Immunogenicity of Mycobacterium ulcerans Hsp65 and protective efficacy of a Mycobacterium leprae Hsp65-based DNA vaccine against Buruli ulcer

Microbes and Infection 8 (2006) 2075e2081 www.elsevier.com/locate/micinf

Original article

Immunogenicity of Mycobacterium ulcerans Hsp65 and protective efficacy of a Mycobacterium leprae Hsp65-based DNA vaccine against Buruli ulcer Emmanuelle Coutanceau a, Pierre Legras b, Laurent Marsollier a, Gilles Reysset a, Stewart T. Cole a, Caroline Demangel a,* a

Unite´ de Ge´ne´tique Mole´culaire Bacte´rienne, Institut Pasteur, 28 rue du Docteur Roux, 75724 Paris Cedex 15, France b Animalerie Hospitalo-Universitaire, CHU, Angers, France Received 30 December 2005; accepted 17 March 2006 Available online 24 May 2006

Abstract Buruli ulcer, a disease caused by Mycobacterium ulcerans, is emerging as an increasingly important cause of morbidity throughout the world, for which surgery is the only efficient treatment to date. The aim of this work was to identify potential vaccine candidates in an experimental model of mouse infection. In BALB/c mice infected with M. ulcerans subcutaneously, Hsp65 appeared to be an immunodominant antigen eliciting both humoral and cellular responses. However, vaccination of mice with a DNA vector encoding Mycobacterium leprae Hsp65 only poorly limited the progression of M. ulcerans infection. In contrast, a substantial degree of protection was conferred by subcutaneous vaccination with BCG, suggesting that BCG antigens that are conserved in M. ulcerans, such as TB10.4, the 19 kDa antigen, PstS3 and Hsp70, may be interesting to consider as subunit vaccines in future prospects. Ó 2006 Elsevier SAS. All rights reserved. Keywords: Mycobacterium ulcerans; Buruli ulcer; Hsp65; Vaccination

1. Introduction Buruli ulcer is the most common mycobacterial infection in humans, after tuberculosis and leprosy, with prevalence rates reaching 22% in some communities of West Africa. In highly endemic districts in Ghana, prevalence exceeds 150/100,000 individuals [1]. Buruli ulcer starts as a painless nodule or papule in the skin which, if left untreated, leads to massive skin ulceration and permanent disabilities. Observations made in animal models suggest that the causative agent Mycobacterium ulcerans is transiently intracellular during the early stages of infection, and then accumulates extracellularly in cutaneous and subcutaneous lesions [2,3]. Destruction of the tissues is closely associated with the production of a macrolide toxin,

* Corresponding author. Tel.: þ33 1 45 68 84 49; fax: þ33 1 40 61 35 83. E-mail address: [email protected] (C. Demangel). 1286-4579/$ - see front matter Ó 2006 Elsevier SAS. All rights reserved. doi:10.1016/j.micinf.2006.03.009

mycolactone, by M. ulcerans [4]. Efforts to design efficient antibiotic therapies have been ineffective so far. Despite promising results obtained in early human lesions with a combination of rifampicin and amikacin [5], the treatment of choice remains surgery [1]. The disease therefore impacts deeply on rural economy and on the scarce health resources of the endemic countries. In this context, research into prevention of Buruli ulcer is urgently needed to control the spread of the disease. No specific vaccine against Buruli ulcer is available yet. Several lines of evidence suggest that the anti-tuberculous vaccine Mycobacterium bovis BCG (BCG) may confer crossreactive protection. BCG vaccination showed a protective effect against severe forms of the disease (osteomyelitis) in children suffering from Buruli disease in Benin [6]. Moreover, randomized controlled trials of BCG vaccination in Uganda have concluded that BCG protects significantly against Buruli ulcer [7]. However, the protection afforded by BCG vaccination was short-term and declined with time. Vaccination

2076

E. Coutanceau et al. / Microbes and Infection 8 (2006) 2075e2081

with BCG is therefore not sufficient to protect against Buruli ulcer and alternative strategies must be developed. Back in 1957, studies performed in mice showed a protective effect of BCG against Buruli ulcer, when the vaccine was administered in one footpad and M. ulcerans in the other [8]. More recently, multiplication of M. ulcerans was reduced in the footpad of mice vaccinated by intravenous injection of BCG [9]. In the latter study, the efficacy of a DNA vaccine encoding Ag85A from BCG showed a marked protective effect, superior to that afforded by BCG. With the aim of defining other potential vaccine candidates, we explored the humoral and cellular immunogenicity of M. ulcerans proteins in BALB/c mice experimentally infected in the tail. This study revealed that the 65 kDa heat shock protein (Hsp65, GroEL2) was a major antigen in this model, eliciting both interferon (IFN)-g production and potent antibody responses. However, when mice were immunized with a DNA vaccine encoding the closely related Mycobacterium leprae Hsp65, minimal protection against Buruli ulcer was induced, compared to BCG. 2. Materials and methods 2.1. Bacteria M. ulcerans (strain 1615, ATCC 35840) was obtained from the Trudeau collection. Mycobacteria were grown on Lo¨wenstein-Jensen medium at 32
C for 6 weeks. BCG Pasteur 1173P2 was from a stock held at the Institut Pasteur, and was cultivated at 37
C in Middlebrook 7H9 supplemented with acidealbuminedextrose 10% (ADC, Becton Dickinson). 2.2. Animals Six-week-old BalB/cbyJIco (BALB/c) female mice were purchased from Charles Rivers Laboratories and housed in the Animalerie Hospitalo-Universitaire (CHU, Angers, France) or in the Institut Pasteur’s animal facilities. 2.3. IFN-g ELISA and ELISPOT assays Single cell suspensions were prepared from pooled spleens by sieving through 200 mm mesh, and resuspending the cells in synthetic HL-1 medium (Cambrex, MD, USA) supplemented with 2 mM L-glutamine, 100 IU penicillin/ml and 100 mg streptomycin/ml. Cultures were seeded in microtitre plates at a density of 4  105 splenocytes per well. Cells were cultivated in the presence of purified Ag85 complex, Hsp65, KatG, PPD (Veterinary Laboratories Agency), Concanavalin A (Sigma), or in the absence of stimulatory agent for 2 days at 37
C. The presence of IFN-g in culture supernatants was then assessed using antibodies from the mouse IFN-g ELISA set (BD Biosciences). For ELISPOT assays, nitrocellulose wells of an Immobilon-P plate (Millipore, Bedford, MA) were coated with an anti-IFN-g monoclonal antibody (mAb) (AN18), washed and coated with PBS containing 2% FCS. Cell suspensions were then plated at 5  105 cells per well

with either medium alone, PPD or Concanavalin A. The plates were incubated at 37
C for 18 h and then extensively washed with PBS. Subsequently, biotinylated XMG1.2 mAb was added to the wells. After 2 h incubation and washing, the plates were incubated with avidinealkaline phosphatase (Sigma). The presence of IFN-g-producing cells was detected using an alkaline phosphatase substrate kit (Bio-Rad Laboratory, Hercules, CA). 2.4. Bacterial and tissue lysates Mycobacterial cell lysates from M. ulcerans or BCG were prepared as follows. Culture samples were centrifuged and bacterial pellets washed twice with 20 mM Tris buffer (pH 7.5) before resuspension in 500 ml of the same buffer with complete EDTA-free protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany). Bacteria were lysed by three consecutive freezeethaw cycles using liquid nitrogen and an 80
C water bath. Ulcerated and healthy tissue extracts were prepared by homogenizing skin biopsies in 20 mM Tris buffer (pH 7.5) supplemented with anti-protease inhibitors, and shaking on an MM300 apparatus for 10 min at maximum speed with 500 ml of acid-washed 106 mm diameter glass beads (Sigma, St Louis, MI). Beads and unbroken cells were removed by centrifugation at 5000  g for 30 min, and the resulting supernatants centrifuged at 15,000  g for 30 min. The protein concentration in both bacterial and tissue lysates was determined using a Bio-Rad protein assay. 2.5. Western blot analysis Immunodetection of Hsp65 was performed with a monoclonal antibody directed against Mycobacterium tuberculosis Hsp65 (CS-44, Colorado State University), or pooled immune anti-sera from eight BALB/c mice, on 10 mg protein lysate samples. Recombinant M. tuberculosis Hsp65 (Colorado State University) was used as control. Briefly, proteins were separated by 12% SDSepolyacrylamide gel electrophoresis and transferred onto a nitrocellulose membrane. Detection used anti-mouse horseradish peroxidase-conjugated IgG (Amersham) and ECLÔ Western Blotting Detection Reagent. 2.6. Immunization procedures BALB/c mice (n ¼ 7) were injected three times, at 2week intervals in both quadriceps with 2  50 mg of CsClpurified pCDNA3-Hsp65, pCDNA3-Ag85B, or pCDNA3-mix (2  50 mg of each plasmid), or control pCDNA3. Mice were challenged by M. ulcerans infection 3 weeks after the last injection. Control mice were vaccinated with 104 BCG at the time of the first DNA injection (7 weeks before challenge). 2.7. Enumeration of M. ulcerans in infected tails Tail tissue specimens were minced with disposable scalpels in a petri dish, and ground with a PottereElvehjem homogenizer (size 22; Kimble/Kontes, Vineland, NJ) in 0.15 M

E. Coutanceau et al. / Microbes and Infection 8 (2006) 2075e2081

NaCl. A 10 ml aliquot of this suspension was stained by the ZiehleNeelsen procedure, in order to determine the content in acid-fast bacilli (AFB). The rest of the suspension was decontaminated with an equal volume of N-acetyl-L-cysteine sodium hydroxide (2%), and 0.2 ml aliquots of serial dilutions used to inoculate Lo¨wenstein-Jensen slants (Sanofi Diagnostics Pasteur). Colonies were counted after incubation of the cultures for 6 weeks at 32
C. Both AFB and colony forming units (cfu) were determined on individual mice.

responses specific for mycobacterial antigens. Animals immunized with equivalent doses of live M. bovis BCG (BCG) via the same route, and unimmunized controls were included. Infection with M. ulcerans led to the generation of weak IFN-g responses specific for mycobacterial antigens in the spleen 4 weeks following infection, as demonstrated by the restimulation of splenocytes in the presence of M. ulcerans lysate or PPD (Fig. 2). In particular, IFN-g responses to Antigen 85B (Ag85B) and to Hsp65 were observed, whereas no reactivity to KatG could be detected. Although significant, these anti-mycobacterial responses were weaker than those developed in BCG-immunized mice (Fig. 2), suggesting that M. ulcerans is less efficient than BCG at stimulating Th1oriented immune responses in vivo. When BALB/c mice infected with M. ulcerans were analysed for serum reactivity to an M. ulcerans bacterial lysate in a Western blot assay, the antibody response appeared to be focused against three major antigens of 65 kDa, 50 kDa and 15 kDa (Fig. 3). The 65 kDa protein was identified as Hsp65, since a monoclonal antibody raised against M. tuberculosis Hsp65 bound the antigen in both M. ulcerans and BCG bacterial lysate preparations (data not shown). Interestingly, the antibody response to Hsp65 was detected as soon as 2 weeks post infection with M. ulcerans, and persisted after 8 weeks (data not shown). Similarly in humans, sera from Buruli ulcer patients reacted with the recombinant protein, suggesting that Hsp65 is produced in significant amounts by M. ulcerans bacilli in vivo (data not shown). Consistently, Hsp65 was detected in tissues harvested from necrotic skin lesions of mice infected with M. ulcerans, whereas the antigen could not be detected in a distant healthy skin preparation (Fig. 3).

2.8. Bioinformatics Amino acid sequences of antigens from M. tuberculosis, M. leprae and M. ulcerans were downloaded from http:// genolist.pasteur.fr/TubercuList, http://genolist.pasteur.fr/Leproma, http://www.pasteur.fr/recherche/unites/Lgmb/mycogenomics.html databases, respectively. Sequence alignments were performed using ClustalV and BLAST analysis softwares. 3. Results 3.1. Multiplication of M. ulcerans in a BALB/c mouse model of infection With the aim of identifying the immunodominant antigens of M. ulcerans, we established a mouse model of infection in BALB/c mice in which live bacilli were injected in the tail via the subcutaneous route. Fig. 1A shows that inoculation of 10 to 104 M. ulcerans AFB resulted in a progressive infection, with bacterial loads ranging from 5  105 to 5  107 cfu 8 weeks later, in an inoculum dose-dependent manner. In mice infected with 104 AFB, multiplication of bacilli was visible after 4 weeks (Fig. 1B). At this time point, the site of infection was swollen and inflamed with no apparent necrotic lesion, skin ulcerations developing after 8 weeks (Fig. 1C).

3.3. M. ulcerans Hsp65 is highly homologous to its M. tuberculosis and M. leprae orthologues The observation that Hsp65 of M. ulcerans cross reacts with a monoclonal antibody raised against M. tuberculosis Hsp65 suggests that the M. ulcerans antigen is closely related to its orthologues in other mycobacterial species. To examine further the degree of homology between Hsp65 of M. ulcerans,

3.2. Hsp65 is a major immunogenic component of M. ulcerans Mice infected with 104 M. ulcerans AFB in the tail were used to investigate the generation of humoral and cellular

B

108

cfus in tail lesion

cfus in tail lesion

A

107

106

2077

C

108

107 106 105 (AFB)

105 10

102

103

Inoculum (AFB)

104

104 0

4

Week

8

Week 4

Week 8

Fig. 1. Multiplication of M. ulcerans in BALB/c mice infected via the subcutaneous route in the tail. (A) Mean cfu per mouse tail 8 weeks post infection with increasing doses of live bacilli; (B) Kinetics of mean cfu in the tail of mice following infection with 104 AFB bacilli. Results are mean  SD. Eight mice were tested individually for each time point. (C) Representative picture of the infection site four and 8 weeks post infection with 104 M. ulcerans AFB.

E. Coutanceau et al. / Microbes and Infection 8 (2006) 2075e2081

2078

Immune sera

medium

kDa

PPD Mu lysate

1200

ulcerated

healthy

180 80 65 50

Ag85B

(pg/ml)

BCG

Hsp65 KatG

IFN-

Mu

-Hsp65

65kDa

900

30 600 15 300

0

unimmunized

BCG

Mu

Fig. 2. IFN-g production in splenocyte cultures from mice immunized subcutaneously with 104 live M. ulcerans or BCG, or from unimmunized controls. Cells were restimulated for 2 days in vitro with PPD, a M. ulcerans bacterial lysate preparation, purified Ag85B, Hsp65 or KatG. Data are means  SD of pg/ml in culture supernatants tested in duplicates from pools of three mice.

M. leprae and M. tuberculosis, their amino acid sequences were aligned. All three Hsp65 antigens shared strong similarity, with 96% and 94% of sequence identity between M. ulcerans Hsp65 and the M. leprae and M. tuberculosis proteins, respectively (data not shown).

Fig. 3. Western blot analysis of BALB/c mice immune sera sampled 4 weeks post infection (diluted 1:100) against M. ulcerans or BCG bacterial lysate, of a monoclonal antibody anti-Hsp65 against ulcerated and healthy skin preparations.

Ag85B were included (pCDNA3-Ag85B) [12], as DNA vaccination with the Ag85 complex A member from BCG was previously reported to confer significant protection against Buruli ulcer [9]. A group of mice immunized with an equimolar mixture of both pCDNA3-Hsp65 and pCDNA3-Ag85B (pCDNA3-Mix) was also examined. Animals vaccinated

A

3.4. Vaccination with a DNA vector encoding Hsp65 confers minimal protection against infection with M. ulcerans

65 kDa

pCDNA3

B Frequency of IFN- producing cells/106 splenocytes

Since Hsp65 is an immunodominant antigen in BALB/c mice experimentally infected with M. ulcerans, we examined whether vaccination with this antigen confers protection against Buruli ulcer. We used a DNA vector encoding M. leprae Hsp65 (pCDNA3-Hsp65), since this vaccine had a pronounced therapeutic action [10] and a potent protective effect [11] against tuberculosis infection in the BALB/c mouse model. Animals immunized with three intramuscular injections of 100 mg pCDNA3-Hsp65 developed an antibody response to Hsp65 2 weeks after the last injection, as evidenced by serum reactivity against purified Hsp65 (Fig. 4A). In contrast, no signal could be detected with sera from mice injected with the control vector pCDNA3. Moreover, immunization with pCDNA3-Hsp65 but not with pCDNA3 induced the generation of IFN-g producing cells specific for Hsp65 in the spleen of the vaccinated animals (Fig. 4B). To investigate the protective efficacy of pCDNA3-Hsp65, mice immunized according to this protocol were challenged by subcutaneous injection of 104 M. ulcerans AFB in the tail 3 weeks after the last injection, and sacrificed 6 weeks later. Mice immunized with a pCDNA3 plasmid encoding

pCDNA3-Hsp65

100 90

pCDNA3 pCDNA3-Hsp65

80 70 60 50 40 30 20 10 0

Post-vaccination

Post-challenge

Fig. 4. (A) Mice vaccinated with plasmid pCDNA3-Hsp65 develop specific antibodies. Western blot detection of purified Hsp65 by sera (diluted 1:100) from mice vaccinated with pCDNA3-Hsp65 or pCDNA3 2 weeks after the last of three injections. (B) Mice vaccinated with pCDNA3-Hsp65 develop a Th1 oriented cellular response against Hsp65. The frequency of IFN-g producing cells in spleens of mice vaccinated with pCDNA3 or pCDNA3-Hsp65 is shown 2 weeks after the last of three injections (post-vaccination), and 4 weeks after subsequent challenge with M. ulcerans (post-challenge).

E. Coutanceau et al. / Microbes and Infection 8 (2006) 2075e2081

with pCDNA3-Hsp65, pCDNA3-Ag85B, or pCDNA3-Mix were compared for M. ulcerans multiplication with mice vaccinated with pCDNA3 control DNA or with BCG. The IFN-g response to Hsp65 was significantly increased in mice vaccinated with pCDNA3-Hsp65 4 weeks after challenge, suggesting that the cellular immune responses to this antigen were boosted during the course of infection with M. ulcerans (Fig. 4B). Six weeks post infection, all DNA vaccinated mice developed lesions, whereas no apparent ulceration was observed in BCG-vaccinated animals. Accordingly, the bacterial load in infected tails was significantly lower in the BCG vaccinated group, with a 3.13 DLog10 difference compared to DNAtreated controls (Fig. 5). Evaluation of the cfu in the tails of DNA vaccinated mice nevertheless revealed differences. Compared to the control DNA-treated group, the multiplication of M. ulcerans was suppressed by DLog10 ¼ 0.66 in mice immunized with pCDNA3-Hsp65, but not significantly with pCDNA3-Ag85B. Co-immunization of pCDNA3-Ag85B with pCDNA3-Hsp65 did not improve protective efficacy. 4. Discussion The mechanisms leading to protective immunity against Buruli ulcer are largely unknown. In contrast to individuals vaccinated with BCG, Buruli ulcer patients are characterized by weak and Th2-oriented systemic cellular responses to mycobacterial antigens [13e15]. When restimulated in vitro with M. ulcerans antigens, PBMCs from Buruli ulcer patients, particularly those with the ulcerative form of the disease, show a defect in the IFN-g response and a remarkably high production of the anti-inflammatory cytokine interleukin

*

108

P=0.05

cfus in tail lesions

107

106

105

104

A

B

C

D

E

Fig. 5. Vaccination with pCDNA3-Hsp65 confers weak protection against infection with M. ulcerans. The bacterial load in skin lesions 6 weeks post infection with 104 AFB of M. ulcerans is shown in mice injected with control plasmid pCDNA3 (A), pCDNA3-Ag85 (B), pCDNA3-hsp65 (C), pCDNA3Mix (D) or vaccinated with BCG (E). Data are means  SD of cfu per mouse tail measured on groups of seven mice. *P < 0.05, ANOVA.

2079

(IL)-10 [16]. These hallmarks of the immune response to infection with M. ulcerans change as the disease progresses through the healing stage, with a proportion of patients showing a positive skin test reaction to Burulin (a crude preparation of M. ulcerans lysate) of 93.8%, compared to 14.3% at the pre-ulcerative stage [17]. Activation of inflammatory responses coincides with clearance of the bacteria, which strongly suggests that stimulation of Th1-oriented cellular immune responses at early stages of infection may accelerate the resolution of the disease. This hypothesis is supported by the observation that BCG, a potent inducer of anti-mycobacterial Th1-oriented cellular responses, confers protection against Buruli ulcer. Considering the phylogenetic relationship between M. ulcerans and the members of the M. tuberculosis complex such as BCG, the generation of cross-reactive immune responses upon vaccination is not surprising. BCG protects mice experimentally infected with M. ulcerans, to variable extents. Tanghe et al. reported a protective efficacy of DLog10 ¼ 1.35 in C57BL/6 mice vaccinated with intravenous BCG and subcutaneous infection with M. ulcerans in the footpad, compared to DNA vaccinated animals [9]. This is a much lower protective efficacy than the one we observed in BALB/c mice vaccinated by subcutaneous BCG and subcutaneous infection with M. ulcerans in the tail (DLog10 ¼ 3.13). Many factors may account for this discrepancy, including the virulence of the M. ulcerans strain, the vaccination route, and the mouse strain as we found that C57BL/6 mice were more susceptible to subcutaneous infection with M. ulcerans than BALB/c mice (data not shown). BCG is a live vaccine expressing a variety of immunogenic components and carrying numerous TLR-stimulating motifs, which contribute to its striking vaccinal properties. The observation that BCG is protective against infection with M. ulcerans in mouse studies and human vaccination trials suggests that mycobacterial antigens expressed by BCG may be of interest as subunit vaccines. Using the recently sequenced genome of M. ulcerans, we can exclude a priori BCG antigens that may be immunogenic and protective against experimental tuberculosis, but are absent from M. ulcerans (Table 1). In particular, no equivalent of MPT70, MPT83 and HspX could be identified in the sequenced M. ulcerans strain. Among the BCG antigens sharing extensive sequence homology with their equivalents in M. ulcerans are Hsp65 and the Ag85 complex members A and B. DNA vaccination with antigen 85A from BCG induced significant Th1-type immune responses to this antigen, and subsequent protection against Buruli ulcer in experimentally infected C57BL/6 mice [9]. No significant protection was obtained in our model following DNA vaccination with M. tuberculosis Ag85B, suggesting that T cell epitopes may differ between Ag85B of M. tuberculosis and M. ulcerans in BALB/c mice [18]. In the present study, we identified the heat shock protein Hsp65 as a major antigen in M. ulcerans infection. Hsp65 has been previously described as an immunodominant antigen of the other pathogenic mycobacteria M. tuberculosis and M. leprae, inducing specific T-cell responses in human and experimentally infected mice [19]. DNA vaccination with

E. Coutanceau et al. / Microbes and Infection 8 (2006) 2075e2081

2080

Table 1 Conservation of selected M. tuberculosis antigens in M. ulcerans, as determined by sequence alignment using the BLAST analysis software Antigen (gene)

Accession numbera

Antigen size (kDa)

Amino acid (aa) number

Specificity

Protectionb

Conservation in BCGc

Conservation in M. ulceransc

Ag85A (fbpA) Ag85B (fbpB) ESAT-6 (esxA)

Rv3804c Rv1886c Rv3875

32 30 6

338 325 95

þþ þþ þ

100% 100% NS

83.4% 87.9% NS

CFP-10 (esxB) PPE68 (PPE68) TB10.4 (esxH ) MPT64 (mpt64)

Rv3874 Rv3873 Rv0288 Rv1980c

10 40 10 26

100 368 96 228

ND ND ND þ

NS NS 100% NS

NS NS 84.4% 66.5%

MPT70 (mpt70) MPT83 (mpt83) PPE18 (mtb39a) Hsp65 ( groEL2)

Rv2875 Rv2873 Rv1196 Rv0440

18 26 39 65

193 220 391 540

þ ND þ þþ

100% 100% 94% 100%

NS NS 56.5% 93.9%

Hsp70 (dnaK )

Rv0350

70

625

þ

100%

93.6%

a-Crystallin (hspX ) PstS1 ( pstS1) PstS3 ( pstS3) 19-kDa Ag (lpqH ) KatG (katG)

Rv2031c

16

144

All mycobacteria All mycobacteria M. tuberculosis, M. bovis, M. smegmatis, M. leprae, M. kansasii, M. marinum Same as ESAT-6 Same as ESAT-6 M. tuberculosis complex, M. leprae M. tuberculosis complex, (absent from some BCG strains) M. tuberculosis complex M. tuberculosis complex M. tuberculosis complex, M. leprae All mycobacteria, homologues in eukaryotic cells All mycobacteria, homologues in eukaryotic cells M. tuberculosis complex

ND

100%

NS

Rv0934 Rv0928 Rv3763 Rv1908c

38 38 19 80

374 370 159 740

þ þ þ ND

100% 100% 100% 100%

NS 67.8% 85% 71%

Homologues in bacteria Homologues in bacteria Slow-growing mycobacteria M. tuberculosis, M. bovis, M. smegmatis, M. avium paratuberculosis, M. marinum

ND, not determined; NS, not significant (amino acid sequence identity level lower than 40%). a As referenced in http://genolist.pasteur.fr/Tuberculist/ web server. b Protection conferred against experimental TB in mouse or guinea pig models, when delivered as protein, DNA vaccine or vaccinia virus (þ, evidence of protection; þþ, protection comparable to that of BCG). c As determined by the percentage of identity.

Hsp65 from M. leprae results in the generation of potent Th1type immune responses in BALB/c mice [11], of greater importance than those elicited by vaccination with BCG (data not shown). Contrary to the members of the Ag85 complex, Hsp65 is not prominent in M. tuberculosis culture filtrate, but rather released in increased amounts by the tubercle bacilli in response to the stress of intracellular existence. As heat shock proteins belong to a group of proteins encoded ubiquitously in all prokaryotic and eukaryotic cells, identification of antibodies or T cells against Hsp65 cannot be indicative of infection with M. ulcerans. We found that Hsp65 is expressed in considerable amounts by M. ulcerans bacilli in vitro and in vivo, and is immunogenic for both B and T cells in mice. The level of protection conferred by vaccination with a plasmid DNA encoding Hsp65 was therefore disappointing. Despite the extensive degree of sequence identity between M. leprae and M. ulcerans Hsp65 antigens (96%), the possibility remains that protective epitopes were not conserved in the M. leprae Hsp65-based vaccine and that vaccination with M. ulcerans Hsp65 may be more effective. Our analysis of the conservation of protective M. tuberculosis antigens in M. ulcerans shows that TB10.4, the 19 kDa Ag and Hsp70 are also antigens present in BCG, that are extensively conserved in M. ulcerans (Table 1). Western blot reactivity of Buruli ulcer patients to M. ulcerans culture filtrates identified three major antigens of 70, 36/38 and 5 kDa,

respectively [17]. Importantly, in contrast to the studies performed by Tanghe et al. in mice and Dobos et al. in humans, which used mycobacterial culture filtrates, serum reactivity was examined against whole bacterial lysates in our study. Discrepancies in B-cell responses between these studies and ours may therefore reflect quantitative and qualitative differences in the sources of antigens [9,17]. The 38 kDa M. ulcerans antigen identified in this work may be the phosphate binding protein PstS3, for which the degree of sequence similarity with BCG is 68%. Therefore, PstS3 and the 70 kDa Hsp70 may be considered as candidate antigens in future vaccination studies.

Acknowledgements We gratefully acknowledge the Association Franc¸aise Raoul Follereau for financial support, and the Colorado State University, NIH (NIAID contract NO1 AI-75320 ‘‘Tuberculosis Research materials and Vaccine testing) for providing the Ag85 complex, the Hsp65 protein and the anti-Hsp65 monoclonal antibody used in this study. We thank Professor W.J. Britton (Centenary Institute, Sydney, Australia) for the gift of the pCDNA3-Ag85B vaccine, and Dr R.E. Tascon (National Institute for Medical Research, London, UK) for the pCDNA3Hsp65 vector.

E. Coutanceau et al. / Microbes and Infection 8 (2006) 2075e2081

References [1] P.D.R. Johnson, T. Stinear, P.L.C. Small, G. Plushke, R.W. Merritt, F. Portaels, K. Huygen, J.A. Hayman, K. Asiedu, Buruli ulcer: New insights, new hope for disease control, Plos. Med. 2 (4) (2005) 282e286. [2] E. Coutanceau, L. Marsollier, R. Brosch, E. Perret, P. Goossens, M. Tanguy, S.T. Cole, P.L. Small, C. Demangel, Modulation of the host immune response by a transient intracellular stage of Mycobacterium ulcerans: the contribution of endogenous mycolactone toxin, Cell. Microbiol. 7 (8) (2005) 1187e1196. [3] M.S. Oliveira, A.G. Fraga, E. Torrado, A.G. Castro, J.P. Pereira, A.L. Filho, F. Milanezi, F.C. Schmitt, W.M. Meyers, F. Portaels, M.T. Silva, J. Pedrosa, Infection with Mycobacterium ulcerans induces persistent inflammatory responses in mice, Infect. Immun. 73 (10) (2005) 6299e6310. [4] K.M. George, D. Chatterjee, G. Gunawardana, D. Welty, J. Hayman, R. Lee, P.L. Small, Mycolactone: A polyketide toxin from Mycobacterium ulcerans required for virulence, Science 283 (5403) (1999) 854e857. [5] S. Etuaful, B. Carbonnelle, J. Grosset, S. Lucas, C. Horsfield, R. Phillips, M. Evans, D. Ofori-Adjei, E. Klustse, J. Owusu-Boateng, G.K. Amedofu, P. Awuah, E. Ampadu, G. Amofah, K. Asiedu, M. Wansbrough-Jones, Efficacy of the combination rifampin-streptomycin in preventing growth of Mycobacterium ulcerans in early lesions of buruli ulcer in humans, Antimicrob. Agents Chemother. 49 (8) (2005) 3182e3186. [6] F. Portaels, J. Aguiar, M. Debacker, C. Steunou, C. Zinsou, A. Guedenon, W.M. Meyers, Prophylactic effect of Mycobacterium bovis BCG vaccination against osteomyelitis in children with Mycobacterium ulcerans disease (Buruli Ulcer), Clin. Diagn. Lab. Immun. 9 (6) (2002) 1389e1391. [7] P.G. Smith, W.D.L. Revill, E. Lukwago, Y.P. Rykushin, Protective effect of BCG against Mycobacterium ulcerans diseasedcontrolled trial in an endemic area of Uganda, Trans. R. Soc. Trop. Med. Hyg. 70 (5-6) (1976) 449e457. [8] F. Fenner, Homologous and heterologous immunity in infections of mice with Mycobacterium ulcerans and Mycobacterium balnei, Am. Rev. Tuberc. 76 (1) (1957) 76e89. [9] A. Tanghe, J. Content, J.P. Van Vooren, F. Portaels, K. Huygen, Protective efficacy of a DNA vaccine encoding antigen 85A from Mycobacterium bovis BCG against Buruli ulcer, Infect. Immun. 69 (9) (2001) 5403e5411.

2081

[10] D.B. Lowrie, R.E. Tascon, V.L.D. Bonato, V.M.F. Lima, L.H. Faccioli, E. Stavropoulos, M.J. Colston, R.G. Hewinson, K. Moelling, C.L. Silva, Therapy of tuberculosis in mice by DNA vaccination, Nature 400 (6741) (1999) 269e271. [11] R.E. Tascon, M.J. Colston, S. Ragno, E. Stavropoulos, D. Gregory, D.B. Lowrie, Vaccination against tuberculosis by DNA injection, Nat. Med. 2 (8) (1996) 888e892. [12] A.T. Kamath, C.G. Feng, M. Macdonald, H. Briscoe, W.J. Britton, Differential protective efficacy of DNA vaccines expressing secreted proteins of Mycobacterium tuberculosis, Infect. Immun. 67 (4) (1999) 1702e1707. [13] T.M. Gooding, P.D.R. Johnson, D.E. Campbell, J.A. Hayman, E.L. Hartland, A.S. Kemp, R.M. Robins-Browne, Immune response to infection with Mycobacterium ulcerans, Infect. Immun. 69 (3) (2001) 1704e1707. [14] T.M. Gooding, P.D.R. Johnson, M. Smith, A.S. Kemp, R.M. RobinsBrowne, Cytokine profiles of patients infected with Mycobacterium ulcerans and unaffected household contacts, Infect. Immun. 70 (10) (2002) 5562e5567. [15] T.M. Gooding, A.S. Kemp, R.M. Robins-Browne, M. Smith, P.D.R. Johnson, Acquired T-helper 1 lymphocyte anergy following infection with Mycobacterium ulcerans, Clin. Infect. Dis. 36 (8) (2003) 1076e1077. [16] G. Prevot, E. Bourreau, H. Pascalis, R. Pradinaud, A. Tanghe, K. Huygen, P. Launois, Differential production of systemic and intralesional IFN-g and IL-10 in nodular and ulcerative forms of Buruli disease, Infect. Immun. 72 (2) (2004) 958e965. [17] K.M. Dobos, E.A. Spotts, B.J. Marston, C.R. Horsburgh, C.H. King, Serologic response to culture filtrate antigens of Mycobacterium ulcerans during Buruli ulcer disease, Emerg. Infect. Dis. 6 (2) (2000) 158e164. [18] S. D’Souza, V. Rosseels, A. Romano, A. Tanghe, O. Denis, P. Jurion, N. Castiglione, A. Vanonckelen, K. Palfliet, K. Huygen, Mapping of murine Th1 helper T-cell epitopes of mycolyl transferases Ag85A, Ag85B, and Ag85C from Mycobacterium tuberculosis, Infect. Immun. 71 (1) (2003) 483e493. [19] C.L. Silva, The potential use of heat-shock proteins to vaccinate against mycobacterial infections, Microbes Infect. 1 (6) (1999) 429e435.