CHEMO-IMMUNOLOGICAL STUDIES ON CONJUGATED CARBOHYDRATE-PROTEINS III. ACTIVE AND PASSIVE AN- WITH %NTH3ZTIC UGAR- PROTEINS I/' BY WIT&TAM S. TIUETT, M.D., OSWALD T. AVERY, M.D.,. AND WALTDER F. GOEBEL, PH.D. (From tie Bos$itd of The Rockefeller Institute fur Medical Research) (Received for publication, July l&1929) In the two preceding papirs of this series Avery and Goebel (1,2) have reported the results of chemo-immunological studies on con- jugated carbohydrate-proteins. In the first of these communications Goebel and Avery (1) described the chemi- cal methods by means of `which the p-aminophenol ghxosides were synthesized from glucose and galactose and the dia&aed substances bound to protein. The two newly synthesized sugar-proteins d&r from one another chemically only in specific rotation and molecular co~tion. In the second paper, Avery and Goebel(2) reported the antigenic properties and serological specifkity of the con- jugated carbohydrates. They found that the glucosidic radical of the compound antigen endowed the new complex with specific reactivity, This was demon- strated in two way& Fit, .when the same glucoside is attached to two d@r& proteins, as serum' globulin or egg albumin, the serum prepared by immuniaation with either one of the antigens is spechically reactive with the other. Second, when the two d$rtwt glucosides,-glucoside and gaktosideke each com- bmed with the iawe protein, the newly formed compounds are serologically dis- tinct. These facts were found to be true even though the individual glucosides are isomers differing only in the spatial con&ration of a single carbon atom. On the other hand the kombiued glucosides alone were found to be non- antigenic. They failed to induce antibody formation in the anhnal body and caused no visible precipitation. when added, to immune sera +c vitro. However, both glucosides possess the capacity of specifically inhibiting the precipitating action of homologous antisugar?protein serum. When galactoside was mired with serum prepared by immuniaation with galacto-globulin,l the subsequent addition l The terms'employed in this paRer to represent the sugar-protein compounds are the same as those .used by Avery and Goebel(2). 551 552 ' CONJUGATED CARBOHYDRATE-PROTEINS. III of gala&-albumin to the mixture did not cause precipitation. Furthermore, when the heterologous glucoside was substituted in the same system, no inhibition of the precipitin reaction occurred. These results demonstrate the specificity of the inhibition phenomenon. Avery and Goebel (2) considered that the uncombined glucosides possess the immunological properties of haptens. The r81e of carbohydrate in auaphylaxis has been a subject of recent experimental investigation. Tomcsik (3) working with B. lactis aerogenes, and later Tomcsik and Kurotch- kin (4) employing B. lactis aerogenes, the pneumobacihus, and a yeast, isolated carbohydrate substances which produced auaphylactic shock in guinea pigs passively sensitized with homologous immune serum. Lancefield (5) also ob- tained from streptococci carbohydrate material with which anaphylaxis could be induced in guinea pigs passively sensitized with antM.reptococcus serum. Be- cause of the presence of small amounts of nitrogen in the products, none of these authors felt justified in concluding that the carbohydrate alone was responsible for the shock. Avery and Tiiett (6) employing the highly purified polysaccharide of the type-specific pneumococci showed that guinea pigs passively sensitized with homologous anti-pneumococcus rabbit serum were thrown into anaphylactic shock by the subsequent injection of the homologous specific carbohydrate. Guinea pigs could not, however, be actively sensitized with the purified poly- saccharides alone. Since the materials used in those experiments were protein- free, and in the csse of the Type II and Type III substances also nitrogen free, the results conclusively demonstrate the capacity of complex;sugars to induce anaphylactic shock in animals passively sensitized with antibacterial sera. . The immunological specificity of pneumococcus polysaccharides has a close anaIogue in the serological specificity exhibited by gluco- protein and galacto-protein. The immunologic behavior of the syu- thesized sugar-proteins led to their use in sensitization experiments. The results, reported in this paper, on anaphylaxis with artifkially prepared carbohydrate-proteins confkn and extend the serological tindings previously reported. The production of both active and passive anaphylaxis was attempted in order to determine the sensitiz- ing properties of the synthetic antigens and to demonstrate the speci- ficity of the reactions. Gluco-globulin represents phenol &glucoside-azo-globulin. Gluco-albumin represents phenol 8-glucoside-azo-albumin. Galacto-globulin represents phenol fi-galactoside-azo-globuliri. Galacto-albumin represents phenol &galactoside-aso-albumin. The globulin was prepared from horse serum, and the albumin from egg white. W. S. TILLETT, 0. T. AVERY, AND W. F. GOEBEL 553 EXPERIMENTAL Guinea pigs weighing 240 to 275 grams were employed. The sensitiaiig dose, whether serum or sugar-protein, was always injected hrtraperitoneally. The &ding dose was unifody injected intravenously into a superficial vein of the hind leg. For details concerning the chemical procedures involved in the synthesis of the materials, the reader is referred to the article by Goebel and Avery (1). The serological characteristics of the serum employed and the method of prepara- tion are described by Avery and Goebel(2). Pas&e Sensitizatio92 A. Results obtained with sugar-pro&h compounds Eight guinea pigs were injected intraperitoneally with the pooled serum of three rabbits immunized with gluco-globulin. Five guhxa pigs each received 5 cc. of serum, one received 3 cc., one received 1 cc., and one 0.5 cc. As previously men- tioned) serum of this character possesses the capacity to precipitate complex anti- gens composed of heterologous protein conjugated with the homologous glucoside. The antigluco-globulin sera used in these experiments had an average titre of specific precipitins, as determined by tests made with gluco-egg-albumin, of 1 to 80,000. By reason of the fact that horse globulin alone when used as antigen does not elicit antibodies reactive with egg-albumin, the high precipitm titre of these sera is obviously dependent upon the conjugated glucoside radical. Twenty-four hours after the administration of gluco-globulin antiserutqeach pig received intravenously 1 cc. of gluco-albumin. From Table I it can be seen that the five pigs sensitized with 5 cc. of serum all died with typical symptoms of anaphylactic shock. The animals passively sensitized with 3 cc. and 1 cc., respectively, of antigluco-globulin sennn had definite and typical symptoms imme- diately following the injection of 1 cc. of gluco-albumin but recovered. Pig No. 6, which received 0.5 cc. of serum, exhibited only a slight reaction. A similar experiment was carried out using antigalacto-globulin squm for sensitization and galacto-albumin as the toxigenic antigen. Sera obtained from three rabbits immuniaed with gab&o-globulin were pooled; the precipitin titre, as determined with galacto-albumin, averaged 1 to 80,000. Eight guinea pigs were injected intraperitoneally as follows: 5 animals received 5 cc. of serum each, one received 3 cc., one received 1 cc., and one, 0.5 cc. The results given in Table II are equally as definite as those shown in Table I. All 554 CONJUGATED CARBOHYDRATE-PROTEINS. III 4 TABLE I Passive ~naphylaxis with Anti-GElu;o-Glob@~ Serum Reactiahs induced by the use of the homologous glucoside conjugated with a deterologous protein-Gluco-Albunk Guinea pirr No. 5 6 i Sameanimal brs. later 8 -' Same animal 4 hrs. later ~!z!ir serum i.p. CG. 5 5 5 3 1 0.5 5 5 hrs. 24 24 24 24 1 24 24 24 24 ihm5? . . `$g/fg 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. laJJ$S 1 cc. ~~ -0Ti-J Marked scratcb- ing, bucking, coughing, ret@- story distress Violent typical Symptoms 0txasi0na1 scRatchillg t3f Itlhltes t31 minute3 t2) minutes Definite symp- toms followed by recovery Definite symp toms followed by recovery No reaction None No reaction Typical t4 minutes None No reaction Typical / I t4+ minutes t Death of animal. i.p. = intraperitoneal. i.v. = intravenous. W. S. TILLETT, 0. T. AVERY, AND W. F. GOEBEL 555 pigs sen&ized with homologous immune serum in amounts from 1 to 5 cc. reacted typically and fatally to the intravpous injection of 1 cc. of galacto-albumin; 0.5 CC. of serum was indicient to sensitize. a TABLE II Passive Amphylak with Ad-GUzcto-GZvb Serum Reactions induced by the use of the homologous galactoside combined with a Guinea pis No. 1 2 `3 4 5 6 7 SamCanimdl hr. later 8 5 24 Sameanima.ll hr. later h&rologous 1 cc. 5 5 5 3 1 0.5 5 h. 24 24 24 24 24 24 24 protein-Gal&t = "9 . . `p&y 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. IL-g- 1 cc. `dado-tsg- dbmir 1 cc. d$f+@&s- 1 cc. Ey?g- 1 cc. Result gz ~~ Typical slight scratching and coughing t3 minutes t8 minutes t3+ minutes t3 minutes t3) minutes Very mild reac- tion with re- covery None No reaction Typical t31 minutes None No reaction p mhlutt?s tDeath of animal. i.p. = intraperitoneal. i.v. = intravenous. In Tables I apd II it is also &own that the reactions in guinea pigs induced with sugar-proteins a&l anti-sera are strictly specific. 556 COVJUGATED CAXBOEIYDFL~TE-PROTEINS. In Animals No. 7 and No. 8 of Table I received antigluco-globulin serum and 24 hours later were injected intravenously with galacto-albumin. No reaction occurred. Four hours later, the introduction of gluco-albumin induced typical fatal shock. Similarly pigs No. 7 and No. 8 of Table II, sensitized with anti- gslacto-globulin serum, were unharmed by gluco-albumin; the subsequent ad- ministration of the homologous galacto-albumin antigen caused anaphylactic death of these animals. The animals also serve to demonstrate the fact that gluco- albumin and galacto-albumin are not primarily toxic. B. Results obtaiwd with uncombined glucosides The specific inhibitory effect exerted by the glucosides on tb precipitin reaction of sugar-protein and anti-sera has been describe& in detail by Avery and Goebel'(2) and has been previously commented upon in this paper. It, therefore, seemed of interest to determine whether anaphylactic shock could be elicited by glucosides alone in passively sensitized guinea pigs, and if not, whether the inhibition which these substtinces have on the precipitin test would also be evident in the anaphylactic reaction. As shown in Table III, two pigs (Nos. 1 and 2), sensitized 24 houm.previously with antigluco-globulin serum, were injected intravenously with 1 cc. of the un- combined homologous glucoside. No reaction occurred. Two hours later 1 cc. of gluco-albumin injected into the same animal caused prompt anaphylactic death. In guinea pigs Nos. 3,4, and 5, the introduction of glucoside alone was followed immediately by an injection of gluco-albumin. Except for slight scratching, no response was elicited. From these results it may be seen that the injection of glucoside into a sensitized animal exerts a definite but transitory protection against the shocking capacity of material which otherwise would be fatal. That the protective action of the glucoside is specific is demon- strated by guinea pigs Nos. 6, 7, and 8. These animals were injected with the heterologous galactoside; when, immediately thereafter they were given gluco-albumin no protection resulted and they died promptly with typical anaphylactic shock. Table IV presents the results obtained with guinea pigs, which, after having been sensitized with antigalacto-globulin serum, were protected by the galactoside from the toxigenic effect of galacto- albumin. Guines Pig No. Interval hti- between lUCO- d injection obulin of Serum acrum and ma- i.p. phylactk test -- G. h. s 24 24 ; 24 !I 24 5 24 -- s 24 5 24 5 24 TABLE III Passive Anafihylads with Anti-Gluco-Globdin Serum Effect of Uncombined Glucoside 3y$Yf . . gfwcosi;ir 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. ga&dosfde 1 cc. 1 cc. 1 cc. Result No reaction No reaction No reaction No reaction No reaction `No reaction No reaction No reaction xaterpal betwcon' injection of &os& and sugar- protein hn. 2 2 Followed inune- diately by Followed bnrne- diately by Followed imroe- diat&ly by Follow@ imme- diately by Followed inune- diatoly by Followed imme- diately by rf+J: 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. symp- Scratches slightly, no other symptoms Scritches slightly, no other symptoms Scratches slightly, no other symptoms Typical Typical Severe symptoms, falh on side. Apn-, on feet Result t4 minutes. 14 minutes No reaction No reaction No reaction t7) minutes t3) minutes Severe shock followed by recovery t De& of auimal. i.p. = intqeritoneal. i.v. = intravenous. Allti- Guinea galacto- Pll No. elobulin sylm `.h -- cc. 7' 5 8 5 Passive Anaphylaxis vith A~i-Gdacto-Gl&&n Scwm Effect of Uncombined Galactoside hts. &tur0~~ 24 1 cc. 24 1 cc. 24 1 cc. 24 1 cc. 24 1 cc. 24 1 cc. -- glum* 24 1 cc. 24 1 cc. Result No reaction No reaction No reaction No reaction No reaction No reaction No reaction No reaction Interval between injcctlon of slucoEfdc and sugar-protein h. 2 4 23 Followed immedi- ately by Followed immedi- ately by Followed immedi- ately by. Followed immedi- ately by Followed immedi- ately by `*F 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. 1 cc. SymPto=J Marked scratching, wughing, baclc- ing, respiratory distress Typical Typical None Typical w-1 Resulr . 8 4 Definite symptoms 8 followed by re- c) covery is t4 minutes 0 t31 minutes No reaction 8 No reaction No reaction g 8 3 . f2f minutes Ef t3mitultes t Death of animal. i.p. = intraperitoneal. i.v. = intravenous. TABLE V Desmsitithon Induced in Passively Sensitized Cuinecr Pigs by Injedim oi Glucoside and Homologous Sugar-Prokin .- Guiaea Pi No. . 1 2 3 5 cc. anti-gluco- , globulin serum 5 cc. anti-gluco- globulin serum 5 cc. anti-gluco- globulin serum 5 cc. anti-galacto- globulin serum , 5 cc. anti-galacto- globulin serum 5 cc. anti-galacta- globulin serum i.p. - intraperitoneal. i.v. = intravenous. prevhs teats. Injectiona made 24 ha. after 7? . . 1 cc. glucoside followed im- mediately by 1 cc. gluco- egg-albumin 1 cc. glucoside followed im- mediately by 1 cc. gluco- egg-albumin 1 cc. glucoside followed fm- mediately by 1 cc. gluco- egg-albumin 1 cc. galactoside followed lmmediate~y by 1 cc. galac~egg-albumin 1 cc. gahctoside followed immediately by 1 cc. gala&-egg-albumin 1 cc. galactide followed immediately by 1 cc. galacto-egg-albumin Reaalt No reaction No reaction No reaction No reaction No reaction No reaction 1 cc. gluco-egg- albumfn 1 cc. gluco-egg- albumin 1 cc. gluco-egg- albumin 1 cc. gala&o-egg albumin 1 cc. `&W-egg- albumin 1 cc. gelactQ-@g- albumin None No reaction None No reaction None No reaction None No reaction None No reaction None No reaction Result 4 " i P 560 * CONJUGATED CARJ3OBXDRATE-PROTEW. m The results are, in every respect, identical with those given in Table III both with regard to the transitory nature of the phenomenon and to its specificity. The mechanism of the protection afforded by the glucosides is not as yet understood. The fact that the protective effect is no longer demonstrable after two hours indicates that "desensitization,"-if such has occurred-is transitory. Instances of what appears to be true desensitization have been observed in these experiments and are recorded in Table V. Guinea pigs Nos. 1, 2, and 3, passively sensitized with antigluco-globulin serum, were protected from the shocking e&t of 1 cc. of gluco-albumin by a pre- vious injection of homQlogous glucoside. Four and one-half hours later the same pigs received a second injection of 1 cc. of &co-albumin. No reaction occurred. The absence of shock following the second injection of khole antigen seems to be dependent upon the first dose of gluco-albumin. That the glucoside alone plays no direct part in the refr~tory state is demonstrated by its ineffectiveness when in- jected singly, two hours prior to the shocking dose (Table III). Pigs Nos. 4,5, and 6 of Table V demonstrate the same principle, the difference being that antigalacto- globnlin serum was used for sensitization, and galactoside and galacto-albumin were empIoyed to complete the test. Avery and Goebel(2) have shown that the serum of rabbits immu- nized with synthetic sugar-proteins (gluco- or galacto-globulin) pos- sesses two distinct antibodies; 1) the specifk precipitin so intimately associated with the carbohydrate radical of the compound; 2) the "common" precipitin, reactive with globulin alone. Passive ana- phylaxis experiments were, therefore, carried out, using pure horse globulin as the toxigenic material. For sensitization, one pig received intraperitoneally 1 cc. of senmr prepared by immunization with gluco-globulin; a second pig received 1 cc. of serum derived from a rabbit immunized with galacto-globulin. Twenty-four hours later each received 12 mgms. of globulin. Both animals diedin typical anaphylactic shock (Table VI). A normal pig receiving the same dose of globulin gave no reaction. W. S. TILLETT, 0. T. AVERY, AND W. F. GOEBEL 561 TABLE VI Passive Ana#tylaxis with A&Glum-Globulin ad Atrti-Cahcto-Globuh SW~ Reactions induced by the use of horse globulin Guinea pig No. 1 1 cc. anti-gluco- globulin serum 2 1 cc. anti-galacto- globulin serum 3 Normal control Time interval .- h. 24 24 - Sho$iy dose . . 1 cc. horse-globulin (12 mpms.) 1 cc. horse-globulin 1 cc. horse-globulin Typical Typical None Result t3 minutes t2* minutes No reaction t Death of animal. i.p. = intraperitoneal, i.v. = intravenous. Active Amzphylaxis For purposes of testing the capacity of the synthetic sugar-proteins to produce active sensitization, 10 guinea pigs were injected intra- peritoneally with 5 cc. of glum-globulin and 10 other animals were similarly inoculated with 5 cc. of galacto-globulin. In the preparations empIoyed, 5 cc. of sugar-globulin contained 50 mgms. of protein. It is estimated (1) that 15 per cent by weight of this complex represents chemically combined giucoside. Consequently each pig received approximately 7.5 mgms. of the synthesized sugar-protein. The 20 pigs were tested 21 days later for active sensitization. Asin the experiments on passive anaphyiaxis, the sensi- tizing dose consisted of materiai in which the giucoside was joined to a protein heterologous to that used for sensitiaation. As previously mentioned, this pre- caution was taken in order to ehminate the possibility of protein-antiprotein reactions entering into the ~izsults. In Table VII, the results of active sensitization obtained by the use of gluco-proteins are given. PigsNos. 1, and 2, previously sensitised with gluco-globulin, were injected in- travenously with 1 cc. of giuco-egg-albumin. Each promptly succumbed with typicai symptoms. Pigs Nos. 3 and 4 of Table VII demonstrate the specitkity of the sensitization; both of the animals when tested with 1 cc. of gala&o-albumin showed no reaction. However, tihen, 3 hours later, 1 cc. of giuco-albumin was introduced, they reacted fatally. Pigs Nos. 5,6,7, and 8 were used to determine the influence of homologous uncombined glucoside on the reaction. In these tests the same relations were found to exist as described in the experiments on passive TABLE VII Active Anaphylaxis in Guinea Pigs Smitid with Gluco-Globulin = I 1st Injection Interval between 1st and 2nd injection 2nd Injection 3rd Injection IntWWl Se%2tw between sensitizing i.p. injection and test GUlIlCa Pi No. 1 2 Material i.v. Result Result ah. - t3 Severe symp- toms. Re covery t3 t24 t44 No reaction No reaction -- _- -- -- - `iKz- 1 cc. 1 cc. min. ::1 gf%" days 5 cc. 21 5 cc. 21 ~- 5 cc. 21 5 cc. 21 -- 5 cc. 21 5 cc. 21 5 cc. 21 5 cc. 21 -- 5 cc. 21 3 4 No reaction 3 No reaction 3 duco-w- dbnmim 1 cc. 1 cc. gkumida 1 cc. 1 cc. 1 cc. 1 cc. l- 1 cc. No reaction No reaction No reaction No reaction 13 2 2 Followed im- mediately No reaction .4 1 cc. 1 cc. 1 cc. 1 cc. "yy+;~- 1 cc. 9 t44 t Death of animal. i.p. = iutraperitoneal. i.v. - intravenous. TABLE VIII Active Anaphylazis in Guiw Pigs Sensitized with Gala&o-Globulin - ,I = P Sd.&S i.p. 2nd Injection ~~~~ 1st Injection ensitiaiig injection and test M;`:M . . - . Result 3rd Injection Interval between 1st and 2nd injection min. t34 t4 ha. No reaction 24 No reaction 2 No rea$ion No reaction No reaction No reaction 14 3 3 Followed im- medidy by No reaction No rektion - 4 4 Yti" . . Result min. t4 t3 t4 t3 t2 Very slight symptoms. Recovery No reaction 1 No reaction 1 .- 1 2 A 3 4 rdodo- ghflin 5 cc. 5 cc. aove `"p&?[t- 21 1 cc. 21 1 cc. -- `g+~- 21 -_ 1 cc. @$&--$- 1 cc. 1 cc. 5 cc. 5 cc. 21 1 cc. m- mLxhi& 21 1 cc. 21 1 cc. 21 1 cc. 5 cc. see. see. 5 cc. 1 dc. 1 cc. 1 cc. 1 cc. 21 1 cc. -l- CJw-rtTr- albumin 1 cc. 1 cc. grvcotida 21 1 ck. 21 1 cc. 5 cc. 5 cc. t Death of animal. i.p. = intraperitoneal. i.v. = intravenous. 564 CON JrtGATED CARBOHkDRATE-PROTEINS. III anaphylaxis. When glucoside was injected immediately before gluco-albumin, complete inhibition of anaphylaxis resulted. However, when glucoside was in- jected one and one-half to two hours before the shocking dose (Pigs Nos. $6. and 7), no protection occurred. Pig No. 9 of Table VII is further evidence of the specificity of active sensitization; in this a&al attempts to inhibit shock with heterologous galactoside and to desensitize with gakto-albumin were ineffectual since the subsequent injection of gluco-albumin produced characteristic death. TABLE IX Active Anaphylaxis ia Guinea Pigs Sen.sitizd with Glum-GlobuJin and Galacto-GloWin Reactions induced by the use of horse globulin-the protein common to both Guinen pip No. scnsi* - i.p 5 cc. gluco-globulin 5 cc. gluco-globulin 5cc. galacto- globulin 5 cc. galacto- globulin Normal control Time lt.ervB davr 21 21 21 21 antigens Shocking dose i.v. 1 cc. horse-globulin (18 mgms.> 0.5 cc. horse-glob- ulin (9 mgmS.) 1 cc. horse-globulin (18 mgms.) 0.5 cc. horse- globulin (9 mgms.) 1 cc. horse-globulin (18 mgms.) symptoms Result Typical t33 minutes Typical t14 minutes Typical t3 minutes Typical t73 minutes None No reaction t Death of animal. i.p. = intraperitoneal. i.v. = intravenous. In Table VIII, a similar group ' of experiments was carried out employing galacto-globulin for senkitization instead of gluco-globulin. Galacto-albumin was the toxigenic agent; galactoside was injected for inhibition tests. Results comparable in every respect to those recorded in Table VII were obtained. Consequently a detailed description need not be given. Table IX presents t&e results obtained in guinea pigs actively sensitized with gluco-globulin (animals Nos. 1 s&d 2) or with galacto- globulin (Nos. 3 and 4) and subsequently injected with horse globulin. W. S. TILLkTT, 0. T. AVERY, ANI3 W; F. GOEBEL 565 All the animals gave typical reactions. Active sensitivity in these pigs was in all probability induced by the uncombined globulin present in the sensitizing material. DISCUSSION The experiments reported in this paper demonstrate the capacity of artificially prepared sugar-proteins to produce both active and passive anaphylaxis.. The tests were devised and carried out in such a manner as to emphasize the significance of the carbohydrate radical. The fact that guinea pigs, passively sensitized with antigluco-globulin serum, or actively sensitized with gluco-globulin, can be subsequently shocked with gluco-albumin, demonstrates that the antigen-antibody specificity in these instances is directly dependent upon the carbo- hydrate fraction of the antigenic compounds. The introduction of the sugar radical into the protein molecule endows the new complex with a sharply defined specific antigenicity. This fact is brought out by experiments in which galactoside was substituted for glucoside in the preparation of sugar-proteins used for sensitization. The same specific relations hold in the production of anaphylaxis with galacto-proteins as that described for gluco- proteins. Attempts to incite anaphylactic shock with heterologous material were ineffectual. The results of the anaphylactic experi- ments conform to the results anticipated by the serological findings of Avery and Goebel (2). Landsteiner (7), employing complex antigens, has reported experi- ments on anaphylaxis of a similar character to those presented in this report. He found that guinea pigs sensitized with one azoprotein could be shocked by the `injection of a second compound containing the same azo-groups attached to a different protein. In addition to the new specificity which the carbohydrate radical confers upon the conjugated proteins, the uncombined glucosides by themselves also exert a detite infIuence on the reactivity of sensitized animals. When sensitized pigs are injected with the homologous glucoside immediately before the introduction of the toxigenic sugar- pro;tein, they are completely protected from shock. However, the protection afforded by the glucoside alone is apparently only transi- tory; for, when the interval between introduction of glucoside and 566 CONPGATED CARBOHYDRATE-PROTEINS. III shocking agent was as long as two hours, the injection of homologous sugar-protein produced prompt and typical anaphylactic death. That the temporary protection just mentioned is specific, was demonstrated by the experiments in which uncombined carbohydrate of heterologous type was shown to exert no such protective action. The transitory, specific protection aEorded by the glucosides alone is not yet understood. Landsteiner (7) found in experimtits on ana- phylaxis with axoproteins that the azo-component, when injected one hour before the conjugated azoprotein, inhibited shock. He con- sidered that a state of anti-anaphylaxis had been induced. In the tests with glucosides and sugar-proteins, sufbcient evidence has not been obtained to interpret the mechanism other than to say that the inhibitory effect of the glucosides disappears in at least two hours. Active and passive anaphylaxis has also been elicited with uncom- bined globulin. Whether animals were passively sensitized with anti- gluco-globulin or antigalacto-globulin serum, the toxigenic action of i globulin was equally effective. Since the sera employed contained . anti-globulin antibodies, these results were to be expected. Guinea pigs actively sensitized with either gluco-globulin or gala&o-globulin, were found to be equally sensitive to uncombined globulin. Since the mated- used to produce active sensitization contained free globulin, the subsequent intoxication with horse globulin is obviously based on a simple protein- anti-protein reaction. CONCLUSIONS 1. Guinea pigs passively .sensitized with the serum of rabbits immunized with an artificially prepared sugar-protein (gluco-globulin) exhibit typical anaphylactic shock when subsequently inoculated with gluco-albumin; the serum of rabbits immunized with a second syn- thetic sugar-protein (galacto-globulin) similarly sensitizes guinea pigs to galacto-albumin. The reactions, in each instance, are specific and depend for their specificity on the carbohydrate component, and not on the protein fraction of the synthesized sugar-protein. 2. Guinea pigs actively sensitized with gluco-globulin or galacto- globulin are similarly subject to anaphylactic shock, when injected, after 21 days, with sugar-proteins containing carbohydrate identical with that present in the sensitizing antigen, regardless of the kind of protein with which it is combined. W. S. TILLETT, 0. T. AVERY, AND W. E. GOEBEL 567 3. `Jhe unconjugated glucosides, although themselves not capable of inducing shock, inhibit the anaphylactic reaction when injected immediately prior to the introduction of the toxigenic sugar-protein. The protective action of the glucosides disappears within ,two hours after injection. In order to elicit the phenomenon, the carbohydrate must be the same as that combined in the sugar-protein complex. 4. Anaphylactic shock may be induced by uncombined globulin in guinea pigs passively sensitized with either antigluco-globulin serum or antigalacto-globulm serum; globulin is similarly &ective in animals actively sensitized with gluco-globulin or galacto-globulin. The reac- tions elicited by globulin alone are dependent upon the common pro- tein present in the antigens, and exhibit only species specificity. L BIBLIOGRAPHY 1. Goebel, W. F., and Avery, 0. T., J. 3&Q. Med., 1929, SO, 521. 2. Avery, O.T., and Goebel, W. F., J. Exp. i&d., 1929,50, 533. 3. Tomcsik, K., &UC. Sac. Exe. Biot. and Med., 1927,24,812. 4. Tom&, K., and Kurotchkin, T. K, J. ExQ. &fed., 1928,47,379. 5. Lance&Id, R. C., J. a?hQ. Med., 1928,47,843. 6. Avery, 0. T., and Tillett, W. S., J. A'zQ. Med., 1929,49,251. 7. Landsteiner, K., J. ExQ. Med., 1924,39,631.