- Split View
-
Views
-
Cite
Cite
B.F. Barrier, B.S. Kendall, C.E. Ryan, K.L. Sharpe-Timms, HLA-G is expressed by the glandular epithelium of peritoneal endometriosis but not in eutopic endometrium, Human Reproduction, Volume 21, Issue 4, April 2006, Pages 864–869, https://doi.org/10.1093/humrep/dei408
- Share Icon Share
Abstract
BACKGROUND: HLA-G is a major histocompatability antigen with documented immune-regulatory function. Various epithelial cancers and tissue allografts have been noted to express HLA-G, which is postulated to aid in their escape from immunosurveillance. We evaluated peritoneal endometriosis and eutopic endometrium for the expression of HLA-G protein and gene transcript. METHODS: Two experiments were performed: (i) archived tissue blocks from peritoneal endometriotic lesions (n = 15) and eutopic endometrium (n = 12) were evaluated for extent of protein immunostaining, and (ii) eutopic endometrial biopsies from women without (n = 17) and with (n = 24) endometriosis, and peritoneal endometriotic lesions (n = 14) were evaluated for presence of RNA transcript by in situ hybridization. RESULTS: HLA-G protein localized in the glandular epithelium of 14 of 15 (93.3%) peritoneal endometriotic lesions, but not in stromal cells. HLA-G protein staining was absent in endometrial biopsies (n = 12). HLA-G gene transcript localized to the glandular epithelium in 13 of 14 (92.8%) peritoneal endometriotic lesions. HLA-G transcript was never observed in eutopic endometrium, regardless of cycle stage or whether from women with (n = 24) or without (n = 18) endometriosis. CONCLUSIONS: HLA-G is expressed by endometriotic glandular epithelium but not by eutopic endometrium under normal conditions. Differential expression of HLA-G suggests that peritoneal inflammation or cellular stress may up-regulate mechanisms to promote ectopic endometrial survival.
Introduction
Endometriosis is defined by the presence of uterine endometrium in ectopic loci. It is associated with pain and infertility in millions of women worldwide. In 1992, the estimated national inpatient cost for hospitalizations in the USA for endometriosis as the primary diagnosis was US $579 million (Zhao et al., 1998), and assuming a 15% annual increase in healthcare costs, likely topped US $3 billion in 2004. The global impact of this disease is enormous. There is no dependable long-term cure for endometriosis. Given the global economic and social impact of endometriosis, interest has grown in the elucidation of basic mechanisms of its pathogenesis.
HLA-G is a non-classical MHC class Ib antigen with limited polymorphism and restricted tissue expression (reviewed in Hunt et al., 2005). Constitutive expression of HLA-G has been described only in thymic epithelial cells and placental chorionic endothelium and trophoblast subpopulations (Kovats et al., 1990; McMaster et al., 1995; Blaschitz et al., 1997; Crisa et al., 1997). HLA-G expression may be up-regulated in otherwise non-expressive tissues through inflammatory challenge. For example, HLA-G is up-regulated in activated monocytes (Yang et al., 1996), mucosal epithelia in transplantation allografts (Lila et al., 2000; Creput et al., 2003), epithelial cancers (Fukushima et al., 1998; Singer et al., 2002; Ibrahim et al., 2004) and inflamed skeletal muscle (Wiendl et al., 2000, 2003). Inflammatory cytokines known to up-regulate expression of HLA-G protein and/or mRNA include interferon-γ, interferon-β and interleukin-10 (Yang et al., 1995; Chu et al., 1999; Moreau et al., 1999).
Through counter-regulation in response to local inflammatory processes, it is possible that HLA-G plays an important role in the maintenance of immune tolerance and prevention of autoimmunity. HLA-G is known to inhibit T-cell- and NK-cell-mediated cytolysis (Le Gal et al., 1999; Riteau et al., 2001; Wiendl et al., 2002) and to suppress alloreactive CD4+ T-cell proliferation in vitro (Bainbridge et al., 2000; Lila et al., 2001). HLA-G transfected antigen-presenting cells promote the development of regulatory CD4+ T cells (LeMaoult et al., 2004). Recombinant soluble HLA-G1 and HLA-G2 have both been demonstrated to decrease the expression of CD8-α in interferon-γ-stimulated CD8+ T-cells (Morales et al., 2003). Dendritic cells expressing HLA-G promote the differentiation of both anergic and regulatory CD4+ and CD8+ T cells (Ristich et al., 2005). It is speculated that these counter-regulatory functions protect HLA-G-expressing tissues from immune recognition and destruction.
Less than 10% of reproductive-aged females develop endometriosis, and the reason for this is currently unknown. Many investigators have reported differences in functional immune cell phenotypes between women with and without the disease (Oosterlynck et al., 1991, 1993; Kanzaki et al., 1993; Ho et al., 1997; Szyllo et al., 2003). Despite this, evidence for a direct mechanism for endometriotic cell escape from immunosurveillance has thus far eluded investigators. Because HLA-G appears to confer protection from cell-mediated cytolysis, we chose to evaluate both endometriotic lesions and eutopic endometrium for expression of this non-classical MHC class Ib antigen.
Materials and methods
Evaluation for HLA-G protein in peritoneal endometriotic lesions and eutopic endometrium was performed at Wilford Hall Medical Center (WHMC) with the approval of the WHMC Institutional Review Board. A pathology database was searched in order to obtain representative cases of histologically confirmed endometriosis. Pathological tissues were evaluated, and therefore it was not necessary to obtain informed consent. A group of 25 peritoneal implants was initially selected for evaluation. Only 15 peritoneal endometriotic lesions were found to contain the unequivocal presence of both glands and stroma and therefore were processed for immunohistochemical evaluation of HLA-G protein. Many of these tissue blocks were sent from outside hospitals, and clinical data were not available for correlation. Further, 12 biopsies of eutopic endometrial tissue containing both proliferative (n = 7) and secretory (n = 5) phase endometrium from women without endometriosis were also evaluated for the presence of HLA-G protein by immunohistochemistry.
Additional eutopic and ectopic endometrial tissues were tested for the presence of HLA-G gene transcript using RNA in situ hybridization at the University of Missouri (MU), Department of Obstetrics, Gynecology and Women’s Health with the approval of the MU Institutional Review Board–Health Sciences Section. Informed consent was obtained from participants prior to enrolment in this study. HLA-G gene transcript was evaluated in two groups of tissues.
First, 42 eutopic endometrial biopsies from women with endometriosis (n = 24; 14 proliferative, nine secretory and one unknown) and from fertile women without endometriosis undergoing elective laparoscopic sterilization (n = 18; nine proliferative and nine secretory) were evaluated by RNA in situ hybridization for the presence of HLA-G gene transcript. Menstrual cycle data were based on the date of the last menstrual period and histological confirmation by the University of Missouri, Department of Pathology. All women had regular menstrual cycles and had not taken reproductive steroid-modulating medications within the past 6 months.
Second, peritoneal endometriotic lesions (n = 14) were evaluated for the presence of HLA-G gene transcript by RNA in situ hybridization including: (i) matched peritoneal endometriotic lesions (n = 6) obtained from the group of women undergoing co-evaluation of eutopic endometrium (n = 3 proliferative, two secretory and one menstrual); (ii) four additional peritoneal endometriotic lesions obtained at MU that did not have matching eutopic endometrial biopsies; and (iii) peritoneal endometriotic lesions from WHMC previously evaluated by immunohistochemistry for the presence of HLA-G protein (n = 4).
All tissues included in the study were formalin-fixed, paraffin-embedded and sectioned at 5 µm. Semi-adjacent tissue sections were placed on charged slides. Immunohistochemistry and in situ hybridization and were performed as described below.
HLA-G immunohistochemistry
Tissue sections were heat-fixed, deparaffinized, and rehydrated by standard methods. Tissues were blocked for endogenous peroxidase activity and then heat-treated (3 min, 120°C at 15–20 psi) in 0.01 mol/l citrate buffer (pH = 6.0). The primary anti-HLA-G antibody 4H84, mouse IgG1 isotype, isolated from mouse ascitic fluid at a concentration of 3.1 mg/ml (generously provided by Dr M.McMaster, University of California, San Francisco, CA, USA) was used at a dilution of 1:6000. This monoclonal antibody is reactive to amino acids 61–83 of the α 1 domain of denatured HLA-G (McMaster et al., 1995) and recognizes all HLA-G isotypes.
Following incubation with the 4H84 antibody for 35 min at room temperature, the tissue sections were washed with phosphate-buffered saline (PBS). A secondary biotinylated anti-mouse antibody (BioCare Medical, Walnut Creek, CA, USA) was added for 10 min at room temperature, then washed with PBS. Streptavidin–horseradish peroxidase complex (BioCare Medical) was then incubated for 10 min at room temperature and washed with PBS. Diaminobenzidine tetrahydrochloride (DAB) chromogen (BioCare Medical) was finally added for 5 min at room temperature. Treated tissue sections were counterstained with haematoxylin. A similarly prepared section of third trimester placenta, exhibiting appropriate staining of intermediate trophoblast, was prepared with each assay as a positive control. A negative control specimen, following all steps except addition of the primary antibody (4H84), was performed from a section of each case. In a subset of 10 samples an additional negative control was run using an identical isotype anti-desmin mouse IgG1 antibody at 1:200 concentration (DakoCytomation, Carpenteria, USA) instead of the primary antibody.
The stained tissues were examined by a pathologist experienced in interpreting immunohistochemical stains (B.S.K.), and evaluated for both extent and intensity of staining. The extent of HLA-G protein staining was assessed utilizing a semi-quantitative scale. Distribution of positive cells was grouped from 0 to 4: 0 for 0%, 1 for 1–10%, 2 for 11–25%, 3 for 26–50%, and 4 for >50%. The HLA-G intensity was scored semi-quantitatively from 0 to 3, with a score of 3 for intensity comparable to the staining of the positive control, 0 comparable to the staining of the corresponding negative control, and 1 and 2 as gradations.
HLA-G in situ hybridization
In situ hybridization was performed by using the GenPoint Tyramide Signal Amplification System for Biotinylated Probes (DakoCytomation, Carpenteria, CA, USA) according to manufacturer’s instructions. Briefly, after tissue sections were deparaffinized and rehydrated they were pretreated with Target Retrieval Solution S1699 (DakoCytomation) at 95°C for 20 min. After cooling to room temperature, the Target Retrieval Solution was removed and the tissues incubated in proteinase K (1:5000 dilution) in 50 mmol/l Tris pH 7.6 for 5 min. Prior to applying the biotinylated probes, endogenous peroxide was quenched by incubation with 0.3% hydrogen peroxide in methanol for 20 min.
Antisense and sense biotinylated cDNA probes were constructed against a sequence found in exon 2 of human HLA-G mRNA as previously described (Houlihan et al., 1992). Exon 2 codes for the α 1 domain common to all isotypes of HLA-G. The biotinylated probes were synthesized and purified by high-pressure liquid chromatography (Synthegen, Houston, TX, USA).
The probes were diluted to a final concentration of 70 ng/ml in the probe dilutent supplied with the system. Hybridization was performed overnight at 37°C. The non-bound probes were removed by rinsing with Tris-buffered saline with Tween-20, pH 7.6 (TBST) (DakoCytomation, CA, USA), and a stringent wash with sodium chloride/sodium citrate (SSC) (DakoCytomation) at 40°C for 20 min. The SSC was removed with TBST and signal amplification was performed with primary anti-streptavidin–horseradish peroxidase (HRP) (15 min), biotinylated tyramide solution (15 min), and secondary streptavidin–HRP (15 min) with alternating washes with TBST. Diaminobenzidine (1:50 dilution) was applied as the chromogen for 5 min and the reaction was stopped by water. Harris haematoxylin (1:4 dilution) was applied for 5 min as a nuclear counterstain.
Cytotrophoblast from first trimester human placenta served as a positive control. Sense strand biotinylated probe served as a negative control on serial tissue sections.
Results
Protein immunohistochemistry
Fourteen of 15 (93.3%) peritoneal endometriotic lesions demonstrated distinct HLA-G immunostaining of glandular epithelium, but not the stroma, using the 4H84 anti-HLA-G antibody. Staining of 12 eutopic endometrial biopsy specimens showed no distinct endometrial glandular or stromal staining in any specimen.
In five peritoneal lesions (38%), the distribution of staining encompassed >50% of the epithelial glandular cells, with moderate to strong intensity of staining. In the remaining lesions, staining was moderate to weak, with varied distribution ranging from <10% to nearly 50% of cells. Data are detailed in Table I. Representative examples of immunohistochemical staining are shown in Figure 1.
Immunohistochemistry . | . | . | In situ hybridization (intensity only) . | . | . | ||||
---|---|---|---|---|---|---|---|---|---|
Case no. . | Distribution . | Intensity . | Case no. . | Glands . | Stroma . | ||||
6914 | 3 | 2 | 4403 | 4 | 1 | ||||
5434 | 4 | 3 | 8934 | 4 | 0 | ||||
9275 | 4 | 3 | 2017 | 3 | 0 | ||||
3868 | 2 | 1 | 3828 | 0 | 0 | ||||
7443 | 1 | 1 | 73 | 3 | 0 | ||||
0494 | 1 | 1 | 109 | 1 | 0 | ||||
2824 | 3 | 1 | 128 | 1 | 0 | ||||
3379 | 4 | 2 | 157 | 3 | 0 | ||||
6160 | 4 | 2 | 250 | 1 | 0 | ||||
1954 | 2 | 1 | 254 | 2 | 0 | ||||
0048 | 2 | 2 | 256 | 1 | 0 | ||||
4403 | 4 | 2 | 327 | 1 | 0 | ||||
8934 | 0 | 0 | 244 | 1 | 0 | ||||
2017 | 2 | 1 | 299 | 1 | 0 | ||||
3828 | 3 | 2 |
Immunohistochemistry . | . | . | In situ hybridization (intensity only) . | . | . | ||||
---|---|---|---|---|---|---|---|---|---|
Case no. . | Distribution . | Intensity . | Case no. . | Glands . | Stroma . | ||||
6914 | 3 | 2 | 4403 | 4 | 1 | ||||
5434 | 4 | 3 | 8934 | 4 | 0 | ||||
9275 | 4 | 3 | 2017 | 3 | 0 | ||||
3868 | 2 | 1 | 3828 | 0 | 0 | ||||
7443 | 1 | 1 | 73 | 3 | 0 | ||||
0494 | 1 | 1 | 109 | 1 | 0 | ||||
2824 | 3 | 1 | 128 | 1 | 0 | ||||
3379 | 4 | 2 | 157 | 3 | 0 | ||||
6160 | 4 | 2 | 250 | 1 | 0 | ||||
1954 | 2 | 1 | 254 | 2 | 0 | ||||
0048 | 2 | 2 | 256 | 1 | 0 | ||||
4403 | 4 | 2 | 327 | 1 | 0 | ||||
8934 | 0 | 0 | 244 | 1 | 0 | ||||
2017 | 2 | 1 | 299 | 1 | 0 | ||||
3828 | 3 | 2 |
Stromal protein staining was not convincingly detected in any endometriotic lesion. In situ hybridization intensity of staining for both glands and stroma, with one specimen containing scattered stromal signal. The HLA-G intensity was scored semi-quantitatively from 0 to 3, with a score of 3 for intensity comparable to the staining of the positive control, 0 comparable to the staining of the corresponding negative control, and 1 and 2 as gradations. Distribution of positive cells was grouped from 0 to 4: 0 for 0%, 1 for 1–10%, 2 for 11–25%, 3 for 26–50%, and 4 for >50%.
Immunohistochemistry . | . | . | In situ hybridization (intensity only) . | . | . | ||||
---|---|---|---|---|---|---|---|---|---|
Case no. . | Distribution . | Intensity . | Case no. . | Glands . | Stroma . | ||||
6914 | 3 | 2 | 4403 | 4 | 1 | ||||
5434 | 4 | 3 | 8934 | 4 | 0 | ||||
9275 | 4 | 3 | 2017 | 3 | 0 | ||||
3868 | 2 | 1 | 3828 | 0 | 0 | ||||
7443 | 1 | 1 | 73 | 3 | 0 | ||||
0494 | 1 | 1 | 109 | 1 | 0 | ||||
2824 | 3 | 1 | 128 | 1 | 0 | ||||
3379 | 4 | 2 | 157 | 3 | 0 | ||||
6160 | 4 | 2 | 250 | 1 | 0 | ||||
1954 | 2 | 1 | 254 | 2 | 0 | ||||
0048 | 2 | 2 | 256 | 1 | 0 | ||||
4403 | 4 | 2 | 327 | 1 | 0 | ||||
8934 | 0 | 0 | 244 | 1 | 0 | ||||
2017 | 2 | 1 | 299 | 1 | 0 | ||||
3828 | 3 | 2 |
Immunohistochemistry . | . | . | In situ hybridization (intensity only) . | . | . | ||||
---|---|---|---|---|---|---|---|---|---|
Case no. . | Distribution . | Intensity . | Case no. . | Glands . | Stroma . | ||||
6914 | 3 | 2 | 4403 | 4 | 1 | ||||
5434 | 4 | 3 | 8934 | 4 | 0 | ||||
9275 | 4 | 3 | 2017 | 3 | 0 | ||||
3868 | 2 | 1 | 3828 | 0 | 0 | ||||
7443 | 1 | 1 | 73 | 3 | 0 | ||||
0494 | 1 | 1 | 109 | 1 | 0 | ||||
2824 | 3 | 1 | 128 | 1 | 0 | ||||
3379 | 4 | 2 | 157 | 3 | 0 | ||||
6160 | 4 | 2 | 250 | 1 | 0 | ||||
1954 | 2 | 1 | 254 | 2 | 0 | ||||
0048 | 2 | 2 | 256 | 1 | 0 | ||||
4403 | 4 | 2 | 327 | 1 | 0 | ||||
8934 | 0 | 0 | 244 | 1 | 0 | ||||
2017 | 2 | 1 | 299 | 1 | 0 | ||||
3828 | 3 | 2 |
Stromal protein staining was not convincingly detected in any endometriotic lesion. In situ hybridization intensity of staining for both glands and stroma, with one specimen containing scattered stromal signal. The HLA-G intensity was scored semi-quantitatively from 0 to 3, with a score of 3 for intensity comparable to the staining of the positive control, 0 comparable to the staining of the corresponding negative control, and 1 and 2 as gradations. Distribution of positive cells was grouped from 0 to 4: 0 for 0%, 1 for 1–10%, 2 for 11–25%, 3 for 26–50%, and 4 for >50%.
RNA in situ hybridization
In situ hybridization demonstrated HLA-G transcript localized to glandular epithelium in 13 of 14 (92.8%) peritoneal lesions tested. HLA-G transcript was not identified in any of the biopsies of eutopic endometrium from women with endometriosis (n = 24) or from disease-free controls (n = 18). Representative examples of in situ hybridization are seen in Figure 1.
Discussion
Both HLA-G protein and RNA gene transcript are commonly present in the glandular epithelium of endometriotic lesions, but not in eutopic endometrium, regardless of whether patients do or do not have endometriosis. HLA-G expression in the context of chronic inflammation is found in other tissues of epithelial origin (Yang et al., 1996; Fukushima et al., 1998; Lila et al., 2000; Singer et al., 2002; Creput et al., 2003; Ibrahim et al., 2004). It is noteworthy that, at least in peritoneal endometriosis, HLA-G expression appears to be limited to glandular epithelium, with little or no stromal involvement.
It could be argued that HLA-G expression by endometriotic lesions offers a means of escape from NK cell cytotoxocity. This argument may at first appear disingenuous, because classical HLA class Ia molecules, known to inhibit NK cells, are expressed by endometrium and by endometriosis. However, the ‘null-cell hypothesis’ may not be as clear as it once appeared. HLA class Ia antigens deliver an inhibitory signal to NK cells through binding KIR (Killer Immunoglobin like Receptors), but the signal transduction appears not to result in a binary, but rather in an analogue, response. Because of this, concurrent binding of activatory ligands may induce cell killing despite presence of HLA class Ia molecules. An example is the binding of the human NK activatory receptor NKG2D by ligands such as MICA, MICB, ULBP1, ULBP2, ULBP3 or ULBP4, which may be produced by stressed cells and which activate NK cell killing. This revisitation of the ‘missing-self hypothesis’ has been recently reviewed (Lanier, 2005). Evidence for this includes the reported NK cell-mediated killing of an MHC class Ia-bearing tumour (Cerwenka et al., 2001) and the observation that NK cells are able to lyse both autologous and heterologous endometrial cells, which express HLA class Ia antigens (Oosterlynck et al., 1991).
Additionally, it has been demonstrated that endometrial cells in culture may down-regulate HLA class Ia molecules and increase their susceptibility to killing by natural killer (NK)-like T lymphocytes (Semino et al., 1995). A paper in the Chinese literature reported that the epithelia of endometriotic lesions themselves have decreased expression of HLA class Ia and increased expression of HLA class II antigens compared to eutopic control endometrium (Xu et al., 2001). Such softening of HLA class Ia expression by endometriotic epithelia may further increase their susceptibility to NK-cell-mediated cytotoxicity, and highlight the potential significance of HLA-G expression by endometriotic lesions.
The observation that HLA-G is expressed differently between eutopic and ectopic endometrial tissues merits discussion. In its eutopic environment, endometrial tissue has unique requirements for survival. Species propagation would be halted if infection and concurrent inflammation were to damage the endometrium, or lead to destruction of the fetus. We speculate that, in order to protect the intrauterine environment, counter-regulatory mechanisms are employed in the presence of inflammation, reinforced by strong selective pressure, to promote endometrial as well as placental survival. In ectopic loci, endometrial implants appear to benefit from protective mechanisms, including: (i) endometrial secretion of interleukin-10 (Zhdanov et al., 2003); (ii) expression of Fas ligand (Garcia-Velasco et al., 1999); (iii) expression of complement regulatory proteins such as decay-accelerating factor (Young et al., 2002) and membrane cofactor protein (CD48) (D’Cruz and Wild, 1992); (iv) expression of soluble intercellular adhesion molecule (ICAM)-1 (Somigliana et al., 1996); (v) secretion of endometrial haptoglobin (Sharpe-Timms et al., 2000, 2002); and (vi) induction of anti-apoptotic proteins such as bcl-2 (Li et al., 2003). Thus the endometriotic tissues expend a significant amount of energy to mitigate the effects of local inflammation, and the up-regulation of HLA-G may be an additional means to this end.
All women appear to shed endometrial cells into the peritoneal cavity, yet only a fraction of these develop endometriosis. The genetic regulation of HLA-G appears to be complex, differing from that of classical HLA class I antigens (Solier et al., 2001; Mouillot et al., 2005). Interestingly, a heat shock element (HSE) within the HLA-G promoter binds to heat shock factor 1 (HSF-1) during heat-shock and arsenite stress conditions in a melanoma cell line in which it is associated with up-regulation of HLA-G (Ibrahim et al., 2000). During menstruation, perhaps the shed endometrial epithelial cells of susceptible women are programmed to up-regulate HLA-G under conditions of stress, while menstrual endometrial epithelial cells from women without endometriosis are not. Future studies of HLA-G regulatory element polymorphisms in women with endometriosis may help to better define this process.
HLA-G may play a role in the limitation of autoimmunity through its recently elucidated role in the development of regulatory T-cells (Morales et al., 2003; LeMaoult et al., 2004; Ristich et al., 2005). Endometriotic implants surviving in the context of chronic inflammation may stimulate weakly self-reactive lymphocytes and thereby lead to the development of autoimmunity. Indeed, anti-endometrial antibodies may be found in 7% of women with mild endometriosis and up to 45% of women with moderate to severe disease (Iborra et al., 2000). Many of these antibodies are specific for a specific carbohydrate epitope, the Thomsen–Friedenreich (T) antigen, shared by a multitude of other tissues in the body (Yeaman et al., 2002). HLA-G expression by ectopic endometriotic glandular epithelium may represent an attempt to escape anti-endometrial adaptive immune responses that could be detrimental to reproductive health and, more broadly, species propagation.
HLA-G is expressed by mucosal epithelia in both transplantation allografts and adenocarcinomas—other tissues that appear to thrive despite local inflammation. Because they lack the physical protection afforded by keratinized epithelia, mucosal epithelia throughout the body must promote the dual purposes of preventing both infection and also immunization against innocuous antigens. Similarly, the endometrium must discriminate seminal and fetal antigens from those that are pathogenic. The default immunological response in the female reproductive tract, like that of other mucosa, seems to be toward tolerance of non-pathogenic foreign antigens (Black et al., 2000; Robertson et al., 2003). In the case of endometriosis, this mucosal epithelium is displaced from an ostensibly protective environment to a less hospitable environment in which innate immune processes are chronically activated. There it may interpret local inflammatory clues in a manner that increases, rather than decreases, its ability to survive.
A previous study used the 4H84 antibody in a sensitive western blot technique and found no evidence of HLA-G protein in stromal cell cultures or peritoneal fluid (Hornung et al., 2001). Our observations confirm the relative absence of stromal involvement. The previous study also evaluated the cell lysate of two suspected endometriotic lesions for HLA-G protein using a sensitive western blot technique and found none. However, histological confirmation of the presence of glandular epithelium was not reported.
A recently published report has suggested that weak acid treatment (defined as pH = 3.0) causes dissociation of β2-microglobulin from the α chain of MHC class I antigens other than HLA-G (Polakova et al., 2004). This could potentially expose an antigenic site that cross-reacts with the 4H84 antibody, leading to false positive results. In our study, antigen retrieval was accomplished by heat treatment in a citrate buffer that kept the pH at 6.0. HLA-G expression was corroborated in a number of cases by the co-localization of HLA-G transcript in endometriotic glandular epithelium.
In summary, we have described the novel expression of HLA-G in human endometriotic lesions. The finding that a known immunomodulatory molecule is expressed by endometriotic lesions may offer new leverage to develop strategies to fight this complex disease.
Acknowledgements
The authors wish to thank Jacob Collins and Christine Yates for their technical assistance during this study.
References
Author notes
1Department of Obstetrics, Gynecology and Women’s Health, Division of Reproductive and Perinatal Research, University of Missouri–Columbia, Missouri and 2Department of Pathology, Wilford Hall Medical Center, Lackland A.F.B., Texas, USA