Immunobiol., vol. 182, pp. 127-134 (1991)
1 Department of Immunology, Erasmus University Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, 2 National Institute of Public Health and Environmental Protection, P.O. Box 1, 3720 BA Bilthoven, 3 Department of Rheumatology, University Hospital Nijmegen, The Netherlands
Natural Antibodies to 65 kD Mycobacterial Heat Shock Protein in Rats do not Correlate with Susceptibility for Mycobacterium tuberculosis Induced Adjuvant Arthritis ARIANNEKEA. GRANDIA 1, Hendrik DE VIssERl,Jan D. A. VANEMBDEN2 , RUURDVANDERZEE 2, WIMB. VAN DEN BERG 3, and MAARTENP. HAZENBERG 1 Received July 9, 1990 . Accepted in Revised Form November 22, 1990
Abstract Natural antibodies to 65 kD heat shock protein (hsp65) of Mycobacterium bovis were found in the sera of Lewis rats. The levels of these natural hsp65 antibodies differed substantially between the individual rats. Each rat was subsequently tested for its susceptibility to develop arthritis following injection of M. tuberculosis in incomplete Freund adjuvant. It was found that the incidence and severity of the induced arthritis did not differ between groups of Lewis rats with relatively high and relatively low natural antibody levels to hsp65. Inoculation of rats without natural antibodies to hsp65 with intestinal contents did not induce hsp65 antibodies, although the rats were able to respond to the antigen.
Introduction A 65 kD heat shock protein (hsp65) from Mycobacterium bovis was prepared by recombinant technique by THOLE et al. (1, 2). Studies of VAN EDEN et al. (3) and VAN DEN BROEK et al. (4) showed that immunization of Lewis rats with hsp65 protects against Mycobacterium tuberculosis adjuvant arthritis (AA) and Streptococcus pyogenes cell wall-induced arthritis. The later model was used by our group while investigating the arthritisinducing properties of cell walls of Eubacterium species from the human intestinal flora (5, 6). A role of the intestinal flora in AA was observed by KOHASHI et al. (7) who found that AA resistant F344 rats became susceptible in germ free (GF) conditions. During our studies on arthritis induction in rats, we observed differences in susceptibility between rats of the same Abbreviations: GF = germ free; PBS = phosphate buffered saline; IP SPF = specific pathogen free; AA = adjuvant arthritis
128 . ARIANNEKE A. GRANDlA et a!.
age and delivery date. This prompted us to investigate the relationship between the intestinal microbial flora, the natural antibody titer against hsp65 and the susceptibility to AA. In this paper, data are presented on the occurrence of such natural antibodies in strains of rats with a different intestinal-microbial status, the induction of anti-hsp65 antibodies by microfloras and the influence of natural antibody titers on arthritis induction.
Materials and Methods Rats Female Lewis rats were obtained from Harlan Olac (Bicester, U.K.) and the Zentral Institut fur Versuchstierkunde (Hannover, Germany). The suppliers of the Lewis rats are further referred to as A and B, respectively. Specific pathogen free (SPF) male Buffalo rats came from the Radiobiological Institute TNO (Rijswijk, The Netherlands) and were raised under strict barrier conditions. During the experiments, all rats were housed under conventional conditions. Sera from germ free (GF) Wag/Rij rats were kindly provided by the Radiobiological Institute TNO, and were obtained by orbital punction.
Immunization and adjuvant arthritis (AA) induction Recombinant M. bovis hsp65 was provided by the National Institute of Public Health and Environmental Protection (Bilthoven, The Netherlands) and prepared as described previously (1, 2). For immunization, 100 [tg hsp65 in 1 ml phosphate buffered saline (PBS) was mixed with 1 ml incomplete Freund adjuvant (ICFA; Difco, Detroit, MI, USA); 0.5 ml emulsion was administered intraperitoneally (IP) to rats. Cecal contents of rats (GF, SPF and conventional) was suspended in PBS and filtered through gauze. Suspensions were adjusted to 20 mg dry weight/m!. Gram stains showed the presence of at least 109 bacteria per ml in the SPF and conventional cecal suspensions. A volume of 0.5 ml was mixed with 0.5 ml rCF A and IP administered to rats. Booster injections were performed with 0.1 ml of the mixture in the tail base. For induction of AA, rats were injected intracutaneously at the base of the tail with 1 mg heat-killed M. tuberculosis (H37RA, Difco) suspended in ICFA (S). Development of AA was followed for 90 days.
Detection of serum hsp65 antibodies by enzyme linked immunosorbent assay (ELISA) Microtiter plates (96 wells, U shaped, Greiner, Niirtingen, Germany) were coated with hsp65 (2 [tg/ml PBS, 125 [tl) during 1 h at 37 dc. After washing with PBS-Tween 0.02 % plates were incubated with 100 ttl serum diluted twice, 1:10 to 1:SO in PBS-Tween 0.5 % during 1 h. After washing, the plates were incubated with peroxidase-conjugated rabbit antibodies to rat immunoglobulins (Dako, Copenhagen, Denmark), 1000x diluted in PBS-Tween 0.5 %. After washing, 40 rug orthophenylene diamine was added as a substrate with 40 fll 30 % H 2 0 2 in a 100 ml buffer of 0.05 M citric acid adjusted to pH 5 with 0.1 M Na2HP04. The reaction was stopped after 45 min with 50 ttl 4 M H 2 S0 4 and read in a Titertek Multiskan (Flow Laboratories, Irvine, Scotland) at 492 nm. Determinations were performed in duplo; mean values were used. Absorbances were corrected for the background absorption of conjugate and antigen. Background absorbance never exceeded O.OS. A serum dilution was considered as positive for antibodies against hsp65 if the absorbance exceeded 0.100. The results are presented as absorbances at 492 nm with 1:30 diluted serum (abs) calculated from the dose response curves.
Western blot Western blotting was performed as previously described (9). In short, protein samples were separated by SDS-polyacrylamide gel electrophoresis on 10.5 % slab gels and transferred on to
Antibodies to 65 kD Heat Shock Protein in Rats and Arthritis . 129 nitrocellulose sheets. Blots were incubated with rat serum at dilutions of 1:25 and 1:100 as the primary antibody. The second antibody was peroxidase-conjugated rabbit antibodies to mouse IgG (DAKO) which is highly crossreactive with rat IgG. Controls with only one conjugate were always negative. As antigen preparations we used: (A) Escherichia coli lysate with an empty expression vector pPlc236 (prepared as B), (B) E. coli lysate containing M. bovis BCG hsp65 (250 ng) expressed from plasmid pRIB1300 (2, 3) and (C) purified hsp65 (100 ng) (2, 3). Evaluation of arthritis The rats were examined 3 times a week for a period of 90 days. A rat was considered as positive for arthritis if within the observation period, one of the footpaws, wrists, ankles or the tail was swollen for at least a week. During an active phase of the arthritis, one rat was sacrificed to confirm the inflammation histologically. Skinned ankle joint specimens were fixed in 1: 10 (v/v) diluted buffered 36 % formaldehyde solution, decalcified in 5 % (v/v) formic acid for five days and embedded in paraffin. Sections were stained with hematoxylin and eosin. Statistics The Mann-Whitney U test was used to calculate the statistical differences between absorbances at 492 nm of sera in the ELISA. The significance of differences in arthritis induction between groups of rats was judged from the X2 -test with Yates correction for continuity.
Results Natural antibodies to hsp65 hsp in rats The presence of antibodies to hsp65 was determined in sera from 10 Lewis rats from supplier A and 10 Lewis rats from supplier B in two samples with 14 days interval. Figure 1A and B present the absorbances with 1:30 diluted sera (solid and open bars). In both samples the absorbances of sera from group A were significantly higher than those from group B (p = 0.04). Absorbances of the two samples from the same animal were not significantly different. In only one serum out of 17 SPF Buffalo rats were antibodies detected in the second sample (abs 0.133). Keeping 10 Buffalo rats 8 weeks under nonbarrier conditions did not result in measurable anti-hsp65 antibodies. In sera from GF Wag/Rij rats anti-hsp65 antibodies were not found. Figure 2 showed the results of Western blots with rat serum with antihsp65 antibodies determined in the ELISA (abs 2.200). The serum recognized purified hsp65 as well as hsp65 in E. coli. E. coli lysate without hsp65, however, was negative. Note that serological crossreactivity with the hsp65 of E. coli is not detectable with this rat serum.
Arthritis induction Induction of AA was attempted in the Lewis rats A and B mentioned above by injection of M. tuberculosis. In the observation period of 90 days, three out of 10 Lewis rats A and five out of 10 Lewis rats B developed arthritis in one of more legs or in the tail. In Figure 1 A and B the rats are marked that developed an arthritis. Insuspectibility for arthritis induction
A RIANNEKE A. G RANOlA
Lewis rats group A
.*n I 9
Lewis rats group B
Figure 1. Antibodies against hsp65 in 30 x diluted serum from 10 Lewis rats from two diffe rent suppliers (designated A and B) determined twice (solid and open bars). Bars marked with an ':' indicate a particular r at developing AA after injection with M. tuberculosis. Dotted bars depict data on anti-h sp65 antibodies after the injection.
was also observed in rats with very low levels of antibodies to hsp65. The results show that relatively high antibody levels to hsp65 did not prevent the arthritis development in group A (rats no. 1 and 3). One of the two rats in group B with natural antibodies to hsp65 (rat 10) also developed arthritis. The incidence of arthritis did neither differ significantly between groups A and B (X 2 =0.21, p=0.65), nor between rats with (At, A3-A5, A7, AS, A10, B9, BIO) and without (A2, A6, A9, BI-BS) hsp65 antibody levels (X2 = 0.01, P = 090). . Two weeks after injection, the presence of antibodies to hsp65 was again determined. Figure I A and B (dotted bars) show that after immunization with M. tuberculosis antibody levels to hsp65 raised in rats independent on the susceptibility for arthritis.
Origin of natural antibodies to hsp65 The absence of antibodies to hsp65 in GF Wag/Rij rats and SPF Buffalo rats led us to investigate the occurrence of the corresponding antigens in the
65 kD Heat Shock Protein in Rats and Arthritis . 131
1 : 100
Figure 2. Western blot with 1:25 and 1:100 diluted ratserum with anti-hsp65 antibodies on the following preparations (A) E. coli lysate only, (B) E. coli lysate with hsp65 and (C) purified hsp65.
conventional flora of Lewis rats with such antibodies. Male Buffalo rats without antibodies to hsp65 were therefore inoculated three times with the cecal content from a GF rat, an SPF rat without antibodies or a Lewis rat with a high level of the antjbodies (four rats per group). Despite the booster injections, antibodies were not found in either group. A fourth group was used as control and injected with PBS only. When this group was later challenged with hsp65, high levels of anti-hsp65 antibodies appeared one week later, showing that the rats were responders to th e antigen (readings above 1.8 with 1:30 diluted serum).
Histology The visual scoring of arthritis induction by M. tuberculosis was confirmed by histological examination of an affected ankle by sacrificing rat B5 42 days after inoculation. As normal control served an unaffected ankle of the same animal. The arthritic ankle showed a florid inflammation (Fig. 3 A). There was an infiltration of both polymorphonuclear and mono-
ARI ANN EKE A. GRAN Ol A
.' 0 ~ ~ , ,
Figure 3. Histological appearance of an arthritic (A) and a normal (B) ankl e o f a emale f Lewis rat, sacrificed 42 days after injection with M. tuberculosis (magnifi cation, x 300).
nuclear cells in the joint cavity, together with precipitation of fibrin and coagulation factors. The sy novial tissue was also infiltrated with inflammatory cells of mixed cellularity. In the unaffected ankle, these features were absent (Fig. 3 B).
Discussion Rheumatoid arthritis patients have raised levels of IgG binding to the hsp65 of M, bovis (10). The relevance of hsp65 in animal models of arthritis is obvious since it can protect against AA as well as cell wall-induced arthritis (3, 4). The importance, however, of the presence of natural or induced antibodies to hsp65 was not implicated in these studies. This is probably due to the observation that the protection is T cell mediated, In the present study we show that like in human sera, natural antibodies to hsp65 could be detected in sera from Lewis rats. The levels of antibodies seemed to be dependent on local environmental conditions, since they were different in Lewis rats from different suppliers. The incidence of arthritis induction by M. tuberculosis was low in both groups and obviously not dependent on the level of hsp65 antibodies. Thus, levels of natural hsp65
Antibodies to 65 kD Heat Shock Protein in Rats and Arthritis . 133
antibodies are not indicative for the presence of sufficient protecting T cells. This study also shows that immunization of rats with hsp65 raised the antihsp65 antibody levels. Similar levels were detected in rats protected by hsp65 for S. pyogenes cell wall-induced arthritis from the study of VAN DEN BROEK et al. (4). Raised antibodies to hsp65 were also found as the result of injection with M. tuberculosis but the response was independent on the development of arthritis. The data show that the determined antibodies are hsp65 directed since the presence of it in Mycobacterium is beyond dispute (1) Definitive proof that the natural antibodies to hsp65 only binds to that molecule and not to some minor contaminant from E. coli in the preparation was given by the Western blot experiments. The absence of natural anti-hsp65 antibodies in germ free rats is not surprising, since the only bacterial antigens such animals encounter are food-derived. More notable was the absence of the antibodies in SPF Buffalo rats. Although the SPF barriers in the laboratory they came from are strict, the animals harbored a complete intestinal flora containing more than 10 10 bacteria per gram cecal contents (data not shown). An eight-week housing period under nonbarrier conditions did not result in anti-hsp65 antibodies. In a previous study, antibodies to human fecal flora bacteria were induced in rats by inoculation with fecal suspensions (11). In the present study, immunization with a flora from animals with high levels of natural antibodies to hsp65 did not induce the antibodies. Apparently the antigenic determinants of hsp65 heat shock protein are not sufficiently present on the countless species, subspecies, and biotypes of the extremely complex indigenous intestinal microflora (12, 13). THOLE et al. (14) pointed on the ubiquity of hsp65 and a trigger role was proposed by RES et al. in the etiology of arthritis (15). The studies presented here do not support an intestinal flora derived origin of hsp65. Acknowledgments This work was supported by the Nederlandse Verenigung voor Reumabestrijding (Dutch Rheumatism Foundation). G. DE KORTE and H. J. DE JAGER are gratefully acknowledged for assistance in the preparation of the manuscript, and T. VAN OS for expert photography.
References 1. THOLE, J. E. R., H. G. DAUWERSE, P. K. DAS, D. G. GROOTHUIS, L. M. SCHOULS, and J. D. A. VANEMBDEN. 1985. Cloning of Mycobacterium bovis BCG DNA and expression of antigens in Escherichia coli. Infect. Immun. 50: 800. 2. THOLE, J. E. R., W. J. KEULEN, A. H. J. KOLK, D. G. GROOTHUIS, L. G. BERWALD, R. H. TIESjEMA, and J. D. A. VAN EMBDEN. 1987. Characterization, sequence determination and immunogenicity of a 64-kilodalton protein of Mycobacterium bovis BCG expressed in Escherichia coli KIl. Infect. Immun. 55: 1466. 3. VAN EDEN, W., J. E. R. THOLE, R. VAN DER ZEE, A. NOORDZlj, J. D. A. VAN EMBDEN, E. J. HENSEN, and I. R. COHEN. 1988. Cloning of the mycobacterial epitope by T lymphocytes in adjuvants arthritis. Nature 331: 171.
134 . ARIANNEKE A. GRANDIA et al. 4. VAN DEN BROEK, M. F., E. J. M. HOGERVORST, M. C. J. VAN BRUGGEN, W. VAN EDEN, R. VAN DER ZEE, and W. B. VAN DEN BERG. 1989. Protection against streptococcal cell wall induced arthritis by pretreatment with the 65 kD mycobacterial heat shock protein. J. expo Med. 170: 449. 5. SEVERIJNEN, A. J., R. VAN KLEEF, M. P. HAZENBERG, and J. P. VAN DE MERWE. 1990. Chronic arthritis induced in rats by cell wall fragments of Eubacterium species from the human intestinal flora. Infect. Immun. 58: 522. 6. SEVERIJNEN, A. J., R. VAN KLEEf, M. P. HAZENBERG, andJ. P. VAN DEMERWE. 1989. Cell wall fragments from major residents of the human intestinal flora induce chronic arthritis in rats. J. Reumatol. 16: 1061. 7. KOHASHI, D., J. KUWATA, K. UMEHARA, F. UEMURA, T. TAKAHASHI, and A. OZAWA. 1979. Susceptibility to adjuvant-induced arthritis among germfree, specific-pathogen-free, and conventional rats. Infect. Immun. 26: 791. 8. HOLOSHITZ, J., Y. MATITIAU, and I. R. COHEN. 1985. Role of the thymus in induction and transfer of vaccination against adjuvant arthritis with a T lymphocyte line in rats. J. Clin. Invest. 75: 472. 9. VAN EMBDEN, J. D., H. J. VANDER DONK, R. V. VAN EIJK, H. G. VAN DERHEIDE, J. A. DE ]ONG, M. F. VAN OLDEREN, A. D. OSTERHAUS, and L. M. SCHOULS. 1983. Molecular cloning and expression of Treponema pallidum DNA in Escherichia coli K-12. Infect. Immun. 42: 187. 10. BAHR, G. M., G. A. W. ROOK, M. AL-SAFFAR, J. VAN EMBDEN, J. L. STANFORD, and K. BEHBEHANI. 1988. Antibody levels to mycobacteria in relation to HLA type: evidence for non-HLA-linked high levels of antibody to the 65 kD heat shock protein of M. bovis in rheumatoid arthritis. Clin. expo Immunol. 74: 211. 11. HAZENBERG, M. P., J. P. VAN DE MERWE, M. BAKKER, and A. VERSCHOOR-BuRGGRAAF. 1982. Induction of serum agglutinins to Eubacterium, Peptostreptococcus and Coprococcus species in mice and rats. FEMS Microbiol. Lett. 13: 17. 12. MORISHITA, Y., and K. MIYAKI. 1979. Effects of age and starvation on the gastrointestinal microflora and the heat resistance of fecal bacteria in rats. Microbiol. Immunol. 23: 455. 13. BORRIELLO, S. P. 1986. Microbial metabolism in the digestive tract, ed. HILL, M. J., CRC Press, Boca Raton, Florida, p. 1. 14. THOLE,]. E. R., P. HINDERSSON,j. DEBRUYN, F. CREMERS,j. VAN DERZEE, M. DE COCK, J. TOMMASSEN, W. VAN EDEN, and J. D. A. VAN EMBDEN. 1988. Antigenic relatedness of a strongly immunogenic 65 kDA mycobacterial protein antigen with a similar sized ubiquitous bacterial common antigen. Microb. Pathogen. 4: 71. 15. RES, P. C. M., C. G. SCHAAR, F. C. BREEDVELD, W. VAN EDEN, J. D. A. VAN EMBDEN, I. R. COHEN, and R. R. P. DE VRIES. 1988. Synovial fluid T cell reactivity against 65 kD heat shock protein of mycobacteria in early chronic arthritis. Lancet ii: 478. Dr. M. P. HAZENBERG, Department of Immunology, Erasmus University Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, The Netherlands