Chlamydia muridarum T Cell Antigens and Adjuvants That Induce Protective Immunity in Mice

Abstract

Major impediments to a Chlamydia vaccine lie in discovering T cell antigens and polarizing adjuvants that stimulate protective immunity. We previously reported the discovery of three T cell antigens (PmpG, PmpF, and RplF) via immunoproteomics that elicited protective immunity in the murine genital tract infection model against Chlamydia infection after adoptive transfer of antigen-pulsed dendritic cells. To expand the T cell antigen repertoire necessary for a Chlamydia vaccine, we evaluated 10 new Chlamydia T cell antigens discovered via immunoproteomics in addition to the 3 antigens reported earlier as a molecular subunit vaccine. We first tested five adjuvants, including three cationic liposome formulations (dimethyldioctadecylammonium bromide-monophosphoryl lipid A [DDA-MPL], DDA-trehalose 6,6′-dibehenate [DDA-TDB {CAF01}], and DDA-monomycolyl glycerol [DDA-MMG {CAF04}]), Montanide ISA720–CpG-ODN1826, and alum using the PmpG protein as a model T cell antigen in the mouse genital tract infection model. The results showed that the cationic liposomal adjuvants DDA-MPL and DDA-TDB elicited the best protective immune responses, characterized by multifunctional CD4⁺ T cells coexpressing gamma interferon (IFN-γ) and tumor necrosis factor alpha (TNF-α), and reduced infection by more than 3 logs. Using DDA-MPL as an adjuvant, we found that 7 of 13 Chlamydia T cell antigens (PmpG, PmpE, PmpF, Aasf, RplF, TC0420, and TC0825) conferred protection better than or equal to that of the reference vaccine antigen, major outer membrane protein (MOMP). Pools of membrane/secreted proteins, cytoplasmic proteins, and hypothetical proteins were tested individually or in combination. Immunization with combinations protected as well as the best individual protein in that combination. The T cell antigens and adjuvants discovered in this study are of further interest in the development of a molecularly defined Chlamydia vaccine.

INTRODUCTION

Chlamydia trachomatis infection is the most common sexually transmitted bacterial infection in humans. It is estimated that more than 92 million new sexually acquired infections occur annually worldwide (35). C. trachomatis infects epithelial cells lining the reproductive tract and can cause severe complications, such as pelvic inflammatory disease, ectopic pregnancy, and infertility in women. Although antibiotic therapy is effective, most cases are asymptomatic, with the potential to escape treatment and cause persistent infection with tissue pathology. Furthermore, public health programs to control C. trachomatis based on case identification, treatment, and contact tracing appear to be failing, since case rates have steadily risen during the past decade, perhaps due to effects on herd immunity (5). Therefore, a vaccine is likely to be the most effective option for control of Chlamydia.

Early vaccine trials in both human and nonhuman primates with whole inactivated C. trachomatis ultimately failed because of incomplete and short-lived protection or worse inflammation postvaccination (16, 39). Due to the failure of whole-cell vaccine, contemporary C. trachomatis vaccine research has focused on molecularly defined subunit vaccines that are administered with an adjuvant or other delivery vehicle to enhance immunogenicity (6, 33). To date, the Chlamydia major outer membrane protein (MOMP) has been the leading Chlamydia vaccine candidate in multiple animal studies, including nonhuman primate models (20, 28, 32). However, protection elicited by MOMP is variably incomplete, which has been partially correlated with the conformational state of MOMP and the adjuvants used. Other candidate antigens that trigger T cell responses in humans and mice have therefore been proposed, including outer membrane protein 2 (OMP2) (12), class I accessible protein 1 (Cap1) (14), cysteine-rich protein A (CrpA) (34), Chlamydia heat shock protein 60 (HSP60) (34), polymorphic membrane protein D (PmpD) (15), homolog of Yersinia pseudotuberculosis YopD (YopD) (15), enolase (15), and Chlamydia protease-like activity factor (CPAF) (8, 24). None of these candidates are demonstratedly better than MOMP, and thus the search for additional protective T cell antigens continues.

Experience has shown that developing vaccines for intracellular pathogens such as Chlamydia that require protective cell-mediated immunity (CMI) is more difficult than doing so for extracellular pathogens that require protective antibody (31). Part of the problem has been the identification of T cell antigens that induce CMI responses, because such antigens need to be presented to T cells by major histocompatibility complex (MHC) molecules, and identifying MHC-bound microbial epitopes has been notoriously difficult (6). Immunity to Chlamydia is currently understood as being mediated by T cells and mainly relies on tissue trafficking CD4⁺ T cells producing gamma interferon (IFN-γ) and tumor necrosis factor alpha (TNF-α) (25, 37, 38). Antibody may play a secondary role in resistance to reinfection. We used an immunoproteomic approach to directly identify MHC class II-bound Chlamydia peptides eluted from Chlamydia muridarum-infected dendritic cells (DCs) (22). We evaluated the proteins containing these MHC class II-bound peptides and found three T cell antigens (PmpG, PmpF, and RplF) that elicited protective immunity in the murine genital tract infection model after adoptive transfer of antigen-pulsed dendritic cells (41). We continued our search for additional T cell antigens using immunoproteomics and selected 13 proteins, including the three mentioned above, for further evaluation.

An additional challenge in developing an effective Chlamydia T cell vaccine is identifying the ideal adjuvant that elicits potent protective cellular immunity in vivo. In this study, we evaluated five adjuvants, including three cationic liposome formulations, Montanide ISA720–CpG-ODN1826 (Mon-CpG), and alum, that are already in use or have the potential to be used in humans. The three cationic liposome formulations are all based on dimethyldioctadecylammonium bromide (DDA) as a vehicle. Two of the formulations (CAF01 and CAF04, developed at the Staten Serum Institut) are based on immunomodulators from the mycobacterial cell wall (trehalose 6,6′-dibehenate [TDB] and monomycolyl glycerol [MMG]). The third cationic formulation is a mixture of DDA and monophosphoryl lipid A (MPL) as an immunomodulator. Montanide is a squalene-based water-in-oil emulsion. CpG oligodeoxynucleotide (CpG-ODN) is a potent Th1-polarizing adjuvant that activates antigen-presenting cells endosomally through Toll-like receptor 9 (TLR9). The Mon-CpG adjuvant has been reported to induce potent protective immunity against Chlamydia genital infection when formulated with MOMP (27). Alum adjuvant is routinely used in human vaccines, including subunit vaccines, to enhance antibody production.

MATERIALS AND METHODS

Chlamydia.

C. muridarum strain Nigg (the mouse pneumonitis strain) was grown in HeLa 229 cells in Eagle's minimal essential medium (MEM) (Invitrogen) supplemented with 10% fetal calf serum (FCS). Elementary bodies (EBs) were purified by discontinuous density gradient centrifugation and stored at −80°C as previously described (7). The infectivity and the number of inclusion-forming units (IFU) of purified EBs were determined by infection of HeLa 229 cells and enumeration of inclusions that were stained by anti-EB mouse polyclonal antibody followed by biotinylated anti-mouse IgG (Jackson ImmunoResearch) and a 3,3′-diaminobenzidine (DAB) substrate (Vector Laboratories) (40). Heat-killed EBs (HK-EB) were prepared by heating to 56°C for 30 min.

Molecular cloning, expression, and purification of C. muridarum recombinant proteins.

The recombinant proteins PmpG, PmpF, RplF, Aasf, TC0420, and MOMP were cloned and expressed as previously described (41, 42). PmpE, PmpH, Ide, RecO, Tarp, AtpE, TC0825, and TC0285 were cloned, expressed, and purified as follows: pmpE, pmpH, ide, recO, tarp, atpE, TC0825, and TC0285 DNA fragments were generated by PCR using genomic DNA isolated from C. muridarum. PCRs were carried out using Herculase Enhanced DNA polymerase (Agilent Technologies). The PCR product was purified with the QIAquick PCR purification kit (Qiagen), and the purified DNA fragments were cloned into the pET32a expression vector (Novagen) after restriction enzyme digestion with SacI/NotI using standard molecular biology techniques. For pmpE, pmpH, and tarp, only the first half of the gene (representing amino acids 26 to 600, 26 to 600, and 42 to 608, respectively) were cloned into the vector for expression. The sequences of the subcloned genes were confirmed by sequencing. Plasmids containing the pmpE, pmpH, ide, recO, tarp, atpE, TC0825, and TC0285 genes were transformed into the Escherichia coli strain BL21(DE3) (Stratagene), where protein expression was carried out by inducing the lac promoter for expression of T7 RNA polymerase using isopropyl-β-d-thiogalactoside pyranoside. The expressed PmpE, PmpH, Ide, RecO, Tarp, AtpE, TC0825, and TC0285 proteins with N-terminal His tags were purified by using a nickel column with the His bind purification system (Qiagen). Lipopolysaccharide (LPS) removal of these proteins was carried out by adding 0.1% Triton-114 in one of the wash buffers during purification.

Adjuvants.

Five adjuvants (DDA-MPL, DDA-TDB [CAF01], DDA-MMG [CAF04], Montanide ISA 720–CpG-ODN1826 [Mon-CpG], and alum) were evaluated in the present study. Dimethyldioctadecylammonium bromide (DDA) (product no. 890810P) and monophosphoryl lipid A (MPL) (product no. 699800P) were purchased from Avanti Polar Lipids (Alabaster, AL). For the DDA-MPL formulation, DDA was mixed into 10 mM Tris buffer at pH 7.4 and heated to 80°C while being stirred continuously on a magnetic hot plate for 20 min and then cooled to room temperature. MPL was suspended in distilled water (dH₂O) containing 0.2% triethylamine. The mixture was heated in a 70°C water bath for 30 s and then vortexed for 60 s. The heating and vortexing procedure was repeated three times. The MPL was mixed with DDA immediately before use. The emulsion consisted of 250 μg DDA and 25 μg MPL per 100 μl. The DDA-TDB (CAF01) and DDA-MMG (CAF04) were both provided from the Statens Serum Institut as stable proprietary formulations and were prepared as previously described (3, 17). For Mon-CpG formulation, Montanide ISA 720 (Seppic, Inc.) was used at a 30:70 volume ratio of immunogen plus CpG-ODN 1826 (5′-TCCATGACGTTCCTGACGTT-3′, phosphorothioate modified; Integrated DNA Technologies, Inc.) to Montanide. Alum (Imject alum; Thermo) was used at 1:1 volume ratio of alum to immunogen.

Mice.

Female C57BL/6 mice (5 to 6 weeks old) were purchased from Charles River Canada (Saint Constant, Canada). The mice were maintained and used in strict accordance with University of British Columbia guidelines for experimental use of animals.

Immunization.

All mice except the live-EB group were immunized three times subcutaneously (s.c.) at the base of the tail at 2-week intervals. Mice intranasally infected with 1,500 inclusion-forming units (IFU) of live C. muridarum (EBs) were used as positive controls. A group of mice immunized with phosphate-buffered saline (PBS) was used as a negative control. Four weeks after the last immunization or 8 weeks after live C. muridarum intranasal infection, mice were intravaginally challenged with live C. muridarum for protection evaluation.

Two mouse trials were conducted in this study. In the first trial, groups of 14 C57BL/6 mice were immunized with the PmpG protein in 100 μl 10 mM Tris buffer (pH 7.4), mixed by vortexing with 100 μl DDA-MPL, DDA-TDB, DDA-MMG, or Alum. Mon-CpG per dose consisted of 30 μl Montanide ISA 720 emulsified with 70 μl immunogen solution of 5 μg PmpG protein plus 10 μg CpG-ODN 1826 in 10 mM Tris buffer (pH 7.4). Two weeks after the last immunization or 6 weeks after live C. muridarum intranasal infection, six mice in each group were sacrificed to isolate splenocytes for lymphocyte multicolor flow cytometry. Four weeks after the last immunization or 8 weeks after live C. muridarum intranasal infection, the remaining eight mice in each group were challenged with live EBs for evaluation of protection.

In the second trial, groups of eight mice were immunized with 5 μg individual Chlamydia recombinant proteins formulated with DDA-MPL (250 μg DDA–25 μg MPL) in 200 μl 10 mM Tris buffer (pH 7.4), respectively. MOMP was used as a reference vaccine antigen. Four weeks after the last immunization or 8 weeks after live C. muridarum intranasal infection, mice in each group were challenged with live EBs for protection evaluation.

Genital tract infection challenge and C. muridarum quantification.

Three weeks after the last immunization or 7 weeks after live C. muridarum intranasal infection, mice were injected s.c. with 2.5 mg of medroxyprogesterone acetate (Depo-Provera; Pharmacia and Upjohn). One week after Depo-Provera treatment, mice were challenged intravaginally with 1,500 IFU of C. muridarum. Cervicovaginal washes were taken at selected dates after infection and stored at −80°C for titration on HeLa cells as previously described (4).

Multiparameter flow cytometry.

Two weeks after the last immunization, mice immunized with PmpG formulated with different adjuvants were sacrificed and splenocytes were stimulated with 2 μg/ml antibody to CD28 and the PmpG-1 protein (1 μg/ml) or HK-EBs (5 × 10⁵ IFU/ml) in complete RPMI 1640 for 4 h at 37°C. Brefeldin A was added at a final concentration of 1 μg/ml, and cells were incubated for an additional 12 h before intracellular cytokine staining. Cells were surface stained for CD3 (peridinin chlorophyll protein [PerCP]-Cy5.5), CD4 (Pacific Blue), and CD8 (allophycocyanin [APC]-Cy7) and with the viability dye, red fluorescent reactive dye (RViD) (L23102; Molecular Probes), followed by staining for IFN-γ (APC) and TNF-α (phycoerythrin [PE]-Cy7) by using a BD Cytoperm Plus fixation/permeabilization (BD Pharmingen) kit according to the manufacturer's instructions. Finally, the cells were resuspended in a 4% formaldehyde solution. All antibodies and all reagents for intracellular cytokine staining were purchased from BD Pharmingen except where noted. We acquired 200,000 live lymphocytes per sample by using an Aria flow cytometer and analyzed the data by using the FlowJo software program (Tree Star).

ELISPOT assay.

The IFN-γ enzyme-linked immunosorbent spot (ELISPOT) assay was performed as described previously (18). Briefly, 96-well MultiScreen-HA filtration plates (Millipore) were coated overnight at 4°C with 2 μg/ml of murine IFN-γ-specific monoclonal antibody (clone R4-6A2; BD Pharmingen). Spleen and draining lymph node cells from mice 3 weeks following genital infection with 1,500 IFU C. muridarum were pooled in complete RPMI 1640 medium (Sigma-Aldrich) and added to the coated plates at 10⁶ cells per well in the presence of individual Chlamydia protein (1 μg/ml) or HK-EBs (5 × 10⁵ IFU/ml). After 20 h of incubation at 37°C and 5% CO₂, the plates were washed and then incubated with biotinylated murine IFN-γ-specific monoclonal antibodies (clone XMG1.2; BD Pharmingen) at 2 μg/ml. This was followed by incubation with streptavidin-alkaline phosphatase (BD Pharmingen) at a 1:1,000 dilution. The spots were visualized with a substrate consisting of 5-bromo-4-chloro-3-indolyl phosphate and Nitro Blue Tetrazolium (Sigma-Aldrich).

Statistical analysis.

All data were analyzed with the aid of the GraphPad Prism software program. The Kruskal-Wallis test was performed to analyze data for IFU (Chlamydia shedding) from multiple groups, and the Mann-Whitney U test was used to compare medians between pairs. Comparison of cytokine production as determined by ELISPOT assay and flow cytometry between groups were analyzed using an independent, two-tailed t test. Data are presented as means ± standard errors of the means (SEM). P values of <0.05 were considered significant.

RESULTS

PmpG formulated with different adjuvants induces different levels of protection.

In order to discover a Th1-polarizing adjuvant that efficiently delivers Chlamydia antigens, we evaluated five adjuvants (DDA-MPL, DDA-TDB, DDA-MMG, Mon-CpG, and Alum) with the same antigen, PmpG. The adjuvants were chosen based on their feasibility for human use. Mice were immunized with PmpG formulated with different adjuvants, with PBS as a negative control and mice recovered from previous intranasal infection as the positive control. Four weeks after the final immunization, mice were challenged intravaginally with C. muridarum. Protection against intravaginal infection was assessed by isolation of Chlamydia from cervicovaginal wash and the determination of the number of IFU recovered from each experimental group at days 6, 13, and 20 postinfection. As shown in Fig. 1a to c, mice immunized with live EB exhibited complete protection against infection at all three time points. Mice vaccinated with PmpG formulated with the three cationic adjuvant formulations demonstrated the best protection, with DDA-MPL and DDA-TDB being somewhat superior at the earliest time point. The Mon-CpG formulation provided less-potent but still statistically significant protection. The alum group did not significantly reduce cervicovaginal shedding at any time point compared to that for the PBS group. We next tested different antigen doses, from 1 μg to 25 μg PmpG in the DDA-MPL formulation, and observed that optimal protection occurred with a 5-μg dose (Fig. 1d).

Fig 1.

Vaccine-elicited protection against Chlamydia muridarum genital tract infection in C57 mice after immunization with PmpG formulated with a variety of adjuvants. Four weeks after the final immunization, mice were challenged intravaginally with 1,500 IFU of C. muridarum. Cervicovaginal washes were taken at day 6 (a), day 13 (b), and day 20 (c) after infection, and bacterial shedding was measured on HeLa 229 cells. Mice immunized with PBS were used as a negative control, and mice infected once with 1,500 IFU of live C. muridarum intranasally were used as a positive control. ** and ***, P values of <0.01 and <0.001, respectively, in comparison to results for the PBS group. (d) Protective efficacy against C. muridarum genital tract infection in C57 mice vaccinated with different doses of PmpG formulated with DDA-MPL adjuvant.

IFN-γ⁺-producing CD4⁺ T cells that highly coexpress TNF-α correlate with protective immunity.

To explore the cellular immune mechanisms for the various levels of protection induced by the five adjuvants, we used multiparameter flow cytometry and assessed antigen-specific IFN-γ- and TNF-α-producing CD4⁺ T cells in mice immunized with PmpG in different adjuvant formulations. As shown in Fig. 2a, each vaccine group exhibited a marked difference in IFN-γ and TNF-α CD4⁺ T cell responses. DDA-MPL and DDA-TDB formulations induced the highest levels of IFN-γ- or TNF-α-producing CD4⁺ T cells. DDA-MMG and Mon-CpG showed moderate levels of IFN-γ- or TNF-α-producing CD4⁺ T cells. No or a very weak IFN-γ or TNF-α CD4⁺ T cell response was observed in the alum group (Fig. 2b). In addition, we assessed Chlamydia antigen-specific CD8⁺ T cell responses but did not detect IFN-γ- or TNF-α-producing CD8⁺ cells in any of the vaccine groups tested (data not shown). We also did not detect Chlamydia-neutralizing antibodies in any of the vaccine groups tested (data not shown). These data demonstrated a correlation between vaccine-mediated protection against Chlamydia and the magnitude of specific CD4⁺ T cell cytokine responses.

Fig 2.

C. muridarum antigen-specific cytokine responses after immunization with PmpG formulated with a variety of adjuvants. Two weeks after the final immunization, mouse splenocytes from different vaccine groups were harvested and stimulated with 1 μg/ml PmpG. IFN-γ- or TNF-α-producing CD4 T cells were analyzed by multiparameter flow cytometry as described in Materials and Methods. Each vaccine group included 6 mice. (a) Flow cytometry analysis of a representative mouse from each vaccine group. (b) Percentage of IFN-γ- or TNF-α-producing CD4 T cells from each vaccine group. (c) Fraction of IFN-γ single- and IFN-γ/TNF-α double-positive cells in total IFN-γ-producing CD4 T cells. (d) MFI for TNF-α of IFN-γ/TNF-α double-positive cells from each vaccine group.

Multiparameter flow cytometry allowed simultaneous analysis of multiple cytokines at the single-cell level. Using the Boolean combination of IFN-γ gating and TNF-α gating, frequencies of three distinct populations (IFN-γ positive/TNF-α negative [IFN-γ⁺ TNF-α⁻], IFN-γ⁻ TNF-α⁺, and IFN-γ⁺ TNF-α⁺) of CD4⁺ T cells were identified from immune splenocytes stimulated with PmpG in each vaccine group as shown in Fig. 2a. The relative percentages of different populations can define the quality of the Th1 response. Figure 2c pictorially represents by pie charts the differences in the fractions of the IFN-γ single-positive and IFN-γ⁺ TNF-α⁺ double-positive cells of the total IFN-γ-producing CD4⁺ T cells between vaccine groups. We observed that the IFN-γ⁺ TNF-α⁺ double-positive CD4⁺ T cells encompassed 87, 87, 83, 69, 0, and 0% of the total IFN-γ-producing CD4⁺ T cells in the DDA-MPL, DDA-TDB, DDA-MMG, Mon-CpG, alum, and PBS groups, respectively. The results showed a correlation between the frequency of multifunctional (IFN-γ and TNF-α-double-positive) CD4⁺ T cells and degree of protection in mice vaccinated with different adjuvants. The mean fluorescence intensity (MFI) is an additional measure of cytokine production on a single-cell basis. We further observed that among all the vaccine groups tested, the higher the frequency of IFN-γ⁺ TNF-α⁺ double-positive CD4⁺ T cells, the stronger the MFI for TNF-α among those cells (Fig. 2d). These results are consistent with previous findings using an infection model that IFN-γ⁺-producing CD4⁺ T cells coexpressing TNF-α correlate best with protective immunity (43). DDA-MPL formulation elicits the best protection, characterized by the highest frequency of IFN-γ⁺ TNF-α⁺ double-positive CD4⁺ T cells and MFI for TNF-α in those cells. Based on these findings, we chose DDA-MPL as the standard adjuvant for subsequent experiments that compare different antigens.

Determination of antigenicity of the Chlamydia proteins based on IFN-γ responses in C57 mice following C. muridarum genital tract infection.

Thirteen proteins containing MHC class II-bound peptides identified by the immunoproteomic approach were evaluated. These included 4 polymorphic membrane proteins (PmpG, PmpF, PmpE, and PmpH), one secreted protein (Tarp), 5 cytoplasmic proteins (RplF, Aasf, Ide, RecO, and AtpE), and 3 hypothetical proteins (TC0420, TC0825, and TC0285) (Table 1). To determine antigenicity in the context of natural infection, we performed an IFN-γ ELISPOT assay using pooled cells from spleens and draining lymph nodes from mice with genital C. muridarum infection. Three weeks after intravaginal inoculation of C. muridarum, mice were sacrificed and splenocytes and cells from draining lymph nodes (iliac) were pooled and stimulated in vitro with the appropriate recombinant protein. An irrelevant protein, glutathione S-transferase (GST), was used as a negative control (Ctr_neg), and heat killed EBs (HK-EB) were used as a positive control. MOMP stimulation was set up as a reference control. As shown in Fig. 3, immune cells exposed to HK-EB developed the greatest numbers of IFN-γ-secreting cells, where more than 1,000 IFN-γ-secreting cells were detected per 10⁶ cells. Cells stimulated with GST as a negative control showed nearly blank background levels, indicating that IFN-γ-secreting cells detected in the experimental system are Chlamydia antigen specific. Immune cells stimulated with individual Chlamydia protein exhibited markedly different levels of IFN-γ responses (Fig. 3). The results demonstrate that IFN-γ responses in immune cells following stimulation with the PmpG, PmpF, PmpH, RplF, Aasf, RecO, and TC0420 protein were strong; MOMP and the other antigens stimulated weak or no IFN-γ responses. Thus, 7 of the 13 proteins were determined to be antigenic based on their moderate to strong IFN-γ responses in the context of natural infection.

Table 1.

MHC class II-bound C. muridarum-derived peptides and their source proteins as identified by immunoproteomics

Chlamydia muridarum locus no.	Protein	Protein abbreviation	Purified protein format
TC0263	Polymorphic membrane protein G	PmpG	25-500-aa N-terminal His tag
TC0262	Polymorphic membrane protein E/F-2	PmpF	25-575-aa N-terminal His tag
TC0261	Polymorphic membrane protein E/F-1	PmpE	26-600-aa N-terminal His tag
TC0264	Polymorphic membrane protein H	PmpH	26-600-aa N-terminal His tag
TC0801	Ribosomal protein L6	RplF	Full-length N-terminal GST tag
TC0707	Anti-anti-sigma factor	Aasf	Full-length N-terminal GST tag
TC0190	Metalloprotease, insulinase family	Ide	26-939-aa N-terminal His tag
TC0755	DNA repair protein	RecO	Full-length N-terminal His tag
TC0741	Translocated actin-recruiting phosphoprotein	Tarp	42-608-aa N-terminal His tag
TC0584	V-type, ATP synthase subunit E	AtpE	Full-length N-terminal His tag
TC0420	Hypothetical protein	TC0420	Full-length N-terminal GST tag
TC0825	Hypothetical protein	TC0825	Full-length N-terminal His tag
TC0285	Hypothetical protein	TC0285	Full-length N-terminal His tag

Fig 3.

Antigenicity of C. muridarum T cell antigens determined by IFN-γ ELISPOT assay. C57BL/6 mice were intravaginally infected with 1,500 IFU of C. muridarum. Three weeks later, splenocytes and cells from draining lymph nodes (iliac) were pooled and stimulated in vitro for 20 h with 1 μg/ml of indicated protein. The irrelevant protein GST was used as a negative control (Ctr_neg), and heat-killed EBs (HK-EB) were used as a positive control. MOMP protein stimulation was also set up as a reference antigen. The results represent the averages for duplicate wells and are expressed as means ± SEM for groups of four mice. These data are representative of two similar experiments.

Chlamydia T cell antigens singly and in combination formulated with the DDA-MPL adjuvant conferred various levels of protection against C. muridarum genital tract infection.

We next investigated the 13 Chlamydia recombinant proteins as vaccine immunogens in the Chlamydia genital infection murine model. Mice were immunized with 5 μg recombinant protein formulated with DDA-MPL. After genital challenge, we tested the Chlamydia inclusion titers in cervicovaginal washes taken at day 6 and day 13. MOMP was again set up as the reference vaccine immunogen. The results showed that the 13 proteins and MOMP conferred various levels of protection as indicated by different levels of Chlamydia shedding in cervicovaginal washes (Fig. 4). On day 6 postchallenge, vaccination with the PmpG, TC0420, PmpE, PmpF, Aasf, MOMP, TC0825, or RplF protein exhibited statistically significant decreases in bacterial shedding compared to results for the PBS-vaccinated group (Fig. 4a). On day 13 postchallenge, besides the eight groups listed above, the mice immunized with four other Chlamydia proteins (Ide, PmpH, Tarp, or TC0285) showed significantly lower Chlamydia shedding than mice who received PBS (Fig. 4b). Two of the 13 Chlamydia proteins, AtpE and RecO, did not confer significant protection. Interestingly, AtpE had been identified when DCs were pulsed with dead but not live EB (43). Vaccination with PmpG demonstrated the greatest level of protection among all antigens tested, resulting in 20 times (Fig. 4a) and 1,000 times (Fig. 4b) reduction of the median cervicovaginal Chlamydia shedding compared to that with PBS vaccination on day 6 and day 13, respectively. Notably, seven Chlamydia proteins (PmpG, TC0420, PmpE, PmpF, Aasf, TC0825, and RplF) conferred protection levels better than or similar to those with MOMP.

Fig 4.

Protective efficacy against C. muridarum genital tract infection in C57BL/6 mice vaccinated with different C. muridarum proteins formulated with DDA-MPL adjuvant. Four weeks after the final immunization, mice were challenged intravaginally with 1,500 IFU of C. muridarum. Cervicovaginal washes were taken at day 6 (a) and day 13 (b) after challenge, and bacterial shedding was measured on HeLa 229 cells. Mice immunized with PBS served as a negative control, and mice intranasally infected once with 1,500 IFU of C. muridarum were used as a positive control. A group of mice immunized with MOMP was set up as a reference. *, **, and ***, P values of <0.05, <0.01, and <0.001, respectively, in comparison to results for the PBS group.

We next evaluated protection induced by three different combinations of T cell antigens: combination 1 consisted of membrane and secreted proteins (PmpG plus PmpE plus PmpF plus PmpH plus Tarp); combination 2 consisted of cytoplasmic proteins (Aasf plus RpIF plus RecO); and combination 3 consisted of hypothetical proteins (TC0420 plus TC0825 plus TC0285). As shown in Fig. 5, all three combination groups induced levels of protection similar to those for the best individual antigen in the corresponding group. Combination 1 may show better protection than other combinations, since four out of eight vaccinated mice completely resolved their infection by day 13 (Fig. 5b).

Fig 5.

Protective efficacies against C. muridarum genital tract infection in C57BL/6 mice vaccinated with different combinations of C. muridarum proteins in comparison to vaccination with individual proteins formulated with DDA-MPL adjuvant. Comb-1, membrane/secreted proteins (PmpG plus PmpE plus PmpF plus PmpH plus Tarp); comb-2, cytoplasmic proteins (Aasf plus RpIF plus RecO); comb-3, hypothetical proteins (TC0420 plus TC0825 plus TC0285). Four weeks after the final immunization, mice were challenged intravaginally with 1,500 IFU of C. muridarum. Cervicovaginal washes were taken at day 6 (a, c, and e) and day 13 (b, d, and f) after challenge, and bacterial shedding was measured on HeLa 229 cells. Mice immunized with PBS served as a negative control, and mice intranasally infected once with 1,500 IFU of C. muridarum served as a positive control. *, **, and ***, P values of <0.05, <0.01, and <0.001, respectively, in comparison to results for the PBS group.

DISCUSSION

Recently Kari et al. (21) reported a live-attenuated plasmid-deficient C. trachomatis vaccine that prevents trachoma in nonhuman primates. The study showed that six macaques vaccinated with an attenuated strain of C. trachomatis cured of its plasmid were either solidly or partially protected after challenge with the virulent plasmid-bearing strain of C. trachomatis. No humoral or cellular immune correlate of protection was detected, although MHC genotyping of monkeys found that all three solidly protected animals shared the same M1 haplotype in one of their MHC class II alleles. In contrast, none of the partially protected monkeys had the M1 haplotype. These data support the notion that protective immunity against C. trachomatis is mediated by CD4 T cell responses and that differences in recognition of MHC class II antigens determine protection against infection. Therefore, identifying CD4 T cell antigens presented by MHC class II molecules is an essential first step in designing a molecular vaccine for Chlamydia.

Difficulty in identifying relevant T cell antigens has led to very few antigens being evaluated in Chlamydia subunit vaccine studies. Reverse vaccinology provides a systems approach to selection of candidate T cell antigens (11, 29). However, such an approach is based on prediction using bioinformatic analysis and often yields many more T cell antigen candidates than can be reasonably validated. For instance, Finco et al. recently evaluated 120 bioinformatically identified C. trachomatis vaccine candidates and identified 21 T cell antigens. Ten of these 21 T cell antigens were tested, and three were found to be protective in the murine model (13). The immunoproteomic approach that we have applied to the problem directly identifies T cell epitopes presented by MHC class II molecules from DCs infected with Chlamydia, resulting in an improvement in the positive validation rate as seen in this study. Another advantage in using immunoproteomics over bioinformatic approaches is that the identified peptides are the result of physiological processing and presentation pathways in DCs, with peptides selected for both the affinity of the MHC molecules and their frequency of presentation. Proteins containing these peptides are likely to be favored in entering the class II pathway during in vivo infection.

Using immunoproteomics, we identified 27 C. muridarum CD4 T cell antigens (41, 43; data not shown), which represent about 3% of the Chlamydia proteome. Excluding those Chlamydia proteins with significant sequence homology to human and other bacterial proteins, we selected 13 antigens for evaluation in this study. We found that among the 13 antigens, 11 proteins (PmpG, TC0420, PmpE, PmpF, Aasf, TC0825, RplF, Ide, PmpH, Tarp, and TC0285) conferred significant protection against Chlamydia genital challenge at an early (day 6) and/or late stage (day 13). Seven proteins (PmpG, TC0420, PmpE, PmpF, Aasf, TC0825, and RplF) induced protection better than or similar to that with the MOMP, a leading Chlamydia vaccine candidate (Fig. 4). Six of the 13 antigens (PmpG, TC0420, PmpE, Aasf, RplF, and PmpH) were recently identified as T cell antigens when C. trachomatis-infected murine DCs were analyzed through immunoproteomics (K. P. Karunakaran, X. Jiang, K. Moon, C. Shen, H. Yu, L. J. Foster, and R. C. Brunham, unpublished data). Importantly, all six proteins were found to confer protective immunity in this study (Fig. 4).

Molecular vaccines generally demonstrate poor immunogenicity owing to the lack of pathogen-associated molecular patterns (PAMPs) and rapid degradation in vivo. Thus, adjuvants are essential to enhance the immunogenicity of molecular antigens. Our studies showed that DDA with MPL or TDB induced the greatest degree of protective immunity among the adjuvants tested (Fig. 1). DDA is a cationic liposome that is able to enhance antigen uptake (23). The liposomal surface charge has a major influence on its adjuvant effect. Most in vivo studies demonstrate that cationic liposomes are superior vehicles to anionic and neutral liposomes and the ability of cationic liposomes to target antigens for endocytosis by APCs plays an important role in their delivery properties (9). DDA does not have a direct effect on the maturation of DCs, and the combination of DDA with MPL or TDB delivers the PAMPs to DCs, thereby potentiating immunostimulation. MPL is a derivative of LPS but more than 100 times less toxic. MPL activates TLR4 and triggers the Trif-dependent pathway (26). MPL is licensed for vaccines against human papillomavirus types 16 and 18 (Cervarix GSK) and hepatitis B virus (Fendrix GSK). TDB is a synthetic analogue of mycobacterial cord factor or trehalose dimycolate (TDM) and shows less toxicity in vivo while retaining adjuvant action (10). Recently two groups reported that the C-type lectin, Mincle, is the essential receptor for TDB that engages a TLR-independent syk-CARD9/Bcl10/Malt-1-dependent pathway (19, 30). DDA-TDB (CAF01) promotes both strong humoral and CMI responses and has been found to be a very useful adjuvant for a number of different vaccines (1, 2). This adjuvant is currently undergoing clinical trials in both a tuberculosis and HIV subunit vaccine.

A previous study showed that IFN-γ⁺-producing CD4⁺ T cells that highly coexpress TNF-α optimally correlated with protective immunity and were preferentially induced by live EB infection (43). In the present study, we found that the greatest level of protection obtained after vaccination with DDA-MPL or DDA-TDB formulation was accompanied by the highest frequency of T cells coexpressing IFN-γ and TNF-α. The frequency of multifunctional CD4⁺ T cells induced by five adjuvant formulations accurately tracked the corresponding pattern of protection against C. muridarum genital tract infection, supporting the hypothesis that IFN-γ and TNF-α-secreting CD4 T cells are the correlate of protective immunity.

In addition, we found that most of proteins that were able to stimulate immune T cells to release strong IFN-γ secretion (Fig. 3) also conferred potent protection in mice when they were used as immunogens (Fig. 4). However, for some proteins, antigenicity as measured by IFN-γ release from immune T cells was not a consistent correlate of protection. For instance, RecO stimulated quite strong IFN-γ secretion but did not provide protection at all. Moreover, PmpE was not recognized by T cells from Chlamydia-infected mice, but vaccination with PmpE engendered substantial protection, especially at early stages after challenge. These results suggest that selection of vaccine candidates based on antigenicity alone may not always be reliable.

Genome analysis of C. muridarum and C. trachomatis revealed a multigene family encoding nine polymorphic membrane proteins (Pmps), designated PmpA to PmpI. Pmps are speculated to augment pathogen-host cell interaction, thereby facilitating the infectious process of Chlamydia, and thus are considered potential vaccine candidates. Four of the nine Pmps were identified as T cell antigens via immunoproteomics and observed to be protective and immunodominant in the mouse model (Fig. 4). While phase variable, Pmps are not hampered by amino acid diversity across different C. trachomatis disease groups (pathology variants) as is the case for the MOMP (36). Taken together, the protective Chlamydia T cell antigens, especially the four Pmps discovered by immunoproteomics, are of interest in the development of a molecularly defined Chlamydia vaccine.

Because a single-component subunit vaccine may not provide protection against infection due to MHC variation or because the antigen may be phase variable (e.g., PmpG), we explored formulations of multiple Chlamydia antigens. We evaluated three combinations consisting of pooled proteins from a membrane/secreted location or the cytoplasm or of unknown function. Protection was similar for a combination and the best single antigen in that formulation. Combination 1, composed of membrane and secreted proteins, may be better than other combinations in accelerating clearance of infection. Multiple-component vaccines are likely to cover a broader range of epitopes for CD4⁺ T cell recognition among different MHC genetic backgrounds and provide cross-protection against multiple antigenic variants of C. trachomatis. Future vaccine studies should target the identification of orthologous antigens from C. trachomatis and evaluate their immunoepidemiological correlates of protection in humans.