Epidemiology and Social: Concise Communications
Bacterial vaginosis, HIV serostatus and T-cell subset distribution in a cohort of East African commercial sex workers: retrospective analysis
Schellenberg, John J.a; Card, Catherine M.a; Ball, T. Blakea,c,d; Mungai, Jane Njeric; Irungu, Erastusc; Kimani, Joshuac; Jaoko, Walterc; Wachihi, Charlesc; Fowke, Keith R.a,b,d; Plummer, Francis A.a,c,d
aDepartment of Medical Microbiology
bDepartment of Community Health Sciences, Faculty of Medicine, University of Manitoba, Winnipeg, Canada
cDepartment of Medical Microbiology, University of Nairobi, Nairobi, Kenya
dNational Microbiology Laboratory, Canadian Science Centre for Human and Animal Health, Winnepeg, Canada.
Correspondence to John J. Schellenberg, Department of Microbiology, 412 Buller Bldg., University of Manitoba, R3T 2N2, Canada. Tel: +1 204 784 8074; fax: +1 204 789 2018; e-mail: firstname.lastname@example.org
Received 3 June, 2011
Revised 21 October, 2011
Accepted 3 November, 2011
Objective: Although bacterial vaginosis is a known correlate of HIV infection, no previous studies have investigated whether women defined as HIV-exposed seronegative (HESN) are less likely to have bacterial vaginosis. Little is known about the effects of bacterial vaginosis on systemic immune activation associated with HIV+ serostatus.
Design: Cohort-based retrospective analysis of bacterial vaginosis in relation to HESN status, HIV+ serostatus and peripheral T-helper cells, with cross-sectional analysis of bacterial vaginosis in relation to peripheral T-regulatory cells (Tregs).
Methods: Bacterial vaginosis diagnosis by Gram stain and determination of systemic CD4+ and CD8+ T-helper cell frequency by flow cytometry for 3504 vaginal samples from 988 commercial sex workers over 4 years. Treg phenotyping by FoxP3 staining and multiparameter flow cytometry in peripheral blood of 97 women at a single time-point.
Results: No differences in bacterial vaginosis diagnosis were observed between HESN and other HIV-negative (HIV-N) controls; however, HIV+ women were more likely to be diagnosed with bacterial vaginosis compared to all HIV-negative women (HESN/HIV-N combined). HIV+ women with bacterial vaginosis had significantly higher CD4+/CD8+ T-helper cell counts and a lower CD4/CD8 ratio, as well as fewer Tregs as a proportion of total T-helper cells, compared to bacterial vaginosis-negative women. The number of bacterial vaginosis diagnoses in this cohort has decreased significantly over time.
Conclusion: Bacterial vaginosis is associated with HIV serostatus and shifts in distribution of T-cell subsets. A concomitant reduction in bacterial vaginosis and HIV infections over time suggests that the elucidation of bacterial vaginosis–HIV interactions will be critical to further understanding of HIV pathogenesis and prevention in this high-risk group.
Resistance to HIV in a cohort of highly exposed commercial sex workers (CSWs) in Nairobi, Kenya, was first described in 1996 . Although several genetic and immune factors correlate with the HIV-exposed seronegative (HESN) phenotype, none are determinative . Mucosal studies have revealed that HESN CSWs may have increased anti-inflammatory factors in mucosal fluids [3,4], mirroring reduced immune activation in peripheral blood .
Generally lower levels of basal gene transcription and down-regulation of pro-inflammatory cytokines in T-helper cells from HESN CSWs suggest that resistant individuals may be characterized by a ‘quiescent’ immune profile . As recently demonstrated, activation of CD4+/CD8+ T cells is reduced and frequency of CD4+CD25+FoxP3+ T-regulatory cells (Tregs) is increased in HESN compared to other HIV-negative (HIV-N) individuals in this cohort . Despite contradictory observations in other groups of HESN [7,8], these findings indicate that increased Tregs may potently down-regulate activated target CD4+ T cells, reducing the likelihood of viral expansion beyond mucosal surfaces .
Since resistance to HIV infection is not absolute, environmental factors may plausibly contribute to this phenomenon. Little is known about vaginal microbiota in Africans, HIV+ women , or CSWs, therefore an evaluation of the dynamics of microbiota and their interactions with HIV at the earliest stages of infection may contribute to an overall understanding of HIV resistance in this cohort. The ‘enigmatic’ clinical entity known as bacterial vaginosis is defined by a reduction in vaginal Lactobacillus populations and overgrowth of anaerobic and Gram-negative organisms [11,12]. Frequently asymptomatic and difficult to prevent or treat, bacterial vaginosis has been associated with increased susceptibility to HIV infection in numerous studies , presumably due to microbiological changes that create a permissive physiological environment for expansion of HIV founder populations.
Although health benefits of vaginal Lactobacillus organisms are well documented , the mechanisms underlying these effects are not fully characterized. The possibility that reduced bacterial vaginosis organisms or increased vaginal Lactobacillus might contribute to immune quiescence observed in some HESN has not been addressed in previous studies. Therefore, we analyzed bacterial vaginosis diagnoses in HESN vs. HIV-N retrospectively over time. We also examined associations between bacterial vaginosis and HIV+ serostatus in the cohort, including differences in systemic T cell subsets, in order to elucidate the well known link between bacterial vaginosis and HIV.
Materials and methods
The study was conducted as part of a larger cohort study of CSWs in Nairobi, Kenya, many of whom remain HIV-seronegative (HIV–) despite years of occupational exposure to HIV-1. Blood and mid-vaginal swabs were collected by clinical staff as part of a semi-annual survey in this group, including samples from HESN (active CSWs remaining HIV– for longer than 3 years of follow-up), other HIV-negative (HIV-N; active HIV– CSWs with less than 3 years of follow-up) and HIV-seropositive (HIV+) women. All experimental procedures were approved by Research Ethics Boards at the University of Manitoba and the University of Nairobi.
Retrospective analysis of laboratory bacterial vaginosis diagnoses in 988 individuals providing a total of 3504 vaginal specimens between November 2004 and July 2008 was carried out. Characterization of Treg frequency was carried out in a cross-section of 97 individuals providing samples in July 2007. All Nairobi participant data, including length of time in cohort and current CSW activity, were derived from an extensive database of survey and laboratory data collected on the same day as biological samples.
Bacterial vaginosis was defined by Nugent scoring of bacterial morphotypes per high-powered field in Gram-stained vaginal smears , resulting in a diagnosis of bacterial vaginosis-negative [bacterial vaginosis (BV−); predominance of Lactobacillus], bacterial vaginosis-intermediate (BVI; reduced Lactobacillus and increased Gram-negative morphotypes) or bacterial vaginosis-positive (BV+; absence of Lactobacillus and predominance of Gram-negative morphotypes). In this study, women with BVI and BV+ diagnoses were combined in a single category for comparison with BV− women, as in similar studies [16,17]. Presence of leukocytes in vaginal smears was also assessed routinely. All slides were analyzed by a trained, experienced evaluator (J.N.M.).
Peripheral blood samples were centrifuged to separate plasma for HIV-1 serology, and peripheral blood mononuclear cells (PBMCs) were isolated using ficoll density gradient centrifugation. Whole blood collected for CD4+ and CD8+ T-cell counts was labeled using antibodies specific for CD4 and CD8 with the Tritest flow cytometry assay (BD Pharmingen, Mississauga, Ontario, Canada) and quantified with a FACScan flow cytometer (BD, Mississauga, Ontario, Canada).
In a subset of 97 patients attending the study clinic between June and July 2007, freshly isolated PBMCs were immunophenotyped using multicolour flow cytometry to determine Treg frequency by staining with anti-CD3 AmCyan, anti-CD4 AlexaFluor700, anti-CD8 Pacific Blue, anti-CD25 PE (BD Biosciences) and anti-FOXP3 APC (eBioscience). Surface markers were stained by incubation of cells with fluorochrome-conjugated antibodies for 30 min at 4°C. Cells were washed in phosphate-buffered saline supplemented with 2% fetal calf serum (Invitrogen). FOXP3 was detected using an intracellular staining set (eBioscience), according to manufacturer's instructions.
Data were acquired on a LSR II flow cytometer (BD Biosciences) and 100 000 events analyzed with FACSDiva software (BD Biosciences). Tregs were phenotypically defined as CD3+CD4+ T cells coexpressing CD25 and FOXP3. Treg frequency was expressed as a percentage of CD4+ T cells. However, since CD4+ T cells are the primary target of HIV-1, Tregs were also expressed as a percentage of total (CD3+) T cells when conducting comparisons of HIV-positive individuals.
For retrospective analyses, multilevel logistic regression modeling was conducted (HLM 6 software package) in order to control for participant variability in the context of repeated measures, as previously reported [18,19]. The chi-square test function in R statistical software was used to determine statistical significance of differences observed between groups at single time-points and the effect of time on sample-level variables. Mean Treg frequencies in HIV+ and HIV– individuals with BVI/BV+ vs. BV− diagnoses were compared using the nonparametric Wilcoxon rank sum test function in R statistical software.
A total of 3504 samples were collected from 988 CSWs over a 4-year period, including 1624 samples from 425 CSWs diagnosed as HIV+. Eleven women, providing a total of 51 samples, seroconverted to HIV+ during the study period. A total of 179 HIV– individuals were categorized as HESN at the end of the study period (i.e. remaining HIV– after 3 years of follow-up), providing a total of 867 samples, whereas 373 HIV– individuals providing 973 samples were categorized as HIV-N controls.
HIV serostatus, bacterial vaginosis diagnosis and mucosal leukocytes
To address the hypothesis that HESN are less likely to have bacterial vaginosis compared to HIV-N, laboratory diagnosis of bacterial vaginosis was analyzed retrospectively (Fig. 1a) over eight ‘re-surveys’, that is bi-annual survey time-points in which cohort members are scheduled for research visits in study clinics. No significant differences in bacterial vaginosis diagnosis were observed between HESN and HIV-N individuals by multilevel analysis. Although higher proportions of HIV-N are BVI/BV+ in earlier re-surveys, these differences are not significant due to the small number of HIV-N sampled at these time-points. Therefore, HESN are just as likely to be diagnosed with bacterial vaginosis compared with HIV-N, contrary to the hypothesis that motivated this study.
HIV+ women were significantly more likely to be diagnosed as BVI/bacterial vaginosis+ compared to all HIV– combined [odds ratio (OR) 1.4, 95% confidence interval (CI) 1.2–1.7, P < 0.001; Fig. 1b]. This difference was observed at every time-point, but was only significant among new recruits and at two re-surveys. The proportion of those diagnosed as BVI/BV+ has also changed over time, most significantly for HIV+ women.
Regardless of HIV serostatus, women with BVI/BV+ diagnoses are consistently more likely to have leukocytes observed on Gram stain, indicating an inflammatory milieu , compared to BV− women (P < 0.001 for all samples and at every time-point, Fig. 1c). The proportion of BVI/BV+ samples in which leukocytes are observed has also changed significantly over time, suggesting an overall improvement in vaginal health in this cohort over time, particularly for HIV+ women with bacterial vaginosis.
Although no significant differences by chi-squared test were observed at individual time-points (not shown), HIV+ individuals had a significantly reduced overall abundance of Lactobacillus morphotypes on Gram stain compared to HIV– individuals (P < 0.001, Fig. 1d). Although the number of seroconverters was small, samples provided by this group had a significantly increased abundance of bacterial vaginosis morphotypes compared to both HIV+ and HIV– groups combined (P < 0.001, Fig. 1d).
Bacterial vaginosis diagnosis and systemic T-cell subsets
Although no significant differences were observed by Wilcoxon rank sum test at individual time-points (not shown), HIV+ women with BVI/BV+ diagnoses had significantly increased CD4+ and CD8+ T-cell counts and a decreased ratio of CD4+ to CD8+ T cells in peripheral blood (P = 0.02, P < 0.001 and P < 0.001 respectively, Fig. 2b) in peripheral blood compared to those who are BV−. No differences of this kind were observed in HIV– women (Fig. 2a).
In a cross-sectional analysis, the proportion of Tregs to total (CD3+) T cells was significantly reduced in HIV+ women with BV+/BVI diagnoses compared to HIV+ BV− women (P = 0.02; Fig. 2c), whereas differences were not significant among HIV– women. These findings indicate an association between BVI/BV+ diagnoses, increased frequency of systemic CD4+ and CD8+ T cells and reduced Tregs in HIV+ women.
No previous studies have addressed possible links between bacterial vaginosis and HIV resistance among highly exposed CSWs. Any investigation of these linkages is complicated by lack of an objective definition or prevailing consensus about the nature of HIV resistance . The time-based definition used in this and many other recent studies, based on length of follow-up between 3 and 7 years [22,23], cannot distinguish women who are in fact HESN but have not been followed for long enough to be defined as such. Therefore, further studies of HESN in relation to alternative controls, such as non-CSW, may clarify whether or not bacterial vaginosis plays a role in HIV resistance in this cohort.
Only a small number of studies have shown an association between bacterial vaginosis and HIV+ serostatus using longitudinal data [16,17,24,25]. Interestingly, the OR we report is identical to that observed in a cohort of Mombasa CSWs 10 years earlier .
The proportion of individuals diagnosed with bacterial vaginosis has changed significantly since 2004. Observed reductions in bacterial vaginosis may have significant implications for HIV incidence in this group. In fact, HIV incidence declined steadily in this cohort between 1985 and 2005, concomitantly with decreases in gonorrhea incidence and increases in reported condom use . These changes may be associated with improved overall vaginal health associated with treatment of sexually transmitted infections, condom use, or other factors. A recent randomized control trial focussed on vaginal health showed reduced bacterial vaginosis, increased Lactobacillus and reduced trichomoniasis in a cohort of Kenyan CSWs .
HIV+ women with bacterial vaginosis had increased levels of CD4+ and CD8+ T cells and reduced Treg cells as a proportion of total T cells in peripheral blood. The extent to which peripheral cells are influenced by bacteriological changes in the vagina is unknown. However, immune cells continuously traffic between mucosal compartments, lymphoid tissues and the periphery, bridging the gap between these sites. Alternatively, bacterial vaginosis might be a microbiological ‘symptom’ of immunological events. Since almost nothing is known about the factors triggering bacterial vaginosis or the transition from one state to another, it seems plausible that some immune mechanism involving changes in T-cell subset frequencies might be the cause of bacterial vaginosis.
Further work will be required to understand how microbiological and host factors interact as part of the normal functioning of mucosal surfaces exposed to a wide variety of endogenous and exogenous stimuli. How these interactions influence the likelihood that HIV introduced at the interface of host and microbiota will cross formidable physicochemical barriers and irreversibly alter host immune function should be a central question of HIV prevention research in coming years.
The authors would like to acknowledge the members of the Majengo CSW cohort for their ongoing participation, as well as the dedicated clinical and laboratory staff in Nairobi and Winnipeg who make this research possible. M. Cheang, Department of Community Health Sciences, University of Manitoba provided multilevel statistical analyses using the HLM software package.
Author contributions: J.S., C.C., T.B., K.F. and F.P. designed the study. J.S., C.C., J.N.M., E.I., J.K., W.K. and C.W. generated data. J.S. and C.C. analyzed data and wrote the paper, with helpful comments from T.B. and K.F.
Conflicts of interest
The authors declare no financial conflicts.
Funding: National Institutes of Health, Bill & Melinda Gates Foundation, Canadian Institutes of Health Research (CIHR). J.S. and C.C. have received CIHR Doctoral Research Awards and fellowships in the CIHR/International Centre for Infectious Diseases (J.S.) and CIHR International Infectious Diseases and Global Health (C.C.) Training Programs. K.F. holds a Manitoba Research Chair from the Manitoba Health Research Council. F.A.P. is a Tier I Canada Research Chair.
1. Fowke KR, Nagelkerke NJ, Kimani J, Simonsen JN, Anzala AO, Bwayo JJ, et al. Resistance to HIV-1 infection among persistently seronegative prostitutes in Nairobi, Kenya
2. Miyazawa M, Lopalco L, Mazzotta F, Lo Caputo S, Veas F, Clerici M. The ’immunologic advantage’ of HIV-exposed seronegative individuals
3. Burgener A, Boutilier J, Wachihi C, Kimani J, Carpenter M, Westmacott G, et al. Identification of differentially expressed proteins in the cervical mucosa of HIV-1-resistant sex workers
. J Proteome Res
4. Iqbal SM, Ball TB, Levinson P, Maranan L, Jaoko W, Wachihi C, et al. Elevated elafin/trappin-2 in the female genital tract is associated with protection against HIV acquisition
5. Card CM, McLaren PJ, Wachihi C, Kimani J, Plummer FA, Fowke KR. Decreased immune activation in resistance to HIV-1 infection is associated with an elevated frequency of CD4(+)CD25(+)FOXP3(+) regulatory T cells
. J Infect Dis
6. McLaren PJ, Ball TB, Wachihi C, Jaoko W, Kelvin DJ, Danesh A, et al. HIV-exposed seronegative commercial sex workers show a quiescent phenotype in the CD4+ T cell compartment and reduced expression of HIV-dependent host factors
. J Infect Dis
2010; 202 (Suppl 3):S339–S344.
7. Lo Caputo S, Trabattoni D, Vichi F, Piconi S, Lopalco L, Villa ML, et al. Mucosal and systemic HIV-1-specific immunity in HIV-1-exposed but uninfected heterosexual men
8. Restrepo C, Rallon NI, del Romero J, Rodriguez C, Hernando V, Lopez M, et al. Low-level exposure to HIV induces virus-specific T cell responses and immune activation in exposed HIV-seronegative individuals
. J Immunol
9. Haase AT. Perils at mucosal front lines for HIV and SIV and their hosts
. Nat Rev Immunol
10. Hummelen R, Fernandes AD, Macklaim JM, Dickson RJ, Changalucha J, Gloor GB, et al.Deep Sequencing of the Vaginal Microbiota of Women with HIV
. PLoS One
11. Larsson PG, Bergstrom M, Forsum U, Jacobsson B, Strand A, Wolner-Hanssen P. Bacterial vaginosis. Transmission, role in genital tract infection and pregnancy outcome: an enigma
12. Westrom L, Evaldson G, Holmes KK. Taxonomy of vaginosis: bacterial vaginosis: a definition
. Scand J Urol Nephrol Suppl
13. Atashili J, Poole C, Ndumbe PM, Adimora AA, Smith JS. Bacterial vaginosis and HIV acquisition: a meta-analysis of published studies
14. Hay P. Life in the littoral zone: lactobacilli losing the plot
. Sex Transm Infect
15. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation
. J Clin Microbiol
16. Martin HL, Richardson BA, Nyange PM, Lavreys L, Hillier SL, Chohan B, et al. Vaginal lactobacilli, microbial flora, and risk of human immunodeficiency virus type 1 and sexually transmitted disease acquisition
. J Infect Dis
17. Myer L, Denny L, Telerant R, Souza M, Wright TC Jr, Kuhn L. Bacterial vaginosis and susceptibility to HIV infection in South African women: a nested case-control study
. J Infect Dis
18. Ross SA, Novak Z, Ashrith G, Rivera LB, Britt WJ, Hedges S, et al. Association between genital tract cytomegalovirus infection and bacterial vaginosis
. J Infect Dis
19. Schellenberg J, Blake Ball T, Lane M, Cheang M, Plummer F. Flow cytometric quantification of bacteria in vaginal swab samples self-collected by adolescents attending a gynecology clinic
. J Microbiol Methods
20. Mardh PA, Novikova N, Niklasson O, Bekassy Z, Skude L. Leukocyte esterase activity in vaginal fluid of pregnant and nonpregnant women with vaginitis/vaginosis and in controls
. Infect Dis Obstet Gynecol
21. Horton RE, McLaren PJ, Fowke K, Kimani J, Ball TB. Cohorts for the study of HIV-1-exposed but uninfected individuals: benefits and limitations
. J Infect Dis
2010; 202 (Suppl 3):S377–S381.
22. Burgener A, Rahman S, Ahmad R, Lajoie J, Ramdahin S, Mesa C, et al.Comprehensive proteomic study identifies serpin and cystatin antiproteases as novel correlates of HIV-1 resistance in the cervicovaginal mucosa of female sex workers
. J Proteome Res
23. Schellenberg JJ, Links MG, Hill JE, Dumonceaux TJ, Kimani J, Jaoko W, et al. Molecular definition of vaginal microbiota in East African commercial sex workers
. Appl Environ Microbiol
24. Taha TE, Hoover DR, Dallabetta GA, Kumwenda NI, Mtimavalye LA, Yang LP, et al. Bacterial vaginosis and disturbances of vaginal flora: association with increased acquisition of HIV
25. van de Wijgert JH, Morrison CS, Brown J, Kwok C, Van Der Pol B, Chipato T, et al. Disentangling contributions of reproductive tract infections to HIV acquisition in African women
. Sex Transm Dis
26. Kimani J, Kaul R, Nagelkerke NJ, Luo M, MacDonald KS, Ngugi E, et al. Reduced rates of HIV acquisition during unprotected sex by Kenyan female sex workers predating population declines in HIV prevalence
27. McClelland RS, Richardson BA, Hassan WM, Chohan V, Lavreys L, Mandaliya K, et al. Improvement of vaginal health for Kenyan women at risk for acquisition of human immunodeficiency virus type 1: results of a randomized trial
. J Infect Dis
bacterial vaginosis; commercial sex workers; HIV resistance; HIV/AIDS; multiparameter flow cytometry; T-helper cell subsets; Treg
© 2012 Lippincott Williams & Wilkins, Inc.
Highlight selected keywords in the article text.