Share this article on:

Lower Genitourinary Tract Sources of Seminal HIV

Coombs, Robert W. MD, PhD*†; Lockhart, David MSc; Ross, Susan O. RN§; Deutsch, Leslie RN§; Dragavon, Joan BSc, MLM*; Diem, Kurt BSc*; Hooton, Thomas M. MD; Collier, Ann C. MD; Corey, Lawrence MD*†; Krieger, John N. MD§

JAIDS Journal of Acquired Immune Deficiency Syndromes: April 1st, 2006 - Volume 41 - Issue 4 - p 430-438
doi: 10.1097/01.qai.0000209895.82255.08
Clinical Science

Objective: To investigate genital tract sources of HIV-1, we conducted extensive genitourinary sampling of 23 seropositive men without urethritis who shed HIV in their seminal plasma.

Design: Semen was collected, then samples were obtained for HIV RNA in blood plasma, urethral fluid, pre-prostate massage fluid/urine (PMF/U) and post-PMF/U, and expressed prostatic secretions. Systematic transrectal ultrasound-guided prostate biopsies obtained from multiple prostate areas were evaluated for HIV RNA and DNA.

Results: Seminal HIV RNA levels correlated with HIV RNA levels in urethral fluid and post-PMF/U and with prostate biopsies HIV DNA, but not with expressed prostatic secretions HIV RNA. However, only the HIV RNA level in post-PMF/U independently predicted that in semen (2.77-fold change in semen for each 10-fold change in post-PMF/U; 95% confidence interval, 1.0-7.7) accounting for one third of the seminal HIV RNA level variation, irrespective of adjustment for antiretroviral therapy.

Conclusions: These data indicate that distal genitourinary sources other than the prostate appear to be the major source of seminal HIV in men without clinical urethritis or prostatitis. Because the HIV RNA level in blood plasma is not reliable as an independent clinical predictor of virus levels in seminal plasma, these findings also extend the concept that the male genital tract is a distinct virological compartment from blood.

From the Departments of *Laboratory Medicine, †Medicine, ‡Biostatistics, and §Urology, University of Washington, Seattle, WA.

Received for publication August 8, 2005; accepted November 16, 2005.

Informed consent was obtained from all subjects following human experimentation guidelines of the United States Department of Health and Human Services and the University of Washington.

The authors have no commercial associations that might pose a conflict of interest.

Presented in part at the 11th Conference on Retroviruses and Opportunistic Infections; February 8 to 11, 2004; San Francisco, CA (abstract 420).

Funding sources: DK-49477; Centers for AIDS Research AI-27757; AIDS Clinical Trials Group AI-27664 and AI-38858, AI-41535, and HD-40540.

Reprints: Robert W. Coombs, MD, PhD, Harborview Medical Center, Research & Training Building, Room 706, 325-9th Avenue, Seattle, WA 98104-2499 (e-mail:

Semen is well established as a primary vehicle for sexual transmission of HIV-1. Surprisingly, the anatomic sites and sources of seminal HIV are largely unknown. Based on earlier studies showing that seminal shedding of HIV RNA and HIV DNA continued after vasectomy, we and others deduced that distal genitourinary anatomic sites are likely important sources of seminal HIV.1-4 We report a pilot study to test the hypotheses that the urethra and prostate are critical viral sources and that these sites remain reservoirs of genital virus in men who continue to shed seminal HIV while on antiretroviral therapy (ART). We localized and quantified HIV viral burdens in specimens systematically obtained from various genitourinary sites by using systematic anatomic sampling of genital fluids and prostate biopsies (Pbx).

Back to Top | Article Outline


Study Design

Our clinical population of HIV-seropositive men was screened between April 1999 and January 2003 to identify men either on no ART or on stable regimens (for at least 3 months) and who had more than 200 HIV RNA copies/mL of blood and seminal plasma. Following written informed consent, each participant agreed to provide 3 separate samples of semen and blood before collecting urethral fluid (UrF; both a urethral swab and/or a less invasive wick), expressed prostate secretions (EPSs), and 6 Pbx. All nonseminal genitourinary specimens were collected during a single clinic visit. The initial study design included only UrF and the prostate biopsy procedure (since we obtained multiple biopsies). However, we modified the protocol to also include pre-prostate massage fluid collected in urine (PMF/U) and post-PMF/U when we noted a low rate of HIV detection in biopsies from the initial participants.

Each subject had a standardized history and physical examination by the same examiner to exclude potential subjects with symptoms or signs of active urethritis, prostatitis, or other active genitourinary tract problems. Semen samples were collected at each subject's home and transported by taxi to the urology laboratory.5 After semen analysis and staining, the specimen was sent to the retrovirus laboratory by our courier service. The semen analysis studies were accomplished within 2 hours, and viral studies were initiated within 4 hours of sample collection.6

Back to Top | Article Outline

Blood Specimens

Peripheral blood was collected into EDTA-containing tubes (Becton Dickinson, Franklin Lakes, NJ), separated within 8 hours of collection into plasma and peripheral blood mononuclear cells by Ficoll-Hypaque gradient centrifugation, then frozen at −70°C. All specimens were tested in batch for HIV RNA at the University of Washington Retrovirology Laboratory, which participates in the National Institute of Allergy and Infectious Diseases (NIAID)-sponsored Virology Quality Assurance Program (Rush Presbyterian Hospital, Chicago, IL). Lymphocyte subsets in fresh blood were determined by standard flow cytometry.

Back to Top | Article Outline

Semen Processing, Staining, and Analysis

Samples were collected by masturbation before the planned clinic visit for the prostate biopsy and after a prescribed 2 to 7 days of abstinence, transported to the laboratory within 4 hours of collection, diluted 1:1 with RPMI-1640 culture medium to reduce viscosity, then centrifuged at 2940g for 2 to 4 minutes to separate the seminal plasma and cells.6 Seminal plasma and cell aliquots were stored at −70°C for later polymerase chain reaction (PCR) analysis within 6 hours of collection.5,7,8 Semen analyses followed our published methods,9,10 including staining to discriminate leukocytes from immature sperm forms. Computerized strict criteria morphology on Papanicolaou-stained thin smears was performed using the Hamilton Thorne Research "Dimensions" program (Hamilton Thorne Biosciences, Inc., Beverly, MA).

Back to Top | Article Outline

Sampling of UrF and EPSs

The specimen sampling sequence is shown in Figure 1. After inserting 3 Sno-strip wicks (Chauvin Pharmaceuticals Ltd., Essex, UK) into the meatus, the pendulous and glandular urethra was "milked" until UrF saturated the wick to the shoulder (∼8 μL per wick). Wick tips were cut at the shoulder and placed into a vial containing 500 μL of guanidinium solution (4 mol/L guanidinium thiocyanate, 25 mmol/L sodium citrate [pH 7], 0.5% N-lauroylsarcosine; 0.1 mol/L 2-mercaptoethanol) and frozen at −70°C within 4 hours of collection.11 In some instances, a Dacron urethral swab (Becton Dickinson, Sparks, MD) was used instead of or in addition to the wicks.



A pre-PMF/U specimen (first void urine, 1-5 mL) was collected into a sterile urine container, and EPSs were obtained by digital rectal massage following our standard procedures.7,12 Briefly, the right prostate lobe was massaged from lateral to medial aspects starting with the bladder base, middle, and then prostate base (apical) segments, with each segment receiving at least 3 massages before proceeding to the next segment; this was repeated on the left side and completed by massaging from the bladder base to the prostate apex along the medial prostate surface twice. The procedure was then repeated until EPS was seen at the meatus: generally, the total procedure lasted 2 to 4 minutes. A Dacron urethral swab (Becton Dickinson) was saturated with EPS (180 μL per swab), placed into 500 μL of guanidinium solution, and processed as above.11 A post-PMF/U was obtained as above, and both urine specimens and swab were sent to the laboratory and frozen at −70°C within 4 to 6 hours of collection.

Back to Top | Article Outline

Transrectal Ultrasound-Guided Prostate Biopsy

Subjects received a single oral dose of either ciprofloxacin or levofloxacin for prophylaxis 1 hour before the time of biopsy. Each subject was placed in the left lateral decubitus position, and the lower genitourinary tract and surrounding structures were systematically evaluated by transrectal ultrasound (TRUS) using a Brüel and Kjaer, model 1846, 7-mHz biplanar transducer (Norcross, GA). Measurements of prostate gland dimensions and volume were calculated for a prolate ellipsoid according to the manufacturer's program. Six transrectal Pbx (left and right prostate, respectively, at bladder base, midprostate, and prostate base) were obtained under real-time TRUS guidance using an 18-gauge needle and Biopsy System (C.R. Bard, Covington, GA).13,14 Samples were placed into an OCT fixative (Sakura Finetek, Torrence, CA) and stored frozen in liquid nitrogen. Subjects were asked to report fever, dysuria, blood in the urine or semen, or rectal pain anytime during the week following the procedure.

Back to Top | Article Outline

HIV Quantification

Genital Fluids

The Amplicor Monitor HIV RNA standard assay (Roche Molecular Systems, Inc., Branchburg, NJ) was used. To remove PCR inhibition, HIV RNA was extracted from each specimen using silica gel.6 The ultrasensitive specimen preparation was used for pre-PMF/U and post-PMF/U. We did not perform an ultracentrifugation step for seminal plasma, UrF, or EPS, which would have increased the sensitivity but would have also increased silica gel-resistant inhibition. The lower limit of quantification was 50 RNA copies/mL of PMF/U (unadjusted for dilution of PMF by urine), 400 copies/mL of seminal plasma (2-fold dilution of semen with RPMI), 600 copies/mL of UrF or EPS collected by swabs (3-fold dilution in final guanidinium solution), and 4000 copies/mL of wick extract collected by Sno-strips (20-fold dilution in final guanidinium solution).

Back to Top | Article Outline

Blood Plasma

The Amplicor Monitor HIV-1 RNA assay with or without ultrasensitive specimen preparation protocol was used for blood plasma (lower limit of HIV RNA quantification was either 50 or 200 copies/mL, respectively).

Back to Top | Article Outline

Prostate Biopsy

Frozen biopsy specimens from the right or left prostate were thawed, pooled, and washed twice with phosphate-buffered saline, and the tissue nucleic acid was extracted in a buffer containing guanidinium according to the Promega Total RNA Isolation System (Promega, Madison, WI) or Qiagen DNA (Qiagen Sciences, Germantown, MD) extraction protocols. HIV nucleic acid was quantified using independently validated real-time PCR amplification for HIV RNA and DNA. The real-time assays met the proficiency standards for HIV RNA established by the Adult AIDS Clinical Trials Group (AACTG) and National Institute of Allergy and Infectious Diseases (NIAID). All laboratory tests were performed blinded to the clinical and other virological results. Forward primer HXB2-gag-F (CAA GCA GCC ATG CAA ATG TT), reverse primer SK431-B (TGC TAT GTC ACT TCC CCT TGG TTC TCT), and probe HXB2-gag (6FAM-AAA GAG ACC ATC AAT GAG GAA GCT GCA GAA-TAMRA) were used. All primers were high- performance liquid chromatography-purified and purchased from Invitrogen (Frederick, MD), and all probes were high-performance liquid chromatography-purified and purchased from Applied Biosystems (Foster City, CA). An armored RNA HIV-1 standard, subtype B (Ambion, Inc, Austin, TX) curve was prepared from a stock preparation containing 30,000 RNA copies/μL and included dilutions down to 30 copies of HIV RNA. An HIV subtype B DNA standard was prepared from ACH-2 cells (National Institutes of Health AIDS Research and Reference Reagent Program, Germantown, MD). DNA was quantified spectrophotometrically, and the HIV DNA copy number per cell was determined by end-point dilution PCR. The HIV RNA standard was verified using the Roche Amplicor Monitor HIV RNA assay. The lower limit of quantification was 25 HIV RNA copies/μg of total RNA and 5 HIV DNA copies/μg of total DNA. Results were normalized for total micrograms of tissue RNA or DNA, respectively.

Back to Top | Article Outline

Statistical Analyses

Before data analysis, HIV RNA and DNA concentrations were log10-transformed to stabilize variances and improve the fit to a normal distribution. The geometric mean was used to summarize multiple measurements for each subject. Values below quantification limits (because these differed for various genitourinary fluids) were assigned values halfway between zero and the lower limit of detection. To evaluate associations with seminal HIV RNA level, Pearson correlation coefficients (r), Student t test for means, paired Wilcoxon test, and multiple linear regression analyses were used. Robust standard errors based on jackknifed residuals were used to account for unequal variances and small samples sizes.15 The relative change in seminal HIV RNA level was determined for each 10-fold change in the variable tested after adjusting for ART.

Back to Top | Article Outline


Demographics and Evaluation at Study Entry

The majority of the 23 subjects were white and not on ART (Table 1). Only 6 (33%) of 18 men with CD4 cell counts less than 350 per microliter were on ART. Seminal plasma HIV RNA log10 copies/mL were lower for men on ART (mean, 3.40; SD, 1.25) compared with those not on ART (mean, 4.74; SD, 0.83; 1-way analysis of variance, P = 0.03). However, significant ART-associated differences were not seen for blood HIV RNA level, CD4 cell count, semen volume, seminal leukocyte count, or sperm count. Semen volume correlated strongly with prostate gland volume (r 2 = 0.51, P = 0.003) and correlated inversely with CD4 cell count (r 2 = 0.23, P = 0.03). Six (40%) of 15 men evaluated had leukospermia (defined by >1 × 106 white blood cells [WBCs]/mL of semen).



Back to Top | Article Outline

HIV Detection Frequency

The 23 men provided 260 total specimens, including 6 men who had a complete set of specimens collected and 17 men who had incomplete specimen collection (Table 2). Prostrate biopsies were obtained from all 23 men and evaluated for HIV in 22 men. No participant experienced untoward effects from the TRUS biopsy procedure (ie, sepsis or urinary tract infection), and only minor and expected traumatic complications such as transient hematuria, hematospermia, or rectal discomfort occurred. Semen samples were provided by 15 men at a median of 19 days (range, 0-105 days) before the prostate biopsy procedure. There was no significant difference in viral detection frequency between blood, semen, or prostate tissue among men who provided a complete set of specimens compared with those who did not. In prostate tissue, detection frequencies of viral RNA or DNA or both in 22 men were RNA positive, 18% or 82%; DNA positive, 6% or 27%; both RNA and DNA positive, 6% or 27%; both RNA and DNA negative, 4% or 18%; and RNA positive and DNA negative, 12% or 55%. Two subjects (13 and 17) had detectable HIV in their semen (geometric mean copies/mL, 31,300), UrF (40,100), pre-PMF/U (25) and post-PMF/U (62), and EPS (46,600), but no detectable HIV RNA or DNA in any of their 12 prostate biopsy specimens.



Back to Top | Article Outline

HIV Levels and Correlations Among Specimens

The mean values of HIV RNA log10 copies/mL were not significantly different between paired seminal plasma and blood plasma (n = 15; mean [SD]: 4.296 [1.145] vs 4.593 [0.765]; t test P = 0.38; Fig. 2) or between paired UrF and EPS (n = 9; mean [SD]: 4.364 [1.063] vs 4.271 [0.906]; t test P = 0.86). However, mean viral RNA levels in post-PMF/U (n = 9; mean, 2.799; SD, 1.011) were higher than for pre-PMF/U (mean, 1.906; SD, 0.653; t test P = 0.011), with higher numerical values present in 8 of 9 subjects. We did not adjust for the estimated 10- to 100-fold volume dilution of the EPS by urine but chose to show the measured values. If we had adjusted for this dilution, then the HIV RNA level in post-PMF/U would have compared with the levels in semen and blood plasma, UrF, and EPS, rather than the lower level shown in Figure 2.



Prostate tissue viral RNA and DNA levels correlated (Table 3). The geometric mean HIV RNA level was detected at more than 50 copies/μg of total RNA for 26 (58%) of 45 prostate biopsy specimens from 11 (73%) of 15 men (1100 copies/μg; range, 59-165,000 copies/μg). Similarly, HIV DNA was detected at more than 10 copies/μg of total DNA for 7 (16%) of 45 specimens from 5 (33%) of 15 men (299 copies/μg; range, 86-3100 copies/μg).



Prostate tissue HIV DNA levels correlated with levels of HIV RNA in prostate tissue and pre-PMF/U and post-PMF/U. Prostate HIV RNA correlated strongly with post-PMF/U (Table 3). The semen WBC counts were associated with prostate tissue levels of HIV RNA (n = 14; r 2 = 0.46, P = 0.005) but not with HIV DNA or seminal HIV RNA levels. Prostate gland volume was not associated with prostate gland HIV nucleic acid levels, although there was a trend with seminal HIV RNA level (r 2 = 0.20, P = 0.06).

HIV RNA levels in post-PMF/U also correlated with those in blood plasma. Seminal plasma HIV RNA levels correlated with those in UrF and approached significance for post-PMF/U and prostate tissue HIV DNA. Post-PMF/U HIV RNA log10 copies per milliliter were lower for men on ART (mean, 1.69; SD, 0.30) than for those not on ART (n = 10; mean, 3.07; SD, 0.84; 1-way analysis of variance P = 0.06).

Back to Top | Article Outline

Predictors of Seminal HIV RNA Shedding

In linear regression analyses, only post-PMF/U and prostate tissue HIV DNA levels predicted seminal HIV RNA level after adjusting for either ART (Table 4) or blood plasma HIV RNA level (not shown). However, only post-PMF/U independently predicted seminal HIV RNA level in multiple linear regression models that included both post-PMF/U and prostate tissue HIV DNA (Table 5). Models with or without an adjustment for ART accounted for 35% of the variation in seminal plasma HIV RNA levels. We did not adjust for blood plasma HIV RNA level in multiple linear regression models because blood plasma HIV RNA was not associated with seminal plasma HIV RNA in this study population (Table 3).





Back to Top | Article Outline


Our study represents the most comprehensive anatomic evaluation of sources of seminal HIV to date, provides important new insights into the pathogenesis of HIV in the male genitourinary tract, and may offer an alternative approach for evaluating seminal HIV shedding in clinical trials. Contrary to our initial expectation, the prostate was not the major source of seminal HIV in these men without clinical urethritis or prostatitis.3 HIV RNA was detected in the wicked or swabbed UrF of all men sampled before prostatic massage; this fluid showed a weak association with, but was not an independent predictor of, seminal HIV RNA level. These findings suggest that HIV is shed continuously into the urethral lumen at levels comparable to seminal plasma levels. These observations support earlier studies suggesting a urethral source of seminal HIV.16,17

Although the prostate gland has been proposed as an important source of seminal HIV,1,4,18 our data suggest that other distal lower genitourinary tract structures were more important in men with no clinical evidence of urethritis or prostatitis. Potential sources of virus include seminal vesicles, periurethral (Littré) glands, bulbourethral (Cowper) glands, and periurethral submucosal mononuclear cells. Virus originating from such anatomic sites would likely involve mucosal and submucosal HIV-infected tissue macrophages or leukocytes, since there are no data to implicate urethral transitional epithelium per se.19 We evaluated distal genitourinary sources of virus by selective genitourinary fluid collection with analysis for HIV RNA and by prostate biopsy for HIV RNA and DNA. Further evaluations would require analysis of split ejaculates (bulbourethral glands), cannulation of the ejaculatory ducts (seminal vesicles and vasa deferentia), or biopsy and in situ detection methods for a more detailed tissue localization of virus in periurethral glands and submucosum.

Support for distal genitourinary HIV sources, other than or in addition to the prostate gland, arises from 4 findings. First, HIV RNA levels in post-PMF/U were not associated with those in EPS, as would be expected if HIV in post-PMF/U originated solely from prostatic fluid. This suggests another source of HIV in the proximal bulbous or membranous urethra or associated genitourinary structure other than the prostate gland.

Second, only the HIV RNA level in post-PMF/U was independently associated with seminal HIV RNA. Neither HIV RNA nor HIV DNA in prostate tissue was independently associated with seminal HIV RNA, as would be expected if the prostate gland represented the major source of seminal HIV. These observations again suggest that there were resident high levels of HIV in the proximal urethra and associated glandular structures in addition to whatever contribution arose directly from the prostate.

Third, additional support for a urethral source of HIV comes from sampling distal UrF. The level of UrF HIV RNA trended toward a significant association with seminal plasma HIV RNA level; moreover, the linear regression model β-coefficient had an upper confidence interval (CI) bound that was similar to that for post-PMF/U (Table 4). We obtained fluid from the posterior (membranous and prostatic) urethra during prostate massage but not the bulbous urethra; therefore, it is unresolved whether massage (milking) of the bulbous urethra before obtaining a UrF specimen would have increased the strength of the association between seminal and UrF HIV RNA levels (Tables 3 and 4).

Fourth, 2 (33%) of 6 men with complete genitourinary specimen collection had no detectable HIV in their Pbx despite having virus detected in all of the other genitourinary fluids. This further supports a significant contribution of nonprostatic sources of seminal HIV. Thus, there appears to be a substantial release of HIV into the distal genitourinary tract in conjunction with either pelvic floor relaxation associated with micturition (after vigorous prostatic massage) or with ejaculation.

There was a moderately strong correlation between prostate HIV RNA, DNA, and post-PMF/U HIV RNA; however, in the multiple regression analyses, prostate HIV nucleic acid levels did not contribute independently to seminal HIV RNA level. The measured prostate HIV nucleic acid levels were unlikely to represent blood contamination because of the exceptionally small interstitial/blood compartment volumes represented by each tissue biopsy and the weak correlation between prostate and blood HIV RNA levels (Table 3). The absence of an independent association between HIV levels in prostate and seminal plasma suggests that cell-free and cell-associated HIV is present largely in the prostate gland stroma and not in the prostatic glandular tissue or fluid per se; however, this distribution could be disrupted by prostatic inflammation.

The low frequency for detecting prostatic HIV DNA and the high frequency for detecting HIV RNA suggest that the prostate was not the most important reservoir for cell-associated HIV in our study population, although our data indicate that replication-competent HIV was present in the prostate tissue. The HIV DNA assay was modestly more sensitive than the HIV RNA assay, but different denominators were used to compare the 2 measurements. We were unable to explain the marked differences in the frequency of HIV nucleic acid detection based on assay sensitivity or efficiency of nucleic acid extraction; as such, there appear to be more cell-free HIV or HIV RNA transcripts in the prostate than there are infected cells as represented by viral DNA detection (intracellular).

Prostate HIV infection was likely focal and subject to biopsy sampling error despite obtaining a minimum of 6 biopsies per subject. The association between seminal WBC count and prostate tissue level of HIV RNA suggests that HIV infection of the prostate may contribute to subclinical prostatitis and leukospermia, although leukospermia itself was not associated with seminal plasma HIV RNA level. Conversely, it is also possible that asymptomatic inflammatory prostatitis (ie, National Institutes of Health category IV, which may be diagnosed by leukospermia) might increase prostate tissue HIV RNA levels.20-23

Based on a total biopsy volume of approximately 0.06 cm3 per subject, the biopsy samples represented only 0.4% of the median 16-cm3 prostate gland volume. To our knowledge, there is no information about the expected prostate size for HIV-infected men. Based on normative data, the age-related prostate volumes for our subjects are in keeping with those for men in the 20- to 50-year age group, although the median prostate size in our study appears to be lower than the reported 20-cm3 prostate volume from an autopsy study of 935 men.24 This difference may reflect the effect of HIV-associated hypogonadism on prostate volume. The correlation between prostate volume and semen volume and the negative correlation between semen volume and CD4 cell count (which may represent increased ejaculatory frequency and lower semen volume for those men with high CD4 cell counts compared with men with low CD4 cell counts) may have implications for the sexual transmission of HIV and suggest that issues of prostate health and physiology in HIV-infected men warrant further study.

Limitations of our study include the small sample size, incomplete sampling of some subjects, and dilution of HIV RNA in the various genital fluids evaluated. Nevertheless, the post-PMF/U HIV RNA level accounted for approximately one third of the observed variation in seminal plasma HIV RNA level and was consistent among the regression models evaluated. Regardless of the models used, the β-coefficient (relative change in viral RNA in semen) remained constant for post-PMF/U, suggesting that post-PMF/U HIV RNA level independently predicts seminal HIV RNA level despite the limited statistical power (Tables 4 and 5). Although prostatitis, urethritis, and other sexually transmitted infections are common in some HIV-infected populations, we did not include men with clinical evidence of urethritis or prostatitis in this study. The prostate and urethra may contribute more HIV to the semen in such situations.25,26

It is clear from this and other studies that the HIV RNA level in blood plasma is generally a poor clinical predictor of seminal HIV shedding in individual men despite the epidemiological association between serum HIV RNA level and sexual transmission.27-31 Among untreated subjects, the risk of HIV transmission through heterosexual contact increases 2.5-fold for each 10-fold increase in blood plasma HIV RNA level; however, the contribution of genital HIV RNA levels to this risk has been only assessed indirectly by modeling.29,31 The inability of the HIV RNA level in blood plasma to predict the level of HIV RNA in semen may be explained by our finding that HIV is compartmentalized anatomically and arises from distal genitourinary sources distinct from the testes, epididymides, and prostate.1,6 Preliminary phylogenetic analysis of genital and blood-associated HIV RNA and DNA from these subjects has provided further support for such compartmentalization.32 Thus, other factors associated with the proximal urethra may explain the overall shedding of HIV in semen and ART penetration into the ejaculate.26,33 Incorporation of either direct or indirect methods to measure seminal and lower genitourinary tract HIV RNA levels might improve on the important epidemiological association between blood HIV RNA level and sexual transmission of HIV.

Our finding that the post-PMF/U HIV RNA level was independently associated with the seminal HIV RNA level may have important implications for the design and execution of clinical studies to evaluate the effect of ART or HIV vaccines on seminal HIV shedding and sexual transmission. The most rigorous way to investigate these issues is to analyze semen directly, but these evaluations have been hampered by the technical difficulty in sampling the male genital tract for HIV. The collection of semen by masturbation is complicated and usually involves separate semen collection at or before clinic visits, logistical and transportation issues that delay the time between semen collection and testing if the semen is not collected on site, and important social and religious barriers to masturbation. Moreover, men with advanced HIV/AIDS disease often have decreased libido and semen volumes and may have neurological disease or drug-related interference with erection and ejaculation. Semen requires relatively rapid processing while it is still fresh and dilution to decrease viscosity for accurate pipetting.5,6 These issues complicate semen sample evaluation in almost all study settings.

To facilitate estimates of seminal HIV shedding, our data suggest that HIV RNA level in the voided urine collected after vigorous prostate massage (approximately 2-4 minutes following a standardized protocol) may serve as a proxy for direct measurement of seminal HIV. This approach may prove critical in evaluating the effect of ART on seminal HIV shedding in discordant couple studies of sexual transmission, where a relatively large sample size with repeated seminal evaluations would be required.

As an alternative to traditional semen evaluation for HIV, we suggest the following algorithm after a recommended abstinence interval of 2 to 7 days: (1) a urethral swab (UrF) or first-void urine (pre-PMF/U) or both are collected for evaluation of inflammation and sexually transmitted infections; (2) the prostate is next massaged until EPS is visualized in the urethral meatus; (3) finally, a post-PMF/U specimen is collected for HIV RNA quantification. The pre-PMF/U and post-PMF/U specimens can be centrifuged and stored frozen for later HIV RNA analysis. Collection, storage, transportation, and testing of post-PMF/U are substantially less complicated than direct semen evaluation. This algorithm requires validation in a clinical trial setting under field collection conditions but is offered as a potential approach to evaluate seminal HIV shedding. The acceptability of prostatic massage versus masturbation for semen collection has not, to our knowledge, been evaluated but may be population dependent.

In summary, the HIV RNA level in blood plasma is not reliable as an independent clinical predictor of virus levels in seminal plasma and extends the notion that the male genital tract is a distinct virological compartment from blood.3,6,27 We were surprised to find that the prostate was not the major source of seminal HIV in men without clinical urethritis or prostatitis. Other distal genitourinary sources of HIV, such as seminal vesicles, periurethral glands, bulbourethral glands, and periurethral submucosal mononuclear cells, are likely to be more important than the prostate gland. The post-PMF/U HIV RNA level accounted for approximately one third of the variation in seminal plasma HIV RNA level; this association was consistent among the linear regression models evaluated. Thus, analysis of the post-PMF/U may prove a suitable proxy for semen specimen evaluation in clinical trials.

Back to Top | Article Outline


Technical support: Carol Glenn (subject enrollment), Charles Muller (semen analysis), Jessica Giesler (prostate volumes), and Socorro Harb, Corey Scherrer, Pax Ortega, Reggie Gausman, Reggie Sampoleo, Alexis Motoshige, and Larry Stensland (retrovirology laboratory support).

Back to Top | Article Outline


1. Krieger JN, Nirapathpongporn A, Chaiyaporn M, et al. Vasectomy and human immunodeficiency virus type 1 in semen. J Urol. March 1998;159(3):430-438. [discussion 825-826].
2. Schwartz GG. Vasectomy and human immunodeficiency virus of mice and men. Fertil Steril. March 1991;55(3):650-651.
3. Paranjpe S, Craigo J, Patterson B, et al. Subcompartmentalization of HIV-1 quasispecies between seminal cells and seminal plasma indicates their origin in distinct genital tissues. AIDS Res Hum Retroviruses. November 20, 2002;18(17):1271-1280.
4. Smith DM, Kingery JD, Wong JK, et al. The prostate as a reservoir for HIV-1. AIDS. July 23, 2004;18(11):1600-1602.
5. Coombs RW, Krieger JN, Collier AC, et al. Plasma viremia and recovery of HIV from semen: implications for transmission and therapy. In: Mélica R, ed. 1st International Symposium on AIDS and Reproduction. Basel: Karger, 1992:145-149.
6. Coombs RW, Speck CE, Hughes JP, et al. Association between culturable human immunodeficiency virus type 1 (HIV-1) in semen and HIV-1 RNA levels in semen and blood: evidence for compartmentalization of HIV-1 between semen and blood. J Infect Dis. 1998; 177(2):320-330.
7. Krieger JN, Egan KJ. Comprehensive evaluation and treatment of 75 men referred to chronic prostatitis clinic. Urology. July 1991;38(1): 11-19.
8. Krieger JN, Coombs RW, Collier AC, et al. Fertility parameters in men infected with human immunodeficiency virus. J Infect Dis. September 1991;164(3):464-469.
9. Krieger JN, Coombs RW, Collier AC, et al. Seminal shedding of human immunodeficiency virus type 1 and human cytomegalovirus: evidence for different immunologic controls. J Infect Dis. April 1995;171(4): 1018-1022.
10. Muller CH, Coombs RW, Krieger JN. Effects of clinical stage and immunological status on semen analysis results in human immunodeficiency virus type 1-seropositive men. Andrologia. 1998;30(suppl 1):15-22.
11. Zuckerman RA, Whittington WL, Celum CL, et al. Factors associated with oropharyngeal human immunodeficiency virus shedding. J Infect Dis. July 1, 2003;188(1):142-145.
12. Berger RE, Krieger JN, Kessler D, et al. Case-control study of men with suspected chronic idiopathic prostatitis. J Urol. February 1989;141(2): 328-331.
13. Ellis WJ, Brawer MK. Repeat prostate needle biopsy: who needs it? J Urol. May 1995;153(5):1496-1498.
14. Melchior SW, Brawer MK. Role of transrectal ultrasound and prostate biopsy. J Clin Ultrasound. October 1996;24(8):463-471.
15. Long JS, Ervin LH. Using heteroscedasticity consistent standard errors in the linear regression model. Am Stat. 2000;54:217-224.
16. Ilaria G, Jacobs JL, Polsky B, et al. Detection of HIV-1 DNA sequences in pre-ejaculatory fluid. Lancet. December 12, 1992;340(8833):1469.
17. Pudney J, Oneta M, Mayer K, et al. Pre-ejaculatory fluid as potential vector for sexual transmission of HIV-1. Lancet. December 12, 1992; 340(8833):1470.
18. Coombs RW, Reichelderfer PS, Landay AL. Recent observations on HIV type-1 infection in the genital tract of men and women. AIDS. March 7, 2003;17(4):455-480.
19. Mullen TE Jr, Kiessling RL, Kiessling AA. Tissue-specific populations of leukocytes in semen-producing organs of the normal, hemicastrated, and vasectomized mouse. AIDS Res Hum Retroviruses. March 2003; 19(3):235-243.
20. Krieger JN, Jacobs RR, Ross SO. Does the chronic prostatitis/pelvic pain syndrome differ from nonbacterial prostatitis and prostatodynia? J Urol. November 2000;164(5):1554-1558.
21. Krieger JN, Nyberg L Jr, Nickel JC. NIH consensus definition and classification of prostatitis. JAMA. July 21, 1999;282(3):236-237.
22. Krieger JN, Ross SO, Deutsch L, et al. Seminal fluid analysis in chronic prostatitis/chronic pelvic pain syndrome. Andrologia. October 2003; 35(5): 266-270.
23. Nickel JC. Classification and diagnosis of prostatitis: a gold standard? Andrologia. June 2003;35(3):160-167.
24. Berry SJ, Coffey DS, Walsh PC, et al. The development of human benign prostatic hyperplasia with age. J Urol. September 1984;132(3): 474-479.
25. Cohen MS, Hoffman IF, Royce RA, et al. Reduction of concentration of HIV-1 in semen after treatment of urethritis: implications for prevention of sexual transmission of HIV-1. AIDSCAP Malawi Research Group. Lancet. June 28, 1997;349(9069):1868-1873.
26. Winter AJ, Taylor S, Workman J, et al. Asymptomatic urethritis and detection of HIV-1 RNA in seminal plasma. Sex Transm Infect. 1999; 75(4):261-263.
27. Zuckerman RA, Whittington WL, Celum CL, et al. Higher concentration of HIV RNA in rectal mucosa secretions than in blood and seminal plasma, among men who have sex with men, independent of antiretroviral therapy. J Infect Dis. July 1, 2004;190(1):156-161.
28. Ragni MV, Faruki H, Kingsley LA. Heterosexual HIV-1 transmission and viral load in hemophilic patients. J Acquir Immune Defic Syndr Human Retrovirol. January 1, 1998;17(1):42-45.
29. Quinn TC, Wawer MJ, Sewankambo N, et al. Viral load and heterosexual transmission of human immunodeficiency virus type 1. Rakai Project Study Group. N Engl J Med. 2000;342(13):921-929.
30. Gray RH, Wawer MJ, Brookmeyer R, et al. Probability of HIV-1 transmission per coital act in monogamous, heterosexual, HIV-1-discordant couples in Rakai, Uganda. Lancet. April 14, 2001; 357(9263): 1149-1153.
31. Chakraborty H, Sen PK, Helms RW, et al. Viral burden in genital secretions determines male-to-female sexual transmission of HIV-1: a probabilistic empiric model. AIDS. March 30, 2001; 15(5): 621-627.
32. Diem K, Motoshige A, Fox A, et al. Phylogenetic comparisons of HIV-1 between the male urogenital tract and blood compartments. Paper presented at: 44th Interscience Conference on Antimicrobial Agents and Chemotherapy [abstract H-196]; October 30 to November 2, 2004; Washington, DC.
33. Taylor S, Pereira AS. Antiretroviral drug concentrations in semen of HIV-1 infected men. Sex Transm Infect. 2001;77(1):4-11.

AIDS; HIV; RNA; DNA; semen; prostate gland; compartmentalization; urethra; sexual transmission; male

© 2006 Lippincott Williams & Wilkins, Inc.