Herpes Simplex Virus Type-2 shedding and genital ulcers during early HIV in Zimbabwean women

Abstract

Background:

Herpes Simplex Virus Type-2 (HSV-2) seropositive persons have a 3–5 fold higher risk of acquiring HIV, possibly due to HSV-2 induced inflammation and recruitment of susceptible immune cells to exposure sites. We hypothesized cervical HSV-2 activation (i.e., viral DNA shedding and/or ulcers) preceded HIV acquisition in the Hormonal Contraception and HIV cohort.

Methods:

Zimbabwean women who acquired HIV were matched to HIV-negative women on visit, age, and bacterial sexually transmitted infections (STIs). Up to 5 cervical swabs bracketing first PCR detection of HIV DNA, the index visit, were selected (t-_6months, t-_3months, t_index, t+_3months, t+_6months). Women with HSV-2 IgG+ before t_index were PCR-tested for viral shedding. Self-reported and clinician-diagnosed ulcers were documented. Multivariable logistic regression, accounting for matching, estimated adjusted odds ratios [aOR] and 95% confidence intervals [CI] at each visit.

Results:

Of 387 HSV-2 seropositive women, most had prevalent as compared to incident HSV-2 (91% vs. 9%, respectively). HSV-2 viral shedding was more common among HIV seroconverters than HIV-negative women (26% vs. 14%, p<0.01). Shedding occurred around HIV acquisition (t-_3months aOR: 2.7, 95% CI: 0.8–8.8; t_index aOR: 2.6, 95% CI: 1.1–6.5; t+_3months aOR: 2.6, 95% CI: 1.0–6.6). Genital ulcers were reported more often among HIV seroconverters than HIV-negative women (13% vs. 7%, p=0.06) and detection was after HIV acquisition (t+_6months aOR: 14.5, 95% CI: 1.6–133.9).

Conclusions:

HSV-2 shedding appeared synergistic with HIV acquisition followed by presentation of ulcers. Evaluating all STIs rather than HSV-2 alone may clarify the relationship between inflammation and HIV acquisition.

INTRODUCTION

Herpes Simplex Virus Type-2 (HSV-2) is highly prevalent in African women, ranging from 30 to 80% ^1,2. HSV-2 is incurable and frequently recurs in the genital tract mucosa ³. These reactivation events may manifest as symptomatic ulcers or asymptomatic viral shedding ⁴. Both range in frequency and duration depending on whether HSV-2 was recently acquired ⁴ and/or was symptomatic ⁵. Subsequent to recurrence, there may be an enrichment of memory CD4+ T cells ^6–8 and an alteration of the epithelial barrier ⁹. This co-localization of target immune cells and compromised immune barrier has been hypothesized to increase susceptibility to HIV transmission.

A systematic review and meta-analysis supports this biological mechanism, finding a higher risk of HIV among recent vs. prevalent HSV-2 antibody positive individuals (incident HSV-2 aRR: 4.7, 95% CI:2.2–10.1; prevalent HSV-2 adjusted relative risk [aRR]: 2.7, 95% CI: 2.2–3.4) ¹⁰. Higher viral activation within the first year of HSV-2 infection may explain this differential risk ¹¹. However, this was not supported by a study that found active HSV-2, defined as self-reported symptomatic genital ulcers from the prior 6 months, was increased concurrent with and after HIV seroconversion ¹². Self-reported data, 10-month visit intervals, and no measurement of viral shedding may have contributed to the null association before HIV acquisition. Therefore, it remains unclear whether subclinical HSV-2 infections (antibody positive, DNA negative) ^10,13–15 create an immune environment conducive for HIV transmission.

We evaluated HSV-2 viral activity at multiple 3-month intervals preceding, during and post HIV acquisition among HSV-2 antibody positive Zimbabwean women. We hypothesized that viral shedding and genital ulcers would peak before HIV acquisition as compared to a time-independent pattern among HIV-negative women. Secondarily, we explored whether HSV-2 activation occurred after HIV acquisition as observed previously ¹².

METHODS

Parent Cohort and Procedures

The Hormonal Contraception and HIV cohort (HC-HIV) recruited women from reproductive health clinics to evaluate hormonal contraceptive use (depot medroxyprogesterone acetate and combined oral contraceptives) on the risk of HIV acquisition between 1999 and 2004 ¹⁶. Women aged 18–35 years, sexually active and HIV-negative were seen quarterly. They completed a structured survey and provided blood and cervical swabs for diagnosis of HIV, HSV-2 and bacterial sexually transmitted infections (STIs). Blood was tested for HIV by an enzyme-linked immunosorbent assay (ELISA). Positive results were confirmed with rapid testing, followed by Western blot or PCR testing. HIV DNA PCR was performed on all visits preceding the HIV-seropositive visit, the first PCR positive date defining HIV incidence. Blood was tested for HSV-2 type-specific IgG anti body ELISA (Focus Technologies, Cypress, CA) ¹³. HSV-2 ELISA testing continued for the non-prevalent cases during study follow-up to identify incident cases. For quality control, 10% of HSV-2 seropositive samples underwent repeat serological testing with Western blot confirmation at an external laboratory ¹⁷. For bacterial STI testing, cervical swabs were rotated in the endocervix for 15–30 seconds before placing in 1 mL of Amplicor lysis buffer (Roche Diagnostics, Somerville, NJ, USA). Cervical swabs were PCR tested for Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG) (Amplicor CT/NG, Roche Diagnostics, Somerville, NJ, USA) ¹⁸. Trichomonas vaginalis (TV) was detected by wet mount ¹⁹.

Sample Selection for Active HSV-2

All Zimbabwean women from the HC-HIV cohort with incident HIV were matched to HIV-negative women on time in study, age, and a composite STI variable. Composite STI positivity was defined as having any CT, NG, or bacterial vaginosis at the HIV seroconversion visit or the visit before ²⁰. For cases, up to five consecutive cervical swab samples that bracketed and included the first HIV DNA-positive visit (index visit) were selected (t-_6mo, t-_3mo, t_index, t+_3mo, t+_6mo). The same selection process was used for the controls but the anchoring of t_index was defined by the incident HIV visit of the matched cases.

In May–September 2008, aliquots of the residual cervical swab samples were shipped on dry ice to Johns Hopkins Bloomberg School of Public Health. Concurrent data, including history of ulcers, HSV-2 serology, TV, human papillomavirus (HPV) and bacterial STIs, were available from the HC-HIV cohort and an HPV sub-study ^16,20. Of the case-control sets, only women with a serology-defined prevalent or incident HSV-2 before the index visit were evaluated for active HSV-2. All visits were tested, including seronegative visits for HSV-2 incident infections at t-_6months.

Laboratory testing for Active HSV-2

HSV-2 DNA was extracted from 250 uL of the cervical swab samples using the QIAamp DNA blood MDx protocol (Qiagen, Valencia, CA). DNA was resuspended in 185 uL of TE Buffer. TaqMan real-time PCR was used to detect HSV-2 using a type-specific probe (GbType2: CGG CGA TGC GCC CCA G with FAM at the 5’-end and TAMRA at the 3’-end) from 5 μL of genomic DNA ²¹. The viral load assay used a reaction volume of 50 μl with Universal PCR Master Mix without UNG (Applied Biosystems, Foster City, CA). Amplification was conducted on the Applied Biosystems 7300 Thermocycler with 10 minutes at 52°C, 12 minutes at 95°C, followed by 50 cycles of 95°C for 15 sec. and 60°C for 1 minute. HSV-2 control DNA plasmid stock was serially diluted, five-fold, in a background of 50ng/μl of human placental DNA in LoTE buffer to create standard curves of known concentrations. Standard curves and negative controls were run in duplicate. PCR testing finished in August 2009.

Ethical considerations

This study was approved by the institutional review boards of the Medical Research Council of Zimbabwe, Case Western Reserve University, FHI 360 and the Johns Hopkins Bloomberg School of Public Health. All women provided written informed consent.

Statistical Analysis

The primary outcome was a binary categorization of incident HIV infection. The independent variables were HSV-2 viral shedding, defined as any PCR detected virus, and self-reported and/or clinician-diagnosed genital ulcers at each visit. Pearson’s Chi-square and Fisher’s exact tests compared crude differences in participant characteristics. Prevalence and 95% confidence intervals (CIs) of active HSV-2 by HIV status and follow-up time were estimated using binomial models. Bivariate and multivariable unconditional logistic regression was used to estimate odds ratios (ORs) and 95% CIs. Because case-control sets were altered after selecting only HSV-2 seropositive women, all models included a priori the matching factors to account for any selection bias. Variables statistically significant in bivariate analyses (p<0.05) that changed the model coefficients by more than 10% were retained in final models. The odds of HSV-2 viral shedding and genital ulcers among cases as compared to controls were adjusted for matching factors, history of ulcers, and prevalent or incident HPV infections at each study visit. A p-value <0.05 was considered statistically significant. Data were analyzed using Stata Statistical Software, Release 13 (StataCorp, College Station, TX).

RESULTS

Of the 632 matched women from HC-HIV, 61% (n=387) had a prevalent or incident seropositive HSV-2 infection detected prior to the index visit. Of the 387, 132 (34%) HIV seroconverted and 255 (66%) remained HIV-negative. HIV seroconverters were less likely to self-report a history of genital ulcers (4% vs. 10%, p=0.03) and have a Nugent score of 0-3 (26% vs. 39%, p=0.03)(Supplemental Table 1). Seroconverters were more likely to have 3 or more prevalent HPVs (27% vs. 13%, p<0.01), NG (16% vs. 2%, p<0.01), and CT (10% vs. 2%, p<0.01). Incident HSV-2 serology positive cases and TV were non-significantly more common among women who HIV seroconverted (13% vs. 8%, p=0.08; 5% vs. 2%, p=0.07, respectively).

Most women had prevalent HSV-2 (91%, n=351) as compared to incident HSV-2 (9%, n=36). Any HSV-2 viral shedding was detected in 18% (70/387) of women, with 16% (60/387) having a single event and 3% (10/387) having two events during follow-up. HIV seroconverters shed HSV-2 more frequently (26% vs. 14%, p<0.01) as did women who were co-infected with NG (39% vs. 17%, p<0.01) while women with CT or TV did not (p>0.05). Ever shedding was also higher in women with incident as compared to prevalent HSV-2 (39% vs. 16%, p<0.01). Prevalence of HSV-2 viral shedding began to differ three months before and through three months post the index visit for cases as compared to controls (t-_3months: 6.5% vs. 2.0%, p=0.03; t_index: 9.2% vs. 4.1%, p=0.05; t+_3months: 10.1% vs. 4.7%, p=0.06)(Figure 1A).

Figure 1A-B.

Prevalence of HSV-2 viral shedding and genital ulcers before, during and after HIV acquisition among HSV-2 seropositive women.

Note: The gray bands are the prevalence and 95% CIs of shedding and genital ulcers among the HIV-negative overall as an indicator of expected background prevalence. HIV seroconversion likely occurred between t_-3months and t with HIV DNA being first detected at t.

Clinician and self-reported genital ulcers were detected in 9% (35/387) of women. Overall, genital ulcers were detected more often among women reporting a history of ulcers (26% vs. 8%, p<0.01) but were non-significantly higher among women who HIV seroconverted (13% vs. 7%, p=0.06). Prevalence of ulcers began to differ after HIV acquisition (t+_3months: 5.6% vs. 2.4%, p=0.14; t+_6months: 11.1% vs. 0.5%, p<0.01) for cases as compared to controls (Figure 1B).

A higher odds of HSV-2 viral shedding occurred around the time of HIV acquisition (t-_3months aOR: 2.7, 95% CI: 0.8–8.8; t_index aOR: 2.6, 95% CI: 1.1–6.5; t+_3months aOR: 2.6, 95% CI: 1.0–6.6). Genital ulcers increased after HIV acquisition (t+_6months aOR: 14.5, 95% CI: 1.6–133.9) (Table 1). In a sensitivity analysis where all 5 visits were available for 217 women, HSV-2 viral shedding was significantly detected at the HIV acquisition visit only (t-_3months aOR: 2.5, 95% CI: 0.6–11.7; t_index aOR: 5.8, 95% CI: 1.7–19.3; t+_3months aOR: 1.7, 95% CI: 0.5–5.3). The genital ulcer model did not converge at the t+_6months visit. When women with incident HSV-2 were excluded, viral shedding was only significantly detected at t+_3months (t-_3months aOR: 1.9, 95% CI: 0.5–6.9; t_index: 2.1, 0.7–6.1; t+_3months: 3.2, 1.2–8.9).

Table 1.

HSV-2 viral shedding and genital ulcer detection before, during and after HIV acquisition.

		Viral shedding			Genital Ulcers
	n/N	Unadjusted OR (95% CI)	Adjusted a OR (95% CI)	n/N	Unadjusted OR (95% CI)	Adjusted a OR (95% CI)
t-6months*
HIV-negative	6/216	1.0	1.0	5/215	1.0	1.0
Before SC	4/110	1.2 (0.3-4.6)	1.5 (0.4-6.3)	3/110	1.2 (0.3-5.1)	1.3 (0.3-6.8)
t-3months*
HIV-negative	5/245	1.0	1.0	3/242	1.0	1.0
Before SC	8/123	3.2 (1.0-10.1)	2.7 (0.8-8.8)	2/120	1.4 (0.2-8.2)	2.4 (0.4-15.6)
t index*
HIV-negative	10/243	1.0	1.0	6/240	1.0	1.0
SC**	12/130	2.4 (1.0-5.7)	2.6 (1.1-6.5)	3/130	0.9 (0.2-3.7)	0.9 (0.2-4.0)
t+3months*
HIV-negative	10/213	1.0	1.0	5/213	1.0	1.0
After SC	11/109	2.4 (1.0-6.0)	2.6 (1.0-6.6)	6/108	2.4 (0.7-8.0)	2.5 (0.7-9.0)
t+6months*
HIV-negative	10/185	1.0	1.0	1/185	1.0	1.0
After SC	5/105	0.9 (0.3-2.7)	0.8 (0.3-2.7)	6/54	20.0 (2.3-176.3)	14.5 (1.6-133.9)

Abbreviations. OR, odds ratio; CI, confidence interval; SC, seroconversion; HPV, human papillomavirus.

^aFinal models adjusted for age, composite sexually transmitted infection, ever symptomatic ulcers, and HPV by logistic regression. Bolded indicates p<0.05.

^*t_-6mo was adjusted for prevalent HPV and t_-3mo through t_+6mo were adjusted for incident HPV.

^**HIV seroconversion likely occurred between t_-3months and t_index with HIV DNA being first detected at t_index.

DISCUSSION

Our findings suggest that active HSV-2 increased around the time of HIV acquisition while genital ulcers were detected 6 months post HIV acquisition. This is consistent with prior work that found symptomatic genital ulcers were detected with and immediately following HIV seroconversion ¹². In another study, symptomatic genital ulcers positive for HSV-2 had a nearly 2-fold higher risk in the 6 months after as compared to before HIV seroconversion ²². Together these findings suggest that acute HIV infection triggers HSV-2 reactivation followed by presentation of genital ulcers. A non-statistically significant ~2-fold increased risk of HIV associated with HSV-2 shedding 3-months prior to HIV acquisition further suggests that reactivated HSV-2 increases risk of HIV acquisition.

After restricting the analysis to women with prevalent HSV-2, viral shedding shifted later to t+_3month. Despite our limited power to stratify, the loss of a significant detection of viral shedding at t_index suggests incident HSV-2 may be contributing events earlier than prevalent HSV-2, consistent with prior work ^10,11,13,15. Since the study design restricted the window of HIV acquisition to between t-_3months and t_index and knowing that HSV-2 reactivation may be of short duration (hours) ²³, capturing the precise window of viral interaction remains a challenge. Only later as HIV becomes chronic and T-cells are depleted is the balance between host and virus lost and a temporal pattern emerges ^24–27.

Randomized controlled trials (RCTs) providing suppressive HSV-2 therapy (i.e., acyclovir and valacycloivir) have been unsuccessful at curbing HIV transmission ^28–30. Ineffective drug doses ³¹, persistence of CCR5+ target cells ⁶, and host differences in the pharmacokinetic availability of the antivirals ³² may have contributed to the null findings. Despite higher doses of antiviral drugs ³¹, viral synergism may still be occurring in the genital mucosa. This may be complicated by other STIs that stimulate local inflammation. Evaluating the combined impact of all STIs on inflammation, parallel to methods evaluating the compositional structure of the microbiome, may be needed to understand increased vulnerability to HIV.

The lack of significant shedding prior to HIV seroconversion may have been limited by power and sampling. Sampling every 3 months, rather than daily or weekly may have misclassified HSV-2 viral shedding ³. Our study was also underpowered to stratify incident vs prevalent HSV-2 although crude analysis suggested that incident HSV-2 shed more and possibly earlier than prevalent HSV-2 cases. HIV susceptibility may still be higher for incident HSV-2 because of earlier and more frequent viral activity as compared to prevalent HSV-2 ^11,13.

In conclusion, HSV-2 shedding appeared synergistic with HIV acquisition followed by presentation of ulcers. Evaluating HSV-2 together with concurrent STIs may clarify the biologic relationship between inflammation and HIV acquisition.