Bacterial Lipoprotein-Based Vaccines Induce Tumor Necrosis Factor-Dependent Type 1 Protective Immunity against Leishmania major

Abstract

Immunity against Leishmania major requires rapid induction of a type 1 immune response in which tumor necrosis factor alpha (TNF-α) plays an essential role. Hence, vaccination strategies that simulate the protective immune response found in hosts that have recovered from natural infection provide a rational approach to combat leishmaniasis. One method for optimizing the qualitative and quantitative immune responses after vaccination is to use an adjuvant. In this study we demonstrate that the OprI lipoprotein (L-OprI) from Pseudomonas aeruginosa induces a long-term cellular (gamma interferon [IFN-γ]) and humoral (immunoglobulin G2a) type 1 immune response against a truncated 32-kDa version (COOHgp63) of the 63-kDa major cell surface glycoprotein gp63. By contrast, immunization with COOHgp63 either fused to OprI nonlipoprotein or with no adjuvant did not result in the induction of type 1 immune responses. The adjuvanticity of L-OprI is strongly dependent on its capacity to induce TNF-α, since generation of type 1 immune responses is clearly delayed and impaired in TNF-α^−/− mice. Vaccination with L-OprICOOHgp63 fusion protein protected BALB/c mice against L. major infection for at least 19 weeks. Vaccinated mice were largely free of lesions or clearly controlled lesion size on termination of the experiment. The control of disease progression in mice vaccinated with L-OprICOOHgp63 was associated with enhancement of antigen-specific IFN-γ production. These data indicate that bacterial lipoproteins constitute appropriate adjuvants to include in vaccines against diseases in which type 1 immune responses are important for protection.

Cutaneous leishmaniasis (CL) caused by Leishmania major usually presents as small, localized lesions that heal within weeks or months without treatment. Individuals who have recovered from CL develop strong immunity against reinfection, suggesting that vaccination against leishmaniasis is feasible. Deliberate infection with L. major is a practice used for immunization and was reported to be highly effective (29). However, deliberate leishmanial infection is not recommended due to obvious safety concerns associated with injection of live virulent organisms (23). Vaccination using preparations of killed Leishmania in the presence of bacillus Calmette-Guérin has also been attempted (30, 36) but was shown to be infective in a controlled trial including several thousand volunteers in Iran (30). These findings suggest that vaccination using defined recombinant antigens may be a more effective and safe alternative for eliciting protective immune responses.

Human CL can be modeled by infecting inbred mice with L. major. Most strains (including C57BL/6) develop self-healing lesions with the development of an interleukin-12 (IL-12)-driven CD4⁺ type 1 helper T-cell (Th1) response. In contrast, BALB/c mice develop Th2 responses manifested primarily by IL-4, IL-10, and antibody production, which render them highly susceptible to L. major (19, 28, 35). Based on these immune correlates, a number of Leishmania antigens, including gp63 (43, 46, 47), LACK (16, 17), and PSA-2 (18, 37), have been used in a range of murine vaccination protocols with various results. In general, these vaccination protocols elicited partial protection against L. major, with resistant mouse strains producing smaller, more rapidly healing lesions and susceptible BALB/c mice having a reduced rate of parasite growth compared to unimmunized mice.

Recently, Wilhelm et al. (45) demonstrated that tumor necrosis factor (TNF) plays an indispensable role in the ultimate control of the infection. C57BL/6 mice lacking the TNF gene rapidly succumb to progressive visceral leishmaniasis after local infection with L. major. The uncontrolled replication of the parasites was due to a delay in generating a T-cell response, indicating that TNF plays a major role in the priming of protective adaptive immune responses. Hence, an immunization approach that incorporates this immune correlate may improve the qualitative and quantitative immune responses after vaccination.

Bacterial lipoproteins, among others, are molecules that stimulate cells of the innate immune system to produce cytokines such as TNF (34, 42) and IL-12 (4). Thereby, bacterial lipoproteins activate innate immune cells via Toll-like receptors (1, 4), and their signaling activity resides in the NH₂-terminal triacylated lipopeptide region (11, 44). The potent capacity of bacterial lipoproteins and/or lipopeptides to induce the production of IL-12 (4), a key signal of the innate immune system, may in turn trigger the development of adaptive immune responses, in particular type 1 T-lymphocyte responses (40). In fact, lipopeptides derived from the outer surface lipoproteins of Borrelia burgdorferi were reported to induce Th1 phenotype development (22). Furthermore, the synthetic lipid moiety analogue of bacterial lipoproteins (i.e., the tripalmitoyl-S-glyceryl-cysteine or Pam3Cys) was reported to increase the immunogenicity of heterologous antigens (2, 3, 10, 26). In a similar approach, we previously reported that fusion proteins between the major OprI lipoprotein of Pseudomonas aeruginosa and heterologous peptides or proteins were found to be highly immunogenic as evidenced by the induction of strong humoral and cytotoxic T-cell responses without the need for adjuvants (7, 25). In the present study, the OprI lipoprotein from P. aeruginosa was used to immunogenize the COOH-terminal part of the leishmanial antigen gp63 (6). Our results demonstrate that lipid-modified OprI-based vaccines elicit a TNF-α-dependent long-term type 1 immune response against the gp63 antigen and that such vaccinations confer significant protection against L. major infections.

MATERIALS AND METHODS

Mice.

Female BALB/c, C57BL/6, and lipopolysaccharide (LPS)-resistant C3H/HeJ mice 6 to 8 weeks of age were obtained from Harlan Nederland (Horst, The Netherlands). C57BL/6 TNF-α^−/− mice were obtained from the National Institute of Animal Health, Tsukuba City, Japan (39) and maintained in our animal facility.

Construction of the expression vector pCIMM2.

The P. aeruginosa mature oprI gene sequence contained in plasmid pVUB3 (8) was amplified by PCR with the primers 5′-GCGCGGATCCTGCAGCAGCCACTCCAAAGAAACCG-3′ and 3′-CTTTTTCGGTCGGCGTTCATTATTCGAACG CG-5′. Amplified DNA was purified, digested with BamHI and HindIII, and cloned downstream of a sequence encoding an oligohistidine peptide of six residues in the expression vector pQE-8 (Qiagen GmbH, Hilden, Germany) devoid of its EcoRI site. The resulting construct, pCIMM2, was transformed into competent JM109 cells and used for further subcloning of heterologous antigens into the lipoprotein I sequence.

Generation of lipidated and nonlipidated recombinant antigens.

The DNA fragment encoding the COOH-terminal part of L. major gp63 (8) (from residue 148 to residue 482 of the mature protein) was used to produce the three different recombinant forms. Generation of the lipidated L-OprICOOHgp63 fusion construct was described in detail previously (8). The recombinant vector producing the nonlipidated NL-OprICOOHgp63 protein was constructed by introducing the BglII-HindIII COOHgp63 DNA fragment (generated by digesting vector pVUB3:COOHgp63) into the His-tagged-OprI-producing pCIMM2 plasmid using standard methods. The recombinant His-tagged COOHgp63 protein was generated by directionally cloning the same BglII-HindIII COOHgp63 DNA fragment into the expression vector pQE32 (Qiagen GmbH) digested with BamHI and HindIII.

Expression and purification of recombinant antigens.

Induction of OprI fusion proteins or the His-tagged COOHgp63 with IPTG (isopropyl-β-d-thiogalactopyranoside) and preparation of outer membrane fractions were performed as described previously (7). Lipidated L-OprICOOHgp63 was purified from outer membrane fractions solubilized in a buffer containing 50 mM Tris-HCl (pH 8.0), 0.6% sodium dodecyl sulfate (SDS), and 10 mM β-mercaptoethanol. The outer membrane proteins were loaded onto a preparative SDS-polyacrylamide column and purified by continuous-elution electrophoresis using a model 491 Prep Cell (Bio-Rad Laboratories, Hercules, Calif.) according to the manufacturer’s instructions. Both His-tagged proteins, NL-OprICOOHgp63 and COOHgp63, were purified by affinity chromatography under denaturing conditions using Ni-nitrilotriacetic acid Superflow resin (Qiagen GmbH) and concentrated by using a VIVASPIN concentrator (VIVASCIENCE, Lincoln, United Kingdom) that was previously treated with 0.02% pluronic acid for 10 min. Both proteins were then repurified by continuous-elution electrophoresis as described above. Finally, the three different recombinant proteins were subjected to two successive gel filtration chromatographies with an AKTA Explorer (Amersham Pharmacia/Biotech, Uppsala, Sweden) using Superdex-75 HR10/30 (Pharmacia/Biotech) in order to remove LPS (21) and eluted in a buffer containing 20 mM Tris-HCl (pH 8.0), 100 mM NaCl, 20 mM glycine, and 0.01% SDS. The protein concentration was determined using the Bio-Rad DC protein assay. LPS in the protein suspension was determined with the Limulus amebocyte lysate assay (Biowhittaker, Inc., Walkersville, Md.), and the levels were comparable for all three proteins (<0.07 ng/μ g of protein).

Immunizations.

Groups of five BALB/c, C57BL/6, and C57BL/6 TNF-α^−/− mice were immunized subcutaneously by one injection or three injections 10 days apart with 1 μ g of either L-OprICOOHgp63, NL-OprICOOHgp63, or COOHgp63. Preimmune sera were taken 1 day before the first immunization. At 7 or 10 days after the first or third immunization, respectively, mice were killed. Sera, spleens, and draining lymph nodes (sacral lymph nodes) were taken to analyze the immune response. To analyze long-term immune responses, sera and spleens were taken 12 weeks after a third immunization.

Cytokine assays.

Homogeneous lymph node and spleen cell suspensions from individual mice were prepared in RPMI 1640 medium supplemented with fetal calf serum (10%), penicillin-streptomycin (100 U and 100 μ g/ml, respectively), l-glutamine (2 mM), 2-mercaptoethanol (5 × 10⁻⁵ M), minimal essential medium amino acid solution (1×), and sodium pyruvate (1 mM). Cells (2 × 10⁶) were separately stimulated with COOHgp63 or without antigen at 37°C in 24-well flat-bottom tissue culture plates (Becton Dickinson, Franklin Lakes, N.J.). Gamma interferon (IFN-γ), IL-4, and IL-10 in culture supernatants taken 24, 48, 72, and 96 h after priming were determined. The cytokine levels were analyzed by a sandwich enzyme-linked immunosorbent assay (ELISA) according to the instructions of the supplier (Pharmingen, San Diego, Calif.). Data are represented as mean cytokine concentrations over 4 days.

Measurement of antibody titers.

Immunoglobulin isotype titers in the preimmune and immune sera were measured by using ELISA (Southern Biotechnology Associates, Inc., Birmingham, Ala.). Briefly, 96-well Nunc-Immuno plates (Nalge Nunc International, Roskilde, Denmark) were coated with 0.5 μ g of COOHgp63 per well, and after exposure to diluted preimmune or immune sera, bound antibodies were detected with horseradish peroxidase-labeled goat anti-mouse IgG1 and IgG2a. ELISA titers were specified as the last dilution of the sample whose absorbance was greater than threefold the preimmune serum value.

Induction of TNF-α production in peritoneal exudate cells after antigen stimulation.

Peritoneal exudate cells from LPS-resistant C3H/Hej mice were harvested by washing the peritoneal cavity with 10 ml of ice-cold sucrose solution (0.34 M). The cells were washed in supplemented RPMI 1640 and left to adhere for 2 h at 37°C in 24-well flat-bottom tissue culture plates (Becton Dickinson) at a concentration of 10⁶ cells/ml. The peritoneal exudate cells were stimulated or not with recombinant murine IFN-γ (100 U/ml) (Life Technologies, Ltd., Paisley, Scotland) in the presence of lipidated and nonlipidated OprICOOHgp63 or COOHgp63. After overnight incubation in a humidified atmosphere of 5% CO₂, supernatants were collected to determine TNF-α concentrations by using the DuoSet ELISA development system (R&D Systems, Abingdon, United Kingdom).

L. major challenge.

Groups of 15 BALB/c mice were subcutaneously immunized three times with 1 μ g of either L-OprICOOHgp63, NL-COOHgp63, or COOHgp63 in the flank. A control group was immunized with the buffer in which the proteins were dissolved. Ten days after the third dose, mice were challenged subcutaneously with 10⁶ stationary-phase L. major promastigotes (L119, I.C. MTAT/KE/00/T4) in the base of the tail. Progress of the disease was monitored weekly by scoring lesion development.

RESULTS

The lipid moiety of OprICOOHGp63 fusion protein is required for the induction of long-term type 1 immune responses.

To evaluate the potential adjuvant capacity of lipoprotein I of P. aeruginosa with heterologous proteins and the contribution of its lipid moiety to the immunogenicity of the chimeric OprICOOHgp63 lipoprotein, three different recombinant proteins were produced: the lipidated L-OprICOOHgp63, the nonlipidated NL-OprICOOHgp63, and COOHgp63 (Fig. 1). All three recombinant proteins contain the COOH-terminal domain of glycoprotein gp63 of L. major, which contains the host-protective T-cell epitopes (47).

FIG. 1.

Formulations of the three recombinant gp63 preparations used in this study. Important amino acids are given in one-letter code. MCS, multiple cloning site; aa, amino acids.

BALB/c mice were immunized subcutaneously once or three times with the recombinant proteins to analyze, respectively, early cellular immune responses in the draining lymph nodes and secondary immune responses, elicited against the heterologous COOHgp63 antigen. In vitro restimulation with COOHgp63 of lymph node cells from BALB/c mice immunized once with either type of lipoprotein construct or COOHgp63 resulted in a clear induction of IFN-γ secretion for L-OprICOOHgp63-immunized BALB/c mice (Fig. 2a). Marginal levels of IL-4 (<30 pg/ml) could be detected in lymph node cells from any immunized group. The induction of a type 1 immune response upon immunization with L-OprICOOHgp63 was sustained after three immunizations as evidenced by high cellular proliferation (data not shown) and the production of high IFN-γ levels in the spleen compartment, recorded 10 days and 12 weeks after the last immunization. The induction of IL-4 was negligible (<30 pg/ml). Three immunizations with NL-OprICOOHgp63 or COOHgp63 resulted in the induction of very low levels of IL-4 (<30 pg/ml) or IFN-γ (<15 pg/ml) when measured 10 days after the last immunization dose (Fig. 2b).

FIG. 2.

T-cell (a and b) and humoral (c) responses of BALB/c mice immunized with L-OprICOOHgp63, NL-OprICOOHgp63, or COOHgp63. The production of IFN-γ and IL-4 from lymph node cells taken 7 days after a single immunization (a) and from spleen cells taken 10 days or 12 weeks (only for L-OprICOOHgp63-immunized mice) after a third immunization dose (b) was quantified. (c) IgG antibody titers against COOHgp63 in sera from BALB/c mice 10 days after mice received the third injection of the indicated protein. Results are expressed as the means ± standard deviations for five individual mice (spleen cells and sera) or as data from pooled samples from five mice (lymph node cells). Data in panels b and c are representative of those from two independent experiments, and data in panel a represent a single experiment.

Since antibody production reflects the ability of the host to respond to antigen efficiently and a proportional distribution of serum isotypes mirrors a bias toward either a type 1 or type 2 immune responses, we investigated the levels of antigen-specific antibodies of the different isotypes postimmunization. Three immunizations with the lipidated OprICOOHgp63 induced high titers of COOHgp63-specific IgG2a and IgG1 antibodies 10 days after the third immunization (Fig. 2c). At 12 weeks, the level of IgG2a antibodies was sustained, whereas the titer of IgG1 antibodies was significantly reduced compared to the titer at 10 days postvaccination (data not shown). Immunization with nonlipidated OprICOOHgp63 or COOHgp63 also resulted in COOHgp63-specific antibody production, but the titers were significantly lower (0.5- to 2-log-unit and 3-log-unit reductions for IgG1 and IgG2a, respectively) than those obtained upon L-OprICOOHgp63 immunization (Fig. 2c).

An effective way of using an adjuvant is to simply mix it with an antigen of interest. In this context, we wondered whether the immune responses induced by L-OprI to a heterologous antigen require covalent linkage of the two entities. To examine this, BALB/c mice were immunized subcutaneously with three injections of 1 μ g of L-OprICOOHgp63 or 0.8 μ g of COOHgp63 mixed with 0.2 μ g of L-OprI. In vitro restimulation of spleen cells from mice immunized with L-OprI mixed with COOHgp63 resulted in a clear induction of IFN-γ, but the induction was about threefold higher for mice immunized with linked L-OprI (Fig. 3). Accordingly, mixing of L-OprI with COOHgp63 elicited production of IgG2a and IgG1 isotype antibodies against COOHgp63 that was 10-fold lower than that induced by immunization with the L-OprICOOHgp63 fusion protein (data not shown).

FIG. 3.

In vitro production of IFN-γ and IL-4 from spleen cells of BALB/c mice immunized with either L-OprICOOHgp63 or COOHgp63 mixed with L-OprI. Results are expressed as the means ± standard deviations for five individual mice. Data are representative of those from two independent experiments.

The type 1 immune response-inducing potential of L-OprICOOHgp63 is TNF-α dependent.

TNF-α, secreted by lipoprotein-activated macrophages (34, 42), has been suggested to be a key molecule, together with IL-12, in the induction of IFN-γ production and amplification of type 1 immune responses (5, 41). Therefore, it was of interest to test whether (i) OprI-based lipoproteins induce TNF-α production by macrophages and (ii) TNF-α contributes to the type 1 adjuvant activity of OprI. Macrophages (unactivated or activated with IFN-γ) from endotoxin-hyporesponsive C3H/HeJ mice were stimulated in vitro with either the lipidated COOHgp63, the nonlipidated COOHgp63, or the COOHgp63 antigen. As shown in Fig. 4, a dose-dependent induction of TNF-α in unprimed macrophages was observed with the lipidated L-OprICOOHgp63. Moreover, the TNF-α-inducing activity of L-OprICOOHgp63 was strongly increased in IFN-γ-primed macrophages (Fig. 4). Similar results were obtained upon incubation of macrophages with L-OprI (data not shown). Under these experimental conditions, both the nonlipidated OprI-COOHgp63 and the COOHgp63 elicited marginal levels of TNF-α synthesis.

FIG. 4.

The lipid tail of L-OprICOOHgp63 is required to induce TNF-α release by peritoneal macrophages either activated or not with IFN-γ.

To test whether the TNF-α-inducing capacity of L-OprICOOHgp63 contributes to its type 1 immune response-inducing potential, one and three immunizations with L-OprICOOHgp63 were performed with C57BL/6 TNF-α^−/− mice. As shown in Fig. 5a, early priming of COOHgp63-specific IFN-γ production was markedly reduced in the culture supernatants of draining lymph node or spleen cells from L-OprICOOHgp63-immunized TNF-α^−/− mice compared to immunized C57BL/6 wild-type mice. Analysis of the antigen-specific cellular immune response after three L-OprICOOHgp63 immunizations revealed the induction of a significant level of IFN-γ production in the spleens and lymph node compartments of TNF-α^−/− mice, although it was markedly lower than that observed in C57BL/6 wild-type mice (Fig. 5b). The induction of IL-4 was negligible (>30 pg/ml).

FIG. 5.

The lipoprotein-induced type 1 immune response is affected in C57BL/6 TNF-α knockout mice (TNF-α^−/− mice). (a and b) IFN-γ and IL-4 production in spleen cells (Sp) from five individual mice or pooled sacral lymph nodes (LN) from these mice immunized with one (a) or three (b) doses of L-OprICOOHgp63. (c) IgG antibody titers against COOHgp63 in sera from wild-type (WT) and TNF^−/− C57BL/6 mice 10 days after mice received the third dose of L-OprI COOHgp63. Results are expressed as the means ± standard deviations for five individual mice (spleen cells) or pooled samples from five mice (lymph node cells and sera).

Analysis of the humoral responses elicited with L-OprICOOHgp63 (after three immunizations) in wild-type and TNF-α^−/− C57BL/6 mice revealed that anti-COOHgp63 IgG2a responses were severely reduced in C57BL/6 TNF-α^−/− mice (Fig 5c). In contrast, the magnitude of IgG1 subclass responses was unaffected in immunized C57BL/6 TNF-α^−/− mice compared to C57BL/6 wild-type mice. Together, these data suggest that the type 1 immune response elicited by L-OprI is strongly TNF-α dependent.

Vaccination with OprI-based COOHgp63 lipoproteins protects highly susceptible BALB/c mice against Leishmania challenge.

During infection with L. major, resistant C57BL/6 mice mount a polarized type 1 cellular immune response mediated by IFN-γ production (28, 35). Recently, Wilhelm et al. demonstrated that TNF-α plays a crucial role in cell recruitment and priming of the immune response in C57BL/6 mice (45). Since lipid-modified OprICOOHgp63 selectively induces an IFN-γ-producing type 1 immune response that is dependent on the presence of TNF-α, we wondered whether vaccinations with this lipoprotein could mirror the antileishmanial immunity developed by C57BL/6 mice and protect susceptible BALB/c mice against Leishmania challenge.

In a preliminary experiment we assessed the role of TNF-α in disease outcome for the L. major strain to be used in the protection experiments. Wild-type and TNF-α^−/− C57BL/6 mice were infected at the tail base with 10⁶ L. major promastigotes, and the course of infection was monitored. Corroborating the results obtained by Wilhem et al. (45), the parasite disseminated rapidly in C57BL/6 TNF-α^−/− mice and formed multiple skin lesions, in sharp contrast to the case for infected BALB/c mice, which developed a single ulcerating skin lesion at the site of infection. Whereas all C57BL/6 TNF-α^−/− mice died from the infection, C57BL/6 wild-type mice controlled the infection and survived, indicating that TNF-α signaling is necessary to confer protection against the L. major strain used (data not shown).

Next, BALB/c mice were immunized with three injections of the lipidated OprICOOHgp63, the nonlipidated counterpart, or COOHgp63. Immunized animals were subsequently challenged by injection of 10⁶ L. major promastigotes at the tail base 10 days after the last immunization, and lesion development was scored weekly. As depicted in Fig. 6, mice vaccinated with L-OprICOOHgp63 were highly protected against the development of cutaneous lesions. Whereas the lesions appeared at week 5 in the unvaccinated mice, no lesions were observed in the L-OprICOOHgp63-vaccinated group until week 9. In contrast, mice vaccinated with either NL-OprICOOHgp63 or COOHgp63 developed lesions similar in size to those of the unvaccinated group, although a slight delay in the onset of lesions was observed (Fig. 6). After 12 weeks of infection, 73% of L-OprICOOHgp63-vaccinated animals still remained without lesions. L-OprICOOHgp63-vaccinated mice that did develop lesions clearly controlled the lesion size compared to the other groups, which developed progressive lesions (Fig. 6). Further follow-up of a subgroup of five mice immunized with L-OprICOOHgp63 showed that an effective immune response was sustained for at least 19 weeks after infection, since no lesions were formed during this time period (data not shown).

FIG. 6.

The OprI-based COOHgp63 lipoprotein protects BALB/c mice against Leishmania challenge. Groups of 15 BALB/c mice were vaccinated subcutaneously three times with the lipidated L-OprICOOHgp63, the nonlipidated NL-OprICOOHgp63, or COOHgp63. Controls were injected with buffer. Mice were infected with 10⁶ live promastigotes 10 days after the last immunization, and lesion development was monitored weekly.

Protection of L-OprICOOHgp63-vaccinated BALB/c mice against Leishmania challenge is correlated with a persistent type 1 immune response.

Susceptibility of BALB/c mice to Leishmania infection is associated with a mixed immune response due to an interaction between Th1 and Th2 responses in complex feedback loops, ultimately resulting in a predominant increase of type 2 cytokines (35). Hence, specific analysis of the immune response of protected mice against the potential vaccine candidate allows determination of whether the type of immune response associated with protection still persists after challenge. Therefore, the ongoing immune response in the protected L-OprICOOHgp63-immunized mice and the unprotected control mice was analyzed at 14 weeks after challenge. As depicted in Fig. 7a, the antigen-specific IFN-γ production in the spleen compartment was sustained in protected L-OprIgp63COOHgp63-vaccinated BALB/c mice, whereas low to undetectable levels of IFN-γ were present in spleen cells from unprotected control mice. At the humoral level (Fig. 7b), protected mice produced substantially higher levels of COOHgp63-specific IgG2a antibody titers (65-fold higher) than unprotected control mice, whereas the increase in levels of IgG1 antibodies was only 10-fold. Collectively, these results show that protected mice produce a persistent type 1 immune response that correlated with a strong IgG2a antibody response and the presence of IFN-γ-secreting Th1 cells.

FIG. 7.

Release of IFN-γ and IL-4 from spleen cells (a) and anti-gp63 antibody subclass responses (b) in protected L-OprICOOHgp63-immunized BALB/c mice (P) and control mice (C). The control mice were immunized with buffer and challenged on the same day as L-OprICOOHgp63-immunized mice. Production of IFN-γ and IL-4 and IgG antibody titers were assayed 14 weeks after L. major challenge. Error bars indicate standard deviations.

DISCUSSION

Analysis of the mechanisms and mediators utilized by pathogens to trigger systemic immune responses upon infection may lead to the development of new immunomodulatory strategies (14). Pathogen-derived lipoproteins are potent immunomodulators that alert innate immune defenses (20). This study aimed at determining whether the OprI lipoprotein from P. aeruginosa could modulate the immune response against a heterologous Leishmania antigen in such a way that type 1 immune responses develop and, consequently, susceptible BALB/c mice can control the infection. Furthermore, the mechanisms underlying the adjuvant activity of OprI were further unraveled in the context of the induction of antileishmanial immunity.

Comparative analysis of lipidated OprICOOHgp63, the nonlipidated counterpart, and COOHgp63 recombinant proteins in immunized mice demonstrated the crucial importance of the lipid tail of P. aeruginosa lipoprotein I in inducing type 1 immune responses against the heterologous antigen as evidenced by the cytokine pattern and profile of antibody subclass production. Indeed, immunization with L-OprICOOHgp63 biased the T-cell response towards IFN-γ production, indicating a preferential induction of a type 1 immune response. Likewise, a type 1 immune response was obtained when L-OprI was simply mixed with the COOHgp63 antigen. However, the induced type 1 immune response was lower in terms of IFN-γ production and IgG2a antibody isotype titers than that with L-OprI covalently linked to the COOHgp63 antigen. Since an effective way of using an adjuvant is to simply mix it with an antigen of interest, further work should be performed to evaluate whether the magnitude of the type 1 immune response induced against COOHgp63 upon mixing with L-OprI in different molar ratios reaches the threshold to mediate protective immunity. Overall, these results corroborate other data showing that synthetic lipopeptides encompassing the N-terminal sequence of B. burgdorferi trigger the development of a type 1 phenotype (22). Besides the induction of IFN-γ-producing cells, our results also demonstrate that the lipid tail of OprI potentiates the induction of humoral responses against a heterologous antigen, since immunizations with this L-OprI-based vaccine increased or triggered IgG2a subclass responses against COOHgp63. Besides the selective triggering of IgG2a isotype antibodies, L-OprICOOHgp63 also triggered the production of IgG1 isotype antibodies. This result at first sight conflicts with the statement of a dominant type 1 immune response. However, Faquim-Mauro et al. (12) recently showed that mouse IgG1 antibodies comprise two functionally distinct types: one type of IgG1 has anaphylactic activity and its synthesis is IL-4 dependent, whereas the other type lacks activity and its synthesis is stimulated by IL-12 or IFN-γ. Moreover, have they shown that adjuvants can modulate the levels of anaphylactic- and non-anaphylactic-type IgG1 antibodies produced in response to a particular antigen (13). Thus, one may consider the possibility that the early IFN-γ production induced upon immunization with L-OprICOOHgp63 accounts for isotype switching to non-anaphylactic-type IgG1 antibodies.

The capacity of L-OprI, either covalently linked to or mixed with COOHgp63, to induce acquired immune responses may reflect its potential to trigger innate immune cells. Corroborating other reports that bacterial lipoproteins are potent inducers of TNF-α production (34, 42), our results show that only lipidated OprI was capable of stimulating significant TNF-α production by either naive or IFN-γ-primed macrophages. Local production of TNF-α may in turn signal the development of type 1 acquired immune responses. Indeed, this cytokine was documented to induce the expression of B7-like costimulatory signals (38), IFN-γ production by T cells (5, 9) and NK cells (41), and type 1 antibody subclass responses (i.e., IgG2a) (33). The involvement of TNF-α in the genesis and/or progression of cellular and humoral type 1 acquired immune responses to leishmanial antigens is further substantiated here, since both type 1 cytokine (IFN-γ) and humoral subclass (IgG2a) responses against the heterologous antigen were severely compromised in L-OprICOOHgp63-immunized TNF-α^−/− mice. Hence, the defective induction of type 1 responses observed in L-OprICOOHgp63-immunized TNF-α^−/− mice most probably reflects the TNF-α-inducing potential of OprI by virtue of its lipid tail. According to our in vivo results and other investigations (5, 32, 41), TNF-α can be considered a component of the innate immune system which, synergistically with or alternatively to IL-12, bridges the gap between innate and acquired immunity. Finally, since the TNF-α-inducing capacity of OprI is strongly increased upon macrophage priming with IFN-γ, TNF-α-mediated induction of IFN-γ production by OprI-based vaccines may further amplify ongoing or subsequent OprI-elicited immune responses.

It is well established that immunological control of L. major infections depends on the production of IFN-γ, which activates macrophages to kill the parasites via induction of NO production (15, 24, 28, 31). Accordingly, the capacity of L-OprICOOHgp63 to elicit COOHgp63-specific IFN-γ-producing memory cells is reflected by the induction of a significant protection against L. major infections in the highly susceptible BALB/c model. Our analysis of the immune response in protected mice further corroborates the ability of L-OprICOOHgp63 to induce a potent type 1 immune response. There was preferential induction of COOHgp63-specific production of IFN-γ before infection (Fig. 2b), and this response was sustained for at least 14 weeks after infection, resulting in control of lesion development even at 19 weeks after infection. It should be noted, however, that the levels of IFN-γ detected in these protected mice were lower than those detected in immunized mice prior to challenge (1,200 versus 8,000 pg/ml). In addition, although protected mice had an increase in IgG2a (consistent with increased IFN-γ), they also had levels of IgG1 similar to those in the unprotected mice. Similar results have been obtained by Gurunathan et al. (17) upon vaccination with DNA encoding LACK parasite antigen. LACK DNA vaccination induced a potent Th1-type response with preferential induction of LACK-specific production of IFN-γ before infection, and this response was sustained for at least 6 weeks after infection in protected mice. The role of IL-12 in mediating both the IFN-γ response and the protective effect induced by LACK DNA vaccination was demonstrated by the fact that neutralizing IL-12 in vivo decreased the LACK-specific IFN-γ production by lymph node cells and completely eliminated protection. Control mice developed progressive disease upon parasite infection and produced no LACK-specific IFN-γ prior to or 6 weeks after infection. On the other hand, IL-4 was not significantly different among the two groups. These observations were further confirmed in vivo by showing that although protected mice had an increase in IgG2a (consistent with increased IFN-γ), they also had levels of IgG1 similar to those in the unprotected mice. Those authors proposed that the fact that IgG1 was increased in healing mice in which IFN-γ was enhanced was consistent with previous work showing that IL-12 treatment could induce an increase in both IgG1 and IgG2a (17).

Taking these findings together with our own data suggests that challenge with L. major parasites may have provoked the emergence of a type 2 immune response but that vaccination with L-OprICOOHgp63 promotes a dominant type 1 immune response sufficient to control infection. Additional studies on parasite numbers at the site of infection and draining lymph nodes should further determine whether the immune response obtained after vaccination resulted in sterile immunity or whether a more complete polarization of the immune response is needed to control parasite growth.

Taking into account that L-OprI-based immunization is highly TNF-α dependent, it is worth mentioning that vaccination with leishmanial antigens together with TNF-α prevents disease enhancement and induces protective immunity against L. major infection in susceptible BALB/c mice (27). In addition, our results show that C57BL/6 TNF-α^−/− mice rapidly succumb to infection with L. major compared to C57BL/6 wild-type mice. Moreover, vaccination of C57BL/6 TNF-α^−/− mice with L-OprICOOHgp63 results in a clear delay in the induction of type 1 cellular responses (IFN-γ) and substantial deficiencies in the IgG response. These findings corroborate the results obtained by Wilhelm et al. (45), using another L. major strain, showing that an absence of TNF-α resulted in a retarded proliferative response of T cells, a delayed or missing switch to antigen-specific, T-cell-dependent serum Igs, and an impaired formation of inducible NO synthase-positive cell clusters.

To conclude, the use of bacterial lipoproteins may represent a general strategy to activate innate antimicrobial defense mechanisms and to amplify type 1 immune responses to a variety of antigens in vaccines.