The prostate is a branched gland accessory to the reproductive tract. In humans, prostate development takes place during the second and third trimester of gestation; whereas in rodents, the prostate gland displays significant growth and morphogenetic changes during the first postnatal week. In this period, the epithelial cells proliferate and differentiate, together with epithelial growth, branching, and canalization (Sugimura et al.,1986a; Donjacour and Cunha,1988; Hayward et al.,1996; Bruni-Cardoso and Carvalho,2007; Bruni-Cardoso et al.,2008). Epithelial growth involves colonization of spaces previously occupied by the mesenchyme/stroma (Bruni-Cardoso and Carvalho,2007).
This early postnatal developmental period involves extensive extracellular matrix (ECM) remodeling and coincides with a physiological window in which the prostate is capable of responding to hormones and endocrine disruptors that can affect prostate physiology and susceptibility to diseases in adult life (Putz et al.,2001; Risbridger et al.,2005).
MMPs are a family of zinc-dependent endopeptidases that preferentially cleave ECM proteins, playing a key role in normal development and physiology (Wiseman et al.,2003; VanSaun and Matrisian,2006; Page-McCaw et al.,2007), as well as in cancer initiation and progression (Matrisian and Bowden,1990; Sternlicht and Werb,2001; Lynch et al.,2005). Recently, in vitro and ex vivo studies by our group have defined roles for MMP-2 in prostate canalization and morphogenesis (Bruni-Cardoso et al.,2010) but to date the exact role for MMP-2 in the development of the organ in vivo remains unexplored. Given the prevalence of MMP-2 in the epithelialization of other organs such as the kidneys (Lelongt et al.,1997) and the mammary gland (Wiseman and Werb,2002), the current study uses mice systemically null for MMP-2 to evaluate the role of MMP-2 in mouse ventral prostate (VP) morphogenesis by comparing some developmental aspects and structure in age-matched neonatal (day 6) and adult (day 60) wild-type and MMP-2−/− mice on a C57BL/6 background. The data presented here demonstrate that the ablation of MMP-2 activity resulted in compromised morphogenesis (less epithelial-cell proliferation, fewer epithelial tips) and reduced weight and epithelial volume in adulthood. Additionally, we also noted increased immunostaining for MMP-9, suggesting a compensatory mechanism, which is, however, not sufficient to compensate for the role of MMP-2 in the growth of the wild-type VP.
MMP-2 Promotes Branching and Epithelial Proliferation in VP Development
Initially, we assessed the localization of MMP-2 in neonatal wild-type and MMP-2−/− VPs. Using immunohistochemistry, MMP-2 was found in the mouse VP on day 6, in both stroma and epithelium, being more concentrated in the epithelial distal regions (Fig. 1A), suggesting a role for this enzyme during epithelial branching and elongation, findings that are in agreement with our earlier in vitro and ex vivo studies in the rat (Bruni-Cardoso et al.,2008,2010). Tissue organization was determined by stereology. No difference was found in the volume density of the epithelium, lumen, and stroma at day 6 after birth (Fig. 1B).
Immunohistochemistry with a pan-cytokeratin (a general epithelial marker) antibody revealed aspects of epithelial organization in the early postnatal VP. Epithelial structures from both the wild-type and the MMP-2−/− ventral prostate were already branched at the end of the first postnatal week, but, unlike the wild-type ventral prostate, the MMP-2−/− samples showed dilated profiles in the distal epithelium. Accordingly, the MMP-2−/− VP showed fewer epithelial tips compared with the wild-type (Fig. 1C).
The normal mouse VP epithelium grows extensively during the first postnatal week (Sugimura et al.,1986b). The MMP-2−/− VP showed fewer proliferating epithelial cells than the wild-type gland (Fig. 1D). No statistical differences in the rate of proliferating mesenchymal cells were found between the wild-type and MMP-2 null groups (results not shown). We could detect no difference between the MMP-2−/− and the wild-type mice, in the rate of epithelial cell death (Fig. 1E). No apoptotic cells were found in the stromal compartment.
MMP-2 Contributes to Adult VP Size and Structure
The MMP-2−/− mice showed a lower body weight, as previously reported (22.5 ± 1.1 g vs. 18 ± 2.7 g; P < 0.05) (Kato et al.,2001). By calculating the relative weight, we noted that the growth of the VP gland was compromised in MMP-2−/− compared with the wild-type VP (0.032 ± 0.004% vs. 0.023% ± 0.002%; P < 0.05), suggesting that MMP-2 is essential for the normal development of the prostate gland.
MMP-2 was identified in the adult mouse VP and was located in the epithelium and around smooth-muscle cells (Fig. 2A). As expected, MMP-2 staining was absent in the VP of the MMP-2−/− mice (Fig. 2A).
The stereological analysis of the adult animals revealed that the volume density of the epithelium was smaller in MMP-2−/− VP than in wild-type samples (Fig. 2B). The resulting volumes of the epithelium and stroma were reduced (Fig. 2B). We also noted that the volume density and volume of the smooth-muscle cells were reduced in the adult VP (results not shown).
Despite the lack of variation in luminal volume, the histological analysis of the adult VP showed that the MMP-2−/− VP epithelial acini were dilated and the epithelium was folded in some microscope fields, aspects not observed in the wild-type mouse (Fig. 2C).
No difference was found in both the number of proliferating cells (Fig. 2D) and in the rate of epithelial apoptosis between the MMP-2−/− and wild-type mice (Fig. 2E). Neither proliferating nor apoptotic cells were found in the stromal compartment.
Enhanced Endogenous MMP-9 in Adult MMP-2 Null VP Does Not Rescue the Wild-Type Phenotype
Previous studies have suggested that the ablation of some MMPs can result in compensatory expression of other MMPs (Rudolph-Owen et al.,1997; Esparza et al.,2004). Because MMP-2 ECM substrate specificities share a significant overlap with those of MMP-9 (Overall,2002; Butler et al.2010), we examined the localization and expression of MMP-9 in the wild-type and MMP-2 null neonatal and adult VPs. Immunohistochemistry revealed the presence of MMP-9 in the early postnatal developing VP (Fig. 3A,B), in agreement with previous results for the rat (Bruni-Cardoso et al.,2008). In adults, MMP-9 protein could not be detected in the VP of the adult mice, as assessed by immunohistochemistry, an observation that is in agreement with our previous results examining adult rat VPs (Bruni-Cardoso et al.,2009; Fig. 3C). However, the MMP-2−/− samples showed staining for this enzyme in the stroma, particularly in the stromal cells (Fig. 3D). This suggests that MMP-9 expression is activated in the absence of MMP-2 activity in the ventral prostate of MMP-2−/− mice but does not rescue the MMP-2 null phenotype because the MMP-2−/− ventral prostates do not reach the same size as their wild-type counterparts.
MMP-2 Is Critical for VP Development and Contributes to the Assembly of the Reticulin Fibers
Considering the critical function of MMP-2 for the processing of ECM components abundant in the prostate stroma, we examined whether the observed effects on the VP in neonatal and adult mice were due to altered reticulin fiber deposition. In wild-type neonatal animals, reticulin fibers were observed surrounding the epithelial cords in the developing mouse ventral prostate (Fig. 4A). Additionally, these fibers were thinner and fewer in the distal regions of the wild-type samples (Fig. 4A). In contrast, the neonatal MMP-2−/− ventral prostate showed thicker and more-abundant reticulin fibers (Fig. 4A), suggesting that the accumulation of these fibers might be responsible for the smaller number of epithelial tips and also the dilated epithelial cords in the distal regions of the developing gland. This altered organization of reticulin fibers in MMP-2−/− ventral prostates persisted into adulthood (Fig. 4A) and also potentially explains why the MMP-2−/− adult ventral prostates are significantly smaller than their wild-type counterparts. Quantitative assessment of these results showed increased content of reticulin per sectional area (Fig. 4B) as well as an increased reticulin/epithelium area ratio (Fig. 4C), which achieved statistical significance at adulthood.
Taken together, these data suggest that MMP-2 is critical for appropriate prostate development by helping the proper assembly of collagen fibers.
We investigated the role of MMP-2 in the mouse VP morphogenesis, by studying the MMP-2−/− mice. We showed that MMP-2 is expressed in the early postnatal and adult VP, and contributes to epithelial cell proliferation and normal branching morphogenesis: the MMP-2 knockout showed a smaller gland with altered tissue organization, despite the detection of MMP-9, in adulthood.
Using an in vitro/ex vivo organogenesis model, we previously reported that the activities of MMP-2 and -9 are higher and localized in the distal ductal region and at the epithelial/stromal interface of rat VP during the first postnatal week (Bruni-Cardoso et al.,2008). Subsequently, we showed that MMP-2 modulates the morphogenesis of the rat ventral prostate in vitro by influencing the epithelial cell proliferation rate, branching, and the canalization of the epithelium, and the remodeling of collagen fibers (Bruni-Cardoso et al.,2010). Although these results indicate that MMP-2 activity and the resulting ECM remodeling are necessary for VP morphogenesis, in particular to accommodate epithelial growth and its projection into the surrounding stroma, a more physiological examination of this hypothesis was needed. Therefore, in the current study, we examined the effect of MMP-2 on prostate development in vivo using MMP-2 null mice.
Studies have reported the involvement of MMPs during the branching morphogenesis of many organs, such as the kidney (Lelongt et al.,1997), mammary gland (Wiseman and Werb,2002), lung (Kheradmand et al.,2002), and submandibular salivary gland (Steinberg et al.,2005; Rebustini et al.,2009).
MMP-2-null mice have smaller body size and reduced neovascularization (Kato et al.,2001). The present findings are in agreement with those examining the role of MMP-2 in the development of other organs such as the mammary gland (Wiseman et al.,2003) and lung (Kheradmand et al.,2002) for the which decreased primary ductal invasion and saccular development were observed, respectively.
As in the rat, the wild-type mouse VP on day 6 after birth expresses MMP-2 in both the mesenchyme and epithelium. In the epithelium, MMP-2 was concentrated at the epithelial tips, indicating a possible contribution of the enzyme to the epithelial invasion through the prostate stroma. As a result of knocking out MMP-2, we found fewer epithelial tips, a reduced rate of epithelial cell proliferation, and accumulation of reticulin fibers around the epithelium, suggesting that MMP-2 contributes to the branching morphogenesis, by remodeling some extracellular matrix proteins and consequently creating spaces for the bifurcation and elongation of the epithelial structures and influencing cell proliferation.
Studies of the adult VP showed that the MMP-2−/− mice have lower ventral-prostate weight and decreased epithelial (and smooth-muscle cell) volume. The epithelial and smooth-muscle cells play crucial roles in the prostate physiology. Luminal epithelial cells are responsible for the gland secretory activity, while smooth-muscle cells help to eliminate the secretion accumulated in the lumen during ejaculation (McNeal et al.,1988). In addition, paracrine signaling between these two cell types is essential in all stages of the gland development and homeostasis (Cunha and Chung,1981). These interactions might be compromised in the MMP-2−/−, as the thickened basement membrane, which normally acts as a physical and chemical barrier between the epithelium and the surrounding smooth-muscle cells, could obstruct the diffusion of regulatory molecules between the two compartments. Because epithelial cell proliferation and apoptosis rates were not affected by the absence of MMP-2 in the adult samples, we conclude that the smaller size is a result of reduced cell proliferation and compromised morphogenesis during development.
MMPs are transcriptionally regulated by a variety of growth factors and cytokines (Qin et al.,1999). In addition, posttranscriptional mechanisms can contribute to this regulation (Borden and Heller et al.,1997). Although MMP-2 expression is constitutive, the levels of this enzyme can change during normal development, inflammation, and tumor progression (Qin et al.,1999). The rodent prostate gland shows three main growth phases. After an initial embryonic stage, early postnatal growth takes place during the first 3 postnatal weeks, in response to a testosterone surge (Corbier et al.,1992). We believe that the high testosterone levels in this period might regulate the expression of MMPs, especially MMP-2. In fact, Liao et al. (2003) showed that testosterone regulates MMP-2 expression in LNCaP and LAPC-4 in a dose-dependent manner. Furthermore, there is an androgen-responsive element (ARE) in the MMP-2 promotor gene (Li et al.,2007), indicating that MMP-2 expression might be regulated by androgens during prostate development.
Our current findings and previous reports clearly demonstrate a role for MMP-2 in mediating prostate gland development, despite de novo expression of proteinases with similar substrate specificities such as MMP-9 (Overall,2002; Butler et al.,2010). It is important to note, however, that MMP-2 may not just contribute to ECM degradation, but specific cleavages in ECM components may reveal neoepitopes within the ECM substrates that provide critical information for the invading epithelial cells. For example, MT2-MMP activity results in the release of the NC1 fragment of collagen IV, which is critical for integrating collagen IV synthesis and proteolysis with epithelial proliferation during branching morphogenesis of submandibular gland (Rebustini et al.,2009). It will be important to characterize the importance of these factors in VP development, by assessing other null mice strains, as well as by the use of pharmacological inhibitors both in vivo and ex vivo.
We have suggested previously that matrix remodeling is essential for the epithelium to grow and invade the stroma (Bruni-Cardoso et al.,2010). In this context, we found that inhibition of MMP-2 by siRNA resulted in a more effective impairment of prostate growth than the use of the broadly specific inhibitor GM6001, and, therefore, we could not rule out the possibility that some MMPs may have an inhibitory effect on branching morphogenesis. The inhibitory effect may result from the degradation of essential extracellular matrix or nonmatrix components such as growth factors and cytokines that contribute to prostate gland development, as demonstrated for tenascin in the lung (Gebb and Jones,2003) and type III collagen in the salivary gland (Nakanishi et al.,1988).
In conclusion, the present in vivo results for the mouse model indicate that MMP-2 plays an important role in epithelial growth and morphogenesis in the mouse VP, through a mechanism involving ECM remodeling.
C57BL/6 wild-type mice and homozygous MMP-2 null mice were housed with IACUC approval to Dr. Matrisian at Vanderbilt University, Nashville, Tennessee. Breeding pairs were mated for 24 hr after which males were removed. Pregnant females were isolated and monitored. Litters were typically born at 21 days postcoitus, and the day of birth was termed day 0. Males were killed on day 6 (neonate) by decapitation, or on day 60 (adult) by CO2 inhalation, and had their VPs removed and processed for paraffin embedding. We used five animals per group in each age point. All adult VPs were weighed, and their relative weight was calculated as a percentage of the animal body weight.
Stereology and Counting of Epithelial Tips
The VPs were fixed in 4% formaldehyde, processed for routine paraffin embedding, stained with hematoxylin–eosin, and submitted to stereological analysis. Volume densities (Vv%) and the volumes of the epithelium, lumen, and stroma were determined by the method of Weibel (1963). A total of 144 dots/72 grid lanes were analyzed as described previously for the VP (Huttunen et al.,1981; Antonioli et al.,2004,2007; Garcia-Florez et al.,2005). Four or five microscopic fields taken at random were analyzed per animal (n = 4), resulting in 16–20 fields per group. Volume density was calculated by considering the number of points falling on a given compartment, epithelium, lumen, nonmuscle stroma (which includes everything in the stroma except smooth-muscle cells), muscle stroma (smooth-muscle cells), and total stroma (which is a summation of nonmuscle and muscle stroma) for the adult VP, and the epithelium, lumen, and stroma for the neonate VP, after conversion to percentages (144 points equal 100%). The number of epithelial profiles in the distal regions (tips) was counted on 2 equatorial cross sections per animal (n = 4), as previously described by Uchida et al. (2007).
Five-micrometer sections were de-waxed and subjected to antigen retrieval by boiling in 10 mM citrate buffer, pH 6.0 for 10 min in a microwave oven, and then treated for 20 min with 20 μg/mL proteinase K in 10 mM Tris-HCl, pH 7.4 buffer at room temperature. Sections were then blocked with blocking solution (1% bovine serum albumin and 5% donkey serum in 10 mM Tris-HCl, pH 7.4 containing 0.1 M MgCl2, 0.5% Tween-20), before incubation with rabbit anti-mouse MMP-2 (cat. ab7052), MMP-9 (catalog no. ab38898; Abcam, Cambridge, MA; diluted 1:500), mouse monoclonal anti-Pan-Cytokeratin-26 (catalog no. ab6401; Abcam; diluted 1:250) in blocking solution overnight at 4°C. After rinsing, except for MMP-9 staining, sections were incubated with Alexa-fluor 488-conjugated donkey anti-rabbit (cat. A11008) or anti-mouse IgG (catalog no. A11059; Invitrogen, Eugene, OR; diluted 1:500) in blocking solution. MMP-9 staining was performed by using the Vectastain Kit (catalog no. pk6100; Vector Laboratories, Burlingame, CA), before developing the reaction with 3,3′-diaminobenzidine tetrahydrochloride and counterstaining with Mayer's hematoxylin. Negative controls were performed by substituting the primary antibody with the appropriate rabbit or mouse IgG at the same concentration as the primary antibody. Sections were then counterstained with 4′-6-diamidino-2-phenylindole (DAPI).
Histological Assessment of VP Growth
Mitotic cells were identified by immunohistochemistry as described above, using a rabbit anti phospho histone H3 (cat. 06570; Millipore, Temecula, CA). For adult VP, at least two equatorial sections per animal (n = 4) were assessed by counting the total number of phospho-histone H3–positive nuclei per section. For animals on day 6, the mitotic index was determined by counting phospho-histone H3–positive nuclei with respect to the total number of epithelial or stromal cell nuclei in five microscope fields taken at random per animal (n = 4) using a ×20 objective, in a Leica DM 2500 microscope.
The TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling) assay was performed in paraffin sections using the in situ cell-death detection kit (Roche Diagnostics, Indianapolis, IN), counterstained with DAPI and analyzed in a Leica DM2500 fluorescence microscope, according to the manufacturer's instructions. Five microscopic fields from equatorial sections of a VP from each of four animals per group were taken at random, and the frequency of apoptotic cells was counted and expressed as a percentage of the total number of epithelial cell nuclei.
Silver Impregnation for Reticulin Fibers
VP paraffin sections submitted to Gömöri's silver impregnation staining were used for the identification of reticulin fibers, as previously described (Bruni-Cardoso et al.,2008). Briefly, the procedure involves sequential treatment with 1% potassium permanganate, 3% oxalic acid, and 1% iron alumen followed by incubation with 10% ammoniacal silver. After the silver impregnation, reticulin fibers (mainly type III collagen fibrils) appear black. Measurements were taken after manual segmentation using the Image J (NIH free software) and the results are presented as the percentage of the sectional are occupied by reticulin fibers and as the reticulin-epithelium area ratio, by using at least 10 microscopic fields from at least three animals.
Measurements were submitted to Student's t-test using the Mini Tab 14 statistical software. Values of P < 0.05 were considered statistically different.
Financial support from the State of Sao Paulo Research Funding Agency and National Research Council (CNPq) through grants to HFC is acknowledged. A.B.-C. was a recipient of a FAPESP fellowship. INFABIC is co-funded by CNPq and FAPESP.