The establishment of a new vascular supply is essential for the survival of endometrial tissue and its development in ectopic locations. We have previously shown that ectopic endometrial cells release an important mitogenic activity for human endothelial cells and identified macrophage migration inhibitory factor (MIF) as one of the principal bioactive molecules involved in endothelial cell proliferation. In the present study, immunohistochemical and dual immunofluorescence analyses showed that MIF is effectively expressed by endometriotic tissue, particularly in the glands, and identified endothelial cells, macrophages, and T lymphocytes as cells markedly expressing MIF in the stroma. Western blot analysis showed a single 12.5-kDa band corresponding to the known mol wt of the molecule. The highest concentrations of MIF protein in endometriotic tissue, as measured by ELISA, were found in flame-like red endometriotic lesions, compared with typical black-bluish (P < 0.01) or with white lesions (P < 0.01). Interestingly, MIF displayed a marked expression in lesions from the initial stage of endometriosis (stage I). Semiquantitative RT-PCR analysis of MIF mRNA levels in the same endometriotic tissues showed a pattern of expression comparable with that of the protein. In view of its potent proinflammatory and angiogenic properties, local production of MIF within endometrial implants, particularly in those that are highly vascularized and representing the earliest and most active forms of the disease, make plausible the involvement of this factor in the local immunoinflammatory process observed in endometriosis and the initial steps of endometriotic tissue growth and development.
ENDOMETRIOSIS, ONE OF the commonest gynecological conditions, is defined as the presence of tissue histologically similar to endometrium at sites outside the uterine cavity. This disease is intriguing and unique in that it is the only known benign disease in which autologous cells, according to the widely accepted transplantation theory (1), can implant and develop in ectopic locations. The etiology of endometriosis is still not clearly defined. Genetic predisposition, environmental toxins, hormonal factors, and immune deficiency may contribute to the susceptibility of a woman to develop this disease (2). However, a key condition for endometrial tissue to survive and grow ectopically following successful adhesion and implantation is the establishment of an effective new blood supply, a process involving the generation of new blood vessels or angiogenesis (3, 4).
Early and most active endometriotic lesions are markedly vascularized; increased vascularization is seen at the implant surface and also in the surrounding peritoneal tissue (5–7), suggesting that endometriotic implant is capable of inducing its own neovascularization by deriving local microvasculature. It is therefore of a crucial interest to elucidate the mechanisms underlying endometriosis-associated angiogenesis and to identify the factors involved in that process. We have recently found that immortalized as well as primary ectopic endometrial epithelial cells release an important mitogenic activity for human endothelial cells and identified macrophage migration inhibitory factor (MIF) as an important mediator of endothelial cell proliferation (8). Originally described as a product of activated T lymphocytes that inhibited the random migration of cultured macrophages, MIF is now known as an important modulator for a variety of cell functions, including inflammatory and immune responses (9, 10). The identification of MIF as a mitogenic factor for endothelial cells released by ectopic endometrial cells is consistent with recent data showing an important role for MIF in tumor growth-associated angiogenesis in vivo and in vitro in autocrine regulation of endothelial cell proliferation (11–13). The present study was undertaken to assess the expression of MIF in endometriotic lesions, identify the site(s) of expression, and investigate whether such an expression varies according to the stage and the activity of the disease.
Materials and Methods
Source and handling of tissue
Endometriotic tissue specimens used in this study were obtained from women who provided informed consent for a research protocol approved by Saint-François d’Assise Hospital Ethics Committee on Human Research. These patients were found to have endometriosis during laparoscopy or laparotomy, had no endometrial hyperplasia or neoplasia, and had not received any antiinflammatory or hormonal medication during a period of at least 3 months before surgery. Endometriosis was staged according to the revised American Fertility Society classification system (14). The cycle phase (proliferative or secretory) was determined according to the patients’ cycle history and to the serum progesterone. The mean age was 34.6 plus or minus 6.0 yr. Endometriotic biopsies were immediately placed at 4 C in sterile HBSS (Life Technologies, Inc., Burlington, Ontario, Canada) containing 100 IU/ml penicillin, 100 μg/ml streptomycin, and 0.25 μg/ml amphotericin transported to the laboratory at which they were immediately washed in HBSS at 4 C, snap frozen on dry ice, or fixed in 10% formalin.
For immunohistochemical analysis, 25 endometriotic biopsies were included. Fifteen were frozen at −80 C in Tissue-Tek OCT compound (Miles Inc., Elkhart, IN), and 10 were embedded in paraffin. These biopsies were from women with endometriosis stage I (two typical black-blue, two red flame-like), stage II (two typical black-blue, three white, and one from the inner wall of ovarian endometrioma), stages III-IV (three typical black-blue, three white, and nine from the inner wall of ovarian endometrioma).
For Western blotting, ELISA, and RT-PCR analyses, 24 biopsies were taken. These biopsies were from women with endometriosis stage I (two typical black-blue, seven red flame-like, and three white), stage II (four typical black-blue, two red flame-like, and three white), and stages III-IV (two red flame-like and one white). Biopsies were all snap frozen and kept at −80 C in microcentrifuge tubes (Eppendorf, Gordon Technologies Inc., Mississauga, Ontario, Canada) until used.
Immunohistochemistry
Five-micrometer cryosections of Optimal Cutting Temperature-frozen endometriotic lesions were mounted on poly-l-lysine-coated microscope glass slides, fixed during 20 min in a 10% buffered formalin phosphate solution (Fisher Scientific, Montreal, Québec, Canada), and washed in PBS. Five-micrometer sections of paraffin-embedded tissues were mounted on poly-l-lysine-coated microscope glass slides, deparaffinized in toluene, rehydrated through graded solutions of ethanol and water, and washed in PBS. The subsequent steps were the same for cryosections and paraffin-embedded tissue sections. Briefly, after permeabilization with Triton-X-100 (1% in PBS) and elimination of endogenous peroxidase with H2O2 (0.3% in absolute methanol), tissue sections were successively incubated at room temperature for 90 min with a goat polyclonal antihuman-MIF antibody (R\|[amp ]\|D Systems, Minneapolis, MN) [0.66 μg/ml in PBS/0.2% BSA/0.01% Tween 20 (PBS/BSA/Tween)], 90 min with a biotin-conjugated rabbit antigoat antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) (1:500 in PBS/BSA/Tween, 45 min with peroxidase-conjugated streptavidin (Jackson ImmunoResearch Laboratories, Inc.) (1:500 dilution in PBS/BSA/Tween) and 20 min diaminobenzidine used as chromogen (3 mg diaminobenzidine/0.03% H2O2 in PBS). Sections were counterstained with hematoxylin and mounted in Mowiol (Calbiochem-Novabiochem Corp., La Jolla, CA). Sections incubated with goat IgGs at the same concentration as the primary antibody were used as negative controls in all experiments. Slides were viewed using a microscope (mikroskopie und systeme GmbH, model DMRB, Leica Corp., Postfach, Wetzlar, Germany) and photographed with 100 ASA film (Eastman Kodak Co., Rochester, NY).
Dual immunofluorescent staining
Cryostat and paraffin-embedded tissue sections were treated and incubated at room temperature for 120 min with a goat polyclonal antihuman-MIF antibody (R\|[amp ]\|D Systems, Minneapolis, MN) at 0.66 μg/ml in PBS/BSA/Tween as described earlier (see the Immunohistochemistry methodology). After a PBS/0.05% Tween 20 rinse, sections were incubated at room temperature for 90 min with one of the following antibodies: mouse monoclonal antihuman-CD68 (DAKO Corp. Diagnostics Canada Inc., Mississauga, Ontario, Canada) (diluted 1:50 in PBS/BSA/Tween) to detect macrophages; mouse monoclonal antihuman-CD3 (a gift from Dr. W. Mourad, Laval University, Quebéc, Canada) (diluted 1:100 in PBS/BSA/Tween) to detect T lymphocytes; and rabbit polyclonal antihuman von Willebrandt factor (vWF) (Sigma-Aldrich Corp. Canada LTD, Oakville, Ontario, Canada) (diluted 1:200 in PBS/BSA/Tween) to detect endothelial cells. After a subsequent wash in PBS/0.05% Tween 20, tissue sections were incubated simultaneously for 60 min at room temperature in the dark with fluorescein isothiocyanate-conjugated donkey antigoat antibody (Jackson ImmunoResearch Laboratories, Inc.) (diluted 1:50 in PBS/BSA/Tween) and rhodamine-conjugated sheep antimouse antibody (Roche Diagnostics, Laval, Quebéc, Canada) (diluted 1:10 in PBS/BSA/Tween) for tissues marked for MIF and CD68 or CD3 or rhodamine-conjugated mouse antirabbit antibody (Jackson ImmunoResearch Laboratories, Inc.) (diluted 1:50 in PBS/BSA/Tween) for tissues marked for MIF and vWF. After a final wash in PBS/0.05% Tween 20, slides were mounted in Mowiol containing 10% para-phenylenediamine (Sigma-Aldrich Corp. Canada Ltd.), an antifading agent, and observed under the microscope (Leica Corp.) equipped for fluorescence with a 100-W UV lamp. Photomicrographs were taken with 400 ASA film (Kodak).
Western blotting
Protein extraction from endometriotic tissue was performed according to our previously described procedure (15), and total protein concentration was determined using the DC protein assay (Bio-Rad Laboratories Ltd., Mississauga, Ontario, Canada). For Western blot analysis, denatured proteins were separated by SDS-PAGE in 15% acrylamide slab gels and transferred onto 0.45-μm nitrocellulose membranes. Equal protein loading was confirmed by staining the membrane with Ponceau S (2%). Nitrocellulose membranes were then cut into strips and incubated overnight at 4 C with a polyclonal goat antihuman MIF antibody (R\|[amp ]\|D Systems) at 2 μg/ml of blocking solution (0.1 m Tris buffer, 0.9% NaCl/0.05% Tween 20 containing 5% nonfat dry milk [wt/vol]) or with normal goat Ig (R\|[amp ]\|D Systems) at the same concentration. Strips were then washed in TBS-0.1% Tween 20, incubated for 1 h at room temperature with a peroxidase-conjugated rabbit antigoat antibody (Jackson ImmunoResearch Laboratories, Inc.), diluted 1:10,000 in the blocking solution, washed again in TBS-0.1% Tween 20, incubated for 1 min with an enhanced chemiluminescence system using BM chemiluminescence blotting substrate (Roche Diagnostics), and exposed to BioMax film (Kodak) for several time intervals allowing for an optimal detection (all bands visible but not overexposed).
MIF ELISA
MIF concentration in endometriotic tissue protein extract was measured by ELISA according to a previously reported procedure (16). Briefly, this technique uses a capture mouse monoclonal antihuman-MIF antibody (R\|[amp ]\|D Systems), a rabbit polyclonal antihuman-MIF antibody for detection, alkaline phosphatase-conjugated goat antirabbit IgGs (Chemicon International Inc., Temecula, CA) and paranitrophenyl phosphate as substrate (Sigma). The optical density was measured at 405 nm and MIF concentrations were extrapolated from a standard curve using recombinant human MIF.
RT-PCR
Total RNA was extracted from endometriotic tissue with TRIzol reagent (Life Technologies, Inc.) according to the manufacturer’s instructions. The Gene Amp PCR core kit (Perkin-Elmer Corp., Foster City, CA) was used to synthesize cDNA with 500 ng total cellular RNA and 2.5 μmol/liter random hexamers in 20 μL of a RT-PCR solution (50 mmol/liter KCl, 10 mmol/liter Tris-HCl, 5 mmol/liter MgCl2, 1 mmol/liter each of dNTPs, 20 IU RNase inhibitor, and 50 IU reverse transcriptase). The reaction was incubated at 25 C for 15 min, 42 C for 30 min, and 99 C for 5 min. For PCR analysis, we used 10% of the reverse transcription (RT) reaction volume as template in a final volume of 50 μL with 50 pmol of each MIF primer (forward primer, 5′-CTCTCCGAGCTCACCCAGCAG-3′; reverse primer, 5′-CGCGTTCATGTCGTAATAGTT-3′; amplimer size, 255 bp), 0.2 mmol/liter dNTP, and 2.5 IU Taq DNA polymerase (QIAGEN, Santa Clarita, CA). Amplification was performed for 30 cycles composed of 1 min denaturation (at 94 C), 1 min annealing (at 60 C), and 1 min primer extension (at 72 C). These optimal conditions were determined by performing linearity tests with 5%, 10%, and 20% of the RT reaction volume and 25 and 30 amplification cycles. Amplification of genomic DNA with these primers did not produce a signal, suggesting that the amplification sites crossed at least one intron/exon boundary. Of the PCR volume, 20% was fractionated by electrophoresis in a 1.8% agarose gel in the presence of ethidium bromide and transferred to a Qiabrane Nylon Plus membrane (QIAGEN). Then the membrane was dehydrated at 37 C for 30 min, prehybridized with a hybridization buffer composed of 5× SSC (0.15 mol/liter sodium chloride and 0.015 mol/liter sodium citrate), 5× Denhardt’s solution, 50 mmol/liter NaH2PO4, 0.5% SDS, 200 μg/ml salmon sperm DNA, and 50% formamide; hybridized with 32P-labeled MIF cDNA in the same buffer except Denhardt’s solution; washed with SSC solutions containing 0.1% SDS, 11× SSC, 0.2× SSC and 0.1× SSC, respectively, and exposed to x-ray film (Eastman Kodak Co.) for different time intervals allowing for an optimal detection (signals visible but not overexposed). As control, glyceraldehyde phosphate dehydrogenase (GAPDH) amplification was used. For PCR analysis, we used 25% of the RT reaction volume as template in a final volume of 50 μl with 25 pmol of each primer (forward primer, 5′-TGATGACATCAAGAAGGTGGTGAAG-3′; reverse primer, 5′-TCCTTGGAGGCCATGTGGGCCAT-3′; amplimer size, 240 bp), 0.2 mmol/liter dNTPs, and 1 IU Vent DNA polymerase. Amplification was performed for 30 cycles of 30 sec denaturation (at 95 C), 30 sec annealing (at 60 C), and 1 min primer extension (at 72 C). These optimal conditions were determined following linearity tests using 10%, 25%, 50%, and 75% of the RT reaction volume. Specificity of the amplification process was verified by Southern blot hybridization. A negative control (PCR in the absence of cDNA) as well as a positive control (cDNA preparation from the human hystiocytic cell line U937, known to secrete MIF) were included in each series of MIF or GAPDH amplification. The intensity of the hybridization signals was determined by computer-assisted densitometry, using Quantity One Quantitation software (Bio-Rad Laboratories, Inc.). The quantity of the PCR products was determined by densitometric analysis of the intensity of the hybridization signal. The relative level of MIF mRNA normalized to GAPDH mRNA was calculated, and the results were expressed as percent of control (positive control).
Statistical analysis
Multiple comparisons of MIF protein concentrations, as measured by ELISA, and mRNA levels, as determined by semiquantitative RT-PCR, in endometriotic lesions according to the type of the lesion or to endometriosis stage were performed using one-way ANOVA and the Tukey’s test. Comparison of two groups was performed using the unpaired t test. All analyses were performed using statistical analysis system (SAS Institute, Inc., Cary, NC). Differences were considered as statistically significant for a P value less than 0.05.
Results
The first objective of this study was to assess the presence of MIF in endometriotic lesions. Western blot analysis of proteins extracted from endometriotic tissue using a goat polyclonal antihuman-MIF antibody showed a single specific 12.5-kDa band corresponding to the known molecular weight of MIF (Fig. 1). RT-PCR and Southern blot analysis showed specific MIF transcripts, thereby confirming MIF expression in endometriotic tissue at the level of mRNA (Fig. 2).
Representative Western blot analysis of MIF expression in endometriotic tissue. Total proteins were subjected to SDS-PAGE analysis and Western blotting using an affinity purified polyclonal goat anti-MIF antibody (lanes 1–3) or an equivalent concentration of normal goat Ig instead of the primary (lanes 4–6). Lanes 1 and 4, 10 μg total proteins; lanes 2 and 5, 20 μg total proteins; lanes 3 and 6, 40 μg total proteins. The detected band has an estimated apparent molecular weight of approximately 12.5 kDa.
Representative after RT-PCR and Southern blot analysis of MIF transcripts. Total RNA obtained from endometriotic tissue was reverse transcribed, amplified with MIF (upper lanes) or GAPDH (lower lanes) primers, and hybridized with 32P-labeled corresponding probes, as described in Materials and Methods. Lane 1, positive control (cDNA preparation from the human hystiocytic cell line U937, known to secrete MIF); lane 2, negative control (PCR in the absence of cDNA); lanes 3–5, linearity test with different RT volumes.
Immunohistochemical analysis of MIF expression showed a specific brownish immunostaining localized to specific compartments of endometriotic tissue. MIF was found to be strongly expressed in glandular epithelial cells and in cells scattered throughout the stroma (Fig. 3A). Incubation of tissue sections with normal goat IgGs used at concentration equivalent to that of the primary goat polyclonal anti-MIF antibody (negative control) did not result in any nonspecific immunostaining (Fig. 3B).
Immunohistochemical analysis of MIF expression in endometriotic tissue (biopsy of a red papular endometriotic lesion from a 37-yr-old woman with stage I endometriosis). Note the intense brownish immunostaining in the glands and cell aggregates throughout the stroma in the presence of a goat polyclonal anti-MIF antibody (A) and the absence of such staining in the presence of goat IgGs used at concentration equivalent to that of the primary antibody (B) (negative control). Scale bar, 20 μm.
To identify cells expressing MIF in the stroma, dual immunofluorescence analysis was performed using antibodies specific to MIF and to CD3, CD68, and vWF. Representative photomicrographs exhibited in Fig. 4 show a marked expression of MIF in CD3-positive T lymphocytes, CD68-positive macrophages, and vWF-positive endothelial cells.
Dual-immunofluorescent staining of MIF (A–C) and CD3 (D), CD68 (E) or vWF (F) in endometriotic tissue. Tissue sections were incubated with goat polyclonal anti-MIF antibody and with mouse monoclonal anti-CD3, mouse monoclonal anti-CD68, or rabbit polyclonal anti-vWF antibody. Sections were then incubated simultaneously with rhodamine-conjugated sheep antimouse antibody and fluorescein isothiocyanate-conjugated donkey antigoat antibody to detect coexpression of MIF with CD3 or CD68 or with rhodamine-conjugated mouse antirabbit antibody and fluorescein isothiocyanate-conjugated donkey antigoat antibody to detect coexpression of MIF with vWF. Note the expression of MIF (green) in CD3-, CD68-, and vWF-positive T lymphocytes, macrophages, and endothelial cells, respectively (red). Superposition of fluorescein (green) and rhodamine (red) signals clearly shows coexpression (yellow signal) of MIF with CD3 (G = A + D), CD68 (H = B + E), and vWF (I = C + F). Scale bars, 20 μm.
To quantify MIF expression in endometriotic tissue and examine whether MIF expression correlates with the type of endometriotic lesion and endometriosis stage, we measured MIF concentrations in total protein extracts by ELISA and determined the levels of mRNA in the same tissues using a semiquantitative RT-PCR analysis. As shown in Fig. 5A, the highest concentrations of MIF protein were found in flame-like red endometriotic lesions, compared with typical black bluish (P < 0.01) or with white lesions (P < 0.01). Otherwise, no significant difference between typical and white endometriotic lesions was found. Furthermore, MIF concentrations appeared to be significantly higher in lesions from endometriosis stage I, compared with those from endometriosis stage II (P < 0.05), whereas no significant difference between stages I and III-IV or II and III-IV was noted. However, only three biopsies from endometriosis stages III-IV were included in this assay, which may have limited the power of statistical analyses including these stages (Fig. 5B). Analysis of MIF protein expression according to the phase of the menstrual cycle showed no statistically significant difference between lesions from the proliferative and the secretory phases. Semiquantitative RT-PCR analysis of MIF mRNA levels in the same endometriotic tissues showed a pattern of expression comparable with that of the protein, but the difference in MIF mRNA levels was found to be significant only between red and white endometriotic lesions (Fig. 6A). On the other hand, no significant difference in MIF mRNA levels according to endometriosis stage (6B) or to the menstrual cycle phase was noted.
Graphical illustration of MIF concentrations (ng/mg of total proteins) as measured by ELISA in endometriotic tissue. A, Endometriotic biopsies were classified according to their appearance at laparoscopy (red, n = 11; typical, n = 6; white, n = 7) (A) or to endometriosis stage (stage I, n = 12; stage II, n = 9; stage III-IV, n = 3) (B). The box-and-whisker plot was used to illustrate the distribution of MIF concentrations. The box delimits values falling between the 25th and the 75th percentiles and the horizontal line within the box refers to the median scores. *, Significant difference between endometriosis stages I and II using the unpaired t test (P < 0.05), **, Significant difference with the red lesions (P < 0.01).
Semiquantitative RT-PCR analysis of MIF mRNA in endometriotic tissue. The quantity of the PCR products was determined by densitometric analysis of the intensity of the hybridization signal. The relative level of MIF mRNA normalized to GAPDH mRNA was calculated, and the results were expressed as percent of control (positive control). A, Types of endometriotic lesions (red, n = 9; typical, n = 6; white, n = 6); B, stage of endometriosis (stage I, n = 10; stage II, n = 7; stage III-IV, n = 3). The box-and-whisker plot was used to illustrate the distribution of MIF mRNA levels. The box delimits values falling between the 25th and the 75th percentiles and the horizontal line within the box refers to the median scores. *, Significant difference with the red lesions (P < 0.05).
Discussion
Endometriosis might be a multifactorial disease, and its etiology remains hypothetical. The presence of tissue structurally and to some extent functionally comparable with the endometrium found outside the uterine cavity suggests that the condition, at least peritoneal endometriosis, results from implantation of exfoliated endometrium following retrograde menstruation (1). Accordingly, the ectopic growth and development of endometrial tissue that is endowed with the capability of adhering to and implanting into the peritoneal tissue requires the genesis of new microvessels, a process called angiogenesis (2–4). Several growth factors, including acidic and basic fibroblast growth factors, platelet-derived endothelial cell growth factor, and vascular endothelial growth factor (VEGF), have the ability to stimulate vascular endothelial cell growth in vitro and in vivo (17) and were found to be expressed by endometriotic lesions (18–21). Our previous studies showed that endometriotic lesions express IL-8 (22), a chemokine having endothelial cell growth promoting activity in vivo (23). In an effort to understand what enables endometrial cells to grow ectopically in some women, we have previously assessed the capability of endometriotic cells to release mitogenic activity for endothelial cells and identified MIF as one of the major factors released by these cells in culture and having the ability to stimulate endothelial cell proliferation in vitro (8). In the present study, we found that MIF was effectively expressed by ectopic endometrial tissue, both at the mRNA and protein levels, as assessed by RT-PCR and Western blotting. Furthermore, we showed that MIF was located in the glands as well as in cell aggregates scattered over the stroma. Dual-immunofluorescence analysis identified endothelial cells, macrophages, and T lymphocytes as cells markedly expressing MIF in the stroma. Furthermore, MIF was found to be highly produced in the endometriotic lesions that were presenting noticeable vascularization and leukocyte infiltration.
These findings clearly demonstrate that MIF can be produced locally in endometriotic tissue, and that by different cell types. In view of its capability to stimulate endothelial cell growth shown in our and other previous studies (8, 11–13) and its faculty to activate and to inhibit macrophage migration (24, 25), it could be suggested that MIF locally produced within endometriotic implants may contribute to their growth and development either by stimulating endothelial cell proliferation or by retaining macrophages that have been shown to secrete VEGF, a potent angiogenic factor (18), and numerous growth factors for endometrial cells (26, 27). Moreover, the marked expression of MIF by endothelial cells makes plausible that this factor can stimulate endothelial cell proliferation via autocrine and paracrine mechanisms, amplifying thereby the angiogenic process.
It is well documented that MIF is a major multifunctional proinflammatory cytokine. The molecule has been shown to be expressed by inflammatory cells such as macrophages and lymphocytes and to stimulate cytokine production by these cells (9, 10, 28, 29). Interestingly, it has been reported that MIF overrides glucocorticoid inhibition of monocyte secretion of TNF-α, IL-1β, IL-6, and IL-8, which are important inducers of immunological and inflammatory responses (16). MIF-stimulated macrophages are in turn shown to secrete bioactive TNF-α and IL-1β (28). These data strongly suggest that MIF could function locally to amplify the inflammatory reaction observed in and around endometriotic lesions, mainly in those considered to represent the most active forms of the disease (5–7, 30).
Endometriotic lesions can be classified according to their appearance and to their activity. In fact, lesions of the peritoneal lining of the pelvis have various macroscopic appearances, which reflect their age and/or activity. The red subtle lesions are more vascularized and have a higher epithelial mitotic index than the typical, puckered black or bluish peritoneal lesions, whereas the vascularization and the mitotic index are lower in the white lesions. Thus, red lesions are thought to correspond to the first, active stage of early implantation of endometrial glands and stroma and would evolve toward the typical black or bluish lesion after enclosure beneath the peritoneal lining. The white lesions, which are believed to correspond to fibrotic quiescent lesions, show less vascularization and/or mitotic activity and represent less active forms of the disease (5, 6, 31, 32).
Interestingly, our present study revealed that MIF expression is significantly higher in red subtle than in typical black-blue or white endometriotic lesions. The protein expression, as measured by ELISA, according to endometriotic lesion type was in keeping with that of mRNA as assessed by semiquantitative RT-PCR, although in the latter case statistical significance was observed only between red and white endometriotic lesions. This indicates that reduced production of MIF in more advanced endometriotic lesions occurs at the mRNA level and is more likely owing to a reduced mRNA synthesis and/or to mRNA instability than to translational and/or posttranslational events. Our findings are consistent with those reported by Donnez et al. (19), who detected a difference in the expression of VEGF between different types of lesions, with the early, highly vascularized red lesions having a greater expression of VEGF than the later more inactive black powder-burn lesions. Thus, the higher expression of MIF in the red lesions might reflect its role in promoting/maintaining a higher degree of vascular development and give support to our hypothesis that MIF plays an important role in endometriosis-associated active angiogenesis and inflammatory processes and as a marker of active disease.
Although angiogenesis appears as critical for endometriosis and that factors that stimulate angiogenesis, such as VEGF (19), and MIF in the present study were found to be overexpressed in initial endometriotic implants, the mechanisms underlying the up-regulated expression of these angiogenic factors in endometriotic cells remain to be more thoroughly investigated. It is well known that endometriosis is a hormone-dependent disease, frequently associated with an immunoinflammatory process both in eutopic and ectopic locations (33–38). Recent interesting studies clearly demonstrated that E2 up-regulates VEGF gene transcription in endometrial cells (39, 40) and that E2-induced gene transcription is ER dependent and is activated through a variant estrogen responsive elements localized within this upstream region of the VEGF gene promoter (41). It also appeared that IL-1, a major proinflammatory cytokine whose peritoneal levels increase in endometriosis (42, 43), induces an angiogenic phenotype in endometriotic cells (44). Hypoxia, which up-regulates VEGF expression in endometrial cells, is thought to be involved in endometrial angiogenesis and to assist revascularization of desquamated endometrial explants when they attach at ectopic sites (45). These factors may activate VEGF expression in endometriotic implants. However, their role in the up-regulation of MIF expression in endometriotic as well as endometrial cells is unknown and currently investigated in our laboratory. Furthermore, it would be interesting to determine whether common transcription factor(s) can be involved in MIF/VEGF production and activation of angiogenesis.
It is noteworthy that MIF was more markedly expressed in lesions from the initial stage of endometriosis (stage I), compared with the more advanced stages, which makes plausible the involvement of this factor in the initial steps of endometriotic tissue growth and development.
Recent studies identified MIF in normal endometrial tissue, predominantly in the glandular epithelium, and showed no significant differences in MIF levels across the menstrual cycle (46). Our current study with ectopic endometrial tissue showed no significant difference in MIF protein and mRNA expression according to the menstrual cycle phase. However, intense immunostaining for MIF was detected both in the stromal and the epithelial compartments, suggesting different expression of MIF in eutopic and ectopic endometrium. Studies evaluating MIF expression in the eutopic endometrial tissue of women with endometriosis in comparison with endometriotic tissue from the same women or from women having a normal gynecological status are, however, in progress in our laboratory, which may shed more light on the functional role of MIF in endometriosis-associated angiogenesis.
In conclusion, the present study showed MIF expression in various cell types throughout endometriotic tissue and its marked expression in lesions representing the earliest and the most active stages of the disease. This suggests that MIF may represent a key effector cell mediator involved in the pathophysiology of endometriosis through its proinflammatory and angiogenic activities.
Acknowledgements
We thank Drs. Jacques Bergeron, Marc Bureau, Jean-Yves Fontaine, Philippe Laberge, André Lemay, Jacques Mailloux, Michel Marois, and Marc Villeneuve for patient evaluation and providing peritoneal fluid samples; François Bigonnesse, Madeleine Desaulniers, Monique Longpré, and Johanne Pelletier for technical assistance; and Dr. Lucile Turcot-Lemay for statistical analysis.
This work was supported by Grant MOP-37921 (to A.A.) from the Canadian Institutes for Health Research and Picower Institute for Medical Research (to C.N.M.).
Abbreviations:
- GAPDH,
Glyceraldehyde phosphate dehydrogenase;
- MIF,
migration inhibitory factor;
- RT,
reverse transcription;
- VEGF,
vascular endothelial growth factor;
- vWF,
von Willebrandt factor.
