399 Diphtheria Antitoxin .
By John Mellanby , M.D. , George Henry Lewes Student .
( Communicated by Professor J. N. Langley , F.R.S. Received June 3 , \#151 ; Read June 25 , 1908 .
) ( From the Welcome Physiological Research Laboratory and the Physiological Laboratory , Cambridge .
) CONTENTS .
PAGE Historical ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... 399 The Relation of Diphtheria Antitoxin to the Normal Proteins of Serum ... 400 The Relation between the Antitoxic Potency of a Serum and the Amount of Protein contained in it ... ... ... ... ... ... ... ... ... ... ... ... ... .
408 The Theory of Antitoxin Production ... ... ... ... ... ... ... ... ... ... ... ... 410 Summary ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... . .
412 Historical .
Many experimenters have endeavoured to isolate diphtheria antitoxin from fluids containing it .
In 1893 Brieger and Ehrlich ( 1 ) investigated the properties of milk obtained from animals which had been immunised against diphtheria toxin .
This milk contained an appreciable amount of antitoxin which could be precipitated between 27- and 38-per-cent , ammonium sulphate .
By this method they were able to increase the antitoxic value of their fluid about five hundred times .
But in milk only two proteins are present which differ widely in their physical properties\#151 ; caseinogen and lact albumin\#151 ; and their results could not be extended to horse serum , which is the main source of diphtheria antitoxin .
Their experiments did not show that antitoxin is not a protein .
Since horse serum is the main source of diphtheria antitoxin , experimental work has been done chiefly on this fluid .
Aronson ( 2 ) stated that globulin precipitated from antitoxic serum by dialysis contained antitoxin .
But Dieudonne showed that globulin precipitated by carbon dioxide from antitoxic serum did not possess this property .
The protein precipitate obtained by both these workers corresponded to the globulin of serum as originally defined .
Dieudonne further showed that if Hammarsten 's extended definition of globulin were adopted , then diphtheria antitoxin was a globulin , since it could be precipitated from serum by saturation with magnesium sulphate .
Dr. J. Mellanby .
[ June 3 , Brodie ( 3 ) found that the same result could be obtained by half saturation with ammonium sulphate .
Belfanti and Carboni confirmed the results of Brodie and Dieudonne .
Freund and Sternberg ( 4 ) attempted the isolation of diphtheria antitoxin from serum by using 1-6-per-cent .
potash alum .
They stated that the filtrate obtained from serum after adding this salt , when treated with ammonium sulphate , gave a precipitate which contained the greater part of the antitoxin associated with only a small quantity of the original protein of serum .
Pick ( 5 ) attempted to differentiate the antitoxic portion of serum by fractional precipitation with ammonium sulphate .
Proscher first digested antitoxic serum with trypsin and then fractionated the resultant liquid by means of ammonium sulphate .
He stated that in this way he obtained a solution containing all the antitoxin free from protein .
The Relation of Diphtheria Antitoxin to the Normal Proteins of Horse Serum.* Three proteins possessing well-defined physical properties may be separated from horse serum .
These proteins are present in serum in the following quantities:\#151 ; In the investigation of the properties oi diphtheria antitoxin the first question to settle was whether it belonged to the protein groups ( a ) , ( h ) , or ( c ) .
The protein ( a)\#151 ; globulin\#151 ; was obtained by precipitating antitoxic serum with 10 volumes of water and neutralising with acetic acid .
The precipitate was adequately washed by decantation and filtration .
This precipitate when dissolved in sodium chloride did not possess antitoxic properties .
It may therefore be concluded that diphtheria antitoxin is not a globulin as rigidly defined .
This result confirms the work of Dieudonnd and Belfanti and Carboni .
The protein ( c)\#151 ; albumin ( ft)\#151 ; is probably that protein of serum which can be crystallised by means of ammonium sulphate and sulphuric acid .
Crystalline albumin was prepared from antitoxic horse serum by this method .
* The experiments detailed in the following pages are based upon results previously described in two papers : " The Physical Properties of Horse Serum , " ' Journ. Physiol. , ' vol. 35 , p. 473 , 1907 , and " The Precipitation of the Proteins of Horse Serum , " 'Journ .
Physiol. , ' vol. 36 , p. 288 , 1907 .
( a ) Globulin ... .
( b ) Albumin ( a ) About 3 per cent. ( c ) Albumin ( ft ) 1908 .
] Diphtheria Antitoxin .
The crystals obtained when dissolved in water did not possess antitoxic properties .
It is evident , therefore , that diphtheria antitoxin belongs to or is associated with that great group of albumins which compose about 85 per cent , of the total protein of serum .
For this class of albumin no method of physical differentiation has been obtained .
It was interesting , therefore , to determine whether a delicate biochemical test such as is involved in the standardisation of diphtheria antitoxin would permit of a further differentiation .
The Freezing of Serum .
Serum containing 400 units per cubic centimetre of diphtheria antitoxin was frozen in a long cylinder and allowed to melt gradually .
The resulting liquid was syphoned off in a series of layers .
Antitoxin tests were made on some of the fractions and the relation of the antitoxic value to the percentage of protein determined .
The following results were obtained :\#151 ; Amount of solid Original serum ... 9*4 in 100 c.c. Antitoxic value .
1*52 grammes 50 + 4-77 \#187 ; ... .
200 + 9-06 ) ) . . . .
350 + 1319 ) ) . . . .
450 + 17*08 5 ) ... .
700 + .
9*4 } } * * * .
400 The + sign above indicates that the animals survived the tests without any serious symptoms .
The results are conclusive , \#151 ; the concentration of antitoxin follows the concentration of the protein on approximately parallel lines .
The Heating of Serum .
The relation of diphtheria antitoxin to the proteins of serum coagulable by heat was determined .
Serum containing 400 units of antitoxin per cubic centimetre y was coagulated at a series of temperatures\#151 ; the coagulum at each temperature being filtered off before further heating .
The following Jesuits were obtained:\#151 ; Solids in each 100 c.c. Antitoxic value .
Original serum 400 Filtrate after heating to 66 ' C ... .
9*27 " 400 " " 77 ' C ... .
3*04 " 50 " 80 ' C ... .
2-6 50 Dr. J. Mellanby .
[ June 3 , Thus heating to 66 ' C. coagulated a little fibrinogen , but destroyed no antitoxin ; but heating to 77 ' C. removed the greater part of the coagulable protein and also the greater part of the antitoxin .
The fact that diphtheria antitoxin can be heated up to 66 ' C. without destruction is interesting .
Ferments and complements are stated to be destroyed when heated to 56 ' C. The greater stability of diphtheria antitoxin to heat differentiates it from these classes of bodies .
The Electrolysis of Antitoxic Serum .
When a constant current is passed into serum through zinc sulphate electrodes , a mass of insoluble protein accumulates at the anode and water comes out at the cathode , the relation between the water and insoluble protein being such that the concentration of solids in the electrolysed serum keeps constant .
The influence of this procedure on antitoxic serum was determined .
A constant current at a pressure of 100 volts was passed into antitoxic serum through zinc sulphate electrodes for three hours .
A considerable mass of insoluble protein accumulated at the anode and an equivalent quantity of water passed out at the cathode .
At the end of this time the antitoxic value of the remaining serum was determined .
This value was found to have remained constant ; the antitoxin had neither accumulated nor diminished in the remaining serum .
This fact is capable of only one explanation\#151 ; that diphtheria antitoxin possesses the same properties as the normal proteins of serum when tested by means of an electric current passed into it through zinc sulphate electrodes .
The protein mass which accumulated at the anode was also tested for antitoxin .
This precipitate consists of a zinc sulphate compound of protein .
The precipitate was dissolved in a minimal quantity of sodium hydroxide , and its antitoxic value determined .
It was found to have an antitoxic value proportional to the amount of protein dissolved .
This fact is interesting\#151 ; the combination of the protein with zinc sulphate , although causing an insoluble compound to be formed , had not produced any chemical change .
The resulting precipitate possessed the antitoxin value of thfe original protein when adequately dissolved .
The Precipitation of Diphtheria Antitoxin .
( A ) By Alcohol.\#151 ; The alcohol precipitation limits of horse serum containing 400 units of antitoxin in each cubic centimetre were determined .
The following results were obtained :\#151 ; 1908 .
] Diphtheria Antitoxin .
403 Percentage of Percentage of protein alcohol .
precipitated .
12-5 3-6 25-0 7-6 37-5 48-7 50-0 97 75-0 99 The greater bulk of the protein is seen to be precipitated between 25 and 50 per cent , alcohol .
Antitoxin tests were made on the filtrate after precipitating with definite percentages of alcohol .
It may be stated that alcohol in the dilutions used does not destroy toxin .
From the antitoxin tests it was found that no antitoxin was precipitated by 25 per cent , alcohol ; 25 per cent , was precipitated by 37'5 per cent , alcohol ; 30 per cent , by 40 per cent , alcohol ; and 66 per cent , by 44 per cent , alcohol .
From these figures it is legitimate to assume that antitoxin has the same precipitation limits as the general mass of protein of the particular serum used .
From the antitoxin tests on the alcohol precipitation of other sera it appeared that the greater part of the antitoxin was precipitated between 35 and 45 per cent , alcohol .
A point of some importance was to see whether the presence of antitoxin in serum had any general influence on the form of the alcohol precipitation curve .
The quantitative precipitations of values were therefore determined .
Antitoxin value four sera of different antitoxin Serum .
per c.c. Total solids in 100 c.c. A 150 9'6 grammes B 250 11*0 " C 600 9-2 " D 800 97 Percentage of alcohol .
Weight of protein precipitated from 100 c.c. of serum .
A. B. c. D. grammes .
grammes .
grammes .
grammes .
30 1-7 1-1 \#151 ; \#151 ; 33 2*53 2 *0 2*0 2-0 36 2 *86 3*4 31 3*3 39 3*8 4*6 4*1 4*7 42 4*83 6*0 5*2 5-7 45 6-7 8-3 6*3 6*9 48 7*96 9-92 7 *3 8-0 50 8*52 \#151 ; 7*7 8*3 VOL. LXXX.\#151 ; B. Dr. J. Mellanby .
[ June 3 , There is some evidence from the form of the curves ( which may be drawn from the above figures ) that the presence of diphtheria antitoxin does affect the precipitation values .
Thus the Curve A is slightly concave to the left , and Curve D is slightly convex to the left .
Curves B and C show features intermediate between A and D. The form of that part of the precipitation curve is affected in which the antitoxin is precipitated , and it may be seen that the degree of variation is roughly proportional to the amount of antitoxin present .
From this we may assume that although diphtheria antitoxin has the same alcohol precipitation properties as the great bulk of the proteins of normal serum , yet it is something added , and not merely a normal protein with an antitoxic group attached to it .
Alcohol as a Protein Coagulant .
An attempt was made to differentiate diphtheria antitoxin from the normal proteins of serum by means of alcohol as a protein coagulant .
In the case of diphtheria antitoxin it was found that although alcohol added to serum did not affect its antitoxic value so long as no protein was precipitated , yet the rate at which precipitated antitoxin was destroyed by alcohol was the same as that at which the normal proteins of serum were coagulated .
Therefore , so far as the coagulating action of alcohol was concerned , there was no evidence of any physical differentiation of diphtheria antitoxin from the proteins of normal serum .
Precipitation by Alcohol below the Critical Point .
The above results were obtained by the action of alcohol on diphtheria antitoxic serum at temperatures above the critical point .
The alcohol precipitation limits of diphtheria antitoxin from serum were obtained at temperatures below the critical point .
As a result of a series of alcohol precipitation experiments at 2 ' C. , it was found that the greater part of the antitoxin was precipitated with other proteins of serum between 16 and 28 per cent , alcohol .
These antitoxin limits , so far as the quantity of precipitated protein is concerned , are practically the same as those obtained at temperatures above the critical point .
The only difference between precipitation above and below the critical point was that in the former case a portion of the antitoxin wTas lost owing to the coagulating power of the alcohol , whilst in the latter case the total protein precipitate was soluble , and the whole of the antitoxin could be recovered .
( B ) Ammonium Sulphate.\#151 ; Brodie precipitated the antitoxin from horse 1908 .
] Diphtheria Antitoxin .
serum in four parts by the gradual addition of ammonium sulphate to half saturation .
He found no accumulation of antitoxin in any one fraction of protein precipitated , the result indicating that the antitoxin was not carried down in a purely mechanical way .
The limits for the precipitation of antitoxin by ammonium sulphate were determined .
It was found that the initial precipitation of the antitoxin was practically coincident with the first precipitation of the proteins of serum , and was complete only after about 22 per cent , of ammonium sulphate had been added .
If these points be examined on the ammonium sulphate precipitation curves , it will be found that they are practically coincident with the alcohol precipitation limits .
( C ) By Neutral Salts in the Presence Acids.\#151 ; The precipitation of antitoxin by salts in the presence of acids was examined ; 5 per cent , acetic acid and 0*25 per cent , hydrochloric acid were found to have no deleterious influence on diphtheria antitoxin at room temperatures .
Controls showed that these acids in dilutions corresponding to the antitoxic value of the serum used did not destroy toxin .
The precipitation of antitoxin by sodium chloride and O'25 per cent , hydrochloric acid started when about 10 per cent , of the total protein of the serum was precipitated , the first 10 per cent , of protein precipitate contained only a minimal amount of antitoxin .
During the rest of the protein precipitation the antitoxin came down in a quantity proportional to the amount of protein precipitate produced .
There was no indication of any differentiation of the antitoxin from the great bulk of the proteid precipitate .
( D ) The Mechanical Precipitation of Diphtheria Antitoxin from Serum.\#151 ; It has been stated above that the temperature at which antitoxin is destroyed differentiates it from ferments .
A few experiments were made to determine whether it possessed the property common to ferments of mechanical precipitation .
Various flocculent substances precipitated in serum did not diminish its antitoxic properties .
This absence of mechanical precipitation might have been due to the simple nature of the preeipitants , e.g. , calcium phosphate and kieselguhr .
A solution of colloidal ferric phosphate was therefore added to antitoxic serum .
The colloidal ferric phosphate was precipitated by the ionised salts present in the serum .
After it had settled to the bottom of the flask as a gelatinous mass , the antitoxic value of the remaining serum was determined .
There was no diminution in its antitoxic properties .
From these experiments it may be concluded that diphtheria antitoxin does not possess the property of associating itself with substances precipitated in a solution of it .
Dr. J. Mellanby .
[ June 3 , The Action of Ferments on Diphtheria Antitoxin .
02 per cent , hydrochloric acid has been stated above to have no destructive action on diphtheria antitoxin .
But when serum is acidified with 0'2 per cent , hydrochloric acid and a quantity of pepsin is added , it is found that the antitoxic value diminishes at a rapid rate .
This destruction of diphtheria antitoxin by pepsin in vitro is in line with the inability to immunise animals against a subcutaneous injection of diphtheria toxin by feeding with antitoxic serum .
Trypsin also destroys diphtheria antitoxin , but much more slowly than pepsin .
With regard to the statement of Proscher concerning the preparation of protein-free antitoxin by digesting antitoxic serum with trypsin , the following experiment is of interest .
400 c.c. of trypsin ( Liquor pancreaticus Benger ) were added to 2000 c.c. of serum and incubated at 37 ' C. After four and nine days respectively the quantitative precipitation of this digested mixture by ammonium sulphate was determined .
The following results were obtained:\#151 ; After four days the amount of solid not precipitated by saturation with ammonium sulphate was 10 per cent , of the total solid ; after nine days the quantity of solid not precipitated had increased to 24 per cent , of the total solid .
The 14 per cent , increase represented digested protein .
The quantitative results for the precipitation of the proteins of serum after four and nine days ' tryptic digestion were determined .
Percentage of Am.2S04 ( solid ) .
Percentage of protein not precipitated .
After 4 days .
After 9 days .
18-75 82 *4 98-2 20-0 74 *5 91 -6 21-25 \#151 ; 77 -0 22 -5 55 *8 70-0 25 -0 42 -0 56 -2 27-5 34 -0 46-6 30 -0 26 *6 40-7 35-0 20 -0 \#151 ; 40-0 \#151 ; 31 -7 There are two results to be noted from these figures:\#151 ; ( 1 ) The action of trypsin on the proteins of serum is slow , but progressive .
( 2 ) The diminution of the amount of protein precipitated by ammonium sulphate after tryptic digestion is constant for all quantities of ammonium sulphate .
Thus , in the above figures the average difference between the amount of protein left in solution by any amount of ammonium sulphate 1908 .
] Diphtheria Antitoxin .
after four and nine days ' tryptic digestion is 15 per cent. The precipitation limits of the digested protein only are altered .
The first result indicates that the slow action of trypsin on the proteins of serum is due to the resistance of this type of protein to tryptic digestion rather than to the presence of an antibody to trypsin in the serum .
The second result emphasises the need for quantitative experiment when precipitating digested mixture of protein with ammonium sulphate , more especially if conclusions as to the antitoxic value of the precipitate are required .
The rate of digestion of the protein as given by the above figures is approximately the rate at which the antitoxin is destroyed by trypsin .
CoTiclusion .
The results of all the above experiments on diphtheria antitoxin may be briefly stated:\#151 ; ( a ) Whenever the main protein of serum is precipitated , diphtheria antitoxin is precipitated .
( b ) Whenever the main protein of serum is destroyed or coagulated , diphtheria antitoxin is destroyed .
These results have only one logical explanation\#151 ; that diphtheria antitoxin is a protein .
There is a possibility that diphtheria antitoxin is a comparatively simple body , and is attached to the protein only by virtue of some chemical affinity .
But in the above experimental work a large number of methods of analysis have been used , and it is reasonable to suppose that if the hypothesis above were correct , some indication would have been obtained of a splitting of the antitoxin from its accompanying protein .
No such indication has been obtained .
If we assume , then , that the diphtheria antitoxin is a protein , what hope is there of isolating it in a pure condition ?
Serum doubtless contains a very large number of proteins which possess different physiological properties .
But a living cell is required to appreciate these physiological differences .
In laboratory work the only available forces are those of a physical and chemical nature .
With these forces it is possible to differentiate the proteids of horse serum into three groups :\#151 ; ( a ) Globulins comprising about 3 per cent , of the total protein .
( b ) Albumins which cannot be crystallised .
These compose about 85 per cent , of the total proteins .
( c ) Albumins which may be crystallised .
These compose about 12 per cent. of the total protein .
408 Dr. J. Mellanby .
[ June 3 , Diphtheria antitoxin belongs to the class ( b ) , composing about 85 per eent .
of the total proteins of serum .
The similarity of diphtheria antitoxin to the main mass of the protein of serum raises the question as to the function of these proteins .
It is generally assumed that the main function of the proteins of serum is to nourish the tissues .
From a general consideration this is improbable .
Foodstuffs are divided into three great classes :\#151 ; proteins , fats , and carbohydrates .
Sugar is present in blood to the extent of about 0-2 per cent. , and usually only a trace of fat can be detected in it .
It is improbable , therefore , that proteins to the extent of 10 per cent , are circulating in the blood to be used as food .
From a study of the processes involved in the coagulation of blood it is fairly clear that globulin originates from a splitting of the fibrinogen molecule into fibrin and globulin when coagulation occurs .
Albumen ( / 3 ) composing 12 per cent , of the total protein may possibly be regarded as concerned in the nutrition of the tissues .
Since diphtheria antitoxin belongs to the protein group composing the remaining 85 per cent , of protein , it may be legitimate to assume that a function of these proteins is to protect the tissues from disease in addition to maintaining the equilibrium of electrolytes in the tissues , and holding carbon dioxide in the blood .
The assumption of such a function would afford an explanation why large injections of normal horse serum after control diseases caused by bacterial infections .
The Eelation between the Antitoxic Potency of a Serum and the Amount of Protein contained in it .
The relation between the antitoxic power of a serum and the solids contained in it has been investigated by various workers .
Szontagh and Wellmann ( 6 ) found only a small increase in the total solid present in serum after immunisation against diphtheria toxin .
Hiss and Atkinson ( 7 ) found a rise of globulin content during the process of immunisation and a corresponding fall in the albumen content .
The globulin was determined by precipitation with magnesium sulphate .
Ledingham ( 8 ) , using half-saturation with ammonium sulphate as a precipitant , also found an increase of globulin content during immunisation .
The work of Hiss and Atkinson , and Ledingham , was based on the assumption that diphtheria antitoxin was a globulin as defined by its precipitation limits with ammonium sulphate and magnesium sulphate .
But from a consideration of the influence of acids and alkalis on the precipitation of the proteins of serum by neutral salts , it is clear that their results could be 1908 .
] Diphtheria Antitoxin .
adequately explained by an assumption of varying alkalinity of the serum during immunisation .
It has been shown in the previous pages that diphtheria antitoxin possesses the same physical properties as nine-tenths of the proteins of serum , and that this antitoxic protein is probably added to serum during the process of immunisation .
The question arises as to whether the theory of the addition of antitoxic protein to the blood during the process of immunisation receives any support from a consideration of the total solids in the blood of a horse during its antitoxic life .
A detailed examination of the quantity of protein in the sera of various horses showed that there was no relation between the total proteins in the blood of a horse and its capacity to produce antitoxin .
From the antitoxic point of view we may consider the proteins present in the serum of an animal as concerned in two great functions : ( ) The maintenance of tissue equilibrium ; ( b ) the protection of the tissues .
Of these two divisions ( a ) is the more active component , and would be more evidently reflected in the amount of protein contained in the serum .
The absence of any relation between the antitoxic producing powers of different animals and their general metabolism reflects the experience of individuals\#151 ; that a strong physical condition does not imply a special capacity to resist disease .
But the question whether the percentage of solids in the blood of a horse at different times of its antitoxic life bears any relation to the quantity of antitoxin in it is of a different nature .
In this case it appears reasonable to assume that , so long as we do not interfere with the general nutrition of the animal by poisoning with excess of toxin , the percentage of solid in the serum will increase with its antitoxic power .
To test the accuracy of this assumption , the percentages of solid in the sera of a number of horses at different periods of their antitoxic life were determined .
In the following table ( p. 410 ) the days are reckoned from the commencement of immunisation .
From the results we may deduce two generalisations :\#151 ; ( a ) That at the start of diphtheria toxin immunisation the percentage of solid increases with the antitoxic power .
( b ) That after a time , depending on the particular horse , there is no connection between the percentage of solid in the serum and its antitoxic power .
This stage is marked by a large decline in the antitoxic producing power of the horse .
These two generalisations are readily appreciated if we consider the protein in the serum of a horse to reflect the two functions stated above : ( i ) the .
maintenance of tissue equilibrium ; ( ii ) the protection of the tissues .
The Dr. J. Mellanby , [ June 3 , Days .
Antitoxic value .
Percentage of solids in serum .
Days* Antitoxic value .
Percentage of solids in serum .
Horse A. Horse B. 0 0 8*93 0 0 7-78 60 225 \#151 ; 97 350 9-02 104 500 10-91 148 1000 10 -54 146 675 11 -98 222 450 9-54 190 575 12-10 266 250 10 -6 235 800 11 -35 313 500 10 -4 254 450 9-88 355 363 10 -3 302 700 11 -53 395 250 10 -i 348 700 11 -52 443 200 10 -77 397 800 12-11 506 200 10 -76 460 575 12 -0 496 900 12 -1 Horse C. 533 500 11 -76 0 0 8 -2 574 800 10 -75 54 250 8 -95 624 250 11 -18 99 660 9 -6 700 333 10 81 143 333 9-04 750 363 10 -85 192 363 9 -43 810 250 11-88 241 571 9-65 292 400 9-8 341 500 9-4 decline of the antitoxin producing power implies the exhaustion of that mechanism by means of which antitoxin is produced .
That this is not reflected in the amount of solid in the serum depends upon the fact that the exhaustion of this special mechanism does not imply a corresponding effect on the general nutrition of the tissues .
Ehrlich 's Theory of Antitoxin Production .
The salient point in the production of diphtheria antitoxin is that toxin in increasing doses is injected under the skin of a suitable animal , and .
, after a variable time , antitoxin appears in the blood .
The explanation of these phenomena which has been most widely accepted is that of Ehrlich .
This theory may be stated in general terms as follows:\#151 ; The toxin combines with the side chain of a cell and so upsets the equilibrium of that cell .
By virtue of its protective mechanism the cell manufactures a series of newT side chains similar to that attacked ; but these side chains are manufactured in greater quantity than is required to neutralise the toxin , and those in excess are cast off and appear in the blood as antitoxin .
This theory demands\#151 ; ( a ) That the tissues which are concerned in the production of antitoxin are those tissues which the toxin attacks .
( b)That antitoxin is produced in tissues remote from the seat of inoculation .
1908 .
] Diphtheria Antitoxin .
( a ) From a study of the lesions produced by diphtheria toxin , it is clear that the main tissues attacked by this poison are the heart and nervous system .
Therefore , on Ehrlich 's theory , these tissues are mainly concerned in the production of antitoxin .
From a general consideration of the metabolism of an animal it is improbable that two master tissues as the heart and nervous system are responsible for the protection of it .
A horse had been immunised against diphtheria toxin for five years .
At the end of this time it was killed , although it was suffering from no obvious disease .
During its antitoxic life it had produced about 200 litres of antitoxic serum , containing , on an average , 500 units of antitoxin in each cubic centimetre , and had been injected with an appropriate amount of diphtheria toxin .
A detailed post-mortem examination of the various organs was made .
There were no marked pathological changes to be noted\#151 ; the heart showed a little brown atrophy , but no changes were found in the nervous system .
It is inconceivable that the heart and nervous system could have been attacked by such enormous quantities of toxin and have produced such a large amount of antitoxin without being more seriously affected .
A consideration of the amount of antitoxin produced by a horse after the injection of a definite quantity of diphtheria toxin is of interest .
During a period of four weeks 800 priifungs doses of toxin were subcutaneously injected into a horse .
At the end of that time the blood contained 400 units of antitoxin per cubic centimetre .
Suppose the horse contained 50 litres of blood .
At the end of four weeks the blood of the animal contained 1000 x 400 x 50 = 2,000,000 units of antitoxin , or 800 priifungs doses of toxin produced 2,000,000 imits of antitoxin ; or each priifungs dose of toxin produced 25,000 units of antitoxin .
If this be interpreted in terms of Ehrlich 's hypothesis , it means that when the side chain of a cell is attacked by a molecule of toxin it manufactures more than 25,000 new side chains to protect itself .
And if Ehrlich 's hypothesis be correct , the cells which are concerned in this extraordinary production are contained in those tissues which perform the main functions of the body\#151 ; namely , in the heart and nervous system .
( b ) The production of antitoxin in a tissue remote from the site of injection of the toxin demands the assumption of other specialised mechanisms .
Let us consider the case of a horse towards the end of its immunisation period .
At the last injection , probably , an amount of toxin containing 200 priifungs doses is injected , and at this stage the blood of the animal contains 400 units of antitoxin in each cubic centimetre .
This toxin , before it can produce any effect , must travel in the lymph stream to the thoracic .
duct , and so via the jugular vein into the general circulation .
When once in VOL. lxxx.\#151 ; b. 2 K Dr. J. Mellanby .
[ June 3 , the blood it will be ultimately carried into the neighbourhood of any tissue .
But before it can attack a special tissue it must first pass through the capillary walls of the blood-vessels into the lymph bathing the cells of that tissue , and such a passage demands a specialised mechanism of chemiotaxis .
Further , from the union of toxin and antitoxin in vitro , it is probable that the toxin would unite with the antitoxin as soon as it got into the blood stream , and the hypothesis of the formation of antitoxin in a distant tissue would demand that this toxin-antitoxin compound should be broken up by the cell before it could stimulate the production of more side chains\#151 ; in fact , that the formation of antitoxin by a cell should be of no use in protecting it from subsequent inoculations of toxin .
From a general consideration of diphtheria toxin immunisation it is probable that the leucocytes are the immediate agents concerned in the production of antitoxin .
After a subcutaneous injection of diphtheria toxin into a horse there is usually produced at the site of inoculation a swelling of varying size .
This swelling is of an oedematous character , and is crowded with finely granular oxyphil cells .
The size of the swelling gives , to some extent , an indication of the degree of antitoxic reaction .
The production of antitoxin by the leucocytes at the seat of inoculation would bring the mechanism involved into line with the facts observed by Metchnikow with pathological bacilli and the more recent work on opsonins .
The antitoxin need not necessarily be produced at the seat of inoculation .
After a leucocyte had ingested a toxin molecule the secretion of antitoxin by it would take place into any fluid in which it was present\#151 ; the blood or lymph .
This hypothesis affords a ready explanation of poisoning by excess of toxin .
If the leucocytes are able to take up the toxin , none of it is carried to distant tissues , such as the heart and nervous system , and no pathological effects are observed .
But if the toxin is injected in too great a quantity to be taken up by the leucocytes , then the excess is carried by the blood to the tissues , and general toxaemia results .
Again , Ehrlich 's hypothesis offers no explanation , why subcutaneous injection of toxin yields much better antitoxic results than injection into the blood stream direct .
The theory that antitoxin is secreted by leucocytes which have ingested toxin molecules affords a ready explanation of these results .
Summary .
All the properties of diphtheria antitoxic serum indicate that diphtheria antitoxin is a protein which possesses characters identical with albumin ( a ) of serum , i.e. , the protein forming 85 per cent , of the total quantity present .
The alcohol precipitation curves of antitoxic sera of varying strength 1908 .
] Diphtheria Antitoxin .
indicate that this protein is added to serum during immunisation , and that the antitoxic property is not due to the addition of a group possessing this character to a normal protein molecule .
This conclusion also follows from a study of the relation between total solid and antitoxic strength of the serum of a horse during its antitoxic life .
It is suggested that one of the functions of the proteins composing the group albumin ( a ) is to protect the tissues against disease .
A theory is advanced that the production of diphtheria antitoxin is due to an active secretion by the leucocytes , this secretion being stimulated by the ingestion of toxin molecules .
REFERENCES .
( 1 ) Brieger and Ehrlich , ' Zeitschr .
f. Hygiene , ' vol. 12 , p. 137 .
( 2 ) Aronson , 'Berl .
klin .
Wochenschr .
, ' vol. 26 , p. 425 .
( 3 ) Brodie , ' Journ. of Pathol , and Bacteriol .
, ' vol. 4 , p. 460 .
( 4 ) Freund and Sternberg , 'Zeitschr .
f. Hyg .
, ' vol. 31 , p. 429 .
( 5 ) Pick , 'Hofmeister 's Beitr .
, ' vol. 1 .
( 6 ) Szontagh and Wellmann , 'Deutsche med .
Wochenschr .
, ' vol. 24 , p. 421 .
( 7 ) Hiss and Atkinson , ' Journ. of Exp. Med. , ' vol. 5 , p. 47 .
( 8 ) Ledingham , 'Journ .
of Hygiene , ' vol. 7 , p , 65 .

