Differential profiles of immune mediators and in vitro HIV infectivity between endocervical and vaginal secretions from women with Chlamydia trachomatis infection: A pilot study

Abstract

Chlamydia trachomatis infection is one of the most prevalent bacterial STIs in the USA and worldwide, and women with C. trachomatis infection are at increased risk of acquiring HIV. Because immune activation at the genital mucosa facilitates HIV/SIV infection, C. trachomatis-mediated cytokine induction may contribute to increased HIV transmission in asymptomatic women. To begin to elucidate the mechanisms, we longitudinally analyzed profiles of innate immune factors and HIV infectivity in genital secretions from anatomically specific sites in asymptomatic women during C. trachomatis infection and post-antibiotic treatment. We found higher levels of cytokines and chemokines in endocervical secretions than vaginal secretions. Compared with the convalescent state, G-CSF, IL-1α, and RANTES were elevated in endocervical secretions, IFN-γ and TNF-α were elevated in vaginal secretions, and IFNγ, IL-1β, and MIP1-α were elevated in cervicolavage fluid (CVL), before adjustment of multiple comparisons. Elevated endocervical levels of IP-10 and MCP-1 were associated with the use of hormonal contraception in infected women after successful treatment, suggesting the role of hormonal contraception in inflammation independent of STIs. Importantly, soluble factors found in endocervical secretions during infection enhanced HIV infectivity while no difference in HIV infectivity was found with vaginal secretions or CVL during infection or at convalescence. Taken together, the profiles of immune mediators and in vitro HIV infectivity indicate that the endocervical and vaginal mucosa are immunologically distinct. Our results underscore the importance of considering anatomical site and local sampling methodology when measuring mucosal responses, particularly in the presence of C. trachomatis infection.

1. Introduction

Sexually transmitted infections (STIs) are known to increase the likelihood of HIV transmission (Galvin and Cohen, 2004; Plummer, 1998; Cohen et al., 1997; Chesson and Pinkerton, 2000; Mabey, 2000). Chlamydia trachomatis infection is the most frequently reported bacterial STI in the United States (Centers for Disease Control and Prevention, 2010). C. trachomatis infection can lead to pelvic inflammatory disease and infertility in women without treatment (Arora et al., 1992; Pal et al., 1998). Over 70% of women with C. trachomatis infection are asymptomatic (Brunham and Rey-Ladino, 2005); hence, the majority of C. trachomatis-infected women remain undiagnosed/untreated and are a population at increased risk of acquiring HIV.

In women, C. trachomatis primarily infects the endocervix (Brunham and Rey-Ladino, 2005), which is also a portal of entry for HIV/SIV (Li et al., 2009; Haase, 2005). C. trachomatis induces robust production of proinflammatory cytokines including IL-8, IL-6, and IL-1α in HeLa cells (an adenocarcinoma cell line from the cervix), but only a very modest induction in endocervical (A2EN cell line) and endometrial epithelial cells (HEC-1B adenocarcinoma cell line) (Rasmussen et al., 1997; Dessus-Babus et al., 2000; Buckner et al., 2011a). In C. trachomatis-positive, infertile women (aged 20–38), the levels of IFN-α, TNF-α, IL-10, and IL-12 are up-regulated in cervical secretions (Reddy et al., 2004). Specific cytokine induction can have an impact on the recruitment of lymphocytes such as CD4+ T cells and their cell surface marker expression. Indeed, increased numbers of endocervical CD4+ T cells have been found in C. trachomatis-infected women (Mittal et al., 2004; Ficarra et al., 2008; Levine et al., 1998). The endocervical cell infiltrate from C. trachomatis-infected women is dominated by activated effector memory T cells with elevated levels of CCR5 expression (Ficarra et al., 2008; Schust et al., 2012; Ibana et al., 2012), which can serve as an ideal HIV target.

Because immune activation at the genital mucosa facilitates HIV/SIV infection (Haase, 2010), C. trachomatis infection may contribute to increased HIV transmission in untreated women. In this pilot study, we analyzed profiles of innate immune factors in genital secretions from anatomically specific sites in women during C. trachomatis infection pre- and post-antibiotic treatment. Their influence on HIV infectivity was also investigated in vitro. Our results indicate that endocervical and vaginal secretions of C. trachomatis-infected young women exhibit differential cytokine/chemokine profiles and differential effects on HIV infectivity. This underscores the importance of appropriate sampling of different sites in the female genital tract to understand how HIV is acquired and transmitted, particularly in the presence of STI co-infections.

2. Materials and Methods

2.1. Study design

Institutional Review Board approvals for this study were obtained from Mount Sinai School of Medicine and Rutgers Biomedical and Health Sciences, New Jersey Medical School. Participants were women attending the Mount Sinai Adolescent Health Center, which provides C. trachomatis screening as part of routine care for all sexually active adolescents and young adults. C. trachomatis-positive women, who met the primary screening criteria as described below and who were willing to participate in the study, signed the informed consent form and were compensated.

The enrollment eligibility criteria were as follows:

1. Healthy, non-pregnant women aged 18–26;

2. Positive nucleic acid amplification screening test for C. trachomatis;

3. If postpartum, at least six weeks postpartum and at least one normal menstrual cycle since delivery;

4. No recent antibiotic treatment (within seven days of enrollment examination);

5. No known or suspected HIV-1 infection;

6. No antibiotic allergies to azithromycin (standard treatment);

7. No known or suspected immunosuppressive illnesses or conditions (e.g., history of splenectomy, history of auto-immune disorders (systemic lupus erythematosus or rheumatoid arthritis), or history of requiring chronic systemic immunosuppressive therapy;

8. No sexual intercourse for 24 h before enrollment visit and willingness to abstain from sexual intercourse for 24 h before follow-up visits, and;

9. Able to give informed consent.

If a participant did not fulfill one of the inclusion criteria after the initial session, their samples were excluded from the primary study analysis and they were not invited for subsequent research visits.

Sexually transmitted pathogens including C. trachomatis, Neisseria gonorrhoeae, Trichomonas vaginalis, HPV, and HSV2 were determined by DNA testing using BD ProbeTec^™ collection kits with the BD Viper^™ System. Samples were taken after collection of endocervical and vaginal secretions as described below. Bacterial vaginosis (BVs) was diagnosed based on the Amsel clinical criteria, which includes an elevated vaginal pH, a positive KOH “whiff test,” clue cells on light microscopy, and appearance of the vaginal discharge. Vaginal pH was measured by using S/P pH indicator strips with pH range of 3.6–6.1 (Cardinal Health, McGaw, IL, USA).

Endocervical and vaginal secretions were collected using Merocel ophthalmic sponges (#405101, Medtronic Xomed, Inc, Jacksonville, FL, USA) (Castle et al., 2004) from women with positive screening tests for C. trachomatis on enrollment and again approximately four weeks after standard antibiotic treatment (azithromycin 1 g orally in a single dose). For endocervical secretions, the sponge was placed in the cervical os for 30 s. The vaginal sample was collected by a 360° sweep of the mid-vaginal wall. Sponges were immediately immersed in 0.7 ml of Keratinocyte-SFM (Invitrogen). Cervicovaginal lavage (CVL) in 10 ml of saline was also collected. After centrifugation, supernatants were aliquoted and stored at −80 °C before analysis. To analyze pretreatment and post-treatment samples at the same time, these samples underwent one freeze–thaw cycle. Samples were analyzed within 12 months of the first specimen being collected. Protein concentrations of cervicovaginal secretions were determined by the Bradford assay (Bio-Rad, Hercules, CA, USA).

2.2. HIV infectivity assay

Pseudotyped HIV-1_JR-FL luciferase reporter viruses were produced as described previously (Connor et al., 1997; Chen et al., 1994). The effect of cervicovaginal secretions on HIV infectivity was analyzed by a single-cycle infection assay (Klotman et al., 2008). Cervicovaginal secretions did not exert any effect on proliferation of HeLa-CD4-CCR5 cells after two hour incubation as determined by MTS assay (Promega). HeLa-CD4-CCR5 cells (clone: JC48) were a gift from David Kabat. Endocervical and vaginal secretions from the swabs were diluted 1:5 with K-SFM and then incubated with R5 HIV-1_JR-FL luciferase reporter virus for one hour at 37 °C before addition to HeLa-CD4-CCR5 cells for additional two hours. CVL was not diluted before incubation with virus in the HIV infectivity assay. Cells were washed with PBS to remove unbound virus and incubated for two days before measuring luciferase activity in the target cells.

2.3. Cytokine/chemokine analysis

Cytokines and chemokines were analyzed using a Luminex 200 (Luminex, Austin, TX, USA) and Millipore multiplex kits (Billerica, MA, USA). Results were analyzed using Millipore Analyst software. Limit of detection (sensitivity) of each analyte is as follows: G-CSF (3.2 pg/ml), IFNγ (0.5 pg/ml), IL-1α (3.2 pg/ml), IL-1β (0.45 pg/ml), IL-6 (0.11 pg/ml), IL-8 (3.2 pg/ml), IL-10 (0.38 pg/ml), IP-10 (3.2 pg/ml), MCP-1 (3.2 pg/ml), MIP-1α (3.2 pg/ml), RANTES (0.3 pg/ml), TNFα (0.2 pg/ml). The choice of these cytokines and chemokines was based on our previous study (Kraus et al., 2010) and the initial screening.

2.4. Statistical analysis

Differences in pH, cytokines, chemokines, and HIV infectivity pre- and post-antibiotic treatment were examined using Wilcoxon signed-rank tests. Differences in cytokines and chemokines between contraceptive users and non-contraceptive users were examined using Wilcoxon rank-sum tests. Multiple comparisons were adjusted by false discovery rates. All statistical analyses were performed using SAS Version 9.2 (SAS Inc., Cary, NC, USA). All P values were two-tailed, with P < 0.05 considered to indicate statistical significance.

3. Results

3.1. Patient characteristics

A total of 19 C. trachomatis subjects meeting the eligibility criteria were enrolled into the study, and of these, 15 subjects were included in the final analysis. One subject was excluded because of a positive chlamydia test at the follow-up (post-treatment) visit, and the other three subjects did not return for a follow-up visit. As expected, according to the demographic characteristics of the recruitment clinic, all subjects were black or Hispanic with a mean age of 19.6 ± 1.4 (Table 1). Seven women (47%) were using hormonal contraception at the time of sample collection. The median vaginal pH at the first visit (C. trachomatis-positive; median 5.0, range 3.7–5.6) was slightly albeit not significantly higher than the pH of samples from subjects post-treatment (median 4.4, range 4.1–5.5; P = 0.074).

Table 1.

Patient characteristics.

Patients	Race/ethnicity	Age	Hormonal contraceptive	pH pre-treatment	pH post-treatment
1	Black/Hispanic	18	None	5.0	5.5
2	Black	19	None	5.0	4.4
3	White/Hispanic	19	NuvaRing	4.4	4.4
4	Black	21	NuvaRing	5.0	4.9
5	Black	18	None	5.0	4.7
6	Hispanic	18	Pill	5.3	4.4
7	Black	20	None	5.5	4.4
8	Black/White/Chinese/Caribbean	19	None	5.5	4.8
9	Hispanic	22	Pill	5.5	5.0
10	Hispanic	20	None	4.3	4.1
11	Black	19	None	4.4	4.1
12	Hispanic	20	NuvaRing	5.0	4.3
13	Black	18	NuvaRing	4.4	NA
14	Hispanic	21	NuvaRing	3.7	4.6
15	Hispanic	22	None	5.6	NA

NA: not available.

The prevalence of laboratory-diagnosed STIs or BVs in these young women at the enrollment was not as high as a previous report on a cohort of high-risk women (Mlisana et al., 2012). Of 15 C. trachomatis-positive subjects, one woman had T. vaginalis, five (33%) had HPV, two (13%) had HSV-2, and four (27%) had BVs (Table 2). None of the women was positive for N. gonorrhoeae infection. Note that azithromycin regimen was prescribed for treating C. trachomatis infection and would not have an effect on BVs (Schwebke and Desmond, 2007).

Table 2.

Concurrent sexually transmitted infections and bacterial vaginosis in patients before treatment.

Patients	CT	Previous CT history	GC	Trich	BV	HPV	HSV2
1	+	No	−	−	−	−	−
2	+	No	−	−	−	−	−
3	+	5 months	−	−	−	−	−
4	+	No	−	−	−	−	−
5	+	22 months	−	−	+	−	−
6	+	No	−	+	−	−	+
7	+	13 months	−	−	+	−	−
8	+	No	−	−	−	−	−
9	+	No	−	−	+	−	−
10	+	No	−	−	−	−	−
11	+	No	−	−	−	+	−
12	+	No	−	−	+	+	−
13	+	No	−	−	−	+	−
14	+	24 months	−	−	−	+	+
15	+	10 months	−	−	−	+	−

Chlamydia trachomatis (CT), Neisseria gonorrhoeae (GC), Trichomonas vaginalis (Trich), bacterial vaginosis (BV), Human papillomavirus (HPV), Herpes simplex virus-2 (HSV-2).

3.2. Cytokines/chemokine profiles of cervicovaginal secretions from C. trachomatis-infected women pre- and post-antibiotic treatment

Overall, the levels of cytokines and chemokines were significantly higher in endocervical secretions than in vaginal secretions or CVL (Tables 3–5). IL-10 and RANTES were not detectable in vaginal secretions or CVL. The levels of G-CSF, IL-1α, and RANTES were significantly higher in pretreatment endocervical secretions than in post-treatment endocervical secretions (Table 3, highlighted in gray, P < 0.05). Although endocervical levels of IL-10, IL-6, IL-1β, and IP-10 were also elevated during chlamydial infection, the difference between pre- and post-treatment was not significant (Table 3). In vaginal secretions, the levels of IFN-γ and TNF-α were significantly higher in pre-treatment than post-treatment (Table 4, highlighted in gray). IFN-γ, IL-1β, and MIP-1α were significantly higher in pretreatment CVL than in post-treatment CVL, whereas TNF-α was unchanged (Table 5, highlighted in gray). Despite a pilot study with 15 subjects, multiple comparisons were adjusted by false discovery rates (Tables 3–5). After adjustment for multiple comparisons, G-CSF and IL-1α in pretreatment endocervical secretions remained significantly higher than in post-treatment endocervical secretions. However, other cytokines/chemokines in vaginal secretions and CVL were not significantly different. Taken together, these results demonstrate that C. trachomatis infection alters the expression pattern of many cytokines and chemokines in the genital milieu, and that anatomically distinct regions of the female genital tract have distinct profiles of these mediators. Thus, while C. trachomatis infection primarily infects and has a significant impact on cytokine responses in the endocervix, there appears to be a more global inflammatory consequence of infection that may also encompass the lower genital tract.

Table 3.

The levels of cytokines/chemokines in endocervical samples from women with chlamydial infection pre- and post-treatment.

Cytokines/chemokines	Pre-treatment median, (IQR), pg/ml	Post-treatment median (IQR), pg/ml	P value	Adjusted *P value
G-CSF	4516.5 (2147–8457)	2345.5 (1544–3477)	0.007	0.042
IFNγ	13.5 (3.6–33.1)	5.4 (4.0–14.0)	0.194	0.259
IL-10	16.5 (7.8–35.5)	7.9 (2.9–18.2)	0.077	0.156
IL-6	251.6 (62.7–435.6)	122.3 (75.4–280)	0.091	0.156
IL-8	1357.5 (651–4337.3)	896.2 (392–2940)	0.463	0.505
IL-1α	342.3 (93–649)	76.1 (25.4–122.4)	0.007	0.042
IL-1β	6.4 (1.6–31.7)	2.95 (1.7–3.8)	0.068	0.156
IP-10	2492.9 (322–9793.5)	980 (309–1407)	0.091	0.156
MCP-1	681.4 (165.7–1378.9)	408 (192.6–778.8)	0.583	0.583
MIP-1α	37.3 (13.6–92.2)	10.9 (3.2–40.3)	0.380	0.456
TNFα	8.5 (4.2–15.3)	4.4 (3.5–5.9)	0.194	0.259
RANTES	79.8 (26.4–182.6)	20.1 (13.1–76.9)	0.022	0.088

There were no significant differences in most cytokine and chemokine levels between hormonal contraceptive users and non-users in either pre- or post-treatment women, except for IP-10 post-treatment, which was 561.9 (408.3–1522) for users and 215.3 (130.5–571.2) for non-users, P = 0.02), and MCP-1, which was 45.4 (3.2–95.3) for users and 5.4 (3.2–14.8) for non-users, P = 0.043 (Wilcoxon rank-sum test). IQR: interquartile range.

^*P values were adjusted by false discovery rates for multiple tests. An adjusted P value < 0.05 was considered significant.

Table 5.

The levels of cytokines/chemokines in CVL from women with chlamydial infection before pre- and post-treatment.

Cytokines/chemokines	Pre-treatment median, (IQR), pg/ml	Post-treatment median, (IQR), pg/ml	P value	Adjusted *P value
G-CSF	643 (279–1223)	265 (103–440)	0.095	0.237
IFNγ	1.3 (1.2–2.3)	1.2 (0.9–1.2)	0.020	0.160
IL-6	4.1 (0.4–7.7)	0.9 (0.4–9.2)	0.638	0.709
IL-8	210 (51.6–366)	280 (92.3–582)	0.169	0.282
IL-1α	393 (83–1159)	230 (45–547)	0.454	0.567
IL-1β	8.5 (1.2–29.5)	4.3 (0.5–12.9)	0.043	0.160
IP-10	43.5 (17.4–174)	85 (22.7–247)	0.978	0.978
MCP-1	20.6 (3.6–34.0)	32.9 (10.4–78.1)	0.268	0.383
MIP-1α	16.3 (14.3–22.7)	13.0 (10.9–18.2)	0.048	0.160
TNFα	0.32 (0.2–3.62)	0.32 (0.2–0.32)	0.156	0.282

No significant differences in GCSF, IFNγ, IL-1α, IL-1β, IL-6, IL-8, IP-10, MCP-1, MIP-1α, and TNFα were found in CVL of hormonal contraceptive users and non-users in either pre- or post-treatment women. IL-10 and RANTES were not detectable in CVL. IQR: interquartile range.

^*P values were adjusted by false discovery rates for multiple tests. An adjusted P value < 0.05 was considered significant.

Table 4.

The levels of cytokines/chemokines in vaginal samples from women with chlamydial infection pre- and post-treatment.

Cytokines/chemokines	Pre-treatment median, (IQR), pg/ml	Post-treatment median, (IQR), pg/ml	P value	Adjusted *P value
G-CSF	1186 (421–10956)	499 (152–3055)	0.208	0.482
IFNγ	2.74 (1.16–8.56)	1.16 (1.16–2.22)	0.027	0.135
IL-6	12.10 (4.57–55.69)	15.29 (4.41–90.34)	0.910	0.910
IL-8	453 (317–5585)	1030 (407–7404)	0.241	0.482
IL-1α	513 (135–2473)	456 (171–1751)	0.762	0.847
IL-1β	14.52 (1.56–88.22)	8.28 (1.23–54.98)	0.169	0.482
IP-10	74.47 (21.35–1396)	233.0 (60.85–492.0)	0.720	0.847
MCP-1	103.0 (11.35–220.0)	55.4 (25.52–140.0)	0.762	0.847
MIP-1α	18.72 (13.62–48.07)	18.47 (14.93–24.46)	0.360	0.600
TNFα	2.02 (0.20–9.97)	0.32 (0.2–1.92)	0.006	0.060

IL-10 and RANTES were not detectable in vaginal specimens. No significant differences in G-CSF, IFNγ, IL-6, IL-8, IL-1α, IL-1β, IP-10, MCP-1, MIP-1α, or TNFα were found in vaginal samples of hormonal contraceptive users and non-users in either pre- or post-treatment women. IQR: interquartile range.

^*P values were adjusted by false discovery rates for multiple tests. An adjusted P value < 0.05 was considered significant.

Hormonal contraception plays an important role in modulating innate immunity in the female genital tract (Hel et al., 2010; Wira et al., 2010). Interestingly, we observed that the levels of cytokines and chemokines were generally not affected by hormonal contraceptive use in either pre- or post-treatment samples. However, in endocervical fluid in the convalescent state, hormonal contraception use significantly increased the expression of IP-10 and MCP-1 (Table 3).

3.3. The effect of cervicovaginal secretions on HIV infectivity

The impact of endocervical secretions, vaginal secretions, and CVL from C. trachomatis-positive women pre- and post-antibiotic treatment on HIV infectivity was investigated using an R5 HIV-1_JR-FL luciferase reporter as described in the Section 2. Endocervical secretions from C. trachomatis-positive women significantly enhanced HIV infectivity compared with samples from the same women post-treatment (Table 6). In contrast, antibiotic treatment did not change the effect of vaginal secretions or CVL on HIV infectivity (Table 6). These results suggest that soluble factors in endocervical secretions might contribute to the increased risk of HIV infection in C. trachomatis-infected women.

Table 6.

The effect of cervicovaginal fluid on HIV infectivity.

	Pre-treatment median (IQR), RLU	Post-treatment median (IQR), RLU	P value
Endocervical	18,166 (7088.5–90,918)	12,710.7 (1548–33,467)	0.002
Vaginal	96,049 (47,081–188,615)	115,850.5 (49,672–568,200)	0.489
CVL	37,599 (11,156–80,547)	54,980 (23,606–290,847)	0.208

The median HIV infection is expressed as relative light units, RLU (interquartile range (IQR)). The average RLU ± SD for HIV infection in PBS from four independent experiments was 12,6721 ± 47898.

4. Discussion

Studies on the vaginal transmission of SIV in the rhesus macaque model indicate that the endocervix and the transformation zone are the preferential, though not exclusive, sites for viral entry (Li et al., 2009; Haase, 2010; Miller et al., 1992). In this longitudinal study, we aimed to determine the profiles of immune mediators in endocervical and vaginal secretions from C. trachomatis-infected women and to examine the effect of these secretions on HIV infectivity in vitro. In agreement with previous reports (Dezzutti et al., 2011; Wira et al., 2005; Fichorova and Anderson, 1999; Bard et al., 2002), our results demonstrated that endocervical secretions were enriched with immune mediators compared with vaginal secretions. Additionally, enhanced HIV infectivity was found in endocervical secretions from C. trachomatis-infected women compared with post-treatment samples. These results suggest the possibility of immunological compartmentalization in the lower genital tract and indicate that endocervical secretions may serve as a valuable source for searching for inflammation biomarkers and for predicting the risk of HIV transmission in C. trachomatis-infected asymptomatic subjects. Similar findings could likely be observed in the context of other STI co-infections, particularly those that target the endocervix and warrant further investigation as we search for appropriate biomarkers for predicting HIV transmission and for developing safe and protective microbicides.

It is quite likely that the differential profile of immune mediators noted between the endocervical and vaginal secretions is because the endocervix is the preferential site of infection for C. trachomatis. Our results also suggest that mediators induced by C. trachomatis infection in the endocervix may also, however, subsequently activate and amplify a more global, but distinct, response in the lower tract. This would likely occur via the natural gravitational flow of cytokine-containing endocervical secretions that could contribute to the activation of a distinct lower tract immune response via their interaction with lower tract epithelial cells. Pretreatment levels of G-CSF, IL-1α, and RANTES in endocervical secretions, IFN-γ and TNF-α in vaginal secretions, and IFN-γ, IL-1β, and MIP-1α in CVL were higher than post-treatment levels. Although levels of these immune mediators differed as well within different regions of the lower genital tract, IFN-γ, RANTES and MIP-1α are involved in Th1 responses (Luther and Cyster, 2001). These results are consistent with a previous prospective analysis of levels of IL-2 and IL-12 in adolescents at high risk of STIs, which demonstrated Th1-associated cytokine responses during C. trachomatis infection (Wang et al., 2005). We did not include IL-2 and IL-12 in this study because of their low abundance in endocervical secretions during an initial screening (data not shown). We speculate that low levels of IL-2 and IL-12 observed in that study might be a consequence of the duration of asymptomatic chlamydial infection, patient selection criteria, and a small sample size compared with the study by Wang et al. (2005). We also speculate that the relatively low levels of IFN-γ found in endocervical pretreatment samples might also be a consequence of the young age of these patients, as higher concentrations of IFN-γ, a key cytokine in the resolution of chlamydial infection, is associated with older age (Arno et al., 1990) and with recurrent infection (Agrawal et al., 2007).

Maxion and Kelly (2002) demonstrated differential gene expression of chemokines within anatomically distinct regions of the female genital tract in mice infected with the C. trachomatis mouse pneumonitis biovar, C. muridarum. In this murine model, induction of CXCL9, IP-10, and RANTES in oviducts (upper genital tract) was much higher than that in the cervicovaginal region (lower genital tract) in response to C. muridarum. Additionally, higher numbers of CD4+ T cells were recruited to the upper genital tract than to the lower genital tract during chlamydial infection in mice (Kelly et al., 2000). In this human study, we observed elevated levels of chemokines such as RANTES and MIP-1α in endocervical secretions from C. trachomatis-positive women compared with post-treatment samples. Induction of these chemokines may contribute to the recruitment of endocervical CD4+ T cells found in women with C. trachomatis infection (Mittal et al., 2004; Ficarra et al., 2008; Levine et al., 1998). In the rhesus macaque model, in response to SIV challenges, the endocervical epithelium produces MIP-3α to recruit pDCs that produce cytokines (IFNα) and chemokines (MIP-1α and MIP-1β), leading to the migration of CCR5-expressing cells such as CD4+ T cells, which are SIV/HIV target cells (Li et al., 2009). Similarly, in asymptomatic C. trachomatis-positive women, local inflammation, soluble endocervical factors, and increased HIV target cells in the genital mucosa may facilitate HIV transmission.

We observed an increase in HIV infectivity of endocervical secretions from C. trachomatis-positive women compared with convalescent samples. This association between C. trachomatis and in vitro HIV infectivity was not found in vaginal secretions or CVL. The HIV infectivity assay in this study was designed to assess factors in the clinical samples that may modulate the early steps of HIV infection, including attachment and entry. The cytokines and chemokines in the clinical specimens were in contact with HIV target cells for only two hours during viral adsorption. Additionally, a freeze–thaw process, which cannot be avoided in a longitudinal study, may reduce the activities of cytokines and chemokines at low concentrations in clinical samples. Thus, it is unlikely that the immune mediators measured in this study had a direct impact on the early steps of HIV infection. For example, the levels of RANTES and MIP-1α, known HIV entry inhibitors by competing for CCR5 co-receptor, were elevated in pre-treatment samples; however, these two chemokines were unlikely to have any effect in the in vitro infectivity assay because of their low concentrations. It remains to be determined whether induction of a HIV-enhancing factor(s) or down-regulation of HIV inhibitory factor(s) by C. trachomatis led to an increase in HIV infectivity in pretreatment samples.

Sex hormones play an important role in innate immunity in the female genital tract (Wira et al., 2010; Hel et al., 2010). We observed that changes in endocervical IP-10 and MCP-1 expression were associated with the use of hormonal contraception in C. trachomatis-positive women post-treatment. This association was not found in vaginal secretions and CVL, despite the fact that these lower genital tract regions are responsive to sex hormones (Wira et al., 1999). Modulation of cytokines and chemokines in response to sex hormones has been primarily studied in the upper genital tract. In women after estradiol treatment, progesterone up-regulates IL-8 and MCP-1 in endometrial epithelium in the pre-receptive (day 18) and receptive (day 21) periods of the secretory phase of the menstrual cycle (Caballero-Campo et al., 2002). However, estrogen, but not progesterone, suppresses Toll-like receptor (TLR3)-mediated IL-6, IL-8, and IP-10 production in an endometrial cell line (RL95-2) following poly I-C stimulation (Lesmeister et al., 2005). Interestingly, in a new polarized endocervical A2EN epithelial cell model, neither estradiol nor progesterone/estradiol affects constitutive cytokine/chemokine production, although the combination of progesterone/estradiol blocks poly I:C-induced secretion of MIP-1β, RANTES, IP-10, TNFα, IL-6, CXCL8, and G-CSF (Buckner et al., 2011b). It is not clear whether the chemokines in endocervical secretions are produced by endocervical epithelial cells or by immune cells at this mucosal site (Wira et al., 2010). Our results demonstrate, however, that hormone contraception modulates the level of some chemokines in the endocervix.

We observed differential profiles of immune mediators and in vitro HIV infectivity between endocervical and vaginal secretions in this longitudinal pilot study. In addition, the concentrations of these immune mediators, with the exception of IL-1 and MIP-1α, were 10- to 100-fold higher in endocervical secretions than in vaginal secretions and CVL. The lower concentrations of mediators in CVL may be due to the larger volume sample collection (10 ml) in comparison to the volume for endocervical and vaginal secretions (0.7 ml). The methods of normalizing cervicovaginal specimens are still a somewhat debatable topic in the field. In this study, we controlled the volume of endocervical and vaginal samples and did not normalize the levels of immune mediators based on concentrations of total protein or human albumin because of the influx of immune cells or proteins to the endocervix that occurs during C. trachomatis infection (Agrawal et al., 2007; Ficarra et al., 2008; Persson et al., 1990). Interestingly, we observed a decrease in HIV infectivity in endocervical secretions compared with CVL or vaginal secretions regardless of the treatment (Table 6), which could be a consequence of the overall higher concentrations of immune mediators including HIV inhibitory factors in the endocervix. A recent ex vivo study indicates that higher HIV replication occurs in cervical tissues than in vaginal tissues from healthy women, possibly because of the availability of HIV target cells in different tissues (Dezzutti et al., 2013). Although the in vitro HIV infectivity assay and the ex vivo HIV infection of tissue biopsies are two distinct methods of addressing the effect of soluble factors or target cells in different tissues on HIV infection, both studies support anatomical sites having a differential impact on HIV infection. Taken together, it is important to consider anatomical sites and local sampling methodology in the interpretation of mucosal responses. Additionally, endocervical specimens should be included in studies that investigate genital mucosal responses involved in HIV spread in the setting of C. trachomatis, and other endocervical co-infections.

A recent large-scale study demonstrated that asymptomatic laboratory-diagnosed STIs, but not symptomatic STI diagnosis (vaginal discharge) were associated with an increased susceptibility to HIV acquisition (Mlisana et al., 2012), highlighting the importance of identifying asymptomatic STIs and treatment for HIV prevention. Mlisana et al. (2012) have shown a weak correlation between elevated levels of inflammatory cytokines IL-1β, IL-6, IL-8, and sCD40L in CVL and an increased risk of HIV infection. Similarly, we found an increase in immune mediators in female genital secretions and enhancement of in vitro HIV infectivity in asymptomatic C. trachomatis-positive women that decreased after antibiotic treatment. Taken together, these results suggest that the effective control of STIs might play an important role in the success of HIV prevention.