Background
The ovarian dynamics comprehended by different phases of the estrous cycle involves extensive tissue remodeling [
1], and major extracellular matrix (ECM) reorganization is a key part of this process. This tissue remodeling is dependent upon cyclic hormonal variation, proteins and peptide growth factors and is a pre-requisite for expansion of the growing ovulatory follicle, breakdown of follicular walls during ovulation, luteinization of the postovulatory follicles, and regression of the
corpus luteum (CL) [
2,
3]. Matrix metalloproteinases (Mmps), a large family of zinc-dependent proteases, and their also numerous inhibitors (Timps) play a major role in ovarian ECM reorganization [
2‐
5]. Together, Mmps and Timps form a network that regulates tissue remodelling and is modulated not only by gonadotropins and steroid hormones, but, also, by the expression of other factors, such as TNF-α, IGF-1 [
5‐
9], and Bsg, which is an inducer of Mmps during the luteinization process [
10]. A complete map of the components of this network, as well as a precise characterization of their expression pattern, may allow us to better understand this tight regulation involved in ovarian tissue remodelling that occur during the estrous cycle.
It is already known from the literature that Mmp2, Mmp9, Mmp14 (Mt1-Mmp), and Mmp19 expression and activity in granulosa and/or theca cells are extensively linked to the disruption of ovarian ECM in rats, which is rich in collagen, laminin, and fibronectin, leading to follicular growth and release of the oocyte during the ovulatory process [
11]. MMP1, MMP2 e MMP9 have been detected in cultures of luteinized granulosa cells [
12‐
14].
The expression and activity of Mmps have been known to be upregulated by gonadotropins during these phases [
15‐
19]. However, hCG also was shown to reduce the expression of MMP2 and MMP9 [
12,
13]. Following ovulation, the ruptured follicle is transformed into a CL by extensive cellular reorganization, migration, and neovascularization [
20‐
22], which is associated with notable gelatinolytic activity, whereas the regression and absorption of CL is marked by an increase in Mmp13 levels and activity [
23,
24].
Tissue inhibitors of metalloproteinases (Timps) have also been extensively studied in ovaries during the estrous cycle. Timp1 is upregulated by gonadotropins during folliculogenesis and ovulation in rats [
25‐
27], providing proteolytic homeostasis to the ECM, through inhibition of metalloproteinases in the corpus luteum, while Timp3 levels decrease slightly [
28].
Timp1 mRNA levels increase during luteal formation and regression [
24,
29], while
Timp3 levels are elevated during luteal maintenance [
29]. TIMP-2 has a role in the local regulation of MMP1 and MMP2 in the corpus luteum [
30]. Moreover, in addition to these inhibitors, there are also proteins, which have already been described as being responsible for inducing the expression of Mmps and ensuring system homeostasis in the ovary. In this context, it is worth mentioning Basigin (Bsg), a protein associated with Mmps induction and also with tumor progression and endometriosis, being expressed in granulosa and theca cells of pre-ovulatory follicles, and also in CL [
10,
31]. However, the expression pattern of Mmp modulators in ovary dynamics is still largely uncharacterized, and there is evidence that members of the network are still to be identified. In particular, two proteins implied in tumor aggressiveness and invasive potential, through modulation of ECM integrity, namely, Sparc and Reck, may have an important role in ovarian tissue remodeling, since they are expressed in ovarian cells and are involved in the regulation of the Mmp system in other tissues [
32‐
36].
Sparc (secreted protein acid and rich in cysteine), also known as Osteonectin, a 43 K protein, and basement-membrane protein 40 (BM-40), is a calcium-binding matrix cellular protein that plays a role in matrix mineralization [
37], modulates TGF-β [
34], and is expressed in the internal theca and follicular basal lamina of ovine follicles and theca-derived small luteal cells [
35]. Even though SPARC is highly expressed in several tumors, it may be able to inhibit tumorigenesis or tumor progression in human ovarian cancer [
38] and its expression is reduced in ovarian cancer cell lines [
38]. Additionally, Sparc is implicated in indirect Mmp modulation and turnover of many physiological processes, but it has not yet been described in whole ovary dynamics. Despite the fact that Bagavandoss et al. described Sparc in the ovary of the late stages of estrous cycle, its expression in the initial phases (folliculogenesis, ovulation and ovulation-luteinization transition) of this process is still unknown [
32]. On the other hand, RECK (REversion-inducing-Cystein-rich protein with Kazal motifs) is a key regulator of extracellular matrix integrity and angiogenesis that is linked to the cell membrane by a GPI-anchor [
36] and inhibits the activity of MMP2, MMP9, and MMP14 at different steps of their activation cascades in humans [
36,
39].
Reck gene is expressed in several mouse tissues during development, including the uterus and ovaries, but at lower levels in most tissues, when compared to those of Timps 1, 2, 3, and 4 [
33]. Reck expression is inhibited by estrogen within the mouse uterus, but this inhibition is partially blocked by progesterone [
40]. To date, no information is available regarding the expression profile and function of Reck in rat ovaries development and maturation.
To analyse the ovarian extracellular matrix remodeling, we analyzed an artificial estral cycle, with a combination of equine chorionic gonadotropin (eCG) and human CG (hCG) in pre-pubertal rats [
41], that have not yet started their estrous cycle, to induce multiple follicular development and ovulation. The rats only start their estrus cycles after the vaginal orifice opens, which tends to occur between postnatal 32 and 36 days [
41]. The literature describes that the administration of 10UI eCG stimulates the folliculogenesis and after 48 h the injection of 10UI of hCG can induce ovulation [
42,
43]. In sequence, it is observed formation of corpus luteum, and 24 and 48 h post-hCG injection, the luteal formation in the ovaries. After 4 and 8 days, a peak of progesterone production is maximal resulting in the luteal maintenance, and corpus luteum regression can be analyzed at 14 days.
Therefore, understanding Reck and Sparc expression during the estrous cycle is essential to understand which role, if any, they play in the regulation of ovarian tissue remodeling, which allows all of the complex physiological processes that take place in the ovary. In this context, our hypothesis is that the Mmps regulators Reck and Sparc are modulated in ovarian dynamics possibly controlling the expression of Mmps, Reck being a possible inhibitor and Sparc a possible inducer of Mmps in this process.
To address these questions, we characterized the spatio-temporal expression pattern of Reck and Sparc proteins, as well as of Mmp2, Mmp9 and Mmp14 by immunohistochemistry in pre-pubertal rat ovaries during different phases of the hCG/eCG induced estrous cycle. To complement these data, we used qRT-PCR to investigate Mmps 2, 9, 13, 14, and 19, Timps 1, 2, and 3, Reck, Sparc, and Bsg mRNA levels and generate a complete expression profile panel of these targets during induced folliculogenesis, ovulation, luteogenesis, and luteolysis in prepubertal rat ovaries.
Discussion
In this study, we evaluated the spatio-temporal expression pattern of several Mmps and their regulating proteins in whole ovaries at several stages of the estrous cycle using immunohistochemistry to generate a panel of these players in rat ovarian dynamics. In addition, we also quantitatively assessed the mRNA expression of
Mmps and their regulators at those same estrous cycle stages. Interestingly, some of the analyzed genes (
Mmp9, Reck and
Sparc) displayed a discrepancy between the observed mRNA and protein levels at specific phases of the induced estrous cycle. A poor correlation between mRNA and protein levels has been extensively reported highlighting the importance of an integrative approach [
45‐
47] or the development of a gene specific RNA-to-protein conversion factor [
48]. Particularly, MMPs and their inhibitors have been shown to display discrepancies in their mRNA to protein correspondence in human prostate tissue and in skeletal muscle cancer in cachexia-associated matrix remodeling [
49,
50] and further studies are necessary to identify the mechanisms that lead to this disparity for this particular targets.
Analysis of Mmp spatial and temporal protein distribution, through immunohistochemistry, revealed that the Mmp2 and Mmp9 gelatinases display an overall higher expression during the rat estrous cycle, when compared to the Mmp14 collagenase. Gelatinases had a widespread spatial distribution in the ovaries, being detected in the theca, stromal, and granulosa cells, oocytes and CL. Mmp14 protein expression, on the other hand, was mainly restricted to theca and stromal cells, but at late stages of the estrous cycle, during luteolysis (G9), being also detected at the CL. The
Mmp14 transcript levels apparently did not vary across ovulation and CL lifespan, similarly to what has been described in the literature [
24,
51]. Nevertheless, high Mmp14 protein levels were detected in CL during its regression, indicating a possible contribution of this MMP to local tissue remodeling.
Mmp2 mRNA levels displayed a 2.8-fold increase during early folliculogenesis (G2), followed by a reduction (G3) and displayed a peak during the ovulation-luteogenesis transition phase (G5). At this same stage (G5), 24 h after hCG injection,
Mmp9 expression displayed an increase, peaking at luteinization (G6). Furthermore, both gelatinases displayed a reduced mRNA expression during CL maintenance (G7-G8) and regression (G9). Our results support previously published gene expression data from other groups, corroborating the increased gelatinolytic and collagenolytic activity during folliculogenesis and
Mmp2 and
Mmp9 expression 24 h post-hCG injection (luteogenesis, G5) in other studies [
16,
23,
24,
51], indicating a role for this enzyme in the tissue remodeling associated with luteolysis.
We observed an increase in
Mmp13 mRNA expression during the hormone-induced estrous cycle, which reached a peak of 150,000-fold 14 days after hCG injection (G9), suggesting that this gene is strongly correlated with luteal regression. These data are supported by others: Nothnick et al. [
23] observed a 15-fold increase in
Mmp13 mRNA expression 12 days after hCG and Liu et al. [
24], using an adult pseudo-pregnant rat model, also showed an increase in
Mmp13 expression during luteal regression [
24].
Our data suggest that Mmp expression is tightly related to the main biological events taking place during the estrous cycle, supporting previously published data and adding evidence to the importance of proteolytic tissue-degradation for ovulation in primates and rodents [
52]. Equally important as determining Mmps expression during the estrous cycle is to evaluate the expression of their inhibitors and inducers, since the balance between all of these proteins is what determines the extent and nature of ECM remodeling.
Timp1 was the only member of the Timp family to show significant mRNA modulation in this work, presenting a sharp increase in mRNA levels during ovulation (G4), 12 h after hGC administration. Conversely, for an Mmp inhibitor, its expression positively correlates with
Mmp2 gene expression. Li and Curry [
53] reported that the expression of both
Timp1 and
Timp3 mRNAs was induced upon treatment with hCG in pre-ovulatory rat granulosa cells in an epidermal growth factor receptor tyrosine kinase and mitogen-activated protein kinase (MAPK) pathway-dependent fashion [
53]. This differential regulation is of extreme interest because Timps are multifunction proteins, whose Mmp-inhibition independent functions might have an important role in this model.
Takahashi and colleagues observed that the RECK MMP inhibitor has a privileged peri-cellular localization in human cells, due to its anchoring to the cell membrane by GPI [
36]. Although
RECK gene is widely expressed in normal organs [
36], no description is available to date about its expression pattern in ovaries, despite the intense MMP-directed ECM remodeling that occurs in these organs. Our analysis of Reck expression in rat ovarian dynamics, by immunohistochemistry, showed no detectable Reck protein expression in early folliculogenesis (G2) and ovulation (G4), while Mmps 2, 9 and 14 levels were elevated. Following the same pattern, Reck protein is expressed at detectable levels in theca and stromal cells at the end of folliculogenesis (G3) whereas expression of Mmps is lower. In addition, during the ovulation-luteogenesis transition phase (G5), an increase in Reck protein expression in theca and stromal cells was observed in parallel to a decrease in Mmp9 expression, but not of Mmp2, in the same cells. Also, at late stages of ovarian remodeling, Reck was expressed in the
corpus luteum, during luteogenesis, CL maintenance and regression (G6-G9). Moreover,
Reck mRNA reached a significant expression peak during the late stages of folliculogenesis (G3), 48 h post-eCG stimulus, opposite to the reduction observed in gelatinases mRNA levels at this point. This orchestrated modulation of Reck and Mmps expression suggests that Reck may be involved in Mmps modulation in this model. In support of this hypothesis, the literature describes that RECK controls MMPs 2, 9, and 14 activities through several mechanisms, including inhibition of their proteolytic activity and activation cascade, as well as their sequestration [
36,
39,
54].
Regarding Mmp inducers, our gene expression analysis shows that
Bsg mRNA levels peak at CL manteinance (G8). In agreement, other authors used a rat hormone-induced estrous cycle model to determine a
Bsg mRNA expression panel, observing that this gene displayed increased expression during luteinization, which persisted until CL regression [
10]. Sparc, another Mmp inducer, was also investigated, revealing a spatio-temporal pattern modulation in rat ovaries during the estrous cycle. Spatial and temporal expression of Sparc protein was widespread, similarly to the pattern observed for Mmps. The Sparc protein profile was the similar to that of Mmps and opposite to the one displayed by Reck. According to Bagavandoss and collaborators, Sparc protein has been detected in theca and interstitial cells and its expression has been shown to increase in rat granulosa cells following hCG administration, during the final phases of ovarian development [
32]. However, in the current study, we also detected Sparc protein expression in theca, stromal, granulosa cells and oocytes in the whole hormone-induced estrous cycle, except at late folliculogenesis. It is important to note that this is the first time that Sparc protein was observed in oocytes and in the initial stages of the ovarian remodeling, namely folliculogenesis, ovulation and ovulation-luteinization transition in rat model. In addition, gene expression analyses indicate high levels of relative
Sparc mRNA expression during late folliculogenesis (G3) and ovulation-luteogenesis transition (G5, G6). Taken together, these results suggest that the Sparc transcript and protein may also play a role in the estrous cycle. In this same direction, Smith et al., using an ovine model, suggested that
SPARC gene is an important modulator of ovarian development and dynamics which is involved in tissue reconstruction during pregnancy [
35].
Therefore, our results generate an expression profile panel of Mmps and their regulators, suggesting that Reck and Sparc seem to play an important role at specific steps of the process, being part of a larger picture in rat ovarian dynamics.