Objectives Sexually transmitted infections (STI) and bacterial vaginosis (BV) cause female genital tract inflammation. This inflammation, which is often present in the absence of symptoms, is associated with increased susceptibility to HIV infection. We aimed to evaluate genital cytokine profiles and the degree of inflammation associated with common STIs and BV.
Methods HIV-uninfected women (n=227) were screened for BV, Chlamydia trachomatis, Neisseria gonorrhoeae, Herpes simplex virus type 2 (HSV-2), and Trichomonas vaginalis. Concentrations of 42 cytokines in cervicovaginal lavages and 13 cytokines in plasma were measured using Luminex. Changes in cytokine profiles were evaluated using Mann–Whitney U test, logistic regression and factor analysis. p Values were adjusted for multiple comparisons using a false discovery rate step-down procedure.
Results Women with chlamydia or gonorrhoea had the highest genital cytokine concentrations, with 17/42 and 14/42 cytokines upregulated compared with women with no infection, respectively. BV was associated with elevated proinflammatory cytokine concentrations, but lower chemokine and haematopoietic cytokine concentrations. HSV-2 reactivation was associated with lower levels of inflammation, while trichomoniasis did not cause significant differences in genital cytokine concentrations. Genital infections did not influence plasma cytokine concentrations. Although certain STIs, in particular chlamydia and gonorrhoea, were associated with high genital cytokine concentrations, only 19% of women with an STI/BV had clinical signs.
Conclusions Chlamydia was associated with the highest genital cytokine levels, followed by gonorrhoea, HSV-2, trichomoniasis, and BV. In regions where HIV is prevalent and STIs are managed syndromically, better STI/BV screening is urgently needed, as certain infections were found to be highly inflammatory.
- GENITAL TRACT INFECT
- BACTERIAL VAGINOSIS
- CHLAMYDIA TRACHOMATIS
- NEISSERIA GONORRHOEA
Statistics from Altmetric.com
Sexually transmitted infections (STI) and bacterial vaginosis (BV) are prevalent in South African women at high risk of HIV infection.1 ,2 The lower female genital tract is a unique, non-sterile environment that, while tolerant to allogeneic sperm, must respond rapidly to infectious agents.3 Inflammation is an immune process essential for microbial control and clearance that is initiated and sustained by cytokine production in response to pathogen recognition.4 STIs are major causes of inflammatory cytokine upregulation and immune cell recruitment to the genital mucosa.5–9 Although inflammation can play an important role in STI clearance, it may also cause destruction of infected epithelial layers, allowing STI-associated microbes to access deeper tissues.4 ,10 Relatively few women spontaneously clear an infection without treatment, and STIs are often recurrent or persistent.11 ,12 If untreated, STIs such as Chlamydia trachomatis, Neisseria gonorrhoea and Trichomonas vaginalis are associated with reproductive complications, including pelvic inflammatory disease (PID), ectopic pregnancy, miscarriage, preterm delivery and infertility.13
BV is a syndrome characterised by displacement of healthy vaginal flora, predominantly consisting of Lactobacillus species, by Gram-positive and Gram-negative bacteria.14 ,15 While BV can upregulate genital proinflammatory cytokines, downregulation of some cytokines may also occur.7 ,16 ,17 Untreated BV is also associated with sequalae including PID, preterm delivery and postpartum infection.13
STIs, BV and associated inflammation influence susceptibility to HIV infection.18–20 We have shown that C. trachomatis, N. gonorrhoeae and Mycoplasma genitalium infections and elevated cervicovaginal lavage (CVL) concentrations of interleukin (IL)-1β, IL-6, IL-8 and soluble CD40L (sCD40L) were associated with increased risk of HIV acquisition.20 Elevated cytokine concentrations in the genital tract may facilitate HIV infection by recruiting and activating HIV target cells, reducing epithelial barrier integrity and promoting HIV replication through NF-kB activation.10 ,21–24
The aims of this study were to evaluate whether common STIs or BV were associated with specific changes in genital cytokine profiles in women at high risk of HIV infection, and to compare the relative degree of cytokine production associated with each infection.
A cohort of 242 HIV-uninfected women at high risk of HIV infection was established in Durban, South Africa.20 For this study, CVLs were available from 227 women and plasma was available from 142 women. This study was approved by the University of KwaZulu-Natal and University of Cape Town Ethics Committees, and all participants provided informed consent.
STI and BV diagnosis
A gynaecological examination was performed and two vulvovaginal swabs were collected.20 Swabs were screened for C. trachomatis, N. gonorrhoeae, M. genitalium, Herpes simplex virus (HSV) and T. vaginalis by PCR. BV was diagnosed by Gram staining using Nugent's criteria (score ≥7 indicating BV; 4–6 indicating intermediate flora) and a subset of slides (n=172) was examined for the presence of fungal hyphae. Blood was collected for syphilis (Treponema pallidum) diagnosis and HSV-2 serology. ELISA was used to detect HSV-2 gG-2 IgG (HerpeSelect, Focus Diagnostics, USA). BD Macro-Vue Rapid Plasma Reagin (RPR) and haemagglutination tests (ImmuTrep TPHA, Omega Diagnostics, UK) were used to screen for T pallidum.
CVLs (10 mL sterile saline) for cytokine measurements were collected, centrifuged and supernatants stored at −80°C.25 CVLs were not collected from menstruating participants. Blood was collected by venepuncture into Acid-Citrate-Dextrose (ACD) vacutainer tubes, plasma isolated and stored at −80°C.
In an exploratory analysis, 42 cytokines were measured in CVLs from 227/242 women, and 13 cytokines were measured in plasma from 142/242 women. CVLs were prefiltered by centrifugation using 0.2 μm cellulose acetate filters (Sigma, USA). The concentrations of all 42 cytokines included in LINCOplex Human Cytokine and High Sensitivity Human Cytokine kits (LINCO Research, USA) were measured in CVLs. These included immunoregulatory cytokines (IL-10 and IL-1Ra), growth factors (vascular endothelial growth factor (VEGF), TGF-α, platelet-derived growth factor (PDGF)-AA, PDGF-AB/BB, FGF-2, epidermal growth factor (EGF)), haematopoietic cytokines (IL-9, IL-7, IL-3, GM-CSF, G-CSF, FLT3 Ligand (FLT3L)), adaptive immune mediators (sIL-2Rα, sCD40L, IL-17, IL-15, IL-13, IL-5, IL-4, IL-2, IFN-γ), pro-inflammatory cytokines and chemokines (IFN-α, RANTES/CCL5, MIP-1α/CCL3, MIP-1β/CCL4, MDC/CCL22, MCP-1/CCL2, MCP-3/CCL7, IP-10, IL-8/CXCL8, growth related oncogene (GRO) family (CXCL1-CXCL3), fractalkine, eotaxin, TNF-β, TNF-α, IL-12p70, IL-12p40, IL-6, IL-1β, IL-1α). The lower limit of detection of these kits ranged between 0.01 and 27.65 pg/mL for each of the cytokines. Data was collected using a Bio-Plex Suspension Array Reader (Bio-Rad Laboratories) and a 5 PL regression formula was used to calculate cytokine concentrations from the standard curves (BIO-Plex manager V.4). Cytokine concentrations below the lower limit of detection were reported as the mid-point between the lowest concentrations measured and zero.
Statistical analyses were performed using GraphPad Prism 5 (GraphPad Software, USA) and STATA V.10 (StataCorp, Texas, USA). Mann–Whitney U test was used for comparisons and Spearman Rank test for correlations. Logistic regression was used to assess associations between STIs/BV and log10-transformed cytokine concentrations, while adjusting for coinfections. p Values were adjusted using a false-discovery rate step-down procedure to reduce false positive results when multiple comparisons were made.26 Confirmatory factor analysis was used to explore underlying associations between cytokines.27 Factor scores, which are linear combinations of concentrations of each cytokine in a factor, weighted according to their factor loadings, were generated for further analysis. Ingenuity Pathway Analysis was used to explore biological functions of the cytokines (Ingenuity Systems). Cytokine concentrations >2-fold above the median in the no STI/BV group were considered upregulated.
A total of 227 HIV-uninfected women were included in this study to evaluate whether common STIs and BV were associated with specific changes in genital cytokine profiles in women at high risk of HIV infection (table 1).20
The median age of these women was 36 years (range 18–58), and most (95%) reported having more than one casual sexual partner within the last 3 months. More than half the women had BV (53%; 120/227; Nugent score ≥7; table 1), 13% (30/227) had intermediate flora (Nugent score 4–6), and 30% (68/227) had a STI (excluding BV and HSV-2 serology). Of those seropositive for HSV-2 (198/227), only six had detectable HSV in their genital tracts and were considered to have reactivated HSV-2 infections, none had visible ulceration. Three women had RPR titres >1:4 with a positive TPHA test, normally indicative of active syphilis but could indicate late-latent or treated syphilis. Three women were infected with M. genitalium. All women with syphilis or M. genitalium also had BV, therefore, cytokine changes in these women could not be evaluated in this study. Of women who had a STI or BV, only 27/141 (19%) had clinical signs of an infection.20
Genital cytokine concentrations were elevated in women with active STIs
Of the common STIs evaluated in this study, women who had chlamydia (10/227) or gonorrhoea (13/227) had the highest CVL cytokine concentrations, despite only 3/10 women with chlamydia and 2/13 women with gonorrhoea having clinical signs of infection (figure 1; see online supplementary figure S1). Concentrations of 17/42 cytokines, including proinflammatory cytokines (IL-1α, IL-1β, IL-6, IL-12p70, TNF-α), chemokines (IL-8, MIP-1β, RANTES), haematopoietic cytokines and growth factors (G-CSF, Flt3L, TGF-α), and adaptive mediators (IL-2, IL-4, IL-5, IL-13, IL-15, IL-17), were higher in CVLs from women with chlamydia compared with women who did not have a STI or BV (n=85), after adjusting for multiple comparisons (figure 1). After adjusting for coinfections, BV and injectable hormone contraceptive use using logistic regression, IL-1β, IL-6, TNF-α, IP-10, MDC, MIP-1α, MIP-1β, RANTES, G-CSF, EGF, FGF-2, PDGF-AA, PDGF-AB/BB, TGF-α, IL-4, IL-5, IL-13, IL-17 and IL-10 concentrations remained significantly associated with chlamydia infections.
Women infected with gonorrhoea displayed a largely overlapping cytokine profile to women with chlamydia, with 11/17 of the same cytokines elevated in women with chlamydia also elevated in women with gonorrhoea (IL-1α, IL-1β, IL-12p70, TNF-α, RANTES, G-CSF, Flt3L, IL-2, IL-5, IL-15 and IL-17). In addition, certain unique cytokines were upregulated in women with gonorrhoea (TNF-β, eotaxin and VEGF; figure 1). Since the majority of women with gonorrhoea also had a coinfection or BV (11/13), none of these cytokines remained significant after adjustment for coinfections, BV and injectable hormone contraceptive use. However, gonorrhoea was associated with high levels of inflammation overall, with a large cumulative β-coefficient (figure 2A).
Although women with active HSV-2 infections (6/227) tended to have elevated CVL concentrations of several proinflammatory cytokines, none was significantly higher than those found in women who did not have an infection, after adjusting for multiple comparisons and/or coinfections (figure 1). However, due to the small number of women who were shedding HSV-2, this analysis was underpowered to detect the changes in proinflammatory cytokine concentrations, particularly after accounting for multiple comparisons. Detection of HSV-2 antibodies in serum alone was not associated with changes in genital cytokine concentrations (data not shown).
Despite being clinically implicated in vaginitis, trichomoniasis (in the absence of BV or coinfections; 13/227) was the least inflammatory infection in this study, with none of the cytokines elevated relative to women with no STI/BV after adjusting for multiple comparisons (figure 1; see online supplementary figure S1). In a logistic regression analysis including all women with trichomoniasis (47/227), no cytokines were significantly associated with this infection after adjustment for coinfections and multiple comparisons (data not shown).
Mixed genital cytokine profile in women with BV
Certain proinflammatory cytokines (IL-1α, IL-1β and TNF-β) were significantly elevated in women who had BV in the absence of a coinfection (n=73), relative to women who had neither BV nor a STI (figure 1; see online supplementary figure S1). However, women with BV simultaneously had lower concentrations of several chemokines (IP-10, GRO, MDC and MIP-1α) and haematopoietic cytokines (IL-7 and GM-CSF), suggesting that cytokine changes associated with BV were complex. Cytokine concentrations in women with intermediate vaginal flora alone (n=18) did not differ from women who did not have a STI/BV (data not shown).
A subset of Gram-stained slides (n=172) was examined for the presence of fungal hyphae. Women with visible fungal hyphae (n=19) had lower genital concentrations of IFN-α, IL-9, IL-15 and IL-17, after adjusting for STIs and BV using logistic regression, with IL-9 remaining significantly lower after adjusting for multiple comparisons.
STIs did not influence cytokine concentrations in plasma
Plasma cytokine concentrations did not differ in women with BV, or active C. trachomatis, N. gonorrhoeae, T. vaginalis, or HSV-2 infections compared to women who did not have a STI/BV. None of the cytokines measured in plasma correlated with those measured in CVLs from the same women (data not shown).
Genital cytokines with similar biological functions clustered together
Confirmatory factor analysis was used to group cytokines according to their primary roles in the immune response in order to reduce the complexity of the dataset and explore the relationships between functional groups of cytokines and STIs/BV (see online supplementary table S1). Cytokines were grouped as proinflammatory (IL-1α, IL-1β, IL-6, IL-12p40, IL-12p70, TNF-α, TNF-β), chemokines (eotaxin, fractalkine, GRO, IL-8, IP-10, MCP-1, MCP-2, MIP-1α, MIP-1β, MDC, RANTES), haematopoeitic (FLT3L, G-CSF, GM-CSF, IL-3, IL-7, IL-9), growth factors (EGF, FGF-2, PDGF-AA, PDGF-AB/BB, TGF-α, VEGF) and adaptive mediators (IFN-γ, IL-2, IL-4, IL-5, IL-13, IL-15, IL-17, sCD40L, sIL-2Rα). Most of the cytokines (34/42) were significantly associated with the functional group to which they were assigned, as indicated by factor loadings >0.32 (which can be interpreted in a similar manner as correlations coefficients).27 This indicates that the concentrations of most cytokines within each functional classification correlated with one another, and are thus likely to be regulated by common underlying influences.
All functional cytokine groups were positively associated with chlamydia after adjusting for coinfections, while all groups except the anti-inflammatory cytokine group were positively associated with gonorrhoea (figure 2B). In support of the mixed cytokine profile observed for BV (figure 1), the proinflammatory cytokine cluster was positively associated with BV, while the chemokine cluster was inversely associated, after adjusting for coinfections and multiple comparisons.
Cytokines upregulated in women with STIs promote immune cell migration, activation, proliferation and differentiation
The Ingenuity Knowledge Base (Ingenuity Systems) was used to explore the effects that cytokine changes in response to STIs and BV may have at the cellular level in the female genital tract (table 2). The cytokines assessed in this study primarily play roles in pathways associated with cellular trafficking, death, survival, growth, proliferation, activation and differentiation. Chlamydia (in particular), gonorrhoea and HSV-2, were each associated with upregulation of the largest number of cytokines involved in these pathways, while BV and trichomoniasis were associated with few upregulated cytokines. Women with chlamydia, gonorrhoea or HSV-2 had upregulated concentrations of cytokines that mediate chemotaxis of T cells, natural killer cells, phagocytes, monocytes, dendritic cells, neutrophils, eosinophils, granulocytes and endothelial cells. Cytokines involved in cellular apoptosis, viability, activation, proliferation and differentiation were upregulated in women with chlamydia, gonorrhoea or HSV-2. These findings indicate that increases in the concentrations of these cytokines in the genital tracts of women with chlamydia and, to a lesser extent, gonorrhoea and HSV-2, would likely result in the accumulation of highly activated and differentiated immune cells.
In women at high risk of HIV infection, chlamydia and gonorrhoea were associated with broadly elevated genital cytokine concentrations compared with women who did not have a STI or BV. Although few women were shedding HSV-2, several key cytokines with inflammatory functions were upregulated in these women compared with those with no STI or BV. By contrast, BV was associated with a mixed inflammatory profile, with upregulated proinflammatory cytokine concentrations, but downregulated concentrations of chemokines and haematopoietic cytokines. While clear inflammatory cytokine signatures were found in genital secretions of women with BV and certain STIs, they did not have similar cytokine signatures in their plasma. No significant correlations were observed between plasma and genital cytokine concentrations, suggesting that inflammatory responses in the female genital tract are independent of blood cytokine concentrations. This is likely because cytokines are locally produced in response to pathogen recognition. Although certain STIs, in particular chlamydia and gonorrhoea, were associated with high concentrations of cytokine biomarkers of genital inflammation, only 19% of women who had a STI/BV had clinical signs (3/10 and 2/13 women with chlamydia and gonorrhoea, respectively). As previously shown in this cohort, women with an STI but no clinical signs had similar cytokine profiles as women with clinical signs.20
The reasons for the mixed cytokine expression profile in women with BV remain unclear, although Ryckman et al17 similarly reported, using a smaller cytokine panel, that women with BV had elevated proinflammatory IL-1α concentrations, but lower chemokine concentrations (IP-10 and MCP-1) than women with no BV. Others have demonstrated elevated genital IL-1β concentrations from women with BV compared with those with normal flora.28 ,29 W1(suppl 3) Cytokine downregulation associated with BV may reflect a state of tolerance to abnormal vaginal microbes, which may be derived from normal gut microflora, may exist as a result of coevolution of these microbes with humans.30 These results demonstrate that BV induces changes in the genital cytokine milieu distinct from those that occur in response to STIs.
T. vaginalis infections, which are associated with vaginitis, cervicitis and PID, and increased risk of HIV acquisition,18 W2 (suppl 3) were prevalent in this cohort of women (20.7%). Although several cytokines were higher in CVL from women with trichomoniasis compared with women without an infection, none was significant. A previous study similarly reported that women with chlamydia or gonorrhoea had increased endocervical CD4 T cells than women with no STI, which was not observed in women with trichomoniasis.5 These findings suggest that asymptomatic trichomoniasis, perhaps involving low pathogen loads only detectable by PCR, are not necessarily inflammatory. Trichomoniasis, identified by culture, was associated with increased IL-8 concentrations in vaginal fluid from pregnant women compared with those with no infection.W3(suppl 3)
Reddy et al8 also demonstrated that IFN-γ, TNF-α, IL-10 and IL-12 were upregulated in the genital secretions of women with chlamydia relative to uninfected women. Cohen et alW4(suppl 3) showed that IL-10 was more often detectable in genital secretions from women with gonorrhoea, chlamydia or BV, but not trichomoniasis, compared with women with no infection. Infection of cervical and vaginal cell lines with N. gonorrhoeae resulted in upregulated IL-1, IL-6 and IL-8 production.6 Contrary to these results, some studies have found no difference in cytokines in genital samples from women with chlamydiaW5(suppl 3) and gonorrhoea,W6(suppl 3) compared with uninfected women.
In this cohort, we previously reported that gonorrhoea, chlamydia and M. genitalium infections and elevated genital concentrations of IL-1β, IL-6, IL-8 and sCD40L were associated with increased HIV risk.20 Genital cytokine upregulation may increase HIV risk by recruiting/activating HIV target cells, reducing epithelial barrier integrity and activating NF-kB.21 ,22–24 Many of the upgregulated cytokines in the CVLs from women with chlamydia, gonorrhoea or HSV-2 are involved in recruitment, activation, proliferation and differentiation of HIV target cells. Increased genital concentrations of these cytokines could result in accumulation of activated/differentiated target cells for HIV infection. Women with BV are also at increased risk of HIV acquisition.19 Although chemokine downregulation in the genital tracts of these women may result in reduced immune cell recruitment, upregulation of proinflammatory cytokines may reduce the integrity of the epithelial barrier, activate HIV target cells that are already present, and promote HIV replication by activating NF-κB.
This study has several limitations: relatively few women had gonnorhoea and chlamydia, possibly limiting our conclusions. Other factors may influence cytokine concentrations in the genital tract. While we showed that women with microscopically evident genital fungal hyphae had lower IFN-α, IL-9, IL-15 and IL-17 concentrations, and the relationships between cytokines and STIs/BV were not influenced by hormone contraceptive use, we have no information on semen exposure, antibiotic use and menstrual cycle phase, which may influence cytokine concentrations. The study was cross-sectional, so long-term implications of these profiles could not be addressed. Finally, this study was largely exploratory and observational, and mechanisms for cytokine differences require further investigation before an intervention can come from these data.
Although chlamydia, gonorrhoea and M. genitalium infections were associated with increased risk of HIV infection in this cohort, most of the women who had one or more of these STIs did not have a clinical sign.20 As a result, women with these infections would likely have remained untreated in settings where STIs are managed syndromically. This study highlights the urgent need for better strategies to manage asymptomatic STIs, as infections such as chlamydia, gonorrhoea and active HSV-2 were found to be highly inflammatory.
Chlamydia was associated with the highest genital cytokine levels, followed by gonorrhoea, herpes simplex virus type 2, trichomoniasis and bacterial vaginosis (BV).
Although women with chlamydia or gonorrhoea had very high genital cytokine biomarkers of inflammation, potentially putting them at risk of HIV infection, most lacked clinical signs.
Women with BV had a mixed genital cytokine profile, with upregulated proinflammatory cytokine concentrations, but also downregulated concentrations of chemokines and haematopoietic cytokines.
Genital infections did not cause changes in plasma cytokine concentrations, and genital and plasma cytokine concentrations did not correlate.
The authors would like to acknowledge the women enrolled and the clinical members of the Acute Infection Study Team at CAPRISA for their contribution to this study.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
Handling editor Jackie A Cassell
Contributors LM performed the laboratory work, analysed the data and prepared the manuscript; KM designed and managed the cohort, analysed the data and contributed to manuscript preparation; FL analysed the data and prepared the manuscript; NNM performed some of the laboratory work, analysed the data and contributed to manuscript preparation; KR analysed the data and prepared the manuscript; HG performed some of the laboratory work, analysed the data and contributed to manuscript preparation; CW analysed the data and prepared the manuscript; LRM analysed the data and prepared the manuscript; GW performed some of the laboratory work and contributed to manuscript preparation; QAK and SAK conceptualised the cohort and prepared the manuscript; J-ASP developed the hypothesis, analysed the data and prepared the manuscript.
Funding This work was supported by grants from the Poliomyelitis Research Foundation (PRF) of South Africa and European and Developing Countries Clinical Trials Partnership (EDCTP). The cohort was supported by grants from the Comprehensive International Program of Research on AIDS (CIPRA) of the Division of AIDS (DAIDS); National Institute of Allergy and infectious Disease (NIAID); National Institutes of Health (NIH) and US Department of Health and Human Services (DHHS) [grant number U19 AI51794]. LM was supported by the PRF; South African Medical Research Council (MRC); the Carnegie Corporation; the National Research Foundation (NRF) of South Africa and the UCT Clinical Infectious Diseases Research Initiative/Wellcome Trust.
Competing interests None.
Patient consent Obtained.
Ethics approval University of KwaZulu-Natal, and University of Cape Town ethics committees.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.