Antibody-Mediated and Delayed-Type Hypersensitivity ... · antigen evoked overlapping antibody-mediated and delayed-type reactions. Lipopolysaccharide andpolysaccharide preparations

INFECriON AND IMMUNITY, Feb. 1975, p. 360-364Copyright 0 1975 American Society for Microbiology

Vol. 11, No. 2Printed in U.S.A.

Antibody-Mediated and Delayed-Type HypersensitivityReactions to Brucella Skin Test Antigens in Guinea Pigs

LOIS M. JONES AND DAVID T. BERMAN*

Department of Veterinary Science, University of Wisconsin, Madison, Wisconsin 53706

Received for publication 8 July 1974

Cutaneous hypersensitivity responses to brucella antigens of different compo-sition were studied in guinea pigs sensitized by infection with smooth brucella orimmunization with killed rough brucella in adjuvant. These animals hadcirculating antibodies to smooth lipopolysaccharide or protein antigens, respec-tively. Intradermal skin tests, active cutaneous anaphylaxis, passive cutaneousanaphylaxis, and immunodiffusion tests were performed. Delayed-type hyper-sensitivity reactions uncomplicated by accompanying antibody-mediated reac-tions were seen only in infected guinea pigs with protein antigen that was entirelyfree of lipopolysaccharide. In the adjuvant-immunized animals, the proteinantigen evoked overlapping antibody-mediated and delayed-type reactions.Lipopolysaccharide and polysaccharide preparations contained varying amountsof protein components. In infected animals, reactions to these antigens wereclearly antibody mediated, but participation of delayed-type hypersensitivitycould not be excluded. In adjuvant-immunized animals, the antibody-mediatedreaction to the lipopolysaccharide preparation was caused by its proteincomponent.

Although the ability of proteins to induce andelicit specific delayed-hypersensitivity reac-tions is generally recognized, there is contro-versy surrounding the role of polysaccharides(PS; 3) and lipopolysaccharides (LPS; 8) thatare unassociated with proteins. Most brucellaallergens contain varying amounts of LPS inaddition to proteins and nucleoproteins (2).Others consist almost entirely of PS (15). Sinceanimals infected with brucella usually haveserum antibodies to PS determinants of the cellwall LPS, it is possible that dermal reactions tosome allergens involve overlapping antibody-mediated and delayed-type reactions.

Preparations of brucella LPS contain lipid A(13) as well as varying amounts of polypeptide,some of which is tightly bound (M. S. Redfearn,Ph.D. thesis, Univ. of Wisconsin, Madison,1960). Previous studies on the dermal responseof sensitized guinea pigs to brucella LPS (10)suggested that the inflammation was a combi-nation of reactions comprising those due to thetoxicity of lipid A and specific delayed hyper-sensitivity to the polypeptide associated withthe antigen. In the present study, a brucella PSantigen that had minimal toxicity for normalguinea pigs was examined concurrently withLPS antigens. These antigens, containing vary-ing amounts of protein, were compared with aprotein allergen that lacked LPS components.The experiments were designed to determine

36C

whether brucella LPS and PS elicit delayedhypersensitivity reactions, or whether the der-mal reactions elicited by these antigens areantibody mediated in whole or in part.

MATERIALS AND METHODSBacterial strains. All brucella cultures have been

used in our previous studies (4, 5). Their significantcharacteristics are as follows. Smooth strains Brucellamelitensis 16M, B. abortus 544, B. abortus 2308, B.abortus 1119-3, and B. abortus 19 all contain smoothLPS and stimulate production of antibody specific forthis surface component (4). B. abortus 544 producesgeneralized brucellosis in guinea pigs inoculated with102 to 103 organisms (11). B. melitensis rough strainB115, B. abortus rough strain llR, and B. abortusrough strain 45/20 lack the smooth LPS (5).

Antigens. Protein (P) antigen was prepared fromB. melitensis rough strain B115 by extraction in coldsaline (2) and characterized as a heat-stable proteinallergen containing no detectable LPS (12). Proteincontent, as determined by the Lowry method (19),was 56%.Two different preparations of LPS antigen, ex-

tracted from B. melitensis smooth strain 16M by themodified hot phenol-water method (14), were used.LPS-1 contained 25% protein. LPS-2 had been sedi-mented at 60,000 x g for 12 h and contained 5%protein.PS antigen was prepared from B. abortus smooth

strain 1119-3 by a method (15) modified by E.Scheibner (personal communication), which con-sisted of trichloroacetic acid extraction of the washed

on June 17, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

BRUCELLA SKIN TEST ANTIGENS

whole cells followed by acetic acid hydrolysis toremove lipids and proteins. The final product con-

tained 1.4% protein.LPS antigens extracted by the hot phenol-water

method from Yersinia enterocolitica types 8 and 9were obtained from R. Diaz. They were includedbecause LPS from smooth brucella strains and LPSfrom Y. enterocolitica type 9 (but not type 8) showcross-reactivity (6, 9) but the protein antigens fromthese genera do not (6).

Antisera. Antisera were obtained from rabbitshyperimmunized with acetone-killed rough B.abortus strain 11R in Freund incomplete adjuvant(FIA), with killed smooth B. abortus strain 2308 inFIA, and by infection with living smooth B. abortusstrain 19 (5). Sera were also obtained from sensitizedguinea pigs before they were tested intradermally.

Sensitization of guinea pigs. Guinea pigs of theHartley strain were sensitized by infection with 10lorganisms of smooth, virulent B. abortus strain 544 or

by immunization with killed rough B. abortus strain45/20 in FIA as previously described (11). They were

tested by intradermal tests or for active cutaneousanaphylaxis (ACA) from 7 to 11 weeks after sensitiza-tion.

Intradermal tests in guinea pigs. A constantvolume of 0.1 ml of antigen was injected per site andthe dose was expressed as micrograms (dry weight) in0.1 ml of pyrogen-free saline. The mean diameter oferythema standard error was the parameter of skinresponse (11).ACA. Infected and adjuvant-immunized guinea

pigs were injected intradermally with 0.1 ml of severaldoses of three or more antigens immediately after theintracardial injection of 0.5 ml of 1% Evan's blue in0.85% saline (16). A blue ring appearing within a fewminutes at the intradermal site of one of the antigensindicated that the intracardial injection had beensuccessful. An intradermal injection of saline was

included on each animal as a control to rule out thepossibility of nonspecific bluing reactions. The ani-mals were observed for 30 min. Normal guinea pigsgiven intracardial injections of Evan's blue followedby intradermal injections with the same antigens didnot develop bluing reactions.

Passive cutaneous anaphylaxis (PCA). Individ-ual or pooled sera from normal, infected, and immu-nized guinea pigs, taken before intradermal or ACAtests, were diluted 1:25, 1:100, and 1:400, and 0.1 mlwas injected intradermally into normal guinea pigs(16). Serum dilutions heated for 30 min at 56 C were

also injected with unheated serum dilutions on some

animals. Three to 4 h later the animals were injectedintracardially with 1 ml containing 100 to 500 jug ofone of the antigens and 0.5 ml of 1% Evan's bluesolution. Only those guinea pigs showing a bluingreaction at at least one of the intradermal sites within30 min were included in the results.

RESULTS

Antigenic analysis. Analysis of the brucellaLPS preparation by immunoelectrophoresiswith hyperimmune anti-smooth brucella serum

showed slowly diffusing lines with the charac-

teristic cathodic mobility that have been identi-fied as LPS antigens (4), as well as severalfast-moving lines associated with the P antigens(4). The PS antigen developed only a single linewith cathodic mobility similar to that observedwith LPS. The P antigen preparation did notdevelop the LPS line with the anti-smoothbrucella serum from infected animals or thosehyperimmunized with either smooth or roughorganisms. P antigen developed a number oflines, as illustrated previously (12), when testedwith either hyperimmune anti-smooth or anti-rough sera.Immunodiffusion tests were set up to deter-

mine the cross-reactivity of the LPS and PSlines, as well as the identity of the P compo-nents, in the various antigenic preparations.Table 1 summarizes the numbers of lines ob-served with all of the combinations of sera andantigens. It shows that sera from rabbits hyper-immunized with killed rough B. abortus did notprecipitate with PS or yersinia antigens andthat sera from rabbits infected with smooth B.abortus did not precipitate with brucella Pantigen. Sera from rabbits hyperimmunizedwith killed smooth B. abortus precipitated withall of the antigens except Y. enterocolitica type8 LPS, which was not precipitated with any ofthe sera.

Figure 1 is from a representative experimentwith sera from rabbits infected with smooth B.abortus. It illustrates the reaction of identity ofone of the components of each of LPS-1 andyersinia type 9 LPS with the single componentin brucella LPS-2 and PS antigens.

Sera from rabbits hyperimmunized withkilled rough brucella developed reactions ofidentity with at least one of the lines in brucellaP, LPS-1, and LPS-2 antigens.The reactions with guinea pig sera were the

same as those with the rabbit sera. That is, serafrom infected guinea pigs had precipitins to the

TABLE 1. Number of lines in immunodiffusion tests ofantigens with rabbit antisera

Immunizing antigen

Test antigena Killed Killed Livingrough smooth smooth

B.abortus B.abortus B.abortus+ FIA ± FIA

Brucella P .............. >3 > 3 0Brucella LPS-1 ......... >3 >3 1-2Brucella LPS-2 ......... 1 2 1-2Brucella PS ............ 0 1 1Yersinia type 9-LPS 0..... 1-2 1-2Yersinia type 8-LPS 0 0 0

a Antigens were prepared as described in Materials andMethods.

VOL. 11, 1975 361

on June 17, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

JONES AND BERMAN

FIG. 1. Immunodiffusion tests of brucella LPS-1, LPS-2, PS, and Y. enterocolitica type 9-LPS (Y9) antigenswith serum from a rabbit infected with living smooth B. abortus.

LPS component but not to the P components,and sera from guinea pigs immunized withrough brucella in adjuvant had precipitins to Pcomponents but not to LPS.Intradermal tests. The characteristics of the

dermal responses of guinea pigs to P and LPSantigens have been described previously (10).To investigate the nature of reactions of PSantigen in the same way, doses of 1, 10, and 100ug were injected singly into non-sensitized con-

trols and guinea pigs sensitized either by infec-tion with living smooth brucella or by immuni-zation with rough brucella in FIA. The 100-,ugdose produced pale erythematous reactions at 3to 6 h in all animals. Only those infected guineapigs with circulating anti-smooth antibody, asdetected by agglutination tests or by immuno--diffusion with LPS, developed larger and more

marked erythematous reactions that persisted

for 24 h. Doses of 10 and 1 ,ug produced minimalreactions that disappeared by 24 h.To establish the qualitative characteristics of

the reactions to the three types of antigens,groups of guinea pigs were injected simultane-ously with two doses of each of the P, LPS-1,and PS preparations (Table 2). Although thenumbers of animals were small, certain trendswere apparent. Within 5 min of intradermalinjection, a red bleb was observed at the site ofLPS and PS injections in infected guinea pigsand at the site of P and LPS injections inadjuvant-immunized guinea pigs. These imme-diate reactions had almost completely disap-peared by 3 h after injection, but erythema wasagain apparent at these sites at 5 h. Infectedguinea pigs showed maximum reactions at 24 hwith P and LPS antigens but at 5 h with PSantigen. In the adjuvant-immunized animals

TABLE 2. Dermal reactionsa to protein, lipopolysaccharide, and polysaccharide antigens of brucella insensitized and normal guinea pigs

Test Dose Infected (4)b Adjuvant immunized (3) Normal (4)antigen C(g) 5min 5h 24h 5 min 5h 24h 5min 5h 24h

P 10 - 8.7 2.8 17.7 2.6 + 14 15.5 + 0.9 - - -

1 - - 13.8 1.9 + 8.3 4 2.3 6.7 ± 2.7 - _ _

LPS-1 10 + 10 ± 2.1 17.4 i 1.7 + 13.7 ± 0.3 13.7 ± 1.8 - 10.3 ± 0.7 11.6 ± 0.91 + 5.7 1.7 10.6 2.2 + 9.3 0.3 12 0.6 - _ 6 1.4

PS 10 + 16.7 0.7 13.5 2.1 - - _ _ _ _1 + 9.5 3.2 4.7 0.7 _ _ _ _ _

a Dermal reactions are designated as + (red bleb at 5 min) and - (no reaction). Figures are mean diameter oferythema (in millimeters) ± standard error of the mean.'Number of animals tested.

INFECT. IMMUN.362

on June 17, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

BRUCELLA SKIN TEST ANTIGENS

there were not significant differences in themagnitude of the reactions between 5 and 24 hwith either the P or LPS antigens, and therewere no reactions with PS antigens. Normalguinea pigs responded to LPS antigen only.ACA. Infected guinea pigs manifested posi-

tive ACA reactions to all antigens containingthe PS components in LPS, PS, or the cross-

reactive yersinia type 9 LPS, but negativereactions to P or yersinia type 8 LPS (Table 3).In contrast, the guinea pigs immunized withrough brucella in adjuvant had positive ACAreactions only with the P and LPS antigenscontaining the brucella proteins.PCA. Sera from infected guinea pigs pro-

duced positive PCA reactions in all animalsinjected intracardially with PS antigen, in someanimals injected with LPS antigen, but in noneof the animals given P antigen (Table 4).

Sera from adjuvant-immunized guinea pigsproduced positive PCA reactions in animalsinjected intracardially with P and LPS anti-gens. Sera from adjuvant-immunized guineapigs were absorbed with P antigen. These ab-sorbed sera did not give PCA reactions witheither P or LPS antigens, although the unab-sorbed sera tested on the same animals were

positive with these antigens. This indicatedthat the P component in LPS was antigenicallythe same as that in P antigen.Serum dilutions that had been heated gave

the same results as nonheated serum dilutionstested simultaneously on the same guinea pigs,i.e., positive at a 1:100 dilution. This minimizesthe probability that immunoglobulin E wasinvolved.

TABLE 3. Active cutaneous anaphylaxis reactions insensitized guinea pigs

Method of sensiti-zation

Test antigen Dose (gg) Adjuvant

Infection immuni-zation

Brucella P .......... 10 0/9a 10/101 0/5 7/8

Brucella LPS-1 ...... 10 8/8 7/71 4/8 7/10

Brucella LPS-2 ...... 10 6/8 3/41 1/6 3/5

Brucella PS ......... 10 9/9 0/101 8/9 0/7

Yersinia LPS-9 ...... 10 5/5 0/2Yersinia LPS-8 ...... 10 0/5 0/2Saline 0/9 0/10

a Numerator, Number of guinea pigs giving positivereactions at site of intradermal injection of testantigens; denominator, number of sensitized guineapigs successfully injected intracardially with Evan'sblue.

DISCUSSIONThe need for improvement of standardization

of skin test antigens in general has been empha-sized in recent World Health Organization re-ports (17, 18). Improvement of the techniques ofstandardization should include evaluation ofthe capacity of antigens used for assessingdelayed hypersensitivity to induce immediatehypersensitivity reactions, as these may inter-fere with the interpretation of the delayedresponse.Our data show that guinea pigs infected with

smooth B. abortus have circulating antibody to

TABLE 4. Passive cutaneous anaphylaxis tests

Source and dilution of guinea pig serum injected intradermallyBrucella test Intracardial dose .

antigen ~ (Mg Infected Adjuvant immunized Normalantigen (,lgJ1:25 1:100 1:400 1:25 1:100 1:400 1:25 1:100

P 100 0/2a 0/2 0/2 2/2 2/2 1/2 0/1 0/1250 0/3 0/3 0/3 3/3 3/3 0/3 0/1 0/1500 0/6 0/6 0/5 8/8 8/8 3/6 0/3 0/3

LPS-1 300 0/1 0/1 0/1 1/1 1/1 1/1 0/1 0/1500 1/4 1/4 1/4 6/6 6/6 0/3 0/3 0/3

LPS-2 200 0/2 0/2 0/2 2/2 2/2 2/2500 2/3 1/3 0/1 2/3 2/3 0/3 0/2 0/2

PS 100 1/1 1/1 0/1 0/1 0/1 0/1 0/1 0/1500 5/5 3/5 1/5 0/5 0/5 0/5 0/5 0/5700 1/1 0/1 0/1 0/1 0/1 0/1 0/1 0/1

a Numerator, Number of guinea pigs giving positive reactions at site of intradermal injection of the specifiedserum; denominator, number of normal guinea pigs that were successfully injected intracardially with the testantigen and Evan's blue.

363VOL. 11, 1975

on June 17, 2020 by guesthttp://iai.asm

.org/D

ownloaded from

JONES AND BERMAN

the PS determinants of the cell wall LPS butnot to the protein components. These animalshave immediate reactions in ACA tests withLPS and PS but not with P antigens, and theirsera give identical results in PCA tests. Thisestablishes that, in infected animals, the reac-tions observed at 24 h in sites injected with Pantigens are not antibody mediated. The reac-tions to LPS and PS antigens clearly have anantibody-mediated component, but the data donot permit the exclusion of an overlappingdelayed-type reaction.Animals immunized with the rough orga-

nisms in adjuvant are not equivalent to infectedanimals in their responses to skin test antigens.They have circulating antibody to the P anti-gens but not to the PS determinants of the cellwall LPS. These animals have immediate reac-tions in ACA tests with P and LPS but not withPS antigens, and their sera give identical resultsin PCA tests. The abolition of PCA reactions toLPS by absorbing sera with P antigen showsthat the antibody-mediated reactions in theseanimals are to the P component. These anti-body-mediated reactions can interfere with thequantification and interpretation of the de-layed-type hypersensitivity response to P anti-gen. It seems clear from these results that it ispreferable to use infected, rather than adju-vant-immunized, guinea pigs for the standard-ization of antigens for delayed-type hypersensi-tivity testing in brucellosis.This and other studies have shown the superi-

ority of LPS-free P antigen for tests of delayedhypersensitivity in brucellosis. There seems tobe no useful purpose served by the inclusion ofLPS in a test antigen, and there are clearly anumber of disadvantages. Lipid A is toxic (7)and mitogenic for B cells (1), and the presenceof the complete LPS or the PS can result incutaneous anaphylaxis and Arthus reactions, orboth, with resultant confusion of the interpreta-tion. Also, the complete LPS in the antigen canstimulate antibody production that can inter-fere with subsequent serological diagnosis.

ACKNOWLEDGMENTSThis study was supported by grants from the World Health

Organization and the Research Committee of the GraduateSchool with funds furnished by the Wisconsin Alumni Re-search Foundation.We are indebted to R. Diaz and G. Dubray for some of the

antigens employed.

LITERATURE CITED1. Andersson, J., O., Sjoberg, and G. Moller. 1972. Induc-

tion of immunoglobulin and antibody synthesis in vitro

by lipopolysaccharides. Eur. J. Immunol. 2:349-353.2. Bhongbhibhat, N., S. S. Elberg, and T. H. Chen. 1970.

Characterization of Brucella skin-test antigens. J.Infect. Dis. 122:70-82.

3. Daniel, T. M., and C. F. Hinz, Jr. 1974. Reactivity ofpurified proteins and polysaccharides from Mycobacte-rium tuberculosis in delayed skin test and culturedlymphocyte mitogenesis assays. Infect. Immun.9:44-47.

4. Diaz, R., L. M. Jones, D. Leong, and J. B. Wilson. 1968.Surface antigens of smooth Brucellae. J. Bacteriol.96:893-901.

5. Diaz, R., L. M. Jones, and J. B. Wilson. 1967. Antigenicrelationship of Brucella ovis and Brucella melitensis.J. Bacteriol. 93:1262-1268.

6. Diaz, R., R. Lacalle, M. P. Medrano, and D. Leong. 1970.Immunological activities of the endotoxin from Yer-sinia enterocolitica strain M. Y., p. 11. In Proc. FifthInt. Congr. Infect. Dis., vol. 2.

7. Galanos, C., E. T. Rietschel, 0. Liideritz, 0. Westphal,Y. B. Kim, and D. W. Watson. 1972. Biologicalactivities of lipid A complexed with bovine-serumalbumin. Eur. J. Biochem. 31:230-233.

8. Greisman, S. E., and R. B. Hornick. 1972. Cellularinflammatory responses of man to bacterial endotoxin:a comparison with PPD and other bacterial antigens. J.Immunol. 109:1210-1222.

9. Hurvell, B. 1973. Serological cross-reactions betweendifferent Brucella species and Yersinia enterocolitica.Biological and chemical investigations of lipopolysac-charides from Brucella abortus and Yersiniaenterocolitica type IX. Acta Pathol. Microbiol. Scand.Sec. B 81:105-112.

10. Jones, L. M. 1974. Specificity of brucella protein antigensand role of lipopolysaccharide antigens in elicitingdelayed hypersensitivity reactions in sensitized guineapigs. Ann. Rech. Vet. 5:189-199.

11. Jones, L. M., D. T. Berman, and R. Diaz. 1973. Factorsinfluencing biometric potency assays of brucella aller-gens in guinea pigs. Br. J. Exp. Pathol. 54:368-379.

12. Jones, L. M., R. Diaz, and A. G. Taylor. 1973. Character-ization of allergens prepared from smooth and roughstrains of Brucella melitensis. Br. J. Exp. Pathol.54:492-508.

13. Lacave, C., J. Asselineau, A. Serre, and J. Roux. 1969.Comparaison de la composition chimique d'une frac-tion lipopolysaccharidique et d'une fraction polysac-charidique isolees de Brucella melitensis. Eur. J. Bio-chem. 9:189-198.

14. Leong, D., R. Diaz, K. Milner, J. Rudbach, and J. B.Wilson. 1970. Some structural and biological propertiesof Brucella endotoxin. Infect. Immun. 1:174-182.

15. Mosimann, W. 1949. Allergene aus Brucella abortusBang. Schweiz. Z. Pathol. Bakteriol. 12:362-379.

16. Ovary, Z. 1964. Passive cutaneous anaphylaxis, p.259-283. In J. F. Ackroyd (ed.), Immunologicalmethods. A symposium. Blackwell, Oxford.

17. Report: Joint FAO/WHO Expert Committee on Brucello-sis, 5th Report. 1971. W. H. 0. Tech. Rep. Ser., no. 464.

18. Report of a WHO Scientific Group. 1973 Cell-mediatedimmunity and resistance to infection. W. H. 0. Tech.Rep. Ser., no. 519.

19. Williams, C. A., and M. W. Chase. 1968. Estimation ofprotein by the Folin-Ciocalteau reaction, p. 273-275. InC. A. Williams and M. W. Chase (ed.), Methods inimmunology and immunochemistry, vol. 2. AcademicPress Inc., New York.

364 INFECT . IMMUN .

on June 17, 2020 by guesthttp://iai.asm

.org/D

ownloaded from