Differential expression of matrix metalloproteinases and their tissue ...

Molecular Human Reproduction vol.3 no.11 pp. 1005–1014, 1997

Differential expression of matrix metalloproteinases and their tissue inhibitors in leiomyomata: a mechanism for gonadotrophin releasing hormone agonist-induced tumour regression* Qingchuan Dou, Roy W.Tarnuzzer1, R.Stan Williams, Gregory S.Schultz and Nasser Chegini2 Department of Obstetrics and Gynecology, Box 100294, University of Florida, College of Medicine, Gainesville, Florida, 32610, USA 1Present 2To

address: Department of Medicine, Division of Endocrinology, University of Florida

whom correspondence should be addressed

Tissue remodelling involving extracellular matrix (ECM) turnover plays a major role in leiomyoma growth and regression, regulated by the combined action of matrix metalloproteinases (MMPs) and the tissue inhibitors of MMPs (TIMPs). We postulated that leiomyomata express MMP and TIMP mRNA and protein, and their expression is inversely regulated during tumour growth and gonadotrophin releasing hormone agonist (GnRHa)-induced regression. We therefore examined the expression of mRNA and protein for MMPs (interstitial collagenase, MMP-1; gelatinases, MMP-2 and MMP-9; and stromelysin, MMP-3) and TIMPs (TIMP-1 and TIMP-2) in leiomyoma and matched unaffected myometrium from GnRHa (lupron)-treated and untreated patients. Reverse transcription–polymerase chain reaction (RT–PCR) and restriction enzyme analysis revealed that leiomyomata and myometrium expressed MMP-1, -2, -3 and -9, as well as TIMP-1 and -2 mRNA. Quantitative RT–PCR indicated that leiomyomata and myometrium during the secretory phase of the menstrual cycle expressed higher levels of MMP and TIMP mRNA compared to the proliferative phase (P < 0.05), with low to undetectable levels of MMP-1, -2 and -3 mRNA in the tumours. GnRHa therapy induced an overall reduction in MMP and TIMP mRNA expression in both leiomyomata and myometrium, but a significant decrease in TIMP-1, and an increase in MMP mRNA expression compared with untreated tumours (P < 0.05). Immunohistochemically, MMP-1, -2, -3 and -9 and TIMP-1 and -2 proteins were localized in leiomyomata and myometrial smooth muscle cells, arteriole wall and connective tissue fibroblasts, with an overall increase in MMP and a decrease in TIMP staining intensity in GnRHa-treated groups. The results suggest that MMP and TIMP expression in leiomyoma and myometrium are hormonally regulated, and that GnRHa-induced tumour regression is accompanied by an increase in MMP expression with a concomitant decrease in TIMP-1 expression, which may potentially provide an environment favouring ECM degradation. Key words: GnRH agonist/leiomyoma/metalloproteinases/myometrium/TIMPs

Introduction Leiomyomata are benign uterine tumours presumably originating from the conversion of normal myometrial smooth muscle cells, which histologically consist mainly of smooth muscle cells and a network of connective tissue (Sullivan and Guzick, 1996). Similar to myometrium, leiomyomata also contain functional oestrogen and progesterone receptors, which are apparently overexpressed in the tumours (Nardelli et al., 1987; Chrapusta et al., 1990; Branden et al., 1993; Bulum et al., 1994; Rein et al., 1995). Because of the ovarian steroid involvement in leiomyoma growth, medical interventions to reverse their growth are centred around the use of gonadotrophin releasing hormone analogues (GnRHa) (Rein et al., 1995; Lemay et al., 1996) and, recently, antiprogestin therapy (Reinsch et al., 1994; Murphy et al., 1995). It appears, however, that the predominant changes in uterine volume due to GnRHa therapy occur in non-leiomyoma tissue, and co-administration *Presented in part at 43rd Annual Meeting of the Society for Gynecologic Investigation, Philadelphia PA, USA, 1996. © European Society for Human Reproduction and Embryology

of medroxyprogesterone acetate (MPA) reverses the beneficial effect of GnRHa therapy (Carr et al., 1993). Modulation of mitotic activity, cellular hypertrophy, and excess accumulation of extracellular matrix (ECM) are considered key factors in leiomyoma growth (Kawaguchi et al., 1989; Stewart et al., 1994; Rein et al., 1995). However, the individual contribution of these parameters to tumour growth and regression are poorly characterized. In regard to ECM, leiomyomata and myometrium have been reported to express a similar level of fibronectin mRNA. However, the ratio and the level of collagen type I and III mRNA expression has been shown to be higher in leiomyomata than myometrium, but only in tissues from the proliferative phase of the menstrual cycle (Puistola et al., 1990; Stewart et al., 1994). GnRHainduced leiomyoma/uterine tissue regression is accompanied by substantial tissue remodelling, presumably involving ECM turnover (Upadhyaya et al., 1990; Gutmann et al., 1994; Rein et al., 1995; Lemay et al., 1996).The ECM turnover is regulated by the rate of synthesis and deposition of various ECM components (Mosher et al., 1992), and a balance between the 1005

Q.Dou et al.

Figure 1. A representative reverse transcription–polyacrylamide chain reaction (RT–PCR) reaction using total RNA isolated from a leiomyoma from the secretory phase of the menstrual cycle indicating the PCR products and predicted 552 base pair (bp) fragment for matrix metalloproteinase (MMP)-1 (lane A), 618 bp for MMP-2 (lane C), 553 bp for MMP-3 (lane E), 616 bp for MMP-9 (lane G), 502 bp for tissue inhibitor of MMP (TIMP)-1 (lane I) and 510 bp for TIMP-2 (lane K), respectively. Digestion of the PCR products with EcoR V for MMP-1 (lane B: 200, 183 and 168 bp), Fok I for MMP-2 (lane D: 380 and 238 bp), Hind III for MMP-3 (lane F: 398, 165 bp), BamH I for MMP-9 (lane H: 433 and 183 bp the extra band may be due to inadequate digestion), Pst I for TIMP-1 (lane J: 298, 175 and 30 bp), and Kpn I for TIMP-2 (lane L: 272, 168 and 70 bp), respectively resulted in the indicated smaller bp fragments. M: DNA marker. Bands corresponding to PCR products lower than 100 bp are not visible due to their reduced capacity to bind ethidium bromide.

Figure 2. Competitive quantitative RT–PCR analysis of total cellular RNA isolated from myometrium from mid–late proliferative and early–mid secretory phases of the menstrual cycle. The synthetic competitor standard at serial dilutions and total RNA were co-amplified by PCR using primers specific to MMPs and TIMPs for 40 cycles; the products were separated on 2% agarose gels, stained with ethidium bromide and photographed. The upper bands are the PCR products generated from the specific message in total cellular RNA, and lower bands from the standard RNA (shown from right to left at dilutions corresponding to 108 to 103 copies/reaction). The far left lanes are the DNA markers. Abbreviatons: see Figure 1.

action of matrix metalloproteinases (MMPs) which degrade ECM, and their tissue inhibitors (TIMPs) (Matrisian, 1992; Woessner et al., 1994). The MMPs are broadly classified based on their ability to specifically degrade various interstitial matrix and basement membrane components, and are inactivated by TIMPs through binding to the active form of these enzymes in 1:1 ratio (Matrisian, 1992; Overall, 1994; Woessner et al., 1994). 1006

The expression of MMPs and TIMPs has been documented in various normal tissues which undergo tissue remodelling, and their overexpression in certain pathological conditions associated with extensive ECM degradation (Matrisian, 1992; Overall, 1994; Woessner et al., 1994). In female reproductive tract tissues, the expression of MMPs and TIMPs has been demonstrated, and associated with events such as regular menstruation, abnormal uterine bleeding, ovulation, implanta-

MMPs/TIMPs and leiomyoma regression

Figure 3. The ratios of template to sample band intensities calculated by digitally scanning the photographs shown in Figure 2 after the values were normalized for their molecular weight. The ratio of the band intensity within each lane was determined and then plotted against the copy number of added template RNA standard. The number of template RNA molecules/cell was calculated based on the constant that there is ~26 µg mRNA/cell. The number of message copies (copy/cell) determined where the ratio equals 1. The levels of MMP-1, -2, -3 and -9 and TIMP-1 and -2 expression in myometrium from proliferative (Pro) (s) and secretory (Sec) (u) phases of the menstrual cycle are shown with equations of best fit lines; MMP-1 (Sec): Y 5 0.418(X) – 0.365 with r2 5 0.97, MMP-2 (Pro): Y 5 0.560(X) – 0.351 with r2 5 0.984, MMP-2 (Sec): Y 5 0.496(X) – 0.478 with r2 5 0.995, MMP-3 (Pro): Y 5 0.862(X) 1 0.891 with r2 5 1.000, MMP-3 (Sec): Y 5 0.533(X) 1 0.102 with r2 5 0.985, MMP-9 (Pro): Y 5 0.378(X) 1 0.055 with r2 5 0.942, MMP-9 (Sec): Y 5 0.525(X) – 0.550 with r2 5 0.973, TIMP-1 (Pro): Y 5 0.385(X) – 0.810 with r2 5 0.959, TIMP-1 (Sec): Y 5 0.320(X) – 0.937 with r2 5 0.996, TIMP-2 (Pro): Y 5 0.467(X) – 0.670 with r2 5 0.996, TIMP-2 (Sec): Y 5 0.496(X) – 1.013 with r2 5 0.996, β-actin (Pro): Y 5 0.352(X) – 1.194 with r2 5 0.971, β-actin (Sec): Y 5 0.399(X) – 1.297 with r2 5 0.996. Only a very weak band was present on the gel for MMP-1 mRNA from proliferative phase myometrium, so that this appears as a single point on the graph. Abbreviations: see Figure 1.

tion, cervical ripening and parturition (Marbaix et al., 1992; Martelli et al., 1993; Rodgers et al., 1993, 1994; Matrisian, 1994; Osteen et al., 1994, Schatz et al., 1994; Bruner et al., 1995). Our hypothesis is that GnRHa-induced leiomyoma/ uterine regression is accompanied by differential expression of MMPs and TIMPs resulting in excessive ECM degradation. To test this hypothesis, the present study was designed to determine the level of mRNA for MMPs and TIMPs in leiomyomata and matched unaffected myometrium from

GnRHa (leuprolide acetate)-treated and untreated patients using competitive quantitative reverse transcriptase–polymerase chain reaction (RT–PCR). The cellular distribution of MMPs and TIMPs protein in these tissues was also determined using specific antibodies and immunohistochemistry.

Materials and methods All the materials for RT–PCR and immunohistochemistry were purchased from commercial sources as previously described (Chegini

1007

Q.Dou et al.

Figure 4. Bar graph representing the level of MMP-1, -2, -3, -9 and TIMP-1 and -2 mRNA expression in myometrium from proliferative and secretory phases of the menstrual cycle, calculated from the bands intensity, where the ratio of template/sample RNA equals 1. The data represent mean 6 SEM of mRNA expression (copy/cell) in three tissues from each phase of the menstrual cycle. *Significantly different from proliferative phase (P , 0.05). Abbreviations: see Figure 1.

et al., 1994; Dou et al., 1996). Portions of leiomyomata and matched unaffected myometrium from 10 premenopausal women aged 21–47 years were collected immediately following abdominal or vaginal hysterectomies for symptomatic uterine leiomyomata. Based on histological dating of endometrium and the patients’ last menstrual period, three specimens were from mid–late proliferative and seven from early–mid secretory phase of the menstrual cycle. Portions of leiomyoma and matched unaffected myometrium (without any trace of endometrium) were collected from nine patients who had received GnRHa (leuprolide acetate) therapy for symptomatic leiomyomata during the previous 3 months prior to surgery. The tissues were collected at the University of Florida affiliated Shands Hospital with the approval of the Institutional Review Board. The tissues were immediately processed for total cellular RNA isolation and immunohistochemistry as previously described (Chegini et al., 1994; Dou et al., 1996).

Isolation of cellular RNA and quantitative competitive RT– PCR Total cellular RNA was isolated from nine leiomyomata and matched unaffected myometrium from the same subject from GnRHa-treated (n 5 3) and untreated (n 5 6) (three from proliferative and three from secretory phase of the menstrual cycle) patients, and individually subjected to standard RT–PCR performed as previously described (Chegini et al., 1994; Dou et al., 1996). The RNA was amplified without the reverse transcription step to detect the presence of any contaminating genomic DNA. The tubes containing all the PCR components except the RT reaction mixture were amplified as a negative control to check for presence of DNA that may have carried over from a prior reaction. For competitive quantitative RT–PCR, a synthetic multiprimer external RNA standard was constructed as previously described (Tarnuzzer et al., 1996; Dou et al., 1996). The external RNA standard contains the complementary sequences corresponding to the 39, 59 and internal probe primers for MMP-1, -2, -3 and -9, as well as TIMP-1 and -2 (Tarnuzzer et al., 1996). Briefly, cDNA was synthesized in a series of standard reactions, each containing 2 µg of total cellular RNA prepared from each tissue and several dilutions of competitive external RNA standard (13108 to 13103 copies/reaction), 2.5 µM oligo(dT)16, 1.5 mM MgCl2, 200 µM of each of dNTPs, 1 U/µl of

1008

human placental ribonuclease inhibitor, 10 mM Tris–HCl (pH 8.3), 50 mM KCl, and 200 U/µg RNA Moloney Murine Leukemia Virus reverse transcriptase (MMLV–RT) in a final volume of 100 µl. The reactions were incubated at 25°C for 10 min, 37°C for 60 min and 92°C for 5 min, and DNA amplification was carried out as previously described (Dou et al., 1996). The PCR products were separated on 2% agarose gels containing 2 ng/ml of ethidium bromide, photographed with polaroid film and scanned on a Hewlett–Packard Scanjet 3C, and stored on computer as TIFF files. The band intensities were determined using NIH-Image version 1.54 and their intensity values were normalized for their molecular weight (Dou et al., 1996). The ratio of band intensities within each lane were determined and then plotted against the copy number of added template RNA standard. The number of mRNA molecules/cell was calculated based on the constant that there is ~26 pg of mRNA/cell (Dou et al., 1996). The quantity of the target messages was determined where the ratio of template/target band intensity was equal to 1 (Dou et al., 1996; Tarnuzzer et al., 1996; Tang et al., 1997). The data were expressed as mean 6 SEM of the band intensity calculated from three separate experiments. The corresponding points on the curves were analysed by Student t-test, and all points on the curves were analysed by ANOVA. P , 0.05 was considered significant.

Immunohistochemistry For immunohistochemical studies small portions of leiomyoma and matched unaffected myometrium from untreated (n5 10) and GnRHatreated (n 5 9) patients were collected and fixed in Bouin’s solution, processed and embedded in paraffin (Chegini et al., 1994). Tissue sections were cut and immunostained using monoclonal antibodies to MMP-1, -2, -3 and -9, and TIMP-1 and -2 (Oncogene Sciences, Cambridge, MA, USA) at a concentration of 2.5–5 µg of IgG/ml for 2–3 h at room temperature as previously described (Chegini et al., 1994; Dou et al., 1996), and visualized with vectastain ABC Elite Kit (Vector Laboratories, Burlington, CA, USA). Omission of the primary antibodies, or incubation of tissue sections with non-immune mouse IgG instead of primary antibodies at the same concentration during immunostaining, were used as controls.

Results Expression of MMP and TIMP mRNA Initially the expression of MMPs and TIMPs mRNA in leiomyomata and matched unaffected myometrium was examined using standard RT–PCR. Total cellular RNA isolated from these tissues and subjected to RT–PCR indicated that MMP and TIMPs mRNA were detectable in myometrium and leiomyomata (Figure 1). The specificity of the reactions was confirmed by appropriate restriction enzyme digestion (Figure 1), as well as by amplification of RNA without the RT step to detect the presence of any genomic DNA contamination, and reactions containing all the PCR components except the RT reaction mixture to check for the presence of DNA that may have carried over from a prior reaction. When competitive quantitative RT–PCR (QRT–PCR) was used to accurately determine the level of MMPs and TIMPs mRNA expression, it appeared that myometrium expresses a significantly higher level of TIMPs than MMPs, with maximal expression occurring during the early-mid secretory phase of the menstrual cycle (Figures 2–4). Leiomyomata from the proliferative phase expressed very low to undetectable levels of MMPs (results are not shown). During the secretory phase, however, leiomyomata

MMPs/TIMPs and leiomyoma regression

Figure 5. A representative of quantitative RT–PCR of total cellular RNA isolated from a gonadotrophin releasing hormone (GnRH) agonisttreated and untreated leiomyoma (A) and matched unaffected myometrium (B). The synthetic competitor standard at serial dilutions and total RNA were co-amplified by PCR using primers specific to MMPs and TIMPs for 40 cycles. Top bands are PCR products generated from the specific message in the specimen’s total RNA, and lower bands from serial dilutions of the standard shown from right to left at dilutions corresponding to 108 to 103 copies/reaction. The far left lanes are the DNA markers. Abbreviations: see Figure 1.

expressed levels of MMPs and TIMPs similar to that seen in myometrium from the proliferative phase, which also expressed low levels of MMPs (compare Figures 4 and 7), representing a 10–100-fold increase compared with proliferative phase leiomyoma (P , 0.05). The level of MMPs and TIMPs mRNA expression was significantly reduced in leiomyomata and myometrium obtained from GnRHa-treated compared to the untreated group from the secretory (Figures 2 and 7A) and proliferative phase (Figure 7B) of the menstrual cycle (P , 0.05). Despite an overall reduction in MMP and TIMP mRNA expression in GnRHa-treated leiomyomata and myometrium, there was a significantly lower TIMP-1, although higher MMP1, -2, -3 and -9, and minimal alteration in TIMP-2 expression (P , 0.05, Figures 5–7). There was no significant difference in the level of β-actin mRNA expression in myometrium and leiomyomata from proliferative and secretory phase of

menstrual cycle, or GnRHa-treated and untreated groups (Figures 4 and 7).

Immunolocalization of MMPs and TIMPs Immunohistochemical observations indicate that leiomyomata and myometrium from untreated groups contained immunoreactive MMP-1, MMP-2, TIMP-1 and TIMP-2, but low MMP3 and MMP-9 protein, associated with the smooth muscle cells, connective tissue fibroblasts and arteriole endothelial and smooth muscle cells (Figure 8). The immunostaining intensity of MMPs and TIMPs was not considerably different in tissues from proliferative compared to secretory phase of the menstrual cycle. However, there was a noticeable increase in MMP and a decrease in TIMP (particularly TIMP-1) immunostaining intensity in leiomyomata (Figure 8) and myometrium (results not shown) from GnRHa-treated com1009

Q.Dou et al.

Figure 6. The ratios of template to sample band intensities calculated from data shown in Figure 5A after normalization for their molecular weight. The ratio of template to samples was calculated and the log of the ratio was plotted versus the log of the input copy number of the template (108 to 103 molecules) and the number of message copies (copy/cell) determined where the ratio equals 1. The level of MMP and TIMP mRNA expression in leuprolide-treated (Tr) (u) and untreated (s) leiomyomata are shown with equations of best fit lines; MMP-1 (Un-Tr): Y 5 0.788(X) 1 0.657 with r2 5 1.00, MMP-1 (Tr): Y 5 0.719(X) – 0.176 with r2 5 1.00, MMP-2 (Un-Tr): Y 5 0.376(X) 1 0.835 with r2 5 0.855, MMP-2 (Tr): Y 5 0.499(X) 1 0.038 with r2 5 0.986, MMP-3 (Un-Tr): Y 5 0.559(X) 1 0.498 with r2 5 0.990, MMP-3 (Tr): Y 5 0.583(X) – 0.002 with r2 5 0.993, MMP-9 (Un-Tr): Y 5 0.510(X) 1 0.372 with r2 5 0.992, MMP-9 (Tr): Y 5 0.537(X) – 0.380 with r2 5 0.975, TIMP-1 (Un-Tr): Y 5 0.384(X) – 0.610 with r2 5 0.995, TIMP-1 (Tr): Y 5 0.348(X) – 0.282 with r2 5 0.996, TIMP-2 (Un-Tr): Y 5 0.431(X) – 0.356 with r2 5 0.996, TIMP-2 (Tr): Y 5 0.534(X) 1 0.139 with r2 5 0.985, β-actin (UnTr):Y 5 0.445(X) – 1.188 with r2 5 0.967, β-actin (Tr): Y 5 0.552(X) – 1.521 with r2 5 0.997. Abbreviations: see Figure 1.

pared with untreated groups. In controls, deletion of the primary antibodies or their replacement with non-immune IgG (Figure 8) resulted in a considerable reduction in immunostaining intensity.

Discussion Leiomyoma growth is considered to be a combination of mitotic activity, ECM accumulation, and cellular hypertrophy (Kawaguchi et al., 1989; Chrapusta et al., 1990; Rein et al., 1995), although the relative individual contribution of each 1010

parameter is poorly defined. Due to fibrotic nature of the leiomyoma, ECM turnover, which is regulated by a balance between ECM deposition, and differential expression of MMPs and TIMPs, may play a critical role. In the present study we have demonstrated that (i) myometrium expresses mRNA and protein for MMP-1, -2, -3 and -9 as well as TIMP-1 and TIMP2, (ii) the myometrial MMPs and TIMPs mRNA expression, but not their immunoreactive proteins, appear to be cycle dependent, with maximal expression occurring during the secretory phase of the menstrual cycle, and (iii) compared with myometrium, leiomyomata from proliferative phase express

MMPs/TIMPs and leiomyoma regression

Figure 7. Bar graphs showing the levels of MMP-1, -2, -3, -9 and TIMP-1 and -2 mRNA expression in untreated leiomyomata from the secretory phase of the menstual cycle and GnRHa-treated leiomyomata (A), and MMP-1 and TIMP-1 in matched myometrium from proliferative phase (B). For comparison with untreated myometrium from secretory phase refer to Figure 4. The data represent mean 6 SEM of mRNA expression (copy/cell) in three different tissues for each group. *Significantly different from untreated group (P , 0.05).

undetectable to low level of MMPs and lower TIMP mRNA. In addition, during the secretory phase, levels of their expression in leiomyoma were comparable to that seen in myometrium of proliferative phase. Collectively the data suggest that MMPs and TIMPs mRNA and protein are expressed in leiomyomata, and at levels lower than myometrium, with maximal mRNA expression occurring during the progesterone-dominated secretory phase. MMPs, which are classified according to their substrate specificity, degrade collagens I–III (MMP-1; interstitial collagenase), collagens IV, V and fibronectin (MP-2; gelatinase A, or type IV collagenase), collagens III, IV, fibronectin, laminin and proteoglycans (MMP-3; stromelysin 1), and collagens IV, V, and elastin (MMP-9; gelatinase B or type V collagenase), and their proteolytic activities are specifically blocked by TIMPs after binding to the active form of these enzymes (Matrisian, 1992; Overall, 1994; Woessner et al., 1994). The level and content of collagen I and III, but not fibronectin, mRNA expression and protein has been reported to be higher in leiomyomata compared with myometrium from the proliferative, but not the secretory phase, of the menstrual cycle (Puistola et al., 1990; Stewart et al., 1994). This coincides with low levels of MMP-1 expression in leiomyomata, which degrades collagens I and III. Other ECM may also be deposited during the proliferative phase due to the low level of MMP2, -3 and -9 expression, similar to MMP-1. In addition to that of MMP, TIMP mRNA expression also appears to be cycle dependent, maximally occurring during the secretory phase, at levels significantly higher than that of MMPs. Despite the cyclic variation in MMP and TIMP mRNA expression, their immunoreactive protein levels determined immunohistochemically appeared not to be menstrual cycle dependent, and differed between leiomyomata and myometrium. Immunohistochemistry is qualitative and the approach does not allow determination of the MMP proteolytic activity in these tissues. However, considering the variations in mRNA expression, the data suggest the existence of an environment favouring

accumulation of ECM in leiomyomata, particularly during the oestrogen-dominated phase. Further work is needed to determine the enzymatic activity of the MMPs as well as the specific ECM degradation in these tissues. The importance of ovarian steroids in leiomyoma growth is well established, and medical interventions to reverse the tumour growth have centred around the use of GnRHa or, recently, antiprogestin therapy. Our data further indicate that GnRHa therapy, which results in ovarian suppression, causes an overall reduction in MMP and TIMP mRNA and protein expression in leiomyomata and myometrium. Despite the reduction in the level of MMP and TIMP mRNA expression, there was an inverse relationship between their expression in GnRHa-treated compared to untreated leiomyomata and myometrium. TIMP-1 has been shown to be secreted as a complex with MMP-9, and specifically inactivates MMP-1, -2, -3 and MMP-9, while TIMP-2, which binds the active form of these enzymes, also binds the latent form of MMP-2 (Woessner et al., 1994; Overall, 1994). The immunohistochemical approach does not allow differentiation between the active and latent forms of MMPs. GnRHa therapy, which often results in leiomyoma regression by reducing the tumour size from 10 to 50% in diameter, is associated with a lack of extensive tissue breakdown. This suggests that the major portion of MMPs in these tissues are in a latent form. The expression of MMPs and TIMPs has been demonstrated in other reproductive tissues (Sato et al., 1991; Marbaix et al., 1992, 1996; Martelli et al., 1993; Rogers et al., 1993, 1994; Waterhouse et al., 1993; Osteen et al., 1994; Schatz et al., 1994; Bruner et al., 1995; Aston et al., 1996; Hulboy et al., 1997). With the exception of MMP-2, which is constitutively expressed in the endometrium throughout the menstrual cycle, MMP-3, MMP-7 (matrilysin), and MMP-11 (stromelysin-3) are found to be expressed during the proliferative phase and at the onset of menstruation, whereas MMP-1, -9 and -10 are expressed only during menstruation (Rodgers et al., 1993, 1994; Salamonsen and Woolley, 1996; Marbaix et al., 1996; Hulboy et al., 1997). These results, as well as the data obtained from experiments in vitro regarding the expression of MMP1 and stromelysin in human endometrial stromal cells, implicate progesterone as a negative regulator of MMP expression (Marbaix et al., 1992, 1996; Schatz et al., 1994; Bruner et al., 1995; Salamonsen and Woolley, 1996; Hulboy et al., 1997). However, the ovarian steroids appear to influence TIMP expression in a different manner to that of MMPs (Sato et al., 1991; Waterhouse et al., 1993; Salamonsen and Woolley, 1996). If, similarly to its effect on endometrium, progesterone negatively regulates the expression of MMPs in leiomyomata/ myometrium, then one would expect a lack of MMP expression in these tissues. The reason for the differences is unclear. However, unlike myometrium/leiomyomata, the endometrium is a dynamic tissue which undergoes rapid and extensive morphological alterations during the menstrual cycle which requires a differentially regulated and higher MMPs expression. Other differences may be due to the sensitivity of the quantitative RT–PCR technique (estimated to be over 1000-fold higher than Northern blot analysis) used in our study to determine endometrial MMP mRNA expression. This enabled us to 1011

Q.Dou et al.

Figure 8. Immunohistochemical localization of MMP-1, -2, -3, -9 and TIMP-1 and -2 in leiomyoma tissue sections from GnRHa-treated and untreated subjects, associated mainly with the smooth muscle cells (SM) and arteriole wall (arrows). Note an overall higher immunostaining intensity for MMPs and a lower intensity for TIMPs in GnRHa-treated compared to untreated leiomyomata. Small arrow heads point to a group of cells possibly of inflammatory origin with strong immunostaining for MMP-3. In controls, replacement of primary antibodies with non-immune mouse IgG resulted in a considerable reduction in immunostaining intensities. Original magnification 3110.

1012

MMPs/TIMPs and leiomyoma regression

measure mRNA expression at low copy numbers which would not have been possible using Northern blot analysis. In addition to ovarian steroids, factors that are locally expressed by leiomyoma and myometrium may also regulate TIMP and MMP expression. These include growth factors and cytokines such as TGF-β, which play a critical role in tissue remodelling and fibrosis (Chegini, 1997), as well as possible GnRHa direct action which has been shown to inhibit the rate of DNA synthesis and TGF-β1 production by myometrial smooth muscle cells (Chegini et al., 1996). The effect of TGF-βs on ECM turnover and tissue fibrosis is mediated via its ability to differentially regulate the expression of ECM, TIMPs and MMPs (Woessner, 1994; Overall, 1994; Border et al., 1994). TGF-β1 up-regulates the expression of procollagen-I, fibronectin and TIMP-1, but down-regulates MMP-1 mRNA in endometrial stromal, glandular epithelial cells, and myometrial smooth cells (Tang et al., 1996). Furthermore progesterone-induced suppression of MMP expression in endometrial epithelial cells has been reported to be mediated through TGF-β production (Bruner et al., 1995). We have shown that GnRHa therapy results in down-regulation of TGF-β mRNA and protein expression in leiomyomata and myometrium (Chegini et al., 1994; Dou et al., 1996). Alternatively, GnRHa-induced TGF-β suppression in leiomyomata may in turn alter the balance between the rate of ECM deposition and degradation through differential regulation of MMPs and TIMPs, resulting in leiomyoma regression. In conclusion, we have shown that leiomyoma/myometrium express MMP and TIMP mRNA and protein, and the patterns of their expression suggest that they may be hormonally regulated during the menstrual cycle and inversely expressed in patients after GnRHa therapy. The data further suggest that GnRHa-induced leiomyomata/uterine reduction in size may be due in part to a mechanism involving MMPs/TIMPs and excess ECM degradation. However, further experiments are required to establish the exact nature of ECM turnover during growth and GnRHa-induced tumour regression.

References Aston, K.E., Stamouli, A., Thomas, E.J. et al. (1996) Effect of gonadotrophin on cell and matrix retention and expression of metalloproteinases and their inhibitor in cultured human granulosa cells modelling corpus luteum function. Mol. Hum. Reprod., 2, 26–30. Border, W.A. and Noble, N.A. (1994) Transforming growth factor β in tissue fibrosis. New Eng. J. Med., 331, 1286–1292. Branden, D.D., Bethea, C.L., Strawn, E.Y. et al. (1993) Progesterone receptor messenger ribonucleic acid and protein are overexpressed in human uterine leiomyomas. Am. J. Obstet. Gynecol., 169, 78–85. Bruner, K., Rodgers, W.H., Gold, L.I. et al. (1995) Transforming growth factor β mediates the progesterone suppression of an epithelial metalloproteinase by adjacent stroma in the human endometrium. Proc. Natl. Acad. Sci. USA, 95, 7362–7366. Bulum, S.E., Simpson, E.R. and Word, R.A. (1994) Expression of the CYP19 gene and its products aromatase cytochrome P450 in human uterine leiomyoma tissues and cells in culture. J. Clin. Endocrinol. Metab., 78, 736–743. Carr, B.R., Marshburn, P.B., Weatherall, P.T. et al. (1993) An evaluation of the effect of gonadotropin-releasing hormone analogs and medroxyprogesterone acetate on uterine leiomyomata volume by magnetic resonance imaging: a prospective, randomized, double blind, placebo-controlled, crossover trial. J. Clin. Endocrinol. Metab., 76, 1217–1223. Chegini, N. (1997) Implication of growth factors and cytokines network in

uterine leiomyoma. In Hill, J.A. (ed.), Cytokines in Reproduction. Medical Intelligence Series. Landes Bioscience, Georgetown, TX, USA (in press). Chegini, N., Zhao, Y., Williams, R.S. et al. (1994) Human uterine tissue throughout the menstrual cycle expresses TGF-β1, TGF-β2, TGF-β3 and TGF-β type II receptor mRNAs and proteins and contain 125I-TGF-β1 binding sites. Endocrinology, 135, 439–449. Chegini, N., Rong, H., Dou, Q. et al. (1996) Gonadotropin-releasing hormone (GnRH) and GnRH receptor gene expression in human myometrium and leiomyomata and the direct action of GnRH analogs on myometrial smooth muscle cells and interaction with ovarian steroids in vitro. J. Clin. Endocrinol. Metabol., 81, 3215–3221. Chrapusta, S., Sieinski, W., Konopka, B. et al. (1990) Estrogen and progestin receptor levels in uterine leiomyomata: relation to the tumor histology and the phase of menstrual cycle. Eur. J. Gynecol. Oncol., 11, 381–387. Dou, Q., Zhao, Y., Tarnuzzer, R.W. et al. (1996) Suppression of TGF-βs and TGF-β receptors mRNA and protein expression in leiomyomata in women receiving gonadotropin releasing hormone agonist therapy. J. Clin. Endocrinol. Metabol., 81, 3222–3230. Hulboy D.L., Rudolph L.A. and Matrisian, L.M. (1997) Matrix metalloproteinases as mediators of reproductive function. Mol. Hum. Reprod., 3, 27–45. Gutmann, J.N., Thronton, K.L. Diamond, M.P. et al. (1994) Evaluation of leuprolide acetate treatment on histopathology of uterine myomata. Fertil. Steril., 61, 622–626. Kawaguchi, K., Fuji, S., Konishi, I. et al. (1989) Mitotic activity in uterine leiomyomas during the menstrual cycle. Am. J. Obstet. Gynecol., 160, 637–641. Lemay, A. and Maheux, R. (1996) GnRH agonists in the management of uterine leiomyoma. Infertil. Reprod. Med. Clin. North Am., 7, 33–55. Marbaix, E., Donnez, J., Courtoy, P.J. et al. (1992) Progesterone regulates the activity of collagenase and related gelatinases A and B in human endometrial explants. Proc. Natl. Acad. Sci. USA, 89, 11789–11793. Marbaix, E., Kokorine, I., Donnez, J. et al. (1996) Regulation and restricted expression of interstitial collagenase suggest a pivotal role in the initiation of menstruation. Hum. Reprod., 11 (Suppl. 2), 134–143. Martelli, M., Campana, A. and Bishoff, P. (1993) Secretion of matrix metalloproteinases by human endometrial cells in vitro. J. Reprod. Fertil., 98, 67–76. Matrisian, L.M. (1992) The matrix-degrading metalloproteinases. Bioassay, 14, 455–463. Matrisian, L.M. (1994) Matrix metalloproteinases gene expression. Ann. NY Acad. Sci., 732, 42–50. Mosher, D.F., Sottile, J., Wu, C. et al. (1992). Assembly of extracellular matrix. Curr. Opin. Cell Biol., 4, 810–818. Murphy, A.A., Morales, A.J., Kettel, L.M. et al. (1995) Regression of uterine leiomyomata to the antiprogesterone RU486: Dose–response effect. Fertil. Steril., 64, 187–190. Nardelli, G.B., Mega, M., Bertasi, M. et al. (1987) Estradiol and progesterone binding in uterine leiomyomata and pregnant myometrium. Clin. Exp. Obstet. Gynecol., 14, 155–160. Osteen, K.G., Rodgers, W.H., Gaire, M. et al. (1994) Stromal–epithelial interaction mediates steroidal regulation of metalloproteinase expression in human endometrium. Proc. Natl. Acad. Sci. USA, 91, 10129–10133. Overall, C.M. (1994) Regulation of tissue inhibitor of matrix metalloproteinases expression. Ann. NY Acad. Sci., 732, 51–64. Puistola, U., Ristili, L., Ristili, J. et al. (1990) Collage metabolism in gynecologic patients: changes in the concentration of the aminoterminal propeptide of type III procollagen in serum. Am. J. Obstet. Gynecol., 163, 1276–1281. Rein, M.S., Barbieri, R.L. and Friedman, A.J. (1995) Progesterone: a critical role in the pathogenesis of uterine myomas. Am. J. Obstet. Gynecol., 172, 14–18. Reinsch, R.C., Murphy, A.A., Morales. A.J. et al. (1994) The effects of RU486 and leuprolide acetate on uterine artery blood flow in the fibroid uterus: A prospective, randomized study. Am. J. Obstet. Gynecol., 170, 1623–1628. Rodgers, W.H., Osteen, K.G., Matrisian, L.M. et al. (1993) Expression and localization of matrilysin, a matrix metalloproteinase, in human endometrium during the reproductive cycle. Am. J. Obstet. Gynecol., 168, 253–260. Rodgers, W.H., Matrisian, L., Giudice, L.C. et al. (1994) Patterns of matrix metalloproteinase expression in cycling endometrium imply differential functions and regulation by steroid hormones. J. Clin. Invest., 94, 946–953. Salamonsen, L.A. and Wooley, D.E. (1996) Matrix metalloproteinases in normal menstruation. Hum. Reprod., 11 (Suppl. 2), 124–133.

1013

Q.Dou et al. Sato, T., Ito, A., Mori, Y. et al. (1991) Hormonal regulation of collagenolysis in uterine cervical fibroblasts. Biochem. J., 275, 645–650. Schatz, F., Papp, C., Totth-Pal, E. et al. (1994) Ovarian steroid-modulated stromelysin-1 expression in human endometrial stromal and decidual cells. J. Clin. Endocrinol. Metabol., 78, 1467–1472. Stewart, E.A., Friedman, A.J., Peck, K. et al. (1994) Relative overexpression of collagen type I and collagen type III messenger ribonucleic acids by uterine leiomyomas during the proliferative phase of the menstrual cycle. J. Clin. Endocrinol. Metab., 79, 900–906. Sullivan, M.W. and Guzick, D.S. (1996) The natural history of uterine myomas. Infertil. Reprod. Med. Clin. North Am., 7, 1–4. Tang, X-M., Ghahary, A. and Chegini, N. (1996) The interaction between transforming growth factor β and relaxin leads to modulation of matrix metalloproteinases expression in human uterus. J. Soc. Gynecol. Invest., 3, Abstr. #205. Tang, X. M., Dou, Q., Zhao, Y. et al. (1997) The expression of transforming growth factor-βs and TGF-β receptor mRNA and protein and the effect of TGF-βs on human myometrial smooth muscle cells in vitro. Mol. Hum. Reprod., 3, 233–240. Tarnuzzer, R.W., Macauley, S.P, Farmerie, W.G. et al. (1996) Competitive RNA templates for detection and quantification of growth factors, cytokines, extracellular matrix components and matrix metalloproteinases by RT– PCR. BioTechniques, 20, 670–674. Upadhyaya, N.B., Doody, M.C. and Googe, P.B. (1990) Histological changes in leiomyomata treated with leuprolide acetate. Fertil. Steril., 54, 811–814. Waterhouse, P., Denhardt, D.T. and Khokha, R. (1993) Temporal expression of tissue inhibitors of metalloproteinases in mouse reproductive tissues during gestation. Mol. Reprod. Dev., 35, 219–226. Woessner, J.F., Jr. (1994) The family of matrix metalloproteinases. Ann. NY Acad. Sci., 732, 11–21. Received on February 18, 1997; accepted on July 22, 1997

1014