Eur. J. Immunol. 1991. 21: 2297-2302
Alexander R. L U S S O W ~ ~ , Christy Barriosmv, Jan van Embdene, Ruur Van der Zeef Antonio S. VerdmioA, Antonello PessioA, Jacques A. Louis.., Paul-Henri Lambert. and Giuseppe Del Giudice. World Health Organization-Immunology Research and Tkainmg Center, Department of Pathology., University of Geneva, Geneva, Institute of Biochemistry., University of Lausanne, Epalinges, National Institute of Public Health and Environmental Protection+, Bithoven and Peptide Synthesis Unito, Sclavo SPA, Rome
Heat-shock proteins as carrier molecules
2297
Mycobacterial heat-shock proteins as carrier molecules* We have previously shown that the priming of mice with live Mycobacterium tuberculosis var. bovis (Bacillus Calmette-Gudrin, BCG) and immunization with the repetitive malaria synthetic peptide (NANP)m conjugated to purified protein derivative (PPD), led to the induction of high and long-lasting titers of anti-peptide IgG antibodies, overcoming the requirement of adjuvants and the genetic restriction of the antibody response to the peptidr: (Lussow et al., Proc. Nutl. Acad. Sci. USA 1990. 87: 2960). This initial work led us to the following observations. BCG had to be live for priming to lead to the induction of anti-peptide antibodies. Surprisingly, priming with other living microorganisms which chronically infect the macrophage (e.g. Salmonella typhimuriurn and Leishmania major) also induced anti-peptide antibodies in mice immunized with PPD-(NANP)40 conjugate. It was, thus, hypothesized that molecules expressed during active infection and also known to be highly conserved between species, namely the heat-shock proteins (hsp), could mediate the Tcell sensitization required for the production of anti-peptide antibodies. In fact, when the PPD portion of the conjugate was replaced by a highly purified recombinant protein corresponding to the 65-kDa (GroEL-type) hsp of M. bovis, this resulted in the production of anti-(NANP) IgG antibodies in BCG-primed mice, irrespective of the major histocompatibility complex-controlled responsiveness to the (NANP) sequence itself. Further, similar induction of anti-peptide antibody response was also obtained with a recombinant 70-kDa (DnaK-type) hsp of M . tuberculosis, but not with a small molecular mass (18kDa) of M . leprue. Finally, an adjuvant-free carrier effect for anti-peptide IgG antibody production in BCGprimed mice, was also exerted by the GroEL hsp of Escherichia coZi.This finding that hsp can act as carrier molecules without requiring conventional adjuvants is of potential importance in the development of vaccine strategies.
1 Introduction
vaccines is the lack of optimal adjuvants acceptable for human use [l].Aluminium hydroxide is the most widely Many of the recombinant or synthetic antigens now being used adjuvant in man, but is limited with respect to the type considered as vaccine candidate antigens are not sufficient- of antigen employed and to the stability of the final vaccine ly immunogenic by themselves to induce a strong and preparation [a]. Hence, many different substances and protective immune response. In order to improve the strategies destined to enhance the human response to immunogenicity of these preparations the addition of an immunogens are being explored as alternative adjuvants. immunostimulant (i.e. adjuvant) is required. Unfortunately, a major obstacle for the development of effective new Intact mycobacteria and mycobacterial components have long drawn the attention of many researchers in this respect [2,3]. Previous work by Lachmann et al. [4] has shown that [I 96011 immunization with antigens conjugated to tuberculin PPD * This work was supported by the World Health Organization- (currently used to measure the Tcell reactivity in people Transdisease Vaccinology Programme, the UNDPlWorld exposed to tuberculosis or vaccination against tuberculosis) B a n N H O Special Programme for Research and Training in of mice primed with BCG (Bacillus Calmette-GuCrin, Tropical Diseases, the Swiss National Science Foundation, and Mycobacterium tuberculosis var. bovis) , could induce a theENI, Italy. strong immune response to these antigens. Recently, we Supported by a fellowship from the Commission F6d6rale des investigated [5] this model for the induction of antibodies Bourses pour Etudiants Etrangers, Switzerland. using a synthetic peptide, consisting of 40 Asn-Ala0 Present address: Ludwig Intitute for Cancer Research, LauAsn-Pro (NANP) repeats [referred to as (NANP)m] from sanne Branch, CH-1066 Epalinges, Switzerland. the repetitive region of the major antigen, the circumsporA Present address: Italfarmaco Research Center, 1-20095 Cinisello ozoite (CS) protein, uniformly covering the external surBalsarno, Italy. face of the sporozoites of the human malaria parasite, Present address: I.R.B.M., I-OOO40 Pomezia, Italy. Plasmodium falciparum [6]. (NANP), peptides, that are Correspondence: Giuseppe Del Giudice, WHO-Immunology considered for the development of human malaria vaccines Research and Training Center, Department of Pathology, Univer- [7,8], are known to be immunogeniconly in mice with the Michel Server 1,CH-1211 Geneva 4, Switzer- H-2bhaplotype [9-111 and to require strong (e.g. Freund's) sity of Geneva, land adjuvants for the induction of specific immune responses. Abbreviations: CS: Circumsporozoite bsp: Heat-shock pro- In our model [5], mice were first primed with BCG and then tein hspR6: Recombinant M. bovis BCG 65-kDa (GroEGtype) immunized with the (NANP)mpeptide conjugated to PPD. hsp hspR70: Recombinant M. tubercuhis 70-kDa (DnaK-type) Following this immunization strategy, anti-parasite IgG antibodies were induced, in the absence of adjuvants, hsp d18: Recombinant M. leprae l&kDa protein 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheh, 1991
0014-2980/91/1010-2297$3.50 + .25/0
irrespective of the murine genetic responsiveness to the (NANP), epitope, at titers similar to those raised with conjugates given in CFA [5]. In the present study, we report the results of experiments carried out to investigate the requirements for the induction of this phenomenon. We found that heat-shock proteins (hsp) could replace PPD in the conjugate and behave as carrier molecules in the induction of anti-peptide antibodies in the absence of adjuvant.
2 Materials and methods 2.1 Bacteria The M. tuberculosis var. bovis used was the BCG-F strain obtained from the BCG laboratory of the Institut Pasteur, Paris, France [12]. Aliquots of 0.5 ml at lo9 CFU/ml of Dubo's medium were stored at -70°C to improve their stability [13]. For priming, mice were infected with lo6CFU given i.p. Heat-killed BCG was obtained by warming a solution of 2 x lo6CFU BCG/ml in a water bath at 80 "C for 20 min; then 0.5 ml of the heated solution was administered per mouse i.p. as the priming stimulus where indicated. Sonicated BCG was prepared by disruption of the bacilli for a total of 10min in a Soniprep 150 cell sonicator (MSE Scientific Instruments, Sussex, GB). The preparation was centrifuged at 20000 x g for 30 min, and the SN (soluble antigens) was removed. After determination of the protein content, 0.25 mg protein of this sonicate preparation was given i.p. per mouse as the priming stimulus, where indicated. Salmonella typhimurium, strain 661/90, was a kind gift of Dr. F. Sadallah [14], Leishmania major was cultured as previously described [15]. A dose of lo6of each viable microorganism was given i.p. to each mouse as the priming stimulus where indicated.
PPD was purchased from the Statens Serum Institut, Copenhagen, Denmark [16]. PPD at a concentration of 1mg/ml in PBS was used throughout the experiments. The (NANP)a synthetic peptide, from the P falciparum CS protein [6], was produced as described in detail elsewhere [17]. The synthesis of the 21-mer peptide YNRNIVNRLLGDALNGKPEEK (referred to as Py l),corresponding to positions 59-79 of the CS protein of the murine malaria parasite l? yoefii,and its characterization as a T cell epitope in H-2d and H-2b mice, have also been reported [18, 191. Lyophilized material was resuspended in sterile distilled water at 10 mglml, aliquoted, and stored at -70 "C until use. M. bovis BCG GroEL-type 65-kDa hsp (hspR65) was expressed from a recombinant E. coli K12 strain harboring plasmid pRIB1300 [20,21] and purified as described [22]. Recombinant M. tuberculosis DnaK-type 70-kDa hsp (hspR70) was obtained and purified by ATPagarose chromatography according to Mehlert and Young [23], and the M. leprae 18-kDa protein (m118), showing homologies with a low-molecular weight hsp of soybean [24], according to Harris et al. [25]. The purified E. coli GroEL protein was a gift from D. Young. One milligram each of PPD, Pyl peptide, hspR65 , hspR70, m118, GroEL was coupled to 1 mg of (NANP)a peptide in the presence of
0.002% glutaraldehyde for 2 h, and then dialyzed against PBS overnight [5].Aliquots of each conjugate were stored at -70°C until use.
2.3 Mice and immunizations Female, 8-12-week-old BALB/c (H-2d) and C57BL/6 (H2b) mice were bred in our animal facilities. The original breeding pairs were obtained from the Jackson Laboratory, Bar Harbor, ME. Each mouse received i.p. 106 CFU of BCG or equivalent doses of heat-killed or sonicated BCG or live S. typhimurium or L. major promastigotes. Each mouse received 20 p,g of the relevant (NANP)m conjugate i.p. in 0.5 ml PBS, as indicated, on days 14 and 35, according to the schedule previously established [5].
Mice were bled weekly from the retroorbital plexus, the serum separated, and anti-(NANP)m antibodies titered by ELISA using the (NANP)40 peptide as solid phase as described in detail [5, 101. Samples with titers < 1:200 were considered as negative.
Having previously shown that priming with BCG followed by immunizations with PPD-(NANP)a peptide conjugate in the absence of adjuvants induced the production of anti-(NANP)M IgG antibodies [5],we addressed the question of whether the model depended on the use of live BCG as the priming stimulus. Three groups of BALB/c mice (H-2d, genetically nonresponder to (NANP) peptides [9, 111) were primed with live, heat-killed or sonicated
Figure 1. Priming with live BCG is required for the induction of anti-peptide antibodies in BCG-primed mice. Three groups of five BALBk mice were primed by i.p. injection with lo6 CFU BCG, with lo6 heat-killed BCG, or with 250 pg of sonicated BCG in 0.5 ml PBS. On days 14 and 35 after priming, all the mice were immunized with 20 pg of the PPD-(NANP)m peptide conjugate. Ten days after the second conjugate injection the mice were bled and the anti-(NANP)40 IgG antibody titers were measured by ELISA using the peptide as solid phase, as described [5, 101. Results are expressed as the geometrical mean f SEM of the loglo reciprocal of the last serum dilution positive for each group of mice, i.e. giving an absorbance 2 0.2 at 492 nm.
Eur. J. Immunol. 1991. 21: 2297-2302
BCG, and then immunized twice with the PPD-peptide conjugate prepared in PBS.The spleen cells from mice in all of the groups exhibited a proliferative response when cultured in the presence of PPD (data not shown). However, only the mice having received live BCG as the priming stimulus were able to produce anti-(NANP) antibodies following two immunizations with the conjugate without adjuvant (Fig. 1). An anti-peptide antibody response was never detectable in the two groups of mice receiving killed BCG. These data demonstrated that priming with live microorganisms was necessary to induce an anti-peptide antibody response in the absence of adjuvants. 3.2 Mycobacterial carrier molecules are required for the induction of anti-peptide antibodies in BCG-primed mice As the mycobacteria are known to be able to stimulate some compartments of the immune system nonspecifically [26], it was decided to evaluate the effect of priming with live BCG on immunization with a conjugate made of a T cell epitope, bearing no relation to the mycobacteria, coupled to (NA")m peptide. A synthetic peptide, referred to as Pyl, from the CS protein of the murine malaria parasite l? yoefii, was chosen for this purpose as it behaves as a strongTcel1 epitope in BALB/c mice [18, 191. As shown in Fig. 2, Pyl-(NANP)m conjugates were able to overcome the genetic restriction of the antibody response to the (NANP)m peptide in BALB/c mice only when the conjugates were prepared in Freunds adjuvants (i.e. first CFA, then IFA). Two immunizations of the Pyl-(NANP)m conjugate in PBS were totally ineffective in eliciting anti(NANP) antibodies, even in mice primed with live BCG.
Heat-shock proteins as carrier molecules
2299
conjugate. Furthermore, our experiments with the live bacilli suggested that mycobacterial antigen(s) actively produced during BCG infection and present in the PPD preparation would be capable of mediating the phenomenon of anti-peptide antibody production in the absence of adjuvant.
3.3 Other parasites of the MQtrcan provide the priming for the induction of anti-peptide antibodies A group of mycobacterial molecules whose synthesis is up-regulated during BCG infection, the hsp, is known to sensitize T cells in the infected host [27]. Furthermore, hsp fragments had been detected among the antigens of PPD [21]. It then seemed possible that mycobacteria hsp could mediate theT cell sensitization that would be recalled upon immunization with the PPD-containing conjugate. As the hsp molecules are the most conserved proteins throughout evolution identified so far [28, 291, we reasoned that other phylogenetically diverse intracellular microorganisms that lead, as BCG does, to long-term infections of the M a , could provide, through their hsp, the T cell sensitization required for further boosting by the PPD in the conjugates. The enterobacterium S.typhimurium and the protozoan L. major were selected because they express different families of hsp [30]. Indeed, when BALB/c mice were infected with live BCG or S. typhimurium or L. major, and subsequently challenged with the PPD-(NANP)m conjugate without adjuvant, a comparable anti-(NANP)4o IgG antibody response was detectable in all the mice (Fig. 3).
These experiments strongly suggested that BCG-primed T cells were providing the "help" required for the production of anti-peptide antibodies. These results were also in agreement with those reported by Lachmann et al. [4] showing that non-mycobacterial carrier proteins, such as KLH, were unable to replace the PPD portion of the Figure 3. Other parasites of the M@ can provide the priming for the induction of anti-peptideantibodies. On day 0, three groups of five BALB/c mice were primed i.p. with 106 BCG, L. major, or S.typhimurium (strain 6 1 m ) . Immunizations with the PPD(NANP)40conjugate and expression of results [anti-(NANP)IgG antibody titers] are as described in the legend of Fig. 1.
1.5 2 2.5 3 Log10 anti-(NANP) IgG titers Figure 2. Mycobacterial carrier molecules are required for the induction of anti-peptide antibodies in BCG-primed mice. Three groups of four BALBk mice were immunized as follows: (squares) mice received the priming with live BCG (106 CFU) i.p. and then two i.p. immunizations with the Pyl-(NANP)m conjugate (20 pg) in PBS; (triangles) unprimed mice received twice the Pyl(NANP)m conjugate in PBS; (circles) unprimed mice received the Pyl-(NANP)a conjugate in complete and then incomplete Freund's adjuvant. Anti-(NANP) IgG antibodies from day 49 are shown as the mean absorbance at 492 nm for each group of mice obtained from serial dilution of the serum samples tested in the
(NANP)40 ELISA [S, lo].
3.4 Mycobacterial hsp act as carrier molecules for the induction of anti-peptide antibodies in BCG-primed mice without adjuvant
To test the hypothesis that hsp present in the PPD preparation were exerting the carrier effect on the production of anti-peptide antibodies, the purified recombinant M. bovis GroELtype hspR65 was coupled to the (NANP)40 peptide and the conjugate used to immunize live BCG-primed C57BL/6 (H-2b, genetically responsive to (NANP), peptides [9-111) and BALBlc nonresponder mice. As illustrated in Fig. 4, the hspR65-(NANP)a conjugate given in PBS was able to induce an anti-(NANP) IgG antibody response in BCG-primed C57BL/6 mice after a
1300
Eur. J. Immunol. 1991. 21: 2297-2302
A. R . Lussow, C. Barrios. J.Van Embden et al
3.5 The E . coli GroEL hsp behaves as a carrier molecule in BCG-primed mice
Figure 4. Carncr effect of the M . h0L.i.c.hspR65. One group of five C57BL/6 and one group o f five BALB!c mice were primed with live BCG, as described for Figure 1. Fourteen days later, both groups were immunized with 30 11g of the hspR65-(NANP),U conjirgnre delivered i. p. in 0.5 nil I’HS. ’l‘lic conjugate irnmrinimioii w a s repeated in the same manner on da? 3.5 after BC’Ci priming. Results [anti-(NANP) IgG antibody titers] from serum samples te5ted after the first (day 1s: hatched columns) or after the second (day 19: solid culumns) conjugate immunizations are shown a n d cxprcssed as in the lcgend of Fig. I .
single immunization. Further. after two immunizations with this conjugate without adjuvant, the genetic rcstriction of the antibody response to t h e peptide was overcome in BALB/c mice, resulting in a strong anti-peptide IgG antibody response. Therefore, immunization with the synthetic peptide linked to the hspR65 moleculc produced results identical to those we previously obtained using the PPD-(NANP)4,1conjug;itc in thcsc two strain4 of micc 151. T h e GroEL-type hsp is not the only class of stress proteins produced by the mycohacteria 131 1. Therefore. t h e role other types o f mycob;ictcriiil hsp coiild Ii;ivc iii this systcm was investigated. .l‘hc recombinant M . tubcrculosi~DnaKtype hspR70 [?3] and the recombinant M . leprae 18-kDa protein (mllX). which has homologies with a lon.-molecular-weight hsp of soybean 1241, were conjugatcd to the ( N A N P ) a peptide and used to immunize BALB/c mice previously primed with livc BCG. The results o f thcsc cxpcrimcntx arc‘ s h o w n in f;ig. 5 . I%oth thc hspK65 and hspR70 molecules from the closely related BCG and M . tuherculmis bacilli. respectively, could mediate the induction of anti-peptide IgG antibodies in the absence of adjuvants. I n contrast, the M . leprue 18-kDa molecule did not act as a carrier for the induction of anti-(NANP) antil)oclics i n BC’G-priiiictl micc. Carrier
2
3
4
Log10 anti-(NANP) IgG titers f i g i t r e 5 Cai ricr cffcct of the ,M ruberculou hsp70 Three groups of BALBk mice wcrc primed ai t h live BCG and immunixd (as clc\cril)ctl for Fig I ) kith 20 pg 01 liy05-(hASl’)qt. hy)70 (NAN1’)q,. or nillH-(NANP)40conjugdtc, given i.p in 0.5 1111 1’1)s on d;i)s 14 dnd 35 after BCG priming Results [anti-(NANP) IgG antibody titers from clay 19l A I L cxprcsscd as in Fig 1
Tdken together, these data demonstrated that at least the mycobacterial hsp belonging t o the G r o E L and DnaK families behaved as carrier molecules for the induction of anti-peptide antibodies in BCG-primed mice without adjuvants. We then ;iddressed t h e question o f whether a non-mycobaclerial hsp belonging to the same class could mediate the same effect.Thc E. coli GroEL hsp was used in these experiments: it mas conjugated to the ( N A N P ) a peptide and the conjugate used t o immunize BALBlc mice previously primed with live BCG. As Fig. 6 shows, mice immunized twicc with thc GroEl ,-(hAhT)u, conjugate without adjuvant produced anti-( NANP) IgG antibodics at titers similar t o those obtaincd with the hspR65-(NANP)al conjugate. Interestingly. similar antibody titers were also obtained when BCG-primed mice were first immunized with the GroEL-(NAhT)lo and then with the hspR65(NANP)al conjugate, or vice versa. C57BL/6 mice gave totally similar results (data not s l i o w i i ) .
E. coli GroEL hsp behaves as a carrier molecule in IN ’( i-primed micc. (;roup\ o f thrcc to Ioiii IMLD/c rnicc wcrc piiirictl w i t h livc H C Y ; :is tlc.icrihcd i n Fig. I , and then irnrnunizccl 14 and 35 days after with 20 big o f the h ~ p R 6 5 - ( S A W ) ~oro Figitre(,.
GroEI.-(KASP)q, conjugate. given i.p. in 0.5 ml PBS. Results [anti-(NA\P) IgCi antibody titen from day 491 are expressed as in Fig. 1.
4 Discussion I lcrc wc have shown that mycolwctcrial hsp t>ch;ivca s carrier molccules for the induction ol anti-peptide antihodies in mice previously primed with BCG. In all cases. the antibod!. production took place with the hsp-peptide conjugate given in the uhsclrtct of adjuvants, and required a previous priming with lille BCG. The data reported in this paper ;idd to those already rcportctl h y u\ conccrniiig the i i w o f I’PI) as :I carrier for the induction of anti-(NANP) antibodics in BCG-primed mice 151. It is then reasonable to hypothesize that the effects mediated by BCG priming and immunization with PPD can be ascribed at least i n part to the hsp of 65 and 70 kDa actively produced hy mycobacteria.
Since in our model the priming of mice had t o be mediated by live microorganisms. it seemed that Stress proteins acre likcl! to mediate the Tcell sensitizarion that would be
Eur. J. Immunol. 1991.21: 2297-2302
recalled to help (NANP)-specific B cells. In fact, hsp production is increased during BCG infection of the MQ, and T cell sensitization is often associated with mycobacterial infection [27,32]. However, priming with heat-killed or sonicated BCG was ineffective in inducing anti-peptide antibodies following immunization with the PPD(NANP)4" conjugate. This was possible due to the activation of hsp-specific Tcells at a lower frequency than that achieved after a long-term infection by BCG which would allow persistent availability of antigen. Alternatively, it is possible that different T cell subsets were activated after priming of mice with live or killed BCG, leading in one case, but not in the other, to the helper effect required for the induction of anti-peptide antibodies. It is interesting to note that microorganisms as phylogenetically diverse as a mycobacterium (BCG), an enterobacterium (S. typhimurium), and a protozoan (L. major),were all able to prime the mice for the production of anti-peptide antibody response after immunization with the PPD (mycobacterial antigens)-(NANP)a conjugate. Since all these organisms express hsp and lead to long-term infections of the MQ, [30,33], and since the hsp molecules are the most conserved proteins throughout evolution [28,29], it is possible that in these experiments priming was mediated through conserved homologous epitopes also present in the PPD preparations: in fact, it has been shown that fragments of the mycobacterial65-kDa hsp are present in PPD [21]. This hypothesis is also supported by our findings showing that BCG-primed mice produced anti-(NANP)a IgG antibodies after immunization with the peptide conjugated with the E. coli GroEL hsp, which has a high degree of homology with the mycobacterial 65-kDa hsp [34]. This would suggest that Tcells induced after priming with live BCG (or other Ma-infecting organisms) can cross-react with hsp of different origins. This would be in line with previous reports by others, showing that indeed this appears to be the case in certain circumstances [35-381. The finding that mycobacterial hsp can successfully substitute for PPD as carriers for the induction of strong antibody responses to defined B cell epitopes in the absence of adjuvant could, on one hand, be expected, since many Tcell epitopes have been identified within them [39]. On the other hand, it was surprising (a) that priming with live BCG was required, and (b) that this, in turn, eliminated the need for the use of adjuvants that are normally required when other carrier molecules are utilized. It is not yet clear through which mechanisms hsp molecules exert their carrier effect. From our data it is evident that the carrier effect was mediated by GroEL- and DnaK-type hsp, but not by the ml18, a 18-kDa protein of M. leprae sharing homologies with a low-molecular weight hsp of soybean [24].These negative results with m118 molecule may be due to the absence of a homologous molecule in M . tuberculosis complex, or to a limited homology between this molecule and possible equivalent hsp from BCG. Some comments have to be made on the possibility of inducing autoimmune reactions, due to the homologies existingbetween some mycobacterial molecules and certain host tissue antigens, and between bacterial and human hsp. For example, antibodies to M. bovis 65-kDa hsp have been reported to cross-react with human antigens [40], and antibodies to the 70-kDa hsp of the malaria parasite
Heat-shock proteins as carrier molecules
2301
R falciparum have been shown to react also with the homologous human hsp [41]. Moreover, T cell epitopes of the 65-kDa hsp, associated with mycobacterial infection or the use of CFA, have been known to induce adjuvant arthritis in certain susceptible strains of rats [42]. However, immunization with the 65-kDa hsp has been shown to protect these animals from such autoimmune reactions [20,43]. Finally, no homology has yet been found between the 70-kDa (DnaK-type) hsp and cartilage proteins [U].
The carrier effect of mycobacterial hsp and the fact that the antibody response induced takes place without requiring adjuvant, make this approach of immunization particularly interesting for the development of vaccine strategies. Since BCG is commonly and widely used very early in life as a vaccine against tuberculosis, the use of hsp in vaccine formulations may present several advantages over other carrier molecules. Many Tcell epitopes have been identified on the M. bovis 65-kDa hsp, and the entire protein has been demonstrated to be immunogenic in a wide variety of mouse strains with differing MHC haplotypes [39]. Furthermore, beyond their possible use in individuals already sensitized to the mycobacteria (through vaccination or natural contact), the hsp molecules may also be able to interact with hsp-specificT cell populations found in apparently healthy adults [36]. Experiments done in our laboratory in order to investigate possible eventual application of this vaccination strategy for human use, have shown that the production of anti-peptide IgG antibodies is induced also when BCG and conjugates are given i.d., and that this approach is also applicable to synthetic peptides shorter than the malaria (NANP)a peptide, i.e. with 24-mer peptides from HIV envelope antigens (manuscript in preparation). Finally, immunizations have been started to investigate whether this approach of immunization also applies to non-human primates. We would like to thank Drs. G. Milon and F! David for critical reading of the manuscript, Ms. C. Tougnefor technical help, Dr. D. Young for the E. coli GroEL hsp, Ms. E Sadallah for S. typhimurium, Ms. K . Hug for L. major parasites, Mrs. M. Gheorghiu for BCG. Received May 27, 1991.
5 References 1 Lussow, A. R., Aguado, M. T., Del Giudice, G . and Lambert, P.-H., Immunol. Lett. 1990. 25: 255. 2 Altman, A. and Dixon, F. J., Adv. Vet. Sci. Comp. Med. 1989.33: 303. 3 Warren, H. S. and Chedid, L. A . , CRC Crit. Rev. Immunol. 1988. 8: 83. 4 Lachmann, l? J., Strangeways, L.,Vyakarmam, A. and Evan, G., Ciba Found. Symp. 1986. 119: 25. 5 Lussow, A. R., Del Giudice, G . , Rknia, L., Mazier, D., Verhave, J. I?, Verdini, A. S., Pessi, A . , Louis, J. A. and Lambert, P.-H., Proc. Matl. Acad. Sci. USA 1990. 87: 2960. 6 Nussenzweig,V. and Nussenzweig, R. S., Adv. Immunol. 1989. 45: 283. 7 Ballou, W. R., Hoffman, S. L., Sherwood, J. A., Hollingdale, M. R.,Neva, F. A., Hockmeyer,W.T., Gordon, D. M., Schneider, I.,Wirtz, R. A. ,Young, J. F.,Wassennan, G. F., Reeve, P., Diggs, C. L. and Chulay, J. D., Lancet 1987. i: 1277.
2302
A. R. Lussow, C. Barrios, J.Van Embden et al.
8 Herrington, D. A., Clyde, D. F., Losonsky, G., Cortesia, M., Murphy, J. R., Baqar, S., Felix, A. M., Heimer, E. I?, Gillessen, D., Nardin, E., Nussenzweig, R. S., Nussenzweig,V., Hollingdale, M. R. and Levine, M. M., Nature 1987. 328: 257. 9 Good, M. F., Berzofsky, J. A., Maloy,W. L., Hayashi,Y, Fujii, N., Hockmeyer,W. and Miller, L. H., J. Exp. Med. 1986.164: 655. 10 Del Giudice, G., Cooper, J. A., Merino, J. ,Verdini, A. S., Pessi, A.,Togna, A. R., Engers, H. D., Corradin, G. and Lambert, P.-H.,J. Immunol. 1986. 137: 2952. 11 Togna, A. R., Del Giudice, G. ,Verdini, A. S., Bonelli, F., Pessi, A., Engers, H. D. and Corradin, G., J. Immunol. 1986. 137: 2956. 12 Gheorghiu, M., Chambon, L., Serie, C. and Juvin-Calmette, D., Dev. Biol. Stand. 1977. 38: 29. 13 Gheorghiu, M. and Lagrange, l?, Ann. Immunol. (Inst. Pasteur) 1983. 134: 125. 14 Sadallah, F., Brighouse, G., Del Giudice, G., Drager-Dayal, R., Hocine, M. and Lambert, P.-H., J. Infect. Dis. 1990.161: 59. 15 Louis, J. A . , Moedder, E., Behin, R. and Engers, H. D., Eur. J. Immunol. 1979. 9: 841. 16 Magnusson, M. and Bentzon, M. W., Bull. W H . 0 . 1958. 19: 829. 17 Esposito, G., Pessi, A. and Verdini, A. S., Biopolymers 1989. 28: 225. 18 Grillot, D., Michel, M., Miiller, I., Tougne, C., Rknia, L., Mazier, D., Corradin, G., Lambert, l?-H., Louis, J. A. and Del Giudice, G., Eur. J. Imrnunol. 1990. 20: 1215. 19 Del Giudice, G., Grillot, D., R h i a , L., Miiller, I., Conadin, G., Louis, J. A., Mazier, D. and Lambert, P.-H., Irnmunol. Lett. 1990. 25: 59. 20 Van Eden, W., Thole, J., Van der Zee, R., Noordzij, A., Van Embden, J., Hensen, E. and Cohen, I., Nature 1988. 331: 171. 21 Thole, J. E., Dauwerse, H. G., Das, P. K., Groothuis, D. G., Schouls, L. M. and Van Embden, J. D., Infect. Immun. 1985. 50: 800. 22 Thole, J. E. ,Van Schooten,W. C., Keulen,W. J., Hermans, P.W., Janson, A. A., DeVries, R. R.,Kolk, A. H. and Van Embden, J. D., Infect. Immun. 1987. 55: 1466. 23 Mehlert, A. and Young, D. B., Mol. Microbiol. 1989. 3: 125. 24 Nerland, A. H., Mustafa, A. S., Sweetser, D., Godal, T. and Young, R., J. Bacteriol. 1988. 170: 5919.
Eur. J. Immunol. 1991. 21: 2297-2302 25 Harris, D. I?, Backstrom, B. T., Booth, R. J., Love, S. G., Harding, D. R. and Watson,J. D., J. Immunol. 1989. 143: 2006. 26 Fromrnel, D. and Lagrange,€?,Parasitol. Today 1989. 5: 188. 27 Kaufmann, S. H. E., Schoel, B., Wand-Wiirttenberger, A., Steinhoff, U., Munk, M. and Koga, T., Curr. Top. Microbiol. Immunol. 1990. 155: 125. 28 Lindquist, S. and Graig, E., Annu. Rev. Genet. 1988. 22: 631. 29 Young, R. A. and Elliott, T. J., Cell 1989. 59: 5 . 30 Buchmeier, N. A. and Heffron, F., Science 1990.248: 730. 31 Shinnick, T. M., Curr. Top. Microbiol. Immunol. 1991. 167: 145. 32 Milon, G., Lebastard, M. and Marchal, G., Immunol. Lett. 1985. 11: 189. 33 Lee, M. G. S., Atkinson, B. L., Giannini, S. H. and Van der Ploeg, L. H. T., Nucleic Acids Res. 1988. 16: 9567. 34 Shinnick, T. M., Vodkin, M. H. and Williams, J. D., Infect. Immun. 1988. 56: 446. 35 Ottenhoff,T. H.,VanEmbden, J. D.,Thole, J. E. and Kiessling, R., J. Exp. Med. 1988. 168: 1947. 36 Munk, M., Schoel, B., Modrow, S., Karr, R.,Young, R. and Kaufmann, S. H. E., J. Immunol. 1989. 143: 2844. 37 Koga,T.,Wand-Wurttenberger,A., DeBruyn, J., Munk, M. E., Schoel, B. and Kaufmann, S. H. E., Science 1989. 245: 1112. 38 Lamb, J. R., Mendez-Sampirio, P., Mehlert, A., So, A., Rothbard, J., Jindal, S.,Young, R. A. and Young, D. B., Int. Immunol. 1989. I : 191. 39 Brett, S. J., Lamb, J., Cox, J., Rothbard, J., Mehlert, A. and Ivanyi, J., Eur. J. Immunol. 1989. 19: 1303. 40 Aguas, A. P., Esaguy, N., Sunkel, C. E. and Silva, M.T., Infect. Immun. 1990. 58: 1461. 41 Mattei, D., Scherf, A., Bensaude, 0. and Pereira da Silva, L., Eur. J. Immunol. 1989. 19: 1823. 42 Van Eden,W., Holoshitz, J., Nevo, Z., Frenkel, A., Klajman, A. and Cohen, I. R., Proc. Natl. Acad. Sci. USA 1985. 82: 5117. 43 Van der Broek, M. F., Hogervorst, E. J. ,Van Bruggen, M. ,Van Eden, W. ,Van der Zee, R. and Van den Berg,W., J. Exp. Med. 1989. 170: 449. 44 McLean, L. ,Vinrow,V. and Blake, D., J. Exp. Pathol. 1990. 71: 295.
Alexander R. L U S S O W ~ ~ , Christy Barriosmv, Jan van Embdene, Ruur Van der Zeef Antonio S. VerdmioA, Antonello PessioA, Jacques A. Louis.., Paul-Henri Lambert. and Giuseppe Del Giudice. World Health Organization-Immunology Research and Tkainmg Center, Department of Pathology., University of Geneva, Geneva, Institute of Biochemistry., University of Lausanne, Epalinges, National Institute of Public Health and Environmental Protection+, Bithoven and Peptide Synthesis Unito, Sclavo SPA, Rome
Heat-shock proteins as carrier molecules
2297
Mycobacterial heat-shock proteins as carrier molecules* We have previously shown that the priming of mice with live Mycobacterium tuberculosis var. bovis (Bacillus Calmette-Gudrin, BCG) and immunization with the repetitive malaria synthetic peptide (NANP)m conjugated to purified protein derivative (PPD), led to the induction of high and long-lasting titers of anti-peptide IgG antibodies, overcoming the requirement of adjuvants and the genetic restriction of the antibody response to the peptidr: (Lussow et al., Proc. Nutl. Acad. Sci. USA 1990. 87: 2960). This initial work led us to the following observations. BCG had to be live for priming to lead to the induction of anti-peptide antibodies. Surprisingly, priming with other living microorganisms which chronically infect the macrophage (e.g. Salmonella typhimuriurn and Leishmania major) also induced anti-peptide antibodies in mice immunized with PPD-(NANP)40 conjugate. It was, thus, hypothesized that molecules expressed during active infection and also known to be highly conserved between species, namely the heat-shock proteins (hsp), could mediate the Tcell sensitization required for the production of anti-peptide antibodies. In fact, when the PPD portion of the conjugate was replaced by a highly purified recombinant protein corresponding to the 65-kDa (GroEL-type) hsp of M. bovis, this resulted in the production of anti-(NANP) IgG antibodies in BCG-primed mice, irrespective of the major histocompatibility complex-controlled responsiveness to the (NANP) sequence itself. Further, similar induction of anti-peptide antibody response was also obtained with a recombinant 70-kDa (DnaK-type) hsp of M . tuberculosis, but not with a small molecular mass (18kDa) of M . leprue. Finally, an adjuvant-free carrier effect for anti-peptide IgG antibody production in BCGprimed mice, was also exerted by the GroEL hsp of Escherichia coZi.This finding that hsp can act as carrier molecules without requiring conventional adjuvants is of potential importance in the development of vaccine strategies.
1 Introduction
vaccines is the lack of optimal adjuvants acceptable for human use [l].Aluminium hydroxide is the most widely Many of the recombinant or synthetic antigens now being used adjuvant in man, but is limited with respect to the type considered as vaccine candidate antigens are not sufficient- of antigen employed and to the stability of the final vaccine ly immunogenic by themselves to induce a strong and preparation [a]. Hence, many different substances and protective immune response. In order to improve the strategies destined to enhance the human response to immunogenicity of these preparations the addition of an immunogens are being explored as alternative adjuvants. immunostimulant (i.e. adjuvant) is required. Unfortunately, a major obstacle for the development of effective new Intact mycobacteria and mycobacterial components have long drawn the attention of many researchers in this respect [2,3]. Previous work by Lachmann et al. [4] has shown that [I 96011 immunization with antigens conjugated to tuberculin PPD * This work was supported by the World Health Organization- (currently used to measure the Tcell reactivity in people Transdisease Vaccinology Programme, the UNDPlWorld exposed to tuberculosis or vaccination against tuberculosis) B a n N H O Special Programme for Research and Training in of mice primed with BCG (Bacillus Calmette-GuCrin, Tropical Diseases, the Swiss National Science Foundation, and Mycobacterium tuberculosis var. bovis) , could induce a theENI, Italy. strong immune response to these antigens. Recently, we Supported by a fellowship from the Commission F6d6rale des investigated [5] this model for the induction of antibodies Bourses pour Etudiants Etrangers, Switzerland. using a synthetic peptide, consisting of 40 Asn-Ala0 Present address: Ludwig Intitute for Cancer Research, LauAsn-Pro (NANP) repeats [referred to as (NANP)m] from sanne Branch, CH-1066 Epalinges, Switzerland. the repetitive region of the major antigen, the circumsporA Present address: Italfarmaco Research Center, 1-20095 Cinisello ozoite (CS) protein, uniformly covering the external surBalsarno, Italy. face of the sporozoites of the human malaria parasite, Present address: I.R.B.M., I-OOO40 Pomezia, Italy. Plasmodium falciparum [6]. (NANP), peptides, that are Correspondence: Giuseppe Del Giudice, WHO-Immunology considered for the development of human malaria vaccines Research and Training Center, Department of Pathology, Univer- [7,8], are known to be immunogeniconly in mice with the Michel Server 1,CH-1211 Geneva 4, Switzer- H-2bhaplotype [9-111 and to require strong (e.g. Freund's) sity of Geneva, land adjuvants for the induction of specific immune responses. Abbreviations: CS: Circumsporozoite bsp: Heat-shock pro- In our model [5], mice were first primed with BCG and then tein hspR6: Recombinant M. bovis BCG 65-kDa (GroEGtype) immunized with the (NANP)mpeptide conjugated to PPD. hsp hspR70: Recombinant M. tubercuhis 70-kDa (DnaK-type) Following this immunization strategy, anti-parasite IgG antibodies were induced, in the absence of adjuvants, hsp d18: Recombinant M. leprae l&kDa protein 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheh, 1991
0014-2980/91/1010-2297$3.50 + .25/0
irrespective of the murine genetic responsiveness to the (NANP), epitope, at titers similar to those raised with conjugates given in CFA [5]. In the present study, we report the results of experiments carried out to investigate the requirements for the induction of this phenomenon. We found that heat-shock proteins (hsp) could replace PPD in the conjugate and behave as carrier molecules in the induction of anti-peptide antibodies in the absence of adjuvant.
2 Materials and methods 2.1 Bacteria The M. tuberculosis var. bovis used was the BCG-F strain obtained from the BCG laboratory of the Institut Pasteur, Paris, France [12]. Aliquots of 0.5 ml at lo9 CFU/ml of Dubo's medium were stored at -70°C to improve their stability [13]. For priming, mice were infected with lo6CFU given i.p. Heat-killed BCG was obtained by warming a solution of 2 x lo6CFU BCG/ml in a water bath at 80 "C for 20 min; then 0.5 ml of the heated solution was administered per mouse i.p. as the priming stimulus where indicated. Sonicated BCG was prepared by disruption of the bacilli for a total of 10min in a Soniprep 150 cell sonicator (MSE Scientific Instruments, Sussex, GB). The preparation was centrifuged at 20000 x g for 30 min, and the SN (soluble antigens) was removed. After determination of the protein content, 0.25 mg protein of this sonicate preparation was given i.p. per mouse as the priming stimulus, where indicated. Salmonella typhimurium, strain 661/90, was a kind gift of Dr. F. Sadallah [14], Leishmania major was cultured as previously described [15]. A dose of lo6of each viable microorganism was given i.p. to each mouse as the priming stimulus where indicated.
PPD was purchased from the Statens Serum Institut, Copenhagen, Denmark [16]. PPD at a concentration of 1mg/ml in PBS was used throughout the experiments. The (NANP)a synthetic peptide, from the P falciparum CS protein [6], was produced as described in detail elsewhere [17]. The synthesis of the 21-mer peptide YNRNIVNRLLGDALNGKPEEK (referred to as Py l),corresponding to positions 59-79 of the CS protein of the murine malaria parasite l? yoefii,and its characterization as a T cell epitope in H-2d and H-2b mice, have also been reported [18, 191. Lyophilized material was resuspended in sterile distilled water at 10 mglml, aliquoted, and stored at -70 "C until use. M. bovis BCG GroEL-type 65-kDa hsp (hspR65) was expressed from a recombinant E. coli K12 strain harboring plasmid pRIB1300 [20,21] and purified as described [22]. Recombinant M. tuberculosis DnaK-type 70-kDa hsp (hspR70) was obtained and purified by ATPagarose chromatography according to Mehlert and Young [23], and the M. leprae 18-kDa protein (m118), showing homologies with a low-molecular weight hsp of soybean [24], according to Harris et al. [25]. The purified E. coli GroEL protein was a gift from D. Young. One milligram each of PPD, Pyl peptide, hspR65 , hspR70, m118, GroEL was coupled to 1 mg of (NANP)a peptide in the presence of
0.002% glutaraldehyde for 2 h, and then dialyzed against PBS overnight [5].Aliquots of each conjugate were stored at -70°C until use.
2.3 Mice and immunizations Female, 8-12-week-old BALB/c (H-2d) and C57BL/6 (H2b) mice were bred in our animal facilities. The original breeding pairs were obtained from the Jackson Laboratory, Bar Harbor, ME. Each mouse received i.p. 106 CFU of BCG or equivalent doses of heat-killed or sonicated BCG or live S. typhimurium or L. major promastigotes. Each mouse received 20 p,g of the relevant (NANP)m conjugate i.p. in 0.5 ml PBS, as indicated, on days 14 and 35, according to the schedule previously established [5].
Mice were bled weekly from the retroorbital plexus, the serum separated, and anti-(NANP)m antibodies titered by ELISA using the (NANP)40 peptide as solid phase as described in detail [5, 101. Samples with titers < 1:200 were considered as negative.
Having previously shown that priming with BCG followed by immunizations with PPD-(NANP)a peptide conjugate in the absence of adjuvants induced the production of anti-(NANP)M IgG antibodies [5],we addressed the question of whether the model depended on the use of live BCG as the priming stimulus. Three groups of BALB/c mice (H-2d, genetically nonresponder to (NANP) peptides [9, 111) were primed with live, heat-killed or sonicated
Figure 1. Priming with live BCG is required for the induction of anti-peptide antibodies in BCG-primed mice. Three groups of five BALBk mice were primed by i.p. injection with lo6 CFU BCG, with lo6 heat-killed BCG, or with 250 pg of sonicated BCG in 0.5 ml PBS. On days 14 and 35 after priming, all the mice were immunized with 20 pg of the PPD-(NANP)m peptide conjugate. Ten days after the second conjugate injection the mice were bled and the anti-(NANP)40 IgG antibody titers were measured by ELISA using the peptide as solid phase, as described [5, 101. Results are expressed as the geometrical mean f SEM of the loglo reciprocal of the last serum dilution positive for each group of mice, i.e. giving an absorbance 2 0.2 at 492 nm.
Eur. J. Immunol. 1991. 21: 2297-2302
BCG, and then immunized twice with the PPD-peptide conjugate prepared in PBS.The spleen cells from mice in all of the groups exhibited a proliferative response when cultured in the presence of PPD (data not shown). However, only the mice having received live BCG as the priming stimulus were able to produce anti-(NANP) antibodies following two immunizations with the conjugate without adjuvant (Fig. 1). An anti-peptide antibody response was never detectable in the two groups of mice receiving killed BCG. These data demonstrated that priming with live microorganisms was necessary to induce an anti-peptide antibody response in the absence of adjuvants. 3.2 Mycobacterial carrier molecules are required for the induction of anti-peptide antibodies in BCG-primed mice As the mycobacteria are known to be able to stimulate some compartments of the immune system nonspecifically [26], it was decided to evaluate the effect of priming with live BCG on immunization with a conjugate made of a T cell epitope, bearing no relation to the mycobacteria, coupled to (NA")m peptide. A synthetic peptide, referred to as Pyl, from the CS protein of the murine malaria parasite l? yoefii, was chosen for this purpose as it behaves as a strongTcel1 epitope in BALB/c mice [18, 191. As shown in Fig. 2, Pyl-(NANP)m conjugates were able to overcome the genetic restriction of the antibody response to the (NANP)m peptide in BALB/c mice only when the conjugates were prepared in Freunds adjuvants (i.e. first CFA, then IFA). Two immunizations of the Pyl-(NANP)m conjugate in PBS were totally ineffective in eliciting anti(NANP) antibodies, even in mice primed with live BCG.
Heat-shock proteins as carrier molecules
2299
conjugate. Furthermore, our experiments with the live bacilli suggested that mycobacterial antigen(s) actively produced during BCG infection and present in the PPD preparation would be capable of mediating the phenomenon of anti-peptide antibody production in the absence of adjuvant.
3.3 Other parasites of the MQtrcan provide the priming for the induction of anti-peptide antibodies A group of mycobacterial molecules whose synthesis is up-regulated during BCG infection, the hsp, is known to sensitize T cells in the infected host [27]. Furthermore, hsp fragments had been detected among the antigens of PPD [21]. It then seemed possible that mycobacteria hsp could mediate theT cell sensitization that would be recalled upon immunization with the PPD-containing conjugate. As the hsp molecules are the most conserved proteins throughout evolution identified so far [28, 291, we reasoned that other phylogenetically diverse intracellular microorganisms that lead, as BCG does, to long-term infections of the M a , could provide, through their hsp, the T cell sensitization required for further boosting by the PPD in the conjugates. The enterobacterium S.typhimurium and the protozoan L. major were selected because they express different families of hsp [30]. Indeed, when BALB/c mice were infected with live BCG or S. typhimurium or L. major, and subsequently challenged with the PPD-(NANP)m conjugate without adjuvant, a comparable anti-(NANP)4o IgG antibody response was detectable in all the mice (Fig. 3).
These experiments strongly suggested that BCG-primed T cells were providing the "help" required for the production of anti-peptide antibodies. These results were also in agreement with those reported by Lachmann et al. [4] showing that non-mycobacterial carrier proteins, such as KLH, were unable to replace the PPD portion of the Figure 3. Other parasites of the M@ can provide the priming for the induction of anti-peptideantibodies. On day 0, three groups of five BALB/c mice were primed i.p. with 106 BCG, L. major, or S.typhimurium (strain 6 1 m ) . Immunizations with the PPD(NANP)40conjugate and expression of results [anti-(NANP)IgG antibody titers] are as described in the legend of Fig. 1.
1.5 2 2.5 3 Log10 anti-(NANP) IgG titers Figure 2. Mycobacterial carrier molecules are required for the induction of anti-peptide antibodies in BCG-primed mice. Three groups of four BALBk mice were immunized as follows: (squares) mice received the priming with live BCG (106 CFU) i.p. and then two i.p. immunizations with the Pyl-(NANP)m conjugate (20 pg) in PBS; (triangles) unprimed mice received twice the Pyl(NANP)m conjugate in PBS; (circles) unprimed mice received the Pyl-(NANP)a conjugate in complete and then incomplete Freund's adjuvant. Anti-(NANP) IgG antibodies from day 49 are shown as the mean absorbance at 492 nm for each group of mice obtained from serial dilution of the serum samples tested in the
(NANP)40 ELISA [S, lo].
3.4 Mycobacterial hsp act as carrier molecules for the induction of anti-peptide antibodies in BCG-primed mice without adjuvant
To test the hypothesis that hsp present in the PPD preparation were exerting the carrier effect on the production of anti-peptide antibodies, the purified recombinant M. bovis GroELtype hspR65 was coupled to the (NANP)40 peptide and the conjugate used to immunize live BCG-primed C57BL/6 (H-2b, genetically responsive to (NANP), peptides [9-111) and BALBlc nonresponder mice. As illustrated in Fig. 4, the hspR65-(NANP)a conjugate given in PBS was able to induce an anti-(NANP) IgG antibody response in BCG-primed C57BL/6 mice after a
1300
Eur. J. Immunol. 1991. 21: 2297-2302
A. R . Lussow, C. Barrios. J.Van Embden et al
3.5 The E . coli GroEL hsp behaves as a carrier molecule in BCG-primed mice
Figure 4. Carncr effect of the M . h0L.i.c.hspR65. One group of five C57BL/6 and one group o f five BALB!c mice were primed with live BCG, as described for Figure 1. Fourteen days later, both groups were immunized with 30 11g of the hspR65-(NANP),U conjirgnre delivered i. p. in 0.5 nil I’HS. ’l‘lic conjugate irnmrinimioii w a s repeated in the same manner on da? 3.5 after BC’Ci priming. Results [anti-(NANP) IgG antibody titers] from serum samples te5ted after the first (day 1s: hatched columns) or after the second (day 19: solid culumns) conjugate immunizations are shown a n d cxprcssed as in the lcgend of Fig. I .
single immunization. Further. after two immunizations with this conjugate without adjuvant, the genetic rcstriction of the antibody response to t h e peptide was overcome in BALB/c mice, resulting in a strong anti-peptide IgG antibody response. Therefore, immunization with the synthetic peptide linked to the hspR65 moleculc produced results identical to those we previously obtained using the PPD-(NANP)4,1conjug;itc in thcsc two strain4 of micc 151. T h e GroEL-type hsp is not the only class of stress proteins produced by the mycohacteria 131 1. Therefore. t h e role other types o f mycob;ictcriiil hsp coiild Ii;ivc iii this systcm was investigated. .l‘hc recombinant M . tubcrculosi~DnaKtype hspR70 [?3] and the recombinant M . leprae 18-kDa protein (mllX). which has homologies with a lon.-molecular-weight hsp of soybean 1241, were conjugatcd to the ( N A N P ) a peptide and used to immunize BALB/c mice previously primed with livc BCG. The results o f thcsc cxpcrimcntx arc‘ s h o w n in f;ig. 5 . I%oth thc hspK65 and hspR70 molecules from the closely related BCG and M . tuherculmis bacilli. respectively, could mediate the induction of anti-peptide IgG antibodies in the absence of adjuvants. I n contrast, the M . leprue 18-kDa molecule did not act as a carrier for the induction of anti-(NANP) antil)oclics i n BC’G-priiiictl micc. Carrier
2
3
4
Log10 anti-(NANP) IgG titers f i g i t r e 5 Cai ricr cffcct of the ,M ruberculou hsp70 Three groups of BALBk mice wcrc primed ai t h live BCG and immunixd (as clc\cril)ctl for Fig I ) kith 20 pg 01 liy05-(hASl’)qt. hy)70 (NAN1’)q,. or nillH-(NANP)40conjugdtc, given i.p in 0.5 1111 1’1)s on d;i)s 14 dnd 35 after BCG priming Results [anti-(NANP) IgG antibody titers from clay 19l A I L cxprcsscd as in Fig 1
Tdken together, these data demonstrated that at least the mycobacterial hsp belonging t o the G r o E L and DnaK families behaved as carrier molecules for the induction of anti-peptide antibodies in BCG-primed mice without adjuvants. We then ;iddressed t h e question o f whether a non-mycobaclerial hsp belonging to the same class could mediate the same effect.Thc E. coli GroEL hsp was used in these experiments: it mas conjugated to the ( N A N P ) a peptide and the conjugate used t o immunize BALBlc mice previously primed with live BCG. As Fig. 6 shows, mice immunized twicc with thc GroEl ,-(hAhT)u, conjugate without adjuvant produced anti-( NANP) IgG antibodics at titers similar t o those obtaincd with the hspR65-(NANP)al conjugate. Interestingly. similar antibody titers were also obtained when BCG-primed mice were first immunized with the GroEL-(NAhT)lo and then with the hspR65(NANP)al conjugate, or vice versa. C57BL/6 mice gave totally similar results (data not s l i o w i i ) .
E. coli GroEL hsp behaves as a carrier molecule in IN ’( i-primed micc. (;roup\ o f thrcc to Ioiii IMLD/c rnicc wcrc piiirictl w i t h livc H C Y ; :is tlc.icrihcd i n Fig. I , and then irnrnunizccl 14 and 35 days after with 20 big o f the h ~ p R 6 5 - ( S A W ) ~oro Figitre(,.
GroEI.-(KASP)q, conjugate. given i.p. in 0.5 ml PBS. Results [anti-(NA\P) IgCi antibody titen from day 491 are expressed as in Fig. 1.
4 Discussion I lcrc wc have shown that mycolwctcrial hsp t>ch;ivca s carrier molccules for the induction ol anti-peptide antihodies in mice previously primed with BCG. In all cases. the antibod!. production took place with the hsp-peptide conjugate given in the uhsclrtct of adjuvants, and required a previous priming with lille BCG. The data reported in this paper ;idd to those already rcportctl h y u\ conccrniiig the i i w o f I’PI) as :I carrier for the induction of anti-(NANP) antibodics in BCG-primed mice 151. It is then reasonable to hypothesize that the effects mediated by BCG priming and immunization with PPD can be ascribed at least i n part to the hsp of 65 and 70 kDa actively produced hy mycobacteria.
Since in our model the priming of mice had t o be mediated by live microorganisms. it seemed that Stress proteins acre likcl! to mediate the Tcell sensitizarion that would be
Eur. J. Immunol. 1991.21: 2297-2302
recalled to help (NANP)-specific B cells. In fact, hsp production is increased during BCG infection of the MQ, and T cell sensitization is often associated with mycobacterial infection [27,32]. However, priming with heat-killed or sonicated BCG was ineffective in inducing anti-peptide antibodies following immunization with the PPD(NANP)4" conjugate. This was possible due to the activation of hsp-specific Tcells at a lower frequency than that achieved after a long-term infection by BCG which would allow persistent availability of antigen. Alternatively, it is possible that different T cell subsets were activated after priming of mice with live or killed BCG, leading in one case, but not in the other, to the helper effect required for the induction of anti-peptide antibodies. It is interesting to note that microorganisms as phylogenetically diverse as a mycobacterium (BCG), an enterobacterium (S. typhimurium), and a protozoan (L. major),were all able to prime the mice for the production of anti-peptide antibody response after immunization with the PPD (mycobacterial antigens)-(NANP)a conjugate. Since all these organisms express hsp and lead to long-term infections of the MQ, [30,33], and since the hsp molecules are the most conserved proteins throughout evolution [28,29], it is possible that in these experiments priming was mediated through conserved homologous epitopes also present in the PPD preparations: in fact, it has been shown that fragments of the mycobacterial65-kDa hsp are present in PPD [21]. This hypothesis is also supported by our findings showing that BCG-primed mice produced anti-(NANP)a IgG antibodies after immunization with the peptide conjugated with the E. coli GroEL hsp, which has a high degree of homology with the mycobacterial 65-kDa hsp [34]. This would suggest that Tcells induced after priming with live BCG (or other Ma-infecting organisms) can cross-react with hsp of different origins. This would be in line with previous reports by others, showing that indeed this appears to be the case in certain circumstances [35-381. The finding that mycobacterial hsp can successfully substitute for PPD as carriers for the induction of strong antibody responses to defined B cell epitopes in the absence of adjuvant could, on one hand, be expected, since many Tcell epitopes have been identified within them [39]. On the other hand, it was surprising (a) that priming with live BCG was required, and (b) that this, in turn, eliminated the need for the use of adjuvants that are normally required when other carrier molecules are utilized. It is not yet clear through which mechanisms hsp molecules exert their carrier effect. From our data it is evident that the carrier effect was mediated by GroEL- and DnaK-type hsp, but not by the ml18, a 18-kDa protein of M. leprae sharing homologies with a low-molecular weight hsp of soybean [24].These negative results with m118 molecule may be due to the absence of a homologous molecule in M . tuberculosis complex, or to a limited homology between this molecule and possible equivalent hsp from BCG. Some comments have to be made on the possibility of inducing autoimmune reactions, due to the homologies existingbetween some mycobacterial molecules and certain host tissue antigens, and between bacterial and human hsp. For example, antibodies to M. bovis 65-kDa hsp have been reported to cross-react with human antigens [40], and antibodies to the 70-kDa hsp of the malaria parasite
Heat-shock proteins as carrier molecules
2301
R falciparum have been shown to react also with the homologous human hsp [41]. Moreover, T cell epitopes of the 65-kDa hsp, associated with mycobacterial infection or the use of CFA, have been known to induce adjuvant arthritis in certain susceptible strains of rats [42]. However, immunization with the 65-kDa hsp has been shown to protect these animals from such autoimmune reactions [20,43]. Finally, no homology has yet been found between the 70-kDa (DnaK-type) hsp and cartilage proteins [U].
The carrier effect of mycobacterial hsp and the fact that the antibody response induced takes place without requiring adjuvant, make this approach of immunization particularly interesting for the development of vaccine strategies. Since BCG is commonly and widely used very early in life as a vaccine against tuberculosis, the use of hsp in vaccine formulations may present several advantages over other carrier molecules. Many Tcell epitopes have been identified on the M. bovis 65-kDa hsp, and the entire protein has been demonstrated to be immunogenic in a wide variety of mouse strains with differing MHC haplotypes [39]. Furthermore, beyond their possible use in individuals already sensitized to the mycobacteria (through vaccination or natural contact), the hsp molecules may also be able to interact with hsp-specificT cell populations found in apparently healthy adults [36]. Experiments done in our laboratory in order to investigate possible eventual application of this vaccination strategy for human use, have shown that the production of anti-peptide IgG antibodies is induced also when BCG and conjugates are given i.d., and that this approach is also applicable to synthetic peptides shorter than the malaria (NANP)a peptide, i.e. with 24-mer peptides from HIV envelope antigens (manuscript in preparation). Finally, immunizations have been started to investigate whether this approach of immunization also applies to non-human primates. We would like to thank Drs. G. Milon and F! David for critical reading of the manuscript, Ms. C. Tougnefor technical help, Dr. D. Young for the E. coli GroEL hsp, Ms. E Sadallah for S. typhimurium, Ms. K . Hug for L. major parasites, Mrs. M. Gheorghiu for BCG. Received May 27, 1991.
5 References 1 Lussow, A. R., Aguado, M. T., Del Giudice, G . and Lambert, P.-H., Immunol. Lett. 1990. 25: 255. 2 Altman, A. and Dixon, F. J., Adv. Vet. Sci. Comp. Med. 1989.33: 303. 3 Warren, H. S. and Chedid, L. A . , CRC Crit. Rev. Immunol. 1988. 8: 83. 4 Lachmann, l? J., Strangeways, L.,Vyakarmam, A. and Evan, G., Ciba Found. Symp. 1986. 119: 25. 5 Lussow, A. R., Del Giudice, G . , Rknia, L., Mazier, D., Verhave, J. I?, Verdini, A. S., Pessi, A . , Louis, J. A. and Lambert, P.-H., Proc. Matl. Acad. Sci. USA 1990. 87: 2960. 6 Nussenzweig,V. and Nussenzweig, R. S., Adv. Immunol. 1989. 45: 283. 7 Ballou, W. R., Hoffman, S. L., Sherwood, J. A., Hollingdale, M. R.,Neva, F. A., Hockmeyer,W.T., Gordon, D. M., Schneider, I.,Wirtz, R. A. ,Young, J. F.,Wassennan, G. F., Reeve, P., Diggs, C. L. and Chulay, J. D., Lancet 1987. i: 1277.
2302
A. R. Lussow, C. Barrios, J.Van Embden et al.
8 Herrington, D. A., Clyde, D. F., Losonsky, G., Cortesia, M., Murphy, J. R., Baqar, S., Felix, A. M., Heimer, E. I?, Gillessen, D., Nardin, E., Nussenzweig, R. S., Nussenzweig,V., Hollingdale, M. R. and Levine, M. M., Nature 1987. 328: 257. 9 Good, M. F., Berzofsky, J. A., Maloy,W. L., Hayashi,Y, Fujii, N., Hockmeyer,W. and Miller, L. H., J. Exp. Med. 1986.164: 655. 10 Del Giudice, G., Cooper, J. A., Merino, J. ,Verdini, A. S., Pessi, A.,Togna, A. R., Engers, H. D., Corradin, G. and Lambert, P.-H.,J. Immunol. 1986. 137: 2952. 11 Togna, A. R., Del Giudice, G. ,Verdini, A. S., Bonelli, F., Pessi, A., Engers, H. D. and Corradin, G., J. Immunol. 1986. 137: 2956. 12 Gheorghiu, M., Chambon, L., Serie, C. and Juvin-Calmette, D., Dev. Biol. Stand. 1977. 38: 29. 13 Gheorghiu, M. and Lagrange, l?, Ann. Immunol. (Inst. Pasteur) 1983. 134: 125. 14 Sadallah, F., Brighouse, G., Del Giudice, G., Drager-Dayal, R., Hocine, M. and Lambert, P.-H., J. Infect. Dis. 1990.161: 59. 15 Louis, J. A . , Moedder, E., Behin, R. and Engers, H. D., Eur. J. Immunol. 1979. 9: 841. 16 Magnusson, M. and Bentzon, M. W., Bull. W H . 0 . 1958. 19: 829. 17 Esposito, G., Pessi, A. and Verdini, A. S., Biopolymers 1989. 28: 225. 18 Grillot, D., Michel, M., Miiller, I., Tougne, C., Rknia, L., Mazier, D., Corradin, G., Lambert, l?-H., Louis, J. A. and Del Giudice, G., Eur. J. Imrnunol. 1990. 20: 1215. 19 Del Giudice, G., Grillot, D., R h i a , L., Miiller, I., Conadin, G., Louis, J. A., Mazier, D. and Lambert, P.-H., Irnmunol. Lett. 1990. 25: 59. 20 Van Eden, W., Thole, J., Van der Zee, R., Noordzij, A., Van Embden, J., Hensen, E. and Cohen, I., Nature 1988. 331: 171. 21 Thole, J. E., Dauwerse, H. G., Das, P. K., Groothuis, D. G., Schouls, L. M. and Van Embden, J. D., Infect. Immun. 1985. 50: 800. 22 Thole, J. E. ,Van Schooten,W. C., Keulen,W. J., Hermans, P.W., Janson, A. A., DeVries, R. R.,Kolk, A. H. and Van Embden, J. D., Infect. Immun. 1987. 55: 1466. 23 Mehlert, A. and Young, D. B., Mol. Microbiol. 1989. 3: 125. 24 Nerland, A. H., Mustafa, A. S., Sweetser, D., Godal, T. and Young, R., J. Bacteriol. 1988. 170: 5919.
Eur. J. Immunol. 1991. 21: 2297-2302 25 Harris, D. I?, Backstrom, B. T., Booth, R. J., Love, S. G., Harding, D. R. and Watson,J. D., J. Immunol. 1989. 143: 2006. 26 Fromrnel, D. and Lagrange,€?,Parasitol. Today 1989. 5: 188. 27 Kaufmann, S. H. E., Schoel, B., Wand-Wiirttenberger, A., Steinhoff, U., Munk, M. and Koga, T., Curr. Top. Microbiol. Immunol. 1990. 155: 125. 28 Lindquist, S. and Graig, E., Annu. Rev. Genet. 1988. 22: 631. 29 Young, R. A. and Elliott, T. J., Cell 1989. 59: 5 . 30 Buchmeier, N. A. and Heffron, F., Science 1990.248: 730. 31 Shinnick, T. M., Curr. Top. Microbiol. Immunol. 1991. 167: 145. 32 Milon, G., Lebastard, M. and Marchal, G., Immunol. Lett. 1985. 11: 189. 33 Lee, M. G. S., Atkinson, B. L., Giannini, S. H. and Van der Ploeg, L. H. T., Nucleic Acids Res. 1988. 16: 9567. 34 Shinnick, T. M., Vodkin, M. H. and Williams, J. D., Infect. Immun. 1988. 56: 446. 35 Ottenhoff,T. H.,VanEmbden, J. D.,Thole, J. E. and Kiessling, R., J. Exp. Med. 1988. 168: 1947. 36 Munk, M., Schoel, B., Modrow, S., Karr, R.,Young, R. and Kaufmann, S. H. E., J. Immunol. 1989. 143: 2844. 37 Koga,T.,Wand-Wurttenberger,A., DeBruyn, J., Munk, M. E., Schoel, B. and Kaufmann, S. H. E., Science 1989. 245: 1112. 38 Lamb, J. R., Mendez-Sampirio, P., Mehlert, A., So, A., Rothbard, J., Jindal, S.,Young, R. A. and Young, D. B., Int. Immunol. 1989. I : 191. 39 Brett, S. J., Lamb, J., Cox, J., Rothbard, J., Mehlert, A. and Ivanyi, J., Eur. J. Immunol. 1989. 19: 1303. 40 Aguas, A. P., Esaguy, N., Sunkel, C. E. and Silva, M.T., Infect. Immun. 1990. 58: 1461. 41 Mattei, D., Scherf, A., Bensaude, 0. and Pereira da Silva, L., Eur. J. Immunol. 1989. 19: 1823. 42 Van Eden,W., Holoshitz, J., Nevo, Z., Frenkel, A., Klajman, A. and Cohen, I. R., Proc. Natl. Acad. Sci. USA 1985. 82: 5117. 43 Van der Broek, M. F., Hogervorst, E. J. ,Van Bruggen, M. ,Van Eden, W. ,Van der Zee, R. and Van den Berg,W., J. Exp. Med. 1989. 170: 449. 44 McLean, L. ,Vinrow,V. and Blake, D., J. Exp. Pathol. 1990. 71: 295.
