We stated at the beginning of the last chapter that various fluid substances of very complicated chemical composition may be absorbed by the tissues and utilised by the organism without requiring to be modified by the digestive juices of the intestinal canal. We must now describe, exactly, the phenomena observed in these cases and endeavour to establish the mechanism of the absorption of fluids in the living organism.

We have already cited the examples of blood serum, milk, and white of egg, all of which are readily utilised by the organism which receives them directly into the peritoneal cavity or below the skin. The proof that these substances are modified—digested by the tissues, is furnished by the observation that their injection necessarily brings about appreciable changes in the properties of the blood. Th. Tchistovitch^[142], in a research carried out in the Pasteur Institute, was the first to demonstrate that the resorption of the blood serums of the eel and horse by the organism of the rabbit, excites in the blood of the latter animal the production of specific precipitates. The blood serum of rabbits that have been vaccinated against the toxic eel’s serum gives a precipitate with eel’s serum; the serum of rabbits treated with horse’s blood gives a similar precipitate with horse’s serum, etc. This property has since been confirmed and studied by several observers, who have made use of it for the recognition of human blood in medico-legal investigations^[143].

Bordet^[144] has made the discovery that intraperitoneal injections of the milk of cows into rabbits provokes in the blood serum of the latter the property of giving a specific precipitate with cow’s milk only. This precipitation bears a great resemblance to the coagulation of casein; which, however, does not justify us in identifying the precipitating substance with rennet. This fact has been confirmed for several other species of milk, and Schütze^[145], in an investigation carried on in the Berlin Institute, essayed to apply it to the differentiation of the various kinds of milk. In the same order of ideas, researches have been made on the artificial precipitins that develop in the blood as the result of injection of white of egg and other albuminoids^[146]. Leclainche and Vallée^[147] have prepared animals in such a fashion that their serum produces a precipitate with urinary albumen. The biological precipitin reactions are more sensitive than any of the chemical reagents properly so called. These specific substances in the serums must be looked upon as belonging to the group of soluble ferments, approximating to the fixatives rather than to the cytases, since they are unaltered by being heated to 56° C. Their action gradually declines after passing 60° C. but is only destroyed at a temperature beyond 70° C.

An analogous soluble ferment has been discovered in the blood serum of animals treated with injections of gelatine. We owe to Delezenne, who has studied this question in his laboratory at the Pasteur Institute, the most important and most complete data on the resorption of gelatine. The blood serum of normal animals possesses only a very feeble power, sometimes even none, of liquefying gelatine. When however this substance is injected several times, the serum, as is the rule for the formed elements and quite a series of fluid substances, acquires a much more pronounced activity. The gelatine, without giving any precipitate, is simply dissolved and will no longer solidify when it is cooled. The ferment of the serum that produces this effect resembles the precipitins in that it withstands the action of a temperature of 56° C. and is only destroyed beyond 60° C. Like the trypsins it acts in a weakly alkaline, neutral, or weakly acid medium; but digestion takes place best in a slightly alkaline medium.

The question of especial interest to us is that of the origin of this ferment which digests gelatine. If several c.c. of a 10% solution of this substance be injected into the peritoneal cavity of a laboratory animal, there is provoked with certainty, within a few hours, a marked leucocytosis of the peritoneal fluid. A considerable afflux of leucocytes, amongst which the microphages are even more numerous than the macrophages, takes place. When to a hanging drop of such an exudation a trace of Ehrlich’s neutral red solution is added, there appears almost at once an intense coloration of the numerous droplets inside the two kinds of leucocytes. It is, therefore, manifest that the gelatine excites a powerful positive chemiotaxis of the mobile phagocytes and that it is absorbed by these cells. This experiment demonstrates that the phagocytes can not only ingest solid bodies, such as the various formed elements, coloured granules, etc., but that they are also capable of absorbing fluid substances introduced into the tissues or cavities of the organism.

The data brought forward by Delezenne demonstrate very clearly the part played by the mobile phagocytes in the digestion of gelatine. He obtained his best results in the dog. We know that it is easy in this animal to provoke an aseptic exudation, very rich in leucocytes. This exudation when deprived of its serum and washed with physiological salt solution gives a solution which exerts a feeble digestive action on gelatine. If the exudation be produced in a dog that has previously received several injections of this substance, we obtain leucocytes whose extract, obtained by the same method, will digest gelatine much more actively. The digestive power of the leucocytes of the treated dog is sometimes five times greater than that of the leucocytes of the normal dog. Here, then, we undoubtedly have an acquired digestive power which reveals a great reinforcement of the phagocytic activity.

In the prepared dogs the leucocytes have a much greater digestive action on gelatine than has the blood serum of the same animals, a fact which indicates that the source of the soluble ferment must be sought for in the phagocytes themselves. The results of these researches are of great service to us in the study of immunity properly so called.

For some time past attempts have been made to show that the soluble ferments, diastases, or enzymes, are closely allied to albuminoid substances. Nencki and Mme Sieber^[148] support this view by their recent researches on the chemical composition of pepsin. In all the above cases there is this in common between the two categories of substances, their absorption by the organism is followed by the appearance in the blood of antagonistic ferments. Just as after the injection of milk, white of egg, serums, etc. into the cavities or tissues, specific precipitins are produced, so the injection of certain enzymes provokes the formation in the organism of antienzymes or antidiastases.

It has been known for some time that the blood serum of many animals prevents the action of certain enzymes. Thus Röden has shown that normal horse’s serum retards or even completely prevents the coagulation of milk by rennet. It has often been observed, too, that normal serums hinder, more or less, the digestion of albuminoids by trypsin. It is only quite recently, however, that we have begun to prepare antienzymes by the injection into animals of corresponding enzymes. Thus, Hildebrand^[149] has succeeded in obtaining an antiemulsin in the serum of rabbits, into which he had injected several separate doses of emulsin. Fermi and Pernossi^[150] have prepared an antitrypsin, and von Dungern^[151] has obtained an antidiastase against the proteolytic enzymes of some bacteria. But of all the antienzymes, the one that has been best studied up to the present is indisputably antirennet, obtained independently by Morgenroth^[152] and Briot^[153]. The former of these investigators treated goats with increasing quantities of rennet and was able to assure himself, by comparative detailed researches, of the appearance and increase in quantity of antirennet in the blood serum. The goat which gave the best result ceasing to develop antirennet it was impossible to make the antirennet potency go beyond a certain point.

Briot also obtained antirennet in rabbits into which he had injected fluid rennet on several occasions. He was able to convince himself that the antirennet of horse’s serum is a non-dialysable substance which is precipitated by alcohol and certain salts. Like the precipitins and the diastase which digests gelatine, antirennet resists a temperature of 55°–56° C.; even heating to 58° C. has no effect on the antirennet serum. At 60° C. however, the heat begins to exert an injurious effect, and after three hours at 62° C. the serum has lost all power to prevent the coagulation of the casein by antirennet. Morgenroth and Briot both state that the antirennet neutralises the rennet by a direct action.

The cell poisons, or cytotoxins, of animal origin which were treated in the preceding chapter, likewise set up the production of special antibodies, or anticytotoxins. The consideration of these latter has a very special interest for those who study the question of immunity from a general point of view. The first discovery of these anticytotoxins was made in connection with the study of the toxic power of the blood serum of eels. Camus and Gley^[154] and, independently of them, H. Kossel^[155] demonstrated that animals when treated with increasing doses of eel’s serum acquire an antitoxic property which protects their corpuscles against the haemolytic action of ichthyotoxin, or the toxic substance of the blood of eels. Th. Tchistovitch^[156] has not only confirmed this discovery, but has added to it new and interesting data.

When antitoxic serum is mixed in vitro with red blood corpuscles of the species which furnished the serum and there is added to it some haemolytic eel’s serum, it will be found that the red corpuscles remain quite unaltered. In the control tubes, however, in which the antitoxic serum is replaced by normal serum of the same species, the red corpuscles are very readily dissolved under the toxic influence of the eel’s serum. In animals (rabbits) that are treated with this latter fluid, there is established not only an antitoxic power of the blood, but the red corpuscles acquire a resisting power more or less pronounced against the ichthyotoxin of eel’s serum. When the red corpuscles are separated from the serum of rabbits (treated with eel’s serum) and some ichthyotoxin is added to them, solution very often does not take place at all. According to the experiments of Tchistovitch there is no direct relation between this acquired resistance and the antitoxic power of the blood. Sometimes even a kind of antagonism is observed between the two properties; that is to say, the red corpuscles of a rabbit whose serum is very antitoxic may be extremely sensitive to the poison of the eel whilst the converse may also hold good [cf. infra p. 120].

The toxic action of the eel’s serum upon the red corpuscles of a great number of Vertebrates is a natural property which demands no previous treatment of the eel. It is the antitoxic power, directed against the ichthyotoxin, which is developed only as a result of the preparation of the animals by the administration of increasing doses of eel’s serum. Nevertheless we also find natural antitoxins present in the blood of man or animals that have not been treated and which act against the cell poisons, cytotoxins, so widely distributed in the blood of a large number of species of animals.

Besredka^[157] has demonstrated that the blood serum of Man and many Vertebrates contains a substance which prevents the solution of red corpuscles under the influence of blood serums of a different species. To reveal the presence of these antitoxins it is useful to heat the serums to 56° C. and then to add to them red corpuscles of the same species and some haemolytic serum of a different species. Under these conditions the solution of the red corpuscles does not take place, whilst their mixture with haemolytic serum alone, inevitably provokes haemolysis.

Along with these natural antihaemolysins there exist a number of artificial antihaemolysins or antihaemotoxins. Jules Bordet^[158] was the first to draw attention to this important subject. He first obtained these antihaemolysins by injecting blood serum of the fowl, which possesses a very great haemolytic power on the red corpuscles of the rabbit, into individuals of this latter species. After some injections, the serum of these treated rabbits was found to be antihaemotoxic against the fowl’s serum. Later^[159], Bordet obtained a serum against an artificial haemotoxin. The serum of the guinea-pig is innocuous to the red corpuscles of the rabbit. But when rabbit’s blood was injected several times into guinea-pigs the serum of the latter became very solvent for the red corpuscles of the rabbit. To prevent this action it is sufficient to inject the haemotoxin of treated guinea-pigs several times into rabbits. The serum of these rabbits becomes antihaemotoxic and protects the red corpuscles of the rabbit against the solvent action of guinea-pig’s serum.

In the normal haemolytic serums, such as the serums of the eel and fowl, the presence of two substances which act by combining could not be demonstrated. On the other hand, in the serums that were obtained as a result of the treatment of animals by the injection of blood from a different species, it was easy to demonstrate, as we have shown in the preceding chapter, the presence of two constituent substances which are: the macrocytase (alexine, complement) and the fixative (amboceptor of Ehrlich, sensibilising substance of Bordet). For this reason the study of the antihaemotoxins obtained against artificial haemotoxins is endowed with special interest. As the solution of the red corpuscles, in this case, can be prevented either by an antitoxic action directed against the cytase, or by a neutralisation of the fixative (for the concurrence of these two substances is indispensable in order that the solution may take place), Bordet asked whether the antitoxic serum, obtained by him in rabbits, is anticytatic or antifixative, or whether it contains both properties. Before resolving this problem it was necessary to establish some of the essential characters of artificial antihaemotoxic serums. The principal one amongst them is the resistance of these antihaemotoxins to a temperature of 55–60° C.; even when heated to 70° C. the antihaemotoxins retain, at least in part, their fundamental property. In this respect these substances differ from the cytases and approach the precipitins, fixatives and agglutinins.

The very exact experiments carried out by Bordet have demonstrated that in the serum of rabbits, treated with the specific haemotoxic serum of guinea-pigs, two substances, an anticytase and an antifixative, are found in combination. The former of these antitoxins is found in abundance, but the amount of antifixative is very small. Bordet was led to this result in the following way. To prevent the solution of the red corpuscles of the rabbit in the haemotoxic serum of the guinea-pig, it was necessary for him to add a considerable dose (10 to 20 times) of the antitoxic serum. When, however, he heated the latter to 55° C. the quantity of this serum necessary to prevent haemolysis could be reduced very considerably. In place of its being necessary to add to the haemotoxic serum 10 or 20 volumes of antitoxic serum, it was sufficient to add three or sometimes only two volumes of this heated serum. As we know already, heating to 55° C. destroys the macrocytase which should be found in the antitoxic blood of the rabbit. This cytase by itself is incapable of dissolving the red corpuscles of the same species; but when it is added to the fixative of the haemotoxic serum of the guinea-pig the macrocytase of the rabbit’s serum dissolves them very readily. Hence the conclusion that in the haemotoxic serum of the guinea-pig there must be present a quantity of fixative sufficient to allow of the solution of the red corpuscles by the macrocytase of the rabbit’s serum. This antitoxic serum, therefore, which only prevents the haemolysis on the condition of being added in comparatively large quantity, contains very little antifixative. When, by heating this serum to 55° C. we destroy the rabbit’s macrocytase, the mixture of antitoxic serum of the rabbit and haemotoxic serum of the guinea-pig, which ordinarily dissolves the red corpuscles of the rabbit, now leaves them intact. The reason is that the free fixative contained in this mixture does not find any available macrocytase: that of the rabbit being destroyed by the heating, and that of the guinea-pig neutralised by the antitoxic serum. The experiment I have just described proves that this antitoxic serum contains specific anticytase. This anticytase is capable of neutralising the guinea-pig’s macrocytase, but is altogether powerless against that of the rabbit. This last circumstance allows us to investigate whether the antitoxic serum of the rabbit contains, in addition to anticytase, a specific antifixative. Bordet prepared a mixture of antitoxic serum of the rabbit, heated to 55° C., with haemotoxic serum of the guinea-pig, also heated to 55° C. In this mixture the two macrocytases (that of rabbit and that of guinea-pig) have been destroyed by heat, but the antitoxins of the rabbit’s serum and the fixative of the haemotoxic serum have remained intact. This mixture owing to its want of macrocytases was incapable of dissolving the red corpuscles of the rabbit. By adding to it some fresh unheated serum from a normal rabbit the rabbit’s macrocytase was introduced. As the latter could not be neutralised by the anticytase of the antitoxic serum and was incapable, by itself, of dissolving the red corpuscles of the rabbit, it was unable to produce haemolysis except on the condition that there is in the mixture a sufficient quantity of unneutralised free specific fixative. As a matter of fact, the red corpuscles of the rabbit are not dissolved in the mixture described; this proves that the fixative had become inactive in consequence of the presence of an antifixative in the antitoxic serum of the rabbit. I need not enter into further details of Bordet’s experiments, which have fully demonstrated the fact that in the antitoxic serum of his rabbits there were really two antitoxins; an anticytase abundant in quantity, and an antifixative present in much smaller amount.

Ehrlich and Morgenroth^[160] quite independently of Bordet have shown that an antihaemotoxic serum is very rich in anticytase. After making a number of injections of normal horse’s serum (very rich in cytase) into a goat, they obtained in the blood serum of the latter an anticytase very active against the cytase of the horse. This antitoxic serum of the goat, as might be anticipated, contains no antifixative, the horse’s serum that served for the injections coming from normal horses which contained no, or very little, fixative. Even in another case, where these investigators^[161] injected a dog with sheep’s serum very rich in fixative specific for the red corpuscles of the dog, they did not succeed in obtaining any antifixative. These observations in no way diminish the value of the discovery of the antifixative by Bordet, though they demonstrate that this antitoxin cannot, in certain cases, be found in the serum. Ehrlich and Morgenroth themselves throw out, in this connection, the suggestion that in these cases the antifixative remains linked to the cell which produces it, without being thrown off into the blood.

The very precise data that we have just summarised do not seem to agree with the statements of certain other investigators. Thus Schütze^[162], from his researches on the antihaemotoxic serum of guinea-pigs, directed against the rabbit’s haemotoxin, has arrived at the conclusion that in the former an antifixative only is produced. As he merely injected into his guinea-pigs haemotoxic rabbit’s serum that had been heated to 60° C. and consequently deprived of the macrocytase, he concluded that in this serum there remained only the specific fixative capable of provoking the formation of an antitoxin. This must consequently be an antifixative. Paul Müller^[163] came to a similar conclusion, after injecting rabbits with the heated haemotoxic serum of fowls. These injections caused the formation in the rabbit’s serum of an antitoxin that Müller regarded as an antifixative.

Ehrlich and Morgenroth^[164] objected to this interpretation, taking their stand on experiments made with the serums of normal animals. They were able to show that these serums, when injected in the fresh state or after being heated to 60° C., caused the production of a corresponding antihaemotoxin which is nothing but an anticytase. When Schütze and Paul Müller concluded that by heating the serums they had entirely deprived them of cytase elements they did not take into account the possibility of the cytases being transformed, under the influence of heat, into other bodies unable to produce haemolysis, but quite capable of provoking the formation of anticytases. Ehrlich and Morgenroth give to these new bodies, derived from cytases under the influence of temperatures between 55°–60° C., the name of complementoids; and these complementoids appear in the experiments of Schütze and Müller to have caused the production of antitoxins—anticytases.

In all the investigations just summarised the anticytases have been obtained by the injection into animals of various blood serums, fresh or heated. Wassermann^[165] has discovered another method of arriving at the same result. He injected into guinea-pigs the leucocytes of rabbits, carefully deprived of all traces of serum. After some time the blood serum of guinea-pigs thus treated became weakly but distinctly anticytatic. From this experiment Wassermann draws the conclusion that, as has been often affirmed by several observers, the leucocytes really contain cytases.

How do the anticytases act upon the cytases? On this point all observers who have studied this question have but one answer, the action of the anticytases is direct. Bordet thinks that the two substances combine so intimately that they cannot be again separated by heat. We know that the cytases are very sensitive to heat and that their haemolytic property is destroyed at 55° C. The anticytases, on the other hand, as already noted, are much more resistant to the action of heat. Bordet has prepared mixtures of haemolytic cytase serum and of antihaemolytic serum, neutral mixtures, that is to say, inactive for red corpuscles or with a very feeble action upon red corpuscles that have been sensibilised by the specific fixative. These mixtures no longer exhibit antihaemotoxic properties or they exercise this power in a very feeble degree. If in these mixtures the cytases remain uncombined alongside the anticytases, it is to be expected that heating them to 55° C. will restore the antihaemotoxic function of the anticytases; the cytases being destroyed at 55° C. there will remain in the mixtures only active anticytase. The experiments made on this point have demonstrated that the heating of these mixtures does not restore the antihaemotoxic action, that is to say, the anticytase is definitely combined with the cytase.

Ehrlich and Morgenroth have satisfied themselves that their antihaemotoxin exerts no influence, either upon the red corpuscles or upon the fixative, and is only capable of preventing the action of the cytase. They introduced red corpuscles of the rabbit into a mixture of goat’s serum, heated to 56° C. and thus only retaining its fixative, and anticytase serum. The fluid bathing the red corpuscles was then removed by centrifugalisation and the corpuscles were mixed with normal haemolytic horse’s serum. Solution of the red corpuscles took place at once as the anticytase had been completely removed during centrifugalisation, being combined with neither the red corpuscles nor the fixative.

These investigators have obtained various anticytases by injecting serum of various species of animals into other mammals. They observed, however, that injections of the serum of an allied species did not bring about the formation of anticytases. Thus the injection of goat’s serum into sheep, or of that of sheep into goats, never produced anticytase serum.

In addition to antihaemotoxic serums several other analogous anticytotoxic serums have now been obtained. Thus Delezenne has prepared serums which prevent the action of neurotoxin and of the cell poison which destroys the liver cells. We^[166] have been able to obtain a rabbit’s serum which prevents the spermatozoa of this rodent being rendered motionless by the specific spermotoxin of the guinea-pig. More recently Metalnikoff^[167], working in my laboratory, has prepared another antispermotoxic serum which prevents the specific spermotoxin of the rabbit from arresting the movement of the guinea-pig’s spermatozoa.

As the history of these antispermotoxins presents certain interesting general features we may with advantage, perhaps, dwell on some of their characters. The two antispermotoxins mentioned above are distinguished by certain peculiarities. When Metalnikoff set to work to inject rabbit’s spermotoxin into guinea-pigs, he thought that he had an easy task before him and that after a few injections the guinea-pig’s serum would become antispermotoxic. This, however, was not the case. The serum from these animals when mixed with spermotoxic serum was powerless to prevent the immobilisation of the spermatozoa of the guinea-pig. It was only when he heated the serum of his treated guinea-pigs to 56° C. that the antispermotoxic power appeared with the greatest distinctness. The inefficacy of the unheated serum must therefore depend on the toxic action of the guinea-pig’s macrocytase, because it is this substance alone that can have been destroyed by the heating process. Now, in order that this macrocytase may act, the presence of the fixative is necessary, which leads us to the conclusion that the serum of the guinea-pigs injected by Metalnikoff contained no antifixative. This hypothesis was fully confirmed by experiment. Metalnikoff introduced a drop of guinea-pig’s serum into a mixture of antispermotoxic serum, heated to 56° C., with spermotoxic serum. The spermatozoa continued their movements in normal fashion. But when afterwards he added a few drops of unheated serum from a normal guinea-pig the motions of the spermatozoa were arrested almost instantaneously. Consequently there was present in the mixture rabbit’s macrocytase which had been neutralised by the anticytase of the prepared guinea-pig’s serum and for that reason the spermatozoa remained motile. But in the same mixture we had also the specific fixative, coming from the rabbit’s spermotoxic serum, which remained free and not neutralised. The motile spermatozoa had become impregnated with this fixative and a little guinea-pig’s macrocytase (against which the anticytase was powerless) was sufficient to make them suddenly cease their movements.

There is no doubt, then, that the serum of guinea-pigs that have been treated with spermotoxin contains anticytase only and no, or almost no, antifixative. Such is not the case with the antispermotoxin obtained by us in rabbits that were treated with spermotoxic toxin of guinea-pigs. Several consecutive injections were sufficient to render the serum of the rabbits so treated capable of preventing the action of the spermotoxic serum of the guinea-pig on the motility of the rabbit’s spermatozoa. In the mixture of antispermotoxic serum and spermotoxic serum these spermatozoa continue to move for a considerable time, whilst in the control mixture prepared with normal rabbit’s serum and spermotoxic serum they become motionless at the end of a few minutes. To obtain this marked effect it was not necessary to heat the antispermotoxic serum as in Metalnikoff’s case. Indeed I have performed almost all my experiments with fresh serums, unheated. As the rabbit’s serum contains macrocytase capable of rendering the spermatozoa, sensibilised by the fixative, motionless and as this macrocytase cannot be neutralised by the anticytase that is active against the guinea-pig’s macrocytase, the fact I have just pointed out indicates that the antispermotoxic serum of my rabbits contains antifixative. The difference between the antispermotoxic serum obtained by Metalnikoff and that prepared by me is similar to that observed between the antihaemotoxic serums. Some contain only anticytase but others undoubtedly contain antifixative also.

As this result appeared to me to be of far-reaching importance I felt bound to verify it by another method. I injected certain rabbits with spermotoxic serum of the guinea-pig and others with normal guinea-pig’s serum. The amount of cytases being about the same in both, the strength of the serums obtained as the result of injections of normal serum and of specific serum should be the same if the antispermotoxic serums contain anticytase only. Experiment demonstrates just the contrary. The antispermotoxic serum of rabbits treated with normal guinea-pig’s serum was on every occasion much less active than the serum of rabbits injected with the spermotoxic serum of prepared guinea-pigs. The former contains anticytase only, whilst the latter contains in addition antifixative. Weichhardt’s^[168] experiments carried out in my laboratory corroborated the conclusion I have just formulated.

Having made ourselves acquainted with the constitution of the anticytotoxins we may now pass to the question of the origin of these bodies and of analogous ferments which act in the resorption of albuminoid substances in the blood and in the tissues.

We have already mentioned that the leucocytes are charged with a soluble ferment which digests gelatine, and that in animals treated with injections of gelatine these cells elaborate a much larger amount of the ferment. Here we have evidence of a kind of education of the leucocytes to produce a greater amount of digestive ferment, in a manner quite analogous to that which has been described in Chapter III in connection with the augmentation of the pancreatic ferments in intestinal digestion. It is, then, quite permissible to look upon leucocytes, and probably phagocytes in general, as the source of the soluble ferment that digests gelatine.

Is this the case with the other substances which take an active part in the resorption of albuminoid substances in the fluids and tissues of the organism? Up to the present the origin of precipitins and antiferments, such as antirennet, has not been studied. The problem being very complex and difficult, it appears to be impossible at present to solve it. It is known indeed that the introduction of these substances into the organism provokes a reaction similar to the one we have described in the case of the injection of gelatine into the peritoneal cavity of guinea-pigs. Thus Morgenroth^[169] observed that in his goats the subcutaneous injection of sterile rennet caused the formation of extensive infiltration at the seat of inoculation, this being accompanied by fever; we are justified in concluding from this that rennet provokes a marked leucocytic reaction. Hildebrandt^[170] has demonstrated by direct experiment that rennet, when enclosed in capillary glass tubes and introduced below the skin of rabbits, induces a marked positive chemiotaxis. This led to the formation of a leucocytic plug several millimetres long. Now we know from Briot that the rabbit is capable of producing antirennet. Hildebrandt has further shown that several other diastases, or hydrolytic ferments, such as sucrase and emulsin, give rise to a similar chemiotactic phenomenon. The leucocytic reaction is consequently a general phenomenon following the introduction into the tissues of substances of complex chemical composition capable of provoking the formation of antibodies. We are tempted from this fact to accept it as a law that the leucocytes are capable of producing these latter substances. Although this hypothesis may be very probable, the number of facts at our disposal is not yet sufficient to justify the statement that its truth is demonstrated.

Since it is the red corpuscles which are affected by the haemotoxins it might be asked whether it may not be that these elements defend themselves by the production of antihaemotoxins the overplus of which is thrown into the blood and fluids in general? The researches that have been made on this point relate especially to the antihaemotoxin of the blood serum of rabbits in relation to the ichthyotoxin of eel’s serum.

We must therefore examine the collected evidence bearing on anticytotoxins and analogous bodies and endeavour to form some idea as to their probable origin. A large accumulation of exact data bearing on the antihaemotoxins does not afford us sufficient information as to the source of these substances.

Let us first examine the question, is it possible to attribute to the red corpuscles the function of producing the antihaemotoxins? If these elements are really the source of the antihaemotoxins it is probable that the red corpuscles of animals whose serum is antihaemotoxic will exhibit marked resistance to the toxins; thus we know that the white corpuscles which produce gelatinase digest gelatine much better than does the serum of the same animals. From the experiments of Tchistovitch (l. c. supra p. 110) on rabbits that have been immunised against eel’s ichthyotoxin, it must be accepted that the red corpuscles of these animals are often very sensitive to the action of the poison at a period when the blood serum of the same rabbits exhibits a marked antihaemotoxic power. It is not until later in the process of immunisation, when the serum loses a great part of this power, that the red corpuscles become resistant to the ichthyotoxin.

But before we abandon the hypothesis of the production of antihaemotoxins by the red corpuscles we must see if it cannot be reconciled with the facts, by the application of Ehrlich’s side-chain theory^[171]. This theory was evolved with the object of explaining the production of antitoxins and their action on bacterial and vegetable toxins. Later, Ehrlich has extended it to the cytotoxins, anticytotoxins and bactericidal substances.

According to Ehrlich the complex molecule of albuminoid substances contains, besides the central stable nucleus, a number of side-chains, or “receptors,” which fulfil various accessory functions and serve especially for the nutrition of the cell. These receptors have a great affinity for the various substances necessary for the maintenance of the life of the cell. Under normal conditions these receptors seize nutritive molecules, as a leaf of Dionaea seizes the fly that serves it as food. Under special conditions these receptors lay hold of complex molecules of albuminoid substances, such as the various toxins. In this case the receptor, instead of combining with a molecule which supports life, fixes a molecule which poisons the cell. According to Ehrlich’s theory on the constitution of toxins their molecules contain an atomic group which poisons—the toxophore, and another group which combines with the receptor—the haptophore. The toxic group of a complex poison, such as ichthyotoxin, cannot penetrate into a red corpuscle except by the help of the haptophore group and of the corresponding receptor. When a red corpuscle has absorbed a large number of molecules of ichthyotoxin, the united action of the toxophore groups renders life impossible and the corpuscle is dissolved. But when a red corpuscle has been touched by only a few toxic molecules, too few to compromise life, there is merely immobilisation of the receptors which are combined with the haptophore groups of the ichthyotoxin. As these receptors fulfil an important function in the nutrition of the red corpuscles, the latter reproduce them in larger numbers than were originally present. We know that in the phenomena of repair an over-production of the new-formed parts often takes place and, according to Ehrlich, to this over-production the presence of antitoxins in the fluids of the body is due. The receptors, developed in excess by the red corpuscles, fill these cells, and no longer finding room therein are extruded from them and overflow into the blood and other fluids of the organism. When a fresh injection of toxin makes its way to the blood it there meets with a number of free receptors, endowed with an affinity for the haptophore group of the molecule of the toxic substance. The chemical combination between the two substances takes place at once in the plasmas, a fact which prevents the haptophore group of the toxin from uniting with the receptor of the red corpuscles and so injuring these cells by introducing the toxophore group into them. According to this theory the same receptors which, in the free state in the fluids, fulfil the antitoxic function become in the interior of the red corpuscles the vehicles of intoxication and consequently fulfil a philotoxic function. This opposite rôle of the receptors has often been compared to a lightning-conductor; so long as the receptors are attached to the molecule of the living protoplasm they attract the toxin just as a lightning-conductor attracts the lightning when it is badly insulated.

So interpreted, it is easy to conceive that the red corpuscles of animals whose fluids are antihaemotoxic may be sensitive to the toxic action of eel’s serum, as has been observed by Tchistovitch. As soon as the protective fluids have been removed from the red corpuscles of the immunised organism, the corpuscles when placed in contact with ichthyotoxin (eel’s serum) attract the haptophore groups of the poison by means of their numerous receptors. These haptophores in their turn introduce the toxophore groups which dissolve the red corpuscles without the slightest difficulty. This theory does not explain the cases, which are numerous, in which the red corpuscles of rabbits that are vaccinated against eel’s poison resist this poison. Camus, Gley, and Kossel, working independently, have arrived at the result that the red corpuscles of immunised rabbits, from which the serum has been carefully removed, are not dissolved when submitted to the action of ichthyotoxin, whilst the red corpuscles of untreated rabbits placed under the same conditions, undergo a rapid solution. Tchistovitch confirming this fact has added to it the observation that the resistance of the red corpuscles of the rabbit is most often found when the serum loses its antitoxic power. If the receptors of the red corpuscles of immunised rabbits, owing to their great affinity for the haptophore group of the ichthyotoxin molecule, only attract the toxophore group of this poison, as the lightning-conductor when badly insulated attracts the lightning, the red corpuscles should never manifest resistance. To explain this contradiction we must not suppose that the red blood corpuscles which have become resistant have got rid of their receptors. In fact, if these receptors are so necessary to the nutrition of the cell that their absence has set up this extraordinary over-production which has inundated the fluids, it is evident that one cannot admit the existence of red corpuscles entirely deprived of corresponding receptors.

When examined from different points of view the hypothesis of the production of antihaemotoxin by the red corpuscles is surrounded with very great difficulties. It appears to be probable, therefore, that the source of this antitoxin must be sought for in other cell elements, and we may be allowed to recall to mind those cells which manifest a general and local reaction of the most constant kind after each injection of ichthyotoxin. Tchistovitch has observed that eel’s serum when introduced into rabbits in non-fatal but immunising doses excites a marked hyperleucocytosis.

The question of the origin of anticytotoxins being so complicated, it has been necessary for its elucidation to seek an experimental method of excluding the organ in which the antibody is supposed to have its origin. As we cannot think of eliminating the red or white corpuscles, nor the greater part of the tissues and organs, there remains only one way of bringing about this result. It is the suppression of the male genital organs. We know already that the injection of semen readily excites the production of a spermotoxin, and that this spermotoxin gives rise to the development of a corresponding antispermotoxin. If it is the spermatozoa, that is to say the elements having a particular affinity for the spermotoxin, which elaborate the antitoxin we must conclude that castrated males would be incapable of producing it. With this in view we have carried out a great number of experiments which have amply proved to us that male rabbits when deprived of their sexual organs are fully as capable of developing antispermotoxin in their fluids as are control rabbits in which the male genital apparatus remains intact. Doe-rabbits, and young, sexually immature rabbits of both sexes, also react to injections of spermotoxin by producing the corresponding antispermotoxin. The specific elements which are sensitive to the action of a cytotoxin undoubtedly are not indispensable for the development of the corresponding anticytotoxin. This result is in complete harmony with the hypothesis above put forward, that the red corpuscles cannot be regarded as the source of the antihaemotoxin. In the case of antispermotoxin this fact can be rigorously established by experiment.

Here arises the following question. We have seen that the anticytotoxins are composed of two different substances: an anticytase and an antifixative. The former is an antitoxin capable of neutralising macrocytase, the soluble ferment which will attack indifferently all kinds of cell elements. It is not to be wondered at, then, that the exclusion of the spermatozoa in no way prevents the production of anticytase by an organism which receives injections of cytotoxins. These latter, as we have already said, contain cytase along with the specific fixative; the macrocytase can attack any kind of animal cell provided that it can find some fixative or any other means to penetrate into the interior of these formed elements. We have seen that the antispermotoxin, obtained by Metalnikoff in guinea-pigs, does not contain any anticytase. Amongst his animals treated with spermotoxin was a castrated male guinea-pig which also produced anticytase. There is nothing astonishing in this fact, the injected cytase must have linked itself to many other cells which were able to develop anticytase.