| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
PHYSIOLOGY, ENDOCRINOLOGY, AND REPRODUCTION |
Department of Veterinary Science, Maxwell H. Gluck Equine Research Center, University of Kentucky, Lexington 40546-0099
 |
Abstract |
|---|
 Top Abstract INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: atresia • equine • follicle growth • granulosa cell • matrix metalloproteinase • tissue inhibitor
|
INTRODUCTION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). Greater than 99% of follicles will undergo atresia, in which they regress and are resorbed by the ovary (Smith et al., 2002). A subset of follicles fail to ovulate or undergo atresia, resulting in persistent anovulatory follicles (PAF; McCue and Squires, 2002) and are linked to obesity and insulin resistance (Vick et al., 2006). It is unclear how insulin may act to inhibit ovulation in this subset of follicles; however, one possible mechanism of action is an effect on matrix metalloproteinases (MMP).
Matrix metalloproteinases are enzymes that accelerate the breakdown of connective tissues and extracellular matrices to facilitate tissue remodeling (Nuttall et al., 2004). Gelatinases, MMP-2 and MMP-9, are investigated because of their involvement in follicular development, ovulation, atresia (see review by Curry and Osteen, 2003), and regulation by insulin (Dandona et al., 2003). Both enzymes are selectively regulated and inhibited by tissue inhibitors of metalloproteinases (TIMP; TIMP-2 and TIMP-1, respectively).
Few studies have been conducted on MMP and TIMP activity in equine follicular development, and there are no known studies regarding atresia. There is also no known characterization of insulin within the follicle of the mare despite its demonstrated role in ovarian dynamics (Sessions et al., 2004). Previous studies provide valuable insights regarding general gelatinase activity in the ovary; however, it is unknown whether MMP system alterations are due to changes in gene expression or enzyme activity. This investigation was conducted to characterize the MMP system and concentrations of insulin during follicular development and atresia in the horse for future study on PAF, and to test the hypothesis that MMP and TIMP gene expression and enzyme activity are altered during follicular growth and atresia both in vivo and in vitro.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Care and Maintenance of Mares
Light horse mares (aged 3 to 19 yr, n = 27) of mixed breeds were selected at random for use in this study. The experimental mares were part of a large herd maintained at the research farm of the University of Kentucky, Department of Veterinary Science, near Lexington (38°2' N latitude, 84°44' W longitude). All mares were allowed access to free-choice pasture and hay and had ad libitum access to water.
Experimental Design
Whole blood (5 mL) was collected via venipuncture of a jugular vein into a disposable glass test tube (Fisher Scientific, Pittsburgh, PA) and stored at 4°C overnight and allowed to clot. Samples were centrifuged at 2,500 x g at 4°C for 20 min and stored at –20°C for hormone analysis. Serum samples were collected every Monday, Wednesday, and Friday from mid-April until mid-October to determine progesterone cycle profiles for each mare. Based on this profile, mares were palpated per rectum and examined via ultrasound during the anticipated follicular phase. Follicle diameter was measured via ultrasound and follicles were classified as medium (15 to 29 mm) or large (
30 mm). Follicles from each size classification were randomly assigned to either the in vivo or the in vitro study. Follicles were further categorized as either healthy, preovulatory, or atretic based on the estradiol-to-progesterone ratio (E:P; Gerard et al., 1999) and on ultrasound data. The E:P used in evaluation was calculated using published data on follicular fluid concentrations of steroid hormones in equine follicles as a function of their state of viability (Kenney et al., 1979), and classification limits for both in vivo and in vitro studies are given in the respective sections. Follicles with elevated estradiol and reduced progesterone were classified as healthy; those with reduced estradiol and increased progesterone were classified as atretic; and those with increased estradiol and progesterone indicated a preovulatory follicle. In addition to hormone classification, only the largest follicle of increasing diameter present was classified as healthy or preovulatory.
Cell Processing and Culture
In Vivo Study.
Equine granulosa cells and follicular fluid were collected via ultrasound-guided transvaginal aspiration as described previously (Vanderwall et al., 2006) with an aspiration medium of Dulbecco’s modified Eagle’s medium-F12 containing 2% (vol/vol) fetal bovine serum, 1% penicillin-streptomycin (Invitrogen, Carlsbad, CA), and heparin (20 U/mL, Butler, Dublin, OH). After aspiration, cells were centrifuged at 700 x g at 27°C for 5 min. Follicular fluid was collected and stored at –20°C for later protein activity and hormone analysis. The remaining cell pellet was resuspended and centrifuged on a Percoll gradient at 700 x g at 27°C for 10 min to separate out contaminating blood cells. The layer containing equine granulosa cells was then washed with serum-free medium. Samples were again centrifuged at 700 x g at 27°C for 5 min. The medium was discarded and the cell pellet was resuspended so that an aliquot could be collected and later counted on a cell counter (Vi-CELL XR, Beckman Coulter Inc., Fullerton, CA) for determination of cell number and viability. Cells were again centrifuged at 27°C for 5 min at 700 x g. The medium was discarded and the cell pellet was homogenized in 1 mL of TRIzol reagent (Invitrogen) by using a 1-mL syringe and a 22-gauge needle. The cell lysate was then pipetted into a 2-mL cryovial and snap-frozen in liquid nitrogen for transport to the laboratory. Samples were stored at –80°C until RNA isolation. The final groups were as follows: medium-healthy (n = 8; E:P >25), medium-atretic (n = 6; E:P <7), large-healthy (n = 9; E:P >140), and large-preovulatory (n = 5; E:P <48).
In Vitro Study. Cells were processed as described above, with the following exceptions. First, rather than being snap-frozen in TRIzol reagent, cells were placed in serum-free medium and maintained in an incubator at 37°C until transported to the laboratory to undergo tissue culture. Subsequent to counting, 3.0 x 106 cells were cultured in 6-well plates in 2 mL of Dulbecco’s modified Eagle’s medium-F12 containing 1% AlbuMAX (BSA, Invitrogen), 1% penicillin-streptomycin (Invitrogen), and 500 ng/mL of testosterone (Sigma Chem. Co., St. Louis, MO) as an estradiol precursor. Cells were cultured in suspension for 48 h in an incubator maintained at 38.5°C and 5% CO2. After cell culture, the suspension was centrifuged at 700 x g at 27°C for 7 min. The conditioned media samples were stored at –20°C for later protein activity and hormone analysis. The granulosa cells were homogenized in 1 mL of TRIzol reagent and stored at –80°C until RNA was isolated. The final groups were as follows: medium-healthy (n = 7; E:P >10), medium-atretic (n = 5; E:P <4), large-healthy (n = 12; E:P >75), and large-preovulatory (n = 5; E:P <11). Steroid hormones used to determine E:P were calculated on a per-million-cell basis, hence the difference between in vivo and in vitro studies.
RNA Isolations and Relative Quantification of MMP and TIMP Expression by Real-Time Reverse Transcription-PCR
The RNA was isolated using the PureLink Micro-to-Midi Total RNA Purification System (Invitrogen) according to a modified protocol by the manufacturer. Briefly, an equal volume of 70% ethanol was added to the thawed sample and mixed by repeated pipetting. The sample was then processed according to the instructions of the manufacturer. A DNase I treatment was performed before elution from the column to eliminate any contaminating genomic DNA. The RNA was stored at –80°C until reverse transcribed.
Total RNA (500 ng) was diluted in 39 µL of nuclease-free water and combined with 41 µL of reverse transcription master mix for each reverse transcription reaction. Reactions were incubated at 42°C for 15 min, 95°C for 5 min, and 3°C for 5 min in a thermocycler. Samples of cDNA were then stored at –20°C until further analyzed by real time-PCR.
Gene expression was measured in cDNA samples with an Applied Biosystems 7500 Fast sequence detection system (Applied Biosystems, Foster City, CA). Equine-specific, intron-spanning MMP-2, MMP-9, TIMP-1, TIMP-2, and β-glucuronidase (β-GUS) primer-probe sets were designed for this purpose; β-GUS was used as a housekeeping gene (Breathnach et al., 2006) and its expression did not vary by more than 2 SD from the mean in medium or large follicles. However, because β-GUS differed between in vivo and in vitro samples, analysis and comparisons were made separately. For each primer-probe combination, a real-time assay using genomic DNA and reverse transcription-negative RNA samples was used to ensure that there was no amplification of genomic DNA. The PCR reactions were incubated at 95°C for 1 min, followed by 40 cycles of 95°C for 3 s and 60°C for 30 s. Each reaction contained 5.5 µL of master mix and 4.5 µL of cDNA template. All reactions were performed in duplicate.
Relative Quantification of MMP-2 and MMP-9 Enzyme Activity by Gelatin Zymography
Zymography was performed by making a 10% SDS-polyacrylamide gel containing 10% gelatin. Pools of samples were made for both conditioned media samples and follicular fluid and were used as quality controls. Medium (12.5 µL) and follicular fluid (2 µL) were loaded into lanes with 10.0 µL of nonreducing loading buffer. Conditioned medium (7.5 µL) from an ovarian cancer cell line, SKOV3, was used as a positive control. Electrophoresis was performed at 90 V until the dye reached the bottom, approximately 1 to 2 h. Gels were washed for 20 min 3 times in 2% Triton + Tris buffer. Gels were then incubated on a shaker at 37°C in Tris incubation buffer with 1 mM ZnCl for 2.5 d. After incubation, gels were stained in Coomassie Brilliant Blue stain. Subsequently, gels were placed in a destaining solution (glacial acetic acid:isopropanol:water, 1:3:6, by vol) on a shaker and remained in the destaining solution until clear bands were visible. The MMP-2 and MMP-9 were visible as clear bands (gelatin degradation) in a blue gel (intact gelatin). Gels were analyzed using MetaMorph (Molecular Devices, Downington, PA) and quantified by setting a threshold of detection according to background staining. Quality controls for both pooled serum and follicular fluid were averaged for each experiment. Data are presented as arbitrary units representing the densitometric concentrations after standardization with a positive control and average values of the quality controls.
Hormone Analysis
Estradiol.
Concentrations of estradiol were determined by RIA (Coat-A-Count, Diagnostic Products Corp., Los Angeles, CA) as described previously (Bailey et al., 2002). In this procedure, conditioned media (1:20, 1:50, 1:200, and 1:500, vol/vol, dilutions) or follicular fluid (1:1,000 and 1:2,000, vol/vol, dilutions) samples were diluted with 0 pg/mL of calibrator. Separate quality control samples were used for reference for follicular fluid and media. Aliquots of the diluted sample were added in 100-µL amounts to an antibody-coated test tube. One milliliter of 125I-estradiol trace was added. The samples were vortexed and incubated at room temperature for 3 h. Samples were decanted and the tubes were counted and analyzed on a Cobra II Auto gamma counter (Perkin-Elmer Packard, Waltham, MA). Intra- and interassay CV of pooled samples were 4.4 and 7.8%, respectively, for 3 assays. The limit of detection for the estradiol assay was 15.0 pg/mL.
Progesterone.
Progesterone concentrations were determined by RIA (Coat-A-Count) as described previously (Silvia et al., 1992). Separate quality control samples were used for reference for serum, follicular fluid, and conditioned media. In this procedure, 100 µL of serum, medium (undiluted and 1:2, vol/vol, dilution in charcoal-stripped serum), or follicular fluid (1:10, 1:20, and 1:40, vol/vol, dilutions in charcoal-stripped serum) were aliquoted to an antibody-coated test tube. One milliliter of 125I-progesterone trace was added. The samples were vortexed and allowed to incubate at room temperature for 3 h. Samples were decanted and the tubes were counted and analyzed on a Cobra II Auto gamma counter. Intra- and interassay CV of pooled samples were 4.1 and 13.1%, respectively, for 22 assays. The limit of detection for the progesterone assay was 0.4 ng/mL.
Insulin.
Insulin concentrations were determined by validated RIA (Coat-A-Count) as reported previously (Powell et al., 2002). Three quality control samples with known insulin concentrations were used for reference. In this study, 200-µL conditioned media or follicular fluid samples were aliquoted to an antibody-coated test tube, and 1.0 mL of 125I-insulin trace was added. The samples were vortexed and allowed to incubate at room temperature overnight. Samples were then decanted and the tubes were counted and analyzed on a Cobra II Auto gamma counter. Intra- and inter-assay CV of pooled samples were 4.7 and 5.4%, respectively, and were determined in 2 assays. Detection limits for the insulin assays were approximately 0.45 µIU/mL.
Statistical Analyses
Real time reverse transcription-PCR data were imported into LinRegPCR software (LinRegPCR 7.0, J. M. Ruijter and C. Ramakers, Academic Medical Center, Amsterdam, the Netherlands) and primer-probe efficiencies were calculated for each sample. Efficiencies for each gene within a follicle size group and health status were averaged to determine the primer-probe efficiency value (E) used in calculations. Changes in expression (rER; relative expression ratio) of the gene of interest (GOI) were calculated for relative quantification using the equation (Schefe et al., 2006)
|
|
CT (gene) = CT (gene; sample of interest) – CT (gene; calibrator). Results are arbitrary units expressed as the mean fold change in gene expression compared with the calibrator. The average cycle threshold (CT) for the smallest healthy follicle was used as a calibrator for each in vivo and in vitro sample. All data are presented as means ± SEM and values for gene expression; enzyme activity of MMP-2 and MMP-9 proteins are in arbitrary units. Protein and hormone data from the conditioned media samples of the in vitro studies were reported on a per-million-cell basis to eliminate any differences caused by disparities in the number of cells within each sample. Gene expression for MMP-2, MMP-9, TIMP-1, and TIMP-2; ratios of MMP-2:TIMP-2 and MMP-9:TIMP-1; enzyme activity of MMP-2 and MMP-9 proteins; estradiol; progesterone; E:P ratio; and insulin were log-transformed and analyzed using the MIXED procedure analysis (SAS Institute, Cary, NC), followed by pair-wise differences of least squares means between medium-healthy, medium-atretic, large-healthy, and preovulatory follicles. The fixed effects for these analyses were follicle size, health status, and follicle size x health status interaction; the random effect was individual mares. Data for in vivo and in vitro studies were always analyzed separately and cannot be compared directly with one another.
Pearson and Spearman correlations (SigmaStat, Stystat Software Inc., Chicago, IL) were calculated to determine linear and nonlinear relationships for both in vivo and in vitro samples. Both medium and large samples classified as healthy were analyzed to observe relationships during follicular development. Follicles classified as atretic or preovulatory were analyzed independently. When variables had a significant correlation with both Pearson and Spearman analysis, the most significant was reported first.
|
RESULTS |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
There was no difference between mean follicle size, health status, and percentage of cells that were viable in vivo or in vitro. Mean follicle diameter was the same for both healthy- and atretic-medium follicles and healthy- and preovulatory-large follicles, both in vivo and in vitro (Table 1).
|
In Vivo. The expression of MMP-2 was greater (P < 0.05) in large-preovulatory follicles than in large-healthy follicles (4.28 ± 1.52 vs. 1.27 ± 0.55, respectively; Figure 1A). There was a follicle size x health status interaction (P < 0.05) for MMP-2 expression. Granulosa cells from large-preovulatory follicles had greater (P < 0.05) expression of TIMP-2 than cells from large-healthy and medium-atretic follicles (4.46 ± 1.07, 1.44 ± 0.48, and 0.66 ± 0.20, respectively; Figure 1B). These alterations in MMP-2 and TIMP-2 expression resulted in a greater (P < 0.05) MMP-2:TIMP-2 ratio in cells from medium-atretic follicles than in cells from medium-healthy or large-preovulatory follicles (17.39 ± 11.02, 3.73 ± 2.72, and 1.68 ± 0.88, respectively; Figure 1C).
|
MMP-9 and TIMP-1 Expression
In Vivo. There was no difference in expression of MMP-9 for any follicle size or health status class (Figure 2A). There was a decrease (P < 0.05) in TIMP-1 expression in large-preovulatory follicles compared with medium-healthy follicles (0.26 ± 0.07 vs. 1.96 ± 0.45; Figure 2B). The MMP-9:TIMP-1 ratio was greater (P < 0.05) in large-preovulatory follicles than in large-healthy follicles (13.01 ± 7.75 vs. 1.98 ± 1.41; Figure 2C), and tended to be greater (P = 0.09) in medium-atretic follicles than in medium-healthy follicles (9.61 ± 5.44 vs. 2.38 ± 1.85).
|
MMP-2 and MMP-9 Protein
In Vivo. Activity of MMP-2 was greater (P < 0.05) in the follicular fluid of medium follicles than in the follicular fluid of large follicles when analyzed independently of health status. However, in pair-wise comparisons, there was only a trend (P = 0.06) toward both large-healthy (21.26 ± 4.12) and large-preovulatory follicles (30.01 ± 14.63) with less MMP-2 enzyme activity in the follicular fluid than toward both medium-healthy (46.32 ± 6.08) and medium-atretic follicles (57.75 ± 9.29; Figure 3A). Activity of MMP-9 was similar for the 4 groups (Figure 3B).
|
Concentrations of Insulin
In Vivo. There was no difference in insulin content of follicular fluid among medium-healthy, large-healthy, and large-preovulatory follicles. However, there was insufficient sample or insulin could not be detected in medium-atretic follicles; therefore, no determination could be made regarding insulin concentration in those samples (Figure 4A).
|
Correlation Analysis
In Vivo. In healthy follicles, diameter was highly correlated with MMP-2 (r = –0.68; P < 0.01, Spearman; r = –0.61, P < 0.05, Pearson) and MMP-9 (r = –0.60, P < 0.05, Pearson) enzyme activities in follicular fluid. Follicle diameter was also correlated with estradiol (r = 0.77, P < 0.001, Pearson; r = 0.79, P < 0.01, Spearman), E:P ratio (r = 0.69; P < 0.01, Spearman; r = 0.62, P < 0.05, Pearson), and insulin (r = 59, P < 0.05, Pearson). Other correlations observed in healthy, growing follicles were between progesterone and TIMP-2 (r = 0.82; P < 0.001; Pearson) and estradiol and enzyme activities of both MMP-2 (r = –0.67; P < 0.05; Spearman) and MMP-9 (r = –0.61; P < 0.05; Pearson; r = –0.59, P < 0.05, Spearman). Activity of MMP-2 was also correlated with concentrations of insulin (r = –0.68; P < 0.05; Spearman).
Among large-preovulatory follicles, diameter showed a trend toward a correlation with TIMP-1 (r = 0.95; P = 0.05, Pearson), progesterone (r = 0.85; P = 0.07, Pearson; r = 0.90, P = 0.08, Spearman), and estradiol (r = 0.83; P = 0.08, Pearson; r = 0.90, P = 0.08, Spearman).
Follicular diameter had no correlation with any of the variables in atretic follicles in vivo. However, in atretic follicles, MMP-9 was correlated with estradiol (r = 0.97; P < 0.01, Pearson) and E:P (r = 0.93; P < 0.01, Pearson). The E:P ratio was also correlated with TIMP-1 expression (r = 0.89, P < 0.05, Pearson).
In Vitro. In healthy follicles, diameter was correlated with MMP-2 enzyme activity (r = –0.59; P < 0.05; Spearman), MMP-9 protein activity (r = –0.60, P < 0.01, Pearson; r = –0.56, P < 0.05, Spearman), estradiol (r = 0.79; P < 0.001, Spearman), and E:P ratio (r = 0.72; P < 0.001; Spearman; r = 71, P < 0.01, Pearson). Other correlations observed in healthy, growing follicles were between progesterone and the MMP-2:TIMP-2 ratio (r = 0.68; P < 0.001; Pearson), and between progesterone and both MMP-2 (r = 0.70; P < 0.05; Pearson) and MMP-9 (r = 0.54; P < 0.05; Pearson). There was also a trend toward a correlation between TIMP-1 and concentrations of insulin in follicular fluid (r = 0.42; P = 0.08; Pearson).
In large-preovulatory follicles, there were correlations between follicle diameter and MMP-9 (r = –0.92; P < 0.05, Pearson), the MMP-9:TIMP-1 ratio (r = –0.90; P < 0.05, Pearson), and the E:P ratio (r = –0.87; P = 0.05, Pearson). Insulin was highly correlated with progesterone in conditioned media (r = 0.94; P < 0.05, Pearson) and tended to be correlated with estradiol (r = 0.84, P = 0.07, Pearson).
In atretic follicles, diameter was not correlated with any of the variables in vitro. Expression of MMP-9 was associated with concentrations of estradiol (r = 1.00; P < 0.05; Spearman). Progesterone was related to MMP-9 enzyme activity (r = –0.96, P < 0.05, Pearson).
|
DISCUSSION |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), pigs (Driancourt et al., 1998), rats (Curry et al., 2001), and cows (Imai et al., 2003). Gene expression varies by species regarding which cell types in the ovary express MMP and TIMP. Because MMP function is regulated by TIMP, it is important to consider research regarding TIMP in follicular development. In addition, both TIMP-1 and TIMP-2 are present in the ovary and demonstrate alterations in expression during follicular development and ovulation in many species, including the rat (Simpson et al., 2001), ewe (Dhar et al., 1998), pig (Smith et al., 1994), and marmoset (Duncan et al., 1996).
Binding of TIMP-2 to MMP-2 occurs in a 1:1 covalent complex and is essential to MMP-2 activation at the proper ratio. When TIMP-2 is present in greater concentrations than MMP-2, it inhibits MMP-2 activity. If TIMP-2 is absent, no complex can be formed, thereby also inhibiting activation (Strongin et al., 1995). When MMP is investigated, it is beneficial to examine the MMP-2:TIMP-2 ratio because that is likely more indicative of the ability of MMP-2 to cleave extracellular matrix components than the evaluation of each protein individually. Whereas TIMP-2 is required for activation of MMP-2, TIMP-1 is primarily an inhibitor of MMP-9; TIMP-1 binds quickly after MMP-9 activation by plasmin and MMP-2 to inhibit the active enzyme (Michaluk and Kaczmarek, 2007). Therefore, if the concentration of TIMP-1 is reduced, MMP-9 is more active.
Granulosa cells were of particular interest in this study because they synthesize most of the components of the basal lamina that surrounds the follicle. In the cow, granulosa cells grown in anchorage-independent conditions synthesize a basal lamina that contains collagen type IV and fibronectin, both substrates of MMP-2 and MMP-9 (Rodgers et al., 1996). The current in vitro study enabled us to study the specific role of granulosa cells in regulating the MMP system separately from theca cell and oocyte involvement during follicular maturation and atresia.
This study confirmed that there are alterations in MMP-2, MMP-9, TIMP-1, and TIMP-2 gene expression in granulosa cells during follicular development and atresia in the horse. In granulosa cells, expression of MMP-2 was low and did not differ from medium to large groups in vivo. However, in contrast to studies in other species showing that MMP-2 gene expression increases with follicle development, the present study revealed that in the horse, MMP-2 protein enzyme activity was negatively correlated with follicle diameter. Whether this finding is due to a species difference or is reflective of influences that occur after gene expression, such as rate of protein breakdown, needs to be determined. These correlations with follicle diameter were not maintained in atretic or preovulatory follicles, indicating a separate role for these enzymes during different follicular processes.
There were also unique alterations in preovulatory follicles. The ratio of MMP-2 to TIMP-2 was decreased in preovulatory follicles, whereas the MMP-9:TIMP-1 ratio was substantially increased in vivo. These alterations could be due in part to breakdown at the apex in the follicle wall for ovulation, but may also be necessary for formation of a corpus luteum. During ovulation, granulosa cells transform from primarily estradiol-producing cells to progesterone-producing cells, whereas the entire follicle changes from an entity with a rather fluid-filled antrum to a dense tissue with vascularization and fibroblast and theca cell infiltration (Curry and Osteen, 2003). The finding that in vitro results were in contrast to observations in vivo could actually support this theory in that cells in vitro had 48 h to continue luteinization and no longer required the same MMP expression.
Large-preovulatory follicles had greater MMP-2 expression, but not enzyme activity, compared with large-healthy follicles, similar to findings in rats, which demonstrates an increase in MMP-2 expression in response to LH before ovulation (Curry et al., 2001). Expression of TIMP-2 was also elevated in large-preovulatory follicles, similar to expression of MMP-2 in vivo. In vitro, these findings were contradicted in that TIMP-2 expression was much less in cells from large-preovulatory follicles compared with both medium-atretic and large-healthy follicles. These in vivo alterations in MMP-2 and TIMP-2 resulted in decreased MMP-2:TIMP-2 ratios in cells from medium-healthy follicles and large-preovulatory follicles. In vivo, elevated expression of TIMP-2 relative to MMP-2 in healthy and large-preovulatory follicles may be attributed to the role that TIMP have in follicular maturation beyond that of MMP regulation, such as stimulating progesterone synthesis in vitro during the periovulatory period (Nothnick et al., 1997; Berkholtz et al., 2006) and autocrine or paracrine functions in cellular differentiation (Wick et al., 1994), proliferation (Hayakawa et al., 1992), or inhibition of apoptosis (Guedez et al., 1998).
In contrast to studies in bovids (McCaffery et al., 2000; Imai et al., 2003), pigs (Besnard et al., 1997), and sheep (Huet et al., 1998), MMP-2 expression and MMP-2 enzyme activity were not increased in follicles classified as atretic in vivo or in vitro. However, because the sample size for atretic follicles was small and no large-atretic follicles were collected, it cannot be definitively concluded that MMP-2 is not involved in atresia in the horse. In addition, although MMP-2 itself was not greater, the MMP-2:TIMP-2 ratio was increased in medium-atretic follicles in vivo. This alteration in the ratio could possibly allow for an increase in activation of the MMP-2 present during the process of atresia and a decrease in possible suppression of apoptosis by TIMP-2. This could also be significant with regard to tissue remodeling in atresia, given that the basal lamina becomes richer in laminin as follicles mature (Rodgers et al., 2003). Laminin is a substrate for MMP-2, and this increase in the MMP-2:TIMP-2 ratio may result in an increase in activation of MMP-2, allowing controlled breakdown of laminin in the basal lamina to begin the folding observed during atresia (Irving-Rodgers et al., 2002).
Expression of MMP-9 did not change in granulosa cells as follicles progressed from medium to large, and was not different between healthy and atretic follicles. However, in vivo, TIMP-1 was less in large-preovulatory follicles than in cells from medium-healthy follicles. This resulted in an increase in the MMP-9:TIMP-1 ratio, possibly allowing more activity from MMP-9 within the follicle as it neared ovulation or remodeling for regression during atresia. In vitro, medium-atretic follicles did show a significantly greater ratio, indicating a possible increase in activity of MMP-9 during atresia.
In agreement with earlier studies in horses, this study demonstrated that MMP-2 and MMP-9 enzyme activities were both present in the follicular fluid of horses. Riley et al. (2001) found that the predominant MMP found in the follicular fluid was latent MMP-2. Also present in follicular fluid was latent MMP-9; however, these amounts were small in small follicles, were increased as follicles became larger, and remained high in the larger follicles.
The present study confirmed that MMP-2 enzyme activity was greater than MMP-9 enzyme activity in follicular fluid (data not shown); however, no difference in MMP-9 was observed between groups classified as medium or large. In fact, follicle diameter was negatively correlated with MMP-9 enzyme activity in the follicular fluid of healthy, growing follicles. As follicles became larger, MMP-9 enzyme activity from granulosa cells in vitro decreased as well. However, follicles classified as large-preovulatory had MMP-9 enzyme activity consistent with medium follicles and greater than large-healthy follicles, possibly indicating a role in the remodeling necessary for follicle growth in smaller follicles and for ovulation or corpus luteum formation in large follicles.
These findings are in contrast to the report of Riley et al. (2004), who found MMP-9 to be decreased in follicular fluid and tissue explants approximately 4 h before ovulation. However, because mares in that study had been administered hCG to induce ovulation, the timing of sample collection relative to ovulation was more tightly controlled than in the present study. Typically, in mares, rather than a surge in LH, as is observed in other species, plasma LH secretion begins to increase 2 to 3 d before ovulation, peaks shortly after ovulation, and then returns to baseline values 3 to 5 d later (Evans et al., 1979). In many species, administration of hCG causes an increase in MMP expression and activity in follicles; however, changes are species specific (for a review, see Curry and Osteen, 2003). Administration of hCG in a 1-dose manner to mimic a surge of LH may also affect MMP regulation differently from the gradual rise typically observed in horses.
In the present study, MMP-2 activity was not detectable in conditioned media of granulosa cells from medium follicles regardless of health status, and was detectable only in small amounts in media samples from large follicles. This may indicate that MMP-2 present in the follicular fluid is predominantly secreted from a cell type other than granulosa cells. This was supported by localization using immunostaining in the horse, which revealed MMP-2 present primarily in the stroma; staining was not detected in the granulosa cell layer (Riley et al., 2001). However, the present results indicate that as follicles near ovulation, MMP-2 enzyme production by granulosa cells increases.
Interestingly, to our knowledge, this is the first study indicating that equine granulosa cells may produce insulin. Insulin was not different between groups in vivo; however, because no samples were available for medium-atretic follicles, no inferences could be made regarding them. Correlations based on individual follicles revealed that insulin was positively related to follicle diameter. In vitro, insulin was least in medium-healthy follicles compared with all other groups. Further studies should be conducted to examine the possible relationship between granulosa cells from medium-atretic follicles and increased production consistent with larger follicles.
Insulin is an important factor to consider, given that in many species and tissue types, it has been linked to alterations in expression and activity of MMP-2, MMP-9, TIMP-1, and TIMP-2 (Boden et al., 2008; Catalyurek et al., 2008; Kappert et al., 2008), and in this study, insulin was negatively correlated with MMP-2 protein activity in vivo. The role of insulin in MMP regulation during equine follicular dynamics is important to consider with regard to PAF. Previous studies have shown that mares that are insulin resistant and hyperinsulinemic have lengthened interovulatory intervals (Sessions et al., 2004), and obesity (Vick et al., 2006) is linked to the formation of PAF. The current study demonstrates that insulin is correlated with MMP-9 gene expression and MMP-2 enzyme activity in follicles. Elevations in insulin could cause deregulation of the MMP system, resulting in the disruption of normal follicle wall breakdown for ovulation, resulting in PAF.
It is important to note that cells from preovulatory follicles were likely beginning to undergo luteinization at the time of collection, as a response to LH stimulation, and are therefore probably no longer true granulosa cells but rather granulosa-luteal cells. In vitro, cells had been in culture for 48 h and had that time to continue luteinization and lose aromatase activity (Veldhuis et al., 1983), thus accounting for the discrepancies between in vitro and in vivo samples. Differences in MMP gene expression, enzyme activity, and steroid secretion patterns in vivo and in vitro may be due in part to a lack of gonadotropin stimulation to cells in vitro and insufficient substrate for hormone synthesis. Granulosa cells collected by transvaginal aspiration will not adhere to tissue culture vessels and were therefore cultured in suspension. Because the cells were grown in suspension, tissue remodeling enzymes may have been altered because of a lack of contact with one another and the tissue culture vessel.
Alterations in steroid hormone production in follicles are important to consider with regard to the MMP system and ovulation. It is believed that steroids have an important role in follicular rupture during ovulation. A study with sows found that progesterone induced distensibility of follicles, whereas FSH and estradiol could not. Administration of drugs that blocked progesterone synthesis resulted in inhibition of follicle rupture (Lipner and Greep, 1971). Many studies have indicated that blockage of follicle rupture is reversed by progesterone administration and that blockage of ovulation occurs when progesterone antagonists are present (for reviews, see Tsafriri, 1995; Tsafriri and Reich, 1999). However, Robker et al. (2000a) established that MMP activation and secretion patterns are not different in progesterone receptor knockout mice compared with control mice, indicating some other pathway of activation and expression.
In the present study, MMP-2 and MMP-9 enzyme activities were negatively correlated with estradiol in vivo, and in vitro were highly positively correlated with progesterone. It is unknown why there was a difference between the relationship of MMP and steroids in vivo vs. in vitro, but this could be due in part to the explanations mentioned above. However, the strong correlation in this study is in agreement with previous findings in the horse that MMP-2 and progesterone are closely linked (Desvousges et al., 2002). In the previous study, inhibition of either progesterone or MMP-2 resulted in a decrease in concentration of both MMP-2 and progesterone, respectively. In the current study, in addition to the relationship between progesterone and MMP-9, progesterone was highly correlated with TIMP-2 expression in vivo, and with MMP-2:TIMP-2 ratios in vitro, further supporting the role of TIMP in steroidogenesis that has been demonstrated previously in other species and cell types (Boujrad et al., 1995; Shores and Hunter, 2000).
At ovulation, bovine follicles have a differential loss in basal lamina components. After the rise in LH, collagen type IV is discontinuously distributed in the basal lamina while laminin remains intact (Irving-Rodgers et al., 2006), as opposed to atresia, in which laminin is broken down for follicle wall folding (Irving-Rodgers et al., 2002). If these events become uncoupled, the oocyte remains trapped within an unruptured luteinizing follicle (Robker et al., 2000b), a potential mechanism of PAF formation.
In conclusion, MMP and TIMP, as well as steroid and insulin production, change in granulosa cells during follicular development and atresia in the horse. The relationships between molecules are complex and tightly regulated. Further research is needed to determine more completely the function of insulin and the MMP system during follicular development in the mare and what possible role this may have in the formation of PAF.
|
Footnotes |
|---|
2 The research reported in this article is published in connection with a project of the Kentucky Agricultural Experiment Station (09-14-037).
3 Corresponding author: bfitz{at}uky.edu
Received for publication April 29, 2009. Accepted for publication August 5, 2009.
|
LITERATURE CITED |
|---|
|
TopAbstractINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
N) of the alpha43-subunit of inhibin on the follicular atresia and expression of tissue inhibitor of matrix metalloproteinase (TIMP-1) in ovarian follicles of sheep. J. Reprod. Fertil. 114:147–155.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
