Human Immunodeficiency Virus Type 1 Shedding Pattern in Semen Correlates
with the Compartmentalization of Viral Quasi Species between Blood
Phalguni Gupta,1,aCaroline Leroux,2,a
Bruce K. Patterson,4Lawrence Kingsley,1
Charles Rinaldo,1,3Ming Ding,1Yue Chen,1
Kathy Kulka,1William Buchanan,1Brian McKeon,2
and Ronald Montelaro2
School of Public Health) and
Northwestern University Medical School, Chicago, Illinois
1Infectious Diseases and Microbiology (Graduate
2Molecular Genetics and Biochemistry
3Pathology (School of Medicine), University of Pittsburgh,
4Department of Gynecology and Obstetrics,
High levels of human immunodeficiency virus (HIV) type 1 have been detected in semen
at all stages of disease. However, it is not clear whether HIV-1 is shed in semen continuously
or intermittently. In a prospective longitudinal study, viral RNA was measured weekly for 10
weeks in semen and blood of HIV-seropositive subjects. Results showed three different pat-
terns of HIV-1 shedding in semen: none (28%), continuous (28%), and intermittent (44%). In
contrast, there was no change in blood plasma virus load during the studyperiod.Phylogenetic
analysis of the envelope sequences of HIV-1 RNA in semen and blood revealed distinct virus
populations in semen and blood of intermittent shedders but similar virus populations in the
semen and blood of continuous shedder. These results indicate for the first time that HIV-1
is shed primarily in an intermittent manner and that shedding patterns of HIV-1 in semen
are related to compartmentalization of HIV-1 between semen and blood.
Human immunodeficiency virus (HIV) type 1 infection in
adults is transmitted predominantly by sexual routes. Semen is
considered to be the major vehicle for such transmission, since
it contains free infectious HIV-1 and/or HIV-1–infected cells
[1–6]. In situ hybridization, immunohistologic, and polymerase
chain reaction (PCR) studies have demonstrated that the prin-
cipal types of HIV-1–infected cells in semen are lymphocytes
and macrophages [3, 7]. The concentration of HIV-1–infected
cells in semen varies substantially, ranging from 0.01% to 5%
of white cells [8, 9]. However, if venereal disease is present,
many more inflammatory cells (and thus virus-infected cells)
are present in seminal plasma . Detection of high levels of
cell-free HIV-1 [10–12] and relatively low levels of HIV-1–in-
fected cells in semen at all stages of disease [2, 8, 9] suggest
that HIV-1 transmission occurs primarily via cell-freevirus.
This hypothesis is supported by the fact that cell-free simian
immunodeficiency virus (SIV) is more efficient in initiating in-
Received 13 January 2000;revised30March2000;electronicallypublished
29 June 2000.
Presented in part: 7th Conference on Retroviruses and Opportunistic In-
fections, San Francisco, January 2000.
Informed consent was obtained from patients in accordance with Uni-
versity of Pittsburgh institutional review board human experimentation
Grant support: NIH (AI-35041, HD-32256, AI-46271).
aP.G. and C.L. contributed equally to this work.
Reprints or correspondence: Dr. Phalguni Gupta, Dept. of Infectious
Diseases and Microbiology, Graduate School of Public Health, University
of Pittsburgh, Pittsburgh, PA 15261 (firstname.lastname@example.org).
The Journal of Infectious Diseases
? 2000 by the Infectious Diseases Society of America. All rights reserved.
trarectal infection than cell-associated SIV . However,treat-
ment with potent antiretroviral therapy reduces virus load in
semen in parallel with that in blood [11, 14].
There is evidence that in some infected subjects HIV-1 var-
iants in semen differ from those in blood. Thisevidenceincludes
the following: a difference in envelope and protease gene se-
quences of HIV-1 proviral DNA in semen and blood [15–17];
lack of correlation between the seminal HIV-1 RNA concen-
tration and plasma virus load[10,18,19];detectionofinfectious
HIV-1 in semen in the absence of plasma viremia [5, 10]; dis-
cordance of the biologic properties of HIV-1 from semen and
blood with respect to syncytia-inducing (SI) properties; and
resistance to reverse transcriptase (RT) inhibitory drugs [5, 10,
20]. These data indicate that in these subjects, HIV-1 in semen
may not originate from blood. In contrast, in other infected
subjects, identical viral variants are present in both blood and
semen [16, 17], indicating a common origin of HIV-1 in these
two body compartments in these subjects.
Although HIV-1 RNA and DNA have been detected in se-
men of HIV-1–infected subjects, it is not clear whether the virus
is shed continuously or intermittently in semen. Intermittent
shedding could explain a lack of detection of HIV-1 in semen
in some infected subjects at different stages of disease .
Furthermore, it is not known whether the shedding pattern is
related to the compartmentalization of HIV-1 between semen
and blood. Krieger et al.  reported intermittent shedding of
HIV-1 in semen as detected by isolation of virus in cell culture.
However, the low sensitivity of the culture technique makes it
difficult to interpret these data. Here we describe a prospective
longitudinal study on the dynamics of HIV-1 shedding insemen
by guest on February 2, 2016
80Gupta et al. JID 2000;182 (July)
virus intermittently; B, those who shed virus persistently; and C, nonshedders.
Human immunodeficiency virus type 1 RNA measurements in semen and blood plasma in 3 groups of subjects: A, those who shed
over a 10-week period and a comparative analysis of the viral
variants found in semen and blood in subjects with different
Study population, semen, and blood samples.
ulation comprised 18 HIV-1–infected subjects from the Pittsburgh
portion of the Multicenter AIDS Cohort Study. Participants were
asymptomatic, with a median of 434 CD4 T cells (range, 117–935),
and were not receiving any potent antiretroviral therapy (e.g., pro-
tease inhibitors) during the study period. Paired semen and he-
parinized blood samples were collected at weekly intervals for 10
weeks. Subjects did not engage in sexual activity for 48–72 h before
donation of semen and did not have other sexually transmitted
Samples were processed within 4 h of collection. Blood plasma
The study pop-
was prepared by centrifuging heparinized blood at 1200 g at room
temperature. Seminal cells were separated from whole semen by
centrifugation for 10 min at 800–1000 g at room temperature. Su-
pernatant (seminal fluid) was stored frozen at ?70?C. Seminal cells
were washed once with Hanks’ balanced salt solution (HBSS), re-
suspended in 5 mL of HBSS, and then subjected to ficoll-hypaque
gradient centrifugation. Seminal mononuclear cells were collected
at the interface, washed, and stored frozen in dimethyl sulfoxide
at ?130?C until use. Blood CD4 cells were counted at the first and
the last visit (visit 10) by our standard flow-cytometric measure-
ment , since no change in CD4 cell number was expected in a
Quantitation of HIV RNA in semen and blood.
HIV-1 RNA in whole semen and blood plasma were quantitated
by a commercial nucleic acid sequence–based amplification assay
(Nuclisens) according to the manufacturer’s (Organon Teknika,
Durham, NC) protocol. This is a modified version of a NASBA
(nucleic acid sequence–based amplification assay) [11, 23–25]. The
The levels of
by guest on February 2, 2016
JID 2000;182 (July) HIV Shedding Pattern in Semen 81
ciency virus (HIV) gag-pol RNA-positive cells in seminal mononuclear
cells from subject 1 who shed HIV-1 intermittently. Data are from visits
2, 4, 7, and 10 (V2–V10). ECD, energy-coupled dye; FITC, fluorescein
Flow cytometry analysis of CD4?human immunodefi-
ficiency virus (HIV) transcription in seminal mononuclear cells.
Correlation of seminal virus load with human immunode-
Seminal mononuclear cells
aPercentage of CD4?cells that are CD4?CD45RO?and CD4?HIV gas-pol
assay had a lower limit of detection of 200–400 copies of HIV-1
RNA/mL and a linear range of detection up to
In situ hybridization and immunostaining of seminal mononuclear
Simultaneous flow cytometric analysis of intracellularHIV-
1 RNA and cellular immunophenotyping of seminal mononuclear
cells was done as described elsewhere . Duplicate samples con-
tainingtoseminal mononuclear cells werelabeled
with optimal concentrations of anti–CD4-ECD (Coulter, Miami)
and anti–CD14-phycoerythrin (PE) and anti–CD45RO-PE/CY5
(PharMingen, San Diego). The cells were washed with PBS and
fixed and permeablized by the addition of 50 mL of PermeaFix
(Ortho Diagnostics, Raritan, NJ). The cells were permeabilized for
?60 min and ?18 h. The cells were then washed twice in PBS (pH
7.4) and once in 2? standard saline citrate. The cells were then
resuspended in 50 mL of hybridization buffer containing a cocktail
of 5-carboxyfluorescein–labeled oligonucleotide probes specific for
HIV RNA at 43?C for 30 min. The cells were washed for 5 min
with buffer I and for 30 min with buffer II at 43?C. Multiparameter
four-color analysis was then performed on labeled cells by flow
cytometer (model XL; Coulter).
Measurement of b-chemokine RANTES in semen.
RANTES was measured by ELISA (R&D Systems, Minneapolis)
in 100 mL of seminal fluid per the manufacturer’s protocol.
Analysis of C2–V5sequence of HIV-1 env from semen and blood
Total RNA was extracted from 100 mL of blood plasma
The level of
or whole semen by the RNA extraction protocol of the Nuclisens
assay as described above. RNA extracted by this procedure re-
moved the RT-PCR inhibitor that is often associated with semen
[11, 23]. In total, 5–25 mL of extracted RNA was reversetranscribed
at 42?C for 50 min with 200 U of SuperScript II RNase H?RT
(Gibco BRL, Grand Island, NY) using the PCR downstream
primer ED1220(20 pmol) as the primer in a reaction mixture con-
taining 8 mM dithiothreitol (Gibco BRL), 0.4 mM each deoxyn-
ucleotide triphosphate (Pharmacia Biotech, Piscataway, NJ), 5?
1st strand buffer (Gibco BRL), and 30 U of RNAguard RNase
inhibitor (Pharmacia Biotech) in a total volume of 25 mL. A nested
PCR was performed with 5-fold serial dilutions of cDNA with
primers ED31/BH2 and DR7/DR8 in the first and second round
of PCR, respectively, as described elsewhere . The conditions
of the PCR assay were as follows: samples were preheated at 94?C
for 2.5 min (1 cycle), denatured at 94?C for 1 min, and annealed
at 55?C for 45 s, with extension at 72?C for 1 min (35 cycles) and
final extension at 72?C for 10 min (1 cycle). Each PCR used 10 mL
of diluted cDNA in the first round and 2 mL of first-roundamplified
DNA in the second-round reaction of the PCR assay. The PCR
reaction mixture contained 10 pmol of each primer, 200 mM of
each deoxynucleotide triphosphate, 2.5 U of AmpliTaq DNA poly-
merase (Perkin-Elmer, Norwalk, CT), 50 mM KCl, 1.25 mM
MgCl2, 10 mM Tris-HCl (pH 8.3), and 5% glycerol ina finalvolume
of 50 L. Amplification was monitored by visualization of an eth-
idium bromide–stained band of 678 bp after electrophoresis in a
2% agarose gel. Multiple HIV-1–negative controls were applied
with each PCR run to detect any possible contamination.
Cloning and sequence analysis.
from ?3 independent PCR reactions, obtained from dilutions of
the input cDNA, were cloned in the pGEM-T Easy vector (Pro-
mega, Madison, WI). Plasmid DNAs were prepared by the Wizard
system (Promega) and digested with appropriate restriction en-
For each sample, PCR products
by guest on February 2, 2016
82Gupta et al.JID 2000;182 (July)
Measurement of RANTES in semen from 4 subjects who shed human immunodeficiency virus (HIV) type 1 intermittently
zymes to confirm the size of the insert. Subsequently, DNA from
9 to 11 clones was purified by using the DNA-pure miniprep kit
(CPG, Lincoln Park, NJ) and sequenced with a Taq Dye Deoxy
Terminator Cycle sequencer kit (Applied Biosystems, Foster City,
CA) in an ABI Prism 377 DNA sequencer (Applied Biosystems).
Sequences from each cloned viral DNA were aligned by the
clustal W multiple sequence alignment program . Genetic dis-
tances were calculated with DNADIST software from thePHYLIP
package  by the maximum likelihoodmethod.Neighbor-joining
trees were constructed with the clustal W package, and a bootstrap
analysis using 100 bootstrap replicates was conducted to assess the
support at each internal nodes.
Nucleotide accession numbers.
GenBank under accession numbers AF256230–AF256286 (pa-
tient 1), AF256287–AF256305 (patient 2), AF256306–AF256322
(patient 4), AF256323–AF256404 (patient 6), and AF256405–
AF256465 (patient 8).
Sequences were deposited in
Virus load in semen and blood.
alence of HIV-1 RNA in semen at all visits, three patterns of
HIV-1 shedding emerged. Representative seminal and blood
virus load data of 4 subjects from each of these patterns are
shown in figure 1. Of the 18 subjects studied, HIV-1 RNA was
detected in semen at all 10 visits in 5 (28%), with levels varying
among subjects (figure 1B); in 5 other subjects (28%), HIV-1
On the basis of the prev-
RNA was not detected in semen at any of the 10 visits (figure
1C); and in the remaining 8 subjects (44%), HIV-1 RNA was
detected intermittently in semen during 15%–70% of all visits
(figure 1A). The difference in the level of seminal virus load
between a positive and a negative visit in this intermittentshed-
ding group ranged between 100- and 10,000-fold. In contrast
to the observed variations in seminal virus levels, blood plasma
virus load did not change significantly during this 10-week
study period, even in the group with intermittent shedding of
HIV-1 in semen (figure 1).
Relationship between seminal virus load and semen volume,
seminal mononuclear cells, blood CD4 cells, and sex partners.
We examined whether the different shedding patterns of HIV-
1 in semen were due to a difference in semen volume or seminal
mononuclear cells. There was no relationship between the pat-
tern of virus load as a group or at a particular visit and the
semen volume or the number of mononuclear cells per milliliter
of semen in that group or at that particular visit (data not
shown). Furthermore, the patterns of seminal virus load did
not correlate with blood CD4 T cell number. These results
indicate that intermittent shedding of HIV-1 in semen was not
due to fluctuation of semen volume, semen mononuclear cells,
or blood CD4 T cell number.
To investigate whether chronic infection or inflammation of
the genitourinary tract was linked to patterns of HIV-1 shed-
ding in semen, several covariates were examined from the last
by guest on February 2, 2016
JID 2000;182 (July) HIV Shedding Pattern in Semen83
blood plasma of subject 1, who shed virus intermittently (A), and from subject 6, who shed virus continuously (B), at visits indicated.
Phylogenetic tree of human immunodeficiency virus type 1 envelope (C2–V5) sequence from HIV-1 RNA isolated from semen and
visit prior tothebaselineevaluationofthisstudy.Nodifferences
were observed among the study groups for physical examina-
tion and self-reported findings of herpes zoster, facial herpes,
genital or anal herpes, syphilis, gonorrhea, nonspecific ure-
thritis, anal warts, molluscum contagiosum, or abdominal par-
asite diseases. In fact, only 1 or 2 persons in the entire study
group reported such findings. Since we did not perform micro-
biologic assays to detect common bacterial urethral pathogens
or collect urethral smears for Gram’s stain microscopy, it is
possible that we missed an association between subclinical ure-
thritis and intermittent shedding of HIV-1 in semen,asreported
by Winter et al. . This seems unlikely, however, because of
the very low (8%) prevalence of subclinical urethritis reported
by Winter et al. . This low prevalence of urethritis would
weaken any association between shedding of HIV-1 in semen
and urethritis. However, we found an association between a
greater number of sex partners (17) and shedding of HIV-1 in
semen, since 4 of 5 subjects with continuous HIV-1 shedding
in semen reported multiple sex partners versus only 1 of 5 study
participants with multiple sex partners who never shed HIV-1
Immunophenotyping and expression of HIV-1 transcripts in
seminal mononuclear cells.
We considered the possibility that
seminal HIV-1 was derived from mononuclear cells present in
semen and that therefore intermittent shedding of HIV-1 in
semen was due to the fluctuation of the viral RNA expression
in seminal mononuclear cells. To test this hypothesis, we ex-
amined the relationship between seminal virus load and the
number of activated CD4?, HIV-1–expressing seminal mono-
nuclear cells. Figure 2 shows a representative flow cytometric
detection of HIV-1 gag-pol RNA in seminal mononuclear cells
from 1 intermittent shedder (subject 1) at visits 1, 3, 5, and 9.
Table 1 summarizes the flow cytometry data of the percentage
of activated CD4?, HIV-1 gag-pol RNA-expressing seminal
by guest on February 2, 2016
84Gupta et al. JID 2000;182 (July)
and blood (A) of 3 intermittent shedders (subject 1: visits 3, 8, 10; subject 2: visit 5; subject 4: visit 8) and (B) from 2 continuous shedders (subject
6: visits 1, 3, 6, 10; subject 8: visits 1, 6, 7, 10).
Phylogenetic tree of human immunodeficiency virus (HIV) type 1 envelope (C2–V5) sequence from HIV-1 RNA isolated from semen
mononuclear cells and the activated CD4?(CD4?, CD45RO?)
seminal mononuclear cells from 3 subjects who intermittently
shed HIV-1 in semen. These results indicate that the percentage
of activated CD4?cells and HIV-1 gag-pol–positive CD4 cells
remained at similar levels at all visits regardless of virus load
in semen. These results indicate that the intermittent shedding
of HIV-1 in semen was not due to fluctuations in levels of viral
RNA expression in seminal mononuclear cells and was not
related to the number of activated CD4?cells in seminal mon-
Virus load and b-chemokine level in semen.
such as RANTES, macrophage inflammatory protein (MIP)–
1b, and MIP-1a inhibit viral infection and thereby controlviral
replication [30–33]. To determine whether theintermittentshed-
ding of HIV-1 was due to fluctuation in the level of b-che-
mokines, we measured the level of RANTES in semen samples
at all visits from 4 subjects who shed HIV-1 intermittently in
semen. As shown in figure 3, there was no change in RANTES
levels in semen during the 10-week study. Thus, intermittent
shedding of HIV-1 in semen was not correlated with the level
of b-chemokines in semen.
Sequence analysis of HIV-1 from semen and blood plasma.
A 650-bp region corresponding to the C2–V5region of HIV-1
gp120 from HIV-1 RNA isolated from semen and bloodplasma
was sequenced to determine the viral genetic diversity in 2 sub-
jects (6 and 8) with continuous shedding and in 3 subjects (1,
2, and 4) with intermittent shedding of HIV-1 in semen. Ac-
cording to the deduced amino acid sequences of the V3 loop,
by guest on February 2, 2016
JID 2000;182 (July)HIV Shedding Pattern in Semen 85
the sequences derived from these 5 subjects belonged to the
subtype B clade, and each set clustered as monophyletic groups
compared with those from other patients in this study. Hence,
these data indicate that there was no evidence of cross-contam-
ination among patient samples.
We initially determined whether intermittent shedding was
due to any rapid genetic variation of HIV-1 between visits. For
this purpose, we analyzed sequences of virus population from
semen and blood plasma samples at multiple visits from 1 in-
termittent shedder (subject 1: visits 3, 8, and 10) where we had
sufficient viral RNA for amplification. From each sample, at
least 10 clones were sequenced at multiple dilutions of the am-
plified DNA, to avoid sampling error. Phylogenetic analysis
(figure 4A) revealed that both blood plasma-andsemen-derived
variants formed a tightly clustered group of sequences between
visits (mean percentages of nucleotide divergences, 0.7?0.4
for semen and for bloodplasma),showingnoindication
of significant genetic diversity among visits. We extended this
observation by analyzing HIV-1 variants from semenandblood
plasma of 1 continuous shedder (subject 6) at visits 1, 3, 6, and
10. As shown in figure 4B, no major difference in genetic di-
versity was observed between visits (mean percentages of nu-
Next, we compared sequences of HIV-1 variants in blood
with those in semen of 3 intermittent shedders (subject 1: visits
3, 8, and 10; subject 2: visit 5; subject 4: visit8)and2continuous
shedders (subject 6: visits 1, 3, 6, and 10; subject 8: visits 1, 6,
7, and 10) at visits with sufficient amplifiable HIV-1 RNA (fig-
ure 5). The percentage of divergence at the nucleotide level
among the plasma-derived clones was
,, , and
3.4?1.8 5?2.8 2.1?1.1
1, 4, and 2, respectively. Overall, such diversity among plasma-
derived clones was higher than the diversity among semen-
derived clones with,
2.3?1.5 2.9?1.3 0.7?0.4 1.8?0.3
and for subjects 6, 8, 1, 4, and 2, respectively. How-
ever, as shown in figure 5, in all 3 intermittentshedders(subjects
1, 2, and 4), semen-derived variants in general formed a tightly
clustered group of sequences distinct from that in their blood
plasma. In general, semen-derived variants were more clustered
than their blood counterparts, as shown by the greater mean
percentage of divergence of HIV-1 variants in plasma (mean,
3.6; range, 2.1–5.0) compared with those in semen (mean, 1.03;
range, 0.6–1.8). In addition, on the basis of the charge of the
amino acids at positions 11, 13, 19, 23, 24, 25, and 32 of the
V3 loop [34–36], all semen and plasma clones from subjects 2
and 4 had the consensus sequence associated with non-
SI–inducing (NSI) phenotype. However, in subject 1, HIV-1
variants with SI genotypes were obtained only in blood plasma
and not in semen.
Phylogenetic analysis of HIV-1 sequences from continuous
shedders (subjects 6 and 8) revealed a distributed pattern of
viral variants between semen and blood plasma, with no evi-
for blood plasma and
from subjects 6, 8,
dence of clustering in either body compartment (figure 5). The
mean percentage of divergence among viral variants in blood
plasma (for subject 6 and 3.2?1.8
very similar to that in semen ( 2.5?1.5
jects 6 and 8, respectively). All HIV-1 variants in both body
compartments of these 2 continuous shedders had NSI
for subject 8) was
In a cross-sectional study, we previously found that HIV-1
can be detected in semen in 66% of seropositive subjects at all
stages of disease . It was not clear whether the observed
lack of detection of HIV-1 in semen in that study was due to
the absence of shedding of HIV-1 or the intermittent shedding
of HIV-1 in semen. Inthepresentstudy,wefoundthreedifferent
HIV-1 shedding patterns in semen: none, continuous, and in-
termittent. Of the subjects who shed HIV-1 in semen, the in-
termittent shedding pattern was predominant (61%).
In the present study, we analyzed various potential factors
as determinants of the different HIV-1 shedding patterns in
semen. We found that the pattern was not related to semen
volume, number of seminal mononuclear cells, or to percentage
of CD4 T cells in blood. We also demonstrated that the inter-
mittent shedding of HIV-1 in semen was not due to fluctuation
in the level of expression of HIV-1 RNA in seminal mono-
nuclear cells. This was further supported by the lack of cor-
relation between the number of activated CD4?(CD4?,
CD45RO?) seminal mononuclear cells and the intermittencyof
HIV-1 in semen. Furthermore, our data indicate that the in-
termittent shedding of HIV-1 is not due to the fluctuation of
the b-chemokine RANTES level, which suppresses HIV-1 in-
fection by a competitive interaction with the viral coreceptor
. This finding is in contrast with the recent report of Iversen
et al. , who found a positive correlation of cervical HIV-1
shedding with genital chemokine secretion. Sequence analysis
of HIV-1 at multiple visits in subjects with intermittent and
continuous shedding of HIV in semen also did not indicate any
rapid genetic variation between visits as the reason for the in-
termittent shedding of HIV-1.
Our data indicate that in subjects with intermittent shedding
of HIV-1 in semen, the virus population in semen was distinct
from that in blood and there was no correlation between the
level of virus in semen and blood. At certain visits, HIV-1 was
not detected in semen but there was a significant amount of
virus in blood. Sequence analysis of the C2–V5region of the
viral envelope gene of HIV-1 variants obtained from semen and
blood plasma RNA at multiple visits from the subjects with
intermittent shedding indicated that HIV-1 variants in semen
differed from those in blood. These results from viral RNA
support our previous data  and those of Zhu et al. ,
which were obtained by sequence analysis of proviral DNA
from semen and peripheral blood mononuclear cells.Moreover,
by guest on February 2, 2016
86 Gupta et al. JID 2000;182 (July)
in 1 intermittent shedder, the SI genotype was detected only in
blood and not in semen. These results suggest that, in inter-
mittent shedders, HIV-1 in semen was not recently derivedfrom
peripheral blood. Furthermore, in these subjects, the seminal
mononuclear cells were not the origin of the HIV-1 in semen,
as evidenced by the lack of correlation between the seminal
virus load and the percentage of HIV-1–expressing seminal
mononuclear cells. In contrast, in continuous shedders, the
HIV-1 RNA population in semen was similar to that in blood,
indicating that HIV-1 in semen in this group of subjects may
come from blood. This was supported by the concordance of
HIV-1 RNA level betweensemenandbloodincontinuousshed-
ders at all visits. This observation of different patterns of com-
partmentalization with different shedding patterns of HIV in
semen could explain our previous findings  and those of
Zhu et al. , in which differential representations of inte-
grated proviral sequences were found between blood andsemen
in 3 subjects but not in 2 others. On the basis of our current
findings, it is conceivable that these 2 subjects with similarvirus
populations in semen and blood could belong to the group of
continuous shedders. In contrast, the other 3 subjects with dis-
tinct HIV-1 species in semen and blood could be grouped with
intermittent shedders. The 60% of subjects with distinct HIV-
1 genotypes in the previous study matches well with 61% of
subjects with intermittent shedding in this study. Of interest,
nonshedders also had very low level of HIV RNA in their
blood, indicating a concordanance of HIV replication between
these two compartments in this group of subjects.
This investigation raises an important question about the
origin of HIV-1 in semen. It is thought that lymphocytes enter
the seminal compartment through the epididymis , because
the number of leukocytes in semen of vasectomized men is
greatly reduced . The prostate and seminal vesicles are the
major contributors of seminal fluid, which contains high levels
of HIV-1 . Vasectomized subjects can transmit HIV-1, and
vasectomization usually occurs at the proximal end of the these
two glands in the male reproductive duct. Therefore, it is pos-
sible that prostate and seminal vesicles could be the major
source of seminal HIV-1. In addition, since other types of mi-
crobial infection have been reported in the prostate [40–42],this
organ appears to be the most likely candidate for a reservoir
for seminal virus.
On the basis of the data reported here, we propose the fol-
lowing model to explain the origin and pattern of HIV-1 shed-
ding in semen. In nonshedders, because of a low virus load in
the blood, very low levels of virus and/or virus-infected cells
enter the male genital organ. Like its counterpart in blood,
viral replication in the genital organ is probably controlled by
the local immune system. In intermittent shedders, at some
point during the course of HIV-1 infection (possibly during
acute infection), HIV-1–infected lymphocytes enter the male
genital organ through the testis or epididymis because of local
inflammation and subsequently reside in specific genital tissue
(e.g., the prostate). In this group of infected subjects,thebarrier
at the blood-testis junction would not allow passage of lym-
phocytes or high-molecular-weight macromolecules (e.g., vi-
ruses) after initial introduction of virus or virus-infected cells.
In addition, the migration of blood lymphocytes through the
epididymis is restricted, giving rise to a sequestered compart-
ment in which HIV-1 could continue to replicate in specific
genital tissue. Since the immune environmentinthemalegenital
organ is thought to differ than that in the blood [3, 38, 43, 44],
the level of HIV-1 replication in the face of immune selection
in the male genital organ would be expected to differ from than
that in blood, giving rise to the diversity of HIV-1 populations
in these two body compartments. Since other types of bacterial
infections are found in the prostate [40–42] and the prostate
contains lymphoid cells [40–42], this organ may be a source of
seminal HIV-1 in intermittent shedders. Thus, intermittent
shedding of HIV-1 in semen would be due to a fluctuation of
viral transcription from infected cells in the prostate. This no-
tion is supported by a study by Kiessling et al. , whoshowed
that, during an episode of asymptomatic prostatitis,theseminal
virus load increased 100-fold over that in peripheral blood and
that protease-resistant HIV-1 variants were found in semen,but
not in blood.
In continuous shedders, inflammation at the testis-blood
junction or at the epididymis would allow continuous passage
of lymphocytes, virus particles, or both to enter from blood to
the male genital organ and thereby contribute to similar viral
representations between semen and blood. The fact that the
continuously shedding subjects in this study had significantly
more sex partners than the nonshedders supports such a pos-
sibility of inflammation in the male genital organ.
In summary, our data indicate that the source of HIV-1 in
semen is complex and is related to pattern ofsheddinginsemen.
Direct examination of viral variants in prostate, semen, and
blood could confirm the prostate origin of HIV-1.
We thank Judy Malenka for secretarial assistance and the partici-
pants and staff of the Pitt Men’s Study for dedication and support.
1. Krieger JN, Coombs RW, Collier AC, et al. Recovery of human immuno-
deficiency virus type 1 from semen: minimal impact of stage of infection
and current antiviral chemotherapy. J Infect Dis 1991;163:386–8.
2. Mermin JH, Holodiny M, KatzensteinDA,MeriganTC.Detectionofhuman
immunodeficiency virus DNA and RNAin semen bythepolymerasechain
reaction. J Infect Dis 1991;164:769–72.
3. Pudney J, Anderson DJ. Orchitis and human immunodeficiency virus type
1 infected cells in reproductive tissues from men with theacquiredimmune
deficiency syndrome. Am J Pathol 1991;139:149–60.
4. Van Voorhis BJ, Martinez A, Meyer K, Anderson DJ. Detection of human
immunodeficiency virus type 1 in semen from seropositive men using cul-
ture and polymerase chain reaction deoxyribonucleic acid amplification
technique. Fertil Steril 1991;55:588–94.
by guest on February 2, 2016
JID 2000;182 (July) HIV Shedding Pattern in Semen87 Download full-text
5. Vernazza P, Eron J, Cohen M, Vander Horst C, Troian L, Fiscus S. Detection
of biologic characterization of infectious HIV-1 in semen of seropositive
men. AIDS 1994;8:1325–9.
6. Vernazza P, Eron J, Fiscus S. Sensitive method for the detection of infectious
HIV in semen of seropositive individuals. J Virol Methods1996;56:33–40.
7. Quayl A, Coston W, Trocha A, Kalams S, Mayer K, Anderson D. Detection
of HIV-1–specific CTLs in the semen of HIV-infected individuals. J Im-
8. Levy J. The transmission of AIDS: the case of the infected cells. JAMA
9. Levy JA. The transmission of HIV and factors influencing progression to
disease. Am J Med 1993;95:86–100.
10. Coombs R, Speck C, Hughes J, 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-1between
semen and blood. J Infect Dis 1998;177:320–30.
11. Gupta P, Mellors J, Kingsley L, et al. High viral load in semen of HIV-1
infected men at all stages of disease and its reduction by therapy with
protease and nonnucleoside RT inhibitors. J Virol 1997;71:6271–5.
12. Vernazza P, Gilliam B, Dyer J. Quantification of HIV in semen: correlation
with antiviral treatment and immune status. AIDS 1997;11:987–93.
13. Miller C. Use of the SIV/rhesus macaque model of the heterosexual trans-
mission of HIV in AIDS vaccine research. Vaccine Res 1992;1:295–301.
14. Vernazza P, Gilliam B, Flepp M. Effect of antiviral treatment on theshedding
of HIV-1 in semen. AIDS 1997;11:1249–54.
15. Byrn RA, Zhang D, Eyre R, Mcgowan K, Kiessling AA. HIV-1 in semen:
an isolated virus reservoir. Lancet 1997;350:1141.
16. Delwart E, Mullins J, Gupta P, et al. Human immunodeficiency virus type
1 populations in blood and semen. J Virol 1998;72:617–23.
17. Zhu T, Wang N, Carr A, et al. Genetic characterization of human immu-
nodeficiency virus type 1 in blood and genital secretions: evidenceforviral
compartmentalization and selection during sexual transmission. J Virol
18. Liuzzi G, Chirianni A, Clementi M, et al. Analysis of HIV-1 load in blood,
semen and saliva—evidence for different viral compartments in a cross-
sectional and longitudinal study. AIDS 1996;10:F51–6.
19. Eron J, Vernazza P, Johnston D. Resistance of HIV-1 to antiretroviralagents
in blood and seminal plasma implications for transmission. AIDS 1998;
20. Kroodsma KL, Kozal MJ, Hamed KA, Winters MA, Merigan TC.Detection
of drug resistance mutations in the human immunodeficiency virus type
1 (HIV-1) pol gene: differences in semen and blood HIV-1 RNA and
proviral DNA. J Infect Dis 1994;170:1292–5.
21. Krieger J, Coombs R, Collier A. Intermittent shedding of human immu-
nodeficiency virus in semen: implications for sexual transmission. J Urol
22. Giorgi JV, Cheng HL, Margolick JB, et al. Quality control in the flow cy-
tometric measurement of T-lymphocyte subsets: the Multicenter AIDS
Cohort Study experience. Clin Immunol Immunopathol 1990;55:173–86.
23. Dyer JR, Gilliam BL, Eron JJ, et al. Quantitation of human immunodefi-
ciency virus type 1 RNA in cell-freeseminalplasma—comparisonofNAS-
BATMreverse transcription–PCR amplification and correlationwithquan-
titative culture. J Virol Methods 1996;60:161–70.
24. Vandamme AM, Schmit JC, Van Dooren S, et al. Quantification of HIV-1
RNA in plasma: comparable results with the NASBA HIV-1 RNA QT
and the AMPLICOR HIV monitor test. J Acquir Immune Defic Syndr
Hum Retrovirol 1996;13:127–39.
25. Van Genman B, Niel PV, Van Beuningen R. In: PCR methods and appli-
cations. Cold Spring Harbor,NY:ColdSpringHarborLaboratories,1995:
26. Patterson B, Mosiman V, Caantarero L, et al. Detection of HIV-
RNA–positive monocytes in peripheral blood of HIV-positive patientsby
simultaneous flow cytometric analysis of intracellular HIV RNA and cel-
lular immunophenotype. Cytometry 1998;31:265–74.
27. Thompson J, Higgins D, Gibson T. CLUSTAL W: improving the sensitivity
of progressive multiple sequence alignment through sequence weighting,
position-specific gap penalties and weight matrix choice. Nucleic Acids
28. Felsenstein J. PHYLIP (phylogeny inference package), version 3.5c. Seattle,
29. Winter AJ, TaylorS,WorkmanJ,etal.Asymptomaticurethritisanddetection
of HIV-1 RNA in seminal plasma. Sex Transm Infect 1999;75:261–3.
30. Bleul C, Farzan M, Choe H, et al. The lymphocyte chemoattractant SDF-
1 is a ligand for LESTR/fusion and blocks HIV-1 entry. Nature 1996;382:
31. Cocchi F, DeVico A, Garzino-Demo A. Identification of RANTES, MIP-1a
and MIP-1B as the major HIV-suppressive factors produced by CD8?T
cells. Science 1995;270:1811–5.
32. Oberlin E, Amara A, Bachelerie F, et al. The CXC chemokine SDF-1 is the
ligand for LESTR/fusin and prevents infection by T-cell-line–adapted
HIV-1. Nature 1996;382:833–5.
33. Pal R, Garzino-Demo A, Markham P, et al. Inhibition of HIV-1 infection
by the B-chemokine MDC. Science 1997;278:695–8.
34. Chesbro B, Wehrly K, Nishio J, Perryman S. Macrophage-tropic human
immunodeficiency virus isolates from different patients exhibit unusual
V3 envelope sequence homogeneity in comparison with T-cell tropic iso-
lates: definition of critical amino acids involved in cell tropism. J Virol
35. Fouchier R, Brouwer M, Broersen S, Schuitemaker H. Simple determination
of human immunodeficiency virus type 1 syncytium-inducingV3genotype
by PCR. J Clin Microbiol 1995;33:906–11.
36. Fouchier R, Groenink M, Koostra N, et al. Phenotype-associated sequence
variation in the third variable domain of the human immunodeficiency
virus type 1 gp120 molecule. J Virol 1992;66:3183–7.
37. Iversen A, Fuggerb L, Eugen-Olsen J, et al. Cervical human immunodefi-
ciency virus type 1 shedding is associated with genital B-chemokine se-
cretion. J Infect Dis 1998;178:1334–42.
38. Anderson D, Pudney J. Mucosal immune defense against HIV-1 in the male
urogenital tract. Vaccine Res 1992;1:143–150.
39. Wolff H, Anderson D. Immunohistologic characterization and quantitation
of leukocyte subpopulations in human semen. Fertil Steril 1988;49:
40. Campbell MF, Retik AB, Vaughan E, Walsh PC, eds. Campbell’s urology.
Philadelphia: WB Sanders, 1997.
41. El-Demiry M, James K. Lymphocyte subsets and macrophages in the male
genital tract in health and disease. A monoclonal antibody-based study.
Eur J Urol 1988;14:226–34.
42. Murphey W. Urology pathology, Philadelphia: WB Sanders, 1996.
43. Quayle A, Xu C, Mayer K, Anderson D. T lymphocytes and macrophages,
but not motile spermatozoa, are a significant source of human immu-
nodeficiency virus in semen. J Infect Dis 1997;176:960–8.
44. Wolff H, Mayer K, Seage G, et al. A comparison of HIV-1 antibody classes,
titers, and specificities in paired semen and blood samples from HIV-1
seropositive men. J Acquir Immune Defic Syndr 1992;5:65–9.
45. Kiessling AK, Zheng G, Eyre RC. Semen producing organs are an isolated
reservoir of HIV which may play a significant role in the development of
drug resistant strains. J Hum Virol 1999;2:193.
by guest on February 2, 2016