The prostate as a reservoir for HIV-1

As the prostate can harbor bacterial and fungal pathogens, it was investigated as a reservoir for HIV. Nine men chronically infected with HIV participated in a crossover trial in which weekly semen samples were collected with and without previous prostate massage (PM). Six of the nine participants had undetectable seminal plasma (SP) HIV RNA in all samples without previous PM, but had detectable SP HIV RNA (≥ 25 copies/ml) in one to three samples collected after PM. Most studies of HIV in the male genitourinary compartment have treated seminal fluid as arising from a single source, when in fact semen consists of secretions from at least four exocrine organs and four mucosal exocrine glands. Krieger et al. [1] showed that the majority of HIV found in semen arises distal to the vas deferens, and Paranjpe et al. [2] demonstrated phylogenetic differences in HIV derived from seminal plasma (SP) and seminal cells. These compartments and sub-compartments of HIV probably arise from local replication in the complex tissues that make up the male genital tract. The prostate contributes a large portion of fluid to semen. It can harbor bacterial and fungal infections that are often difficult to treat, because many antimicrobial agents penetrate the prostate poorly [3,4]. Because semen levels of most antiretroviral medications are relatively low compared with blood levels, it may be that these drugs are also inadequate in the prostate [5–7]. To investigate the prostate as a reservoir for HIV, nine men with chronic HIV infection were enrolled in this self-controlled study. Their mean baseline CD4 cell count was 483 cells/ml (range 150–796 cells/ml) and their mean baseline blood plasma viral load (BPVL) was 20 800 copies/ml (range < 50–46 393 copies/ml). Seven participants had been on stable antiretroviral regimens for more than 3 months before starting the trial; two were taking no antiretroviral medications. Three had undetectable BPVL (< 50 copies/ml) throughout the study, with their HIV suppressed for a mean of 6 months before study entry. There were no changes in HIV medications during the trial. Participants were screened for sexually transmitted infections (STI) and were found to be negative for urethral gonorrhea and chlamydia (LCx STD system; Abbott Laboratories, Abbott Park, IL, USA), syphilis (rapid plasma reagin), and active genital herpes (physical examination). No participant reported symptoms to suggest the acquisition of an STI during the study. Participants were asked to abstain from sex for at least 48 h before specimen collection. Semen samples were collected in the morning, by masturbation without lubrication, weekly for 11 weeks. Samples from weeks 1–4 and week 11 were collected without previous prostate massage (PM) and samples from weeks 5–10 were submitted after PM, performed by digital massage of the prostate per the rectum (similar to a digital rectal examination) for 30 s by the author (D.S.). At weeks 5 and 8 samples were collected immediately after PM; at weeks 6 and 9 they were collected 30 min after PM; and at weeks 7 and 10 they were collected 1 h after PM. At the time of collection, 2 ml viral transport medium (80% RPMI 1640, 9% fetal bovine serum, 9% penicillin/streptomycin and 2% nystatin) was added to each sample. SP was separated from the seminal cells by centrifugation at 700g for 12 min within 2 h of specimen collection. HIV RNA was extracted from 1 ml SP from each sample using the Boom method (NucliSens Extractor; BioMeriéux, Inc., Durham, NC, USA) and then quantified (Amplicor HIV-1 Monitor test; Roche Molecular Diagnostics, Pleasanton, CA, USA). HIV-RNA quantities were normalized to ejaculate volume by subtracting out the volume of transport media. SP HIV RNA (≥ 25 copies/ml) was detected with significantly greater frequency when SP was collected after PM. Only 11 out of 44 samples collected without previous PM had a detectable seminal plasma viral load (SPVL) versus 24 out of 53 samples collected after PM (Fisher's exact test, P = 0.035) [8]. There was no correlation between the time lapsed from PM to specimen collection and the detection of SP HIV RNA (data not shown). There was also no correlation between the volume of SP produced and previous PM (data not shown; P = 0.96). Three subjects (nos. 2, 3, and 8) with undetectable baseline BPVL (< 50 copies/ml) had undetectable SPVL (< 25 copies/ml) at all collections without previous PM, but all had at least one sample with detectable HIV RNA after PM (Fig. 1). Three participants (nos. 4, 5, and 9) with detectable baseline BPVL were undetectable in their SPVL without previous PM, but had detectable SP HIV RNA levels after PM (Fig. 1). The remaining subjects (nos. 1, 6, and 7) had baseline detectable BPVL and detectable SPVL at most collections, which did not change with previous PM (data not shown).Fig. 1.: Six of the nine participants had detectable seminal plasma viral loads only after prostate massage. PM, Prostate massage. —♦— Subject 2; —▪— subject 3; —▴— subject 4; —▪— subject 5; — ▪ — subject 8; — • — subject 9.BPVL at baseline and at the end of the study were unchanged in all participants except subject 1, who showed a significant decrease from 42 102 to less than 50 copies/ml during the study. The SPVL levels for this subject varied greatly (3041–285 732 copies/ml), suggesting a different source for his SP virus or a drug sanctuary effect. He had a greater SPVL than BPVL at all collection timepoints, and the SPVL in this participant did not correlate with previous PM. This subject had been treated with amprenavir and lamivudine for 11 months before the start of the study, which he reported was the only regimen that he could tolerate. Medication compliance was not assessed, although the observed decline in BPVL suggests that this participant's compliance may have improved during the course of the study. No resistance-associated mutations (ViroSeq HIV-1 Genotyping System; Celera Diagnostics, Alameda, CA, USA) were identified in virus isolated from his week 2 and week 10 SP, which is not consistent with non-adherence to lamivudine. These data are further evidence of the significant compartmentalization between the blood and the genital tract. They also raise the question of how many patients being treated for HIV have undetectable BPVL, but continue to have high SPVL. The recovery of cell free virus after PM demonstrates that the prostate acts as a sanctuary site for HIV, even among individuals on antiretroviral therapy with virological suppression (< 50 copies/ml) in the blood plasma. These findings extend the report from Zhang et al. [9] of the isolation of latently infected cells harboring replication-competent virus in the genital secretions from men on antiretroviral medication. There are significant transmission implications related to the prostate as a reservoir for HIV. If prostate massage simulates receptive anal intercourse in men who have sex with men, then anal intercourse may increase the amount of HIV RNA in the receptive partner's semen, which he could then transmit in his next sexual encounter. This mechanism could contribute to the high HIV infection rates historically seen among men who have sex with men. Although these data identify a potentially significant variable related to HIV shedding in the male genital tract, further investigation is needed among transmission partner pairs to evaluate both the anatomical source of the infecting virus and the attributable risk of prostate stimulation to HIV transmission. Acknowledgements The authors are grateful to the study participants for their unwavering generosity and to Nancy Keating, Heather Overton, Sherri Rostami, and Melissa Moore for their technical assistance. They would also like to thank Matthew C. Strain, PhD, for statistical assistance; F. Ramzi Asfour, MD, Tari Gilbert, FNP, Paula Potter, RN, and Joanne Santangelo, FNP, for clinical assistance; Gary Dyak, LVN, and Laureen M. Copfer for administratrative assistance; and James Ellzy, MD, for insightful comments. Sponsorship: This work was supported by grants 5K23AI055276, AI27670, AI38858, AI43638, UCSD Center for AIDS Research (AI36214) and AI29164 from the National Institutes of Health, and the Research Center for AIDS and HIV Infection of the San Diego Veterans Affairs Healthcare System. Human Research Protections: These studies were conducted with appropriate subject consent and were approved by the Human Research Protections Program at the University of California, San Diego, CA, USA.