- Email: [email protected]
Glucocorticoid receptor-mediated regulation of MMP9 gene expression in human ovarian surface epithelial cells
REPRODUCTIVE ENDOCRINOLOGY Glucocorticoid receptor-mediated regulation of MMP9 gene expression in human ovarian surface epithelial cells Michael T. Rae, Ph.D., Deborah Price, M.S., Christopher R. Harlow, Ph.D., Hilary O. D. Critchley, M.D., and Stephen G. Hillier, Ph.D. Centre for Reproductive Biology, The Queen’s Medical Research Institute, The University of Edinburgh, Edinburgh, United Kingdom
Objective: To obtain proof-of-concept that locally produced anti-inflammatory steroids suppress ovulation-associated extracellular matrix proteases in human ovarian surface epithelial (OSE) cells. Design: Primary OSE cell cultures treated with interleukin-1a (IL-1a) (500 pg/mL) as proxy for inflammation, with/without anti-inflammatory steroid (cortisol or progesterone [P], 0.01–1.0 mM). Setting: Academic medical center. Patient(s): Sixteen premenopausal women (29–46 years) undergoing surgery for nonmalignant gynecological conditions. Main Outcome Measure(s): Semiquantitative extracellular matrix protease gene expression profiling with verification by real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) and gelatinase zymography. Result(s): Treatment with IL-1a stimulated messenger RNA (mRNA) expression of several ovulation-associated matrix metalloproteinase genes by OSE cell cultures, including gelatinase B (MMP9) but not gelatinase A (MMP2). The IL-1a-stimulated MMP9 mRNA production was suppressed by cortisol but not P. Cortisol but not P also dose-dependently suppressed IL-1a-stimulated MMP9 gelatinase activity and this effect was blocked by the glucocorticoid receptor antagonist RU486. Conclusion(s): In human OSE cells, stimulation of MMP9 gene expression and proteolytic activity by IL-1a is suppressed by anti-inflammatory cortisol through a glucocorticoid receptor-mediated mechanism. Because IL-1a also generates cortisol formation in OSE by stimulating cortisone reductase activity, these results support a role for intracrine cortisol in minimizing proteolytic damage to the OSE at ovulation. (Fertil Steril 2009;92:703–8. 2009 by American Society for Reproductive Medicine.) Key Words: Ovarian surface epithelium, inflammation, ovulation, steroids, extracellular matrix, ovarian cancer
The human ovarian surface epithelium (OSE) contains a local (intracrine) system for generating anti-inflammatory glucocorticoids through increased conversion of cortisone to cortisol by 11b-hydroxysteroid dehydrogenase type 1 (11bHSD1), which can be activated by inflammatory cytokines (1). Based on in vitro evidence that cortisol suppresses inflammation-associated (e.g., cyclooxygenase-2) gene expression by OSE cells we have proposed that cytokine Received May 9, 2008; revised June 16, 2008; accepted June 23, 2008; published online August 22, 2008. Centre for Reproductive Biology, The Queen’s Medical Research Institute, The University of Edinburgh, Edinburgh, United Kingdom. M.T.R. has nothing to disclose. D.P. has nothing to disclose. C.R.H. has nothing to disclose. H.O.D.C. has nothing to disclose. S.G.H. has nothing to disclose. Supported by Medical Research Council Programme Grant G0500047 (S.G.H. and H.O.D.C.) and European Commission EU Contract EUK1-CT-2002-00128 (EURISKED) (S.G.H.). Reprint requests: Stephen G. Hillier, Ph.D., Centre for Reproductive Biology, The Queen’s Medical Research Institute, The University of Edinburgh, 47 Little France Crescent, Edinburgh, EH16 4TJ, United Kingdom (FAX: 44-131242-6441; E-mail: [email protected]).
0015-0282/09/$36.00 doi:10.1016/j.fertnstert.2008.06.040
activation of 11bHSD1 can provide a mechanism for antiinflammatory ‘‘protection’’ of ovary during ovulation (2). Ovulation is a form of natural injury (3–5) in which gonadotropin induces an inflammatory cascade within the ovary that terminates in proteolytic breakdown of the follicle wall and overlying OSE, allowing follicular rupture and oocyte release (6). Proteases involved in ovulation include the matrix metalloproteinases MMP2 (gelatinase A) and MMP9 (gelatinase B) (7), which are also implicated in ovarian tumorigenesis through roles in cell adhesion (8) and angiogenesis (9). Epithelial ovarian cancer begins in the OSE and its incidence is positively linked to ovulation (10, 11), most likely due to genetic damage that accumulates during serial inflammation-associated injury and repair. We hypothesize that ovulation-associated tissue injury due to extracellular proteases is naturally constrained by antiinflammatory regulators that localize and limit ovarian proteolysis. Likely anti-inflammatory agents include intracrine cortisol or paracrine progesterone (P) produced by the periovulatory follicle. Here we assess the inflammation-associated
Fertility and Sterility Vol. 92, No. 2, August 2009 Copyright ª2009 American Society for Reproductive Medicine, Published by Elsevier Inc.
703
proteolytic potential of human OSE cells and evaluate the antiproteolytic actions of cortisol and P on gelatinase activity. Results support a role for intracrine cortisol in the suppression of inflammation-associated proteolytic remodeling of the ovarian surface. MATERIALS AND METHODS Patients Human OSE cells were obtained (with informed consent after local ethical committee approval) from the ovaries of 16 premenopausal women aged 29–46 years who were undergoing surgery for nonmalignant gynecological conditions, as described previously (12). None of these subjects were receiving medication that would have been likely to influence ovarian function. Briefly, cells were collected at laparotomy by gentle scraping of the ovarian surface with a sterile wooden spatula, and rinsing into sterile, warmed human OSE1 culture medium for establishment of cell monolayers (discussed later). Cells were collected as close to the beginning of the surgical procedure as possible/practicable to avoid potential blood cell contamination and prevent inadvertent loss of OSE during surgery. Collections were immediately examined by phase-contrast microscopy to verify that a representative biopsy of the OSE had been recovered. Cell Culture Unless stated otherwise tissue culture reagents were obtained from GIBCO BRL (Life Technologies Ltd., Renfrewshire, United Kingdom) and Sigma Chemical Co. (Poole, Dorset, United Kingdom). The OSE scrapings were cultured in donor calf serum-precoated flasks (75 cm2; Corning Inc., Glass Works, Corning, NY). Culture medium (OSE1) comprised Medium199:MCDB105 (1:1 vol/vol) supplemented with fetal calf serum (15% vol/vol), streptomycin (50 mg/mL), penicillin (50 IU/mL), and L-glutamine (1 mM) (1, 2, 13). Cells were incubated at 37 C in a humidified incubator under an atmosphere of 5% CO2 in air for up to 28 days, with medium renewal every 7 days. An advantage of 28-day propagation in the presence of serum is that any theoretical influence of the stage of the menstrual cycle at which cells were collected is nullified, as evidenced by remarkably uniform responses of individual cell cultures to treatment (2, 13). Cell monolayers were routinely examined by phase-contrast microscopy for contaminating fibroblasts and cell purity was confirmed in selected cases with immunocytochemical staining for cytokeratins 5, 6, 8, and 17 (10, 12) using a monoclonal antihuman cytokeratin antibody (Dako; Ely, Cambridgeshire, United Kingdom). Experimental Treatments Confluent OSE cell monolayers from individual patients were treated with trypsin-ethylenediaminetetraacetic acid (EDTA) in Hank’s balanced salt solution (0.05% wt/vol trypsin, 0.5 mM EDTA; Invitrogen Ltd., Paisley, Renfrewshire, United Kingdom) at 37 C for 5 minutes. Cells were then collected by centrifugation (800 x g, 5 minutes). The cell pellet 704
Rae et al.
Quelling periovulatory ovarian injury
was washed in fresh OSE1 media and resuspended in fresh medium. Cell number and viability (routinely >95%) were determined by trypan blue dye (Sigma) exclusion using a hemocytometer. Cell suspension volume was adjusted to yield 50,000 viable cells per 500 mL of medium. Portions (500 mL) of cell suspension were allocated to 24-well polystyrene culture plates (Corning) and incubated for 24 hours. The medium was then aspirated and replaced with 500 mL of serum-free OSE1 medium (OSE2) containing 0.01% bovine serum albumin (BSA) (Sigma) instead of serum. After 24 hours, experimental treatments were begun and culture continued for an additional 48 hours. Interleukin (IL)-1a (R&D Systems Europe Ltd., Abingdon, Oxon, United Kingdom) was used at a final concentration of 500 pg/mL based on previously published data showing this concentration to be maximally effective in induction of 11bHSD type 1 expression (1). Cortisol in ethanol vehicle was added at final concentrations in medium of 10 nM–1 mM, reflecting those of serum and follicular fluid around the time of ovulation (14). The final concentration of ethanol (mirrored in control cultures) was 0.02% (vol/vol). Identical concentrations of P were used for comparison. In selected experiments a 10-fold molar excess of the glucocorticoid (GR) and P receptor (PR) antagonist RU486 was also included to test steroid specificity. Cultures were then routinely continued for 48 hours, unless for time-dependence studies. Spent media were removed and stored at -20 C. Each set of experiments was performed at least three times (i.e., using primary cell cultures from at least three different patients). RNA Extraction and Quality Analysis RNAwas extracted from human OSE cells using RNeasy minispin columns (Qiagen, Crawley, West Sussex, United Kingdom) as per manufacturer’s protocols. Portions (1 mL) of purified RNA were removed for microfluidic quantification and quality assessment using the Agilent 2100 Bioanalyser system for total RNA in combination with RNA6000 nano chips (Agilent Technologies, Cheshire, United Kingdom). Only RNA displaying intact 18S and 28S ribosomal RNA peaks was reverse-transcribed to complementary DNA (cDNA) for real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis. This quality control step was included for each experimental run to avoid generation of falsenegative results due to RNA degradation and as a quantification method to ensure that comparable amounts of RNA were reverse-transcribed in each real-time reaction. Extracellular Matrix Protease Gene Profile Total RNA from the OSE cell cultures of three women, each treated for 48 hours with and without 500 pg/mL IL-1a (discussed previously), was taken for extracellular matrix (ECM) protease gene profiling. The RT2 Profiler Human Extracellular Matrix and Adhesion Molecules PCR Array (SuperArray Bioscience PAHS-013, Frederick, MD) was used to examine the mRNA levels of 84 genes important for cell–cell and cell– matrix interactions. This array contains ECM proteins Vol. 92, No. 2, August 2009
including basement membrane constituents, collagens, and genes playing a role in ECM structure, including five ‘‘housekeeping genes’’ in 96-well plates. Each first strand cDNA synthesis reaction included 80 ng of total RNA with the appropriate negative controls (no reverse transcription, no template) according to the manufacturer’s instructions. Only data for the ECM protease subset in the array are presented in this article: a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS)1, ADAMTS13, ADAMTS8, MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP10, MMP11, MMP12, MMP13, MMP14, MMP15, MMP16, tissue inhibitor of metalloproteinase (TIMP)1, TIMP2, and TIMP3. Quantitative RT-PCR Analysis Real-time qRT-PCR was performed as previously described (1, 2, 13). Primers and probes for MMP2 and MMP9 were purchased prevalidated from Applied Biosystems (Warrington, Cheshire, United Kingdom). Target mRNA was quantified in relation to the abundance of 18S ribosomal RNA in each sample. Positive controls were human liver total RNA (Ambion; Huntingdon, Cambridgeshire, United Kingdom) and in-house prepared human placental mRNA. Negative controls consisted of real-time negative (RNA template with no reverse transcriptase enzyme) and real-time H2O (water in place of RNA template). A Taqman reaction negative control (water in place of cDNA) was also routinely included. Gelatin Zymography Enzymatic activity associated with MMP2 and MMP9 was assessed by gelatinase zymography. Spent culture medium (500 mL) was lyophilized and reconstituted in 125 mL of 0.01% sodium dodecyl sulfate (SDS) (4x concentration step). Portions (10 mL) of concentrate were electrophoresed under native, nondenaturing conditions on a 7.5% polyacrylamide gel containing 10% gelatin as substrate. Zymographic analysis was then carried out by previously published methods (15). Intensity of digested bands was quantified using Open lab image analysis software (Improvision, Coventry, United kingdom) after densitometric scanning (BioRad Laboratories Ltd., Hemel Hempstead, United Kingdom). Statistical Analyses Treatment differences were analyzed by one-way analysis of variance (ANOVA). Specific treatment differences were identified by Fisher’s least significant difference post-hoc testing and confirmed by paired, two-tailed Student’s t-test (Bonferroni corrected). Contrasts between treatments and untreated controls were considered significant at P<.05. RESULTS Gene-Expression Profiling of Inflammation-Associated ECM Proteases Several ECM protease genes expressed in cultured OSE cells were strongly up-regulated by treatment with IL-1a (Fig. 1), Fertility and Sterility
FIGURE 1 Interleukin-1a-stimulated extracellular matrix protease genes in ovarian surface epithelial cells. The ovarian surface epithelial cell cultures from three individual patients were treated with and without interleukin-1a (500 pg/mL) for 48 hours and total RNA taken for expression profiling by polymerase chain reaction array analysis, as described in the Materials and Methods section. Bars represent mean results (n ¼ 2–3 patients) for each gene expressed as mRNA level in interleukin-1a-treated culture relative to the untreated control. The broken horizontal line denotes arbitrary threefold threshold cutoff stimulation.
Rae. Quelling periovulatory ovarian injury. Fertil Steril 2009.
most notably MMP1 (interstitial collagenase), MMP8 (neutrophil collagenase), MMP9 (gelatinase B), and MMP13 (collagenase 3). MMP9 was selected for follow-up along with MMP2 (gelatinase A) because of their prior association with ovulation (16, 17).
IL-1a Stimulates MMP9 but not MMP2 mRNA Expression and Enzymatic Activity Real-time qRT-PCR analysis confirmed stimulation of MMP9 but not MMP2 mRNA by IL-1a (Fig. 2). The stimulatory effect of IL-1a on MMP9 mRNA was maximum at 24–48 hours (data not shown). IL-1a also significantly increased MMP9 but not MMP2 gelatinase activity (Fig. 3). Glucocorticoid Receptor Mediates Suppression of MMP9 Gene Expression and Enzymatic Activity Presence of cortisol in OSE cell culture medium blocked stimulation of MMP9 mRNA by IL-1a (P<.05) (Fig. 2). Cortisol also dose-dependently suppressed IL-1a-stimulated MMP9 gelatinase activity (P<.05) (Fig. 3). Conversely, P failed significantly to alter either MMP9 mRNA expression or gelatinase activity (data not shown). The inhibitory action of cortisol on IL-1a-stimulated MMP9 gelatinase activity was completely reversed by a 10-fold molar excess of the GR/PR antagonist RU486 (P<.05) (Fig. 4), establishing cortisol action by the GR. MMP2 gelatinase activity was unaffected by any of the treatments applied (Figs. 3 and 4). 705
FIGURE 2
FIGURE 3
Interaction of interleukin (IL)-1a and cortisol (CORT) on mRNA expression of MMP9 (closed bars) and MMP2 (open bars) in human ovarian surface epithelial cells. Ovarian surface epithelial cells from three individuals were cultured for 48 hours in the presence or absence of IL-1a (500 pg/mL) and cortisol (0–1.0 mM). Results from qRT-PCR analysis of MMP9 and MMP2 mRNA are presented as foldincrease relative to the corresponding untreated control. Data (mean SEM, n ¼ 3) are from primary ovarian surface epithelial cell cultures donated by three patients. *P< .05 compared to nonstimulated control culture.
Interaction of interleukin (IL)-1a and cortisol (CORT) on secretion of MMP2 and MMP9 gelatinase activities by ovarian surface epithelial cells. Ovarian surface epithelial cells were cultured for 48 hours in the presence or absence of IL-1a (500 pg/mL) and cortisol (0–1.0 mM) then spent medium was taken for gelatinase zymographic analysis as described in the Materials and Methods section. Top: Representative gel from a single experiment/patient. Bottom: Quantified MMP9 gel scans from three individual experiments/patients (mean SEM, n ¼ 3). *P< .05 compared to untreated control.
Rae. Quelling periovulatory ovarian injury. Fertil Steril 2009. Rae. Quelling periovulatory ovarian injury. Fertil Steril 2009.
DISCUSSION We demonstrate that cortisol suppresses IL-1a-stimulated MMP9 gelatinase activity in human OSE cells in vitro. Because we have previously shown that IL-1 increases the potential for local generation of cortisol through up-regulation of 11bHSD1 in OSE cells (1), these results suggest a novel intracrine mechanism for localizing and limiting proteolytic damage to the ovarian surface during ovulation. Our finding that IL-1a stimulates MMP9 mRNA in primary OSE cell cultures extends an earlier finding that proinflammatory cytokines (IL-1b and tumor necrosis factor-a [TNF-a]) stimulate MMP9 enzyme secretion by an immortalized human OSE cell line (16). The marginal (twofold) stimulation by IL-1a of MMP2 mRNA suggested by gene expression profiling of patient-derived OSE cells (Fig. 1) proved to be nonexistent with rigorous qRT-PCR and gelatinase activity analysis. Interleukin-1a also had no measurable effect on MMP2-associated gelatinase activity in primary OSE cell cultures. Thus our data confirm that MMP2 secretion by OSE cells is ‘‘constitutive’’ (16). This does not, however, rule out a role for MMP2 in ovulation (17). Rather it implies regulation by other means in vivo. Cytokine-induction of MMP9 mRNA and gelatinase activity in primary human OSE cell cultures with suppression by cortisol is an entirely novel finding with potential clinical relevance. Our data are based on independent OSE cell collections from 16 patients. Because of the need to propagate 706
Rae et al.
Quelling periovulatory ovarian injury
sufficient OSE cells for study from each patient during 2–4 weeks, experimental design was inevitably limited. We also acknowledge that the properties of cultured OSE cells might not faithfully reflect those of OSE in situ. However, the model represents an advance over genetically modified cell lines to study MMP secretion (16) and the high patient-to-patient reproducibility we record suggests our results are of physiological relevance. Subject to these caveats we interpret the data as further support for an anti-inflammatory mode of cortisol action associated with its regeneration from cortisone due to IL-1a-stimulated 11bHSD1 enzymatic activity in human OSE (1, 2, 13). Suppression of IL-1a-stimulated MMP9 mRNA and gelatinase activity by cortisol is reversible by RU486, which establishes ligand-activated GR signaling as the principal mode of anti-inflammatory glucocorticoid action in OSE cells. Although RU486 is also a PR antagonist, P itself was ineffective in suppressing IL-1a-induced MMP9 gene expression or enzymatic activity in cultured OSE cells. As the ovulation-associated wave of local cortisol regeneration at the ovarian surface subsides, increasing amounts of P become available for paracrine action secreted by granulose–lutein cells. We have previously demonstrated an anti-inflammatory action of P in suppressing IL-1a-induced cyclooxygenase-2 expression by OSE cells (2). Progesterone, therefore, seems able to suppress an inflammatory gene that ‘‘threatens’’ luteinization (i.e., cyclooxygenase-2 through generation of luteolytic Vol. 92, No. 2, August 2009
FIGURE 4 Glucocorticoid receptor-mediated suppression of interleukin (IL)-1a stimulated secretion of MMP9 gelatinase activity by ovarian surface epithelial cells. Ovarian surface epithelial cells were cultured for 48 hours in the presence or absence of IL-1a (500 pg/ mL) or cortisol (CORT, 1.0 mM) with or without 10 mM glucocorticoid receptor antagonist (RU486). Spent medium was taken for gelatinase zymographic analysis as described in the Materials and Methods section. Cells from three different patients were cultured in three separate experiments. Top: Representative gelatin zymogram. Bottom: quantified MMP9 gel scans from three individual experiments (mean SEM, n ¼ 3). *P< .05 compared to untreated control.
Rae. Quelling periovulatory ovarian injury. Fertil Steril 2009.
prostanoids) while sparing genes such as MMP2 and MMP9 that are necessary for ECM remodeling and corpus luteum (CL) formation. This implies differential action of cortisol and P and suggests handover from the more generalized anti-inflammatory action of cortisol to a more ‘‘ovary friendly’’ action of P consistent with minimization of periovulatory ovarian tissue damage and promotion of subsequent CL formation and function. Our results are consistent with multiple members of the MMP gene family being susceptible to inflammatory regulation in OSE cells (6). A recent review by Curry and Smith (7) documents increased ovarian expression of numerous members of the MMP family during ovulation, including mRNA for Mmp1, Mmp2, Mmp9, Mmp13, Mmp14, and Mmp19 in rodent; MMP1, MMP13, and MMP14 in bovine; and MMP1, MMP2, MMP7, and MMP9 in humans. These earlier measurements were generally on whole ovarian tissues or isolated granulosa cells (GC), whereas our study was on isolated human OSE cells. The OSE gene expression profile we undertook indicated that along with gelatinase B Fertility and Sterility
(MMP9), interstitial collagenase (MMP1), neutrophil collagenase (MMP8), and collagenase 3 (MMP13) are strongly responsive to IL-1a in human OSE cells. However, we only rigorously followed up MMP9. It therefore remains to be confirmed that these other MMP genes are inducible by proinflammatory cytokines in OSE cells and whether their expression is also subject to anti-inflammatory modulation by glucocorticoids (or P). The clinical relevance of our results is potentially threefold. First, it is likely that MMP9 secreted by OSE cells in response to inflammatory cytokine stimulation is involved in the remodeling of basement membrane integrity required for ovulation. Such remodeling may lead to basal lamina dissolution associated with aberrant processes such as earlystage tumorigenesis (16, 18). Second, our demonstration that cortisol suppresses MMP9 expression and activity provides proof-of-concept that intracrine generation of cortisol through cytokine activation of 11bHSD1 is a viable physiological mechanism for localizing and limiting proteolytic damage to the ovarian surface during ovulation. Third, both MMP2 and MMP9 are implicated in the progression of metastatic ovarian cancer through promoting tumor spread and angiogenesis (8, 9). Ovarian cancer has an inflammatory basis (19), therefore anti-inflammatory regulation of MMP9 expression through locally regenerated glucocorticoids has both physiological (ovulation) and pathological (cancer spread) implications. We have noted previously that human ovarian cancer cell lines are generally deficient in 11bHSD1 gene expression relative to normal OSE cells (20). If this translates into deficient glucocorticoid suppression of MMP9 in primary ovarian cancer, this could be a mechanism of promoting metastatic tumor spread in vivo. In conclusion, human OSE cells express multiple ECM proteases up-regulated by the proinflammatory cytokine IL-1a in vitro. Gelatinase A (MMP2) and gelatinase B (MMP9) are both secreted by OSE cells but only MMP9 is consistently upregulated by IL-1a. Cortisol produced in increased amounts at the time of ovulation through stimulation of 11bHSD1 (cortisone reductase) activity in OSE cells potentially suppresses IL-1a-induced MMP9 gene expression and enzymatic activity through a GR-mediated mechanism. Thus OSE cells possess the molecular machinery necessary to minimize proteolytic injury of the ovarian surface during ovulation. Acknowledgments: The authors thank Catherine Murray and Sharon Donaldson for help with tissue collection.
REFERENCES 1. Yong PY, Harlow C, Thong KJ, Hillier SG. Regulation of 11beta-hydroxysteroid dehydrogenase type 1 gene expression in human ovarian surface epithelial cells by interleukin-1. Hum Reprod 202;17:2300–6. 2. Rae MT, Niven D, Critchley HOD, Harlow CR, Hillier SG. Anti-inflammatory steroid action in human ovarian surface epithelial cells. J Clin Endocrinol Metab 2004;89:4538–44. 3. Espey LL. Ovulation as an inflammatory reaction—a hypothesis. Biol Reprod 1980;22:73–106. 4. Espey LL. Current status of the hypothesis that mammalian ovulation is comparable to an inflammatory reaction. Biol Repro 1994;50:233–8.
707
5. Hillier SG, Tetsuka M. An anti-inflammatory role for glucocorticoids in the ovaries? J Reprod Immunology 1988;39:21–7. 6. Murdoch WJ, Gottsch ML. Proteolytic mechanisms in the ovulatory folliculo-luteal transformation. Connect Tissue Res 2003;44:50–7. 7. Curry TE Jr, Smith MF. Impact of extracellular matrix remodeling on ovulation and the folliculo-luteal transition. Semin Reprod Med 2006;24:228–41. 8. Kenny HA, Kaur S, Coussens LM, Lengyel E. The initial steps of ovarian cancer cell metastasis are mediated by MMP-2 cleavage of vitronectin and fibronectin. J Clin Invest 2008;118:1367–79. 9. Huang S, Van Arsdall M, Tedjarati S, McCarty M, Wu W, Langley R, et al. Contributions of stromal metalloproteinase-9 to angiogenesis and growth of human ovarian carcinoma in mice. J Natl Cancer Inst 2002;94:1134–42. 10. Auersperg N, Wong AST, Choi KC, Kang SK, Leung PCK. Ovarian surface epithelium: biology, endocrinology and pathology. Endocrinol Rev 2001;22:255–88. 11. Cannistra SA. Cancer of the ovary. N Engl J Med 2004;351:2519–29. 12. Hillier SG, Anderson RA, Williams AR, Tetsuka M. Expression of oestrogen receptor a and b in cultured human ovarian surface epithelial cells. Mol Hum Reprod 1998;4:811–5. 13. Rae MT, Niven D, Ross A, Forster T, Lathe R, Critchley HOD, et al. Steroid signaling in human ovarian surface epithelial cells: the response to interleukin-1 alpha determined by microarray analysis. J Endocrinol 2004;183:19–28.
708
Rae et al.
Quelling periovulatory ovarian injury
14. Andersen CY. Possible new mechanism of cortisol action in female reproductive organs: physiological implications of the free hormone hypothesis. J Endocrinol 2002;73:211–7. 15. Riley SC, Gibson AH, Leask R, Mauchline DJ, Pedersen HG, Watson ED. Secretion of matrix metalloproteinases 2 and 9 and tissue inhibitor of metalloproteinases into follicular fluid during follicle development in equine ovaries. Reproduction 2001;121:553–60. 16. Yang W-L, Godwin AK, Xu X-X. Tumour necrosis factor-a-induced matrix proteolytic enzyme production and basement membrane remodelling by human ovarian surface epithelial cells: molecular basis linking ovulation and cancer risk. Cancer Res 2004;64:1534–40. 17. Gottsch ML, Van Kirk EA, Murdoch WJ. Role of matrix metalloproteinase 2 in the ovulatory folliculo-luteal transition of ewes. Reproduction 2002;124:347–52. 18. Yang D, Smith ER, Cohen C, Wu H, Patriotis C, Godwin AK, et al. Molecular events associated with dysplastic morphologic transformation and initiation of ovarian tumorigenicity. Cancer 2002;94: 2380–92. 19. Ness RB, Cottreau C. Possible role of ovarian epithelial inflammation in ovarian cancer. J Nat Cancer Inst 1999;91:1459–67. 20. Gubbay O, Guo W, Rae MT, Niven D, Langdon SP, Hillier SG. Inflammation-associated gene expression is altered between normal human ovarian surface epithelial cells and cell lines derived from ovarian adenocarcinomas. Br J Cancer 2005;92:1927–33.
Vol. 92, No. 2, August 2009
Objective: To obtain proof-of-concept that locally produced anti-inflammatory steroids suppress ovulation-associated extracellular matrix proteases in human ovarian surface epithelial (OSE) cells. Design: Primary OSE cell cultures treated with interleukin-1a (IL-1a) (500 pg/mL) as proxy for inflammation, with/without anti-inflammatory steroid (cortisol or progesterone [P], 0.01–1.0 mM). Setting: Academic medical center. Patient(s): Sixteen premenopausal women (29–46 years) undergoing surgery for nonmalignant gynecological conditions. Main Outcome Measure(s): Semiquantitative extracellular matrix protease gene expression profiling with verification by real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) and gelatinase zymography. Result(s): Treatment with IL-1a stimulated messenger RNA (mRNA) expression of several ovulation-associated matrix metalloproteinase genes by OSE cell cultures, including gelatinase B (MMP9) but not gelatinase A (MMP2). The IL-1a-stimulated MMP9 mRNA production was suppressed by cortisol but not P. Cortisol but not P also dose-dependently suppressed IL-1a-stimulated MMP9 gelatinase activity and this effect was blocked by the glucocorticoid receptor antagonist RU486. Conclusion(s): In human OSE cells, stimulation of MMP9 gene expression and proteolytic activity by IL-1a is suppressed by anti-inflammatory cortisol through a glucocorticoid receptor-mediated mechanism. Because IL-1a also generates cortisol formation in OSE by stimulating cortisone reductase activity, these results support a role for intracrine cortisol in minimizing proteolytic damage to the OSE at ovulation. (Fertil Steril 2009;92:703–8. 2009 by American Society for Reproductive Medicine.) Key Words: Ovarian surface epithelium, inflammation, ovulation, steroids, extracellular matrix, ovarian cancer
The human ovarian surface epithelium (OSE) contains a local (intracrine) system for generating anti-inflammatory glucocorticoids through increased conversion of cortisone to cortisol by 11b-hydroxysteroid dehydrogenase type 1 (11bHSD1), which can be activated by inflammatory cytokines (1). Based on in vitro evidence that cortisol suppresses inflammation-associated (e.g., cyclooxygenase-2) gene expression by OSE cells we have proposed that cytokine Received May 9, 2008; revised June 16, 2008; accepted June 23, 2008; published online August 22, 2008. Centre for Reproductive Biology, The Queen’s Medical Research Institute, The University of Edinburgh, Edinburgh, United Kingdom. M.T.R. has nothing to disclose. D.P. has nothing to disclose. C.R.H. has nothing to disclose. H.O.D.C. has nothing to disclose. S.G.H. has nothing to disclose. Supported by Medical Research Council Programme Grant G0500047 (S.G.H. and H.O.D.C.) and European Commission EU Contract EUK1-CT-2002-00128 (EURISKED) (S.G.H.). Reprint requests: Stephen G. Hillier, Ph.D., Centre for Reproductive Biology, The Queen’s Medical Research Institute, The University of Edinburgh, 47 Little France Crescent, Edinburgh, EH16 4TJ, United Kingdom (FAX: 44-131242-6441; E-mail: [email protected]).
0015-0282/09/$36.00 doi:10.1016/j.fertnstert.2008.06.040
activation of 11bHSD1 can provide a mechanism for antiinflammatory ‘‘protection’’ of ovary during ovulation (2). Ovulation is a form of natural injury (3–5) in which gonadotropin induces an inflammatory cascade within the ovary that terminates in proteolytic breakdown of the follicle wall and overlying OSE, allowing follicular rupture and oocyte release (6). Proteases involved in ovulation include the matrix metalloproteinases MMP2 (gelatinase A) and MMP9 (gelatinase B) (7), which are also implicated in ovarian tumorigenesis through roles in cell adhesion (8) and angiogenesis (9). Epithelial ovarian cancer begins in the OSE and its incidence is positively linked to ovulation (10, 11), most likely due to genetic damage that accumulates during serial inflammation-associated injury and repair. We hypothesize that ovulation-associated tissue injury due to extracellular proteases is naturally constrained by antiinflammatory regulators that localize and limit ovarian proteolysis. Likely anti-inflammatory agents include intracrine cortisol or paracrine progesterone (P) produced by the periovulatory follicle. Here we assess the inflammation-associated
Fertility and Sterility Vol. 92, No. 2, August 2009 Copyright ª2009 American Society for Reproductive Medicine, Published by Elsevier Inc.
703
proteolytic potential of human OSE cells and evaluate the antiproteolytic actions of cortisol and P on gelatinase activity. Results support a role for intracrine cortisol in the suppression of inflammation-associated proteolytic remodeling of the ovarian surface. MATERIALS AND METHODS Patients Human OSE cells were obtained (with informed consent after local ethical committee approval) from the ovaries of 16 premenopausal women aged 29–46 years who were undergoing surgery for nonmalignant gynecological conditions, as described previously (12). None of these subjects were receiving medication that would have been likely to influence ovarian function. Briefly, cells were collected at laparotomy by gentle scraping of the ovarian surface with a sterile wooden spatula, and rinsing into sterile, warmed human OSE1 culture medium for establishment of cell monolayers (discussed later). Cells were collected as close to the beginning of the surgical procedure as possible/practicable to avoid potential blood cell contamination and prevent inadvertent loss of OSE during surgery. Collections were immediately examined by phase-contrast microscopy to verify that a representative biopsy of the OSE had been recovered. Cell Culture Unless stated otherwise tissue culture reagents were obtained from GIBCO BRL (Life Technologies Ltd., Renfrewshire, United Kingdom) and Sigma Chemical Co. (Poole, Dorset, United Kingdom). The OSE scrapings were cultured in donor calf serum-precoated flasks (75 cm2; Corning Inc., Glass Works, Corning, NY). Culture medium (OSE1) comprised Medium199:MCDB105 (1:1 vol/vol) supplemented with fetal calf serum (15% vol/vol), streptomycin (50 mg/mL), penicillin (50 IU/mL), and L-glutamine (1 mM) (1, 2, 13). Cells were incubated at 37 C in a humidified incubator under an atmosphere of 5% CO2 in air for up to 28 days, with medium renewal every 7 days. An advantage of 28-day propagation in the presence of serum is that any theoretical influence of the stage of the menstrual cycle at which cells were collected is nullified, as evidenced by remarkably uniform responses of individual cell cultures to treatment (2, 13). Cell monolayers were routinely examined by phase-contrast microscopy for contaminating fibroblasts and cell purity was confirmed in selected cases with immunocytochemical staining for cytokeratins 5, 6, 8, and 17 (10, 12) using a monoclonal antihuman cytokeratin antibody (Dako; Ely, Cambridgeshire, United Kingdom). Experimental Treatments Confluent OSE cell monolayers from individual patients were treated with trypsin-ethylenediaminetetraacetic acid (EDTA) in Hank’s balanced salt solution (0.05% wt/vol trypsin, 0.5 mM EDTA; Invitrogen Ltd., Paisley, Renfrewshire, United Kingdom) at 37 C for 5 minutes. Cells were then collected by centrifugation (800 x g, 5 minutes). The cell pellet 704
Rae et al.
Quelling periovulatory ovarian injury
was washed in fresh OSE1 media and resuspended in fresh medium. Cell number and viability (routinely >95%) were determined by trypan blue dye (Sigma) exclusion using a hemocytometer. Cell suspension volume was adjusted to yield 50,000 viable cells per 500 mL of medium. Portions (500 mL) of cell suspension were allocated to 24-well polystyrene culture plates (Corning) and incubated for 24 hours. The medium was then aspirated and replaced with 500 mL of serum-free OSE1 medium (OSE2) containing 0.01% bovine serum albumin (BSA) (Sigma) instead of serum. After 24 hours, experimental treatments were begun and culture continued for an additional 48 hours. Interleukin (IL)-1a (R&D Systems Europe Ltd., Abingdon, Oxon, United Kingdom) was used at a final concentration of 500 pg/mL based on previously published data showing this concentration to be maximally effective in induction of 11bHSD type 1 expression (1). Cortisol in ethanol vehicle was added at final concentrations in medium of 10 nM–1 mM, reflecting those of serum and follicular fluid around the time of ovulation (14). The final concentration of ethanol (mirrored in control cultures) was 0.02% (vol/vol). Identical concentrations of P were used for comparison. In selected experiments a 10-fold molar excess of the glucocorticoid (GR) and P receptor (PR) antagonist RU486 was also included to test steroid specificity. Cultures were then routinely continued for 48 hours, unless for time-dependence studies. Spent media were removed and stored at -20 C. Each set of experiments was performed at least three times (i.e., using primary cell cultures from at least three different patients). RNA Extraction and Quality Analysis RNAwas extracted from human OSE cells using RNeasy minispin columns (Qiagen, Crawley, West Sussex, United Kingdom) as per manufacturer’s protocols. Portions (1 mL) of purified RNA were removed for microfluidic quantification and quality assessment using the Agilent 2100 Bioanalyser system for total RNA in combination with RNA6000 nano chips (Agilent Technologies, Cheshire, United Kingdom). Only RNA displaying intact 18S and 28S ribosomal RNA peaks was reverse-transcribed to complementary DNA (cDNA) for real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis. This quality control step was included for each experimental run to avoid generation of falsenegative results due to RNA degradation and as a quantification method to ensure that comparable amounts of RNA were reverse-transcribed in each real-time reaction. Extracellular Matrix Protease Gene Profile Total RNA from the OSE cell cultures of three women, each treated for 48 hours with and without 500 pg/mL IL-1a (discussed previously), was taken for extracellular matrix (ECM) protease gene profiling. The RT2 Profiler Human Extracellular Matrix and Adhesion Molecules PCR Array (SuperArray Bioscience PAHS-013, Frederick, MD) was used to examine the mRNA levels of 84 genes important for cell–cell and cell– matrix interactions. This array contains ECM proteins Vol. 92, No. 2, August 2009
including basement membrane constituents, collagens, and genes playing a role in ECM structure, including five ‘‘housekeeping genes’’ in 96-well plates. Each first strand cDNA synthesis reaction included 80 ng of total RNA with the appropriate negative controls (no reverse transcription, no template) according to the manufacturer’s instructions. Only data for the ECM protease subset in the array are presented in this article: a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS)1, ADAMTS13, ADAMTS8, MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP10, MMP11, MMP12, MMP13, MMP14, MMP15, MMP16, tissue inhibitor of metalloproteinase (TIMP)1, TIMP2, and TIMP3. Quantitative RT-PCR Analysis Real-time qRT-PCR was performed as previously described (1, 2, 13). Primers and probes for MMP2 and MMP9 were purchased prevalidated from Applied Biosystems (Warrington, Cheshire, United Kingdom). Target mRNA was quantified in relation to the abundance of 18S ribosomal RNA in each sample. Positive controls were human liver total RNA (Ambion; Huntingdon, Cambridgeshire, United Kingdom) and in-house prepared human placental mRNA. Negative controls consisted of real-time negative (RNA template with no reverse transcriptase enzyme) and real-time H2O (water in place of RNA template). A Taqman reaction negative control (water in place of cDNA) was also routinely included. Gelatin Zymography Enzymatic activity associated with MMP2 and MMP9 was assessed by gelatinase zymography. Spent culture medium (500 mL) was lyophilized and reconstituted in 125 mL of 0.01% sodium dodecyl sulfate (SDS) (4x concentration step). Portions (10 mL) of concentrate were electrophoresed under native, nondenaturing conditions on a 7.5% polyacrylamide gel containing 10% gelatin as substrate. Zymographic analysis was then carried out by previously published methods (15). Intensity of digested bands was quantified using Open lab image analysis software (Improvision, Coventry, United kingdom) after densitometric scanning (BioRad Laboratories Ltd., Hemel Hempstead, United Kingdom). Statistical Analyses Treatment differences were analyzed by one-way analysis of variance (ANOVA). Specific treatment differences were identified by Fisher’s least significant difference post-hoc testing and confirmed by paired, two-tailed Student’s t-test (Bonferroni corrected). Contrasts between treatments and untreated controls were considered significant at P<.05. RESULTS Gene-Expression Profiling of Inflammation-Associated ECM Proteases Several ECM protease genes expressed in cultured OSE cells were strongly up-regulated by treatment with IL-1a (Fig. 1), Fertility and Sterility
FIGURE 1 Interleukin-1a-stimulated extracellular matrix protease genes in ovarian surface epithelial cells. The ovarian surface epithelial cell cultures from three individual patients were treated with and without interleukin-1a (500 pg/mL) for 48 hours and total RNA taken for expression profiling by polymerase chain reaction array analysis, as described in the Materials and Methods section. Bars represent mean results (n ¼ 2–3 patients) for each gene expressed as mRNA level in interleukin-1a-treated culture relative to the untreated control. The broken horizontal line denotes arbitrary threefold threshold cutoff stimulation.
Rae. Quelling periovulatory ovarian injury. Fertil Steril 2009.
most notably MMP1 (interstitial collagenase), MMP8 (neutrophil collagenase), MMP9 (gelatinase B), and MMP13 (collagenase 3). MMP9 was selected for follow-up along with MMP2 (gelatinase A) because of their prior association with ovulation (16, 17).
IL-1a Stimulates MMP9 but not MMP2 mRNA Expression and Enzymatic Activity Real-time qRT-PCR analysis confirmed stimulation of MMP9 but not MMP2 mRNA by IL-1a (Fig. 2). The stimulatory effect of IL-1a on MMP9 mRNA was maximum at 24–48 hours (data not shown). IL-1a also significantly increased MMP9 but not MMP2 gelatinase activity (Fig. 3). Glucocorticoid Receptor Mediates Suppression of MMP9 Gene Expression and Enzymatic Activity Presence of cortisol in OSE cell culture medium blocked stimulation of MMP9 mRNA by IL-1a (P<.05) (Fig. 2). Cortisol also dose-dependently suppressed IL-1a-stimulated MMP9 gelatinase activity (P<.05) (Fig. 3). Conversely, P failed significantly to alter either MMP9 mRNA expression or gelatinase activity (data not shown). The inhibitory action of cortisol on IL-1a-stimulated MMP9 gelatinase activity was completely reversed by a 10-fold molar excess of the GR/PR antagonist RU486 (P<.05) (Fig. 4), establishing cortisol action by the GR. MMP2 gelatinase activity was unaffected by any of the treatments applied (Figs. 3 and 4). 705
FIGURE 2
FIGURE 3
Interaction of interleukin (IL)-1a and cortisol (CORT) on mRNA expression of MMP9 (closed bars) and MMP2 (open bars) in human ovarian surface epithelial cells. Ovarian surface epithelial cells from three individuals were cultured for 48 hours in the presence or absence of IL-1a (500 pg/mL) and cortisol (0–1.0 mM). Results from qRT-PCR analysis of MMP9 and MMP2 mRNA are presented as foldincrease relative to the corresponding untreated control. Data (mean SEM, n ¼ 3) are from primary ovarian surface epithelial cell cultures donated by three patients. *P< .05 compared to nonstimulated control culture.
Interaction of interleukin (IL)-1a and cortisol (CORT) on secretion of MMP2 and MMP9 gelatinase activities by ovarian surface epithelial cells. Ovarian surface epithelial cells were cultured for 48 hours in the presence or absence of IL-1a (500 pg/mL) and cortisol (0–1.0 mM) then spent medium was taken for gelatinase zymographic analysis as described in the Materials and Methods section. Top: Representative gel from a single experiment/patient. Bottom: Quantified MMP9 gel scans from three individual experiments/patients (mean SEM, n ¼ 3). *P< .05 compared to untreated control.
Rae. Quelling periovulatory ovarian injury. Fertil Steril 2009. Rae. Quelling periovulatory ovarian injury. Fertil Steril 2009.
DISCUSSION We demonstrate that cortisol suppresses IL-1a-stimulated MMP9 gelatinase activity in human OSE cells in vitro. Because we have previously shown that IL-1 increases the potential for local generation of cortisol through up-regulation of 11bHSD1 in OSE cells (1), these results suggest a novel intracrine mechanism for localizing and limiting proteolytic damage to the ovarian surface during ovulation. Our finding that IL-1a stimulates MMP9 mRNA in primary OSE cell cultures extends an earlier finding that proinflammatory cytokines (IL-1b and tumor necrosis factor-a [TNF-a]) stimulate MMP9 enzyme secretion by an immortalized human OSE cell line (16). The marginal (twofold) stimulation by IL-1a of MMP2 mRNA suggested by gene expression profiling of patient-derived OSE cells (Fig. 1) proved to be nonexistent with rigorous qRT-PCR and gelatinase activity analysis. Interleukin-1a also had no measurable effect on MMP2-associated gelatinase activity in primary OSE cell cultures. Thus our data confirm that MMP2 secretion by OSE cells is ‘‘constitutive’’ (16). This does not, however, rule out a role for MMP2 in ovulation (17). Rather it implies regulation by other means in vivo. Cytokine-induction of MMP9 mRNA and gelatinase activity in primary human OSE cell cultures with suppression by cortisol is an entirely novel finding with potential clinical relevance. Our data are based on independent OSE cell collections from 16 patients. Because of the need to propagate 706
Rae et al.
Quelling periovulatory ovarian injury
sufficient OSE cells for study from each patient during 2–4 weeks, experimental design was inevitably limited. We also acknowledge that the properties of cultured OSE cells might not faithfully reflect those of OSE in situ. However, the model represents an advance over genetically modified cell lines to study MMP secretion (16) and the high patient-to-patient reproducibility we record suggests our results are of physiological relevance. Subject to these caveats we interpret the data as further support for an anti-inflammatory mode of cortisol action associated with its regeneration from cortisone due to IL-1a-stimulated 11bHSD1 enzymatic activity in human OSE (1, 2, 13). Suppression of IL-1a-stimulated MMP9 mRNA and gelatinase activity by cortisol is reversible by RU486, which establishes ligand-activated GR signaling as the principal mode of anti-inflammatory glucocorticoid action in OSE cells. Although RU486 is also a PR antagonist, P itself was ineffective in suppressing IL-1a-induced MMP9 gene expression or enzymatic activity in cultured OSE cells. As the ovulation-associated wave of local cortisol regeneration at the ovarian surface subsides, increasing amounts of P become available for paracrine action secreted by granulose–lutein cells. We have previously demonstrated an anti-inflammatory action of P in suppressing IL-1a-induced cyclooxygenase-2 expression by OSE cells (2). Progesterone, therefore, seems able to suppress an inflammatory gene that ‘‘threatens’’ luteinization (i.e., cyclooxygenase-2 through generation of luteolytic Vol. 92, No. 2, August 2009
FIGURE 4 Glucocorticoid receptor-mediated suppression of interleukin (IL)-1a stimulated secretion of MMP9 gelatinase activity by ovarian surface epithelial cells. Ovarian surface epithelial cells were cultured for 48 hours in the presence or absence of IL-1a (500 pg/ mL) or cortisol (CORT, 1.0 mM) with or without 10 mM glucocorticoid receptor antagonist (RU486). Spent medium was taken for gelatinase zymographic analysis as described in the Materials and Methods section. Cells from three different patients were cultured in three separate experiments. Top: Representative gelatin zymogram. Bottom: quantified MMP9 gel scans from three individual experiments (mean SEM, n ¼ 3). *P< .05 compared to untreated control.
Rae. Quelling periovulatory ovarian injury. Fertil Steril 2009.
prostanoids) while sparing genes such as MMP2 and MMP9 that are necessary for ECM remodeling and corpus luteum (CL) formation. This implies differential action of cortisol and P and suggests handover from the more generalized anti-inflammatory action of cortisol to a more ‘‘ovary friendly’’ action of P consistent with minimization of periovulatory ovarian tissue damage and promotion of subsequent CL formation and function. Our results are consistent with multiple members of the MMP gene family being susceptible to inflammatory regulation in OSE cells (6). A recent review by Curry and Smith (7) documents increased ovarian expression of numerous members of the MMP family during ovulation, including mRNA for Mmp1, Mmp2, Mmp9, Mmp13, Mmp14, and Mmp19 in rodent; MMP1, MMP13, and MMP14 in bovine; and MMP1, MMP2, MMP7, and MMP9 in humans. These earlier measurements were generally on whole ovarian tissues or isolated granulosa cells (GC), whereas our study was on isolated human OSE cells. The OSE gene expression profile we undertook indicated that along with gelatinase B Fertility and Sterility
(MMP9), interstitial collagenase (MMP1), neutrophil collagenase (MMP8), and collagenase 3 (MMP13) are strongly responsive to IL-1a in human OSE cells. However, we only rigorously followed up MMP9. It therefore remains to be confirmed that these other MMP genes are inducible by proinflammatory cytokines in OSE cells and whether their expression is also subject to anti-inflammatory modulation by glucocorticoids (or P). The clinical relevance of our results is potentially threefold. First, it is likely that MMP9 secreted by OSE cells in response to inflammatory cytokine stimulation is involved in the remodeling of basement membrane integrity required for ovulation. Such remodeling may lead to basal lamina dissolution associated with aberrant processes such as earlystage tumorigenesis (16, 18). Second, our demonstration that cortisol suppresses MMP9 expression and activity provides proof-of-concept that intracrine generation of cortisol through cytokine activation of 11bHSD1 is a viable physiological mechanism for localizing and limiting proteolytic damage to the ovarian surface during ovulation. Third, both MMP2 and MMP9 are implicated in the progression of metastatic ovarian cancer through promoting tumor spread and angiogenesis (8, 9). Ovarian cancer has an inflammatory basis (19), therefore anti-inflammatory regulation of MMP9 expression through locally regenerated glucocorticoids has both physiological (ovulation) and pathological (cancer spread) implications. We have noted previously that human ovarian cancer cell lines are generally deficient in 11bHSD1 gene expression relative to normal OSE cells (20). If this translates into deficient glucocorticoid suppression of MMP9 in primary ovarian cancer, this could be a mechanism of promoting metastatic tumor spread in vivo. In conclusion, human OSE cells express multiple ECM proteases up-regulated by the proinflammatory cytokine IL-1a in vitro. Gelatinase A (MMP2) and gelatinase B (MMP9) are both secreted by OSE cells but only MMP9 is consistently upregulated by IL-1a. Cortisol produced in increased amounts at the time of ovulation through stimulation of 11bHSD1 (cortisone reductase) activity in OSE cells potentially suppresses IL-1a-induced MMP9 gene expression and enzymatic activity through a GR-mediated mechanism. Thus OSE cells possess the molecular machinery necessary to minimize proteolytic injury of the ovarian surface during ovulation. Acknowledgments: The authors thank Catherine Murray and Sharon Donaldson for help with tissue collection.
REFERENCES 1. Yong PY, Harlow C, Thong KJ, Hillier SG. Regulation of 11beta-hydroxysteroid dehydrogenase type 1 gene expression in human ovarian surface epithelial cells by interleukin-1. Hum Reprod 202;17:2300–6. 2. Rae MT, Niven D, Critchley HOD, Harlow CR, Hillier SG. Anti-inflammatory steroid action in human ovarian surface epithelial cells. J Clin Endocrinol Metab 2004;89:4538–44. 3. Espey LL. Ovulation as an inflammatory reaction—a hypothesis. Biol Reprod 1980;22:73–106. 4. Espey LL. Current status of the hypothesis that mammalian ovulation is comparable to an inflammatory reaction. Biol Repro 1994;50:233–8.
707
5. Hillier SG, Tetsuka M. An anti-inflammatory role for glucocorticoids in the ovaries? J Reprod Immunology 1988;39:21–7. 6. Murdoch WJ, Gottsch ML. Proteolytic mechanisms in the ovulatory folliculo-luteal transformation. Connect Tissue Res 2003;44:50–7. 7. Curry TE Jr, Smith MF. Impact of extracellular matrix remodeling on ovulation and the folliculo-luteal transition. Semin Reprod Med 2006;24:228–41. 8. Kenny HA, Kaur S, Coussens LM, Lengyel E. The initial steps of ovarian cancer cell metastasis are mediated by MMP-2 cleavage of vitronectin and fibronectin. J Clin Invest 2008;118:1367–79. 9. Huang S, Van Arsdall M, Tedjarati S, McCarty M, Wu W, Langley R, et al. Contributions of stromal metalloproteinase-9 to angiogenesis and growth of human ovarian carcinoma in mice. J Natl Cancer Inst 2002;94:1134–42. 10. Auersperg N, Wong AST, Choi KC, Kang SK, Leung PCK. Ovarian surface epithelium: biology, endocrinology and pathology. Endocrinol Rev 2001;22:255–88. 11. Cannistra SA. Cancer of the ovary. N Engl J Med 2004;351:2519–29. 12. Hillier SG, Anderson RA, Williams AR, Tetsuka M. Expression of oestrogen receptor a and b in cultured human ovarian surface epithelial cells. Mol Hum Reprod 1998;4:811–5. 13. Rae MT, Niven D, Ross A, Forster T, Lathe R, Critchley HOD, et al. Steroid signaling in human ovarian surface epithelial cells: the response to interleukin-1 alpha determined by microarray analysis. J Endocrinol 2004;183:19–28.
708
Rae et al.
Quelling periovulatory ovarian injury
14. Andersen CY. Possible new mechanism of cortisol action in female reproductive organs: physiological implications of the free hormone hypothesis. J Endocrinol 2002;73:211–7. 15. Riley SC, Gibson AH, Leask R, Mauchline DJ, Pedersen HG, Watson ED. Secretion of matrix metalloproteinases 2 and 9 and tissue inhibitor of metalloproteinases into follicular fluid during follicle development in equine ovaries. Reproduction 2001;121:553–60. 16. Yang W-L, Godwin AK, Xu X-X. Tumour necrosis factor-a-induced matrix proteolytic enzyme production and basement membrane remodelling by human ovarian surface epithelial cells: molecular basis linking ovulation and cancer risk. Cancer Res 2004;64:1534–40. 17. Gottsch ML, Van Kirk EA, Murdoch WJ. Role of matrix metalloproteinase 2 in the ovulatory folliculo-luteal transition of ewes. Reproduction 2002;124:347–52. 18. Yang D, Smith ER, Cohen C, Wu H, Patriotis C, Godwin AK, et al. Molecular events associated with dysplastic morphologic transformation and initiation of ovarian tumorigenicity. Cancer 2002;94: 2380–92. 19. Ness RB, Cottreau C. Possible role of ovarian epithelial inflammation in ovarian cancer. J Nat Cancer Inst 1999;91:1459–67. 20. Gubbay O, Guo W, Rae MT, Niven D, Langdon SP, Hillier SG. Inflammation-associated gene expression is altered between normal human ovarian surface epithelial cells and cell lines derived from ovarian adenocarcinomas. Br J Cancer 2005;92:1927–33.
Vol. 92, No. 2, August 2009