A Live Experimental Vaccine against Burkholderia pseudomallei ...

MAJOR ARTICLE

A Live Experimental Vaccine against Burkholderia pseudomallei Elicits CD4+ T Cell–Mediated Immunity, Priming T Cells Specifi for 2 Type III Secretion System Proteins Ashraful Haque,1,a Karen Chu,1 Anna Easton,1 Mark P. Stevens,2 Edouard E. Galyov,2 Tim Atkins,3 Rick Titball,3 and Gregory J. Bancroft1 1 London School of Hygiene and Tropical Medicine, London, 2Institute for Animal Health, Compton, Berkshire, 3Defence Science and Technology Laboratories, Porton Down, United Kingdom

Burkholderia pseudomallei is the etiological agent of melioidosis, a serious human disease for which no vaccine is available. Immunization of susceptible BALB/c mice with the live attenuated mutant B. pseudomallei ilvI (referred to as “2D2”) generated significant although incomplete, immunity. Splenic B. pseudomallei–specifi T cells, detected in immunized mice, proliferated and produced interferon-g in vitro in response to dead bacteria. Assessment of T cell antigen specificit indicated that subpopulations of B. pseudomallei–reactive T cells were responsive to BopE, a type III secretion system (TTSS) effector protein, and to a lesser extent to BipD, a TTSS translocator protein. Increased survival of severe combined immunodeficien mice adoptively transferred with T cells from immunized mice, compared with that of naive T cell recipients, demonstrated that immunization with 2D2 generated T cell–mediated immunity. CD4+ and CD8+ cell depletion studies demonstrated that CD4+ cells, but not CD8+ cells, mediated this protection in vivo. Thus, CD4+ T cells can mediate vaccine-induced immunity to experimental melioidosis. Burkholderia pseudomallei, the etiological agent of melioidosis, is a gram-negative, intracellular bacterial pathogen that is endemic in Southeast Asia and northern Australia [1]. Clinical manifestations vary from acute, lethal sepsis to chronic, localized abscess formation to latent infection, which can reactivate decades later [2].

Received 12 May 2006; accepted 20 June 2006; electronically published 25 September 2006. Presented in part: 4th World Melioidosis Conference, 16–18 September 2004, Singapore. Potential conflicts of interest: none reported. Financial support: Defence Science and Technology Laboratory, United Kingdom (grants RD032-0469 and RD013-0931417 to G.J.B.); National Institutes of Health (grant AI-61363, subaward 2004-1489, to G.J.B.); Biotechnology and Biological Sciences Research Council (grant C20021 to M.P.S. and E.E.G.). a Present affiliation: Queensland Institute for Medical Research, Herston, Brisbane, Australia. Reprints or correspondence: Dr. Ashraful Haque, Queensland Institute for Medical Research, 300 Herston Rd., Herston, Brisbane, QLD 4006, Australia (ashraful [email protected]). The Journal of Infectious Diseases 2006; 194:1241–8 2006 by the Infectious Diseases Society of America. All rights reserved. 0022-1899/2006/19409-0009$15.00

Mortality rates in acute cases can exceed 40%, with 10%–15% of survivors relapsing despite prolonged antibiotic treatment [2]. B. pseudomallei is classifie as a class B potential agent for biological warfare and terrorism. There is no vaccine against B. pseudomallei, and knowledge of mechanisms of resistance to B. pseudomallei is limited. A better understanding of immunity to B. pseudomallei is needed for the generation of vaccine or immunotherapy strategies against this infection. Knowledge of mechanisms of resistance to B. pseudomallei infection has largely derived from murine models of melioidosis [3–6]. Our previous work highlighted the importance of interferon (IFN)–g, interleukin (IL)–12, and IL-18 as early mediators of protection [6, 7]. In addition, we demonstrated a role for T cells in resistance to primary infection [7]. Studies of melioidosis patients demonstrated elevated levels of IFN-g, IL-12, and IL-18 in serum samples [8] and have also detected B. pseudomallei–specifi T cells in the blood [9]. Thus, there is growing evidence that T cell–mediated immune responses occur during B. pseudomallei infection.

CD4+ T Cell–Dependent Immunity to B. pseudomallei

• JID 2006:194 (1 November) • 1241

Most examples of protection against B. pseudomallei have relied on antibodies against flagellin capsule, or lipopolysaccharide [10–13]. However, given the ability of B. pseudomallei to persist inside host cells [14], effective vaccine-mediated immunity to B. pseudomallei will most likely require the combined actions of antibody and T cell–mediated immunity. We, and others, have reported significan protection by immunization with live B. pseudomallei [15, 16], but in neither case were the immune mechanisms of protection studied. Two recent studies demonstrating antibody and cellular immune responses in mice vaccinated with antigen-pulsed dendritic cells or DNA-encoding B. pseudomallei flagelli did not assess the direct contribution to protection of B. pseudomallei–specifi T cells [17, 18]. Thus, to date, the generation of specifi T cell responses against B. pseudomallei proteins has not been proven as a vaccine strategy. We generated immunity to B. pseudomallei by immunizing genetically susceptible BALB/c mice with a live attenuated mutant, 2D2 [15]. 2D2 was highly attenuated in mice, was effectively cleared from various organs, and protected mice in the short term from subsequent virulent challenge [15]. Here, we demonstrate that immunization with 2D2 (hereafter, “2D2 immunization”) primes B. pseudomallei–specifi CD4+ and CD8+ T cells and that elicited protection is mediated by CD4+ T cells, but not by CD8+ T cells. Using this experimental vaccination strategy, we identify B. pseudomallei type III secretion system (TTSS) proteins as targets of T cell–mediated immunity. MATERIALS AND METHODS Bacterial strains and culture conditions. B. pseudomallei 576, originally isolated from a patient with human melioidosis in Thailand, was obtained from T. Pitt, Health Protection Agency, London. B. pseudomallei 576 ilvI (referred to as “2D2”) was generated by transposon-based mutagenesis [15]. Bacteria were cultured in tryptone soy (TS) broth or TS agar. Bacteria were stored as described elsewhere [7]. Killed bacteria, prepared by gamma irradiation, were confi med to be nonviable by culture and stored as described elsewhere [7]. All procedures involving live bacteria were conducted under Advisory Committee on Dangerous Pathogens containment level 3 conditions. Expression and purificatio of BipD and BopE. Constructs for inducible expression of BipD–glutathione-S-transferase (GST) and BopE-GST fusion proteins in Escherichia coli have been described elsewhere [19, 20]. Proteins were purifie using glutathione sepharose 4B resin (Amersham Pharmacia Biotech), in accordance with the Amersham Pharmacia GST gene fusion protocol. BipD and BopE proteins were cleaved from the fusion partner, GST was removed as described above, protein concentration in the eluate was estimated using bicinchonic acid assays (Pierce Biotechnology), and purity was assessed by sodium-dodecyl-sulphate polyacrylamide gel electrophoresis. 1242 • JID 2006:194 (1 November) • Haque et al.

Infection of mice and determination of splenic bacterial burden. Female 8–10 week-old BALB/c mice were housed under specifi pathogen-free conditions with free access to food and water. Mouse experiments were performed in accordance with the Animals (Scientifi Procedures) Act 1986 and were approved by the local ethical review committee. Bacteria were thawed from frozen stocks, diluted in pyrogen-free saline, and administered intraperitoneally (ip; 0.2 mL). 2D2 immunization was performed by ip injection of 106 cfu of B. pseudomallei 2D2, followed 5 weeks later by challenge or removal of spleens. Boosting with 2D2 was performed by administering 2D2 as before, 2 weeks after the firs dose. Boosted mice were then left a further 3 weeks before challenge or removal of spleens. No viable bacteria could be recovered from the spleens of mice after the full period of immunization, as determined by culturing tissue homogenates on agar plates (data not shown). Challenge with virulent B. pseudomallei was performed by ip injection of 106 cfu of B. pseudomallei 576, unless otherwise stated. Verificatio of inoculation doses was performed as described elsewhere [7]. Splenic bacterial burdens were determined as described elsewhere [7]. Preparation and in vitro stimulation of murine spleen cells. Spleen cells were prepared as described elsewhere [7]. Cells in round-bottom 96-well plates (2.5 ⫻ 10 6 cells/mL) were stimulated in triplicate with dead bacteria (1 bacterium/10 spleen cells) or recombinant proteins at 37C in 5% CO2 for 24 h (intracellular cytokine staining [ICCS]), 48 h (ELISA), or 5 days (carboxyflu orescein diacetatesuccinimidyl ester [CFSE] proliferation assays). Responses to restimulations using dead bacteria or recombinant proteins were confi med to be dose dependent (data not shown). Flow cytometry for cell surface markers and intracellular IFN-g staining. Cell surface and intracellular cytokine staining was performed as described elsewhere [7]. Cells were analyzed using a FACScalibur (BD Biosciences) with FlowJo software (Tree Star) under containment level 3. Viable lymphocytes were gated using forward and side scatter parameters. ELISA analysis of IFN-g in cell culture supernatants. Anti–IFN-g antibodies, AN18 and biotinylated-R46A2, purchased from Mabtech, were used in ELISAs, in accordance with the manufacturer’s guidelines. ELISAs were developed with streptavidin-conjugated peroxidase (1 mg/mL; Sigma) and SureBlue substrate (KPL Laboratories). Each sample from triplicated restimulation groups was tested in duplicate wells, and the mean value of duplicate wells was taken for subsequent analysis. CFSE-based lymphocyte proliferation assays. Splenocytes, washed twice in Ca2+/Mg2+-free PBS, were incubated in 5 mmol/ L CFSE (Sigma) for 10 min at room temperature. CFSE-labeled cells were washed twice in PBS before being plated out and stimulated in triplicate, as described above, for 5 days. Cells were then processed for flo cytometry, as described above. Viable T cells were gated using forward and side scatter pa-

rameters and CD4+ or CD8+ T cell expression. A minimum of 25,000 CD4+ or CD8+ T cell events were analyzed per sample. Isolation and adoptive transfer of T cells into severe combined immunodeficienc (SCID) mice. T cells isolated from splenocytes by use of Dynal Mouse T cell negative isolation kits (Invitrogen), in accordance with the manufacturer’s instructions, were confi med to be 195% pure by flow-cytometri analysis with fluo escently conjugated anti-CD4 monoclonal antibody (MAb) RM4-5 and anti-CD8 MAb 53-6.7 (BD Biosciences). A concentration of 2 ⫻ 107 cells in 200 mL of saline was injected into female CB-17.SCID mice via a tail vein. Antibodies and in vivo cell depletion. Anti-CD4+ (YTS191) and anti-CD8+ (YTS169) MAbs and isotype control Mac-5 antibodies were donated by R. Lukaszewski, Defence Science and Technology Laboratories, Porton Down, United Kingdom. Dosage regimes and determination of efficienc of cell depletions were performed as described elsewhere [7]. Efficienc of depletions at the time of infection and at 3–4 days after infection were 199% for CD4+ T cells with YTS191 and 197% for CD8+ T cells with YTS169. Statistical analysis. Survival curves were compared using log rank Kaplan-Meier tests. Student’s t test was employed for all other statistical tests. P ! .05 was considered to be statistically significant RESULTS Significant although incomplete, protection against challenge with virulent B. pseudomallei 576 via 2D2 immunization. Using an acute model of infection, in which genetically susceptible BALB/c mice are challenged with ∼10,000 times the median lethal dose of B. pseudomallei 576, we have shown previously that 2D2 immunization robustly protects the majority of mice against challenge over a 35-day time frame [15]. To investigate whether 2D2 immunization provides longer-term protection against lethal melioidosis, 2D2-immunized and unimmunized BALB/c mice were challenged with wild-type B. pseudomallei 576 and monitored for survival. Unimmunized mice displayed signs of illness within the firs few days after challenge and typically succumbed to infection within the firs week after challenge, with a median survival time (MST) of 3 days. 2D2-immunized mice displayed no clinical signs of illness within the firs week after challenge and were significantl protected, compared with unimmunized mice (MST, 52 days; P ! .0001) (figu e 1). 2D2immunized mice began to succumb to infection from day 30 after challenge (figu e 1). Postmortem analysis of 2D2-immunized mice that became moribund after challenge revealed the presence of splenic abscesses, often 12 mm in diameter and found to contain 1107 bacteria (data not shown). Kinetics of splenic bacterial loads in 2D2-immunized mice after challenge with B. pseudomallei 576. We determined the kinetics of bacterial accumulation after challenge in the spleens

of unimmunized and 2D2-immunized mice. In addition, because 2D2 immunization provided incomplete protection against infection of susceptible mice, we tested whether a second, boosting dose of 2D2 might enhance the protection provided by a single immunizing dose. Naive, 2D2-immunized and 2D2-boosted mice were infected with B. pseudomallei 576, and splenic bacterial loads were monitored at various time points (figu e 2). Immunized groups exhibited median splenic bacterial burdens of 30 and 1100 cfu, compared with 2.2 ⫻ 104 cfu for naive mice at day 1 after challenge. 2D2-boosted mice had lower splenic bacterial burdens than did 2D2-immunized mice at day 1 after challenge (P ! .02), but no statistically significan differences between these 2 groups were observed at any other time points. By day 2 after challenge, median splenic bacterial burdens in naive mice had increased to 2.5 ⫻ 10 5 cfu, whereas those in the 2 immunized groups were 200 and 1400 cfu. Naive mice began to succumb to infection from day 3 after challenge. Analysis of splenic bacterial burdens in both immunized groups from day 6 after challenge revealed 2 distinct patterns. The majority of mice had low bacterial loads (100–1000 cfu), with no macroscopically observable abscesses. In contrast, at day 13 after challenge 3 of 10 2D2-immunized mice and 4 of 10 2D2-boosted mice possessed numerous abscesses containing 1107 bacteria. Taken together, these data indicate that 2D2 immunization prevents rapidly disseminated infection within the firs few days after challenge. In this genetically susceptible model of infection, single and double immunizing doses of 2D2 can delay but not prevent the eventual development of abscesses, which lead to death in the later stages of infection. Priming by 2D2 immunization of antigen-specific IFN-g– producing T cells. To investigate a role for antigen-specifi T

Figure 1. Survival of Burkholderia pseudomallei 2D2–immunized BALB/ c mice after challenge with virulent B. pseudomallei 576. BALB/c mice (5/ group) were immunized intraperitoneally (ip) with 106 cfu of 2D2 each. Unimmunized mice were given saline. Five weeks after immunization, mice were challenged ip with 106 cfu of B. pseudomallei 576 each and monitored daily for survival. Data are representative of 5 independent experiments.

CD4+ T Cell–Dependent Immunity to B. pseudomallei • JID 2006:194 (1 November) • 1243

Figure 2. Kinetics of bacterial control in the spleens of 2D2-immunized and 2D2-boosted mice. 2D2-immunized, 2D2-boosted BALB/c mice and unimmunized mice were each challenged intraperitoneally with 106 cfu of Burkholderia pseudomallei 576. At 1, 2, 6, or 13 days after challenge, spleens from groups of 5–10 mice were removed, and their bacterial loads were determined. Five mice per group were used for days 1, 2, and 6, and 10 mice per group were used for day 13. Horizontal lines indicate the median value per group. The limit of detection for this assay was 100 cfu/spleen (horizontal dotted line). Data are representative of 3 independent experiments. *P ! .05. NS, not significant.

cells in the protection elicited by 2D2 immunization, we attempted to detect such cells in the spleens of 2D2-immunized and 2D2-boosted mice. Splenocytes from 2D2-immunized and 2D2-boosted mice were stimulated in vitro with dead B. pseudomallei 576. CD4+ T cells and CD8+ T cells were then assessed for IFN-g production by ICCS. T cells from naive mice produced low levels of IFN-g when cultured with dead B. pseudomallei (figu e 3A). However, a significantl greater frequency of CD4+ T cells and CD8+ T cells from 2D2-immunized and 2D2-boosted mice produced IFN-g in response to dead B. pseudomallei 576, compared with naive splenocytes (figu e 3A and 3B). The in vitro IFN-g response to dead bacteria was completely abrogated in the presence of cyclosporin A (1 mg/mL), a fungal metabolite that blocks antigen-specifi activation of T cells (data not shown). Frequencies of IFN-g–producing T cells were ∼50% greater in 2D2-boosted mice than in single-dose immunized mice (figu e 3A). Taken together, these data indicate that 2D2 immunization primes CD4+ T cells and CD8+ T cells specifi for B. pseudomallei antigens and that boosting with 2D2 increases their frequency in the spleen. Priming by 2D2 immunization of IFN-g–producing T cells specifi for 2 TTSS proteins, BipD and BopE. We previously demonstrated that proteins from a B. pseudomallei TTSS can act as B cell immunogens [19]. To determine whether 2D2 immunization primes T cells against TTSS proteins, splenocytes from unimmunized, 2D2-immunized, and 2D2-boosted mice were stimulated in vitro with the TTSS proteins BopE or BipD 1244 • JID 2006:194 (1 November) • Haque et al.

and assayed for IFN-g production by ICCS (figu e 3A and 3B). IFN-g production was detected in !0.1% of either CD4+ or CD8+ T cells from unimmunized mice (figu e 3A and 3B). In spleens from 2D2-boosted mice, 1% of CD4+ T cells and 0.5% of CD8+ T cells produced IFN-g in response to BopE (figu e 3A and 3B). The frequencies of IFN-g–producing BopE-reactive T cells were ∼50% greater in the spleens of 2D2-boosted mice than in mice given a single dose of 2D2. Low frequencies (0.1%–0.2%) of T cells from 2D2-boosted mice produced IFN-g in response to BipD, but these frequencies were not statistically significantl different, compared with those for unimmunized groups. This observation suggested either that BipD-reactive T cells were absent from the spleens of immunized mice or that their frequency was too low to be detected by ICCS. To resolve this issue, IFN-g was detected by ELISA in the supernatants of spleen cells stimulated with either BipD or BopE (figu e 3C). A dose response to BipD, observed in splenocytes from 2D2-boosted mice, was completely absent in supernatants from spleen cells from unimmunized mice, demonstrating a BipD-specifi cellular immune response in 2D2boosted mice (figu e 3C). A dose response was also seen for BopE, from 1 to 50 mg/mL, with the mean SD concentrations of IFN-g in supernatants ranging from 8 0.5 to 17 1.0 ng/ mL (figu e 3C; data not shown). Responses to BipD and BopE were completely abrogated in the presence of cyclosporin A, thus confi ming them to be antigen specifi (data not shown). To determine whether BopE- and BipD-specifi T cells from

Figure 3. In vitro detection of antigen-specific interferon (IFN)–g production by splenic T cells from 2D2-immunized mice. For panels A and B, spleen cells from 2D2-immunized, 2D2-boosted, and unimmunized mice were stimulated overnight with dead Burkholderia pseudomallei 576 (1 bacterium/10 spleen cells), BopE protein (10 ug/mL), or BipD protein (10 ug/mL). Viable lymphocytes were gated and analyzed for CD4, CD8 and IFN-g expression. A, Mean + SD values from triplicate samples. B, Representative plots, gated on viable lymphocytes, of the IFN-g response by in vitro–stimulated CD4+ and CD8+ splenic lymphocytes from 2D2-boosted mice. Data presented are representative of 2 independent experiments. For panel C, spleen cells from 2D2boosted or naive mice were stimulated for 2 days with BopE (1 mg/mL) or BipD (1, 10, or 50 mg/mL), and supernatants were assayed for IFN-g expression by ELISA. Data are representative of 3 independent experiments. ***P ! .0001. NS, not significant.

immunized mice could proliferate in response to cognate antigen, splenocytes from unimmunized and 2D2-boosted mice were labeled with the fluo escent vital dye CFSE and stimulated in vitro. Unstimulated T cells from either group did not proliferate under these conditions (figu e 4). Cells stimulated with E. coli lipopolysaccaride (LPS; up to 1 mg/mL) did not proliferate, demonstrating that small quantities of contaminating LPS in recombinant protein preparations would not induce nonspecifi lymphocyte proliferation (data not shown). CD4+ and CD8+ splenic lymphocytes from 2D2-boosted mice, but not unimmunized mice, proliferated in response to dead B. pseudomallei or BopE (figu e 4). BipD induced low but statistically significan CD4+, but not CD8+, lymphocyte proliferation in the 2D2-boosted group, but not in the unimmunized group (figu e 4). Taken together, these data indicate that 2D2 immunization primes B. pseudomallei–reactive CD4+ T cells, some of which display specificit for BopE and BipD, and also B.

pseudomallei–reactive CD8+ T cells, some of which are specifi for BopE but not BipD. CD4+ T cell–dependent, but CD8+ T cell–independent, protection by 2D2 immunization. To determine whether antigen-specifi splenic T cells from immunized mice were protective against virulent challenge, splenic T cells from naive and 2D2-boosted mice were isolated and adoptively transferred into SCID mice, which possess no T or B cells and are extremely susceptible to B. pseudomallei infection. Mice were challenged intraperitoneally with 5 ⫻ 10 4 cfu of B. pseudomallei 576 and monitored for survival. Recipients of naive T cells were more protected than control mice (MST, 5 vs. 4 days; P ! .05), demonstrating some background protection by naive T cells in these acutely susceptible immunocompromised mice (figu e 5A). However, recipients of T cells from 2D2-immunized mice were significantl more protected than naive T cell recipients (MST, 11 vs. 5 days; P ! .014) (figu e 5A). These data demonstrate


Figure 4. In vitro proliferation of antigen-specific T cells from 2D2-boosted mice. Spleen cells from 2D2-boosted mice (black bars) or unimmunized mice (white bars) were labeled with carboxyfluorescein diacetatesuccinimidyl ester (CFSE) and stimulated in vitro with dead Burkholderia pseudomallei 576 (1 bacterium/10 spleen cells), BopE protein (10 mg/mL), or BipD protein (10 mg/mL). Viable CD4+ or CD8+ lymphocytes were then analyzed by flow cytometry for proliferation, as indicated by the loss of the CFSE stain. Plots indicate CFSE profiles for CD4+- and CD8+-viable lymphocytes and are representative of triplicate samples. The vertical dashed line in each plot represents the threshold expression of CFSE below which a cell was considered to be CFSE negative. Graphs indicate the mean percentage (+1 SD for triplicate samples) of total CD4+- or CD8+-viable CFSE-negative lymphocytes. Data are representative of 2 independent experiments. *P ! .05; **P ! .001; ***P ! .0001. NS, not significant.

that 2D2 immunization primes a population of splenic T cells that are protective against experimental melioidosis and can act independently of B cells. To determine the relative roles played by CD4+ and CD8+ T cells in the protection elicited by 2D2 immunization, 2D2immunized mice were antibody depleted of their CD4+ cells or CD8+ cells before and after challenge with B. pseudomallei 576. Immunized, CD4+ cell–depleted mice demonstrated substantially increased susceptibility to infection, compared with immunized mice given isotype control antibodies (MST, 3 vs. 28 days; P ! .0001) (figu e 5B). CD4+ T cell–depleted mice succumbed at the same time as unimmunized control mice (MST, 3 vs. 3 days; P ! .0291) (figu e 2), indicating that 2D2-mediated protection was abolished by CD4+ T cell depletion. In contrast, CD8+ T cell depletion had no effect on 2D2-mediated protection (figu e 5C). Taken together, these data demonstrate that protection elicited by 2D2 immunization is mediated by CD4+ T cells, but not by CD8+ T cells. DISCUSSION There is currently no vaccine against B. pseudomallei, and little is known of how acquired immunity to this pathogen may be 1246 • JID 2006:194 (1 November) • Haque et al.

generated. We demonstrate here that immunization with the live attenuated mutant of B. pseudomallei, 2D2, generates CD4+ T cell–mediated immunity in mice. These data are consistent with those from murine models of other intracellular bacterial infections, where CD4+ T cells are important for immunity [21–23]. The work presented here supports our studies of resistance to primary infection with B. pseudomallei, in which we reported a greater dependency on CD4+ T cells, rather than on CD8+ T cells, for protection against infection [7]. In the current study, CD4+ T cell–depleted immunized mice were as susceptible to challenge with virulent bacteria as naive mice, indicating that B. pseudomallei–specifi CD4+ T cells mediated the majority of protection elicited by 2D2 immunization. B. pseudomallei–specifi CD4+ T cells were detected in the spleens of immunized mice, and their frequency was increased by administering a boosting dose of 2D2. The mechanism by which B. pseudomallei–specifi CD4+ T cells mediate protection is unknown, although our adoptive transfer experiments indicate that B. pseudomallei–specifi CD4+ T cells can mediate protection independently of specifi antibody or B cells. In addition, their ability to secrete IFN-g in vitro, a cytokine essential for resistance to infection [6], is consistent with a role in macro-

Figure 5. Dependency of protection elicited by 2D2 immunization on CD4+ T cells. For panel A, severe combined immunodeficiency (SCID) mice (n p 6) were adoptively transferred with 2 ⫻ 107 T cells from naive BALB/c mice or 2D2-boosted BALB/c mice or with saline alone. The following day, SCID mice were challenged intraperitoneally (ip) with 5 ⫻ 104 cfu of Burkholderia pseudomallei 576 each and monitored for survival. For panel B, BALB/c mice (n p 6–10) were immunized ip with 106 cfu of B. pseudomallei 2D2 each. Naive mice were given saline. Five weeks after immunization, mice were treated with CD4+ T cell–depleting antibodies (Abs), CD8+ T cell–depleting Abs, or isotype control Abs. Mice were then challenged ip with 106 cfu of B. pseudomallei 576 each and monitored daily for survival. Data are representative of 2 independent experiments.

phage activation. Thus, it appears that, in both primary and secondary B. pseudomallei infections, CD4+ T cells play a crucial role in protection, whereas CD8+ T cells are less important. The absence of CD8+ T cell involvement in 2D2-mediated protection was not due to a lack of CD8+ T cell priming. Antigenspecifi CD8+ T cells were readily detected in immunized mice and were shown in vitro to produce IFN-g. The reasons why B. pseudomallei–specifi CD8+ T cells appear to play no role in protection under these conditions are not known. The genome of B. pseudomallei is predicted to encode 15000 proteins; yet, to date, few T cell antigens have been define for B. pseudomallei in either humans or mice [18]. We now add to our previous observations of B. pseudomallei–specifi T cell priming during primary infection [7] to demonstrate that antigen-specifi CD4+ and CD8+ T cell responses to 2 B. pseudomallei TTSS proteins, BipD and BopE, are readily detected in 2D2-immunized mice. BipD is a homologue of the Salmonella TTSS1 translocator protein SipD, a component of the molecular complex that interacts with host cells [19, 24, 25]. BopE is a guanine nucleotide exchange factor and homologue

of the Salmonella protein SopE, which is injected by Salmonella TTSS1 into host cells [19, 20]. There have been 2 other reports of TTSS-specifi T cell responses, to a Salmonella translocator protein, SipC [26], and to a Yersinia effector protein, YopE [27]. Here, the frequency of BopE-specifi T cells was substantially greater than the frequency of BipD-specifi T cells in immunized mice. In addition, the BopE-specifi T cell response included CD4+ and CD8+ T cells, whereas the BipD-specifi T cell response was composed of only CD4+ T cells. The reasons for these differences are unknown but may become apparent when responses to other B. pseudomallei TTSS proteins are analyzed. Furthermore, it will be interesting to investigate the independent protective capacity of TTSS-reactive T clones generated in vitro and adoptively transferred into naive hosts for subsequent challenge. The generation of sterile cure in murine models of melioidosis has remained a challenging goal for researchers. Many studies, including ours, report significant although incomplete, protection against infection [12, 15–18]. Here, protection of 2D2-immunized mice was mediated by B. pseudomallei–specifi


T cells, but boosting their numbers by 50% did not enhance long-term protection or effect sterile cure. It is unclear why additional CD4+ T cells did not enhance protection. It will be important to further increase B. pseudomallei–specifi CD4+ T cell numbers in immunized mice and to determine whether currently undefine immune mechanisms of protection in other recent models of partial protection might be employed in addition to CD4+ T cell–mediated immunity, to effect sterile cure [16–18]. In conclusion, we believe this to be the firs demonstration that acquired cellular immunity to B. pseudomallei can be generated by priming a population of CD4+ T cells with specificit for B. pseudomallei proteins. Our initial assessment of the antigen specificit of these CD4+ T cells has identifie BopE, a TTSS effector protein, and to a lesser extent BipD, a TTSS translocator protein, as T cell immunogens. Further analysis of the antigen specificit of CD4+ T cells primed by 2D2 immunization provides an opportunity to identify other T cell– reactive antigens within the B. pseudomallei proteome. Such information will facilitate the development of vaccine strategies that, in light of our data, should aim to induce CD4+ T cell– mediated immunity for optimal protection against this important infection.

Acknowledgments We thank members of the London School of Hygiene and Tropical Medicine Biological Services Facility for animal husbandry and Heidi Alderton for supervision of work performed at containment level 3. We are grateful to Dr. Roman Lukaszewski for providing CD4+- and CD8+depleting antibodies and thank Michael Wood and Terry Field for purificatio of proteins.

References 1. Dance DA. Melioidosis. Curr Opin Infect Dis 2002; 15:127–32. 2. White NJ. Melioidosis. Lancet 2003; 361:1715–22. 3. Ulett GC, Ketheesan N, Clair TW, McElnea CL, Barnes JL, Hirst RG. Analogous cytokine responses to Burkholderia pseudomallei strains contrasting in virulence correlate with partial cross-protection in immunized mice. Infect Immun 2002; 70:3953–8. 4. Ulett GC, Ketheesan N, Hirst RG. Cytokine gene expression in innately susceptible BALB/c mice and relatively resistant C57BL/6 mice during infection with virulent Burkholderia pseudomallei. Infect Immun 2000; 68:2034–42. 5. Ulett GC, Ketheesan N, Hirst RG. Proinflammato y cytokine mRNA responses in experimental Burkholderia pseudomallei infection in mice. Acta Trop 2000; 74:229–34. 6. Santanirand P, Harley VS, Dance DA, Drasar BS, Bancroft GJ. Obligatory role of gamma interferon for host survival in a murine model of infection with Burkholderia pseudomallei. Infect Immun 1999; 67: 3593–600. 7. Haque A, Easton A, Smith D, et al. Role of T cells in innate and adaptive immunity against murine Burkholderia pseudomallei infection. J Infect Dis 2006; 193:370–9.

1248 • JID 2006:194 (1 November) • Haque et al.

8. Lauw FN, Simpson AJ, Prins JM, et al. Elevated plasma concentrations of interferon (IFN)–g and the IFN-g–inducing cytokines interleukin (IL)–18, IL-12, and IL-15 in severe melioidosis. J Infect Dis 1999; 180: 1878–85. 9. Ketheesan N, Barnes JL, Ulett GC, et al. Demonstration of a cell-mediated immune response in melioidosis. J Infect Dis 2002; 186:286–9. 10. Brett PJ, Mah DC, Woods DE. Isolation and characterization of Pseudomonas pseudomallei flagelli proteins. Infect Immun 1994; 62:1914–9. 11. Brett PJ, Woods DE. Structural and immunological characterization of Burkholderia pseudomallei O-polysaccharide-flagelli protein conjugates. Infect Immun 1996; 64:2824–8. 12. Jones SM, Ellis JF, Russell P, Griffi KF, Oyston PC. Passive protection against Burkholderia pseudomallei infection in mice by monoclonal antibodies against capsular polysaccharide, lipopolysaccharide or proteins. J Med Microbiol 2002; 51:1055–62. 13. Nelson M, Prior JL, Lever MS, Jones HE, Atkins TP, Titball RW. Evaluation of lipopolysaccharide and capsular polysaccharide as subunit vaccines against experimental melioidosis. J Med Microbiol 2004; 53: 1177–82. 14. Jones AL, Beveridge TJ, Woods DE. Intracellular survival of Burkholderia pseudomallei. Infect Immun 1996; 64:782–90. 15. Atkins T, Prior RG, Mack K, et al. A mutant of Burkholderia pseudomallei, auxotrophic in the branched chain amino acid biosynthetic pathway, is attenuated and protective in a murine model of melioidosis. Infect Immun 2002; 70:5290–4. 16. Ulett GC, Labrooy JT, Currie BJ, Barnes JL, Ketheesan N. A model of immunity to Burkholderia pseudomallei: unique responses following immunization and acute lethal infection. Microbes Infect 2005; 7: 1263–75. 17. Elvin SJ, Healey GD, Westwood A, Knight SC, Eyles JE, Williamson ED. Protection against heterologous Burkholderia pseudomallei strains by dendritic cell immunization. Infect Immun 2006; 74:1706–11. 18. Chen YS, Hsiao YS, Lin HH, Liu Y, Chen YL. CpG-modifie plasmid DNA encoding flagelli improves immunogenicity and provides protection against Burkholderia pseudomallei infection in BALB/c mice. Infect Immun 2006; 74:1699–705. 19. Stevens MP, Wood MW, Taylor LA, et al. An Inv/Mxi-Spa-like type III protein secretion system in Burkholderia pseudomallei modulates intracellular behaviour of the pathogen. Mol Microbiol 2002; 46:649–59. 20. Stevens MP, Friebel A, Taylor LA, et al. A Burkholderia pseudomallei type III secreted protein, BopE, facilitates bacterial invasion of epithelial cells and exhibits guanine nucleotide exchange factor activity. J Bacteriol 2003; 185:4992–6. 21. Mittrucker HW, Kaufmann SH. Immune response to infection with Salmonella typhimurium in mice. J Leukoc Biol 2000; 67:457–63. 22. Pamer EG. Immune responses to Listeria monocytogenes. Nat Rev Immunol 2004; 4:812–23. 23. Kaufmann SH. How can immunology contribute to the control of tuberculosis? Nat Rev Immunol 2001; 1:20–30. 24. Stevens MP, Haque A, Atkins T, et al. Attenuated virulence and protective efficac of a Burkholderia pseudomallei bsa type III secretion mutant in murine models of melioidosis. Microbiology 2004; 150: 2669–76. 25. Collazo CM, Galan JE. The invasion-associated type III system of Salmonella typhimurium directs the translocation of Sip proteins into the host cell. Mol Microbiol 1997; 24:747–56. 26. Musson JA, Hayward RD, Delvig AA, Hormaeche CE, Koronakis V, Robinson JH. Processing of viable Salmonella typhimurium for presentation of a CD4 T cell epitope from the Salmonella invasion protein C (SipC). Eur J Immunol 2002; 32:2664–71. 27. Falgarone G, Blanchard HS, Virecoulon F, Simonet M, Breban M. Coordinate involvement of invasin and Yop proteins in a Yersinia pseudotuberculosis-specifi class I-restricted cytotoxic T cell-mediated response. J Immunol 1999; 162:2875–83.