Schacker T, Ryncarz AJ, Goddard J, Diem K, Shaughnessy M, Corey L. Frequent Recovery of HIV-1 From Genital Herpes Simplex Virus Lesions in HIV-1–Infected Men. JAMA. 1998;280(1):61-66. doi:10.1001/jama.280.1.61
From the Departments of Medicine (Drs Schacker and Corey) and Laboratory Medicine (Drs Ryncarz and Corey, Messrs Goddard and Diem, and Ms Shaughnessy), University of Washington, and the Program in Infectious Diseases, Fred Hutchinson Cancer Research Center (Drs Ryncarz and Corey), Seattle, Wash. Dr Schacker is now with the Division of Infectious Diseases, Department of Medicine, University of Minnesota, Minneapolis.
Context.— Genital ulcer disease has been epidemiologically linked as a risk factor
in the transmission of the human immunodeficiency virus 1 (HIV-1). While herpes
simplex virus 2 (HSV-2) is the most common cause of genital ulcers, no study
has systematically evaluated the frequency or titer of HIV-1 virus in HSV-2
Objective.— To compare lesional HIV-1 RNA levels during and after genital HSV-2
reactivation and to evaluate the frequency, titer, and duration of HIV-1 RNA
shedding in lesions due to HSV-2.
Design.— Convenience sample.
Setting.— Sexually transmitted disease research clinic at the University of Washington,
Patients.— Twelve HIV-infected men with a history of symptomatic HSV-2 infection
who underwent daily sampling of genital lesions for HIV-1 RNA by polymerase
chain reaction assay and HSV-2 by culture.
Main Outcome Measure.— Detection of lesional HIV RNA and HSV-2.
Results.— HIV-1 RNA was detected from lesional swabs in 25 of 26 consecutively
studied HSV-2 episodes and on 67% of days in which genital lesions were noted.
The HIV-1 RNA titers in lesional swabs exceeded 10000 copies/mL of swab sample
in 75% of samples (range, 2.2-3.2×105 copies/mL of swab sample).
HIV-1 RNA in genital lesion swabs was seen in persons with high and low titers
of plasma HIV-1 RNA and was not associated with plasma HIV-1 RNA levels.
Conclusions.— HIV-1 virions can consistently be detected in genital ulcers caused
by HSV-2, which suggests that genital herpes infection likely increases the
efficiency of the sexual transmission of HIV-1.
SEVERAL epidemiologic studies have shown an association between the
acquisition of the human immunodeficiency virus 1 (HIV-1) and the reported
presence of genital lesions in the presumed source contact.1- 6
In 1 study of discordant couples with HIV-1 infection, presence of a genital
ulcer in the source partner was associated with a 5-fold greater risk in the
transmission of HIV-1.7 Worldwide, the most
common infectious etiologies of genital ulcers are herpes simplex virus (HSV),
syphilis, and chancroid,8,9 with
HSV the most common in North America and Europe, accounting for 75% to 80%
of infectious genital ulcers.10 One of the
first reports of a common source cluster of sexually transmitted HIV-1 infection
in Europe emanated from an HIV-1–infected individual who on retrospective
questioning was noted to have genital HSV-2 infection.4
Recent studies have shown that genital herpes is the most frequent sexually
transmitted disease among HIV-1–seropositive persons,11
and most HSV-2–seropositive persons (whether HIV-1 infected or immunocompetent)
intermittently reactivate the virus on mucosal surfaces.12- 14
Clinically, HSV-2 causes lesions that are smaller and of shorter duration
than those caused by chancroid or syphilis, and HSV-2 ulcerations often occur
in locations difficult to visualize, such as the cervix or perineal region.13,15 As such, the true prevalence of clinically
described genital lesions due to HSV among persons suspected of transmitting
HIV-1 is likely to be underestimated. We undertook a study to evaluate the
frequency, titer, and duration of HIV-1 in genital HSV-2 lesions; to relate
lesional titers of HIV-1 RNA to plasma titers of HIV-1 RNA; to examine the
association between lesional HIV-1 RNA with HSV-2 in culture; and to compare
HIV-1 quasi species in lesions and plasma. As the HIV-1 epidemic in Seattle,
Wash, is largely present in the homosexual/bisexual male population, our study
enrolled gay men.
We recruited patients who were HIV-1 positive and had HSV-2 antibodies
by Western blot assay through newspaper advertisements and referrals from
local physicians into a protocol approved by the University of Washington
Institutional Review Board. Patients were excluded at the screening visit
if they had been receiving daily anti-HSV therapy in the previous 30 days.
Of 33 persons referred to our clinic for the study, we enrolled 12. The reasons
for lack of participation included desire to use antiviral therapy for herpes
recurrences (n=8), clinically unstable HIV-1 disease that required immediate
use of other medications or new antiretrovirals (n=7), or the patient felt
the frequent visits to the clinic during HSV recurrences were too time-consuming
(n=6). We enrolled all patients who fit our serologic and clinical criteria.
The study was conducted in 1995, when many HIV-1–positive patients were
untreated or receiving monotherapy.
Upon entry, informed consent was obtained, and a standardized questionnaire
describing the subject's history of HSV and HIV-1 infection was administered.
Blood was obtained at entry for HSV and HIV-1 antibody assays, plasma HIV-1
RNA, and CD4 cell counts. Patients returned to the clinic within 24 hours
of lesion onset and at every-other-day intervals until healed. At each visit
a genital examination was done, including measurement of the lesion.16 Swabs of the lesion for HSV culture and HIV-1 polymerase
chain reaction (PCR) assay were obtained at each clinic visit. Patients were
also allowed to obtain cultures for HSV-2 at home on the days between visits.
Home cultures for HIV-1 were not obtained, as the medium for HIV-1 RNA isolation
contained guanidinium thiocyanate. The last 6 patients also collected swab
samples from the genital region after the lesions healed.17
Lesional swabs for HIV-1 RNA were obtained by rubbing a Dacron swab
over the base of the lesion and immediately placing it in 500 mL of guanidinium
thiocyanate denaturing buffer for transportation and storage. Samples were
stored at−70°C within 3 hours of collection. A separate swab for
HSV culture was obtained and placed into 1 mL of standard viral transport
medium.18 In 2 patients, a separate lesional
swab from the initial and subsequent 3 clinic visits during an HSV-2 reactivation
was also collected for HIV-1 isolation.19
HSV Serology, Isolation, and PCR. The HSV antibody determinations were performed by Western blot analysis.15 Isolation of HSV from lesional swabs and detection
of HSV DNA in lesional swabs by PCR were performed as previously described.20
HIV—1 RNA Polymerase Chain Reaction and Culture. Plasma HIV-1 RNA was assayed using the branched chain DNA method (Chiron
Diagnostics, Emeryville, Calif). This assay was performed according to the
manufacturer's specifications and at the time of this study had a lower limit
of detection of 10000 copies/mL.21,22
The HIV-1 RNA levels in HSV lesions were determined by reverse transcriptase–polymerase
chain reaction (RT-PCR) with gag-specific primers
and oligonucleotide probes using a modification of the assay described by
Piatak et al.23 Briefly, total RNA was extracted
and complementary DNA (cDNA) was synthesized from the purified lesion RNA
using random hexamers, prime inhibitor (5 Prime to 3 Prime), and Superscript
II (SSII) (Gibco-BRL) reverse transcriptase in a SSII 1×reaction buffer.
The HIV-1 gag- specific primers (gag 4 and gag 6) were used to amplify (45
cycles) the HIV-1 cDNA.
Amplified HIV-1 gag sequences were detected
by liquid hybridization using a phosphorus 32–labeled oligonucleotide
probe (5′-CCAGGCCAGATGAGAGAACCAAGGG-3′) specific for a conserved
internal region of the amplified HIV-1 gag PCR product.
Electrophoresis was performed with 20 µL of the hybridization products
in a 6% polyacrylamide gel. Visual inspection of the autoradiographs, along
with comparisons to known amounts of amplified HIV-1 cDNA, were used to determine
the approximate original number of HIV-1 RNA copies in the lesion swab. Each
autoradiograph band signifying a positive PCR assay was assigned a score of
1 to 4+approximating the concentrations of a dilution curve containing from
5 to 5000 copies of HIV-1 RNA, respectively. Each cDNA and PCR reaction contained
both positive and negative controls. Known amounts of HIV-1 RNA and cDNA were
used for positive controls and for quantification. All samples that were PCR
negative for HIV-1 were confirmed to be true negatives, not attributable to
nonspecific inhibition of the reaction by performing an additional PCR with
103 copies of HIV-1 cDNA. Samples that were unable to support amplification
of the input substrate were denoted as inhibitory. All others were reported
as samples void of HIV-1 RNA. For calculations estimating HIV-1 copies per
200 µL of specimen by PCR, samples with titers between 500 and 5000
were given a titer of 2500, those between 50 and 500, a titer of 250, and
those below 50, a titer of 25.
Heteroduplex Tracking Assays. Heteroduplex tracking assays (HTAs) to HIV-1 envelope were performed
to compare viral quasi species in genital lesions and plasma.24- 27
Nested PCR primers and amplification conditions were performed as previously
described.24,25 Primer ED-12 was
used for reverse transcription. Primers ED-5 and ED-12 were used for the first
round and primers ES-7 and ES-8 for the second round of PCR amplification.26,27
Probes were generated and then reannealed with the PCR products. The
resulting reactions were run on 6% nondenaturing polyacrylamide gels, autoradiographed,
scanned using an Agfa Arcus II scanner, and analyzed with National Institutes
of Health image software.
Plasma and Lesion Cloning. The ES-7– and ES-8–derived PCR products from lesion samples
or plasma samples in which HSV-1 cDNA was detected were ligated into the PCR
2.1 vector (Invitrogen, Carlsbad, Calif) and transformed into Escherichia coli with blue/white screening according to the manufacturer's
directions. Plasmid DNA was isolated by standard technique. Clones were confirmed
by PCR with ES-7 and ES-8 primers.
The PCR products were cloned into a PCR 2.1 vector and then reassayed
in the HTA to ensure their similar mobility to the parent clone.
Statistical analyses were performed using χ2 and nonparametric
methods as described.
We studied 12 male subjects; their mean age was 39 years (range, 30-56
years), and all were men who had sex with men. The CD4 cell counts at entry
ranged from 0.018 to 0.541×109/L (18-541/µL); 3 of
the 12 were receiving antiretroviral therapy at entry (Table 1). All 12 patients were HSV-2 seropositive, 10 (83%) of 12
also had HSV-1 antibodies. While 9 reported a history of clinically diagnosed
genital or anal HSV at the enrollment visit, the other 3 subjects who denied
any history of clinically diagnosed genital herpes reported intermittent episodes
of genital ulcerations that would spontaneously heal within 5 to 14 days.
No patient was receiving suppressive therapy for HSV infection, and none routinely
used antiviral therapy for their herpetic recurrences. The median plasma HIV-1
RNA level at entry was 35000 copies/mL (range, 13000-1600000 copies/mL) (Table 2).
The 12 patients were followed up through 26 episodes of HSV-2 reactivation;
8 through a single HSV-2 episode, 1 during 2 episodes, 1 during 4 episodes,
and 2 during 6 consecutive episodes. The anatomic site of reactivation of
the 26 episodes was penile shaft in 4 episodes (15%), perirectal or inner
gluteal fold region in 19 episodes (73%), facial region in 1 episode (4%),
and the outer buttock or lumbosacral region in 2 episodes (8%). The size of
lesions (median, 47.5 mm2) and the duration
of the episode (median, 15 days) were typical for genital lesions for mildly
immunocompromised persons.13 HSV-2 was isolated
from the lesions in 24 (92%) of the 26 episodes; the 2 HSV culture-negative
episodes had HSV-2 DNA detected by PCR in the lesional swab at the initial
visit for that episode. Thus, all 26 episodes sampled were due to HSV-2; of
the 26 episodes, 23 healed without antiviral therapy.
A total of 175 lesional swab samples were collected at the clinic visits
(median, 5.5 samples per episode; range, 2-18 samples per episode). Five lesional
swabs from 5 separate persons repeatedly inhibited the PCR reaction and were
thus not evaluable for HIV-1 RNA detection. HIV-1 RNA was detected in lesional
swabs in 25 of the 26 episodes of genital herpes (Table 2), and HIV-1 RNA was detected in 108 (64%) of the 169 assayable
lesional swab samples. The median percentage of days that HIV-1 RNA was detected
from lesional swabs was 66.7%; HIV-1 RNA was detected at a level of 10000
copies/mL or higher in 75% of the of HIV-1 RNA–positive lesional swab
samples. To estimate the maximal titer of HIV-1 RNA in the lesion samples
with more than 10000 copies/mL in the initial screening assay, additional
10-fold dilution of the cDNA isolated from the swabs was performed on 8 samples
from 6 patients who had a lesional swab titer of more than 5000 copies/200
µL of swab specimen. The titer of HIV-1 RNA per milliliter of swab sample
in the 8 samples ranged from 20000 to 320000 copies/mL of swab specimen (median,
40000 copies/mL) with 3 of the samples having end point titers between 2.2×105 and 3.2×105 copies of HIV-1 in the lesional swab
samples. These titers of HIV-1 RNA in the lesional swab specimens exceeded
the patient's plasma RNA taken at the same time by at least 2-fold in 4 of
the 6 patients.
A representative autoradiogram from the lesional swabs is shown in Figure 1. The PCR reaction was performed
both with and without the addition of RT. Even among samples in which the
level of HIV-1 RNA was greater than or equal to 5000 copies/200 µL of
lesion swab material, a signal was detected only in the reaction in which
RT was placed in the sample, indicating that all the PCR signal was attributable
to the presence of virion RNA (Figure 1).
Proviral DNA was not detected in any of the 14 HIV-1 RNA–positive samples
run in parallel with and without RT.
HIV-1 RNA was detected at equal frequency in all anatomic areas sampled,
and all 4 patients who were sampled through multiple recurrences of HSV-2
had HIV-1 RNA detected from lesional swab samples in all the recurrences.
A representative autoradiogram of the lesion samples from sequential episodes
of 1 of these 4 patients is shown inFigure
We next evaluated whether HIV-1 RNA could be detected on intact genital
skin from a recently healed HSV-2 episode. We tested 9 swabs obtained from
3 of the 6 patients who collected these samples after lesional healing. All
samples were negative for HIV-1 by viral isolation, and 8 of the 9 swabs from
intact skin were also negative for HIV-1 RNA by PCR. One swab had approximately
10 copies of HIV-1 RNA detected in the swab sample obtained from a site where
a herpes lesion had healed within the previous 4 days.
Swabs were submitted for HIV-1 isolation in cell culture in 2 patients
(4 swabs each). No HIV-1 was isolated from any of the lesional swabs. HIV-1
RNA was, however, detected in all 8 of these same samples (4 swabs were associated
with HIV-1 RNA titers ≥5000 copies/200-µL sample, and 4 were associated
with titers of between 500-5000 copies/200-µL sample).
HIV-1 RNA was detected in lesions from persons with relatively low titers
of HIV-1 RNA in plasma (patients 4 and 11), those with high HIV-1 RNA titers
(patients 1, 2, and 6), and from persons taking and not taking antiretroviral
therapy (Table 2).
Detection of HIV-1 RNA in lesional swabs was significantly associated
with isolation of HSV-2 by culture. HIV-1 RNA was present in 63 (80%) of the
79 swabs taken on days from which HSV-2 was also isolated vs 45 (50%) of the
90 lesional swab samples on days from which HSV-2 was not isolated (odds ratio
[OR], 4.6; 95% confidence interval [CI], 1.8-8.7; χ2; P<.01). Figure 3
illustrates the HSV-2 and HIV-1 RNA shedding in genital lesion swabs during
the course of a genital lesion. High titers of HSV DNA and HIV-1 RNA were
detected in lesional swabs early in the episode. By day 6, within 2 days after
the initiation of acyclovir (400 mg orally every 8 hours), HSV-2 was no longer
isolated from the lesion, and by day 11 HSV DNA as detected by PCR was no
longer detected. The HIV-1 RNA titer in genital lesions persisted at high
levels until HSV replication as detected by PCR fell to low levels and lesion
healing occurred. HIV-1 RNA became undetectable once the genital lesion reepithelialized.
We also evaluated the genetic relationship of the HIV-1 RNA detected
in lesional swabs with plasma RNA. For these analyses we took samples from
patient 4 and used HTA to compare the viral quasi species in the genital lesion
with those in plasma samples collected at the time of reactivation and shortly
thereafter. Initially, we cloned HIV-1 envelope sequences from the lesional
swab sample on the initial day of HSV reactivation. For these studies we used
the ES-7 and ES-8 probes, which span a 700–base pair region of the HIV-1
envelope between regions C2-V5, including the V3 loop region. Two variants
were identified in the lesional swab: a predominant variant and a minor variant
labeled variant 2. Variant 2 was cloned, radiolabeled, and used in an HTA
assay of sequential serum taken after the HSV reactivation. Figure 4 shows increasing concentrations of variant 2 between March
10, the initial onset of the lesion, and May 2, which was 52 days later, 45
days after healing of the lesions. Using Image Analysis software, we found
that variant 2 increased 2.5-fold in plasma over the 7-week follow-up period.
These data were corroborated by cloning of variant 2 from plasma on March
10 and May 2. Of 18 clones derived from the plasma obtained, on March 10,
the day of onset of the lesion, 2 of 18 were variant 2. On May 2, five of
the 18 clones derived from plasma were variant 2. Thus, by both methods a
2.5-fold increase in variant 2 was seen. These data suggest that variant 2
found in genital lesions is a replication competent clone of HIV-1.
We consistently detected HIV-1 RNA by both gag
and envelope primers in HSV-2 lesions. The HIV-1 nucleic acid we detected
in lesional swabs was virion associated, underscoring the infectious potential
of these lesions. Heteroduplex tracking assay demonstrated that an HIV-1 variant
present in the lesional swab specimen was present in increasing amounts in
plasma over the course of follow-up, suggesting that replication competent
virions of HIV-1 can persist in herpetic lesions.
Most genital herpes episodes were associated with at least 1 specimen
with an HIV-1 RNA titer of more than 5000 copies/200 µL of lesional
swab specimen. In some specimens, end point dilution analysis revealed titers
of more than 2×105 copies/mL of lesional swab sample. The
titer of HIV-1 that is associated with transmission of HIV-1 during sexual
activity is unknown. The presence of HIV-1 in semen is also an important component
in the transmission of HIV-1 infection, and it is clear that HIV-1 transmission
can occur without genital lesions. Thus, the relative contribution of HIV-1
shedding from genital lesions vs that from semen cannot be determined from
our study. However, epidemiologic data support the intermittent nature of
transmission of HIV-1 among stable sexual partners.28
The "bursts" of HIV-1 virions that accompany genital lesions may help explain
the variable transmission rates of HIV-1 among couples.28
The "clusters" of rapidly spreading HIV-1 that occur in association with source
contacts with reported genital lesions suggest that shedding of HIV-1 in such
lesions may be an underappreciated factor in the overall efficiency of sexual
transmission of HIV-1.1- 7
Tetanus toxoid, influenza, and pneumococcal immunizations transiently
increase plasma HIV-1 RNA concentrations via in vivo activation of T cells
from antigenic stimulation.29- 32
Our data support the hypothesis that antigenic stimulation on mucosal sites
by reactivation of HSV, an unrelated infectious pathogen, can potentially
increase HIV-1 replication on mucosal surfaces. This concept is underscored
by the consistency of our findings, irrespective of the anatomic site of the
HSV-2 lesions we sampled. While herpetic lesions often have serosanguinous
fluid in them, raising the possibility that HIV-1 RNA detected was "contamination"
from blood, we found no correlation between frequency of detection of HIV-1
in genital lesions and HIV-1 plasma RNA titers. As shown in Table 2, high titers of HIV-1 in lesional swabs were not related
to levels in plasma, and subsequent reduction of HIV-1 RNA in genital lesions
was not related to plasma RNA levels (Figure
3). The correlation between the detection of HIV-1 RNA in lesional
swabs and the ability to isolate HSV-2 in the lesion along with the prompt
reduction in HIV-1 RNA and HSV-2 titer in mucosal swabs associated with acyclovir
therapy support the hypothesis that HSV-2 reactivation may play an important
role in the titer of HIV-1 on mucosal surfaces.
The HSV-infected lesions are associated with an influx of CD4-bearing
lymphocytes that carry activation markers.33,34
Activation of latently infected CD4 cells has been shown to markedly up-regulate
HIV-1 replication,35 and several HSV regulatory
proteins up-regulate HIV-1 replication through their interaction with the
HIV-1 long terminal repeat region.36- 40
Coinfection of HSV and HIV-1 in lymphocytes has been described both in vitro
and in vivo.41,42 Thus, the high
titers of virion HIV-1 that we detected in genital herpetic lesions may be
the result of an influx of activated, HIV-infected CD4-bearing cells into
the mucosal ulceration, as well as a potential in vivo interaction between
the 2 viral agents. Additional evidence of this biologic interaction of the
2 agents is that similar results were seen irrespective of the anatomic site
of the HSV-2 reactivation. We were unable to demonstrate presence of proviral
DNA in the lesion. However, our extraction methods were designed to maximize
recovery of RNA from lesional swabs, and it is possible that some provirally
infected cells were present in genital lesions, but escaped detection.
The duration, extent, and clinical severity of the HSV reactivations
in the patients we studied did not differ from persons with similar CD4 cell
counts in other studies of the natural history of genital herpes in HIV-1–infected
persons.12,13 If anything, our
population was skewed to persons with milder disease who did not feel antiviral
therapy of their lesions was necessary or whose health care providers did
not routinely prescribe antiviral therapy. Thus, we feel it likely our results
would apply to most episodes of HSV-2 reactivation among HIV-1–infected
persons. The association that we saw between treatment of the HSV episode
with acyclovir and reduction of HIV-1 expression in the lesion warrants further
study. For, if the findings are consistent, they suggest more attention is
warranted to the treatment of genital ulceration due to HSV-2, especially
among persons with early HIV infection.