Viral-vectored boosting of OmcB- or CPAF-specific T-cell responses fail to enhance protection from Chlamydia muridarum in infection-immune mice and elicits a non-protective CD8-dominant response in naïve mice

Abstract

A vaccine is needed to combat the Chlamydia epidemic. Replication-deficient viral vectors are safe and induce antigen-specific T-cell memory. We tested the ability of intramuscular immunization with modified vaccinia Ankara (MVA) virus or chimpanzee adenovirus (ChAd) expressing chlamydial outer membrane protein (OmcB) or the secreted protein, chlamydial protease-like activating factor (CPAF), to enhance T-cell immunity and protection in mice previously infected with plasmid-deficient Chlamydia muridarum CM972 and elicit protection in naïve mice. MVA.OmcB or MVA.CPAF increased antigen-specific T cells in CM972-immune mice ~150 and 50-fold, respectively, but failed to improve bacterial clearance. ChAd.OmcB/MVA.OmcB prime-boost immunization of naïve mice elicited a cluster of differentiation (CD) 8-dominant T-cell response dominated by cluster of differentiation (CD)8 T cells that failed to protect. ChAd.CPAF/ChAd.CPAF prime-boost also induced a CD8-dominant response with a marginal reduction in burden. Challenge of ChAd.CPAF-immunized mice genetically deficient in CD4 or CD8 T cells showed that protection was entirely CD4-dependent. CD4-deficient mice had prolonged infection, whereas CD8-deficient mice had higher frequencies of CPAF-specific CD4 T cells, earlier clearance, and reduced burden than wild-type controls. These data reinforce the essential nature of the CD4 T-cell response in protection from chlamydial genital infection in mice and the need for vaccine platforms that drive CD4-dominant responses.

INTRODUCTION

Chlamydia trachomatis is the leading sexually transmitted bacterium in the world¹. Ascension of infection from the cervix to the uterus and fallopian tubes in women can result in pelvic inflammatory disease² and reproductive sequelae. Despite screening programs, the rates of chlamydial infection continue to rise³, demonstrating the necessity for a vaccine to combat the ongoing epidemic.

Primary chlamydial infection in rodent models induces partial immunity, resulting in abbreviated challenge infection with lower bacterial burden⁴, and concordant data have been reported from human studies⁵. Interferon-γ (IFNγ)-producing CD4 T cells are essential for resolving infection and resisting reinfection^6,7. CD8 T cells can contribute to resolution through the production of IFNγ^8,9 and antibodies can augment protection¹⁰.

The recombinant replication-deficient viral vectors chimpanzee adenovirus (ChAd) and modified vaccinia Ankara (MVA) have been given to thousands of people, with an excellent safety profile^11–15. Clinical testing with ChAd and MVA vectors alone or as heterologous ChAd/MVA prime-boost regimens have consistently induced high frequencies of cluster of differentiation (CD)4 and CD8 T-cell responses^11,12 and have afforded protection against a range of infectious diseases in humans and challenge models^12,15,16. Heterologous boosting of two HIV vaccine candidates with MVA was safe in human volunteers and led to increased HIV-specific CD4 and CD8 T cells¹⁷. Immune readouts from a single dose of ChAdOx1 nCoV-19 (AZD1222) vaccine in a phase 1/2 trial were characterized by induction of IFNγ and tumor necrosis factor-α (TNFα) secretion by CD4 T cells, with coincident induction of polyfunctional and cytotoxic CD8 T cells¹⁸. The consistent immunogenicity of these ChAd and MVA vectors against multiple pathogens, including the intracellular bacterial pathogen, Mycobacterium tuberculosis^15,19,20, led us to hypothesize that use of these vectors for expression of chlamydial antigens would drive protection.

Humans are commonly reinfected with Chlamydia, demonstrating the low level of natural immunity induced by genital infection, but short-lived sterilizing immunity can develop after chronic exposure^21,22. Because human chlamydial vaccine efficacy studies will require vaccination of persons at risk for sexual exposure to C. trachomatis, many enrollees will have a level of preexisting immunity. The use of attenuated Chlamydia muridarum (CM) strains can mimic this scenario in mice. Intravaginal inoculation of female mice with attenuated, plasmid-cured CM (strain CM972) results in a 2-log₁₀ reduction in burden and prevents oviduct pathology after challenge with wild-type CM. This protection is associated with induction of a robust polyfunctional CD4 T-cell response^23,24. Owing to the successful use of MVA as a heterologous vaccine booster²⁵, we examined its ability to enhance immune responses and further reduce burden in CM972-immune mice. In addition, we examined the ability of MVA, ChAd, and ChAd/MVA as antigen delivery platforms in naïve mice. We used the intramuscular (IM) route because this has been shown to be safe and induce protection against mucosal pathogens in humans^13,26–30.

We chose two immunogenic proteins for testing, chlamydial outer membrane complex protein B (OmcB), a highly expressed and conserved adhesin³¹, and the highly abundant and secreted protein, chlamydial protease-like activating factor (CPAF)^32,33. Although viral vectors expressing these antigens significantly boosted peripheral T-cell responses in CM972-immune mice, they failed to enhance protection. Viral vector delivery of OmcB or CPAF in naïve mice induced a robust IFNγ-producing CD8 T-cell response that failed to reduce burden upon challenge when OmcB was the antigen but led to a marginal but statistically significant reduction in burden when CPAF was the antigen. The examination of challenge infections in vaccinated wild-type mice and mice deficient in CD4 or CD8 T cells revealed that burden reduction was completely dependent on the CD4 T-cell response. These data reinforce the essential nature of using chlamydial vaccine platforms that generate high frequencies of antigen-specific CD4 T cells. However, these vectors have potential utility for boosting preexisting anti-chlamydial CD4 T-cell immunity in humans based on their ability to boost T-cell responses in CM972-immune mice.

RESULTS

Chlamydia infection induces OmcB-specific T-cell responses

Splenocytes harvested from BALB/c and C57BL/6 mice that had previously cleared a genital tract infection with the virulent CM001 strain were examined for IFNγ-producing T cells by ELISpot and were detected at similar levels in both murine strains (Fig. 1A). Epitope mapping performed using splenocytes from infection-immune C57BL/6 mice revealed the immunodominant epitope OmcB_506–523 (SKETVEFSVTLKAVSAGD) (Figs. 1B and 1C).

Fig. 1.

OmcB is immunogenic in mice and vaccination with MVA.OmcB boosts the T-cell response primed by natural infection with live-attenuated CM (CM972). (A) Splenocytes from female BALB/c (n = 4) and C57BL/6 mice (n = 7) were analyzed for responses to a pool of OmcB OLPs 3 months after intravaginal C. muridarum (CM001) infection. (B) Peptide mapping with immune C57BL/6 splenocytes using whole length and quarter-length OmcB fragments (left) revealed the immunodominant epitope OmcB_506–523 (SKETVEFSVTLKAVSAGD) (right). Results are expressed as SFUs per million splenocytes. Mean with SD is depicted. (C) Timeline for CM972 infection and MVA. OmcB immunization. (D) Splenocytes from C57BL/6 mice (n = 5/group) that received CM972, MVA.OmcB, or CM972/MVA.OmcB were analyzed for OmcB-specific T-cell responses by ELISpot. Results are expressed as SFUs per million splenocytes. Mean with SD is depicted. Statistical significance determined by one-way ANOVA. *p < 0.05, **p < 0.01, ***p < 0.0001 (E) Gating strategy and (F) representative splenic CD4 and CD8 T-cell responses to the OmcB overlapping peptide pool and individual OmcB_506–523 peptide from CM972/MVA.OmcB immunized mice were determined by ICS. ANOVA = analysis of variance; CD = cluster of differentiation; CM = Chlamydia muridarum; ELISpot = enzyme-linked immunosorbent spot; ICS = intracellular cytokine staining; MVA = modified vaccinia virus Ankara; OLP = overlapping peptide; OmcB = outer membrane complex protein B; SD = standard deviation; SFU = spot forming unit.

MVA.OmcB immunization boosts the OmcB-specific T cells induced by genital infection but fails to enhance clearance

We tested the ability of MVA.OmcB to boost T-cell responses in mice previously immunized intravaginally with CM972 (Figs. 1C and 1D). OmcB-specific IFNγ-producing T cells were not detectable in CM972-immune mice but were detected in mice that received a single intramuscular (IM) immunization with MVA. OmcB and were significantly increased in mice that received intravaginal CM972, followed by an IM MVA.OmcB booster (Fig. 1D). Intracellular cytokine staining of splenocytes detected OmcB-specific IFNγ+TNFα+ T cells that were mostly (75%) CD4-biased. T cells targeting OmcB_506–523 were completely CD4 T cell–biased (Figs. 1E and 1F).

Because we observed that one MVA.OmcB booster dose enhanced OmcB-specific T helper (Th)1 cells, we next tested the ability of two IM MVA.OmcB boosters to further enhance this response and protect from a virulent CM001 challenge infection in mice primed intravaginally with attenuated CM972 (Fig. 2A). Although serial homologous MVA.OmcB vaccination significantly increased frequencies of OmcB-specific T cells compared with CM972 alone or CM972, followed by one MVA.OmcB boost (Fig. 2B), neither the duration of infection or chlamydial burden were significantly reduced. At best, a marginal and transient reduction in bacterial burden was observed compared with the other CM972/MVA.OmcB boosted groups (Fig. 2C).

Fig. 2.

Two doses of MVA.OmcB significantly boosts the T-cell response primed by CM972 infection but does not significantly improve protection after challenge. (A) Timeline for CM972 infection, followed by MVA. OmcB booster immunization. (B) Splenocytes from C57BL/6 mice (n = 3–5/group) that received CM972 and/or MVA. OmcB immunizations were analyzed for OmcB-specific T-cell responses by ELISpot. Results are expressed as SFUs per million splenocytes. Mean with SD is depicted. Statistical significance determined by one-way ANOVA. ***p < 0.001, ****p < 0.0001. (C) Course of infection in immunized mice and controls (n = 5/group). Error bars represent SEM. Statistical significance determined by two-way RM ANOVA with a post hoc Tukey test. p < 0.0001 for naive or MVA.OmcB + MVA.OmcB versus CM972 primed groups; p = 0.409 for CM972 + MVA.OmcB + MVA.OmcB versus CM972 + MVA.OmcB + MVA.GFP; p = 0.876 for CM972 + MVA. OmcB + MVA.OmcB versus CM972 + MVA.GFP + MVA.OmcB; p = 0.214 for CM972 + MVA.OmcB + MVA.OmcB versus CM972 + MVA.GFP + MVA. GFP. ANOVA = analysis of variance; CM = Chlamydia muridarum; ELISpot = enzyme-linked immunosorben spot; GFP = green florescence protein; MVA = modified vaccinia virus Ankara; OmcB = outer membrane complex protein B; RM = repeated measure; SD = standard deviation; SEM = standard error of the mean; SFU = spot forming unit.

We also tested the ability of an IM prime-boost regimen with MVA.OmcB to induce OmcB-specific T-cell responses and protect from CM001 challenge in naïve mice. Although this regimen induced OmcB-specific IFNγ-producing T cells comparable to those in mice primed with CM972 and boosted once with MVA.OmcB (Fig. 2B), the chlamydial burden and duration of infection were not reduced upon challenge compared with unimmunized mice (Fig. 2C).

Heterologous IM prime-boost with ChAd.OmcB/MVA.OmcB induces a robust antibody and CD8 T-cell response that fails to reduce burden and minimally reduces the frequency of oviduct hydrosalpinx

Groups of C57BL/6 female mice were primed IM with ChAd. OmcB and boosted 28 days later with MVA.OmcB or with vector.GFP controls. Mice were sacrificed 7 days after boost for immunogenicity studies and 42 days after challenge for analysis of oviduct pathology (Fig. 3A). Heterologous ChAdOx1.OmcB/MVA.OmcB prime-boost led to high numbers of IFNγ+ T-cell responses (mean = 2524) that were five- and 10-fold greater than those induced by either vector alone after OmcB pool stimulation and elicited the highest frequencies among all groups after stimulation with the CD4-biased OmcB_506–523 epitope (Fig. 3B). Intracellular cytokine staining of splenocytes stimulated with OmcB OLPs revealed that IFNγ and TNFα responses were dominated by CD8 T cells, whereas OmcB_506–523 peptide-specific responses were again biased to CD4 T cells (Fig. 3C). Serum antibody enzyme-linked immunosorbent assays (ELISAs) revealed that a single dose of MVA.OmcB elicited detectable titers of OmcB-specific immunoglobulin (Ig)G, with subtype-specific antibodies being mostly undetectable. In contrast, a single dose of ChAd.OmcB elicited high titers of OmcB-specific total IgG and IgG subtypes, with ChAd.OmcB/MVA.OmcB boost also inducing high IgG and IgG subtype titers to OmcB, and the highest levels of OmcB-specific IgG2b (Fig. 3D).

Fig. 3.

Heterologous prime-boost immunization with ChAd.OmcB/MVA.OmcB elicits a CD8-dominant T-cell response that marginally reduces the frequency of hydrosalpinx but does not significantly reduce chlamydial burden after challenge. (A) C57BL/6 mice were intramuscularly vaccinated with ChAd.OmcB or GFP control and boosted 28 days later with MVA.OmcB or GFP control for immunogenicity (n = 5) and challenge experiments (n = 5–10). (B) Spleens were harvested 1 week after MVA boost to measure the response to the omcB overlapping peptide pool (left) and OmcB_506–523 peptide (SKETVEFSVTLKAVSAGD) (right) by IFNγ ELISpot (n = 5/group). Results are expressed as SFUs per million splenocytes. Mean with SD is depicted. Statistical significance determined by one-way ANOVA. *p < 0.05, **p < 0.01, ****p < 0.0001. (C) Representative splenic CD4 and CD8 T-cell–specific responses to the OmcB peptide pool and OmcB_506–523 peptide from ChAd.OmcB/MVA.OmcB immunized mice by ICS. (D) OmcB-specific EPT for IgG, IgG1, IgG2b, and IgG2c from immunized mice. Mean with SD is depicted. (E) Course of CM001 infection from a representative experiment in C57BL/6 mice immunized intramuscularly with ChAd.OmcB/MVA.OmcB (n = 10) or controls (n = 5/group). Error bars represent SEM. Statistical significance determined by two-way RM ANOVA with a post hoc Tukey test. (F) Frequency of oviduct hydrosalpinx between immunized mice and negative controls (naive plus GFP controls) from two independent experiments. Mean with SD is depicted. Statistical significance determined by Fisher’s exact test, *p = 0.036. ANOVA = analysis of variance; CD = cluster of differentiation; ChAd = chimpanzee adenovirus; CM = Chlamydia muridarum; EPT = end point titer; ELISspot = enzyme-linked immunosorbent spot; GFP = green florescence protein; ICS = intracellular cytokine stainings; IFN = interferon; Ig = immunoglobulin; MVA = modified vaccinia virus Ankara; OmcB = outer membrane complex protein B; RM = repeated measure; SD = standard deviation; SEM = standard error of the mean.

Despite elicitation of low but detectable OmcB-specific CD4 T cells, high numbers of OmcB-specific CD8 T cells, and high titers of anti-OmcB-specific antibodies after heterologous immunization, the course of challenge infection with CM001 was similar to challenged naïve mice and control mice inoculated with ChAd.GFP/MVA.GFP (Fig. 3E). These responses were associated with a marginally decreased frequency of oviduct hydrosalpinx formation in mice that received heterologous OmcB vaccination compared with naïve and GFP-vaccinated control groups. (Fig. 3F). We observed a similar lack of robust protection in mice intranasally immunized with the ChAd.OmcB/MVA.OmcB regimen (Supplementary Fig. 1).

CPAF is an immunodominant CD4 T-cell antigen in CM972-immune mice

We directly compared the immunogenicity of OmcB and CPAF in C57BL/6 mice that received primary infection with attenuated CM972, followed by a challenge with CM001 (Fig. 4A), and found that CPAF elicited ~20-fold more IFNγ+ T cells than OmcB (Fig. 4B), and the CPAF-specific T-cell response was completely CD4-biased (Fig. 4C).

Fig. 4.

MVA.CPAF significantly boosts the T-cell response primed by live-attenuated CM972. (A) Timeline for infection with live-attenuated CM972 and virulent CM001 infections. (B) Splenocytes from C57BL/6 mice (n = 4/group) infected with CM972, followed by CM001, were analyzed for OmcB- and CPAF-specific IFNγ responses by ELISpot. Results are expressed as SFUs per million splenocytes. Mean with SD is depicted. Statistical significance determined by unpaired t test, ***p = 0.0003. (C) Gating strategy and frequency of CPAF-specific CD4 and CD8 T-cell responses. (D) Timeline for CM972 infection and MVA.CPAF immunization. (E) Spleens (n = 5/group) were harvested 1 week after MVA boost to measure CPAF-specific IFNγ responses by ELISpot. Mean with SD is depicted. Statistical significance determined by one-way ANOVA, ****p < 0.0001. ANOVA = analysis of variance; CD = cluster of differentiation; CM = Chlamydia muridarum; CPAF = chlamydial protease-like activating factor; ELISspot = enzyme-linked immunosorbent spot; IFN = interferon; MVA = modified vaccinia virus Ankara; OmcB = outer membrane complex protein B; SD = standard deviation; SFU = spot forming unit.

MVA.CPAF boosts CM972-indued immune responses but does not significantly improve protection

Intravaginal CM972 infection, followed by IM MVA.CPAF, led to high numbers of IFNγ+ T cells. Frequencies of CPAF-specific T cells were comparable with those induced by infection with CM972, followed by CM001 (Fig. 4E compared with Fig. 4B). The ~50-fold increases of CPAF-specific T cells produced by IM MVA.CPAF boosting (Figs. 4D and 4E) did not impact the courses of infection after challenge with CM001 or less virulent CM006 (Supplementary Fig. 2). Oviduct hydrosalpinx detection was low (0—1/group) for all mice that were primed with CM972 ± MVA.CPAF booster, demonstrating the marked level of protection induced by CM972 infection. In contrast, hydrosalpinges were detected in three of 10 oviducts from naïve mice challenged with CM001 and six of 10 oviducts in the MVA.GFP control group challenged with CM006 (Supplementary Fig. 2).

Homologous vaccination with ChAd.CPAF prime-boost elicits a robust CD8 T-cell response that minimally reduces burden upon challenge and fails to prevent oviduct pathology

ChAd.CPAF prime-boost IM immunization (Fig. 5A) induced high frequencies of IFNγ+ T cells detected specifically with CPAF OLP stimulation (Fig. 5B and Supplementary Fig. 3) and CPAF-specific IFNγ+TNFα+ T cells were dominated by CD8 T cells. (Fig. 5C) as we had observed for ChAd.OmcB/MVA.OmcB (Fig. 3C). Titers of anti-CPAF antibodies ranged from ~3 × 10³ to 1 × 10⁴ for total IgG and subtypes (Fig. 5D). Prime-boost vaccination with IM ChAd.CPAF (Fig. 6A) led to a statistically significant but marginal reduction (−0.6 log₁₀) in CM006 burden over the course of infection but no reduction in infection duration or oviduct dilatation scores compared with naïve controls (Figs. 6B and 6C). In contrast, CM972-immune mice exhibited the expected reduction in burden (−2 log₁₀), duration of infection (Fig. 6B), and absence of oviduct dilatation (Fig. 6C). This robust CM972-elicited protection was associated with detectable CPAF-specific CD4 T cells in the female genital tract that were absent in ChAd.CPAF immunized mice (Supplementary Fig. 4).

Fig. 5.

ChAd.CPAF prime-boost immunogenicity in female mice. (A) Female C57BL/6 (n = 5) mice were immunized intramuscularly with ChAd. CPAF and boosted 28 days later. (B) Spleens were harvested 1 week after boost to measure the CPAF-specific T-cell response by IFNγ ELISpot. Results expressed as SFUs per million splenocytes. Mean with SD is depicted. (C) Gating strategy (left) and splenic CD4 and CD8 T-cell–specific responses to CPAF were determined by ICS (right). Frequencies of IFNγ/TNF double-positive CD4 and CD8 T cells graphed as percentage of total CD4 or CD8 T cells. (D) CPAF-specific antibody end point titers (IgG, IgG1, IgG2b, and IgG2c) from immunized mice. Mean with SD is depicted. CD = cluster of differentiation; ChAd = chimpanzee adenovirus; CPAF = chlamydial protease-like activating factor; ELISpot = enzyme-linked immunosorbent spot; ICS = intracellular cytokine staining; IFN = interferon; Ig = immunoglobulin; SD = standard deviation; SFU = spot forming unit; TNF = tumor necrosis factor.

Fig. 6.

Protective efficacy of ChAd.CPAF/ChAd.CPAF immunization requires CD4 T cells. (A) Timeline for challenge experiment. (B) Course of CM006 infection in ChAd.CPAF immunized mice (n = 10) and controls (n = 5/group). Error bars represent SEM. Statistical significance determined by two-way RM ANOVA with a post hoc Tukey test. ChAd.CPAF/ChAd.CPAF versus naive (−0.6 log) ***p = 0.0005; CM972-immune versus ChAd.CPAF/ChAd.CPAF (−1.4 log) ****p < 0.0001; CM972-immune versus naive (−2.0 log) ****p < 0.0001. Figure is representative of two independent experiments. (C) Oviduct dilatation scores from immunized mice and controls. (D) CPAF-specific IFNγ T-cell responses from wild-type and immunodeficient C57BL/6 mice determined by ELISpot. Mean with SD depicted. Statistical significance determined by one-way ANOVA. (E) Representative cytokine-positive responses (top). Frequency of CPAF-specific CD4 (left) and CD8 (right) T-cell IFNγ responses in wild-type, CD8^−/−, and CD4^−/− mice (n = 5/group) determined by ICS. Mean with SD depicted. Significance determined by Student’s t test; ***p = 0.0002, *p = 0.021. (F) Course of CM006 infection in naive, immunized, and immunized immunodeficient mice (n = 6–8 mice/group). Error bars represent SEM. Statistical significance determined by two-way RM ANOVA with post hoc Tukey test. *p = 0.0448 for wild-type immunized versus naive (−0.29 log); ****p < 0.0001 for CD4^−/− immunized versus wild-type immunized (1.14 log); ***p = 0.0004 for CD8^−/− immunized versus wild-type immunized (−0.44 log). ****p < 0.0001 for CD8^−/− immunized versus naive (−0.74 log). ****p < 0.0001 for all other comparisons. Figure is representative of two independent experiments. (G) Representative cytokine-positive responses (top). Frequency of CPAF-specific CD4 (left) and CD8 (right) T-cell responses in wild-type and CD8^−/− mice (n = 5/group) determined by ICS. Mean with SD depicted. Significance determined by Student’s t test; *p = 0.0197, **p = 0.0015. (H) Comparison of IFNγ CD4 T-cell responses in immunized wild-type and CD8^−/− mice before and after challenge (left) and comparison of IFNγ CD8 T-cell responses in immunized wild-type mice before and after challenge (right). ANOVA = analysis of variance; CD = cluster of differentiation; ChAd = chimpanzee adenovirus; CM = Chlamydia muridarum; CPAF = chlamydial protease-like activating factor; ELISpot = enzyme-linked immunosorbent spot; ICS = intracellular cytokine staining; IFN = interferon; RM = repeated measure; SD = standard deviation; SEM = standard error of the mean.

CD4 T cells are responsible for protective effects of ChAd. CPAF vaccination against genital Chlamydia infection

We explored the immune mechanism contributing to the burden reduction observed after challenging ChAd.CPAF/Chad. CPAF immunized mice by examining mice genetically deficient for B cells, CD4, or CD8 T cells, immunized according to the same regimen for wild-type mice (Fig. 6A). Challenge of wild-type mice that received IM ChAd.CPAF/ChAd.CPAF confirmed a modest reduction in burden (−0.29 log₁₀) over the course of infection compared with naïve mice (Fig. 6F). We observed a significant decrease in burden and shortened course of infection in immunized B-cell deficient mice compared with wild-type controls that was associated with a significantly increased T-cell response (Supplementary Fig. 5). Specifically, CD4 deficiency had a major impact, with sustained high levels of infection being detected through day 40 (Fig. 6F). This occurred despite detection of high numbers of CPAF-specific IFNγ+ CD8 T cells (Figs. 6D and 6E). Although CD8 deficiency led to significantly decreased numbers of total IFNγ+ T cells, reflecting induction of a CD8-dominant response (Figs. 6D and 6E), immunized CD8-deficient mice had 4-fold higher IFNγ+ CD4 T cells compared with wild-type controls (Fig. 6E). This was associated with significantly reduced burden (−0.44 log) and shortened duration of challenge infection compared with immunized wild-type controls (Fig. 6F). The frequency of CPAF-specific IFNγ+ CD4 T cells in immunized wild-type mice was significantly higher than IFNγ+ CD8 T cells 42 days post-challenge (Fig. 6G). Further, the CD4 T-cell response was 10-fold higher after challenge compared with pre-challenge levels (Fig. 6H, left), while the CD8 T-cell response was over 10-fold lower in wild-type mice (Fig. 6H, right). We also observed that the frequency of IFNγ+ CD4 T cells was still significantly increased in immunized CD8-deficient mice post-challenge compared with wild-type controls (Fig. 6H, left). There was no significant difference in the number of hydrosalpinges between immunized wild-type mice (8 of 16) and immunized CD8^−/− mice (11 of 16) (data not shown).

DISCUSSION

Viral vectors were originally used for their ability to induce potent CD8 T-cell responses to viral antigens. However, in recent years, analyses have revealed the ability of MVA to boost both CD4 and CD8 T-cell responses^14,25 and ChAd regimens to induce robust antibody and Th1 responses¹⁸. These data led us to hypothesize that expression of immunogenic chlamydial proteins in either MVA or ChAd vectors would enhance protection induced by natural infection and protect naïve mice.

Genital tract infection with plasmid-deficient CM972 leads to non-sterilizing immunity in female mice that significantly protects from burden and oviduct pathology upon challenge with virulent CM^23,24. We successfully boosted the CD4 T-cell response induced by CM972 infection in female mice using IM MVA.OmcB (1250 IFNγ SFUs/10⁶ splenocytes after two boosters). Unfortunately, even two MVA.OmcB boosters failed to reduce burden above that induced by CM972 infection alone. Since we determined that virulent CM infection led to a >20-fold increase in CPAF-specific IFNγ+ T cells compared with OmcB, we hypothesized that boosting with MVA.CPAF might lead to enhanced protection. However, MVA.CPAF boosting failed to further reduce bacterial load after challenge with highly virulent CM001 or less virulent CM006. Possibilities for these negative results include the potential failure of IM boosting to induce CD4 T cells that traffic efficiently to the genital tract upon challenge, need for a multi-antigen booster, our use of high inoculums for challenge infections, or that the immunity induced by genital infection with CM972 is simply too robust to allow for observance of improved protection. Evidence for this last possibility is manifest in the extremely low level of CM006 burden detected in CM972-immune mice. The absence of improved protection in mice does not exclude the possibility that these vaccines might enhance protection in humans given their ability to significantly boost antigen-specific Th1 responses.

Although IM MVA.OmcB prime-boost in naïve C57BL/6 mice led to readily detectable OmcB-specific IFNγ-producing T cells, the course of infection upon challenge with CM001 was no different than challenged naïve mice and no reduction in hydrosalpinx formation was observed. Replacing CM972 with an IM ChAd.OmcB priming dose followed by an IM MVA.OmcB booster led to a CD8 dominant T-cell response that failed to protect from challenge despite coincident induction of high titers of OmcB-specific IgG1, IgG2b, and IgG2c antibodies. OmcB-specific IFNγ/TNFα -producing CD4 T cells were induced at 10-fold lower frequencies compared with CD8 T cells. Chlamydia specific CD8 T cells producing TNFα have been associated with oviduct tissue damage following natural infection³⁴, but we observed a marginal decrease in the frequency of hydrosalpinx in ChAd. OmcB/MVA.OmcB vaccinated mice. This result could potentially be explained by reduced ascension of CM to the oviducts due to enhanced antibody opsonization in conjunction with a minor induction of CD4 T cells.

Finally, we tested IM ChAd.CPAF/ChAd.CPAF, which led to the robust induction of CPAF-specific IFNγ-producing T cells (>2500 SFUs/10⁶ splenocytes) and high levels of CPAF-specific antibodies that were balanced for IgG1 and IgG2b/c. Although the response was CD8-dominant, as was observed when OmcB was used as the antigen, a marginal [−0.6 log₁₀ inclusion-forming units (IFUs)] but statistically significant decrease in burden was observed over the course of infection in mice challenged with CM006. The incidence of hydrosalpinx was similar in naïve and vaccinated mice despite the CD8 dominant response. We investigated the immune mechanism for the reduced bacterial burden using CD4- and CD8-deficient mice. Protection was enhanced in CD8-deficient mice immunized with ChAd.CPAF/ChAd.CPAF and was associated with a significant increase in CPAF-specific IFNγ+ CD4 T cells before and after challenge.

A recent paper reported the induction of IFNγ+ OmcB-C-terminal segment-specific responses from splenocytes of mice immunized intraperitoneally with recombinant OmcB-C-terminal segment combined with cytosine phosphoguanine (CpG) oligodeoxynucleotides, a Th1-inducing adjuvant. Burden reduction in challenged mice was <1 log₁₀ lower than controls, with no reduction in duration of infection³⁵. Another study used a recombinant fusion protein of OmcB (CT443) and CT521 (rl16) combined with a liposomal adjuvant CAF01 via subcutaneous injection in mice and reported a statistically significant 1.1–1.9 log₁₀ reduction in shedding of CM upon challenge. A polyfunctional CD4 T-cell response was induced, and protection was eliminated by depletion of CD4 T cells. No pathology data were reported³⁶. These papers reveal minor levels of protection can be induced with OmcB as a component immunogen if adjuvants are used that drive CD4 Th1 responses and contrast with our data where a CD8 T-cell–dominant response to OmcB failed to lead to any protection from burden in challenged mice.

The essential nature of the CD4 response for protection against chlamydial infection is reinforced by our results using ChAd.CPAF, where we observed only a 0.3–0.6 log₁₀ reduction in burden and no reduction in time to infection resolution or pathology, which was associated with high levels of CPAF-specific IFNγ-producing CD8 T cells. Furthermore, burden reduction was enhanced in ChAd.CPAF-immunized mice that were CD8-deficient. The lack of CD8 T-cell responses to CPAF in infection-immune mice suggests that CPAF is poorly presented on major histocompatibility complex class I during natural infection, which would prevent CPAF-specific CD8 T cells induced by vaccination from effectively targeting infected epithelial cells.

These data contrast with reports of protection from oviduct pathology and abbreviated infections in mice immunized intranasally with CPAF adjuvanted with interleukin-12 or CpG^37,38. Follow-up studies revealed that protection mediated by recombinant CPAF + CpG was completely dependent on CD4 T cells³⁹. Our use of the IM route for ChAd.CPAF inoculation may have contributed to the lack of protection because we were unable to detect genital tract resident CPAF-specific CD4 T cells. Mucosal vaccination has been shown to be essential for protection from chlamydial genital tract infection with some regimens⁴⁰, although systemic vaccination routes with CM native major outer membrane protein, adjuvanted with CpG-1826 plus Montanide resulted in significant protection⁴¹. Future experiments could investigate the efficacy of intranasal viral vector delivery. However, due to the dominance of the CD8 T-cell response, switching to a mucosal route of delivery is unlikely to significantly enhance efficacy.

Despite reports of IM ChAd-vectored COVID-19 vaccines leading to Th1-biased responses in humans¹⁶, our studies in naïve mice revealed a clear bias toward CD8 T cells and a lack of protection. This was despite the viral vector preparations used in our murine studies containing tissue plasminogen activator signal sequences, which have been shown to enhance immunogenicity by driving a target protein into the cellular secretion pathway with higher expression⁴². Our detection of robust IgG responses to OmcB- and CPAF-vectored regimens indicate the vectors successfully induced antibody-producing B cells, but these antibodies failed to protect in the absence of a robust CD4 T-cell response and B cells were completely dispensable for protection in ChAd.CPAF-immunized mice. The enhanced immunogenicity of CPAF compared with OmcB in infection-immune mice and the detection of minor levels of protection with CPAF regimens suggest that a heterologous vaccination regimen of ChAd.CPAF/MVA.CPAF might induce enough CD4 T cells to provide improved protection compared with the homologous ChAd.CPAF/ChAd.CPAF regimen.

The documented safety profile of these Food and Drug Administration–approved viral vector regimens in humans continues to make them attractive platforms for consideration. Unfortunately, recent data from a phase 1 trial of an intranasally delivered ChAdO×1 vaccine against COVID-19 was safe and well-tolerated but led to relatively weak and inconsistent immune responses⁴³. Improvements may be induced with the use of engineered vectors, addition of mucosal adjuvants, higher doses, or nebulization to enhance delivery to the lower airways. The immunogen, route, and mode of vaccine delivery are all important variables for consideration. Iterative testing is required to move the field forward, and much can be learned from examinations of immune responses in preclinical vaccine studies where protection fails to be achieved.

METHODS

Generation of vaccine constructs

Replication-deficient ChAd⁴⁴ vector and MVA⁴⁵ expressing the OmcB or CPAF antigens were constructed by the Viral Vector Core Facility within the Pandemic Sciences Institute at the University of Oxford. CPAF contained a serine to alanine mutation at position 491 to inactivate proteolytic activity⁴⁶.

An ampicillin-resistant plasmid was generated (GeneArt, Regensburg, Germany) containing the following: 5′ KpnI site>Kozak and tissue plasminogen activator sequence>Chlamydia antigen TC0727>Xho1 site>NotI site 3′. The sequence was codon-optimized to avoid runs of three or more nucleotides. The plasmid (pC1246) was resuspended in molecular grade water and transformed into New England Biolabs (NEB) DH5a cells and grown on lysogeny borth (LB;Vegitone) agar plate to recover a clone for MidiPrep.

For ChAd.OmcB and ChAd.CPAF generation, pC1246 was digested in parallel with the adenovirus entry vector containing the long cytomegalovirus (CMV) promoter and required size products were gel extracted after electrophoresis. The insert and backbone were ligated using T4 ligase and resulting plasmid deoxyribonucleic acid (DNA) transformed into NEB10 beta Escherichia coli and grown on LB agar. Clones were based on antibiotic selection and colony polymerase chain reaction (PCR), and a single clone was selected. LR clonase reaction was performed with ChAd destination vector and entry vector to form the shuttle vector. A single clone was selected as previously described and large-scale cultures were grown. Bacterial artificial chromosome (BAC) DNA was sequenced to ensure the correct antigen sequence before enzyme digestion for transfection of HEK392A T-Rex cells using linearized plasmid. The virus generated was titered and assayed by spectrophotometry to quantify the concentration of virus particles.

For MVA.OmcB and MVA.CPAF generation, pC1246 was digested in parallel with the MVA shuttle vector backbone and required size products were recovered by gel electrophoresis and extraction. Ligation of the backbone and insert was performed using T4 ligase and resulting plasmid DNA transformed into NEB10 beta E. coli and grown on LB agar plates. Clones were selected based on antibiotic selection and colony PCR, and a single clone was selected. Midiprep was performed to recover the required plasmid quantity and sequenced. The plasmid was recombined with MVA-mCherry stock. The cycle lysate from this recombination was harvested and used to infect Douglas Foster-1 (DF-1) cells. These cells were MoFlo^® (Beckman Coulter, Brea, CA) single-cell sorted into 96-well plates and used to culture recombinant virus by addition to DF-1 cells. The wells containing suitably infected cells were harvested and screened by PCR to confirm identity and test purity. Plate picking was performed until the culture was free of parental virus, as determined by PCR. The virus was then bulk-prepped and purified using a sucrose cushion.

Strains, cell lines, and culture conditions

CM Nigg stock (AR Nigg) was obtained from Roger Rank and has been previously described⁴⁷. Plaque-purified CM Nigg strain CM001, CM006⁴⁸, and CM972²³, derived from CM001-plasmid curing, were propagated in mycoplasma-free L929 cells and titrated by inclusion-forming units⁴⁹ using a fluorescently tagged anti-chlamydial lipopolysaccharide monoclonal antibody (Bio-Rad, Hercules, CA).

Animals, immunizations, and infections

Female C57BL/6J (strain #: 000664), B6.129S2-Ighm^tm1Cgn/J (muMT negative/; B-cell KO; strain no. 002288), B6.129S2-Cd4^tm1Mak/J (CD4 KO; strain no. 002663), B6.129S2-Cd8^tm1Mak/J (CD8 KO; strain no. 002665), or BALB/c (stock #: 000651) mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). Mice were given food and water ad libitum in an environmentally controlled pathogen-free room, with a cycle of 12 hours of light and 12 hours of darkness. Age-matched mice were used between 8 and 12 weeks of age. All experiments were approved by the Institutional Animal Care and Use Committee at the University of North Carolina. Experimental mice were randomized into different groups by distributing mice into different cages and assigning treatment in a non-blinded manner. All experiments were performed unblinded. Treatment and sampling were performed in the same order and sample sizes were based on previous studies. C57BL/6J mice were immunized intramuscularly in the musculus tibialis with 1 × 10⁸ infectious units ChAdOx1 vectors and/or 1 × 10⁷ plaque-forming units MVA vectors. Female mice were subcutaneously injected with 2.5 mg medroxyprogesterone (Depo-Provera; Upjohn, Hastings, MI) 5–7 days before infection to induce a state of anestrous. In some experiments, mice were intravaginally inoculated with 5 × 10⁵ IFU live-attenuated plasmid-deficient CM972 to induce protective immunity. Mice were intravaginally challenged with 5 × 10⁵ IFU CM006 or 1 × 10⁵ CM001 diluted in 30 μL SPG buffer 7 or 30 days after the final booster immunization. Before reinfection, mice were treated intraperitoneally with 0.3 mg of doxycycline for 5 days. Genital tract gross pathology was examined and recorded at sacrifice.

Chlamydia infection quantification

Mice were monitored for cervicovaginal shedding via endocervical swabs and IFUs were calculated, as described previously⁵⁰.

Genital tract pathology

Genital tract gross pathology, including hydrosalpinx development, was examined and recorded at sacrifice on day 42 after challenge. Genital tracts were removed en bloc, fixed in 10% buffered formalin, and embedded in paraffin. Longitudinal 4-μm sections were cut and stained with hematoxylin and eosin. Oviduct dilatation was assessed using a four-tiered semi-quantitative scoring system by a pathologist blinded to the experimental design⁵⁰.

ELISpot assay and ICS analysis

Single-cell suspensions of murine splenocytes were prepared by passing cells through 70-μM cell strainers and hypotonic lysis buffer before resuspension in complete media, as described previously⁵⁰. Splenocytes (1—2.5 × 10⁵ cells/well) were stimulated with pools of 18 amino acid peptides overlapping by 15 amino acids (Sigma-Aldrich, St. Louis, MO) spanning the CM OmcB or CPAF antigens, or with the OmcB_506–523 peptide (SKETVEFSVTLKAVSAGD) on polyvinyliden difluoride (PVDF)-membrane plates (Millipore, Burlington, MA) coated with 5μg/ml anti-mouse IFNγ (AN18). After 18–20 hours of stimulation, IFNγ spot-forming cells were detected by staining membranes with anti-mouse IFNγ biotin (1 μg/mL; R46A2), followed by streptavidin-alkaline phosphatase (1 μg/mL) and developed with 1-step NBT/BCIP substrate solution (Thermo Fisher, Waltham, MA). Spots were enumerated on an AID ELISpot reader (AID Autoimmun Diagnostika GmbH, Strasbourg, Germany).

Mouse female genital tracts were harvested and minced after fat removal. The tissue was then digested using a lamina propria gentleMACS^™ dissociation kit and gentleMACS^™ dissociator (Miltenyi Biotec, Bergisch Gladbach, Germany). Briefly, approximate 0.5-cm pieces of tissue were incubated with the digestion cocktail in a shaking incubator at 220 rpm and 37 °C for 30 minutes. After incubation, the tissue was passed through the dissociator. The dissociated sample was spun quickly, and the resulting pellet was resuspended in 50 mL of phosphate buffered saline (PBS) with 1% fetal bovine serum (FBS). The tissue suspension was passed through a 70-mm filter using a plunger to disrupt remaining tissue fragments. The filtered cells were pelleted by centrifugation and resuspended for subsequent flow cytometric analysis.

For analysis of intracellular cytokines, cells were stimulated as previously described in a round-bottom 96-well plate at a concentration of 1 × 10⁶ cells/well ± 5 μg/mL peptide(s) for 12–16 hours at 37 °C in the presence of GolgiPlug (BD Biosciences, Franklin Lakes, NJ). Control samples were stimulated with PMA/ionomycin (eBioscience, San Diego, CA) and GolgiPlug. After stimulation, cells were stained with Live/Dead-UV (Life Technologies, Carlsbad, CA) and combinations of the surface markers anti-CD45 (30-F11), anti-CD3 (17A2), anti-TCRb (H57–597), anti-CD4 (RM4–5 or GK1.5), and anti-CD8 (53–6.7) from BD Biosciences. After surface staining, cells were fixed in BD Bioscience Cytofix/Cytoperm for 20 minutes. For detection of intracellular cytokines, cells were incubated for 30 minutes in BD Bioscience Perm/Wash with anti-TNFα (MP6-XT22) and anti-IFNγ (XMG1.2) from BD Bioscience. Brilliant Stain Buffer was used when cells were stained with BD Horizon Brilliant fluorescent polymer dyes. Samples were analyzed on an LSR II flow cytometer (BD Bioscience), and data were analyzed with FlowJo version 10 software (FlowJo LLC, Ashland, OR). Fluorescence minus one (FMO) controls were used to establish precise gating for positivity. A minimum of 50,000 events were acquired for CD4 and CD8 T cells.

Antibody ELISA

OmcB- or CPAF-specific IgG, IgG1, IgG2b, and IgG2c antibodies were measured in serum of C57BL/6 mice. ELISA plates were coated overnight at 4 °C, with 10 μg/mL recombinant maltose binding protein (MBP)-CPAF or MBP-OmcB (AA 212–554) (from John Harris, QUT) diluted in 0.5 M sodium bicarbonate. Plates were washed with PBS-Tween. After blocking with 2% BSA PBS-Tween for 1 hour at 37 °C, the samples were serially diluted and incubated for 1 hour at 37 °C. Internal controls were generated using reference serum. Plates were washed with PBS-Tween. A goat anti-mouse horseradish peroxidase-conjugated secondary antibody was added (Southern Biotech, Birmingham, AL) and incubated for 1 hour at 37 °C. Goat anti-mouse IgG was added at a 1:10,000 dilution and IgG1, IgG2b, and IgG2c were added at a 1:2000 dilution. After washing, the plates were developed using TMB substrate solution (Thermo Fisher, Waltham, MA) for a maximum of 15 minutes. The reaction was stopped using 0.5 M sulfuric acid and optical density was read at 450 nm. The threshold for detection of binding antibody was defined by a value that was greater than the mean +3 standard deviations of the non-specific antibody values.

Statistical analysis

The differences between the means of experimental groups after infection were calculated using two-way repeated measures analysis of variance. Significant differences in ELISA and ELISpot data were determined by unpaired t test for comparison of two groups or one-way analysis of variance for comparing three or more groups. Comparison of oviduct hydrosalpinx frequency was determined by Fisher’s exact test. Statistical differences in oviduct histopathology were determined by the Kruskal–Wallis test. Prism software (GraphPad, La Jolla, CA) was used for statistical analyses, and p ≤ 0.05 were considered significant.