| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
From the Fearing Research Laboratory, Department of Obstetrics, Gynecology and Reproductive Biology, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts 02115
ABSTRACT
Cell-mediated immunity (CMI) is key to defense against intracellular pathogens such as Chlamydia trachomatis and viruses that infect the lower female genital tract, but little is known about CMI at this site. Recent studies indicate that there are immunological microenvironments within the female genital tract, and that immune functions are affected by hormones as well as infections and inflammatory processes. To determine the distribution of mediators of CMI within the lower female genital tract, we have enumerated and characterized T-lymphocyte subsets and natural killer and antigen presenting cells (APCs; macrophages and dendritic cells) in the introitus, vagina, ectocervix, endocervix and cervical transformation zone (TZ) from healthy women, and have examined the effects of the menstrual cycle, menopause and inflammation on these parameters. In women without inflammation, T cells and APCs were most prevalent in the cervical TZ and surrounding tissue. Intraepithelial lymphocytes were predominantly CD8+ T cell+; most CD8+ cells in the TZ and endocervix, and a proportion of cells in the ectocervix, expressed T-cell internal antigen-1, a marker of cytotoxic potential. In contrast, the normal vaginal mucosa contained few T cells and APCs. Cervicitis and vaginitis cases had increased numbers of intraepithelial CD8+ and CD4+ lymphocytes and APCs. The menstrual cycle and menopause had no apparent effect on cellular localization or abundance in any of the lower genital tract tissues. These data indicate that the cervix, especially the TZ, is the major inductive and effector site for CMI in the lower female genital tract. Because CD4+ T cells and APCs are primary host cells for human immunodeficiency virus type 1 (HIV-1), these data also provide further evidence that the cervix is a primary infection site of HIV-1, and that inflammation increases the risk of HIV transmission.
cervix, female reproductive tract, menstrual cycle, vagina
INTRODUCTION
Humoral immunity in the human female genital tract has been well characterized (reviewed in Kutteh [1]). IgG and IgA-secreting plasma cells are abundant in the lamina propria of the endocervix and scarce in the vagina, providing evidence that immunological microenvironments exist in the lower female genital tract. Recent studies have also documented mechanisms of innate immunity, including the production of antimicrobial molecules such as lysozyme, lactoferrin, serine leukocyte protease inhibitor, and alpha- and beta-defensins, by epithelial cells at various sites within the female genital tract (reviewed in Cohen and Anderson [2]). In contrast, relatively little is known about the distribution or function of T lymphocytes, natural killer (NK) and antigen presenting cell (APC) populations in the human female reproductive tract, or the nature of local cell-mediated immune (CMI) responses to sexually transmitted pathogens. Mucosal tissues sequester the largest proportion of T lymphocytes in the body [3]. T cell populations in the gastrointestinal tract have been well characterized, and express unique phenotypic and functional characteristics (reviewed in Kohne et al. [4]). It has been known for some time that lymphocytes and antigen-presenting cells also reside in the female genital tract mucosa. Characterization of immune cell populations in the lower genital tract, including those responsible for the inductive and effector arms of cellular immunity, remains an essential step in designing appropriate vaccine strategies to induce local protective immunity.
The lower genital tract in women is comprised of four discrete anatomical regions: 1) the introitus, which is covered by a keratinized stratified squamous epithelium resembling skin, 2) the vaginal mucosa, which is covered by an aglandular, nonkeratinized stratified squamous epithelium, 3) the ectocervix, which is covered by a mucosal layer histologically similar to that of the vagina, and 4) the endocervix. which consists of a simple columnar epithelium with numerous glands (pseudoglands). The transformation zone (TZ) represents an abrupt transition between the ectocervix and the endocervix. The susceptibility of these regions to infectious organisms differs. Candida albicans and Trichomonas vaginalis colonize the vagina, whereas the cervix is susceptible to infection by Chlamydia trachomatis and Neiseria gonorrhoea. The transformation zone is the main target of human papilloma virus (HPV) infection. Although the primary initial infection site of human immunodeficiency virus type 1 (HIV-1) in the female genital tract is still unknown, HIV-1+ cells have been recovered from the endocervix of women with established infections [5] and memory CD4+ T cells in cervical explants are infectable with HIV-1 in vitro [6]. Vaginal inoculation experiments in macaques have implicated both endocervical T cells and vaginal dendritic cells as initial targets in simian immunodeficiency virus (SIV) transmission [7, 8].
Descriptive studies using immunohistochemistry or flow cytometry have shown that T cells and APCs are present throughout the human cervical and vaginal mucosae [9–15], but these studies have mostly included small sample numbers, and have failed to quantify and precisely locate the cells. Furthermore, few have attempted to characterize these cell populations in terms of functional or homing markers. Recent studies have shown that antigen-reactive CD8+ T cells can be harvested from surgical tissues [16, 17] and endocervical cytobrush specimens [18–24], but the optimal site(s) in the lower genital tract for sampling mucosal CD8+ and CD4+ lymphocytes with specific immune functions has yet to be determined.
The primary purpose of this study was to further define sites of CMI responses within the female lower genital tract. We used classical markers that identify lymphocyte subsets (CD8+ suppressor/cytotoxic cells, CD4+ helper/inducer cells, and CD56+ and CD57+ natural killer cells) and APC (CD68+ macrophages and CD1a+ dendritic cells), and further characterized the cells by expression of the following markers: CD45RO (memory T cells), CD45RA (naive T cells), and TIA1, an RNA binding protein associated with cytotoxic granules in cytotoxic T cells and natural killer cells. We also investigated expression of homing- and tissue-specific leukocyte adhesion molecules, specifically: CD62 (L-selectin), which mediates homing of naive T cells to peripheral lymph nodes and Peyer's patches and to areas of inflammation; cutaneous lymphocyte-associated antigen (CLA), an integrin expressed on skin-homing lymphocytes; and CD103 (alpha E integrin), which is expressed on mucosal-associated T cells. We therefore provide a comprehensive description of T lymphocyte subpopulations, NK cells, dendritic cells, and macrophages within the epithelium and lamina propria at various sites in the female lower genital tract. A large sample size also enabled the analysis of variables such as the menstrual cycle and inflammation, which could potentially influence leukocyte distribution and function. Because many of the cell types involved in CMI responses are host cells for HIV-1, this study also provides information on the location and abundance of HIV-1 target cells in the lower female genital tract.
MATERIALS AND METHODS
Protection of Human Subjects
This research was conducted under Institutional Review Board-approved protocols. Discarded surgical samples were coded, and after review of medical records, patient identifiers were eliminated from all laboratory records to ensure patient confidentiality.
Tissues
Vaginal tissue proximal to the cervix (n = 21) and cervical tissues (n = 38) were obtained from women undergoing vaginal repair (mean age 42.8 ± 8.8 yr, range 27–54 yr) and/or hysterectomy (mean age 43.3 ± 7.0 yr, range 27–63 yr) for nonmalignant indications (leiomyoma, prolapsed uterus, or pelvic pain). Both vaginal and cervical tissues were removed from the same patient in seven cases. Four patients were diagnosed with vaginal inflammation and three with cervical inflammation. The menstrual cycle stage was determined from medical records and endometrial dating. Vaginal tissues were obtained during the proliferative phase (n = 9) and the secretory phase (n = 4) of the menstrual cycle. Menstrual cycle data were unavailable for four patients. Cervical tissues were removed at the proliferative (n = 21) and secretory (n = 9) stages of the menstrual cycle, with one sample obtained at menses. Data for one sample were unavailable. Four vaginal and six cervical samples were obtained from postmenopausal women not receiving hormone replacement therapy. Two samples of wax-embedded vaginal tissues containing the introitus and urethral opening were also studied.
Tissue samples were obtained within 30 min of surgical removal and placed in ice-cold RPMI-1640 medium for transport. Large vaginal samples were sliced into smaller fragments, and cervical specimens were divided, when possible, into ectocervix, endocervix, and transformation zone. All tissues were embedded, unfixed, in ornithine-carbomoyl transferase compound (Miles Laboratory Inc.), rapidly frozen, and stored at –70°C. Surgical specimens of pediatric tonsil and adult large bowel were similarly collected and processed as tissue controls for immunocytochemistry.
Immunochemistry
Frozen sections (5 µm) were placed on Superfrost slides (Fisher Scientific), fixed in 100% acetone at –20°C for 10 min, air-dried, and stored at –70°C. Slides were rehydrated in Tris buffer and incubated with a blocking solution of normal goat serum (Biogenix) for 20 min. Tissues were then reacted for 30 min with a panel of primary antibodies (Table 1). Primary antibodies were diluted in 0.02M Tris buffer, pH 8.2, containing 1% crystalline BSA and 0.02 M sodium azide. Following a TRIS buffer rinse, the blocking step was repeated for 15 min, and sections were incubated with an alkaline phosphatase detection system (Biogenix) with Fast Red (DAKO) as the substrate that stains positive cells red. Sections were counterstained with haematoxylin and mounted under a water-soluble medium (Accergel; Accurate). The tissue distribution of macrophages and dendritic cells was investigated in further detail using a double labeling technique. In brief, the immunocytochemical procedure was carried out twice and dendritic cells visualized using a substrate that stained positive cells blue (Vector Blue, Vector Labs) and macrophages were detected using Fast Red (DAKO). Frozen sections of tonsil and intestine served as positive controls for the immunocytochemical technique. For negative controls, tissue sections were processed without primary antibody, which was replaced with either the antibody diluent, an isotype-matched irrelevant antibody (MOPC-21, Sigma) or normal mouse serum (Biogenex).
|
Antibodies recognizing epitopes following formalin fixation and wax embedding were used to identify immune cells in the mucosa of the distal vagina: CD8+ lymphocytes (1:30), CD45RO+ and CD45RA+ lymphocytes (1:100), CD68+ macrophages (1:100), CD15+ granulocytes (1:100), and HLA-DR+ cells (1:20), all obtained from DAKO. An antigen-retrieval step was needed to expose epitopes before staining with antibodies against CD8A and HLA-DR. This was carried out by heating sections in a proprietary antigen-retrieval solution (DAKO) at 95°C for 20 min. Sections in the antigen-retrieval solution were then removed from the water bath and allowed to stand at room temperature for 20 min, followed by washing in Tris buffer (5 min x 2).
Quantitative Assessment
Intraepithelial CD4+, CD8+, CD103+, CDla+ and TIA1+ cells were counted in sections of vagina and ectocervix using an eyepiece reticule (0.16 mm2) at an objective magnification of 20x. At least three separate regions of the epithelium were analyzed for each sample. The counts were averaged and presented as number of cells/mm2 epithelium. Because the endocervical epithelium is a single layer of cells, intraepithelial lymphocytes (IELs) in this tissue were expressed as number of positive cells per 100 epithelial cells observed at 20x magnification. As with other tissues, three endocervical regions were analyzed per sample, and epithelial regions (including glandular epithelium) were chosen where the morphology allowed identification of individual epithelial cells.
Statistical Analysis
Because data did not satisfy the conditions of normal distribution and homogeneity of variance, group differences were analyzed by nonparametric Mann-Whitney tests. Statistical tests were performed using StatView statistical software (version 5.0.1; SAS Institute Inc.).
RESULTS
There was a high level of interpatient and intrapatient variability in numbers of T cells and APCs at different sites within the lower genital tract. Despite this, significantly different localization patterns were observed for several immune cell populations.
Vagina
Introitus (vaginal opening) This region is covered by a reflexion of skin. The histological appearance was that of a keratinized stratified squamous epithelium with few lymphocytes, macrophages, or dendritic cells within the epithelial layer. In the lamina propria, small focal aggregations of lymphocytes were observed that stained positive for CD8A, CD4, CD45RO, CD45RA, CD103 and TIA1. CLA+ (cutaneous-associated) T lymphocytes were also detected in the lamina propria. In samples of the introitus we also observed tissue surrounding the opening to the urethra. This site was characterized by sparse populations of both CD4+ and CD8+ lymphocytes and CD68+ macrophages scattered throughout the stroma. A small number of CD8+ lymphocytes and macrophages were detected within the urethral epithelium. Moderate concentrations of granulocytes were also found in this region. Paraurethral glands were also present and contained CD8+ lymphocytes and CD68+ macrophages in the epithelium lining both glands and ducts.
Distal vagina Histologically, the distal vaginal mucosa is extremely irregular with many, often deep, infoldings and villous projections. Dermal papillae are exceptionally abundant and often are thrust so deep into the epithelium that they are covered by only 1–2 layers of epithelial cells. CD8+ lymphocytes were present in high concentrations in both cases studied. Most of the CD8+ population in one sample were IELs, and were predominantly CD45RO+. In contrast, the majority of CD8+ lymphocytes in the second sample were CD45RA+ and occurred in the lamina propria as extensive accumulations. Few CD4+ lymphocytes were detected in either of the cases, and these were restricted to the lamina propria. CD68+ macrophages were present in small numbers in the lamina propria and epithelium along with low concentrations of HLA-DR+ cells with a dendritic morphology.
Proximal vagina T lymphocytes were observed in all samples, but their concentrations varied widely (Fig. 1). For 50% of normal (no inflammation) samples, the majority of CD8+ cells were found in the lamina propria, either scattered throughout the stroma (Fig. 2A) or concentrated along the base of the epithelium (Fig. 2B). In the remaining 50% of normal samples, the majority of CD8+ cells were detected within the epithelial layer (Fig. 2C), often with many located in the basal layer of epithelial cells. In contrast, CD4+ lymphocytes were rarely observed in the epithelium, but were consistently present in the lamina propria (Fig. 2D). Some large, irregularly shaped CD4+ cells were noted in the vaginal epithelium and were presumed to be dendritic cells (Fig. 2E).
|
|
The majority of T lymphocytes in normal vaginal tissues were CD45RO+ (memory) cells, but most samples also contained a small proportion of CD45RA+ naive lymphocytes, restricted to the lamina propria. CD103+ (mucosal-associated) T lymphocytes were found throughout the vaginal mucosa, occurring in greater numbers within the epithelium than in the lamina propria (Fig. 2F). However, even within the IEL population, only a small proportion of T cells were CD103+ (Fig. 1). Most samples also contained a few CD62L-selectin+ lymphocytes that were restricted to the lamina propria. Few CD8+ T cells in the vagina expressed TIA1, an RNA binding protein associated with cytolytic granules present in cytotoxic T cells and NK cells (Figs. 1 and 2G). Few CD57+ and CD56+ cells and no gamma-delta T lymphocytes were detected in the vaginal mucosa.
Concentrations of macrophages in the vaginal mucosa varied from woman to woman, with most samples containing small numbers in the lamina propria and fewer in the epithelium (Fig. 2H). CD1a+ cells in the vagina exhibited a typical dendritic morphology and were mostly present in low concentrations restricted to the epithelium (Figs. 1 and 2I). Few granulocytes were detected in any vaginal tissues examined.
A notable feature of the human vaginal wall is its papillated lamina propria, which produces vascularized projections that extend deep within the epithelial layer. A topographical relationship existed between subpopulations of lymphocytes and the papillae. Lymphocytes expressing CD62 L-selectin, CD45RA, and TIA1 often appeared to almost exclusively reside within the stroma of these structures (Fig. 2J). Aggregates of CD8+ lymphocytes were often observed in the epithelium surrounding the papillae (Fig. 2K).
In summary, the predominant IEL immune cell profile in the normal vaginal epithelium consisted of CD8+ T lymphocytes, a proportion of which were TIA1+, CD103+, and CD45RO+. These cells were especially numerous around dermal papillae. Few CD4+ lymphocytes, CD68+ macrophages, or CD1a + dendritic cells were observed within the epithelial layer.
Four patients were diagnosed with chronic vaginal inflammation. In all of these cases, there was an increased concentration of lymphocytes in the vaginal epithelium compared to normal samples. The epithelial immune profile of two of these cases consisted of numerous CD8+ but few CD4+ lymphocytes, with some TIA1+ and a larger number of CD103+, plus many CD1a+ dendritic cells (Fig. 3C). The epithelial immune profiles of the other two cases were different, and also varied slightly from each other. Both samples contained numerous CD8+ as well as CD4+ IELs (Fig. 3A). Many of these IELs were CD103+. For one sample, only a few were TIA1+, whereas for the other, abundant TIA1+ cells were present (Fig. 3B). The sample with few TIA1+ IELs contained many CD1a+ dendritic cells, whereas the sample with numerous TIA1+ IELs contained few CDla+ dendritic cells that were mostly rounded in shape.
|
Cervix
Ectocervix Significantly more IELs staining positive for CD4, CD103, and TIA1 were detected in the normal ectocervix as compared to the vagina (Fig. 1). CD8+ lymphocytes were still the most abundant T cell population, and their distribution patterns could be divided into three groups. Group 1 (n = 9) had a fairly equal distribution of CD8+ lymphocytes between the lamina propria and epithelium (Fig. 4A). Group 2 (n = 14) had a greater proportion of CD8+ lymphocytes in the lamina propria than in the epithelium (Fig. 4B). Group 3 (n = 7) had more CD8+ lymphocytes in the epithelium compared to the lamina propria (Fig. 4C). Differences in spatial distribution of CD8+ lymphocytes were not related to stage of the menstrual cycle or menopausal status. As in the vagina, CD8+ IELs often formed a distinct layer at the base of the epithelium (Fig. 4D). The ectocervical mucosa, like the vagina, is papillated, and in some samples numerous CD8+ lymphocytes were present in the stroma of these papillae and in the epithelium surrounding them (Fig. 4E). The abundance of CD4+ lymphocytes varied greatly in the ectocervical mucosa, with the majority occurring in the lamina propria (Fig. 4F).
|
Most of the T lymphocytes in the ectocervix had a CD45RO (memory) phenotype. CD45RA+ lymphocytes were sparsely represented and were mainly restricted to the lamina propria. The majority of CD103+ lymphocytes were preferentially located in the epithelium, either at the base (Fig. 4G) or distributed throughout the width of the epithelium. As in the vagina, only a proportion of IELs in the ectocervix were CD103+ (Fig. 1). For most cases, few CD62L-selectin+ lymphocytes occurred in the ectocervical mucosa, and these were restricted to the lamina propria. TIA1+ lymphocytes varied from few to abundant and were distributed between the epithelium and lamina propria. Although more TIA1+ lymphocytes were found in the epithelium of the ectocervix than in the epithelium of the vagina, this still represented only a small proportion of the total lymphocyte population (Figs. 1 and 4H). Most of the ectocervical samples contained few CD57+ and CD56+ cells. For several cases, however, higher concentrations of these cells were present, located mainly in the lamina propria or at the base of the epithelium. No 
TCR lymphocytes were detected in the ectocervical mucosa. All ectocervical samples contained abundant macrophages in the lamina propria (Fig. 4I). Significantly more CD1a+ dendritic cells were present in the ectocervical mucosa than in the vaginal mucosa (Fig. 1); they were usually distributed either in the suprabasal region (Fig. 4I) or throughout the width of the epithelium (Fig. 4J).
In summary, for the majority of patients, significantly more CD1a+ dendritic cells, CD4+ T cells, and CD8+ cells expressing TIA1 and CD103 were present in the ectocervical epithelium compared with the vaginal epithelium.
Three patients were diagnosed with cervicitis. Of these samples, the ectocervical epithelium of one patient with chronic inflammation contained numerous CD8+ IELs and CD1a+ dendritic cells. The other two patients, diagnosed with both acute and chronic inflammation of the ectocervix, displayed a different immune profile, with numerous CD8+ and CD4+ IELs (Fig. 5A), many of which expressed TIA1 and CD103, but few CDla+ dendritic cells, which were mostly rounded in shape (Fig. 5B).
|
Transformation Zone Of all the female genital tract tissues examined, the TZ contained the highest concentrations of macrophages and CD4+ and CD8+ lymphocytes. Numerous granulocytes were also observed in this region. CD8+ (Fig. 6A), CD45RO+, and CD103+ (Fig. 6B) lymphocytes often occurred as focal accumulations in the lamina propria of the TZ. Although CD45RA+, CD62L-selectin+, and CD57+ cells were rarely observed, for several samples small accumulations of CD56+ cells were present in this region. CD1a+ dendritic cells were observed on the ectocervical side of the TZ, and were present at approximately the same density as in the ectocervix. No CD1a+ dendritic cells were observed on the endocervical side of the TZ.
|
Endocervix Like the TZ samples, endocervical samples often contained granulocytes that in some cases were very numerous and located in both the stromal and epithelial layers. For most specimens, few CD8+ and CD4+ lymphocytes were detected in the lamina propria and in the lumenal and glandular epithelium (Fig. 7, A and B). Concentrations of CD8+ IELs ranged from 1 to 5 per 100 epithelial cells (median of 3) and CD4+ IELs ranged from 0 to 4 (median of 1). Endocervical lymphocytes expressed a combination of CD45RO+, CD45RA+, CD62L+, and CD103+ phenotypes (Fig. 7C). In contrast to the CD8+ T cells in the vagina, which did not express TIA, most if not all CD8+ cells in the endocervix were TIA1+ (Fig. 7D). These cells were distributed throughout the lamina propria and were also detected in the lumenal epithelium of the cervix and cervical glands (range = 0–3, median = 1). A few CD57+ cells were detected in the endocervical lamina propria. CD56+ cells, however, were consistently present in all samples and in general, although not abundant, appeared to be present in greater numbers than CD57+ cells (Fig. 7E). Macrophages occurred in the lamina propria of the endocervical mucosa in varying concentrations (Fig. 7F), with intraepithelial macrophages a consistent component of the lumenal and glandular epithelium (Fig. 7G). CD1a+ dendritic cells were not detected in any endocervical specimens.
|
Three patients diagnosed with chronic inflammation of the endocervix had a different immune cell profile. Abundant CD8+ and CD4+ lymphocytes, often occurring as focal accumulations, were detected in these endocervical tissues. Further, in several of these samples, the CD8+ and CD4+ T cells tended to be localized just beneath the thin layer of columnar epithelium lining the endocervical mucosa (Fig. 8, A and B).
|
Hormonal Effects
Concentrations of T lymphocyte and APC populations were compared in normal vaginal and cervical samples collected at different stages of the menstrual cycle and postmenopause. No significant differences were detected in the abundance of any IEL subpopulation or CD1a+ cells when luteal and follicular phases of the menstrual cycle were compared, or in postmenopausal samples. It should be noted, however, that the standard deviations were high in all three groups (Fig. 9).
|
DISCUSSION
We believe that this is the first study to quantify lymphocyte subpopulations and APCs in different regions of the lower human female genital tract. In addition to distinguishing the different anatomical regions of the lower genital tract, we also differentiated between intraepithelial and subepithelial cell populations at each site. We encountered dramatic intrasubject variation in the abundance and intraepithelial vs. subepithelial localization of immune cell populations, even when variables such as inflammation and menstrual cycle were taken into account. Intrasubject variation may be attributable to recent use of vaginal products, unprotected intercourse, or undiagnosed subclinical genital infections, all of which can affect levels of lymphocytes in the lower genital tract [25–27]. Women with clinical diagnoses of cervicitis and vaginitis as a group had elevated levels of T lymphocytes and APCs within the epithelial layer of the affected tissue, but even in this group there were striking differences in types and activation status of cells within the intraepithelial infiltrates. Widely varying intraepithelial T lymphocyte profiles detected between patients could explain previously reported contradictory results on the prevalence of lymphocyte subpopulations in the cervix in studies involving small numbers of subjects [9–11].
Despite the high degree of intrasubject variation, because of the large number of samples studied, several observations could be made concerning regional differences in cellular distribution within the lower genital tract. The introitus and vagina were remarkable for their paucity of leukocytes, particularly within the epithelium. The few immune cells present consisted primarily of CD1a+ dendritic cells and CD8+ lymphocytes. Only a small proportion of vaginal IELs expressed CD103, an integrin almost exclusively expressed by IELs at many mucosal sites [28]. CD103 is the ligand for E-cadherin, which is expressed by mucosal epithelial cells; this interaction is thought to retain lymphocytes within the epithelium [29]. Because they do not express CD103, the majority of T-cells in vaginal mucosae may not be classical mucosal IELs. It is unknown how long these immune cells are retained within the epithelium or what factors induce and/or regulate the migration of these cells.
Leukocytes in the vaginal lamina propria consisted of CD4+ and CD8+ T cells and macrophages. In several patients, small focal accumulations of lymphocytes were detected, but there was no evidence of organized lymphoid nodules. This is in contrast to reports of lymphoid cell aggregates in the mucosa of the simian vagina [30]. In our study, T cells in both the epithelium and the lamina propria were predominantly of the memory (CD45RO+) phenotype, indicating most of these cells had already encountered antigen. A striking feature of the CD8+ T cell population in the vaginal mucosa was the absence of TIA1, a marker that indicates cytotoxic potential in T lymphocytes and NK cells [31]. Although functional studies were not carried out on vaginal IELs, the lack of TIA1 suggests that the majority of CD8+ T cells in the normal vagina may not be able to perform cytolytic functions; it is possible that these cells represent a population of regulatory T cells such as have been described in the gut mucosa [32, 33].
In most women, the ectocervical mucosa contained significantly higher concentrations of CD4+ IELs (P < 0.036) and CD1a+ dendritic cells (P < 0.0002) than the vaginal mucosa. Because these cell types are implicated in the sexual transmission of HIV-1, this finding provides evidence that HIV infection may be more common in the ectocervix than in the vagina in women without inflammatory conditions. A different cellular pattern has been observed in reproductive tissues of rhesus macaques, in which dendritic cells are more numerous in the vaginal epithelium than in the cervical epithelium [30]. This correlates with data indicating that the vagina is an important SIV transmission site in macaques [8]. In light of these differences, caution should be used when extrapolating data from SIV transmission studies to HIV transmission health issues. Another notable difference between the ectocervix and the vagina, in our study, was the significantly higher percentage of TIA1+ IELs in the ectocervix than in the vagina (P = 0.001). This suggests that the ectocervix may be more active in cytotoxic cellular immune defense functions.
In both the vaginal and ectocervical mucosae, the majority of lymphocytes were often found along the stroma-epithelium interface. Although the functional significance of this topographical pattern is unknown, this location places these lymphocytes in a favorable position for migration into the epithelium. Also of note was the preferential location of several lymphocyte subpopulations within the stroma of epidermal papillae and the epithelium surrounding these structures. Dermal papillae were present throughout the vagina and ectocervix, and were especially numerous and pronounced in the distal vagina, near the vaginal opening. The stratified epithelium lining the vagina and ectocervix presents a formidable physical barrier to lymphocyte migration. The vascular dermal papillae, containing a central capillary arcade and possibly lymphatic vessels, could provide a rapid and efficient portal of entry and exit for leukocytes, and facilitate their movement through the epithelium.
CD57+ and CD56+ NK cells kill virus-infected as well as neoplastic cells, and are important first line mediators of immune defense at a number of mucosal sites [34]. CD57+ cells were detected in both vaginal and ectocervical tissue. For most patients these cells were sparse, but several samples contained high numbers of these cells, especially in the ectocervical lamina propria. A similar pattern was observed for CD56+ cells, with an increased number in the ectocervix compared to the vagina. In a few samples, large numbers of CD56+ cells were also found in the epithelium. The cases with large numbers of NK cells may have had an underlying subclinical infection; NK cells have been reported in the TZ of women infected with HPV [13].
The most striking observation from our study was the preferential localization of lymphocytes and APCs to the cervical transformation zone. Accumulations of CD8+, TIA1+ T cells at this site provide evidence that the TZ functions as an immunologically dynamic barrier to ascending pathogens. At the same time, a higher concentration of CD4+ cells may also make this a preferred site of HIV-1 infection.
The endocervical epithelium contained the lowest concentrations of T cells and macrophages of any of the genital tract tissues evaluated in this study. Endocervical T cells did, however, express TIA1, suggesting functional effector activity. It is possible that less reliance is placed on cellular immunity at this site because of its ability to mount a vigorous mucosal secretory immune response (reviewed in Kutteh [1]). Further, no CD1a+ dendritic cells were observed in the endocervical epithelium, glands, or lamina propria in any of our cases. A recent report [35] citing CD1a+ cells in human endocervical samples relied on cytobrush sampling, which can easily pick up cells from the TZ, where CD1a+ cells are numerous. A study on the number and distribution of immune cells in genital tract tissues of Rhesus macaques also reported CD1a+ cells in endocervical tissues [30]; this apparent species difference could have an effect on HIV/SIV infection mechanisms as well as on the location of immune responses following infection or vaccination.
Like several other studies [14–16], our study did not detect any significant differences in leukocyte numbers or localization between the proliferative vs. secretory phases of the menstrual cycle, or in postmenopausal women. However, the female sex hormones have been shown to affect a variety of immune functions [36], and reportedly affect cytotoxic T cell function in the uterus [16]. More research is needed on this important topic.
Inflamed vaginal and cervical samples contained higher concentrations of several IEL subpopulations (TIA1+, CD8+ T cells; CD4+ T cells; and CD103+ lymphocytes) compared to noninflamed tissues. Further, immature immune cell phenotypes characteristically found in the stroma, i.e., CD45RA and CD62L, were also present in higher concentrations in inflamed tissues. In contrast, fewer CDla+ dendritic cells were found in the epithelium of inflamed vaginal and ectocervical mucosae, and, when present, these cells often had a rounded shape. These changes in morphology and abundance of CD1a+ cells are similar to those reported for CD1a+ cells in the ectocervical epithelium of women infected by HPV [37, 38] or HIV [39], as well as in malignant tumors of the skin [40]. It has been suggested that loss of CDla+ cells from the epithelium is caused by migration of antigen-primed dendritic cells to the lymph nodes [38].
We have shown distinct regional differences in abundance and phenotype of immune cells in the lower female genital tract. This information has relevance for vaccine development against sexually transmitted pathogens such as HIV-1. The high concentration of CD8+, TIA1+ T lymphocytes in the ectocervix and TZ indicates that these regions could be major effector sites for cytotoxic T lymphocytes responses. The ectocervix and TZ contain the highest number of antigen-presenting cells and therefore also appear to be the best location for the induction of CMI immune responses in the lower genital tract. In contrast, the normal endocervical epithelium contains fewer T lymphocytes, but expresses the poly-IgR and contains numerous IgA+ and IgM+ plasma cells that mediate humoral immune responses [1]. Based on this information, we propose that vaccines intended to elicit immunity in the genital tract of women be targeted to the ectocervix/TZ to elicit a specific cytotoxic T lymphocytes response, and to the endocervix to stimulate a local secretory mucosal immune response. Our observation of a preponderance of HIV-1 host cells in the cervix also suggests that a barrier method such as the diaphragm that covers the cervix during intercourse could reduce the sexual transmission of HIV-1.
ACKNOWLEDGMENTS
We are indebted to the doctors of the Obstetrics and Gynecology Department at Brigham and Women's Hospital, in particular Drs. Joseph Hill and Eboo Versi, for providing us with fresh surgical specimens. We also thank Dr. Gary Russell (Department of Pediatrics, Massachusetts General Hospital) for providing a biopsy of large intestine, and Drs. Paul Anderson, Michael Brenner (Department of Rheumatology, Brigham and Women's Hospital) and L.J. Picker (Oregon Health Sciences University, Beaverton, OR) for their kind gifts of antibodies. A special thanks goes to Dr. Joseph Politch for his considerable assistance in statistical analysis and preparation of the manuscript.
FOOTNOTES
1 Supported in part by NIH grant R01HD33205 and by contract CSA-88–020 from the Contraceptive Research and Development Program (CONRAD) under a cooperative agreement with the United States Agency for International Development (USAID), which in turn receives funds for AIDS research from an interagency agreement with the National Institutes of Child Health and Human Development. The views expressed by the authors do not necessarily reflect the views of USAID or CONRAD. 
2 Correspondence. FAX: 617 414 8481; jeffrey.pudney{at}bmc.org
3 Current address: Division of Reproductive Biology, Department of Obstetrics and Gynecology, Boston University School of Medicine, 715 Albany St.-Evans 230, Boston, MA 02118.
4 Current address: Department of Microbiology, Immunology, and Parasitology, LSU Health Sciences Center, 1901 Perdido St., New Orleans, LA 70112-1393.
Received: 22 April 2005.
First decision: 18 May 2005.
Accepted: 2 August 2005.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  O. Heikinheimo and P. Lahteenmaki Contraception and HIV infection in women Hum. Reprod. Update, November 1, 2008; (2008) dmn049v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Schulke, F. Manconi, R. Markham, and I.S. Fraser Endometrial dendritic cell populations during the normal menstrual cycle Hum. Reprod., July 1, 2008; 23(7): 1574 - 1580. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Van Herrewege, L. Morellato, A. Descours, L. Aerts, J. Michiels, L. Heyndrickx, L. Martin, and G. Vanham CD4 mimetic miniproteins: potent anti-HIV compounds with promising activity as microbicides J. Antimicrob. Chemother., April 1, 2008; 61(4): 818 - 826. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Cassone, F. De Bernardis, and G. Santoni Anticandidal Immunity and Vaginitis: Novel Opportunities for Immune Intervention Infect. Immun., October 1, 2007; 75(10): 4675 - 4686. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
