Increase in Endocervical CD4 Lymphocytes among Women with Nonulcerative
Sexually Transmitted Diseases
William C. Levine, Victoria Pope, Asha Bhoomkar,
Pradnya Tambe, Joel S. Lewis, Akbar A. Zaidi,
Carol E. Farshy, Shannon Mitchell,
and Deborah F. Talkington
Epidemiology and Surveillance Branch and Statistics and Data
Management Branch, Division of Sexually Transmitted Disease
Prevention, National Center for HIV, STD, and TB Prevention, Division
of AIDS, STD, and TB Laboratory Research and Division of Bacterial
and Mycotic Diseases, National Center for Infectious Diseases, Centers
for Disease Control and Prevention, and Fulton County Health
Department, Atlanta, Georgia
To assess associations of nonulcerative sexually transmitted diseases (STDs) with human immuno-
deficiency virus (HIV)–susceptible leukocytes on female genital mucosa, cervicovaginal specimens
from 32 HIV-negative STD clinic patients with gonorrhea, chlamydial infection, or trichomoniasis
were compared with specimens from 32 clinic patients without these infections. Twenty-eight pa-
tients had single infections (15 gonorrhea, 10 chlamydial infection, 3 trichomoniasis), and 4 had
dual infections. A saline vaginal wash and saline suspensions of vaginal wall scrapings, ectocervical
vix had the highest proportions of lymphocytes, monocytes, and Langerhans’ cells. The median
number of endocervical CD4 lymphocytes/10,000 cells was greater among patients with STDs than
among those without (476 vs. 245; P õ .001). These data suggest that the endocervix may have a
particularly important role in heterosexual HIV transmission and that nonulcerative STDs may
facilitate HIV transmission by increasing the presence of CD4 lymphocytes at this site.
Although epidemiologic studies of sexually transmitted dis-
eases (STDs) as risk factors for human immunodeficiency virus
(HIV) infection have demonstrated strong associations of HIV
with ulcerative STDs, they have shown less consistent associa-
tions of HIV infection with the nonulcerative STDs—gonor-
rhea, chlamydial infection, and trichomoniasis . A limitation
of these studies is that they have not been able to characterize
the extent to which STDs may facilitate HIV transmission by
augmenting HIV infectivity of persons with HIV versus in-
creasing susceptibility of persons who are exposed to HIV.
While investigators are now examining the role of nonulcera-
tive STDs in increasing infectivity by increasing viral shedding
[2–4], the question of increased susceptibility has received less
attention. Epidemiologic evidence for a role of these diseases
in increasing susceptibility to HIV infection comes from one
well-designed prospective study of female sex workers who
were at increased risk for seroconverting to HIV if they had
nonulcerative STDs  and from two prospective studies of
HIV-discordant sex partners, in which HIV seroconversion was
associated with STDs (including nonulcerative genital infec-
tions or syndromes) in the previously HIV-negative sex partner
[6, 7]. However, the specific mechanisms by which these STDs
may increase susceptibility to HIV infection have not been
Cervical inflammation in women is a factor that may affect
HIV transmission risk. However, most studies of the role of
nonulcerative STDs in HIV transmission have used imprecise
methods to assess the cervix. Clinical observations of friability,
mucopurulent cervicitis, or numbers of polymorphonuclear leu-
kocytes present on Gram’s stain of an endocervical smear are
clinically useful but do not provide a detailed picture of the
inflammatory process, particularly in terms of quantifying the
presence of HIV-susceptible cells. One detailed study that sys-
tematically examined Papanicolaou-stained cervical smears
demonstrated associations of Chlamydia trachomatis and Tri-
chomonas vaginalis infections with greater numbers of lym-
phocytes , and one study of cervical histopathology demon-
strated that cervical chlamydial infection is associated with
underlying infiltration of the endocervix with lymphocytes and
macrophages . However, in those studies, lymphocyte sub-
sets possibly associated with HIV transmission (e.g., CD4 lym-
phocytes) were not assessed. One study that examined lympho-
cyte immunophenotypes in vaginal wall and cervical scrapings
and in endocervical exudate of women with HIV infection
found no association of prior STDs with presence of CD4
at the time the cell specimens were collected .
Weperformed astudy toaddressone aspectof thesecomplex
issues,focusing onfactorsthat maycorrelate withsusceptibility
Received 17 December 1996; revised 9 July 1997.
Presented in part: X International Conference on AIDS/International Confer-
ence on STD, Yokohama, Japan, August 1994.
Informed consent was obtained from all study participants according to
institutional guidelines. Human experimentation guidelines of the US Depart-
ment of Health and Human Services, the Centers for Disease Control and
Prevention, and the Fulton County Health Department were followed.
Reprints or correspondence: William C. Levine, Epidemiology and Surveil-
lance Branch, Division of STD Prevention, Centers for Disease Control and
Prevention, 1600 Clifton Rd, NE (Mailstop E02), Atlanta, GA 30333.
The Journal of Infectious Diseases
? 1998 by The University of Chicago. All rights reserved.
/ 9d3d$$ja2611-26-97 14:25:14jinfaUC: J Infect
by guest on July 13, 2011
168 Levine et al.JID 1998;177 (January)
men for examination of endocervical cells. Vaginal pool specimens
from the posterior fornix were taken for Gram’s stain and for T.
vaginalis and Candida cultures. A heparinized peripheral blood
sample was collected for lymphocyte immunophenotyping. Gonor-
rhea Gram’s stains and cultures were done at the Fulton County
Health Department and HSV cultures at the Georgia State Public
Health Laboratory. Samples for the remaining tests were delivered
to the Centers for Disease Control and Prevention within a few
hours of collection; the Chlamydia culture transport tube was kept
on ice, and the remaining specimens were kept at ambient tempera-
Clinical specimens for culture of C.
trachomatis were placed in 2-sucrose-phosphate transport medium
. If specimens were processed within 24 h, they were stored
at 4?C; if ú24 h, they were stored at 070?C. The shell vial culture
method of Ripa was used to culture clinical specimens . The
cell monolayers were stained with Pathfinder Culture Confirmation
Reagent (Kallestad Diagnostics, Chaska, MN). Clinical specimens
were passaged blindly before culture results were reported. Endo-
cervical specimens were assayed for C. trachomatis DNA by the
Amplicor C. trachomatis test (Roche Diagnostics, Branchburg,
NJ). Martin-Lewis medium (Carr-Scarborough, Atlanta) was used
for the inoculation of endocervical samples for the cultivation of
N. gonorrhoeae . Vaginal specimens were obtained for isola-
tion of T. vaginalis and Candida albicans by use of modified
Diamond’s medium prepared at the Centers for Disease Control
and Prevention . M. hominis was cultured by standard tech-
niques in SP4 broth medium  and on A8 agar . U. urealyti-
cum was cultured in 10B broth  and on A8 agar. Specimens
were also tested for M. hominis by assaying for a 334-bp portion
of the 16S rRNA gene  and for U. urealyticum DNA by
assaying for the urease genes . Specimens for HSV determina-
tion were inoculated into the A549 cell culture line, which was
examined for cytopathic effect. A Gram’s stain of vaginal secre-
tions from each participant was evaluated for the presence of bacte-
rial vaginosis by use of morphologic criteria .
Processing of cell suspensions.
cessed within 2–6 h of specimen collection. They were first centri-
fuged at 500 g for 5 min to pellet the cells. The supernatants
were removed by vacuum. The cell pellets from the vaginal wall
scrapings and ecto- and endocervical samples were resuspended
in 0.3 mL of physiologic saline. The cells from the vaginal wash
were suspended in 1.0–1.5 mL of physiologic saline. Monoclonal
antibodies were added to each of 3 tubes. One hundred microliters
of sample was added and incubated for 15 min at room temperature
in the dark. Two milliliters of lysing reagent (1:10 dilution of
FACS Lyse buffer [Becton Dickinson, San Jose, CA]) was added
and incubated for 15 min at room temperature in the dark. Cells
were then centrifuged at 300 g for 3 min and resuspended in 1.5
mL of wash buffer (2.0% AB serum, 0.1% bovine serum albumin,
and 0.2% sodium azide in 0.01 M PBS, pH 7.2). The cells were
centrifuged as above, then resuspended in 300 mL of 2% buffered
formalin. The samples were stored at 4?–8?C overnight to mini-
mize cell aggregation.Samples were analyzed ona FACScan (Bec-
ton Dickinson) by three-color analysis. The fluorescent markers
used were anti-CD45–peridinin chlorophyll protein (PerCP),
-CD14–phycoerythrin (PE), and -CD1–fluorescein isothiocyanate
(FITC) in tube 1; anti-CD3–PerCP, -CD19–FITC, and –T cell
by using immunophenotyping methods to examine mucosal
inflammation in HIV-negative women who were infected with
nonulcerative STDs. We attempted to address the following
question: What are the effects of nonulcerative STDs (gonor-
rhea, chlamydial infection, and trichomoniasis) on the numbers
of HIV-susceptible cells (CD4 lymphocytes, monocyte/macro-
phages, and Langerhans’ cells) on female mucosal surfaces
and in cervicovaginal secretions? Each of these cell types can
become infected with HIV, and infection of monocytes and
Langerhans’ cells can lead to infection of CD4 lymphocytes.
Because of potential effects of other types of genital infections
on cervicovaginal mucosa, we also examined associations of
Candida species,bacterial vaginosis,Mycoplasma hominis,and
Ureaplasma urealyticum with numbers of mucosal white blood
offered enrollment in the study if they were sex partners of a man
with documented gonorrhea, chlamydial infection, or nongonococ-
cal urethritis or if they had a chief complaint of vaginal discharge.
Patients were included only if they were 18–30 years of age, were
not pregnant or menstruating, had taken no antibiotics during the
previous 14 days, and within the past 2 days had not douched,
used vaginal creams or suppositories, or had sex. Patients found
on physical examination to have ulcerative lesions were excluded.
Patients were confidentially tested for HIV and included in this
analysis only if they were HIV-negative. On study days, the first
1 or 2 eligible patients per day who fit criteria were offered enroll-
ment. Two patients declined entry into the study.
All patients were interviewed and examined by one of two study
clinicians, who used a standard form to collect demographic and
clinicaldata. Apatient wasconsideredto havea nonulcerativeSTD
if they had endocervical infection with C. trachomatis identified by
culture or polymerase chain reaction (PCR), Neisseria gonor-
rhoeae by culture, or T. vaginalis by culture. These patients were
compared with eligible patients who had none of these infections.
Cervical ectopy was estimated by the clinician as a percentage of
the radius of the cervix. All patients were examined for presence
of mucopus at the os and for the presence of pus on the first
endocervical swab used to collect microbiologic specimens. Pa-
tients were considered to have easily induced endocervical bleed-
ing if a moderate amount of blood (blood-soaked or pink discolor-
ation) was present on the first endocervical swab; a trace amount
of blood was recorded if few discrete spots were present.
A vaginal wash with physiologic saline
was done before collecting other specimens; 5 mL of saline was
injected onto the vaginal walls and then aspirated. An ectocervical
scraping with a wooden spatula was agitated in 5 mL of physio-
logic saline, and an endocervical cytobrush was also agitated in 5
mL of physiologic saline. On the first 25 patients, a scraping of
the vaginal wall was obtained after the vaginal wash by use of a
spatula and agitated in 5 mL of physiologic saline. Endocervical
samples for gonorrhea Gram’s stain and culture, herpes simplex
virus (HSV) culture, and C. trachomatis, M. hominis, and U. urea-
lyticum culture and PCR were taken after collection of the speci-
Women presenting to an STD clinic were
Cell suspensions were pro-
/ 9d3d$$ja2611-26-97 14:25:14jinfaUC: J Infect
by guest on July 13, 2011
169JID 1998;177 (January)Endocervical CD4 Lymphocytes and STDs
receptor (TCR) g/d–PE in tube 2; and anti-CD4–PerCP, -CD8–
PE, and -CD38–FITC in tube 3.
List mode data were collected by use of FACScan Research
software (Becton Dickinson) and analyzed with Lysis software
(Becton Dickinson). Ten thousand events or the contents of the
tube, whichever came first, were collected per tube. The 3-tube
panel was run on the heparinized blood sample and used as a
control for each patient and to set the initial lymphocyte gate based
on light scatter and CD45 and CD14 markers. In tube 1, CD45
was used for identifying white blood cells, CD14 for monocytes,
and CD1 for Langerhans’ cells; in tube 2, CD3 was used for T
cells, CD19 for B cells, and the g/d marker for g/d T cells; and
in tube 3, which included CD4 and CD8 markers, CD38 was used
as an indicator of T cell activation. In all tubes, the gate was set
to minimize the number of monocytes, and the fluorescent pattern
of the CD45 tube indicated that there were few monocytes in the
gate. The heparinized blood tube was also used to set the initial
granulocyte gate for each patient. Gate parameters were set using
forward scatter versus side scatter and checked with CD45 versus
CD14. The number of events inside the gate was designated as
the number of granulocytes. Peripheral blood samples were run in
our standard HIV panel  as well as the same 3-tube panel that
was used for the cervical and vaginal samples.
Assessment of blood in specimens.
presence of blood immediately after initial centrifugation was
done. Because blood cells are lysed during preparation of speci-
mens for flow cytometry, presence of blood in specimens was
assessed semiquantitatively by the following method: 0, no cells
observed; 1/, suspension with no pink tinge but with a thin line
of red cells visible at the periphery of the cell pellet; 2/, suspen-
sion pink-tinged and a layer of red cells seen on the top of the
cell pellet; and 3/, obvious gross contamination with blood of the
suspension and cell pellet.
To facilitate comparisons, cell counts for
tubes with õ10,000 cells were represented as number of cells per
10,000 cells counted ([n/N] 1 10,000). Because most cell counts
were not normally distributed, medians and interquartile ranges
were used as summary statistics for cell counts. Medians were
compared by the Kruskal-Wallis nonparametric test of signifi-
cance. The x2test with Yates’s correction or Fisher’s exact test
(when expected cell values were õ5) were used to compare pro-
portions. Power calculations were done on the basis of observed
or hypothetical differences in proportions. Because of inadequate
or missing specimens, or lack of reagents, not all denominators
STD clinic patients with and without Neisseria gonorrhoeae, Chla-
mydia trachomatis, or Trichomonas vaginalis infections.
Characteristics and coexistent genital infections of female
(n Å 32)
(n Å 32) Patient characteristic
Age õ25 years, %
Sex partners in past year, median no.
Current oral contraceptive use, %
Time since last menstrual period,
Douched in previous 2 weeks, %
Cervical ectopy ú10%,‡%
Other genital infection, %
* Gonorrhea, chlamydial infection, or trichomoniasis.
†Medians were compared with Kruskal-Wallis nonparametric test; propor-
tions with x2test using Yates’s correction or Fisher’s exact test (when expected
cell number õ5).
‡Cervical ectopy estimated by clinician as proportion of radius of cervix.
§Tests for U. urealyticum infection were done on 27 patients with gonorrhea,
chlamydial infection, or trichomoniasis and 29 patients without these infec-
Samples were assessed for
of such patients (P Å .15). The remaining patients had a chief
complaint of vaginal discharge. Patients with infection were
younger than those without infection, but there were no sig-
nificant differences in infection status by number of sex part-
ners or oral contraceptive use (table 1). There were also no
significant differences in the number of days since last men-
strual period, the proportion who douched in the previous 2
weeks, or the presence of cervical ectopy noted by visual exam-
ination of the cervix. Many study participants with and without
STDs had positive assays for Candida species, M. hominis,
and U. urealyticum and had bacterial vaginosis. None of the
endocervical culture specimens from the 64 patients grew HSV.
Proportional cell counts by site of specimen collection.
Among all patients, the highest cell counts of lymphocytes,
monocytes, and Langerhans’ cells were in the endocervix, com-
pared with cell numbers in the ectocervix and vaginal wash
specimens (table 2). In the endocervix, CD4 lymphocytes were
more common than monocytes and Langerhans’ cells, whereas
in the ectocervix and vaginal wash, monocytes were somewhat
more common. The vaginal wall samples had few detectable
lymphocytes, monocytes, or Langerhans’ cells. Granulocytes
predominated at all sites.
Proportional cell counts by infection with nonulcerative
The median number of CD4 lymphocytes in the endo-
cervix among patients with STDs was 476/10,000 cells, com-
pared with 245 for patients without STDs (P õ .001; table
3). Twenty-three (72%) of 32 patients with STDs had a CD4
lymphocyte proportion that was above the overall median of
385, compared with 9 (28%) of the 32 patients without STDs
tive STDs, 28 had single infections (15 gonorrhea, 10 chlamyd-
ial infection, and 3 trichomoniasis), and 4 patients had mixed
infections. These patients were compared with 32 patients with
no infection identified at the clinic. Because of the small num-
bers for each disease, infected patients were grouped together
for most analyses. Of the patients with STDs, 27 (84%) were
enrolled because they were a contact of a patient with urethral
infection, and of those without STDs, 21 (66%) were contacts
Of the 32 patients with nonulcera-
/ 9d3d$$ja2611-26-97 14:25:14jinfaUC: J Infect
by guest on July 13, 2011
170Levine et al. JID 1998;177 (January)
Median (interquartile range) of lymphocytes, monocytes, Langerhans cells, and granulocytes per 10,000 cells by site of specimen
B lymphocytesMonocytes LangerhansGranulocytes
monocytes, and granulocytes on 61 of these patients; and Langerhans cells at each site on 54 of these patients. Collection and testing of vaginal wall scrapings
was limited to first 25 patients enrolled (14 of these patients had STD).
Determinations of endocervical and vaginal wash lymphocytes, monocytes, and granulocytes were made on 64 patients; of ectocervical lymphocytes,
(P Å .001). On the basis of these proportions, the study had a
power of .94 to detect a significant difference (a Å .05). The
median number of CD4 lymphocytes per 10,000 cells in ecto-
cervical and vaginal wash specimens did not differ significantly
by the presence of STDs (table 3). Among patients with STDs,
44% (14/32) had §1 CD4 lymphocyte detected in the vaginal
wash/10,000 cells counted, compared with 41% (13/32) of pa-
tients without STDs (P Å .9). The statistical power to detect
a 100% increase in the proportion with CD4 lymphocytes in
the vaginal wash (i.e., from 41% to 82%) was .90; however,
to detect a 50% increase (i.e., from 41% to 62%) the power
was only .27.
At the endocervix, the proportions of other lymphocyte sub-
sets were also increased in patients with STDs. The median
number of endocervical CD8 lymphocytes was 415/10,000
cells among patients with STDs, compared with 257 in patients
with no STD (P Å .001); for g/d T-lymphocytes, the medians
were 47 and 24, respectively (P Å .004). The number of endo-
cervical CD4 lymphocytes that carried the activation marker
CD38 was also greater in patients with STDs than in patients
without: 187 versus 74/10,000 cells (P Å .007). The percentage
ofCD4lymphocytes thatcarriedthismarker wasgreateramong
patients with STDs than among those without, but not signifi-
cantly (41% vs. 33%; P Å .15). Monocytes and Langerhans’
cells showed no significant differences at the endocervix or
ectocervix or in vaginal wash specimens between patients with
and without infection (data not shown). In the presence of
STDs, granulocyte counts were significantly greater at the en-
docervix (3642 vs. 2903/10,000 cells; P Å .01) but not at the
ectocervix or in the vaginal wash.
The significant association of STDs with endocervical CD4
lymphocytes persisted when patients were stratified by age.
Among women õ25 years of age, the 22 women with an STD
had a median cell count of 494 endocervical CD4 lymphocytes/
10,000 cells, compared with 247/10,000 for the 12 women
without one of these infections (P Å .02). Among participants
§25 years of age, the 10 women with an STD had a median
cell count of 456 CD4 lymphocytes/10,000, compared with
218 for the 20 patients without an STD (P Å .006). Within
each of the age categories, median age did not differ signifi-
cantly between those with and without STDs (data not shown).
Patients with chlamydial infection or gonorrhea had signifi-
cantly greater proportions of endocervical CD4 lymphocytes
than did patients without an STD (table 4). Although the data
cytes per 10,000 cells by sexually transmitted infection among en-
Median (interquartile range) of endocervical CD4 lympho-
No Neisseria gonorrhoeae, Chlamydia
trachomatis, or Trichomonas
C. trachomatis and N. gonorrhoeae
N. gonorrhoeae and T. vaginalis
C. trachomatis and T. vaginalis
10,000 cells by sampled sites among patients with and without infec-
tion with nonulcerative STD (gonorrhea, chlamydial infection, or
Median (interquartile range) of CD4 lymphocytes per
32 245 (125–399)
(n Å 32)
(n Å 32)Site sampled
* P Å .008 vs. patients with none of 3 infections.
†P õ .001 vs. patients with none of 3 infections.
Medians were compared by Kruskal-Wallis nonparametric test.
* Medians were compared by Kruskal-Wallis nonparametric test.
†n Å 29.
/ 9d3d$$ja2611-26-97 14:25:14jinfa UC: J Infect
by guest on July 13, 2011
171JID 1998;177 (January)Endocervical CD4 Lymphocytes and STDs
phocytes per 10,000 cells of patients with and without cervicitis,
easily induced cervical bleeding, and blood in endocervical saline
suspensions, by infection with nonulcerative STD (gonorrhea, chla-
mydial infection, or trichomoniasis).
Comparisons of median number of endocervical CD4 lym-
are sparse, mixed infections did not seem to further increase
No significant associations were found between the presence
of Candida species, bacterial vaginosis, M. hominis, or U. urea-
lyticum and any of the white blood cell types for any of the
sites, when analyzed together or stratified by the presence and
absence of chlamydial infection, gonorrhea, and trichomonia-
sis. Significant associations of these latter three STDs with
endocervical CD4 lymphocytes were observed even when data
were stratified by these other genital infections (data not
Patients with nonulcerative STDs had significantly greater
median ratios of endocervical white blood cells to all other
endocervical cells, of endocervical lymphocytes to granulo-
cytes, and of endocervical CD4 lymphocytes to Langerhans’
cells (table 5). There were no significant differences in ratios
of endocervical T lymphocytes to B lymphocytes, endocervical
CD4 lymphocytes to CD8 lymphocytes, or CD4 lymphocytes
Absolute cell counts by infection with STDs.
32 patients with STDs, 14 (44%) had §10,000 total cells in
tube 3 of the endocervical specimen (used to determine the
number of endocervical CD4 lymphocytes), compared with
only 4 (13%) of the 32 patients without STDs (P Å .01).
Among the 46 patients with õ10,000 cells in this tube, the 18
patients with STDs had a median of 3880 cells (interquartile
range, 2315–4558), and the 28 patients without STDs had a
median of 4432 cells (interquartile range, 2645–5575) (P Å
.4). If 10,000 is assumed to be the maximum absolute number
of cells per tube, then the median absolute number of CD4
lymphocytes in tube 3 of the endocervical specimen of patients
with STDs was 243 (interquartile range, 126–540), compared
with 78 for those without STDs (interquartile range, 42–209;
P õ .001). For patients with chlamydial infection only, the
median was 311 (interquartile range, 114–564; P Å .008 com-
pared with no STD); for gonorrhea only, 361 (interquartile
range, 136–511; P Å .001); and for the 3 patients with tricho-
moniasis only, 43 (interquartile range, 39–669; P Å .9). Be-
cause of the high proportions of patients with §10,000 cells
in tubes prepared from the other sampled sites (89% for ecto-
Median no. of CD4
STD No STD
Induced endocervical bleeding
Blood in endocervical cell suspension
Moderate (2/ or 3/)
None/trace (0 or 1/)
* Medians were compared by Kruskal-Wallis nonparametric test.
†4 patients with incomplete observations were excluded (3 with STDs and
1 without STD).
cervical specimens and 100% for vaginal lavage and vaginal
wall specimens), differences in absolute cell numbers between
persons with and without STDs could not be assessed.
In further analyses that controlled for the differences in abso-
lute numbers of endocervical cells counted, the associations
between the CD4 lymphocytes and STDs remained highly sig-
nificant (data not shown).
Associations of cervicitis and induced cervical bleeding with
endocervical CD4 lymphocytes.
ship of clinically apparent cervicitis to endocervical CD4 lym-
phocytes in the presence and absence of nonulcerative STDs.
For this analysis, patients were considered to have cervicitis if
yellow pus was seen on the first endocervical swab, if pus was
seen at the os, or if §30 polymorphonuclear leukocytes per
high-power field were seen on the endocervical Gram’s stain.
Among patients without cervicitis, STDs were associated with
a greater proportion of endocervical CD4 lymphocytes. Among
patients without STDs, cervicitis was also significantly associ-
ated with a greater proportion of endocervical CD4 lympho-
cytes (table 6).
We also examined associations of easily induced cervical
bleeding (indicated by a moderate amount of blood on the first
endocervical swab) and of blood in the endocervical saline
suspension with endocervical CD4 lymphocytes. Eight patients
had easily induced endocervical bleeding; 7 of these had an
STD, and all had clinical cervicitis. Among patients without
STDs, presence of blood in suspension (2/ or 3/) was signifi-
cantly associated with increased endocervical CD4 lympho-
cytes, compared with patients with no or only trace blood (0
We examined the relation-
among STD clinic patients with and without nonulcerative STDs
(gonorrhea, chlamydial infection, or trichomoniasis).
Median ratios between selected endocervical cell types
Ratio of cell typesSTDNo STD
White blood cells/all other cells
T lymphocytes/B lymphocytes
CD4 lymphocytes/CD8 lymphocytes
CD4 lymphocytes/Langerhans cells
NOTE.Medians were compared by Kruskal-Wallis nonparametric test.
/ 9d3d$$ja2611-26-97 14:25:14 jinfa UC: J Infect
by guest on July 13, 2011
172Levine et al.JID 1998;177 (January)
or 1/) (table 6). Presence of trace blood alone was not signifi-
cantly associated with STDs (data not shown). As noted in
table 6, among patients without easily induced bleeding or
without blood in suspension, the association of endocervical
CD4 lymphocytes with STDs remained significant.
Among the 7 patients with none of the three STDs studied,
no evidence of cervicitis, and no blood detected in the speci-
men, the median proportion of endocervical CD4 lymphocytes
was 57 cells/10,000 (interquartile range, 18–130), and the me-
dian absolute number of endocervical CD4 lymphocytes in tube
3 was 18 (interquartile range, 8–59). The median proportion
of ectocervical CD4 lymphocytes was 2 (interquartile range,
1–5), and of vaginal wash CD4 lymphocytes, 0 (interquartile
range, 0–2). Five of these 7 vaginal wash specimens had no
CD4 lymphocytes. Among these 7 patients, 5 had bacterial
vaginosis, 5 were positive for M. hominis, 2 had Candida spe-
cies, and 2 of 6 tested were positive for U. urealyticum.
Patient enrollment category (contact with STD patient vs.
chief complaint of vaginal discharge) was not significantly
associated with presence of clinical cervicitis, with presence
of blood on the swab or in the specimen, or with differences
in endocervical CD4 lymphocytes (data not shown).
Peripheral blood CD4:CD8 ratio and lymphocyte counts.
The patients with STDs had a lower median peripheral blood
CD4:CD8 ratio than did patients without infection (1.5 vs. 1.8;
P Å .01). The absolute peripheral blood CD4 lymphocyte count
was lower among infected patients, but not significantly (878
vs. 1030 cells/mL; P Å .6). The absolute CD8 lymphocyte
count was somewhat higher among patients with STDs (665
vs. 542 cells/mL; P Å .2). This difference in CD4:CD8 ratios
between patients with and without STDs was similar for the
patients who had gonorrhea (1.4 vs. 1.8; P Å .02) and chlamyd-
ial infection (1.5 vs. 1.8; P Å .15), with no similar trend among
the few patients with trichomoniasis. No association was found
between peripheral blood CD4:CD8 ratio and presence of M.
hominis, U. urealyticum, Candida species, or bacterial vag-
cytes per 10,000 cells counted and 3 times the median absolute
number of CD4 lymphocytesin analyzed specimens. Compared
with the subset of patients who had no STD, cervicitis, or blood
in specimens, patients with STDs had Ç8 times the median
proportional number of endocervical CD4 lymphocytes and
ú10 times the median absolute number.
These data must be interpreted cautiously. First, it is un-
known if superficial mucosal white blood cells, sampled ac-
cording to the methods we used, accurately reflect presence of
these cells in the mucosa or submucosa, or if these superficial
cells could actually become infected with HIV and circulate
to local lymph nodes. Second, the few studies that have been
done to investigate potential mechanisms of heterosexual HIV
transmission suggest that Langerhans’ cells and dendritic cells
may be the first cells to become infected with HIV, which then
interact with CD4 cells in draining lymph nodes [23, 24]. Our
data showed no significant increase in the proportion of endo-
cervical Langerhans’ cells in the presence of nonulcerative
STDs but did show an increase in the ratio of CD4 lymphocytes
to Langerhans’ cells, a factor that could possibly influence host
susceptibility to infection. Third, mucosal epithelial cells may
also be directly susceptible to HIV infection . In addition,
we found that patients with STDs had greater numbers of endo-
cervical CD4 lymphocytes that carried CD38, a marker for
cellular activation. Activated CD4 lymphocytes on mucosal
surfaces may increase the infectiousness of persons who are
HIV-positive and may also have implications for susceptibility
The finding that clinically apparent cervicitis was associated
with a greater proportion of endocervical CD4 lymphocytes,
even in the absence of detectable gonorrhea or chlamydial
infection, suggests that this entity may also be associated with
increased susceptibility to HIV infection. Both infectious and
noninfectious factors may be important in the development of
cervical inflammation . Although nongonococcal, nonchla-
mydial cervicitis is common, we know very little about its
etiology, natural history, or role in HIV transmission. The find-
ing that the association of STDs with endocervical CD4 lym-
phocytes persists even when patients with clinical cervicitis,
endocervical friability, or blood in specimens are excluded
suggests that increases in CD4 lymphocytes among patients
with STDs do not occur solely when these factors are present.
C. trachomatis infection appears to increase local secretion of
proinflammatory cytokines , which may recruit leukocytes
to the cervical mucosa; such inflammation may not always be
manifested as clinically apparent cervicitis or friability. How-
ever, the significant association of blood in the specimen with
increased endocervical CD4 lymphocytes, even in the absence
of STDs, emphasizes that systematic measurement of blood
must be done as part of any assessment of mucosal white blood
cells that uses flow cytometry.
In this study, CD4 lymphocytes appeared to be rare in vagi-
nal lavage fluid, even in the presence of STDs. This finding is
Our findingthat HIV-susceptible white bloodcells, including
CD4 lymphocytes, monocytes, and Langerhans’ cells, are rela-
tively more common on the endocervix than on the ectocervix,
on the vaginal wall, or in vaginal lavage fluid suggests that the
endocervix may be a mucosal site that is particularly suscepti-
ble to HIV infection. These findings from flow cytometry are
consistent with previous studies done using immunohistochem-
ical techniques . We also observed greater proportions and
numbers of endocervical CD4 lymphocytes in the presence
of nonulcerative STDs (gonorrhea, chlamydial infection, and
trichomoniasis), suggesting a possible mechanism through
which these STDs may increase susceptibility to HIV infection.
Compared with patients with none of these STDs, the patients
/ 9d3d$$ja26 11-26-97 14:25:14 jinfa UC: J Infect
by guest on July 13, 2011
173 JID 1998;177 (January)Endocervical CD4 Lymphocytes and STDs
to detect any T lymphocytes in the vaginal lavage fluid of the
5 women they studied, except during menses, and also with
the study of Bardeguez et al. , who found no leukocytes
in scrapings of the vaginal wall in 5 women without HIV
infection. Also consistent with the relative paucity of lympho-
cytes, monocytes, and Langerhans’ cells in the vaginal wall
scrapings and in the vaginal wash, Clemetson et al.  found
that in HIV-infected women, HIV DNA was significantly less
likely to be identified in vaginal wall specimens than in endo-
cervical specimens; in that study, the endocervical specimen
was obtained with a swab. In patients with HIV infection,
Bardeguez et al. found no significant difference in CD4 lym-
phocytes between vaginal and cervical specimens sampled with
a wooden spatula or in aspirated cervical mucus, but that study
did not examine endocervical swab or cytobrush specimens. It
is possible that the different specimen collection techniques
yield different findings.
In this sample of patients who were HIV-negative, gonorrhea
and chlamydial infection appeared to be associated with de-
creased peripheral blood CD4:CD8 ratios, with absolute CD4
counts tending to be decreased and CD8 counts to be increased.
This finding is consistent with other reports of decreased
CD4:CD8 ratios among patients with a variety of other acute
and persistent infectious conditions . The potential effects
of chlamydial and gonococcal infection on peripheral blood
lymphocyte numbers and ratios should also be considered in
studies that assess associations of these diseases with mucosal
lymphocytes in persons with HIV infection.
This study had several limitations related to sample size and
patientselection.The samplesizewassmall fordisease-specific
comparisons; specifically, the numbers of patients with tricho-
moniasis and with dual infections were too few to draw conclu-
sions. However, the observations of increased endocervical
CD4 lymphocytes among patients with gonococcal and chla-
mydial infections, when analyzed separately, were significant.
The findings concerning the associations of cervicitis with cell
counts must be considered preliminary, both because of the
small numbers and because of the characteristics of the enrolled
patients. Patients without STDs who were enrolled in this study
were not representative of the general population, as they were
also at increased risk for STDs, and some may have had infec-
tions thatwere notdetected. Thesepatients hada relativelyhigh
rate of cervicitis, which probably diminished the probability of
finding significant differences between groups and may also
have introduced other biases. It is possible that the small subset
of patients without any evidence of STDs, cervicitis, or friabil-
ity better reflects the general population of healthy women, but
even most of these patients tested positive for other organisms.
In this cross-sectional study, our findings are limited to de-
scribing associations; we must therefore be particularly cau-
tious in their interpretation. Longitudinal studies that evaluate
genital mucosal white blood cells among patients before, dur-
ing, and after STDs are acquired and treated would provide
important information on the evolution of inflammatory
changes. A further consideration is that methods for examining
cervicovaginal cell immunophenotypes using flow cytometry
are experimental and more work needs to be done to refine
them. The method of examining cell proportions in female
genital specimens needs to be examined critically and further
comparisons made with absolute cell numbers. To better under-
stand the utility of these methods and mechanisms of HIV
acquisition, it may be particularly useful to directly compare
the results of flow cytometric analyses with colposcopic exami-
nation and with immunohistochemical staining of biopsy speci-
mens from patients with and without STDs.
Thisstudy demonstratesthepotentialfor expandingepidemi-
ologic studies of STD-HIV interactions to include systematic
assessment of mucosal immunity and inflammation by use of
modern immunophenotyping methods. Immunophenotyping of
mucosal cells may also be useful for assessing the effects of
vaginal microbicides on genital mucosa and for investigating
the possible role of other factors on HIV transmission.
We acknowledge the important contributions of George P.
Schmid, who initially proposed the use of flow cytometry for
immunophenotyping of genital exudates in this study, and who
made useful comments on the manuscript; William E. Brady, for
assisting in the design and pilot phase; Riduan Joesoef, for inter-
preting the vaginal Gram’s stains; Stuart M. Berman, for his im-
portant insights regarding the interpretation of these data; Ray-
mond Ransom, for assistance with data management; Michael E.
St. Louis, for critical review of the manuscript; and Stephen A.
Morse, Sandra Larsen, and Carolyn Black, for their thoughtful
suggestions and guidance on the laboratory aspects of this study.
1. Wasserheit JN. Epidemiological synergy. Interrelationships between hu-
man immunodeficiency virus infection and other sexually transmitted
diseases. Sex Transm Dis 1992;19:61–77.
2. Clemetson DBA, Moss GB, Willerford DM, et al. Detection of HIV DNA
in cervical and vaginal secretions. Prevalence and correlates among
women in Nairobi, Kenya. JAMA 1993;269:2860–4.
3. Moss GB, Overbaugh J, Welch M, et al. Human immunodeficiency virus
DNA in urethral secretions in men: association with gonococcal urethri-
tis and CD4 cell depletion. J Infect Dis 1995;172:1469–74.
4. 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. Lancet 1997;349:1868–73.
5. Laga M, Manoka A, Kivuvu M, et al. Non-ulcerative sexually transmitted
diseases as risk factors for HIV-1 transmission in women: results from
a cohort study. AIDS 1993;7:95–102.
6. De Vincenzi I. A longitudinal study of human immunodeficiency virus
transmission by heterosexual partners. European Study Group on Het-
erosexual Transmission of HIV. N Engl J Med 1994;331:341–6.
7. Deschamps MM, Pape JW, Hafner A, Johnson WD Jr. Heterosexual trans-
mission of HIV in Haiti. Ann Intern Med 1996;125:324–30.
8. Kiviat NB, Paavonen JA, Brockway J, et al. Cytologic manifestations of
cervical and vaginal infections. I. Epithelial and inflammatory cellular
changes. JAMA 1985;253:989–96.
/ 9d3d$$ja26 11-26-97 14:25:14 jinfa UC: J Infect
by guest on July 13, 2011
174Levine et al. JID 1998;177 (January) Download full-text
9. Kiviat NB, Paavonen JA, Wølner-Hanssen P, et al. Histopathology of
endocervical infection caused by Chlamydia trachomatis, herpes sim-
plex virus, Trichomonas vaginalis, and Neisseria gonorrhoeae. Hum
10. Bardeguez AD, Skurnick JH, Perez G, Colon JM, Kloser P, Denny TN.
Lymphocyte shedding from genital tract of human immunodeficiency
virus–infected women: immunophenotypic and clinical correlates. Am
J Obstet Gynecol 1997;176:158–65.
11. Gordon FB, Harper IA, Quan AL, Treharne JD, Dwyer RS, Garland JA.
Detection of chlamydia (bedsonia) in certain infections of man. I. Labo-
ratory procedures: comparison of yolk sac and cell culture for detection
and isolation. J Infect Dis 1969;120:451–62.
12. Ripa KT, Mardh PA. Cultivation of Chlamydia trachomatis in cyclohexi-
mide treated McCoy cells. J Clin Microbiol 1977;6:328–31.
13. Martin JE Jr, Lewis JS. Anisomycin: improved antimycotic activity in
modified Thayer-Martin medium. Public Health Lab 1977;35:53–62.
14. Kreiger JN, Tam MR, Stevens CE, et al. Diagnosis of trichomoniasis.
Comparison of conventional wet-mount examination with cytologic
studies, cultures, and monoclonal antibody staining of direct specimens.
15. Tully JG, Whitcomb RF, Clark HF, Williamson DL. Pathogenic mycoplas-
mas: cultivation and vertebrate pathogenicity of a new spiroplasma.
16. Shepard MC. Culture medium for ureaplasmas. In: Tully JG, Razin S, eds.
Methods in mycoplasmology. Vol I. New York: Academic Press, 1983:
17. Shepard MC, Lunceford CD. Serological typing of Ureaplasma urealyti-
cum isolates from urethritis patients by an agar growth inhibition
method. J Clin Microbiol 1978;8:566–74.
18. Blanchard A, Yanez A, Dybvig K, Watson HL, Griffiths G, Cassell GH.
Evaluation of intraspecies genetic variation within the 16S rRNA gene
of Mycoplasma hominis and detection by polymerase chain reaction. J
Clin Microbiol 1993;31:1358–61.
19. Blanchard A, Hentschel J, Duffy L, Baldus K, Cassell GH. Detection of
Ureaplasma urealyticum by polymerase chain reaction in the urogenital
tract of adults, in amniotic fluid, and in the respiratory tract of newborns.
Clin Infect Dis 1993;17(suppl 1):S148–53.
20. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial
vaginosis is improved by a standardized method of Gram stain interpre-
tation. J Clin Microbiol 1991;29:297–301.
21. Pope V, Larsen SA, Rice RJ, Goforth SN, Parham SE, Fears MB. Flow
cytometric analysis of peripheral blood lymphocyte immunophenotypes
in persons infected with Treponema pallidum. Clin Diagn Lab Immunol
22. Edwards JNT, Morris HB. Langerhans’ cells and lymphocyte sub-
sets in the female genital tract. Br J Obstet Gynecol 1985;92:
23. Spira AI, Marx PA, Patterson BK, et al. Cellular targets of infection and
route of viral dissemination after an intravaginal inoculation of simian
immunodeficiency virus into rhesus macaques. J Exp Med 1996;183:
24. Ayehunie S, Groves RW, Bruzzese AM, Ruprecht RM, Kupper TS,
Langhoff E. Acutely infected Langerhans cells are more efficient
than T cells in disseminating HIV type 1 to activated T cells follow-
ing a short cell-cell contact. AIDS Res Hum Retroviruses 1995;11:
25. Tan X, Pearce-Pratt R, Phillips DM. Productive infection of a cervical
epithelial cell line with human immunodeficiency virus: implications
for sexual transmission. J Virol 1993;67:6447–52.
26. Cohen MS, Heine RP. Mucosal defenses and microbial pathogens of the
genital tract. In: Quinn TC, ed. Sexually transmitted diseases. New
York: Raven Press, 1992:39–55.
27. Critchlow CW, Wo ¨lner-Hanssen P, Eschenbach DA, et al. Determinants
of cervical ectopia and of cervicitis: age, oral contraception, specific
cervical infection, smoking, and douching. Am J Obstet Gynecol 1995;
28. Rasmussen SJ, Eckmann L, Quayle AJ, et al. Secretion of proinflammatory
cytokines by epithelial cells in response to chlamydia infection suggests
a central role for epithelial cells in pathogenesis. J Clin Invest 1997;
29. Hill JA, Anderson DJ. Human vaginal leukocytes and the effects of vaginal
fluid on lymphocytes and macrophage defense functions. Am J Obstet
30. Laurence J. T-cell subsets in health, infectious disease, and idiopathic
CD4/T lymphocytopenia. Ann Intern Med 1993;119:55–62.
/ 9d3d$$ja26 11-26-97 14:25:14jinfa UC: J Infect
by guest on July 13, 2011