Elevated expression of Par3 promotes prostate cancer metastasis by forming a Par3/aPKC/KIBRA complex and inactivating the hippo pathway

Prostate cancer cell lines PC3, DU145 and normal prostate epithelial cell line PNT1B were purchased from the American Type Culture Collection (ATCC, Rockville, MD, USA). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) with 10% fetal bovine serum (FBS, Gibco) and maintained at 5% CO₂ at 37 °C.

The Par3 knock down and control plasmids were obtained from Origene (Rockville, MD, USA). Plasmids were transfected with Lipofectamine3000 (Thermo Fisher Scientific). Puromycin (0.5 μg/ml) was used for selecting stable Par3 knockdown and relevant control subclones, which were named as PC3-shPar3, PC3-con, DU145-shPar3 and DU145-con respectively. The lentiviral vector expressing one of the Par3 isoforms (NM_019619.3, 150 kDa) or a non-phosphorylatable YAP mutant, YAP(S127A); and control lentiviral vector were constructed and packaged by Genomeditech Comp (Shanghai, China). 2 × 10⁶ cells were seeded in 6-well plates and infected using relevant lentiviral vector (MOI = 10 for each) concomitant with 5 μg/ml polybrene. Western blot was employed to detect Par3 expression levels. IF was employed to detect YAP subcellular location. MOI: multiplicity of infection.

In vitro migration and invasion assays were performed using 24-well Cell Migration and Invasion Assay kit (Cell BioLabs, San Diego, CA, USA), according to the manufacturer’s instructions. Briefly, after serum starvation for 24 h, 1 × 10⁵ cells were suspended in 100 μl DMEM basic medium and seeded in the upper chamber, and 700 μl DMEM medium with 10% FBS was added to the lower chamber. After incubation for 6 h (for migration assay) or 8 h (for invasion assay), cells on the lower surface of the membrane were fixed with 4% paraformaldehyde, stained with 0.2% crystal violet, photographed and counted under a microscope in five random fields.

Six-week-old male BALB/C athymic nude mice (SLAC, Shanghai, China) were housed and manipulated according to the protocols approved by the Renji Hospital Medical Experimental Animal Care Commission. For orthotopic inoculation, 1 × 10⁶ cells were first injected under the subcutaneous and tumor mass was harvested after 4 weeks. Xenografts were digested with collagenase IV for 30mins, 0.05% Trypsin for 10mins and then normalized by DMEM medium with 10% FBS to collect cells for orthotopic transplantation. Cell suspension (1 × 10⁵ cells in 20 μl) was mixed with 20 μl matrigel and injected into the left ventral anterior of mouse prostate. PET-CT (Siemens Inveon) detection was performed using ¹⁸F–FDG (750 μCi/100 g) by intravenous injection 7 weeks after orthotopic inoculation. Orthotopic tumor growth and potential tumor metastasis were evaluated in living animals by the absorption of ¹⁸F–FDG. Mice were sacrificed 1 week after PET-CT detection. Orthotopic tumor mass and metastatic nodes were collected for H&E and immunofluorescence staining.

Investigation has been conducted in accordance with the ethical standards and according to the Declaration of Helsinki and national and international guidelines. Human tissues used in this study were reviewed and approved by the Committee for Ethical Review of Research Involving Human Subjects at Renji Hospital. PCa (n = 14 for qRT-PCR and western blot, n = 2 for immunohistochemical) and normal tissues (n = 7 for qRT-PCR and western blot, n = 1 for immunohistochemical) were obtained from the Renji Biobank, Shanghai Jiao Tong University School of medicine [25]. Written informed consent was obtained from all patients.

Cells were seeded on cover slide placed in 24-well plate and cultured in DMEM medium supplemented with 10% FBS and maintained at 5% CO₂ at 37 °C for 48 h. Adherent cells on cover slide were then fixed in 4% paraformaldehyde for 30mins at room temperature. Cells were washed with PBS and blocked with 10% normal goat serum (Vector, Burlingame, CA, USA) for 1 h at room temperature for IF staining. Tissues were fixed with 4% paraformaldehyde for 24 h and embedded in paraffin. H&E staining was performed by Runnerbio biotech. Comp (Shanghai, China). Paraffin sections were dewaxed in xylene for 5 min, sequentially hydrated in ethanol with concentrations of 100%, 95%, 85% and 70% for 3 min each, respectively, and rinsed three times in water. Sections were treated with disodium-hydrogen phosphate-2-hydrate for 15 min to inactivate endogenous peroxidase and then blocked with 10% normal goat serum for 1 h at room temperature for IF and IHC staining. After serum blocking, both cells and sections were incubated with relevant primary antibodies (1:200, diluted in PBS with 1% normal goat serum) overnight at 4 °C. For IF staining, cells and sections were washed with PBS for three times (10mins each) and then incubated with Alexa Fluor-546 or Alexa Fluor-488 conjugated secondary antibody (Thermo Fisher Scientific) for 1 h at room temperature. Cells and sections were washed with PBS for three times again before mounted with DAPI containing mounting medium (Vector). For IHC staining, sections were washed with PBS for three times (10 min each) and then incubated with horseradish peroxidase-conjugated secondary antibody (Vector) for 1 h at room temperature. Sections were washed with PBS for three times again before DAB staining (Sangon Biotech, Shanghai, China) and hematoxylin counterstaining (Beyotime, Shanghai, China). Primary antibody for YAP (#4912) was purchased from Cell Signaling Technology (Beverly, MA, USA) and for Par3 (07–330) and human nuclei (MAB1281) from Millipore (Billerica, MA, USA). IF and IHC staining was visualized under a microscope (Leica DFC420C) and images were merged using the ImageJ software.

Total RNA was extracted using Trizol (Thermo Fisher Scientific) according to the method previously described [26]. For mRNA relative expression analysis, total RNA was reversed transcribed into cDNA with Prime-Script RT kit (Takara, Shiga, Japan) and amplified with SYBR-Green Real-time PCR Master Mix (Applied Biosystems, Thermo Fisher Scientific). The mRNA expression level of GAPDH was used as an internal normalization control. Comparative quantification was performed by using the 2^-ΔΔCt method. All primers are available in the Additional file 1: Table S1.

For western blot, 1 × 10⁷ cultured cells were lysed with nuclear and cytoplasmic protein extract kit (Cat. No. PP1082, Bioteke, Shanghai, China) or membranous and cytoplasmic protein extract kit (Thermo Fisher Scientific) supplemented with a Complete Protease Inhibitor Cocktail and a final concentration of 2 mM PMSF (Thermo Fisher Scientific) to isolate and purify nuclear vs. cytoplasmic or membrane vs. cytoplasmic protein respectively. Clinical normal or tumor tissues were lysed with RIPA (Millipore). The concentration of total proteins was measured by BCA (bicinchoninic acid) kit (Thermo Fisher Scientific) respectively. Forty micrograms of total proteins from relevant samples above were separated by 10% SDS-PAGE and then transferred to polyvinylidene fluoride membrane (Millipore) under the same experimental condition. The membrane was blocked with TBST containing 5% BSA for 1 h, incubated with specific primary antibodies overnight at 4 °C and then probed with HRP-conjugated secondary antibody at room temperature for 1 h. Proteins were detected with HRP Substrate (Millipore) and photographed using ECL (electro-chemiluminescence) detection instrument (Thermo Fisher Scientific).

For immunoprecipitation, 1 × 10⁷ cultured cells were lysed with RIPA (Millipore) lysate and 100 μg total protein were incubated on ice with 5 μg specific antibodies and 40 μl protein G-sepharose beads (Thermo Fisher Scientific) overnight. Beads-antibody complex was washed three times with Chilled PBS and boiled for 10 min to obtain protein supernatant captured by the protein G-sepharose beads. Antibodies for western blot and co-IP analysis are listed in the Additional file 1: Table S2.

Cell extraction was prepared using Chromatin IP kit (Cell Signaling Technology) according to the manufacturer’s protocol. Briefly, 1 × 10⁷ cells were cross-linked with 37% formaldehyde, collected and digested to produce chromatin fragments for incubation with IgG (Cell Signaling Technology) or specific antibodies for Tead1 (ab133533, Abcam, Cambridge, MA, USA), Tead2 (H00008463-M01A, Abnova, Taipei, Taiwan), Tead4 (ab58310, Abcam) and YAP (#4912, Cell Signaling Technology) respectively. ChIP DNA was amplified and analyzed by qPCR and sequencing. Relative enrichment of specific factors was assessed by the formula provided in the protocol. ChIP primers are listed in the Additional file 1: Table S3.

Independent Student’s t-test and analysis of variance (ANOVA) were used to compare the differences between two groups. Difference of Par3 expression between normal and PCa tissues was analyzed using Prism GraphPad5 (GraphPad Software, La Jolla, CA, USA). All data were represented as mean ± SD from triplicate experiments. Results were considered statistically significant when p < 0.05.

In order to investigate the relationship between Par3 expression and PCa, we analyzed two independent datasets from Oncomine database (Tomlins et al. [27] and Luo et al. [28]). Both of the datasets exhibited a significant upregulation of Par3 expression in PCa tissues compared to normal controls, indicating that enhanced expression of Par3 is positively associated with PCa (Fig. 1a). Consistently, analysis of the TCGA PRAD database also revealed that Par3 expression level was elevated in PCa tissues (Additional file 2: Figure S1). To determine whether expression of Par3 is further upregulated in metastatic prostate tumors, we compared Par3 expression in primary tumor tissues derived from patients with (n = 8) or without metastasis (n = 6) to that in prostatic tissues (n = 7) derived from normal population at both mRNA and protein levels (Fig. 1b, c, Additional file 3: Table S1). We found that expression of Par3 was enhanced in all tumor samples compared to normal tissues. In addition, expression of Par3 was further enhanced in tumor tissues from patients with metastasis, indicating a positive correlation of Par3 expression with metastasis (Fig. 1b, c). By IHC staining in clinical samples, we confirmed that expression of Par3 was enhanced in tumor tissues especially in those with metastasis (Fig. 1d, Additional file 3: Table S2). These observations together indicate that enhanced expression of Par3 is positively correlated with human PCa metastasis.

In order to provide direct experimental evidence that Par3 regulates PCa metastasis, we performed a Par3 knockdown experiment in PCa cell lines. For this purpose, we first measured endogenous expression of Par3 in normal prostatic cell lines PNT1B and PCa cell lines PC3 and DU145. We found that compared to PNT1B cells, Par3 expression was significantly elevated in PC3 and DU145 cells, both of which possess migratory capability (Additional file 2: Figure S2a). Thus we established Par3 stable knockdown subclones or controls with these two cell lines respectively to evaluate the effect of Par3 on migration in vitro and metastasis in vivo (Additional file 2: Figure S2b, Figure S2c). In addition, we also established Par3-overexpressing PC3 cells (PC3-Par3OE) for further confirmation of relevant observations (Additional file 2: Figure S2d).

To determine whether downregulation of Par3 affects PCa cell migration and invasion in vitro, we performed transwell assays. As shown in Fig. 2a , the migratory capability was decreased to about 50% in PC3 (50.6 ± 3.36 cells/field in PC3-shPar3 vs 103.4 ± 8.68 cells/field in PC3-con) and to about 46% in DU145 (42.6 ± 6.95 cells/field in DU145-shPar3 vs 92.2 ± 13.71 cells/field in DU145-con) after Par3 knockdown respectively (Fig. 2a, Additional file 2: Figure S3). The invasive capability was also decreased to about 48% in PC3 (37.8 ± 5.57 cells/field in PC3-shPar3 vs 78.1 ± 7.42 cells/field in PC3-con) and to about 38% in DU145 (24.4 ± 5.95 in DU145-shPar3 vs 64.2 ± 5.73 in DU145-con) after Par3 knockdown respectively (Fig. 2a, Additional file 2: Figure S3). Conversely, after overexpession of Par3 in PC3 cells by lentiviral infection, both migratory and invasive capabilities were improved more than 1.5 folds (Fig. 2b). These results together indicate that stable Par3 knockdown inhibits the migration and invasion of PCa cells in vitro.

To further confirm whether Par3 knockdown leads to an inhibition on metastasis of PCa cells in vivo, we conducted orthotopic implantation experiments using PC3-shPar3 vs PC3-con cells and DU145-shPar3 vs DU145-con cells. Relevant cells were first inoculated subcutaneously in 6-week old male BALB/C athymic nude mice and then dissociated from tumor mass 4 weeks later prior to orthotopic implantation via injection of 1 × 10⁵ cells into the left anterior of prostate (Fig. 3a). Orthotopic tumor growth and micro- or macro-metastases were evaluated in living animals by PET-CT 7 weeks post-implantation. We found a significant inhibition on distant liver metastasis after orthotopic implantation of PC3-shPar3 (#4 to #6) compared to PC3-con cells (#1 to #3, Fig. 3b). Serious invasion to peritoneal lymph nodes was also observed in one of the mice after inoculation of PC3-con cells but not PC3-shPar3 cells (#1 vs. #4 in Fig. 3b and Additional file 2: Figure S4a). The paravertebral lymph nodes invasion and kidney metastasis were found to be inhibited when using DU145-shPar3 cells for orthotopic implantation (Additional file 2: Figure S4b). Mice were sacrificed 1 week after PET-CT analysis and the orthotopic tumor growth, lymph nodes invasion and metastasis in the liver or kidney were examined and then confirmed by H&E staining (Fig. 3c, d, Additional file 2: Figure S4c). Orthotopic grafts were dramatically larger along with tissue adhesions and necrosis after inoculation of PC3-con or DU145-con cells (#1 to #3 in Fig. 3c left panel and #1, #2 in Additional file 2: Figure S4c) but restricted in the anterior prostate after inoculation of PC3-shPar3 or DU145-shPar3 cells (#4 to #6 in Fig. 3c left panel and #3, #4 in Additional file 2: Figure S4c), indicating that stable knockdown of Par3 inhibits orthotopic tumor growth. A large number of micro-metastatic nodes with several large ones were observed not only on the surface but also inside of the liver in PC3-con cells inoculated mice (Fig. 3c right panel #1 to #3, Fig. 3d), but only a limited number of nodes were found in PC3-shPar3 groups (Fig. 3c right panel #4 to #6, Fig. 3d). We further compared Par3 expression in metastatic nodes in livers of PC3-shPar3 to that of PC3-con inoculated mice, by co-staining Par3 with anti-human nuclei antibody to distinguish the metastatic human PCa cells from mouse liver cells. While maintained in those from control, Par3 expression was significantly suppressed in the liver metastatic nodes from PC3-shPar3 inoculated mice (Fig. 3e). Collectively, these results indicate that knockdown of Par3 significantly inhibits PCa cell migration in vitro and metastasis in vivo.

Given that Hippo-YAP pathway has been reported to be involved in tumor metastasis [24], we herein suspect that Par3 might exert its effects on tumor metastasis via the Hippo-YAP pathway as a novel candidate upstream regulator. To provide evidences for this possibility, we investigated expression of LATS1/2, the major component of the Hippo-YAP pathway, in PC3-shPar3 cells and made a comparison to that in PC3-con cells. We found that after stable knockdown of Par3, expression levels of LATS1/2 were similar to that of the control group, but the level of phosphorylated LATS was significantly increased in PC3-shPar3 cells, indicating an activation of the Hippo pathway (Fig. 4a). As expected, the phosphorylation of YAP was consequentially enhanced (Fig. 4a) and its arrest in cytoplasm was increased in PC3-shPar3 cells (Fig. 4b). As a result, decreased nuclear localization of YAP was observed in PC3-shPar3 cells in vitro (Fig. 4b) as well as in PC3-shPar3 cells derived orthotopic grafts in vivo compared to relevant controls respectively (Fig. 4c). Expressional level of intranuclear YAP was also significantly decreased in liver metastatic nodes derived from PC3-shPar3 inoculated mice compared to the control mice (Fig. 4d). By IHC staining, we also observed that nuclear translocation of YAP was enhanced in tumor tissues, especially in those with metastasis (Fig. 4e, Additional file 3: Table S2). All these findings indicate that knockdown of Par3 inhibits PCa metastasis via activation of the Hippo pathway and antagonizing YAP through phosphorylation and arrest of YAP in the cytoplasm. To provide additional support for this notion, we also compared expression of LATS1/2 and the activation of LATS and YAP in PC3-Par3OE cells with the control cells. We found that forced overexpression of Par3 still had no effect on LATS1/2 expression but repressed the phosphorylation of both LATS and YAP, and eventually enhanced the nuclear translocation of YAP (Additional file 2: Figure S5a). Similar expression patterns of LATS1/2, phosphorylated LATS and intranuclear YAP were observed in DU145-shPar3 vs DU145-con cells in vitro (Additional file 2: Figure S5b,c), suggesting again that knockdown of Par3 can inhibit PCa cell migration in vitro and metastasis in vivo, associated with activation of the Hippo pathway.

In order to further determine whether knockdown of Par3 inhibits PCa metastasis by activating the Hippo pathway and attenuating the nuclear translocation of non-phosphorylated YAP as a response, we carried out a rescue experiment by exogenous overexpression of YAP(S127A), a non-phosphorylatable YAP mutant [29], in PC3-shPar3 cells to evaluate its effect on restoration of metastasis against Par3 knockdown (Additional file 2: Figure S6a). As expected, after overexpression of YAP(S127A), the effect of Par3 knockdown on inhibition of primary tumor growth and liver metastasis was significantly reversed in vivo (Additional file 2: Figure S6b, c, d). Moreover, increased nuclear translocation of YAP was also observed in metastatic nodes in liver, indicating no response to the activation of Hippo pathway by Par3 knockdown (Additional file 2: Figure S6e). Taken together, our results revealed that elevated expression of Par3 promotes PCa metastasis by inactivation of Hippo pathway and enhanced nuclear translocation of oncogenic YAP.

In order to demonstrate how knockdown of Par3 activates the Hippo pathway, we paid attention to KIBRA, which was recently identified as an activator of the Hippo pathway by forming a KIBRA/Merlin/FRMD6 complex to interact with and sequentially to phosphorylate Hippo components, LATS1 and LATS2 [30, 31]. In addition, it has been demonstrated that KIBRA can co-localize and interact with the PAR complex in mammalian epithelial cells [32, 33]. Based on these reports, we hypothesize that elevated expression of Par3 may inactivate the Hippo-YAP pathway via forming a complex with KIBRA to interfere its function on phosphorylation of the Hippo pathway.

To test this possibility, a co-immunoprecipitation (co-IP) assay was carried out to determine potential interactions among KIBRA, the PAR complex and the Hippo components in non-tumorigenic PNT1B prostatic cells and PC3 and DU145 PCa cell lines respectively (Fig. 5a, b, c). Our results revealed that besides interaction with aPKC to form a canonical PAR complex in both normal prostatic and PCa cell lines, a non-canonical interaction of Par3 with KIBRA was significantly enhanced in PC3 and DU145 cells (Fig. 5a). Consistently, we observed an enhanced interaction of KIBRA with Par3 and aPKC and an attenuated interaction with its canonical partners Merlin and FRMD6 in both PC3 and DU145 cells but not PNT1B cells (Fig. 5b). Furthermore, although KIBRA was found to interact with LATS1/2 in all of the three cell lines, the interaction was significantly weakened in both PC3 and DU145 cells (Fig. 5c). As a result, the phosphorylation of both LATS and downstream YAP was repressed and the nuclear translocation of YAP was consequentially enhanced in both PC3 and DU145 cells (Fig. 5d). These results raised a possibility that elevated expression of Par3 may sequestrate KIBRA from its partners Merlin and FRMD6, and thus dissociate canonical KIBRA/Merlin/FRMD6 complex so that the Hippo pathway is inactivated due to impaired function of the KIBRA/Merlin/FRMD6 complex on the phosphorylation of LATS [30, 31]. In order to further investigate this possibility, we repeated the above co-IP assay in either Par3 knockdown (PC3-shPar3) or overexpression (PC3-Par3OE) PC3 cells to identify the sequestration of KIBRA by Par3 (Fig. 5e, f).

Considering that Par3 can be localized in both the cytoplasm and the membrane of cancer cells and that cytoplasmic Par3 may function differently from membranous Par3 [34], we carried out the following experiments. First, we determined expression levels of Par3 in both the cytoplasm and the membrane of PCa cell lines vs. control cells by IF staining. As shown in Figure S7a, an elevated expression of Par3 was observed in both the cytoplasm and the membrane in PC3 and DU145 cells compared to PNT1B cells (Additional file 2: Figure S7a). Similar results were obtained in 35 PCa clinical samples compared to normal samples from a tissue chip (HProA100PG01, Shanghai Outdo Biotech Co., data shown in Additional file 2: Figure S7b and Additional file 4: Table S1). Next, we separated membranous protein extraction from cytoplasmic protein extraction and checked the interaction of membranous or cytoplasmic Par3 with KIBRA respectively. As shown in Fig. 5e, knockdown of Par3 decreased the interaction of both membranous and cytoplasmic Par3 with KIBRA. As a response to the release of KIBRA from Par3, the formation of canonical KIBRA/Merlin/FRMD6 complex was restored (Fig. 5e). In contrast, overexpression of Par3 enhanced the sequestration of KIBRA by both membranous and cytoplasmic Par3 so that the interaction of KIBRA with Merlin and FRMD6 was further impaired (Fig. 5f). Therefore, the findings together indicate that both of membranous and cytoplasmic Par3 have the same function on the regulation of the Hippo pathway. They both can sequestrate KIBRA to impair the formation of the canonical KIBRA/Merlin/FRMD6 complex, which in turn causes a decrease of the phosphorylation of LATS and YAP to promote metastasis.

Given that Par3 promotes PCa metastasis via an inactivation of the Hippo pathway and a consequential enhancement of nuclear translocation of YAP by sequestration of KIBRA from its canonical KIBRA/Merlin/FRMD6 complex, we attempt to investigate potential downstream effectors of YAP, which are associated with metastasis. It has been known that YAP activates gene transcription by conjunction with the TEAD transcription factors family (mainly refer to TEAD1, 2 and 4, which share the same binding site) [14, 35‐37]. Thus, we attempted to screen and identify potential TEAD binding sites by bioinformatics analysis using the PROMO online software [38] and chromatin immunoprecipitation (ChIP) assay in the promoter region of well-known pro-metastatic genes such as the matrix metalloproteinase (MMP) family, Zeb1, et al. We found that TEAD binding sites were harbored in the promoter region of MMP1, MMP9, Zeb1, Snail1 and Twist1 respectively (Fig. 6a-e), indicating that these pro-metastatic genes could be activated at the transcriptional level by YAP. Consistent with the results above, expression levels of these five genes was downregulted after Par3 knockdown, as a response to the decreased nuclear translocation of YAP (Fig. 6f, Additional file 2: Figure S8a). Conversely, opposite results were observed after Par3 overexpression in these cells (Additional file 2: Figure S8b). Collectively, the data demonstrate that elevated expression of Par3 increase expression of pro-metastatic genes via promotion of nuclear YAP-mediated activation of TEAD transcription factors.