Interleukin-17 promotes prostate cancer via MMP7-induced epithelial-to-mesenchymal transition.

Chronic inflammation has been associated with a variety of human cancers including prostate cancer. Interleukin-17 (IL-17) is a critical pro-inflammatory cytokine, which has been demonstrated to promote development of prostate cancer, colon cancer, skin cancer, breast cancer, lung cancer and pancreas cancer. IL-17 promotes prostate adenocarcinoma with a concurrent increase of matrix metalloproteinase 7 (MMP7) expression in mouse prostate. Whether MMP7 mediates IL-17's action and the underlying mechanisms remain unknown. We generated Mmp7 and Pten double knockout (KO) (Mmp7-/-) mouse model and demonstrated that MMP7 promotes prostate adenocarcinoma through induction of epithelial-to-mesenchymal transition (EMT) in Pten-null mice. MMP7 disrupted E-cadherin/β-catenin complex to upregulate EMT transcription factors in mouse prostate tumors. IL-17 receptor C and Pten double KO mice recapitulated the weak EMT characteristics observed in Mmp7-/- mice. IL-17 induced MMP7 and EMT in human prostate cancer LNCaP, C4-2B and PC-3 cell lines, while small interfering RNA knockdown of MMP7 inhibited IL-17-induced EMT. Compound III, a selective MMP7 inhibitor, decreased development of invasive prostate cancer in Pten single KO mice. In human normal prostates and prostate tumors, IL-17 mRNA levels were positively correlated with MMP7 mRNA levels. These findings demonstrate that MMP7 mediates IL-17's function in promoting prostate carcinogenesis through induction of EMT, indicating IL-17-MMP7-EMT axis as a potential target for developing new strategies in the prevention and treatment of prostate cancer.

INTRODUCTION

Chronic inflammation has been associated with a variety of human cancers. Approximately 15% of all human cancers have been suggested to result from infection and chronic inflammation.[1] Almost all surgical prostate specimens contain evidence of inflammation.[2-4] Chronic inflammation invokes proliferative inflammatory atrophy of prostate – a potential precursor lesion to prostatic intraepithelial neoplasia (PIN) and carcinoma.[5] The cause of prostatic inflammation includes infection, urine reflux, diet, estrogen, and physical trauma.[6,7] Inflammation is a complex response involving many immune cells, chemokines, and cytokines as well as matrix-degrading enzymes.

Interleukin-17 (IL-17, also named IL-17A) is a key pro-inflammatory cytokine that plays critical roles in many inflammatory and autoimmune diseases.[8] IL-17 has been demonstrated to promote development of colon cancer,[9-12] skin cancer,[13,14] breast cancer,[15] prostate cancer,[16,17] lung cancer,[18,19] and pancreas cancer.[20] IL-17 is secreted by T helper 17 (TH17) cells, γδ T cells, natural killer cells, and other immune cells.[21] IL-17 acts on IL-17RA/IL-17RC receptor complex to recruit nuclear factor-κB (NF-κB) activator 1 (Act1). Act1 activates tumor necrosis factor receptor-associated factor 6 (TRAF6),[22] and subsequently activates transforming growth factor-β-activated kinase 1 (TAK1) and IκB kinase (IKK) complex, resulting in activation of NF-κB pathway that initiates transcription of a variety of chemokines and cytokines, such as C-X-C motif ligand 1 (CXCL1), C-C motif ligand 20 (CCL20), IL-1β, and IL-6.[8] These IL-17-downstream factors promote cancer formation through increased cellular proliferation, attenuated apoptosis, and sustained angiogenesis, as well as creation of an immunotolerant microenvironment.

We have previously generated an IL-17 receptor C (Il-17rc) and prostate-specific conditional phosphatase and tensin homolog (Pten) double knockout (KO) mouse model.[16,17] IL-17RC-deficient (IL17-RC− or RC−) mice display smaller prostates and develop a reduced number of invasive prostate adenocarcinomas, compared to IL-17RC-sufficient (IL17-RC+ or RC+) mice. Further, matrix metalloproteinase 7 (MMP7) expression is increased in RC+ mice compared to RC−mice.[16] However, whether MMP7 mediates IL-17’s action and the underlying molecular mechanisms remain unknown. MMP7 (also known as putative metalloproteinase I or matrilysin) is exclusively expressed in the epithelial cells.[23] MMP7 is overexpressed in human prostate cancer,[24] but not expressed in normal prostate glands.[25] Here, we investigated the role of MMP7 in mediating IL-17’s action, using an Mmp7 and Pten double KO mouse model. Our findings demonstrate that MMP7 mediates IL-17’s function in promoting prostate carcinogenesis through induction of epithelial-to-mesenchymal transition (EMT).

EMT involves changes in epithelial cells to behave more like mesenchymal cells.[26] Cells undergoing EMT switch from a polarized epithelial phenotype to a highly mobile mesenchymal phenotype.[27] Expression of epithelial markers such as E-cadherin, claudin, and zona occludens 1 (ZO-1) is decreased, whereas expression of mesenchymal markers such as vimentin and N-cadherin is increased. EMT has been associated with cellular invasiveness[28] and cancer metastasis.[29-31]

RESULTS

MMP7 is the main active MMP in mouse prostate tumors

Mmp7 traditional KO mice[32] were crossbred with Pten conditional KO mice[33] to generate Mmp7+/+;Pten;Cre+ (Mmp7 in abbreviation) mice, Mmp7;Pten;Cre+ (Mmp7 in abbreviation) mice, and Mmp7;Pten;Cre+ (Mmp7 in abbreviation) mice (Figure 1a). Male mice were genotyped at 3 weeks of age (Figure 1b). MMP7 protein in mouse prostates was confirmed by immunohistochemical (IHC) staining (Figure 1c) and Western blot (Figure 1d). To assess MMP enzyme activity in mouse prostates, MMPSense™ 750 FAST Fluorescent Imaging Agent (PerkinElmer, Inc., Waltham, MA) was injected intravenously into 30-week-old Mmp7+/+ and Mmp7 mice. This agent is optically silent and produces fluorescent signals after cleavage by active MMPs including MMP2, 3, 7, 9, 12, and 13. The animals were scanned with IVIS® Lumina XRMS imaging system (PerkinElmer, Inc.).[34] Mmp7+/+ mice, but not Mmp7 mice, showed MMP activities in the prostate region (Figure 1e). Scanning of the freshly dissected genitourinary blocs (GU-blocs) confirmed that the fluorescent signals came from Mmp7+/+ prostates, but not Mmp7 prostates (Figure 1f). Together, these results indicated that MMP7 was the main active MMP in mouse prostate tumors.

Figure 1

Establishment of Mmp7 and Pten double KO mouse model. (a) Strategy of animal breeding. (b) Representative gel images of PCR genotyping. WT, wild-type; HT, heterozygous; KO, knockout. (c) IHC staining of MMP7 in dorsal lobes of 30-week-old mouse prostates. (d) Western blot analysis of MMP7 protein expression in 30-week-old mouse prostates. (e, f) Fluorescence imaging of MMP activities in Mmp7 and Mmp7 mice or mouse prostates. Arrows indicate the fluorescent signals.

Mmp7 mice develop smaller prostate tumors than Mmp7

The GU-bloc weight has often been used to represent the prostate tumor burden.[16] Prostate tumors of Mmp7+/+ mice were obviously larger than those of Mmp7 mice at 30 weeks of age (Figure 2a). At 9 weeks of age, the GU-bloc weight showed no significant differences among the three groups of animals (P > 0.05). However, at 30 weeks of age, the GU-bloc weight of Mmp7+/+ mice was significantly heavier than that of Mmp7 mice (P < 0.05, Figure 2b). The GU-bloc weight of Mmp7+/− mice was slightly heavier than that of Mmp7 mice (P > 0.05, Figure 2b). These results indicated that Mmp7 mice developed smaller prostate tumors than Mmp7 mice.

Figure 2

Mmp7 KO decreases formation of invasive prostate adenocarcinoma in mice. (a) Representative photographs of the GU blocs. (b) GU-bloc weight. The number of animals in each group is shown under the abscissa. *P < 0.05. (c) Representative sections of dorsal prostatic lobes stained with H&E or for laminin. Arrows indicate invasive sites in Mmp7 and Mmp7 mice and non-invasive site in Mmp7 mice. (d) Percentages of normal, PIN and cancer (invasive prostate adenocarcinoma) in ventral, dorsal, and lateral prostatic lobes at 9 and 30 weeks of age. The number of animals in each group is shown under the abscissa. **P < 0.01 compared to Mmp7 mice.

Mmp7 KO decreases formation of invasive prostate adenocarcinoma

We and other researchers have reported that Pten mice develop invasive prostate adenocarcinoma at 9 weeks of age.[16,33] Here, we found that invasive prostate adenocarcinomas were formed at different rates among Mmp7, and Mmp7 mouse prostates at 9 and 30 weeks (Figures 2c and d). At 30 weeks of age, 33% and 27% of prostatic glands presented with invasive prostate adenocarcinomas in Mmp7 and Mmp7 mice, respectively. In contrast, only 11% of prostatic glands showed invasive prostate adenocarcinomas in Mmp7 mice. The differences in the percentages of lesions were statistically significant between Mmp7 and Mmp7 mice at 9 and 30 weeks and between Mmp7and Mmp7 mice at 30 weeks (P < 0.01, Figure 2d). These results suggested that Mmp7 KO decreased formation of invasive prostate adenocarcinoma.

Mmp7 KO decreases cellular proliferation and increases apoptosis in the prostate lesions

To reveal the underlying cause of the differences in prostate tumor burden among the animals at 30 weeks of age, we found that there were significantly more Ki-67-positive cells in Mmp7 and Mmp7prostates than in Mmp7prostates (P < 0.05 or 0.01, Figures 3a and b). In addition, there were significantly fewer apoptotic cells in Mmp7 and Mmp7 prostates than in Mmp7 prostates (P < 0.05 or < 0.01, Figures 3c and d). These results suggested that the decreased prostate tumor burden in Mmp7 mice was due to decreased cellular proliferation and increased apoptosis in mouse prostates.

Figure 3

Mmp7 KO decreases proliferation, increases apoptosis, and inhibits angiogenesis in 30-week-old mouse prostates. (a) Representative prostate sections stained for Ki-67. Arrows indicate the positive cells. (b) Percentages of Ki-67-positive cells in mouse prostates. Data are represented as mean ± SEM, n = 3 animals per group, *P < 0.05 and **P < 0.01. (c) Representative prostate sections stained for apoptosis (TUNEL assay). Arrows indicate the positive cells. (d) Percentages of apoptotic cells in mouse prostates. Data are represented as mean ± SEM, n = 3 animals per group, *P < 0.05 and **P < 0.01. (e) Representative prostate sections stained for VEGFA. Arrows indicate the positive cells. (f) Western blot analysis of VEGFA expression in mouse prostates. (g) Representative prostate sections stained for CD31. Arrows indicate microvessels. (h) Density of microvessels in mouse prostates. Data are represented as mean ± SEM, n = 3 animals per group, *P < 0.05.

Mmp7 KO decreases angiogenesis in the prostate lesions

MMP7 has been associated with angiogenesis through stimulating proliferation of endothelial cells in human colon cancer[35] and human umbilical vein endothelial cells,[36] thus we assessed angiogenesis in mouse prostate tumors using IHC staining of CD31. We found that there were significantly more blood vessels in Mmp7+/+ prostates than in Mmp7+/− or Mmp7−/− prostates (P < 0.05), which was accompanied with higher levels of vascular endothelial growth factor A (VEGFA) in Mmp7+/+ prostates than in Mmp7+/− or MmpP7−/− prostates (Figures 3e to h). These results indicated that reduced VEGFA expression and angiogenesis in MMP7−/− prostates contributed to the decreased prostate tumor burden in MMP7−/− mice.

Mmp7 KO inhibits epithelial-to-mesenchymal transition (EMT) in the prostate lesions

To further understand the molecular mechanisms underlying the reduced prostate tumor formation in Mmp7 KO mice, we examined expression of several epithelial and mesenchymal marker proteins in the prostate lesions. We found that Mmp7 mice had obviously increased expression of epithelial markers such as E-cadherin, claudin, and ZO-1, compared to Mmp7+/+ and Mmp7+/− mice (Figures 4a and b; and Supplementary Figure S1f). In contrast, Mmp7 mice had reduced expression of mesenchymal markers such as β-catenin, vimentin, and N-cadherin, compared to Mmp7+/+ and Mmp7+/− mice (Figures 4c and d; and Supplementary Figure S1e). Western blot analysis of the prostate tissue proteins confirmed the IHC results (Figure 4e). Since EMT is induced by certain transcription factors such as Snail, Slug, Twist, and ZEB1, we also assessed their expression levels. We found that the expression levels of Snail, Slug, Twist, and ZEB1 were reduced in Mmp7 prostates, compared to Mmp7+/+ and Mmp7+/− prostates (Supplementary Figures S1a to d). Together, these results suggested that Mmp7+/+ and Mmp7+/− prostates showed higher expression levels of mesenchymal markers and transcription factors, whereas Mmp7 prostates showed higher expression levels of epithelial markers, implying that Mmp7 KO weakened EMT characteristics in the mouse prostate tumors.

Figure 4

Mmp7 KO inhibits epithelial-to-mesenchymal transition (EMT) in the prostate lesions in 30-week-old mice. (a-d) Representative prostate sections stained for EMT markers. Arrowheads indicate the positive cells. (e) Western blot analysis of EMT markers in mouse prostates.

MMP7 induces EMT by disrupting E-cadherin/β-catenin complex

E‐cadherin interacts with a β-catenin-based complex to act on actin cytoskeleton and mediate adhesion-dependent signaling, and several proteinases including MMP7 are known to be able to cleave E-cadherin.[37-39] Thus, we tested if MMP7 could cleave E-cadherin in three human prostate cancer cell lines LNCaP, C4-2B, and PC-3. We generated MMP7-overexpressing cell lines LNCaP-MMP7, C4-2B-MMP7, and PC-3-MMP7. We observed that LNCaP-MMP7 and C4-2B-MMP7 cells appeared like mesenchymal (spindle-shaped) cells, whereas the parental LNCaP and C4-2B cells appeared like epithelial (cobblestone-like) cells (Figures 5a to d). In addition, E-cadherin staining changed from cytoplasmic membrane localization in LNCaP and C4-2B cells to cytoplasmic localization in LNCaP-MMP7 and C4-2B-MMP7 cells (Figures 5e to h). β-catenin staining also changed from cytoplasmic membrane localization in LNCaP and C4-2B cells to nuclear localization in LNCaP-MMP7 and C4-2B-MMP7 cells (Figures 5i to l). Further, Western blot analysis showed that full-length E-cadherin and ZO-1 levels were reduced, while soluble E-cadherin (sE-cadherin, a cleaved fragment that is 40 KDa shorter than the full-length E-cadherin) levels were increased in LNCaP-MMP7 and C4-2B-MMP7 cells, compared to LNCaP and C4-2B cells (Figures 5m and n). Of note, the levels of sE-cadherin were also higher in Mmp7+/+ and Mmp7+/− prostates than Mmp7−/− prostates (Figure 4e). As expected, the levels of β-catenin, vimentin, Snail, and Slug were increased in LNCaP-MMP7 and C4-2B-MMP7 cells, compared to LNCaP and C4-2B cells (Figures 5m and n). Similar results were obtained from PC-3 and PC-3-MMP7 cells (Supplementary Figure S3f). It has been reported that E-cadherin and β-catenin form a complex,[40] and we confirmed this in LNCaP, C4-2B, and PC-3 cells (Supplementary Figure S3i and j). Together, these results suggested that MMP7 cleaved E-cadherin to release β-catenin from E-cadherin/β-catenin complex, leading to nuclear translocation of β-catenin and subsequently activation of downstream transcription factors Snail and Slug, hence inducing EMT.

Figure 5

MMP7 induces EMT in LNCaP and C4-2B cells by disrupting E-cadherin/β-catenin complex. (a-d) Phase-contrast photomicrographs of human prostate cancer cells in monolayer culture. (e-h) Human prostate cancer cells were stained for E-cadherin (in green color) and the nuclei (in blue color). (i-l) Human prostate cancer cells were stained for β-catenin (in red color) and the nuclei (in blue color). (m-n) Western blot analysis of EMT markers in human prostate cancer cells.

MMP7 increases prostate cancer cell invasion

Since EMT is known to promote cancer cell invasion,[41] we assessed if MMP7-induced EMT could promote prostate cancer cell invasion. We found that LNCaP-MMP7, C4-2B-MMP7, and PC-3-MMP7 cells invaded through Matrigel®-coated porous membranes in significantly larger numbers than LNCaP, C4-2B, and PC-3 cells (P < 0.05 or 0.01, Figures 6a to d, and Supplementary Figure S3a to e). On the other hand, mouse prostate cancer cells isolated from Mmp7−/− prostates invaded through Matrigel®-coated porous membranes in significantly smaller numbers than mouse prostate cancer cells isolated from Mmp7 prostates (P < 0.01, Figures 6e and f). These results indicated that MMP7 expression increased prostate cancer cell invasion.

Figure 6

MMP7 increases prostate cancer cell invasion. (a-b) Representative photomicrographs of human prostate cancer cells invaded through the Matrigel®-coated porous membrane. (c-d) The number of invasion cells. Data are represented as mean ± SEM, n = 3 wells per group, *P < 0.05 and **P < 0.01. The experiment was repeated 3 times. (e) Representative photomicrographs of mouse prostate cancer cells invaded through the Matrigel®-coated porous membrane. (f) The number of invasion cells. Data are represented as mean ± SEM, n = 3 wells per group, **P < 0.01. The experiment was repeated 3 times.

IL-17 induces MMP7 to enhance EMT

Since we have previously demonstrated that IL-17 induces MMP7 expression in mouse prostate,[16] we hypothesized that IL-17 could induce MMP7 to enhance EMT. To test this hypothesis, we first assessed EMT markers in RC+ and RC− mouse prostate tumors. These tumors were developed due to Pten KO and differed in either expressing Il-17rc (RC+) or not expressing Il-17rc (RC−).[16] We found that RC− prostate tumors had increased expression of E-cadherin, claudin, and ZO-1, but had decreased expression of MMP7, β-catenin, vimentin, and N-cadherin, compared to RC+ prostate tumors by IHC staining (Figures 7a to e; and Supplementary Figures S2e and f). Western blot analysis confirmed the IHC results (Figure 7f). Further, RC− prostate tumors had reduced expression of Snail, Slug, Twist, and ZEB1, compared to RC+ prostate tumors by IHC staining (Supplementary Figures S2a to d). These results showed that Il-17rc KO mice recapitulated the weak EMT characteristics in Mmp7 KO mice, suggesting existence of IL-17-MMP7-EMT axis. To test this axis, we treated LNCaP and C4-2B cells with recombinant human IL-17 and found that IL-17 indeed induced MMP7 expression in both LNCaP and C4-2B cells (Figure 7g). Although the levels of full-length E-cadherin were little affected by increased MMP7 expression, the levels of sE-cadherin were dramatically increased (Figure 7g). As expected, the levels of β-catenin (in the nuclei), Snail, Slug, and Twist were also increased along with the increase in MMP7 expression (Figure 7g). To verify if MMP7 indeed mediated IL-17’s action, we transfected LNCaP and C4-2B cells with either control small interfering RNA (siRNA) or MMP7-specific siRNA and then treated the cells with IL-17. We found that MMP7-specific siRNA knocked down the levels of MMP7 expression induced by IL-17. At the same time, the full-length E-cadherin levels were slightly increased compared to the control siRNA-transfected cells (Figure 7h). In contrast, sE-cadherin levels were dramatically reduced in MMP7-specific siRNA-transfected cells compared to control siRNA-transfected cells (Figure 7h). Further, the levels of β-catenin, Snail, Slug, and Twist were also decreased in MMP7-specific siRNA-transfected cells compared to control siRNA-transfected cells (Figure 7h). Similar results were obtained from PC-3 cells (Supplementary Figure S3g and h). Together, these results indicated that IL-17 induced MMP7 expression in prostate cancer cells to cleave E-cadherin and then release β-catenin to induce EMT transcription factors Snail, Slug, Twist, and ZEB-1, leading to increased expression of mesenchymal markers (vimentin and N-cadherin) and decreased expression of epithelial markers (E-cadherin, claudin, and ZO-1). Therefore, these results supported the existence of IL-17-MMP7-EMT axis in prostate carcinogenesis.

Figure 7

IL-17 induces MMP7 to enhance EMT. (a-e) Representative mouse prostate sections stained for EMT markers. RC+, Pten KO mice expressing IL-17RC. RC−, Pten KO mice not expressing IL-17RC. (f-h) Western blot analysis of EMT markers. Of note, the proteins were from either mouse prostate tissue homogenates (f) or whole cell lysates (g, h), except those used for β-catenin (g, h) that were from nuclear extracts.

Selective MMP7 inhibitor decreases development of prostate cancer

Pten single knockout male mice were treated with Compound III, a selective MMP7 inhibitor[42] for 5 to 6 weeks. Although there was no significant difference in GU-bloc weight between the control and MMP7 inhibitor treatment groups, the percentage of invasive prostate cancer was significantly less in the MMP7 inhibitor treatment group than the control group (Figure 8a to d). IHC staining showed increased expression of E-cadherin and ZO-1, but decreased expression of β-catenin, vimentin, Snail, Slug, Twist, and ZEB1 in MMP7 inhibitor treatment group, compared to the control group (Figure 8e to l).

Figure 8

MMP7 inhibitor decreases formation of invasive prostate adenocarcinoma in mice. Pten single conditional KO male mice at 6-week-old were randomized into treatment group (treated with MMP7 inhibitor) or control group (treated with vehicle DMSO); animals were euthanized at 12-week-old. (a) Representative photographs of the GU blocs. (b) GU-bloc weight. The number of animals in each group is shown under the abscissa. (c) Representative sections of dorsal prostatic lobes stained with H&E. Arrows indicate invasive sites in the control group. (d) Percentages of PIN and cancer (invasive prostate adenocarcinoma) in ventral, dorsal, and lateral prostatic lobes at 12 weeks of age. (e-l) Representative prostate sections stained for EMT markers. Arrows indicate the positive cells.

IL-17 mRNA levels are positively correlated with MMP7 mRNA levels in human normal prostates and prostate tumors

We analyzed the Cancer Genome Atlas (TCGA) data and found that both IL-17 and MMP7 mRNA levels were slightly increased in human primary prostate tumors compared to human normal prostates, however, the differences were not statistically significant (Figure 9a and b). Nevertheless, we found that IL-17 mRNA levels were positively correlated with MMP7 mRNA levels in human normal prostates and prostate tumors (Figure 9c).

Figure 9

IL-17 mRNA levels are positively correlated with MMP7 mRNA levels in human normal prostates and prostate tumors. Complete Clinical Dataset (level 2) and RNASeqV2 gene expression dataset (level 3) under the category of human prostate adenocarcinoma (PRAD) were downloaded from the TCGA website https://tcga-data.nci.nih.gov/tcga/. Log2 transformation was applied to gene expression levels in the included samples. (a, b) IL-17 and MMP7 mRNA levels in normal prostates and primary tumors. (c) Correlation between the mRNA levels of IL-17 and MMP7.

DISCUSSION

Prior to the use of spontaneous cancer models, the role of IL-17 in carcinogenesis was quite controversial.[43,44] IL-17 was proposed to have both pro-tumorigenic role[45,46] and anti-tumorigenic role.[47,48] The discrepancies arose from using nude mice, over-expression of IL-17, or grafting of different tumor types.[49,50] However, using animal models of autochthonous cancer, many independent groups have demonstrated that IL-17 promotes development of colon cancer,[9-12] skin cancer,[13,14] breast cancer,[15] prostate cancer,[16,17] lung cancer,[18,19] and pancreas cancer.[20] IL-17 has been shown to cause marked epithelial hyperproliferative responses and inflammatory infiltrates in a colon cancer model.[9] IL-17 has also been shown to promote chemical-induced inflammation and keratinocyte proliferation in a skin cancer model.[13] IL-17 increases cellular proliferation, decreases apoptosis, and enhances angiogenesis in the animal tumors.[16-18] IL-17 recruits myeloid-derived suppressor cells (MDSCs) and increases the suppressive function of MDSCs on T cells, thus creating an immunotolerant tumor microenvironment.[15,18] A recent study has demonstrated a hematopoietic-to-epithelial IL-17 signaling axis as another important driver of pancreatic carcinogenesis.[20]

The present study provides several lines of evidence to support the existence of IL-17-MMP7-EMT axis in prostate carcinogenesis. First, Mmp7 KO and Il-17rc KO mice showed similar prostate tumor phenotype in Pten-null mice. Second, both Mmp7 KO and Il-17rc KO prostates showed similarly weak EMT characteristics compared to wild-type mouse prostates. Third, IL-17 directly induced MMP7 expression in three human prostate cancer cell lines and in ex-vivo cultured mouse prostate tissues,[16] and IL-17 induced expression of EMT markers in three human prostate cancer cell lines. Finally, MMP7 knockdown blocked IL-17-induced EMT in three human prostate cancer cell lines, which strongly supports that MMP7 mediates IL-17’s action in induction of EMT. In human lung cancer cells, IL-17 activates NF-κB to upregulate ZEB1 expression, thus inducing EMT and enhancing lung cancer cell migration.[51] Our previous study has shown that Mmp9 (rather than Mmp7) is induced by IL-17 to promote tumor cell motility in a mouse model of lung cancer.[19] In two human nasopharyngeal cancer cell lines, IL-17 induces expression of MMP2 and MMP9 through activation of p38 kinase and NF-κB pathways, which is accompanied with decreased full-length E-cadherin and increased vimentin levels.[52] It is not clear whether MMP2 and MMP9 expression plays any role in regulating E-cadherin and vimentin levels in that study.[52] Nevertheless, our present study clearly shows that MMP7 cleaves E-cadherin into sE-cadherin in both human prostate cancer cell lines and mouse prostate tumors, resulting in disruption of E-cadherin-β-catenin complex and release of β-catenin that activates the downstream EMT transcription factors. Loss of membrane-associated E-cadherin has been shown to trigger β-catenin nuclear localization with subsequent c-Myc expression and cellular proliferation in human colon cancer cells.[53,54] β-catenin signaling also inhibits apoptosis.[55,56] Therefore, it is possible that IL-17-induced MMP7 expression not only triggers EMT through activation of β-catenin signaling, but also promotes cellular proliferation and inhibits apoptosis. Further, VEGFA has been shown to be a direct downstream target gene of β-catenin signaling,[57] thus the observed angiogenesis phenotype may also be linked to MMP7-activated β-catenin signaling.

In summary, our findings demonstrate that IL-17 induces MMP7 expression to disrupt E-cadherin/β-catenin complex and release β-catenin, thus enhancing EMT and tumor cell invasion, which indicates IL-17-MMP7-EMT axis as potential targets for developing new strategies in the prevention and treatment of prostate cancer. Our proof-of-principle animal study showed that a selective MMP7 inhibitor decreased expression of EMT markers and significantly reduced the number of invasive prostate cancer, suggesting that IL-17-MMP7-EMT axis is targetable. We also found that IL-17 mRNA levels were positively correlated with MMP7 mRNA levels in human normal prostates and prostate tumors. However, further studies are warranted to assess the prognostic potential of IL-17 and MMP7 in human prostate cancer.

MATERIALS AND METHODS

Mice

Animal study was approved by the Animal Care and Use Committee of Tulane University. Pten(Pten) and PB-Cre4 mice were described previously.[16,33] Mmp7mice (strain name: B6.129-Mmp7tm1Lmm/J; genetic background: B6.129*C57BL/6J)[32] were obtained from the Jackson Laboratory, Bar Harbor, ME. The breeding strategy for generating Mmp7−/− and Pten double KO mice is shown in Figure 1a. Tail DNA of 3-week-old pups was used for PCR genotyping and PCR primers are shown in Supplementary Table S1. Il-17rc and Pten double KO mouse prostates were obtained from our previous study.[16] In a separate experiment, 15 Pten single KO male mice[16,33] at 6-week-old were randomized by flipping a coin into treatment group (n = 10) and control group (n= 5). The treatment group was treated daily with 60 mg/kg body weight of Compound III, a highly selective MMP7 inhibitor[42] (a gift from AstraZeneca R&D Mölndal, Mölndal, Sweden). Compound III was first dissolved in dimethylsulfoxide (DMSO), then dissolved in phosphate-buffered saline (PBS) up to 200 μl, and injected intraperitoneally once a day from 6-week-old to 12-week-old. It was pre-established that animals should be included if they survived to 12-week-old and they should be excluded if they died before reaching 12-week-old. Two animals died after about 4 weeks of treatment due possibly to obvious adhesion of peritoneal organs, which were excluded from the analysis. Compound III treatment was discontinued at 11-week-old in 3 animals due to abdominal distention, and these 3 animals survived to 12-week-old and were included in the analysis. The control group was injected intraperitoneally with equal volume of DMSO dissolved in PBS once a day up to 12-week-old. All animals survived to 12-week-old were euthanized for necropsy.

In Vivo fluorescence imaging

Two nmols of MMPSense™ 750 FAST Fluorescent Imaging Agent (Cat#NEV10168, PerkinElmer, Inc.) were injected into the tail vein of 30-week-old male mice 12 hours prior to imaging using IVIS® Lumina XRMS imaging system (PerkinElmer, Inc.). The animals were anesthetized with XGI-8 gas anesthesia system (PerkinElmer, Inc.) according to the manufacturer’s instructions. Appropriate optical filters were set to collect fluorescence signals from both Mmp7+/+ and Mmp7−/− mice side-by-side simultaneously. Immediately after whole animal imaging, animals were euthanized and then the dissected GU-blocs were imaged for fluorescence signals.

Histopathology

Mice were euthanized and weighed at 9, 12, or 30 weeks of age. The GU-blocs were photographed, weighed with an empty bladder, and fixed as described.[16,58] Thirty-two consecutive 4-μm sections of each prostate were cut and eight sections (from every 8th section) per sample were H&E stained for histopathologic assessment in a genotype-blinded fashion according to the Bar Harbor Classification.[58] The prostatic glands were assessed under low- and high-power magnifications, and approximately 17 to 370 prostatic glands in each prostate were counted.

IHC, immunofluorescence, and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining

IHC staining was performed according to previously established protocols,[16,17] using VECTSTAIN ABC kits and DAB Substrate Kits (Vector Laboratories, Burlingame, CA) according to the manufacturer’s instructions. The primary antibodies used were: rabbit anti-E-cadherin (#3195), anti-β-catenin (#8480), anti-vimentin (#5741), anti-N-cadherin (#13116), anti-claudin-1 (#13255), anti-ZO-1(#8193), anti-Snail (#3879), anti-Slug (#9585), and anti-TCF8/ZEB1 (#3396) (Cell Signaling Technology, Beverly, MA); mouse anti-β-catenin (#05-613) and anti-Ki-67 (#AB9260) (Millipore, Billerica, MA); rabbit anti-VEGFA (#sc-152, Santa Cruz Biotechnology, Dallas, TX); rabbit anti-CD31 (#Ab28364, Abcam, Cambridge, MA); rabbit anti-laminin (#L9393, Sigma-Aldrich, St. Louis, MO); goat anti-MMP7 (#AF2967, R&D Systems, Minneapolis, MN). For immunofluorescence staining, monolayer cultured cells were fixed in methanol and blocked with 5% normal donkey serum in PBS; the cells were incubated with anti-E-cadherin or anti-β-catenin antibodies overnight at 4°C; after 3 washes, the cells were incubated with Cy™3-conjugated anti-mouse IgG (#715-485-150) or DyLight™488-conjugated anti-rabbit IgG (#711-485-152, Jackson ImmunoResearch Laboratories, West Groove, PA) for 1 hour at room temperature; and the nuclei were stained with Hoechst33342. TUNEL staining was performed using TACS.XL® Blue Label In Situ Apoptosis Detection Kits (Trevigen, Gaithersburg, MD) according to the manufacturer’s instructions. Quantification of the staining was performed as previously described.[16,17]

Cell Culture

Human prostate cancer cell lines (LNCaP, C4-2B, and PC-3) were authentic and free of mycoplasma as described previously[59] and cultured in monolayer in a humidified 5% CO2 incubator at 37°C. The cell lines were transfected with human MMP7 plasmid (Addgene, Cambridge, MA, USA) using Lipofectamine® 2000 Transfection Reagent (Invitrogen, Grand Island, NY, USA) according to the manufacturer’s instructions. Positive cells were selected with 1 mg/ml Geneticin® (Invitrogen) for two weeks and the cells stably overexpressing MMP7 were named LNCaP-MMP7, C4-2B-MMP7, PC-3-MMP7, respectively. The cells were treated without or with 20 ng/ml recombinant human IL-17 (R&D Systems) for 0.5, 2, and 8 hours. In another experiment, the cells were first transfected with 5 nM control siRNA (catalog #4390843, Silencer® Select Negative Control) or 5 nM siRNA targeting MMP7 (catalog #4392420, Silencer® Select Pre-designed siRNA) (Fisher Scientific) using Lipofectamine® 2000 Transfection Reagent (Invitrogen) according to the manufacturer’s instructions; 24 hours later, the cells were treated without or with IL-17 for 2 and 8 hours.

Western Blot Analysis

Mouse prostate tissues and cultured cells were homogenized for protein isolation. For IL-17-treated cells, nuclear protein was isolated for detecting nuclear β-catenin, using ProteoExtract® Subcellular Proteome Extraction kit (Millipore) according to manufacturer’s instructions. Whole cell lysate was used for detecting other proteins as described previously.[60]

Invasion Assay

Invasion assay was performed using Corning® BioCoat™ Matrigel® Invasion Chambers (Corning Inc., Corning, NY) following the manufacturer’s instructions. Approximately 2 × 105 cells (human prostate cancer cells and mouse prostate cancer cells isolated from Mmp7 and Mmp7mice) were seeded in the upper chamber in serum-free medium in triplicate wells per group, while the lower chamber contained medium with 10% FBS; 48 and 72 hours later, non-invaded cells were removed from the upper chamber with a cotton swab; the cells invaded through the Matrigel®-coated porous membrane were fixed with methanol, stained with 0.5% crystal violet, and counted under a microscope.

Immunoprecipitation

Whole cell lysates were used for immunoprecipitation (IP) and Western blot (WB) using control IgG, anti-β-catenin, or anti-E-cadherin antibodies, according to our previously described protocol.[61]

The Cancer Genome Atlas (TCGA) Data Analysis

Complete Clinical Dataset (level 2) and RNASeqV2 gene expression dataset (level 3) under the category of human prostate adenocarcinoma (PRAD) were downloaded from the TCGA website https://tcga-data.nci.nih.gov/tcga/. TCGA aligned the sequencing reads using the MapSplice software[62] and quantified expression with upper quartile normalized RNASeq by Expectation Maximization (RSEM) count estimates.[62,63] The mRNA expression dataset contains 52 normal prostate tissues and 497 primary prostate tumors. Samples without IL-17 expression data were first filtered out and the remaining normal prostates (N = 29) and prostate tumors (N = 183) were analyzed. Log2 transformation was applied to gene expression levels. The differences of IL-17 and MMP7 expression levels between normal prostates and primary tumors were evaluated using the Student’s t test. Pearson’s correlation analysis was performed between the expression levels of IL-17 and MMP7.

Statistical Analysis

Statistical analysis was performed using the R package “stat” (https://www.r-project.org/). All in vitro experiments were replicated three times. Sample sizes were selected on the basis of preliminary results to ensure an adequate power and animal number estimate was based on our previous studies.[16,17] The examiners were blinded to the grouping of animals when assessing the outcomes. Quantitative data are presented as mean ± standard error of the mean (SEM, error bar). Comparisons of the GU-bloc weight and other quantitative data were analyzed using Student’s t test (two-sided) with the assumption of normal distribution of data and similar variance between the groups. The χ2 test was used to compare the incidences of PIN and invasive adenocarcinoma. Statistical significance was defined as P < 0.05.