and Schepilewsky conclude “that not only with the substance of the nerve centres, but also with various other substances, they were able to obtain a certain protective effect against the toxin of botulism” (p. 221). Their experiments with cholesterin and tyrosin were suggested to them by the previous researches of Phisalix[628] who demonstrated that the bile salts, as well as the two substances I have just mentioned, would protect animals against the venom of the viper.

Bearing all these facts in mind, it appears to be probable that in the above cases it is principally the fatty matters of the nerve centres that temporarily fix these toxins, and allow the animal organism to divert the poisons from their morbific action. From this point of view, it is interesting to note that the toxic action of the tetanus poison can also be prevented by other substances than the emulsion of the nerve centres. Thus Stoudensky[629] demonstrated, in an investigation carried out in Roux’s laboratory, that carmine fixes the tetanus toxin and prevents its action on the guinea-pig. As in the case of the cerebral substance, this fixation by carmine is very unstable. When the carmine that has fixed the tetanotoxin is macerated in distilled water it gives up the poison to the water which is then capable of producing tetanus. Such fixation does not end, any more than in the case of the cerebral substance, in the destruction or disappearance of the toxin. Carmine if first dissolved or macerated in water (especially if heated) loses its fixative power and can no longer prevent tetanus poisoning. Sterilisation, at 120°, 100° and even at 60° C., of the carmine, suspended in physiological salt solution, caused it to lose its protective action, although dry heat applied to it in closed tubes did not destroy this power.

[Sidenote: [408]]

In many respects carmine, which is derived especially from the adipose body of the cochineal insect, exerts an antitoxic influence analogous to that of maceration with the nerve centres. If fats play a special part in this action, we can readily understand how a brain, such as that of the frog, poor in fatty matters, cannot fix the tetanus toxin and prevent its morbific action. In any case the fact that certain substances of diverse nature, acting on toxins, exert an influence similar to that of the pounded mass of the nerve centres, does not allow us to accept Wassermann and Takaki’s experiment as proving the nervous origin of tetanus antitoxin. The analogy with the facts bearing on the anticytotoxins, collected and described in the fifth chapter, also tells against this hypothesis. We would here remind the reader that the two constituent parts of the antispermotoxin, the anticytase and the antispermofixative, develop in castrated animals and are consequently produced outside the spermatozoa, elements susceptible to the spermotoxin. The facts collected concerning the antihaemotoxins indicate also that these substances have some other origin than the red blood corpuscles.

[Sidenote: [409]]

[Sidenote: [410]]

This latter supposition appears to be in contradiction to Ransom’s[630] very interesting researches on the haemolytic action of saponin, carried out in Meyer’s laboratory at Marburg. This glucoside, owing to its property of fixing itself on the stroma of these corpuscles dissolves the red corpuscles of many vertebrates. The cholesterin of this stroma combines with the saponin, as the result of which the red corpuscles become altered and allow the haemoglobin to diffuse. But this same substance, cholesterin, which causes the poison to penetrate into the red blood corpuscles, prevents the solution of these elements when they are bathed in blood-serum. This fluid, in fact, acts as the antitoxin to saponin and does so just because it contains cholesterin. The cholesterin of the serum, fixing the saponin, prevents it from affecting the red corpuscles, thus fulfilling the function of a well fitted lightning conductor. On the other hand, when the cholesterin of the stroma of these corpuscles is linked on to the saponin, it renders them the disservice of a defective lightning conductor. The accord between these facts and the postulates of Ehrlich’s theory led Ransom to suppose that in the haemolysins and antihaemolysins, cholesterin perhaps played a similar part. His experiments convinced him that this was not the case. As it is generally accepted, after Calmette’s[631] experiments and according to Ehrlich’s view, that the alkaloids and the glucosides in general are incapable of setting up the formation of antitoxins, we might regard the attempts to find an antisaponin and to settle whether it is identical with cholesterin as useless. But in regard to these delicate questions we must be careful not to give too great weight to _a priori_ arguments. It was believed until quite recently that substances with very complex molecules, such as the albuminoids, toxins and soluble ferments, must always give rise to the production of antibodies in the animal; whilst the simpler substances whose chemical nature was better defined could never lead to this. Facts acquired in recent years have led to a modification of this view. In our fifth chapter we have already spoken of the fruitless attempts of Ehrlich and Morgenroth to obtain certain antifixatives. And yet the fixatives, as is shown by the results of the researches of Bordet and myself, belong to the category of substances which are quite capable of setting up the formation of antibodies. Again, certain mineral poisons, quite unexpectedly, gave rise to the development of the counterpoison in the animal body. This fact forced itself upon Besredka[632] in his researches on the adaptation to arsenic made in my laboratory. His experiments were undertaken for the purpose of studying the mechanism of the refractory condition against a poison, apart from any antitoxic action whatever, which, according to previous investigations, seemed excluded. This action, however, was exhibited in such a degree that it could not be ignored. The serum of animals immunised against arsenious acid was found to possess both protective and antitoxic properties against a dose of this poison killing a rabbit in 48 hours. It is true that Morishima[633], in a research carried out in Heyman’s laboratory at Ghent, has thrown doubt upon these results. His objections, however, cannot refute the statements of Besredka which rest on very precise and numerous experiments which I witnessed. Morishima left out of account several important circumstances and carried out his experiments without any continuous check by means of control animals. It must be said also that the resistance of the rabbit against arsenic depends on many different factors and that, at certain seasons, it is much more difficult to adapt them to the poison than at others. It is only by numerous researches extending over a very long period that we can arrive at precise and conclusive results.

From these observations there is every inducement for us to attempt to ascertain whether, by subjecting animals to repeated injections of saponin, it is possible to augment the antisaponic power of their blood-serum and whether, if this takes place, the antitoxic action is due to a rise in the amount of cholesterin in this serum. I therefore requested Besredka to carry out some experiments bearing on this point. Guinea-pigs, injected with progressive doses of saponin for more than two months, at the end of this period showed no increase in the antisaponic power of their serum. They followed the rule established by Ehrlich; they developed no antitoxin against a glucoside. Moreover, they gave us no new information as to the origin of these antibodies.

[Sidenote: [411]]

In his first memoir in which the theory of side-chains is treated, Ehrlich insists on the nervous origin of antitetanin as an example of the production of antitoxins by animals susceptible to poisons. Now, however, that he has come to distinguish haptophore and toxophore groups in the toxic molecule, it is to the side-chain, which fixes the first group, that Ehrlich attributes prime importance. “The formation of antitoxins”—he says[634] in the opening address at his Institute at Frankfort—“would, therefore, be absolutely independent of the action of the toxophore elements.” In other words, for a cell to be capable of producing antitoxin, it is not at all necessary that it should be susceptible to the toxic influence of the poison; it is only necessary that it should possess receptors, or side-chains, capable of combining with the haptophore group of the toxin. Thus it is possible, as we have described above, to produce antitoxins, with modified toxins whose toxic action is _nil_ or almost so, but which have retained their power of combining with antitoxic substances. According to Ehrlich, these modified toxins are _toxoids_, in which the toxophore group is completely destroyed; “whilst the haptophore group, the producer of immunising substances, is retained in its integrity.” It is evident then that, under such conditions, the tetanus antitoxin might be developed elsewhere than in the nerve centres. For that it would be sufficient that outside the nerve cells there should be other living elements capable of fixing the tetanus toxin, or, to use Ehrlich’s phraseology, elements, possessing side-chains, having an affinity for the haptophore group of the tetanus poison.

Dönitz[635] has already expressed the view that in the rabbit the tetanus toxin may be fixed not only by the nerve elements but also by the various other cells.

[Sidenote: [412]]

[Sidenote: [413]]

The existence of such cells, outside the nervous system, is not merely hypothetical. It is shown very clearly in Roux and Borrel’s experiments on cerebral tetanus. In order to produce this disease in the rabbit, it is sufficient to introduce a very small dose of toxin directly into the brain. When inoculated subcutaneously with much larger quantities of the same tetanus poison, the rabbit remains in good health or exhibits merely a slight and transient tetanus. “The resistance of the rabbit against the tetanus toxin, injected under the usual conditions”—conclude Roux and Borrel[636]—“is not due, then, to a relative insusceptibility of the nerve centres, but to the fact that much of the poison introduced does not reach the nerve cells and is destroyed in some part of the animal.” In the guinea-pig, as shown by the same investigators, the difference of the dose of tetanus poison, necessary to produce fatal tetanus by intracerebral or by subcutaneous injection, is minimal or nil, from which it may be argued that in this very susceptible animal there is no destruction of toxin outside the nerve centres and that the whole of the poison introduced makes its way without hindrance as far as these organs. Ehrlich, in his report to the International Congress of Medicine in Paris (August, 1900), accepted these results, as seen from his tenth and eleventh propositions: “The receptors exist, sometimes in certain tissues only, sometimes in the majority of the organs (action of tetanus poison in the guinea-pig and in the rabbit),” “... the presence of numerous receptors in the organs of less vital importance may bring about—thanks to a kind of diversion of the toxin molecules—a diminution in the susceptibility of the animal to this toxin[637].” We must here recall the differences between the susceptibility of the guinea-pig and that of the rabbit to small doses of tetanus toxin frequently repeated as in Knorr’s experiments already referred to. The guinea-pig, subjected to these injections, dies in a tetanic condition long before it has received the minimal lethal dose for this species when injected in a single dose. The rabbit, on the other hand, is very tolerant of repeated doses and even rapidly acquires an immunity against five minimal lethal doses for the rabbit (injected at once). Knorr explained this difference as due to the hypersusceptibility of the nerve centres in the guinea-pig and to their acquired insusceptibility in the rabbit. The experiments of Roux and Borrel on the cerebral tetanus of rabbits vaccinated against tetanus, have demonstrated that this insusceptibility is not produced in these animals. We must, therefore, seek some other explanation. In rabbits subjected to small repeated doses, the poison is more and more prevented by certain living elements from reaching the nerve centres. Further, it is neutralised by the antitoxin which is rapidly produced. We find from Knorr’s[638] researches that in rabbits antitoxin appears in the blood in cases where, affected with a transitory tetanus, their limbs remain contracted for weeks. In guinea-pigs, affected with the same form of tetanus, antitoxin in appreciable quantity is never found, even after complete recovery. All these facts accord with the hypothesis that there exist, outside the nervous system, certain living cells which absorb the tetanus toxin and produce antitoxin. Rabbits and fowls possess this property in a much greater degree than do guinea-pigs. The fowl, according to Knorr, develops “a large quantity of antitoxin, whilst the tetanic symptoms are still augmenting.” In this animal, as we have been able to show[639], a portion of the tetanus toxin is absorbed by the leucocytes. By exciting aseptic exudations in fowls into which I had previously injected this toxin, I was able to convince myself that these exudations, much richer in leucocytes than was the blood, were also much more tetanigenic than was the blood. I observed also a more or less pronounced leucocytosis after the injection of non-lethal doses of tetanus toxin into fowls. It is possible that the leucocytes were actual agents in protecting the animal against the penetration of this poison to the susceptible nerve centres.

The great susceptibility of leucocytes to microbial toxins serves to indicate that these cells are of some importance in the struggle of the animal against these poisons. Their injection usually sets up a marked hyperleucocytosis of the blood. On this point Chatenay[640], working in my laboratory, has carried out a series of experiments on animals poisoned by bacterial (tetanus and diphtheria), phanerogamic (ricin and abrin) and animal (snake venom) toxins. He was able to demonstrate a striking analogy between them and the phenomena which occur in bacterial infections. When death takes place at the end of a very short period, the number of leucocytes markedly diminishes; if the animal lives beyond 24 hours or resists completely, a hyperleucocytosis, often of very marked character, is produced. In the guinea-pig, which is so susceptible to tetanus, the leucocytosis observed occurs even after injections of quantities of tetanus toxin equal to several lethal doses, and it is only after the introduction of an amount equal to one hundred times the lethal dose that the number of leucocytes remains stationary or shows a diminution. Here we have something analogous to what takes place against the anthrax bacillus in the same animal. The penetration of this deadly organism sets up a marked leucocytosis, but the accumulated leucocytes are incapable of seizing the bacilli or of preventing their noxious action. In other species of animals, such as the rabbit and the fowl, the intervention of the leucocytes against the anthrax bacillus, as well as against the tetanus toxin, is more effective.

[Sidenote: [414]]

If this toxin, instead of being injected in solution, be introduced along with the bodies of the micro-organisms which contain it, the struggle on the part of the animal takes place under more favourable conditions and even very susceptible animals may afford evidence that they offer a high resistance. Vaillard and Vincent[641] have shown that if we inject living tetanus bacilli, or the spores of these bacilli, deprived of free toxin, into guinea-pigs a great accumulation of leucocytes, which prevent the production of infection and intoxication by devouring the bacilli and their spores, takes place. The toxin contained in the ingested bacilli remains innocuous; this affording evidence of the protective part played by the leucocytes against the toxin. The same interpretation may be offered to explain the survival of animals very susceptible to tetanus, when the tetanus poison, mixed with pounded cerebral substance or with carmine powder, is injected. In these mixtures the toxin, as mentioned above, becomes attached to certain substances of the triturated brain or to the grains of carmine. This fixation is very unstable, the toxin is readily set free; but, when introduced into the body of the animal, the mixture induces a great accumulation of leucocytes which seize the cerebral particles and the grains of carmine and along with them take possession of the toxin. Wassermann and Takaki’s experiments and those of Stoudensky are easily explained if we assume two protective acts: the first of these consists in fixing the toxin, thus preventing it from diffusing and rapidly reaching the living nerve cells; the second is the absorption of the toxin fixed by the leucocytes,—cells endowed with receptors for the haptophore group of the toxin, but insusceptible to its toxophore group. When one of the two factors is absent, tetanus cannot be prevented. It is for this reason that in Courmont and Doyon’s experiments with emulsion of the frog’s brain, mixed with tetanus toxin, the inoculated animals died from tetanus in spite of an accumulation of leucocytes. This fact affords additional proof that, under these conditions, the toxin does not become anchored to the particles of the pounded cerebral substance, this anchoring being a condition necessary for the effective reaction of the leucocytes.

[Sidenote: [415]]

[Sidenote: [416]]

The absorption of the tetanus toxin becomes evident when we study in detail the phenomena produced in the experiments carried out according to Vaillard’s methods with tetanus spores and those of Wassermann and Takaki with poison to which cerebral emulsion has been added, or according to Stoudensky’s method with grains of carmine. When, however, it is desired to bring forward rigorous proof of the presence of the tetanus toxin inside the leucocytes charged with spores, with granules of cerebral substance or with grains of carmine, very great difficulties are encountered. How, indeed, is it possible to demonstrate this poison fixed upon these various bodies, a poison, the presence of which cannot be demonstrated except by its injection into the animal? For this, in the study of the reaction of the organism of the animal against the poisons, it is very important to have recourse to substances, whose presence can be demonstrated more easily than can the microbial toxins. We must first have recourse to the alkaloids, especially atropin, which, in this respect, present numerous advantages. We know that rabbits resist considerable doses of sulphate of atropin, even when this poison is injected directly into the blood. On the other hand, when it is introduced into the brain, according to Roux and Borrel’s method, even small quantities are quite sufficient, as demonstrated by Calmette[642], to produce a fatal poisoning. The intracerebral injection of the one-hundredth part of a dose which, when introduced into the circulation of the rabbit, produces no disturbance, in the same animal at the end of a few minutes sets up an enormous pupillary dilatation with symptoms of very lively excitation, increase of the reflexes, and general anaesthesia. These phenomena are succeeded by paralysis and death, which supervenes three or four hours after the injection. The natural immunity of the rabbit against atropin falls therefore into the same category as that against morphin. It is not due to the innate insusceptibility of the nerve cells, but to something which prevents the alkaloid from reaching these living elements. With the object of ascertaining the mechanism of this immunity, Calmette injected into the veins of rabbits a fairly large quantity of sulphate of atropin (0·2), he then bled these animals and collected from their blood the plasma and the white corpuscles, separating them by centrifugalisation. When injected into the brain of other rabbits, these constituents of the blood did not act in the same way. Whilst large doses of plasma set up merely a short period of excitation and a very transitory pupillary dilatation, corresponding quantities of leucocytes caused grave disturbances, sometimes followed by death in from seven to twelve hours. Calmette concludes from his researches that the atropin does not remain in the fluid part of the blood, since mere traces of it are found in the serum, but that it is seized and absorbed almost immediately by the leucocytes[643]. This result has been confirmed by Lombard[644] by another series of experiments. After injecting very large quantities of sulphate of atropin into rabbits and guinea-pigs, he bled these animals and separated out the elements of their blood. Instead of introducing these elements into the brain of rabbits, he injected them into cats, animals very sensitive to atropin. The cats which received the red corpuscles and the plasma exhibited very insignificant symptoms of poisoning. Those, on the other hand, which were injected with a corresponding quantity of leucocytes, had much graver symptoms of intoxication, such as photophobia with maximal pupillary dilatation, dysphagia and persistent diarrhoea.

It is, therefore, to the absorption of the atropin by the leucocytes that naturally refractory animals owe their immunity, an immunity which is very marked in spite of the susceptibility of the nervous elements of these animals. We have been able to obtain this result thanks to the delicate physiological reactions obtained with certain alkaloids. As regards arsenic the demonstration could be pushed even further, for the absorption of this mineral poison by the leucocytes has been established by chemical analysis.

[Sidenote: [417]]

When engaged in my researches on the leucocytic phenomena in intoxications I succeeded[645] in showing that in rabbits subjected to rapidly fatal doses of arsenious acid, there is a marked diminution in the number of white corpuscles in the blood. On the other hand, in rabbits habituated to arsenic, the same doses which brought about hypoleucocytosis and death of the control rabbits, induced a considerable rise in the number of leucocytes. Later, Besredka[646] made continuous and detailed researches upon this subject and obtained most interesting results. In order to simplify the conditions of experiment, he studied the reaction of the organism of the animal after the introduction of a red trisulphide of arsenic[647], a not very soluble salt, easily recognisable by its colour and markedly toxic. When non-lethal doses of this salt were injected into the peritoneal cavity of the guinea-pig, there was, first a transitory fall in the number of the white corpuscles in the peritoneal fluid, followed by a hyperleucocytosis of the most marked character. Of the leucocytes accumulated in the exudation the macrophages almost exclusively seized the yellowish-red granules of the trisulphide of arsenic. Very shortly, the whole of the salt injected was found within the peritoneal leucocytes, and the animals in which this marked phagocytosis occurred remained in good health. The ingested granules could be observed for several days in the macrophages; but in course of time, these arsenical particles were broken up into very small granules and ultimately disappeared. Here, then, we have an intraphagocytic solution of the trisulphide of arsenic and very probably a transformation of this salt into some other arsenical combination, innocuous to the animal. This soluble substance escapes from the macrophages and is finally excreted by the urinary passages.

[Sidenote: [418]]

Since the phagocytes ingest the trisulphide of arsenic and render it innocuous, it was to be anticipated that the elimination of these protective cells would lead to a fatal poisoning by doses which, under normal conditions, are readily withstood by guinea-pigs. When Besredka used sacs of reed pith containing non-fatal quantities of the red trisulphide and introduced them into the peritoneal cavity of guinea-pigs these animals were not long in exhibiting symptoms of poisoning and died at the end of a longer or shorter period, this varying with the amount of poison introduced. Even when the phagocytic reaction had been impaired as the result of a previous injection of carmine powder, the guinea-pigs died after doses of trisulphide of arsenic which, under ordinary conditions, did not kill them. The phagocytes in this experiment devoured numerous grains of carmine and were rendered incapable of ingesting enough of the trisulphide of arsenic to save the animal. On the other hand, when Besredka set up a previous accumulation of macrophages in the peritoneal cavity of his guinea-pigs, he succeeded in rendering these animals resistant to doses of trisulphide of arsenic that, under normal conditions, were fatal. The whole of these facts converge to show that the phagocytes, thanks to their power of seizing the trisulphide of arsenic and of modifying it within them, exercise a beneficent and immunising action on the organism of the animal. The analogy of the main facts concerning this protective influence with that observed in the immunity against infective micro-organisms is indeed very considerable.

Having determined the part played by the macrophages in the resistance of the organism of the animal against a not very soluble salt of arsenic, Besredka proceeded to study the leucocytic phenomena in poisoning by soluble arsenical compounds. In his experiments he made use of potassium arsenite and he found that when lethal doses were injected the guinea-pigs showed a diminution of leucocytes in the blood in less than 24 hours, whilst with non-lethal doses, he produced a marked hyperleucocytosis. When he injected lethal doses into rabbits accustomed to arsenic, these animals manifested an increase of white corpuscles, just as in animals injected with non-lethal doses. These oscillations in the number of leucocytes, like those which have been observed after poisoning by trisulphide of arsenic, certainly indicate that the organism and its defensive cells behave in the same way to both slightly soluble and very soluble salts of arsenic. In the first case it was easy to demonstrate that the accumulation of leucocytes in the blood and in the peritoneal exudation terminated in the ingestion of the granules of trisulphide. With potassium arsenite, it was not so easy to prove the point; a chemical analysis of the elements of the blood, however, has given a decisive answer. After injecting the lethal dose of this soluble salt into rabbits accustomed to arsenic, Besredka bled them in order to separate the plasma, leucocytes and red corpuscles. Several experiments made on these rabbits gave a concordant result which this observer sums up thus: “Although the bulk of plasma and of red corpuscles was much greater than that of the leucocytes, it was in the latter only that arsenic was found” by chemical analysis. It was only in those cases where the animals survived, and manifested hyperleucocytosis, that Besredka succeeded in demonstrating the presence of arsenic in the white corpuscles.

[Sidenote: [419]]

These experiments, excluding any doubt as to the protective part played by the leucocytes against arsenical intoxication, of course suggested the idea of investigating whether the nerve elements, submitted to the direct influence of potassium arsenite, exhibit any real susceptibility to this poison. The injection of solutions of this salt into the brain demonstrated that the one-hundredth part of an ordinary lethal subcutaneous dose was sufficient to cause fatal poisoning. This fact, then, falls into line with other facts, already numerous, as to the susceptibility of the nerve centres to microbial toxins, alkaloids and other poisons. But in the case of potassium arsenite, it was even more easily demonstrated than in the other cases that immunity natural or acquired, is connected with the absorption of the poison by the leucocytes. These cells, themselves much less susceptible to the toxic action than are the nerve elements, protect them from contact with the poison.

It is manifest that arsenic is not the only mineral substance capable of being absorbed by the phagocytes, and there are already on record well established facts in support of this thesis. Some time previous to the researches on arsenical poisoning just summarised, Kobert, then in Dorpat, set his pupils, Stender, Samoïloff, Lipsky and others[648] to make systematic researches on the fate of iron in the animal organism. For this purpose these observers made use of a very soluble preparation of iron—or better expressed, as soluble as possible—Dr Hornemann’s _ferrum oxydatum saccharatum solubile_, which does not precipitate in alkaline media. They proved that a small quantity of the iron introduced into the animal is eliminated by the kidneys and the wall of the intestine, but that the greater part of the metal is arrested in the organs, especially the liver, spleen and bone marrow. The iron is there absorbed by the leucocytes which hold it for some time and then throw it into the intestine.

[Sidenote: [420]]

I have had the opportunity of observing this circulation of Dr Hornemann’s soluble salt in the organism of several species of vertebrates. Some time after its introduction into the organism by the blood vessels, peritoneally or subcutaneously, the iron may be found (by means of the microchemical reaction with potassium ferrocyanide) accumulated in the various phagocytes, especially the leucocytes, the stellate Kupffer’s cells of the liver and the macrophages of the splenic pulp. The non-phagocytic cells, as, for example, Ehrlich’s basophile leucocytes, so abundant in the lymph of rats, take up very little of this iron, although the macrophages and microphages are full of it[649]. Against these facts Weigert[650] has advanced the objection that the leucocytes absorb only the iron precipitated in the form of granules, but my own researches allow of no doubt that not only granular but dissolved iron is absorbed. This discussion, however, loses much of its importance in view of the results obtained with potassium arsenite.

According to Samoïloff[651], soluble salts of silver in the animal organism undergo a fate similar to that of Hornemann’s soluble iron salt and are absorbed by the phagocytic elements. It must be noted, further, that according to the experiments of Arnozan and Montel[652], the leucocytes absorb such drugs as calomel and salicylate of soda.

[Sidenote: [421]]

These observations all clearly show that the phagocytes must not be looked upon as cells capable of seizing merely the dead bodies of micro-organisms and of animal cells, always fearing and avoiding poisons and only able to come forward when protected by some other antitoxic function. The phagocytes no doubt often exhibit a negative susceptibility for many poisons, when these are introduced into the animal organism in too large a quantity. But these cells are most resistant to toxic substances and protect the higher elements from the poison. Under these conditions, it is quite natural to assign to the phagocytes the rôle of the fighting agents of the animal organism against poisons and we may even enquire whether these elements do not produce the antitoxins. It has been pointed out that it is very difficult to attribute this function to the cells susceptible to the toxic action,—the spermatozoa in the production of antispermotoxin, the red blood corpuscles in the development of antihaemotoxin, or the nerve cells in the production of tetanus antitoxin. Moreover since, according to Ehrlich’s theory, it is only the haptophore group which excites the formation of antitoxins on the part of the elements which possess the corresponding receptors, it is quite possible that the phagocytes, thanks to the facility with which they absorb the poisons, occupy an important place as producers of antitoxins. I have already formulated this hypothesis, and several investigators, amongst whom may be cited Gautier[653] and Courmont[654], have received it favourably, though in the imperfect state of our knowledge, it cannot, as yet, be fully demonstrated. It might perhaps be objected against this hypothesis that in many instances, after the injection of micro-organisms living or dead, in spite of a vigorous leucocytic reaction the organism of the animal does not produce any antitoxin. In such cases, there is clearly a development of antibodies, such as the fixatives, whose phagocytic origin may reasonably be claimed, but no true antitoxins. It must not be forgotten, however, that the various kinds of phagocytes present, amongst themselves, great differences, and that perhaps certain only of these elements are capable of producing antitoxins. When micro-organisms, living or dead, are introduced into an animal it is found that antitoxins do not as a rule appear in the fluids; in these cases the reaction is set up mainly by the microphages. The macrophages represent the principal source of antitoxins. In cases where these phagocytes ingest the micro-organism the blood exhibits an undoubted antitoxic power. Such is the case with bubonic plague in the human subject, where the micro-organism is readily ingested by the macrophages. Here we obtain antitoxic serums even after the introduction of living or dead organisms into the animal, a fact observed by Roux and his collaborators. Another fact in favour of the hypothesis I am defending is furnished to us by the cayman. As noted above, this reptile, of all known animals, supplies antitoxins most quickly and easily. In the cayman the leucocytic system is composed of eosinophile microphages filled with granules, and of macrophages. As the eosinophile cells are only very weakly phagocytic, it is the macrophages almost exclusively which intervene in the reaction against the micro-organisms. It is probable, then, that in the cayman and in animals inoculated with the plague bacillus the exclusion of the microphages from the struggle constitutes a factor favourable to the production of antitoxins and at the same time favourable to the manifestation of the activity of the macrophages.

[Sidenote: [422]]

If these latter phagocytes play the primary rôle in the excretion of antitoxins in the fluids of the body we should expect to find this function exercised not only by the motile macrophages of the blood and lymph, but also by the fixed macrophages, so widely diffused through almost all the organs.

I advance this hypothesis for what it is worth, simply as a guiding idea for new researches in this field, of which so much is still unknown[655]. The brief account of the actual state of the question of artificial immunity against toxins, has indicated to us that this is a problem far more difficult of solution than is that of acquired immunity against micro-organisms. The mere fact that these latter can still be found some hours or even days after their entry into the refractory animal, affords a great advantage in these researches as compared with those on toxins which are lost, often almost immediately, after their injection. Consequently our knowledge of antimicrobial immunity is more advanced than is that on immunity against the soluble products of micro-organisms.

The facts narrated in this chapter support the thesis I have defended on the subject of immunity against micro-organisms—that antimicrobial immunity in no way depends on a previous resistance against the toxins. As a general rule the immunity against micro-organisms is developed more readily than the immunity against their toxic products and at an earlier stage.

Although much still remains to be done in the elucidation of the mechanism of antitoxic immunity, the principal data acquired on the subject of this immunity have undoubtedly led to applications of the highest importance, as will be set forth in one of the following chapters.