Vaginal complaints including malodor and abnormal discharge are among the most common reasons for seeking gynecologic care (Anderson et al., 2004). Bacterial vaginosis (BV) and Trichomonas vaginalis (TV) are two of the most prevalent etiologies of vaginitis. The estimated prevalence of TV in US women aged 15-59 is 2.1% (Flagg et al., 2019) with 3.5 million incident cases in this age group in 2018, the most recent year for which these data are available (Lewis et al., 2021). Prevalence estimates for TV are less certain, since trichomoniasis is not a reportable infection (Lewis et al., 2021). General population prevalence of BV has been estimated to range between 23-29% globally (Peebles et al., 2019), however prevalence is highly variable by geographic region, race and ethnicity. Beyond the management of symptoms and associated healthcare costs, TV and BV are associated with a variety of adverse reproductive health outcomes even among asymptomatic individuals including: pelvic inflammatory disease and infertility (Wiringa et al., 2019; Ravel et al., 2021), preterm birth (Van Gerwen et al., 2021; Gudnadottir et al., 2022), and increased risk of HIV acquisition and transmission (Kissinger and Adamski, 2013; Mtshali et al., 2021a). BV and TV are frequently co-occurring in the vagina, where 40-60% of TV positive individuals are also BV positive (Simhan et al., 2007; Gatski et al., 2011; Huang et al., 2023). The mechanistic link between infection and the notable increased risk of adverse clinical outcomes, however, is incompletely understood. Understanding how vaginal infections influence specific cytokine and cellular immune responses may help elucidate the mechanism by which they contribute to adverse health outcomes.

The immune system and inflammation likely contribute to the effect of vaginal infections on the adverse reproductive health outcomes detailed above. Several prior studies have shown increased cytokine concentrations, especially IL-8, but also MIP3α, and sTNFr1 with TV infections in vitro (Fichorova et al., 2006; Fichorova et al., 2013) and in vivo (Simhan et al., 2007; Jarrett et al., 2015). BV has been associated with multiple elevated cytokine concentrations, but especially IL-1β (Dabee et al., 2021; Mtshali et al., 2021b). BV has been associated with increased expression of CCR5+ T cells in the genital tract in some studies (Thurman et al., 2015; Dabee et al., 2019), however this has not been consistently demonstrated in all studies (Mitchell and Marrazzo, 2014; Shannon et al., 2017). There are few studies describing the distinct cytokine and T cell profile associated with specific vaginal infectious etiologies and several studies analyze immune profile in relation to sexually transmitted infections (STIs) as a composite of multiple etiologies (Mlisana et al., 2012; Jarrett et al., 2015). The objective of this study was to characterize the independent effect of TV and BV infection on inflammatory makers in the vagina among reproductive-aged women enrolled in a cohort study in Atlanta, Georgia.

Data for this analysis come from a prospective cohort examining the impact of progestin contraceptive initiation (Etonogestrel implant, Levonorgestrel intrauterine device, depot-medroxyprogesterone acetate injection) on HIV target cells and inflammatory markers in the lower genital tract (Haddad et al., 2020). Cis-gender women interested in initiating one of the contraceptive methods were recruited from Grady Memorial Hospital clinics or fliers posted in the Atlanta area. Eligibility criteria included: age 18-45 years, normal length of three prior menstrual cycles (defined as 22 to 35 days); HIV seronegative (rapid testing via OraQuick™); not using a copper intrauterine device or hormonal contraceptive for six months prior to enrollment; medically eligible to initiate desired contraception (according to the US Centers for Disease Control and Prevention Medical Eligibility Criteria (Curtis et al., 2016)); and no signs or symptoms of cervicitis on examination (as determined by examining clinician) at the time of enrollment.

Each participant was scheduled for four study visits. The first visit was scheduled to coincide with the luteal phase of the menstrual cycle and the second visit during the follicular phase. The contraceptive method of choice was initiated at the completion of the second visit. The third visit was scheduled approximately three months after visit two, and the fourth visit was scheduled two weeks later. Pelvic exam and specimen collection (as described below) were conducted at all four visits. If the participant reported bleeding on the day of their appointment, they were rescheduled for after bleeding resolved.

At the baseline visit, a questionnaire was administered, and data were collected on participant’s demographic characteristics, prior contraception history, medication use, intravaginal practices (such as douching or application of products to the vagina), substance use, reproductive/sexual/menstrual history, and medical and STI history. At follow-up visits, a short questionnaire related to recent medication exposures and sexual history was administered. Ethical approval for this study was obtained from the Emory University Institutional Review Board and the Grady Memorial Hospital Research and Oversight Committee.

A total of 80 individuals were enrolled in the study. Of these, 20 had no CVL samples available for analysis and were excluded, leaving an analytic sample of 60 individuals, 43 of whom had follow-up data available. Participants were median age 30 years (IQR 24, 37), half (52%) reported a household income of <$10,000 per year, and most (78%) were of self-reported Black race. Other demographic characteristics are listed in Table 1 . At the baseline visit, there were 9 individuals (15%) with TV in the absence of co-infections, 22 (37%) with BV and no other infections, 23 (38%) who were negative for all vaginal infections (TV, CT, NG, BV), and 6 (10%) who had CT, NG, or a combination of infections. At follow-up, 5 (12%) had TV only (all were positive at baseline), 14 (33%) had BV (5 were incident among women negative for BV at baseline), 18 (42%) were negative for all four infections, and 6 (14%) had CT, NG, or a combination of infections. All participants with BV at the baseline visit were asymptomatic and none were prescribed antibiotics. Fifteen participants with TV, CT, NG, or a combination were prescribed antibiotics at the baseline visit. Of these, 6 had no follow-up visit, 8 had an infection at the follow-up visit, and 1 had no infection at follow-up. Individuals receiving antibiotics were included in the primary analysis. Given the small sample sizes, it was not possible to analyze new infections separately from ongoing/re-infections. See Table 1 .

In this study of reproductive-aged women in Atlanta, GA we found evidence that TV infection, but not BV, was associated with the presence of mucosal vaginal T cells in comparison to those who were negative for TV, BV, NG, and CT. Those with “other” infections (CT, NG, or a combination) also had higher T cell concentration. Neither TV nor BV were associated with a score representing increased concentration of inflammatory cytokines. The vaginal environment is complex, and there has been significant heterogeneity in reported cytokine and cellular immune profiles in the presence of BV and TV (Dabee et al., 2021; Bongiorni Galego and Tasca, 2023). The immune profile can be influenced by multiple factors including individual bacterial species, viral infections, phase of the menstrual cycle, use of hormonal contraception, vaginal practices, lifestyle factors (e.g., diet, stress, smoking), host genetics, and others (Dabee et al., 2021).

We found increased T cell concentration, which included CD4+CCR5+ cells in the presence of TV infection. This has biologic plausibility, as TV infection adheres to and induces a cytotoxic response in host epithelial cells resulting in host inflammatory responses (Mercer and Johnson, 2018). Increased number of CCR5+ T cells in the presence of TV could explain the epidemiologic association between TV infection and HIV acquisition (Masha et al., 2019). In mice models, TV infection has been shown to increase trafficking of CD4+ but not CD8+ T cells into vaginal tissue (Paintlia et al., 2002; Smith and Garber, 2015). In a cross-sectional study of 65 women in Chicago (TV prevalence 32%), TV infection was not associated with increased expression of CD4+CCR5+ or CD8+CCR5+ T cells among women who had been exposed to HIV but remained HIV-seronegative (Jarrett et al., 2015). HIV-exposed seronegative women have previously been shown to have a relative immune quiescence (Lajoie et al., 2014) and may have differential mucosal immunologic expression in response to infectious challenge. Overall, there is a lack of published literature reporting the effect of TV infection on female genital tract T cell subsets in vivo (Nemati et al., 2018).

We found no difference in cytokines with BV and TV infection. Participants with CT, NG, or a combination of infections, however, did have significantly higher cytokine score and this pattern is evident across multiple individual cytokines. Sample size did not allow us to examine if a specific infection or combination of infections was driving this difference. Prior studies have shown increases in individual cytokines with TV (especially IL-8) and BV infection (multiple cytokines, often with the exception of IP-10) (Simhan et al., 2007; Jarrett et al., 2015; Dabee et al., 2021; Mtshali et al., 2021b). There is evolving understanding of BV as a heterogenous condition with differing types and concentrations of bacterial taxa present (Lev-Sagie et al., 2022), which could result in heterogeneity in cytokine signature. To avoid multiple testing with a large panel, we utilized PCA for dimension reduction. Individual cytokine differences may have been obscured by this technique. Future studies could address cytokine signatures in relation to specific vaginal infections.

The presence of blood was associated with both cellular and cytokine immune response in multivariable modeling. Presence of blood in the vaginal sample increased from 7% at baseline to 35% at follow-up despite study protocol to re-schedule clinic visits if the participant reported menses/clinically significant bleeding. This is most likely due to the initiation of progestin contraception, which increases risk of unscheduled bleeding (Zigler and McNicholas, 2017). The presence of blood in the vagina influences inflammation and infection acquisition risk. Menstrual bleeding increases risk of BV (Tamarelle et al., 2022) which, in turn, increases risk of TV acquisition (Sena et al., 2021). Trichomonas vaginalis infection directly damages the cervicovaginal mucosa, which could result in detection of blood in the sample (Mercer and Johnson, 2018). The presence of blood, in turn, can result in alterations in immune expression and function (Monin et al., 2020). Future research could address effects of microscopic versus macroscopic bleeding and source of bleeding (i.e., uterine, cervical, vaginal).

We utilized gram stain with Nugent scoring for BV diagnosis, which has been widely applied in research and clinical contexts (McKinnon et al., 2019). Nugent diagnosis has improved specificity in comparison to Amsel’s criteria, but symptoms of the infection do not contribute to the score (Muzny et al., 2023). All but one individual diagnosed with BV in our study would be classified as asymptomatic. Individual perception of symptoms is highly variable and many women with BV deny any symptoms. For example, in a study of nearly 3,000 women negative for CT, NG, and TV in Alabama, 58% of those with Nugent-diagnosed BV endorsed no vaginal discharge (versus 57% of women without BV) and 75% endorsed no odor (vs 82% without BV) in the prior 6 months (Klebanoff et al., 2004). Although there is controversy surrounding treatment of asymptomatic BV in a clinical context (Muzny and Schwebke, 2020), substantial evidence supports asymptomatic BV as a risk factor for the same adverse reproductive health outcomes as symptomatic BV (Myer et al., 2005; Brotman et al., 2010; Thurman and Doncel, 2011; Mlisana et al., 2012).

Strengths of this study include highly sensitive and specific diagnostic testing for STIs (nucleic acid amplification testing), longitudinal sampling, and broad assessment of immunologic markers. Importantly, the cohort under study comprises HIV negative women in Atlanta, the majority of whom identify as Black or African American; women of African descent bear disproportionate burden of HIV and STIs in the United States making this a priority population for understanding underlying drivers of increased risk for HIV/STI. Limitations include inability to assess other factors that contribute to vaginal inflammation, including the vaginal microbiome and other vaginal pathogens which have previously been shown to affect female genital tract inflammatory markers (Masson et al., 2014; Ntuli et al., 2022) (e.g., human papilloma virus, herpes simplex virus, Mycoplasma hominis, trichomonas vaginalis virus), although we did assess important covariates, including semen exposure and presence of blood. While the menstrual cycle phase has been shown to impact vaginal inflammatory markers (Hughes et al., 2022), the design of our study (pre/post initiation of progestin contraception) precluded control for cycle phase, although influence of contraception type was evaluated in model-building for multivariable analyses and was not statistically significant and therefore omitted from final models. Similarly, other factors known to affect the vaginal microbiome including intravaginal practices, smoking, presence of semen, and sexual behavior (Kwon and Lee, 2022) were assessed but not significant in final models.

Our study found increased T cell immune activation with TV infection and “other” infection (CT, NG, or a combination of infections) versus those negative for TV, BV, CT, and NG in cis-gender women in Atlanta. These findings suggest a mechanism for increased risk of HIV acquisition with TV infection. Understanding the immunology of the vaginal mucosal environment could lead to better strategies for improving sexual and reproductive health outcomes.