Human immunodeficiency virus receptor and coreceptor expression on human uterine epithelial cells: regulation of expression during the menstrual cycle and implications for human immunodeficiency virus infection

Abstract

Human immunodeficiency virus-1 (HIV-1) is primarily a sexually transmitted disease. Identification of cell populations within the female reproductive tract that are initially infected, and the events involved in transmission of infection to other cells, remain to be established. In this report, we evaluated expression of HIV receptors and coreceptors on epithelial cells in the uterus and found they express several receptors critical for HIV infection including CD4, CXCR4, CCR5 and galactosylceramide (GalC). Moreover, expression of these receptors varied during the menstrual cycle. Expression of CD4 and CCR5 on uterine epithelial cells is high throughout the proliferative phase of the menstrual cycle when blood levels of oestradiol are high. In contrast, CXCR4 expression increased gradually throughout the proliferative phase. During the secretory phase of the cycle when both oestradiol and progesterone are elevated, CD4 and CCR5 expression decreased whereas CXCR4 expression remained elevated. Expression of GalC on endometrial glands is higher during the secretory phase than during the proliferative phase of the menstrual cycle. Because epithelial cells line the female reproductive tract and express HIV receptors and coreceptors, it is likely that they are one of the first cell types to become infected. The hormonal regulation of HIV receptor expression may affect a woman's susceptibility to HIV infection during her menstrual cycle. Moreover, selective coreceptor expression could account for the preferential transmission of R5-HIV-1 strains to women. In addition, these studies provide evidence that the uterus, and potentially the entire upper reproductive tract, are important sites for the initial events involved in HIV infection.

Introduction

Heterosexual transmission of human immunodeficiency virus-1 (HIV-1) accounts for 70–80% of new infections worldwide, with 80% of transmissions occurring from male to female.¹ It is not well understood what promotes or contributes to the establishment of HIV infection within the female reproductive tract, or whether infection is transmitted by free virus or by virus-infected cells, both of which are present in semen.² It is thought, however, that the amount of virus, the presence of other genital tract infections, and the effectiveness of the innate and adaptive mucosal immune systems within the female reproductive tract all contribute to the establishment of HIV-1 infection.^3–6

Chemokine receptors are recognized as essential coreceptors for HIV infection and are expressed on several types of haematopoietic and non-haematopoietic cells, including epithelial cells.^7–11 CCR5 was identified as a major coreceptor for macrophage-tropic (M-tropic) non-syncytium inducing strains of HIV-1 because infection is blocked by the presence of the CCR5 ligands macrophage inflammatory protein-1α (MIP-1α), MIP-1β and regulated on activation, normal T-cell expressed, and secreted (RANTES).^7,12 Similarly, CXCR4 was identified as the major coreceptor for infection by T-tropic, syncytium-inducing strains of HIV-1, as infection by these strains of HIV-1 is blocked by the CXCR4 ligand stromal cell-derived factor-1 (SDF-1).^7,12 These findings have led to the renaming of M-tropic strains of HIV-1 as R5 strains, T-tropic strains as X4 strains, and dual tropic strains as R5X4.^7,12 Of particular interest is the finding that despite the presence of both R5- and X4-utilizing strains of HIV in infected individuals, it appears that R5 strains preferentially establish infection following transmission across mucosal surfaces.^13,14 In addition to selectivity in transmission, R5 and X4 viruses elicit different pathogenic sequela,^15,16 suggesting that both the tropism of the virus and the cell population that is initially infected may determine important clinical features.

Menstruation, pregnancy and immunity in the female reproductive tract are regulated by the female sex hormones oestradiol and progesterone. The onset of menstruation, typically denoted as day 1 of the menstrual cycle, is characterized by low levels of both progesterone and oestradiol. During the first 14 days of the menstrual cycle, referred to as the proliferative phase, oestradiol levels slowly rise whereas progesterone levels remain low. Immediately prior to ovulation (days 13–14), oestradiol levels rapidly peak. This results in a rise in the levels of luteinizing hormone that initiates ovulation. The second half of the menstrual cycle (days 14–28) referred to as the secretory phase, is characterized by a gradual rise in both oestradiol and progesterone. If the ovum is fertilized, progesterone levels continue to rise to promote and maintain implantation of the fetus, whereas oestradiol levels decline. If fertilization does not occur, then progesterone and oestradiol levels both decline prior to the onset of menstruation on day 28 of the cycle.

In addition to the regulation of menses, oestradiol and progesterone have been shown to influence immune cell trafficking and functional activity within the female reproductive tract. Previous findings by our group have demonstrated the existence of lymphocyte aggregates (LA) within the submucosa of the endometrium that are comprised of a core of B cells surrounded by CD8⁺ T cells, and encircled by macrophages.^17,18 As LA are significantly smaller during the proliferative phase of the menstrual cycle than during the secretory phase, it is likely that they arise as a result of cell trafficking rather than from clonal expansion of pre-existing cells within the endometrium.¹⁸ In addition to regulating the trafficking of leucocytes into the female reproductive tract, oestradiol and progesterone also regulate immune cell function. White et al. showed that although CD8⁺ CD3⁺ cytotoxic T-cell numbers were maintained within the upper female reproductive tract throughout the menstrual cycle of premenopausal women, they were functionally active only during the proliferative phase and not during the secretory phase.¹⁹ It is hypothesized that the high levels of progesterone present during both the secretory phase and throughout pregnancy serve to inhibit T-cell cytolytic activity in order to prevent destruction of the allogeneic fetus. Thus, both progesterone and oestrogen serve important roles to regulate mucosal immunity within the female reproductive tract.

As many of the same receptors important in immunity are also involved in HIV infection, we examined the expression of HIV-1 receptors and coreceptors on uterine epithelial cells during various stages of the menstrual cycle. Although there are a few reports of CD4 expression on human epithelia, including human uterine endometrium²⁰ there is little known about regulation of CD4 expression as a function of menstrual cycle status. In addition, CXCR4 and CCR5 expression on epithelia in the gastrointestinal tract and lung has also been described.^9–11,21 Moreover, expression of chemokines including interleukin (IL)-8 and RANTES has been reported in human endometrium.^22–24 It is known that HIV infection of epithelial cell lines derived from the colon is dependent on chemokine receptor expression, as infection with an X4 and galactosylceramide (GalC) binding strain of HIV-1 required interaction with CXCR4 in addition to CD4 or GalC binding.⁸ In these studies, virus attached to GalC-positive cells, but was not internalized if the cell did not also express CXCR4.⁸ We recently demonstrated infection of CXCR4⁺ uterine epithelial cell lines with X4, but not R5, strains of HIV²⁵. Thus, the correlation between chemokine and chemokine receptor expression, and susceptibility to infection by HIV-1 strains of different tropism, would provide valuable information about the mechanism by which HIV-1 infects cells within the female reproductive tract and the regulation of such infection by oestradiol and progesterone.

Although the lower reproductive tract tissues (vagina and cervix) are generally thought of as primary sites of infection in the human female reproductive tract, a growing body of evidence suggests that the upper reproductive tract may also serve as a site for the initial infection by HIV-1. In fact, the cervix and uterus appear to have more cells expressing HIV receptors and coreceptors than the vagina (D. Anderson, personal communication). Within minutes following in vitro fertilization, spermatozoa were found in the uterus and Fallopian tubes.²⁶ More recently, radio-opaque dye deposited into the vagina was found in the uterus and Fallopian tubes within two hours (A. Parsons, personal communication). Dye ascended into the upper reproductive tract irrespective of whether women were premenopausal, on oral contraceptives, or postmenopausal at the time of analysis. In other studies, both bacterial and viral sexually transmitted disease (STD)-causing organisms, including Clamydia trachomatis,^27,28 Neisseria gonorrhoeae^29,30 and HIV-1³¹ were found to travel freely in reproductive tract secretions. By binding to sperm, these micro-organisms exploited the rapid transit of sperm from the vagina to the uterus and Fallopian tubes.³² In the case of HIV-1, virions bind to the sperm through interaction of the envelope glycoprotein gp120 with galactosyl-alkylacyl-glycerolipid (GalAAG).^31,33,34 GalAAG is structurally and antigenically related to GalC which, when expressed on epithelial and neuronal cells, is a CD4 independent infectivity receptor for HIV-1, HIV-2 and simian immunodeficiency virus (SIV).^35–37 Thus, both bacterial and viral STD organisms in semen are likely to reach the upper female reproductive tract tissues within minutes of deposition in the vagina. Our previous studies demonstrating that epithelial cells from the uterus and the Fallopian tubes can be infected by HIV-1³⁸ further support our hypothesis that epithelial cells in the uterus are likely to be a primary site for HIV infection.

In this study, we evaluated uterine epithelial cells for expression of CCR5, CXCR4, CD4 and GalC using immunofluorescent staining of frozen endometrial sections obtained from patients undergoing hysterectomy. These tissues were obtained from women at different stages of the menstrual cycle to correlate the level of receptor and coreceptor expression with menstrual cycle stage. The results presented here demonstrate that all four receptors are expressed on uterine epithelium and that their levels of expression vary with the stage of the menstrual cycle. These findings support the hypothesis that HIV-1 can likely infect cells in the upper reproductive tract through these receptors, and suggest that hormonal variation in HIV receptor and coreceptor expression may regulate a woman's susceptibility to infection. Moreover, the selective expression of HIV receptors and chemokine coreceptors in the uterus, and their regulation during different stages of the menstrual cycle, may in part explain the mechanism responsible for the selective transmission of CCR5-utilizing strains of HIV-1 to women.

Materials and methods

Tissues

Uterine tissue was obtained from patients undergoing hysterectomy following informed consent. Most patients included in this study were diagnosed as having leiomyomata, prolapsed uterus, or benign ovarian disease. None had a postoperative diagnosis of malignant uterine disease. Tissues to be used for immunophenotyping were dissected out from sites distal to any gross pathology and processed for preparation of frozen sections. The menstrual stage of the endometrium was determined by examining haematoxylin/eosin stained paraffin sections³⁹ and were classified as early (four tissues), mid (three tissues), or late proliferative phase (two tissues), or early (three tissues), mid (three tissues), or late secretory phase (two tissues).

Antibodies

Mouse anti-human CXCR4 (immunoglobulin G2b (IgG2b), clone 44716.111), and IgG2b isotype control (clone 20116.11) were purchased from R & D Systems (Minneapolis, MI) and labelled with cyanine 5 (Cy5).¹⁷ Mouse anti-human CD4 (IgG2b, clone OKT4, American Type Culture Collection, Rockville, MD), mouse anti-epithelial cell antibody (IgG1, clone BerEP4, Dako, Carpinteria, CA) and a IgG1 isotype control (1711.11, R & D Systems) were labelled with cyanine 3 (Cy3).¹⁷ Mouse anti-human CCR5 (IgG2b, clone 45549.111, R & D Systems), mouse anti-human CCR5 (IgG2a, clone 2D7, Pharmingen, San Diego, CA), IgG2b isotype control (clone 20116.11, R & D Systems) and IgG2a isotype control (catalogue IC003F, R & D Systems) were purchased labelled with fluoroscein isothiocyanate (FITC). Anti-GalC (IgG3, clone mGalC, Boehringer-Mannheim, Indianapolis, IN) and IgG3 isotype control (clone FLOPC-21, Caltag, San Francisco, CA) were purchased unlabelled.

Immunofluorescent phenotyping of frozen sections

Frozen tissue sections were fixed in acetone for 15 min at room temperature and then rehydrated in a humidity chamber at 4° for 48 hr prior to staining. Combinations of three monoclonal antibodies (mAb) were applied to the sections at a concentration of 1 µg/100 µl in blocking buffer (phosphate-buffered saline (PBS)/1% bovine serum albumin (BSA)/0·1% azide containing 4 mg/ml human immunoglobulin) and incubated in humidity chambers for 2 hr at 4°. The addition of unlabelled antibodies and the fluorochrome-labelled anti-mouse IgG antibodies were performed prior to incubation of tissue sections with directly labelled antibodies. Following indirect staining, directly labelled antibodies were added in the presence of 5% normal mouse serum to prevent binding of anti-mouse antibodies to the directly conjugated mouse mAb. Completeness of blocking was controlled for as described previously.¹⁷ Unbound antibody was removed from the sections by aspiration followed by three 15-min washes in PBS/1% BSA/0·1% azide. Washed sections were fixed overnight at 4° in PBS containing 1% paraformaldehyde. Stained sections were wet-mounted in Prolong™ antifade (Molecular Probes Inc., Eugene, OR), sealed with nail varnish, and stored at 4° in the dark for up to 10 days prior to confocal imaging.

Confocal scanning laser microscopy

Antibody-stained sections were imaged using a Bio-Rad MR1000 Confocal Scanning Laser Microscope system equipped with a krypton/argon laser and a 40× oil immersion objective (Hercules, CA). Laser power, photomultiplier tube gain and enhancement factors were determined for the FITC, Cy3 and Cy5 channels using the single fluorochrome-stained sections to ensure effective cross-channel compensation.

Image analysis and fluorescence quantitation

Images were analysed on an IBM PC compatible computer using Adobe^® Photoshop^® version 5.5 (Adobe Systems Inc., San Jose, CA) equipped with plugins from the Image Analysis Tool Kit version 3.0 (Reindeer Games Inc., Raleigh, NC). To estimate fluorescence intensity for individual glands, the path tools of Photoshop were used to segment out each gland using morphology to define glandular margins. The average fluorescence/µm² of epithelial cells in each gland was then determined for each of the three fluorescent channels.

Statistical analysis

Differences in the average fluorescence intensity on glands were evaluated by Mann–Whitney tests using GraphPad Prism^® software.⁴⁰ Co-expression of receptors on glands was evaluated by Spearman rank correlation using the same software.

Results

Identification of epithelial cells within endometrial sections

The mAb BerEP4 was used as a marker to identify epithelial cells within uterine sections. The intensity of staining of the mAb BerEP4 varied between individual uterine glands within the same section (Fig. 1a and Fig. 2), and between samples from patients who were at different menstrual cycle stages (Fig. 3). BerEP4 reactivity was generally higher in the stratum basalis (lower one-third of the uterus, closest to the myometrium) than in the stratum functionalis (upper two-thirds of the uterus, closest to the lumen; Fig. 2). The median fluorescence intensity of BerEP4 staining on glandular tissue increased gradually during the proliferative phase of the menstrual cycle, and dropped during the secretory phase to levels equal to or below that present in the early proliferative phase (data not shown). Because the level of expression of BerEP4 varied considerably in different regions of the same gland and between glands from patients in different menstrual cycle stages, the fluorescence intensity of BerEP4 staining could not be used to normalize the fluorescence intensity of other cell surface markers. Rather, BerEP4 staining together with morphology were used to identify individual glands within the uterine tissue.

Figure 1.

Expression of BerEP4, CD4, GalC, CCR5 and CXCR4 on uterine epithelial cells. Scale is indicated on each panel by a bar representing 50 µm. (a–d) Distribution of BerEP4 (epithelial cell), CCR5 and CXCR4 expression in a late proliferative phase endometrium. Frozen sections were directly stained with three fluorescently tagged monoclonal antibodies: an epithelial cell specific antibody (Cy3-labelled clone BerEP4, red colour in (a–d)), anti-CCR5 (FITC-labelled clone 2D7, Green colour in (a–d)) and anti-CXCR4 (Cy5-labelled 12G5, blue colour in (a–d). (a) A montage of overlapping images extending from the myometrial interface (far left) to the luminal epithelium (right-hand side). (b–d) Higher magnification fields from the same region (as indicated by the boxes in (a)). BerEP4 reactivity decreases progressively from the myometrial interface towards the luminal epithelium (red colour in (a)). The luminal epithelium, in contrast to the adjacent glands, retains a relatively intense BerEP4 expression (red colour in (a) and (d)). CCR5 expression is prominent on the lymphoid aggregates located in the stratum basalis immediately adjacent to the myometrium (‘LA’ in (b)). Epithelial expression of CCR5 is low in the stratum basalis and increases in the proximal third of the stratum functionalis, thereafter decreasing towards the lumen (a) and Fig. 2). Again, the luminal epithelium is the exception to this pattern where CCR5 shows intense reactivity (green colour in (d). With the exception of the occasional gland, CXCR4 expression is low in the stratum basalis and proximal third of the stratum functionalis but increases in the outer two thirds of the stratum functionalis (a) and Fig. 2) and is maintained on the luminal epithelium (d), L indicates the position of the uterine lumen). (e) CCR5 is localized on the lateral and apical aspects of the epithelial cells and expression is low or absent on the basolateral aspects (green) whereas CXCR4 expression is more uniformly distributed on all aspects of the epithelial cells (blue). (f) CD4 expression on glandular epithelial cells. The CD4 expression is primarily localized to the apical surface of the epithelial cells (red), CXCR4 staining is evident on all aspects of the epithelial cells in this late proliferative tissue (blue). Apical staining appears purple indicating colocalization of CD4 and CXCR4. (g) GalC expression on glandular epithelial cells. GalC expression is present on patches of epithelial cells on some individual endometrial glands, in secretory phase endometria (green, GalC, red, BerEP4).

Figure 2.

Epithelial cell BerEP4 and chemokine receptor expression varies between glands and by location. Image analysis of the composite figure shown in Fig. 1a shows the differential expression of BerEP4 (a), CCR5 (b) and CXCR4 (c) across the endometrium. Total pixel intensity for each fluorochrome for individual glands is plotted against the distance from the myometrial–endometrial interface in the direction of the lumen. These results confirm the subjective impression obtained from visual examination of Fig. 1: BerEP4 is high in the stratum basalis and decreases in the functionalis, whereas CCR5 and CXCR4 are low in the stratum basalis and increase in the stratum functionalis.

Figure 3.

CCR5 and CCR4 expression on epithelial cells varies with the stage of the menstrual cycle. Each dot represents the average fluorescence intensity per µm² epithelium on an individual gland. (a) CCR5 expression during the early, middle and late stages of the proliferative phase and the early middle and late stages of the secretory phase. The median bars indicate that the level of expression of CCR5 increases with progress through the proliferative phase but is lower during the early, mid and late secretory phase; (b) this is reflected in a higher overall CCR5 expression in the proliferative phase. (c) CXCR4 expression during the early, middle and late stages of the proliferative phase and the early, middle and late stages of the secretory phase. The median bars indicate that the CXCR4 expression also increases during the early, middle and late stages of the proliferative phase but does not decline significantly during the secretory phase; (d) this is reflected in an overall slightly higher expression during the secretory phase. EP, early proliferative; MP, mid proliferative; LP, late proliferative; ES, early secretory; MS, mid secretory; LS, late secretory. Bracketed pairs indicate that the differences in median intensities between those adjacent pairs were statistically significant (Mann–Whitney test, * = P < 0·05, ^**P < 0·005, ^***P < 0·0005).

Epithelial cell expression of chemokine receptors

Using confocal microscopy, we determined that glandular and luminal epithelium from the uterus express both CCR5 and CXCR4 (Fig. 1). Expression was found to be independent of the disease diagnosis (data not shown).

The intensity of epithelial staining of CCR5 and CXCR4 varied among individual glands within each uterus (Fig. 1a–d). Generally, expression of these chemokine receptors was lower in the glandular epithelium of the stratum basalis and increased progressively through the stratum functionalis towards the lumen (Fig. 2bc). The luminal epithelium generally expressed high amounts of both CCR5 and CXCR4 (Fig. 1a and d). As an internal control for mAb binding, we found that CCR5 was also expressed on T cell aggregates in the stratum basalis (Fig. 1a and b), but was largely absent on macrophages, stromal T cells and intraepithelial lymphocytes (IEL). In contrast, CXCR4 expression was absent from T-cell aggregates as well as stromal macrophages and T cells, but was present on some IEL (data not shown).

To determine whether chemokine receptor expression varied during the menstrual cycle, multiple individual glands from patients at different menstrual stages were compared for CCR5 and CXCR4 expression. CCR5 expression on epithelial cells increased during the proliferative phase and then decreased during the secretory phase (Fig. 3). In contrast, CXCR4 staining rose during the proliferative phase and remained elevated throughout the secretory phase of the menstrual cycle (Fig. 3). CCR5 reactivity was most prominent on the apical and lateral aspects of the epithelial cells (Fig. 1e), whereas CXCR4 staining was more evenly distributed on the cell surfaces. However, CXCR4 expression was more apically distributed during the secretory phase of the menstrual cycle (Fig. 1f).

Epithelial CD4 expression

We examined uterine epithelial cells for CD4 expression, and compared expression among tissues from various menstrual cycle stages. CD4 was primarily localized on the apical side of epithelial cells and staining was punctate in nature (Fig. 1f). CD4 staining intensity was significantly higher in proliferative phase compared to the secretory phase of the menstrual cycle (Fig. 4a).

Figure 4.

CD4 and GalC expression on epithelial cells in proliferative and secretory endometria. Each dot represents the average fluorescence intensity per µm² epithelium on an individual gland. (a) CD4 expression during the proliferative phase and the secretory phase. The median bars indicate that the level of expression of CD4 was higher in the proliferative phase than in the secretory phase. No significant differences in expression were observed between the early, middle and late stages of the proliferative phase or between the early, middle and late stages of the secretory phase. (b) GalC expression during the proliferative and secretory phases. Although the median levels of expression of GalC were similar in both phases, several glands in secretory phase tissues expressed high amounts of GalC. Bracketed pairs indicate that the differences in median intensities between menstrual phases were statistically significant (Mann–Whitney test, * = P < 0·05, ^**P < 0·005, ^***P < 0·0005).

Epithelial GalC expression

GalC, an alternative infectivity receptor, was also expressed in uterine epithelium. Expression was found on relatively few endometrial glands, and was usually restricted to the luminal aspects of the cells (Fig. 1g). A small number of individual glands stained more intensely in some secretory endometria than that seen in tissues from the proliferative stage of the menstrual cycle (Fig. 4b).

Coexpression of chemokine receptors and CD4 on epithelial cells

The coexpression of the HIV receptor CD4 and coreceptors CXCR4 and CCR5 at different stages of the menstrual cycle was determined by Spearman rank correlation analysis. The relative fluorescence intensity of CXCR4 and CCR5 staining on individual glands was compared to that of CD4. When the coexpression of CXCR4 and CD4 was evaluated, there was a significant positive correlation between these markers during the late proliferative and late secretory phases (Table 1). In contrast, CCR5 and CD4 coexpression was only correlative in the early secretory phase. CXCR4 and CCR5 were coexpressed on individual glands only in the mid-secretory phase of the menstrual cycle, suggesting that the expression of these coreceptors on epithelial cells is relatively independent.

Table 1.

Co-expression of chemokine receptors and CD4 on uterine epithelial cells

	Proliferative			Secretory
	Early	Mid	Late	Early	Mid	Late
CXCR4 versus CD4	NS	NS	0·6154(P = 0·0252)	NS	NS	0·6103(P = 0·0093)
CCR5 versus CD4	NS	NS	NS	0·7275(P = 0·0032)	NS	NS
CCR5 versus CXCR4	NS	NS	NS	NS	0·676(P = 0·0021)	NS

Discussion

The vast majority of new HIV-1 infections in women are the result of heterosexual transmission.¹ Understanding the mechanism by which HIV-1 infects cells within the female reproductive tract and the conditions that regulate infection would be helpful in the design of intervention strategies to reduce the rate of new infections. It is possible that HIV-1 could infect cells within the female reproductive tract by a different mechanism than that used in the peripheral blood. This would include the possibility that cells other than T lymphocytes and macrophages are the initial targets of infection. Moreover, the expression of cellular receptors and coreceptors used by HIV-1 for infection may be differentially regulated on target cells in the mucosae than in the peripheral blood. Because R5 strains of HIV-1 are selectively transmitted across mucosal surfaces during the initial infection¹³ it is important to determine whether this selection takes place at the level of the target cell, the virus, or both.

Our studies focused on defining the expression of HIV-1 receptors and coreceptors on epithelial cells that line the lumen of the uterus, as a first step in defining potential cell targets within the female reproductive tract. We reasoned that epithelial cells would be one of the first cell types to come in contact with HIV-1, and might play a role in the replication of virus and in transmission of the infection to leucocytes present in the submucosa. Our findings demonstrate that key HIV receptors and coreceptors are present on these cells, and that their expression varies as a function of the menstrual cycle. The hormonal regulation of receptor and coreceptor expression may be an important mechanism that controls susceptibility to HIV-1 infection within the female reproductive tract.

The expression of CXCR4 and CCR5 on haematopoietic cells has been shown to be essential for HIV-1 infectivity.^7,41,42 Recognizing that 70–80% of all new cases of HIV infection occur as a result of sexual transmission prompted us to determine the extent to which CD4, CXCR4, CCR5 and GalC are expressed on epithelial cells within the female reproductive tract, and whether their expression is dependent on the stage of the menstrual cycle. The likelihood that HIV-1 can establish an initial primary infection in the uterus is high based on the potential for this virus to enter the uterine cavity within minutes after deposition in the vagina. We found that the glandular and luminal epithelium of the human uterus express both the major HIV-1 receptor CD4 and the epithelial cell associated receptor GalC, as well as the major HIV coreceptors CCR5 and CXCR4.

The role of the sex hormones progesterone and oestradiol in regulating chemokine receptor expression at mucosal sites is not well known. Our studies indicate that cell surface expression of these receptors is likely to be under hormonal control because expression varies with the stage of the menstrual cycle. In support of this hypothesis, are findings that progesterone inhibits IL-2-induced up-regulation of CCR5 and CXCR4 expression on activated T cells from the peripheral blood.⁴³ Moreover, this regulation of coreceptor expression correlated with a reduced susceptibility to infection by both R5 and X4 tropic strains of HIV-1. In addition to regulating expression of chemokine receptor expression, sex hormones can also affect HIV-1 transmission by their effects on the recruitment of lymphocytes and monocytes to mucosal tissues. Our previous studies^17,18 as well as studies by others^44–47 demonstrated that the distribution of leucocyte populations within reproductive tract tissues varied with the stage of the menstrual cycle, during pregnancy, and after menopause.

This report is the first to examine CD4 expression on human uterine epithelial cells during the menstrual cycle, and extends the previously reported expression of CD4 on human uterine epithelium and endometrial-derived cell lines.²⁰ In the present study, CD4 was most highly expressed on epithelial cells from uterus during the proliferative phase of the menstrual cycle, and expression was reduced on these cells during the secretory phase. CCR5 expression on glandular epithelium was also highest during the proliferative phase of the menstrual cycle, and expression was also found to be lower during the secretory phase. In contrast, CXCR4 expression increased during the proliferative phase and remained high throughout the secretory phase. Taken together, these results suggest that during the proliferative phase of the menstrual cycle, epithelial cells could bind and become infected preferentially with R5 and dual tropic HIV strains. In contrast, the high expression of CXCR4 during the secretory phase in the absence of CCR5 suggests that epithelial cells may be more susceptible to X4 or dual tropic HIV strains during the second half of the menstrual cycle. If, however, epithelial cells require the coexpression of CD4 and one or the other of the chemokine receptors for infection, we would speculate that infection by X4 strains of HIV would occur most likely during late proliferative phase, and by R5 strains during the early secretory phase of the menstrual cycle.

During the secretory phase of the menstrual cycle, however, CD4 levels are relatively low and may not be expressed at sufficient levels for HIV to infect epithelial cells through this receptor. That infection can occur during the secretory phase is suggested by our finding that the alternative HIV receptor, GalC is expressed on uterine epithelial cells. GalC expression was evident on the luminal aspects of some individual glands within secretory phase tissues and was low or absent in proliferative phase endometria. This difference in expression may be the result of changes in the expression of the enzymes involved in GalC synthesis, or in the degree to which GalC is capped with sialic acid. Expression levels of galactosyltransferase and sialyltransferase, enzymes that are involved in the synthesis of GalC are hormonally controlled and their expression in the uterus varies during the menstrual cycle.⁴⁸ Similar to CD4, GalC was largely restricted in its expression to the apical surface of the epithelial cells, suggesting that it would be accessible in vivo for HIV binding. HIV can bind both to GalC and to its sulphated derivative sulphatide.³⁴ Sulphatide expression is increased in the human endometrium during the secretory phase, although the cell types and subcellular localization of this molecule have not been determined.⁴⁹ Further studies would be required to determine whether HIV is able to infect uterine epithelial cells through GalC and its sulphated derivative.

Evidence that progesterone increases HIV infectivity in mucosal tissues comes from both epidemiological studies and from SIV infectivity studies in primates, where administration of depo-provera and progesterone containing oral contraceptives increased the risk of infection, whereas oestrogen had a protective effect.^5,50–53 Our findings that CD4, GalC, CXCR4 and CCR5 are expressed on uterine epithelial cells at different levels during the menstrual cycle, provides further evidence that susceptibility to HIV infection may be influenced by the female sex hormones through regulation of these critical receptors. Alternatively, oestradiol and/or progesterone may influence HIV susceptibility by increasing epithelial cell proliferation, thus providing additional target cells for infection. Studies are currently in progress to further elucidate the effects of sex hormones on HIV-1 infection of epithelial cells from the female reproductive tract.

Another mechanism that is likely to control HIV infection within the female reproductive tract, and that appears to be under hormonal control, is secretory leucocyte protease inhibitor (SLPI) production by epithelial cells and macrophages. SLPI has been shown to have potent anti-HIV-1 activity.⁵⁴ We previously reported SLPI production and secretion by primary cultures of uterine epithelial cells from premenopausal women but not from epithelial cells from post-menopausal women.⁵⁵ SLPI produced by epithelial cell cultures inhibited colony formation from both Gram-negative and Gram-positive bacteria, and this activity was completely neutralized with an anti-SLPI antisera. Thus, the hormonal regulation of SLPI production together with its anti-HIV activity are likely to be important factors in the susceptibility of women to HIV infection. For example, fluctuations in SLPI levels within the female reproductive tract may partly explain why the risk of infection (1 : 1000–1 : 10 000 per act of sexual intercourse) remains low under conditions of repeated viral exposure. Studies are needed to define the role of SLPI and other innate virucidal agents in inhibiting the spread of HIV-1 infection.

In summary, these findings suggest that the upper female reproductive tract is likely to be a site susceptible to HIV infection. Further, they indicate that differing hormonal conditions, as seen during the menstrual cycle, may alter susceptibility to infection. Studies are underway to evaluate the effect of altering hormone levels on HIV-1 infection, chemokine receptor expression, and proliferation of uterine epithelial cells.

Abbreviations

GalC

galactosylceramide

GalAAG

galactosyl-alkylacyl-glycerolipid

Cy3

cyanine 3

Cy5

cyanine 5

IEL

intraepithelial lymphocyte