INFECTION AND IMMUNITY, Mar. 2002, p. 1538–1546
0019-9567/02/$04.00?0 DOI: 10.1128/IAI.70.3.1538–1546.2002
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Vol. 70, No. 3
Chemokine Expression Patterns Differ within Anatomically Distinct
Regions of the Genital Tract during Chlamydia trachomatis Infection
Heather K. Maxion and Kathleen A. Kelly*
Department of Pathology and Laboratory Medicine, UCLA Medical Center, Los Angeles, California 90095-1732
Received 15 August 2001/Returned for modification 2 October 2001/Accepted 27 November 2001
Untreated infections with Chlamydia trachomatis commonly result in ascending infection to fallopian tubes
and subsequent immune-mediated tubal pathology in females. The proposed immune-mediated injury may be
associated with the increased recruitment of CD4 cells to the upper genital tract (GT) (oviducts) in comparison
to the lower GT (cervix) during infection, as shown in animal models. To understand the mechanisms
responsible for this biased recruitment of CD4 cells within the GT, we characterized chemokine expression
patterns in the upper and lower GTs in mice during infection with the murine pneumonitis biovar of Chlamydia
trachomatis. Enzyme-linked immunosorbent assays of supernatants from GT homogenates revealed that the
levels of the Th1-associated chemokines CXCL9 (monokine induced by gamma interferon), CXCL10 (inter-
feron-inducible protein 10), and CCL5 (RANTES) were significantly higher in the upper GT than in the lower
GT after infection, while the CCL3 (macrophage inflammatory protein 1?) level was not increased. In contrast,
the level of chemokine CCL11 (eotaxin) was significantly elevated in the lower GT later in the course of
infection. Increased levels of mRNA confirmed the selective differences in chemokine expression within the
upper and lower GTs. The increased levels of Th1-inducible chemokines in the upper GT were not due to
differences in the magnitude of infection or progesterone pretreatment. These data demonstrate that the upper
and lower regions of the GT respond differently to Chlamydia infection.
The gram-negative bacterium Chlamydia trachomatis is sex-
ually transmitted and infects squamous epithelial cells in the
female cervix. From here, the bacteria ascend and establish
infection in columnar cells of human fallopian tubes or ovi-
ducts in mice. Left untreated, human infection can lead to
pelvic inflammatory disease, fallopian tube injury, and infertil-
ity (22). Oviduct injury, characterized by hydrosalpinx, and
infertility are also seen in murine infections (33). The devel-
opment of T-cell-mediated immunity is one of the most crucial
elements required for the effective clearance of this pathogen.
A CD4 Th1 response is necessary for Chlamydia eradication,
and this finding has been demonstrated by prolonged infection
in gamma interferon knockout mice (23) as well as in mice
injected with blocking antibodies against the Th1-inducing cy-
tokine interleukin 12 (IL-12) (23). Additionally, the ability to
clear infection in nude mice is restored following the adoptive
transfer of an anti-C. trachomatis murine pneumonitis biovar
(MoPn) CD4 Th1 clone (15). Similarly, IL-10 knockout mice
exhibit a shorter course of infection (37). Thus, the regulation
of Th1 and Th2 responses in the genital tract (GT) during
Chlamydia infection is a crucial factor controlling the duration
of infection and subsequent tubal pathology.
Chemokines are a rapidly growing family of small chemo-
tactic molecules that are specific for various subsets of lym-
phocytes as well as other types of leukocytes. Increasing evi-
dence suggests that chemokines play an important role in the
regulation of Th1 and Th2 responses in vivo. These responses
appear to be directed by the differential expression of chemo-
kine receptors on Th1- and Th2-cell subsets (21). Many studies
have demonstrated patterns of either Th1- or Th2-associated
chemokines in diseased tissues previously shown to contain
large infiltrates of either Th1 or Th2 cells (9). For instance, in
the Th1-mediated disease multiple sclerosis, high levels of che-
mokines CXCL10 (interferon-inducible protein 10 [IP-10]),
CXCL9 (monokine induced by gamma interferon [MIG]), and
CCL5 (RANTES) are found in the cerebrospinal fluid (30).
These data suggest that the chemokine profile plays a central
role in determining the predominant T-cell subset associated
with a particular disease or infection.
Chemokines also provide fine specificity for the direction of
cellular recruitment to discrete anatomical regions within a
given tissue. For example, site specificity has been noted at
mucosal surfaces, where CCL25 (thymus-expressed chemo-
kine) has been shown to localize to the epithelium of the small
intestine but not the large intestine (19). Many chemokines
have been detected in the endometrial epithelium within the
female GT in humans, including CCL3 (macrophage inflam-
matory protein 1? [Mip-1?]) (1), CCL5 (RANTES) (2), CCL2
(monocyte chemotactic protein 1 [MCP-1]) (4), and CCL11
(eotaxin) (13). However, it is not known whether chemokine
expression differs within functionally discrete regions of the
GT. It was previously shown that a significantly larger number
of CD4 cells are recruited to the oviducts (upper GT) than to
the cervical-vaginal region (lower GT) of mice infected with
MoPn (18). To investigate the basis for the increased recruit-
ment of CD4 cells to the upper GT, we evaluated the expres-
sion of chemokines associated with Th1 and Th2 responses in
the upper and lower GTs during infection.
MATERIALS AND METHODS
Infection. Female BALB/c mice, 4 to 6 weeks old, were purchased from Harlan
Sprague-Dawley (Indianapolis, Ind.) and were housed according to American
* Corresponding author. Mailing address: UCLA Medical Center,
Department of Pathology and Laboratory Medicine, 10833 Le Conte
Ave., Mailroom A2-179 CHS, Los Angeles, CA 90095-1732. Phone:
(310) 206-5562. Fax: (310) 794-4863. E-mail: firstname.lastname@example.org.
Association of Accreditation of Laboratory Animal Care guidelines. Experimen-
tal procedures were approved by the UCLA Institutional Animal Care and Use
Committee. All mice were first injected subcutaneously with 2.5 mg of medroxy-
progesterone acetate (Depo-Provera; Upjohn, Kalamazoo, Mich.) in 100 ?l of
sterile phosphate-buffered saline. Medroxyprogesterone acetate drives mice into
a state of anestrus, thus eliminating the variability in the rate and severity of
infection due to the estrous cycle (26). Seven days later, while under sodium
pentobarbital anesthesia, all mice were inoculated with 107inclusion-forming
units (IFU) (50% infective dose, 1.5 ? 103IFU) of MoPn grown in McCoy cells.
Mice were killed on days 3, 7, 14, 21, and 35 after inoculation. Infection was
monitored by examining cervical-vaginal swab samples (Dacroswab type 1; Spec-
trum Labs, Houston, Tex.) obtained immediately before mice were killed and
homogenates of tissue samples obtained from the upper and lower GTs. Swab
samples were stored in sucrose-phosphate buffer at ?70°C until analyzed. Tissue
homogenate samples were also frozen at ?70°C until analyzed.
Tissue homogenates. GT tissues were divided into the cervical-vaginal region
(lower GT) and oviducts (upper GT) with the ovaries removed. Uterine horns
were not included in our analysis. Tissue sections from individual mice were
placed in 1 ml of a protease inhibitor buffer (1 ?g each of antipain, aprotinin,
leupeptin, and pepstatin A/ml and 2 mM phenylmethylsulfonyl fluoride in sterile
phosphate-buffered saline) (Sigma, St. Louis, Mo.) and homogenized as previ-
ously described (3) by using a hand-held homogenizer (Omni International,
Warrenton, Va.). Aliquots of each homogenate were removed for isolation of
chlamydiae. The remaining homogenate volumes were sonicated at 20 kHz for 1
min and then centrifuged at 900 ? g for 15 min at 10oC to remove cellular debris.
Supernatants were filtered through 0.2-?m-pore-size Acrodisks (Gelman Sci-
ences, Ann Arbor, Mich.) to remove chlamydiae, and samples were stored at
?70oC until analyzed.
Isolation of chlamydiae from cervical-vaginal swab and tissue homogenate
samples. Swab samples were prepared as previously described. McCoy cell
monolayers in individual wells of 96-well plates were inoculated with 200 ?l of
the solution from vaginal swabs or homogenized GT tissue (18), followed by
centrifugation at 1,900 ? g for 1 h. The plates were incubated for 2 h at 37°C. At
this time, the isolation solutions were removed, fresh cycloheximide medium was
added, and the plates were incubated for an additional 32 h. The cultures were
then fixed with methanol. MoPn inclusions were identified by the addition of
anti-MoPn immune sera and anti-mouse immunoglobulin G conjugated to flu-
orescein isothiocyanate (ICN Immunobiologicals, Irvine, Calif.). The inclusion
bodies within 20 fields (?40) were counted under a fluorescence microscope, and
numbers of IFU per milliliter were calculated (17). Data were adjusted for IFU
per milligram of crude homogenized GT tissue (upper or lower).
Chemokine ELISAs. Recombinant protein and antibodies against CCL3 (Mip-
1?), CCL11 (eotaxin), CXCL9 (MIG), CXCL10 (IP-10), and CCL5 (RANTES)
were purchased from R&D Systems (Minneapolis, Minn.) and those against
CCL2 (MCP-1) were purchased from PharMingen (San Diego, Calif.) for use in
enzyme-linked immunosorbent assays (ELISAs). Upper and lower GT homog-
enates were added to duplicate wells of microtiter enzyme immunoassay plates
(Costar/Corning, Acton, Mass.) and assayed according to the manufacturer’s
protocol with the following exceptions. CXCL10 and CXCL9 primary antibody
concentrations were 1 and 2 ?g/ml, respectively, and secondary antibody con-
centrations were 0.5 ?g/ml. The recommended substrate was replaced with
1-StepTM Turbo TMB-ELISA substrate (Pierce Chemical Co., Rockford, Ill.).
The optical densities were read at 450 nm with a microplate reader (model 550;
Bio-Rad, Hercules, Calif.). Chemokine values were determined from a standard
curve generated with recombinant chemokines by using microplate reader soft-
ware. Chemokine values were corrected for total protein by using a micro-
bicinchoninic acid protein assay kit (Pierce).
Serum progesterone levels. Serum was collected from mice administered or
not administered medroxyprogesterone acetate. Progesterone levels were deter-
mined by a competitive electrochemiluminescence immunoassay with an Elecsys
2010 automated analyzer (Roche, Berkeley, Calif.).
mRNA isolation and SuperArray analysis. Total RNA was isolated from
paired GT tissues of mice according to the manufacturer’s protocol following
homogenization of tissues in RNAzol B (Tel-Test, Inc., Friendswood, Tex.) and
stored at ?80°C until use. Nonrad-GEArray kits specific for chemokine analysis
were purchased from SuperArray Inc. (Bethesda, Md.). Each kit provides a
matched set of membranes containing 23 chemokines plus controls. Probe syn-
thesis was carried out by using 10 or 7 ?g of mRNA per sample. The manufac-
turer’s protocol was followed for all steps. Following substrate addition, mem-
branes were exposed to X-ray film (Fuji, Tustin, Calif.) for 5 to 10 min. Data were
quantified by using a laser densitometer and ImageQuaNT software (Molecular
Dynamics, Sunnyvale, Calif.) to calculate the average integrated volumes of dots.
Data were expressed as the average integrated volume of a sample relative to the
average integrated volume of a positive control (glyceraldehyde-3-phosphate
Immunohistochemical analysis. Tissues were harvested and prepared as pre-
viously described (18). Staining was carried out as previously described with the
following exceptions. After a tissue blocking step with rabbit serum, the primary
antibodies (goat anti-mouse CXCL10 [IP-10] and goat anti-mouse CCL11
[eotaxin]) (R&D Systems) were incubated on tissue sections for 45 min at room
temperature in a humidified chamber, and then the sections were washed. A
rabbit anti-goat immunoglobulin G antibody conjugated to biotin at 30 ?g/ml
(Antibodies Inc., Davis, Calif.) and streptavidin conjugated to horseradish per-
oxidase (Zymed, San Francisco, Calif.) were added next, and the tissue sections
were incubated for 45 min. Slides were developed as previously described (18).
Photographs were generated by scanning the microscope slides with an Olympus
DP10 color digital video camera.
Statistics. Statistical differences in chemokine protein levels were tested by
using two-way analysis of variance (ANOVA) and Tukey’s post hoc test. Statis-
tical differences in chemokine message levels were determined by using a paired
Student t test. The above statistical tests were suggested by and performed with
SigmaStat software based on the distribution of the data and sample size (Jandel
Scientific, San Rafael, Calif.). Groups were considered statistically different at a
P value of ?0.05.
Differential expression of chemokines in the upper and
lower GTs in response to C. trachomatis infection. To deter-
mine whether chemokine production differed between the up-
per and lower GTs during MoPn infection, we measured pro-
tein levels in tissue homogenates at weekly intervals that
spanned the induction phase (0 to 14 days) and the resolution
phase (14 to 35 days) of infection. We evaluated the expression
of chemokines that are generally associated with a Th1 re-
sponse, CXCL10 (IP-10), CXCL9 (MIG), CCL3 (Mip-1?),
and CCL5 (RANTES), or a Th2 response, CCL11 (eotaxin).
We also evaluated CCL2 (MCP-1), which has not been shown
to associate with any particular T-cell subset. As shown in Fig.
1A, the Th1-associated chemokines CXCL10 (IP-10), CXCL9
(MIG), and CCL5 (RANTES) were all induced by infection in
the upper GT. CXCL10 (IP-10) was measured at a significantly
elevated level on day 3 compared to the level in uninfected
mice. The level of CXCL9 (MIG) was significantly elevated on
day 7 and again later in infection, on day 35, whereas an
elevated level of CCL5 (RANTES) was maintained in the
upper GT from day 7 throughout the course of infection. In the
lower GT (Fig. 1B), the levels of both CXCL9 (MIG) and
CCL5 (RANTES) were elevated early in the course of infec-
tion and the CCL5 (RANTES) level was significantly increased
late in the course of infection compared to the results for
controls. Finally, CCL3 (Mip-1?) was not induced by infection
in either the upper or the lower GT. However, CCL3 (Mip-1?)
was expressed at a constitutively higher level in the upper GT
than in the lower GT.
We next examined the protein levels of CCL11 (eotaxin) and
CCL2 (MCP-1) in the upper and lower GTs with ELISAs. As
shown in Fig. 2A, CCL11 (eotaxin) expression was induced by
infection in the upper GT but only later in the course of
infection (day 14). Furthermore, the level of CCL11 (eotaxin)
was comparatively lower than the peak levels of Th1-associated
chemokines in the upper GT (Fig. 1A). CCL11 (eotaxin) was
also induced in the lower GT (Fig. 2B) by day 14 and, inter-
estingly, was expressed to a significantly higher degree in the
lower GT than in the upper GT during the resolution phase of
infection (Fig. 2). There was no difference in CCL2 (MCP-1)
expression between uninfected and infected tissues and be-
VOL. 70, 2002 CHEMOKINE EXPRESSION AND CHLAMYDIA GENITAL INFECTION1539
tween the upper and lower GTs. Taken together, these data
show that the overall production of chemokines was greater in
the upper GT than in the lower GT during infection. Further-
more, higher concentrations of Th1-associated chemokines
were found in the upper GT, while the concentration of the
Th2-associated chemokine CCL11 (eotaxin) was higher in the
lower GT than in the upper GT.
Course of C. trachomatis infection. To rule out the possibility
that a larger bacterial load in the upper GT could account for
differences in chemokine production, we quantitated chlamyd-
iae from GT homogenates. As shown in Fig. 3, the magnitudes
of infection were similar for both GT regions. In fact, larger
numbers of chlamydiae were detected in the lower GT 3 days
following infection, a time when elevated chemokine levels
were seen in the upper GT. Resolution of infection began in
both regions between days 14 and 21 but appeared to occur
more rapidly in the lower GT than in the upper GT. Thus, the
levels of infection were similar in the upper and lower GTs
during the induction phase of the infection, when marked
differences in chemokine levels appeared.
Effect of progesterone on chemokine expression. To ensure
that the administration of progesterone 7 days prior to MoPn
FIG. 1. Th1-associated chemokine protein levels during MoPn in-
fection. Chemokines were measured in upper and lower GT tissue
homogenates in uninfected (day 0) and infected (days 3, 7, 14, 21, and
35) mice by ELISAs. The data are compiled from two separate exper-
iments and are expressed as the mean and standard error of the mean
(SEM) for four mice per group. Number signs indicate values that
were significantly elevated in the upper GT versus the lower GT
(CXCL10, days 3 and 7; CXCL9, days 3 and 35; CCL5, days 7, 14, 21,
and 35; and CCL3, days 0, 3, 14, 21, and 35). Plus signs indicate values
that were significantly different from those on day 0 in the upper GT
(CXCL10, day 3; CXCL9, days 7 and 35; and CCL5, days 7, 14, 21,
and 35). Asterisks indicate values that were significantly different from
those on day 0 in the lower GT (CXCL9, day 7; and CCL5, days 7 and
35). The P value determined by ANOVA was ?0.005; the P value
determined by the post hoc Tukey test was ?0.05 ? SEM.
FIG. 2. CCL11 and CCL2 protein levels during MoPn infection.
Chemokines were measured in upper and lower GT tissue homoge-
nates in uninfected (day 0) and infected (days 3, 7, 14, 21, and 35) mice
by ELISAs. The data are compiled from two separate experiments and
are expressed as the mean and standard error of the mean (SEM) for
four mice per group. A number sign indicates a value that was signif-
icantly elevated in the lower GT versus the upper GT (CCL11, days 21
and 35). A plus sign indicates a value that was significantly different
from those on day 0 in the upper GT (CCL11, day 14). Asterisks
indicate values that were significantly different from those on day 0 in
the lower GT (CCL11, days 14 and 21). The P value determined by
ANOVA was ?0.005; the P value determined by the post hoc Tukey
test was ?0.05 ? SEM.
1540MAXION AND KELLYINFECT. IMMUN.
infection did not influence chemokine expression profiles in
the upper and lower GTs, we used ELISAs to measure the
levels of chemokines CXCL10 (IP-10) and CCL11 (eotaxin) in
the upper and lower GTs in medroxyprogesterone acetate-
treated, uninfected mice at weekly intervals matching those of
infected mice. We found that baseline levels of CXCL10 (IP-
10) and CCL11 (eotaxin) remained unchanged for at least 6
weeks after progesterone injection (data not shown). We also
measured serum hormone levels in uninfected mice following
the medroxyprogesterone acetate treatment; serum progester-
one levels were elevated and fluctuated between 300 and 1,500
pg/ml, while serum estradiol concentrations were less than 10
pg/ml during the same time period (data not shown). There-
fore, the increases in chemokine expression seen in Fig. 1 and
2 are a result of inoculation with MoPn and not hormonal
Measurement of chemokine message levels by SuperArray
analysis. We further confirmed our finding of differential che-
mokine expression between the upper and lower GTs by mea-
suring mRNA levels by SuperArray analysis. The SuperArray
system is designed to semiquantitatively compare the levels of
mRNA expression of two matched samples by using paired
membranes containing equal amounts of each probe. In addi-
tion, the SuperArray assay enabled us to measure mRNA
expression for 17 additional chemokines. The comparison of
mRNAs isolated from infected (day 7) and uninfected upper
GT tissues showed the expression of a number of chemokines
(Fig. 4). Notable increases were seen in the intensities of spots
for CXCL10 (IP-10) (spots 3A and 3B) and CXCL9 (MIG)
(spots 5A and 5B) in infected tissues compared to uninfected
To quantitate the data, a laser densitometer was used to
determine the average integrated intensity of each dot. Aver-
age integrated intensities of duplicate chemokine dots were
compared with the average integrated intensity of six GAPDH
dots for each paired hybridization experiment. As shown in
Fig. 5A, a significant increase in CXCL10 (IP-10) and CXCL9
(MIG) mRNA levels was found in infected upper GT tissues
compared to uninfected tissues and was consistent with signif-
icantly elevated protein levels. The level of CCL3 (Mip-1?)
mRNA was also increased in infected tissues, although not
significantly. Surprisingly, only a moderate increase in the level
of CCL5 (RANTES) mRNA was detected, despite an elevated
protein level. Finally, we did not observe a difference in mRNA
expression for CCL11 (eotaxin) and CCL2 (MCP-1) in the
upper GT following infection. These data further support the
differential expression of Th1-associated chemokines CXCL10
(IP-10) and CXL9 (MIG) following C. trachomatis infection in
the upper GT.
The microarray analysis revealed the expression of other
chemokines not evaluated at the protein level. The levels of
FIG. 3. Quantitation of viable chlamydiae in different regions of
the GT throughout the course of MoPn infection. The lower and upper
GT regions from individual mice were homogenized and cultured for
the isolation of chlamydiae. The data are compiled from two separate
experiments and are expressed as the mean and standard error of the
mean (SEM) for seven or eight mice per group. In the upper GT,
protein levels ranged from 0.25 mg/ml in uninfected mice to 2 mg/ml
between days 7 and 14 of infection. Lower GT protein levels ranged
from 1.1 to 2.2 mg/ml in uninfected and infected mice. A number sign
indicates a value that was significantly elevated in the lower GT versus
the upper GT (the ANOVA P value was ?0.001; the post hoc Tukey
test P value was ?0.05 ? SEM). Asterisks indicate values that were
significantly elevated in the upper GT versus the lower GT (the
ANOVA P value was ?0.001; the post hoc Tukey test P value was
?0.05 ? SEM).
FIG. 4. SuperArray analysis of uninfected and infected (day 7)
upper GT tissues. (A) Total RNA was isolated from the oviducts of
three mice, and a 10-?g pool of RNAs (3.3 ?g/mouse) was converted
to cDNA by using biotinylated dUTP. Hybridized products were de-
tected by using avidin-alkaline phosphatase and a chemiluminescent
substrate. (B) Chemokine template. GAPDH and ?-actin served as
internal controls (spots 3G to 8G and 8E and F, respectively). Bacterial
plasmid pUC18 (1G and 2G) served as a negative control.
VOL. 70, 2002 CHEMOKINE EXPRESSION AND CHLAMYDIA GENITAL INFECTION 1541
chemokines CCL21 (secondary lymphoid tissue chemokine
[SLC]), growth-regulated oncogene 1 (Gro-1), T-cell activation
gene 3 (TCA-3), and XCL1 (lymphotactin) were elevated in
upper GT tissues from infected mice compared to uninfected
mice (Fig. 5B). We also detected mRNA expression of CXCL4
(platelet factor 4), CXCL12 (stroma-derived factor 1?/? [SDF-
1?/?]), and SDF-2. However, the levels of expression were
similar between infected and uninfected tissues, suggesting
that these chemokines may be constitutively expressed in mu-
We compared the levels of chemokine mRNAs between the
upper and lower GTs during infection (Fig. 6A). Data com-
piled from two experiments showed a significant increase in the
CXCL9 (MIG) mRNA level in the upper GT compared to the
lower GT 7 days after infection. Similarly, CXCL10 (IP-10)
and CCL3 (Mip-1?) mRNA levels were elevated in the upper
GT in comparison to the lower GT. We again observed only
low levels of CCL5 (RANTES) mRNA, which did not differ
between the upper and lower GTs. Also, the levels of mRNA
for CCL11 (eotaxin) and CCL2 (MCP-1) were slightly higher
in the lower GT than in the upper GT of infected mice. As
noted for the previous microarray analysis, CCL21 (SLC),
CXCL12 (SDF-1?/?), and SDF-2 mRNAs were again found in
the upper and lower GTs (data not shown), but the levels did
not differ significantly between the tissues.
FIG. 5. Quantitation of mRNA data obtained by SuperArray anal-
ysis of uninfected and infected (day 7) upper GT tissues as pictured in
Fig. 4. Spot intensities were determined by laser densitometry and by
using ImageQuaNT software. Data are expressed as the mean (Ave.)
integrated volume (Int. Vol.) of duplicate chemokine spots relative to
the mean integrated volume of six GAPDH spots; error bars indicate
the standard error of the mean (SEM). (A) Chemokines previously
measured by ELISAs. Asterisks indicate values that were significantly
elevated in infected upper GTs (the Student’s t test P value was ?0.05
? SEM). (B) Other chemokines with detectable message levels.
FIG. 6. Quantitation of mRNA data obtained by SuperArray anal-
ysis from infected upper and lower GT tissues. Spot intensities were
determined by laser densitometry and by using ImageQuaNT software.
Data are expressed as the mean (Ave.) integrated volume (Int. Vol.) of
duplicate chemokine spots relative to the mean integrated volume of
six GAPDH spots; error bars indicate the standard error of the mean
(SEM). (A) Day 7 of infection. Results are an average from two
experiments with 10 or 7 ?g of mRNA. An asterisk indicates a value
that was significantly elevated in the upper GT versus the lower GT
(the ANOVA P value was ?0.001; the post hoc Tukey P value was
?0.05 ? SEM). (B) Day 21 of infection.
1542 MAXION AND KELLYINFECT. IMMUN.
Finally, we compared mRNA expression levels in the upper
and lower GTs of day 21 infected mice (Fig. 6B). In general, we
found that mRNA levels had decreased by this time point, as
was observed for chemokine protein levels. Specifically, mes-
sage levels for CXL10 (IP-10) and CXCL9 (MIG), which
peaked between days 3 and 7 of infection, returned to the
levels seen in uninfected mice by day 21. Likewise, mRNA
levels for CCL5 (RANTES), CCL11 (eotaxin), and CCL2
(MCP-1) also dropped to almost undetectable values in both
the upper and the lower GTs. CCL3 (Mip-1?) was the only
chemokine with message levels that remained elevated in both
the upper and the lower GTs on day 21. Together, these data
support the notion of differential chemokine expression be-
tween the upper and lower GTs during Chlamydia infection.
Localization of chemokines by immunohistological analysis.
To determine which cells within the GT are responsible for
chemokine production, we used immunohistochemical analysis
to identify CXCL10 (IP-10)- and CCL11 (eotaxin)-producing
cells in the upper and lower GTs. As shown in Fig. 7, CXCL10
(IP-10) was found on columnar epithelial cells, endothelial
cells, and stromal cells within the oviduct (Fig. 7, upper left
panel). Following infection, the same cell types stained posi-
tively for CXCL10 (IP-10) but with greater intensity on day 7
(Fig. 7, upper middle panel). In the lower GT region, CCL11
(eotaxin) staining was not found in uninfected mice (Fig. 7,
lower left panel), but squamous epithelial cells stained posi-
tively on days 7 (data not shown) and 21 (Fig. 7, lower middle
panel). Interestingly, CXCL10 (IP-10) staining in the lower GT
was also confined to squamous epithelial cells (data not
shown). These data suggest that the high CXCL10 (IP-10)
protein levels noted in the upper GT may result from increased
production by multiple cell types that are not associated with
an inflammatory response, while in the lower GT, chemokine
production is confined to the epithelium following infection.
The expression of chemokines within tissues regulates the
recruitment of specific subsets of lymphocytes to distinct tissue
sites. Chemokines are therefore responsible, in part, for direct-
ing the immune response that ensues following bacterial inva-
sion. Our results are the first to demonstrate that there are
regional differences in chemokine expression within the female
reproductive tract in response to Chlamydia infection. Al-
though homeostatic differences in CCL25 (thymus-expressed
chemokine) expression have been demonstrated within regions
of the intestinal tract (19), this is the first report demonstrating
regional chemokine differences in response to infection. Stud-
ies measuring cellular influx and adhesion molecule expression
during Chlamydia infection first suggested that there were re-
FIG. 7. Cellular localization of CXCL10 and CCL11. Immunohistochemical staining of CXCL10 in the upper GT (upper panels) and CCL11
in the lower GT (lower panels) in uninfected (day 0) and infected (day 7 or 21) mice is shown. Arrowheads denote columnar (upper GT) or
squamous (lower GT) epithelial cells, closed arrows denote endothelial cells, and open arrows denote stromal cells. Magnification, ?400. A low
magnification was used to emphasize the differences in cell types producing CXCL10 and CCL11.
VOL. 70, 2002CHEMOKINE EXPRESSION AND CHLAMYDIA GENITAL INFECTION1543
gional differences in the immune response between the cervi-
cal-vaginal region and oviducts of mice (18, 25). Our data
further support this theory, as we found that chemokines as-
sociated with Th1 responses were present at significantly
higher levels in the oviducts than in the cervical-vaginal tissues
of mice during infection. Namely, CXCL10 (IP-10) and
CXCL9 (MIG) protein levels peaked early in infection in the
upper GT and then returned to the baseline, whereas the
CCL5 (RANTES) level remained elevated for the duration of
infection. These results were confirmed at the mRNA level,
although CXCL9 (MIG) but not CXCL10 (IP-10) mRNA lev-
els were significantly higher in the upper GT than in the lower
GT. This finding may have been due to the fact that the
CXCL10 (IP-10) protein level in the upper GT peaked at a
time point earlier than the one at which we measured mRNA.
Although we have not yet confirmed the functions of these
chemokines, there are data to suggest that the concentrations
reached during infection are sufficient for lymphocyte recruit-
ment (29, 32). Experiments are under way to demonstrate that
the induction of CXCL9 (MIG), CXCL10 (IP-10), and/or
CCL5 (RANTES) is responsible for the selective recruitment
of Th1 cells to the upper GT during infection.
Compared to the results for the upper GT, the chemokine
expression patterns differed quantitatively and kinetically in
the cervical-vaginal region during infection. First, only low
levels of Th1-associated chemokines were present in the lower
GT. Second, CCL11 (eotaxin) levels were significantly in-
creased late in the course of infection. Immunohistochemical
staining supported these findings by showing that CCL11
(eotaxin) expression was confined to epithelial cells during the
resolution phase of infection (day 21). However, the mRNA
expression of CCL11 (eotaxin) increased in the lower GT rel-
ative to the upper GT early after infection but not at later time
points, when the expression of CCL11 (eotaxin) protein was
significantly elevated. Although mice cleared infection in the
lower GT, the diminished production of Th1-associated che-
mokines in that region may have been responsible for the
reduced numbers of CD4 cells observed in the lower GT. It is
possible that ascending infection correlates with smaller num-
bers of CD4 Th1 cells in the lower GT.
Considering the anatomical and functional differences be-
tween the oviducts and cervical regions of the GT, it is not
unanticipated to find immunologically distinct responses at
these sites. For instance, epithelial cells are different at the two
sites. Squamous epithelial cells are found in the cervical region,
while ciliated columnar epithelial cells line the oviducts. Epi-
thelial cells play a central role in directing the immune re-
sponse, since they host Chlamydia and secrete cytokines, such
as IL-8, early after infection (27). Moreover, endocervical but
not endometrial cell lines secrete IL-8 in response to Chla-
mydia infection (35), suggesting that epithelial cells at these
discrete sites respond differently to infection. In addition, we
found that CXCL10 (IP-10) was expressed on a wider array of
cell types in the upper GT than in the lower GT, further
supporting the concept that chemokine secretion differs be-
tween the upper and lower GTs.
The differences in chemokine expression in the upper and
lower GTs cannot be explained by simple differences in the
level of infection between these two regions. Our data show
that the level of infection was significantly higher in the lower
GT early in infection, at a time when the levels of Th1-asso-
ciated chemokines were significantly higher in the upper GT.
Likewise, chlamydia levels were similar in the upper and lower
GTs during the second week of infection, although resolution
of infection occurred more quickly in the lower GT. By day 35,
the lower GT was negative for chlamydiae, while the upper GT
was either negative or had minimal numbers of inclusions.
These results are similar to those previously reported (18),
although our data indicate more variability in the numbers of
chlamydiae detected in the upper and lower GTs throughout
infection and suggest that there may be a small lag in the
clearance of chlamydiae from the upper GT. Also, to rule out
the possible influence of inoculating dose on chemokine levels,
we found no differences in the levels of CXCR10 (IP-10) and
CCL11 (eotaxin) in mice infected with 1.5 ? 105IFU of MoPn
(data not shown). These data, coupled with the results of the
immunohistochemical analysis showing that chemokine ex-
pression occurs in noninflammatory cell types early after in-
fection, suggest that chemokine expression in the upper GT
precedes the recruitment of inflammatory cells and is not in-
fluenced by the inoculating dose. These conclusions are not
surprising, since all somatic cells produce chemokines and in
other models, chemokine production has been shown to pre-
cede the influx of inflammatory cells (20).
Our results showing that steady, basal levels of CXCL10
(IP-10) and CCL11 (eotaxin) are maintained in the upper and
lower GTs of uninfected mice treated with medroxyprogester-
one acetate indicate that the chemokine differences seen be-
tween the upper and lower GTs of infected mice are not due to
progesterone treatment. Female reproductive hormones have
been reported to alter cytokine production (10, 24), and other
data have shown that the expression of some chemokines var-
ies with hormonal fluctuations during normal menstruation.
For example, increased immunoreactivity to CCL11 (eotaxin)
has been observed in endometrial epithelium during the luteal
phase (high progesterone) of the mouse menstrual cycle com-
pared to the follicular phase (low progesterone) (13). In con-
trast, Saavedra and colleagues (28) reported that estrogen
treatment did not alter CCL2 (MCP-1), Mip-2, or CCL5
(RANTES) levels over a 21-day period. In our model, proges-
terone treatment did not appear to influence CXCL10 (IP-10)
and CCL11 (eotaxin) levels, verifying that the increases ob-
served were produced in response to infection.
CCL5 (RANTES) was the only chemokine to stay at signif-
icantly elevated levels in the upper GT throughout the course
of infection. However, mRNA expression was low on all days
that were evaluated. The presence of CCL5 (RANTES) pro-
tein in tissue has generally been shown to correlate with
mRNA expression, making our results somewhat surprising. A
possible explanation is that CCL5 (RANTES) protein is deliv-
ered to the tissue from other sites. CCL5 (RANTES) is found
at picogram levels in the blood of healthy humans (8) and is
known to be released from thrombin-stimulated platelets (16).
It is therefore possible that the CCL5 (RANTES) protein
measured in the GT following Chlamydia infection is blood
derived rather than locally produced. Upon secretion, CCL5
(RANTES) may then directly bind to the activated genital
SuperArray analysis allowed the evaluation of additional
chemokines in the local genital mucosa of infected mice. Other
1544 MAXION AND KELLYINFECT. IMMUN.
chemokines detected by this analysis include CCL21 (SLC),
Gro-1, TCA-3, XCL1 (lymphotactin), CXCL4 (platelet factor
4), CXCL12 (SDF-1?/?), and SDF-2. Most notably, there was
an increase in the level of CCL21 (SLC) mRNA in infected
upper GT tissues compared to uninfected tissues. CCL21
(SLC) is important for T-cell migration across high endothelial
venules within secondary lymphoid tissues, as demonstrated
for mice deficient in CCL21 (SLC) (11) or the chemokine
receptor CCR7 (38). However, CCL21 (SLC) has also been
shown to bind to CXCR3, the receptor for CXCL9 (MIG) and
CXCL10 (IP-10) in mice but not humans (31). Preliminary
data obtained with reverse transcription-PCR for whole GT
homogenates have shown that CXCR3 and CCR5 are ex-
pressed only in the GTs of infected mice and not until 14 days
after infection (unpublished observations). In addition, we
found that the levels of CXCL12 (SDF-1?/?) mRNA expres-
sion were similar between infected and uninfected tissues but
were approximately twofold higher in upper GT tissue than in
lower GT tissue (data not shown). CXCL12 (SDF-1?/?) in-
duces rapid adhesion of CD4 cells to CD54 (5). Thus, CXCL10
(IP-10), CXCL9 (MIG), CCL5 (RANTES), CCL21 (SLC),
and CXCL12 (SDF-1?/?) are most likely involved in the che-
motaxis of Th1 cells to the upper GT during Chlamydia infec-
To date, there have been very few reports of chemokine
induction in response to Chlamydia infection in the GT. Pre-
vious reports have examined chemokine induction in vitro and
have focused on chemokines of the CXC class, which are
important for neutrophil chemotaxis. Namely, IL-8 (27, 35),
CXCL1 (Gro-?), and CXCL5 (epithelial neutrophil activating
protein 78 [ENA-78]) (35) were produced by epithelial cells
infected with human serovars of C. trachomatis. Interestingly,
IL-8 was not found in vaginal secretions of women with C.
trachomatis infection (12). However, Mip-2 and CCL2
(MCP-1) were found at increased levels in the lungs of mice
during infection with Chlamydia psittaci (14). In this study, we
noted an increase in Gro-1 but not Mip-2, the functional ho-
molog of murine IL-8. We also found that CCL2 (MCP-1)
mRNA expression was consistently low in both the upper and
the lower GTs, supporting our protein data. CCL2 (MCP-1)
has been shown to be upregulated in vaginal tissues of mice
following infection with Candida albicans in vivo (28). These
differences in Mip-2 and CCL2 (MCP-1) expression may re-
flect differences between tissue sites or specific features of the
The factors that lead to ascending Chlamydia infection in a
subset of individuals are currently unknown. Our data showing
differential chemokine expression in the upper and lower GTs
support increasing evidence that the inflammatory response in
the lower GT may be prematurely terminated even in the
presence of an active C. trachomatis infection. Perhaps Chla-
mydia-infected cells secrete immunosuppressive factors which
hamper antichlamydial immunity in the lower GT. Alterna-
tively, early termination of inflammatory responses in the
lower GT may be an inherent response of a site that is com-
monly exposed to nonpathogenic organisms. For example, us-
ing another mucosal tissue that is exposed to commensal flora,
Yamamoto et al. showed that intestinal epithelial cells inhibit
T-cell responses through a novel, non-transforming growth
factor ?-dependent mechanism (36). Interestingly, the early
production of gamma interferon (6) and tumor necrosis factor
? (7) in vaginal secretions and the expression of adhesion
molecules in the lower GT early after infection (18) diminished
to nearly baseline levels by day 7 in the presence of viable
chlamydiae. Therefore, we hypothesize that delayed eradica-
tion of chlamydiae in the lower GT early after infection may
facilitate upper GT infection. Future studies will therefore be
aimed at selectively boosting the antichlamydial immune re-
sponse in the cervical-vaginal region.
We thank Robert Strieter for helpful discussions and advice and
Ann Chan for histochemical staining.
This work was supported by PHS grant AI26328 from NIH. H.K.M.
was supported by Microbial Pathogenesis training grant 5-T32-AI-
1. Akiyama, M., H. Okabe, K. Takakura, Y. Fujiyama, and Y. Noda. 1999.
Expression of macrophage inflammatory protein-1alpha (MIP-1alpha) in
human endometrium throughout the menstrual cycle. Br. J. Obstet. Gynae-
2. Altman, G. B., A. M. Gown, D. L. Luchtel, and C. Baker. 1999. RANTES
production by cultured primate endometrial epithelial cells. Am. J. Reprod.
3. Arenberg, D. A., S. L. Kunkel, P. J. Polverini, S. B. Morris, M. D. Burdick,
M. C. Glass, D. T. Taub, M. D. Iannettoni, R. I. Whyte, and R. M. Strieter.
1996. Interferon-gamma-inducible protein 10 (IP-10) is an angiostatic factor
that inhibits human non-small cell lung cancer (NSCLC) tumorigenesis and
spontaneous metastases. J. Exp. Med. 184:981–992.
4. Boucher, A., W. Mourad, J. Mailloux, A. Lemay, and A. Akoum. 2000.
Ovarian hormones modulate monocyte chemotactic protein-1 expression in
endometrial cells of women with endometriosis. Mol. Hum. Reprod. 6:618–
5. Campbell, J. J., J. Hedrick, A. Zlotnik, M. A. Siani, D. A. Thompson, and
E. C. Butcher. 1998. Chemokines and the arrest of lymphocytes rolling under
flow conditions. Science 279:381–384.
6. Darville, T., C. W. Andrews, Jr., and R. G. Rank. 2000. Does inhibition of
tumor necrosis factor alpha affect chlamydial genital tract infection in mice
and guinea pigs? Infect. Immun. 68:5299–5305.
7. Darville, T., J. Sikes, C. W. Andrews, Jr., and R. Rank. 2000. Cytokine
response profile that determines early eradication of chlamydial genital tract
infection in a murine model, p. 171. In P. Saikku (ed.), Immunology and
pathogenesis. Proceedings for the Meeting of the European Society for
Chlamydia Research. Universitas Helsingiensis, Helsinke, Finland.
8. Fiore, J., M. Di Stefano, L. Lepera, A. Saracino, L. Monno, G. Angarano,
and G. Pastore. 1999. Evidence for a local synthesis of B-chemokines within
the genital tract of both HIV-1-infected and uninfected women. J. Acquir.
Immune Defic. Syndr. 21:255–257.
9. Gerard, C., and B. J. Rollins. 2001. Chemokines and disease. Nat. Immunol.
10. Gilmore, W., L. P. Weiner, and J. Correale. 1997. Effect of estradiol on
cytokine secretion by proteolipid protein-specific T cell clones isolated from
multiple sclerosis patients and normal control subjects. J. Immunol. 158:446–
11. Gunn, M. D., S. Kyuwa, C. Tam, T. Kakiuchi, A. Matsuzawa, L. T. Williams,
and H. Nakano. 1999. Mice lacking expression of secondary lymphoid organ
chemokine have defects in lymphocyte homing and dendritic cell localiza-
tion. J. Exp. Med. 189:451–460.
12. Hedges, S. R., D. A. Sibley, M. S. Mayo, E. W. Hook, and M. W. Russell.
1998. Cytokine and antibody responses in women infected with Neisseria
gonorrhoeae: effects of concomitant infections. J. Infect. Dis. 178:742–751.
13. Hornung, D., K. Dohrn, K. Sotlar, R. R. Greb, D. Wallwiener, L. Kiesel, and
R. N. Taylor. 2000. Localization in tissues and secretion of eotaxin by cells
from normal endometrium and endometriosis. J. Clin. Endocrinol. Metab.
14. Huang, J., M. D. Wang, S. Lenz, D. Gao, and B. Kaltenboeck. 1999. IL-12
administered during Chlamydia psittaci lung infection in mice confers imme-
diate and long-term protection and reduces macrophage inflammatory pro-
tein-2 level and neutrophil infiltration in lung tissue. J. Immunol. 162:2217–
15. Igietseme, J. U., K. H. Ramsey, D. M. Magee, D. M. Williams, T. J. Kincy,
and R. G. Rank. 1993. Resolution of murine chlamydial genital infection by
the adoptive transfer of a biovar-specific TH1 lymphocyte clone. Reg. Im-
16. Kameyoshi, Y., A. Dorschner, A. L. Mallet, E. Chrisophers, and J. M.
VOL. 70, 2002 CHEMOKINE EXPRESSION AND CHLAMYDIA GENITAL INFECTION 1545
Schroder. 1992. Cytokine RANTES released by thrombin-stimulated plate-
lets is a potent attractant for human eosinophils. J. Exp. Med. 176:587–592.
17. Kelly, K. A., and R. G. Rank. 1997. Identification of homing receptors that
mediate the recruitment of CD4 T cells to the genital tract following intra-
vaginal infection with Chlamydia trachomatis. Infect. Immun. 65:5198–5208.
18. Kelly, K. A., J. C. Walker, S. H. Jameel, H. L. Gray, and R. G. Rank. 2000.
Differential regulation of CD4 lymphocyte recruitment between the upper
and lower regions of the genital tract during Chlamyida infection. Infect.
19. Kunkel, E. J., J. J. Campbell, G. Haraldsen, J. Pan, J. Boisvert, A. I. Roberts,
E. C. Ebert, M. A. Vierra, S. B. Goodman, M. C. Genovese, A. J. Wardlaw,
H. B. Greenberg, C. M. Parker, E. C. Butcher, D. P. Andrew, and W. W.
Agace. 2000. Lymphocyte CC chemokine receptor 9 and epithelial thymus-
expressed chemokine (TECK) expression distinguish the small intestinal
immune compartment: epithelial expression of tissue-specific chemokines as
an organizing principle in regional immunity. J. Exp. Med. 192:761–768.
20. Kunkel, S. L., T. Standiford, K. Kasahara, and R. M. Strieter. 1991. Inter-
leukin-8 (IL-8): the major neutrophil chemotactic factor in the lung. Exp.
Lung Res. 17:17–23.
21. O’Garra, A., McEvoy, L. M., and A. Zlotnik. 1998. T cell subsets: chemokine
receptors guide the way. Curr. Biol. 8:R646-R649.
22. Paavonen, J., and M. Lehtinen. 1996. Chlamydial pelvic inflammatory dis-
ease. Hum. Reprod. Update 2:519–529.
23. Perry, L. L., K. Feilzer, and H. D. Caldwell. 1997. Immunity to Chlamydia
trachomatis is mediated by T helper 1 cells through IFN-gamma-dependent
and -independent pathways. J. Immunol. 158:3344–3352.
24. Piccinni, M. P., M. G. Giudizi, R. Biagiotti, L. Beloni, L. Giannarini, S.
Sampognaro, P. Parronchi, R. Manetti, F. Annunziato, C. Livi, et al. 1995.
Progesterone favors the development of human T helper cells producing
Th2-type cytokines and promotes both IL-4 production and membrane
CD30 expression in established Th1 cell clones. J. Immunol. 155:128–133.
25. Rank, R. G., A. K. Bowlin, and K. A. Kelly. 2000. Characterization of
lymphocyte response in the female genital tract during ascending chlamydial
genital infection in the guinea pig model. Infect. Immun. 68:5293–5298.
26. Rank, R. G., M. M. Sanders, and A. T. Kidd. 1993. Influence of the estrous
cycle on the development of upper genital tract pathology as a result of
chlamydial infection in the guinea pig model of pelvic inflammatory disease.
Am. J. Pathol. 142:1291–1296.
27. Rasmussen, S., L. Eckmann, A. J. Quayle, L. Shen, Y. Zhang, D. J. Ander-
son, J. Fierer, R. S. Stephens, and M. F. Kagnoff. 1997. Secretion of proin-
flammatory cytokines by epithelial cells in response to Chlamydia infection
suggests a central role for epithelial cells in chlamydial pathogenesis. J. Clin.
28. Saavedra, M., B. Taylor, N. Lukacs, and P. L. Fidel, Jr. 1999. Local pro-
duction of chemokines during experimental vaginal candidiasis. Infect. Im-
29. Sebastiani, S., P. Allavena, C. Albanesi, F. Nasorri, G. Bianchi, C. Traidl, S.
Sozzani, G. Girolomoni, and A. Cavani. 2001. Chemokine receptor expres-
sion and function in CD4? T lymphocytes with regulatory acitivity. J. Im-
30. Sorensen, T. L., M. Tani, J. Jensen, V. Pierce, C. Lucchinetti, V. A. Folcik,
S. Qin, J. Rottman, F. Sellebjerg, R. M. Strieter, J. L. Frederiksen, and R. M.
Ransohoff. 1999. Expresion of specific chemokines and chemokine receptors
in the central nervous system of multiple sclerosis patients. J. Clin. Investig.
31. Soto, H., W. Wang, R. M. Strieter, N. G. Copeland, D. J. Gilbert, N. A.
Jenkins, J. Hedrick, and A. Zlotnik. 1998. The CC chemokine 6Ckine binds
the CXC chemokine receptor CXCR3. Proc. Natl. Acad. Sci. USA 95:8205–
32. Taub, D. D., K. Conlon, A. R. Lloyd, J. J. Oppenheim, and D. J. Kelvin. 1993.
Preferential migration of activated CD4? and CD8? T cells in response to
MIP-1alpha and MIP-1beta. Science 260:355–358.
33. Tuffrey, M., F. Alexander, C. Inman, and M. E. Ward. 1990. Correlation of
infertility with altered tubal morphology and function in mice with salpingitis
induced by a human genital-tract isolate of Chlamydia trachomatis. J. Re-
prod. Fertil. 88:295–305.
34. Wiedermann, C. J., E. Kowald, N. Reinish, C. M. Kaehler, I. von Luettichau,
J. M. Pattison, P. Huie, R. K. Sibley, P. J. Nelson, and A. M. Krensky. 1993.
Monocyte haptotaxis induced by the RANTES chemokine. Curr. Biol.
35. Wyrick, P. B., S. T. Knight, T. R. Paul, R. G. Rank, and C. S. Barbier. 1999.
Persistent chlamydial envelope antigens in antibiotic-exposed infected cells
trigger neutrophil chemotaxis. J. Infect. Dis. 179:954–966.
36. Yamamoto, M., K. Fujihashi, K. Kawabata, J. R. McGhee, and H. Kiyono.
1998. A mucosal intranet: intestinal epithelial cells down-regulate intraepi-
thelial, but not peripheral, T lymphocytes. J. Immunol. 160:2188–2196.
37. Yang, X., J. Gartner, L. Zhu, S. Wang, and R. C. Brunham. 1999. IL-10 gene
knockout mice show enhanced Th1-like protective immunity and absent
granuloma formation following Chlamydia trachomatis lung infection. J. Im-
38. Yoshida, R., M. Nagira, M. Kitaura, N. Imagawa, T. Imai, and O. Yoshie.
1998. Secondary lymphoid-tissue chemokine is a functional ligand for the CC
chemokine receptor CCR7. J. Biol. Chem. 273:7118–7122.
Editor: J. D. Clements
1546MAXION AND KELLYINFECT. IMMUN.