Histone Demethylase JMJD1A Promotes Tumor Progression via Activating Snail in Prostate Cancer

This article is featured in Highlights of This Issue, p. 669

Prostate cancer is one of the most commonly diagnosed malignancies and the second leading cause of cancer-related death in American men (1). Androgen deprivation therapy (ADT) remains the primary clinical treatment for patients in the early stage of prostate cancer (2). However, most prostate cancer tumors become resistant to ADT and progress to a lethal stage called castration-resistant prostate cancer (CRPC), which is characterized by aggressive growth and distal organ metastasis, and is incurable (3). Current therapies for CRPC, including treatment with new-generation androgen receptor (AR)-pathway inhibitors (abiraterone or enzalutamide) and chemotherapy extend patient life by only a few month and benefit a little for the patients with metastatic CRPC (4, 5). Thus, exploring the underlying mechanism in metastasis of prostate cancer is crucial to develop therapeutic targets to treat or prevent prostate cancer metastasis.

H3K9 methylation is the mark of heterochromatin and often functions as a repressive epigenetic modification to regulate gene expression (6). JMJD1A (also named KDM3A) demethylates H3K9me1 and H3K9Me2 to increase gene transcription, and with wide functional roles in spermatogenesis (7), sex determination (8), stem cell renewal (9), and cancer (6, 10). JMJD1A overexpression has been observed in a subset of primary human cancers including liver cancer and sarcoma (11). We previously found that JMJD1A promoted prostate cancer cells' growth as an AR coactivator or by upregulating levels of AR-V7, an AR splice variant 7 (6, 11). However, the tumor-promoting role and molecular mechanisms of JMJD1A in prostate cancer remains largely unknown.

Metastasis is a process by which cancer cells invade and migrate, and can be induced by regulators such as twist, Snail, and slug (12). Snail, a snail family zinc finger protein, has been found to promote cell survival and movement through induction of epithelial–mesenchymal transition (EMT) or independent of EMT (13–15). Snail has been found to overexpress in a set of cancers, one study reported that overexpression of Snail was related with poor prognosis in patient with breast cancer (16). Moreover, nuclear Snail expression was found to be related to tumour progression (17, 18). In prostate cancer, several studies have reported that Snail expression was elevated in metastatic prostate cancer cell lines, Snail transcriptional regulatory network was important for the progression and metastasis of cancer (19, 20) and Snail overexpression could induce EMT in prostate cancer cells (21). Importantly, the mechanism for upregulating Snail in prostate cancers is still unclear.

Here, we found that JMJD1A protein and mRNA levels are elevated in prostate cancer compared with normal tissues. Furthermore, JMJD1A overexpression significantly promotes prostate cancer cell proliferation, migration, and invasion, while JMJD1A knockdown exhibits the opposite effects. More importantly, these biologic effects mediated by JMJD1A might be dependent on the Snail transcription factor. Identifying these mechanisms highlights JMJD1A is closed with prostate cancer development through Snail, suggesting that JMJD1A–Snail axis could be targeted for prostate cancer therapies.

RV1 and LNCAP cells were purchased from ATCC and cultured in DMEM (Gibco, 110 Life Technologies) supplemented with 10% FBS (Biological Industries), 100 U/mL penicillin/streptomycin (Gibco, Life Technologies). Cells were grown at 37°C with 5% CO₂. All cell lines were authenticated by short tandem repeat analysis at China Center for Type Culture Collection and all the cell lines were tested and authenticated by karyotyping analysis on January 3, 2018, and confirmed by National Infrastructure of Cell Line Resources of China. Cell lines were tested Mycoplasma free. Cells used for experiments were between 20–30 passages from thawing. Flag-JMJD1A and Myc-JMJD1A were described previously (6). Flag-Snail (Plasmid #16218) was purchased from Addgene.

JMJD1A (12835-1-AP) was purchased from Proteintech. EMT marker antibodies (#9782) were purchased from Cell Signaling Technology. H3K9me2 (07-441) was purchased from EMD Millipore. H3K9me1 (ab9045) antibodies, anti-Ki67 antibody (ab15580), and anti-Cleaved Caspase-3 antibody (ab2302) were purchased from Abcam. Anti-β-actin antibodies (A1978) were purchased from Sigma.

The indicated cells (RV1, 1 × 10⁴ cells and LNCAP, 1.5 × 10⁴ cells) were mixed with agar to a final concentration of 0.4% and layered on top of 0.8% agar in 6-well plates. Cells were maintained in the normal growth media. Triplicate plates were incubated at 37°C for 3 weeks. Photographs of colonies were obtained using a microscope. The independent experiments were repeated at least three times. The results are shown as means ± SD (n = 3), and ANOVA was used for statistical analysis.

Cells were cross-linked using 1% formaldehyde for 10 minutes at room temperature, and then quenched with 5 mol/L glycine. Cell nuclei were extracted with buffer A (10 mmol/L HEPES, pH 7.4, 10 mmol/L KCl, 1.5 mmol/L MgCl₂, 10% glycerol, 1 mmol/L DTT, and 1× protease cocktail). Nuclear extracts were solubilized in the lysis buffer (50 mmol/L Tris-HCl, pH 8.0, 1% SDS, 10 mmol/L EDTA, and 1× protease cocktail), and sonicated to obtain 500-bp chromatin fragments. Chromatin (100 μg) was 10-fold diluted and incubated with 5 μg of antibodies overnight followed by incubation with 30 μL of protein A/G beads for 4 hours at 4°C. After four washes, cross-linking was reversed, and DNA was purified using spin columns and subjected to qPCR analysis for the indicated genes. The primers are listed in Supplementary Materials and Methods. The chromatin immunoprecipitation (ChIP) assays were performed in triplicate and independent experiments were repeated at least three times. Data were calculated as percentage of input and presented as means ± SD (n = 3), and ANOVA was used for statistical analysis.

Prostate tumor tissue samples and matched adjacent nontumoral prostate tissue samples were collected at The Second Clinical Medical College of Jinan University (Guangdong, P.R. China). Tissue sample collection was approved by the medical ethics committee of Shenzhen People's Hospital (Shenzhen, P.R. China). Tissue microarray chips containing adjacent nontumoral prostate tissue samples and prostate tumor tissue samples were obtained from Shanghai OUTDO Biotech Co., Ltd.. IHC staining and quantification were conducted as described previously (6). The score of JMJD1A expression at the protein level was classified semiquantitatively as follows: no staining, 0 points; weak staining, 1 point; moderate staining, 2 points; and strong staining, 3 points; while 2 or 3 points were defined as high expression. The proportion of tumor cells was graded from 0 to 2: 0 points (<5%), 1 point (5%–50%), and 2 points (>50%). The sum of scores represented the expression levels: 0 to 2 points as low expression and 3 to 4 points as high expression. The mean score from two pathologists was used as the final immunostaining score. The IHC score was calculated as described previously (6). The χ² test was used for statistical analysis of JMJD1A expression.

Seven-week-old female BALB/c nude mice were purchased from Charles River Laboratories in China and housed under pathogen-free conditions in the animal experiment center of The Second Clinical Medical College of Jinan University (Shenzhen, China). RV1 cells (1 × 10⁶) stably expressing the control vector, stably expressing JMJD1A, control short hairpin RNA (shRNA), or shJMJD1A#1 were injected subcutaneously into female BALB/c nude mice. Every 4 days, the tumor volume and weight were measured following the standard protocol. ANOVA test was used to analysis the data. For the blinding procedures, two researchers in each group performed all the mice experiments. All protocols involving live mice were approved by the Animal Care and Use Committee of The Second Clinical Medical College of Jinan University. Mice were sacrificed after subcutaneous injection on day 22 (control vector and stably expressing JMJD1A) or day 26 (control shRNA and shJMJD1A#1).

Data are represented as the mean ± SEM of three independent experiments for cell proliferation and cell clone assays. Data are represented as the mean ± SD of 5 mice for the tumor growth study in xenograft. The two-tailed, unpaired Student t test, ANOVA, and χ² test were utilized for the statistical analyses (*, P < 0.05; **, P < 0.01). Statistical analysis was performed using Prism 5.0 (GraphPad Software). Data were considered significant when P < 0.05.

We previously found that JMJD1A played a key role in AR-FL coactivation and AR-V7 alternative splicing (6, 11). Here, we analyzed the data from Oncomine and found JMJD1A mRNA levels were significantly upregulated in prostate cancer tissues compared with normal tissues (Fig. 1A). Surprisingly, we found elevated JMJD1A mRNA was positively associated with more advanced tumor (T) stage in prostate cancer specimens (Fig. 1B). To confirm this finding, JMJD1A levels in 14 pairs of prostate tumor tissue samples and matched adjacent nontumoral prostate tissue samples were analyzed. As shown in Fig. 1C,–E, JMJD1A protein levels and mRNA levels were elevated in most prostate cancer tissues. Supplementary Table S1 shows the clinical pathologic variables and JMJD1A expression in 14 prostate cancer cases for detection by Western blot and RT-PCR analyses. To confirm the relevance of our findings to prostate cancer tissues, we evaluated the expression of JMJD1A in a human prostate cancer tissue microarray (TMA) containing 77 tumor samples and 25 adjacent nontumoral prostate tissue samples. As shown in Fig. 1F, JMJD1A was primarily nuclear localized. IHC results were consistent with prior observations that JMJD1A is significantly increased in prostate cancer samples (Fig. 1G). On the basis of patient data from TMA, we found that a high JMJD1A expression was related with a higher preoperative PSA level (P = 0.00589), Gleason score (P = 0.028), and or lymph node metastasis (P = 0.00585). The results are shown in Supplementary Table S2.

To investigate the function of JMJD1A in prostate cancer progression, JMJD1A was overexpressed in two prostate cancer cell lines. As shown in Fig. 2A, Western blot results revealed JMJD1A was overexpressed in RV1 and LNCAP cells. Next, we evaluated the role of JMJD1A in prostate cancer proliferation. We found, using MTT assays, that the cells of JMJD1A overexpression promoted cell growth (Fig. 2B). To further determine the role of JMJD1A in prostate cancer proliferation, we overexpressed JMJD1A in prostate cancer cells and measured the cell proliferation by colony formation assay and soft-agar assay. Concordantly, JMJD1A overexpression also increased the number and size of cell colonies, as determined by colony formation assay or soft-agar assay (Fig. 2C and D). These data demonstrate that JMJD1A overexpression promotes prostate cancer cell proliferation.

Given that overexpression of JMJD1A was associated with T stage in prostate cancer (Fig. 1B) and metastasis seriously limits the treatments available for prostate cancer. Thus, we assessed whether JMJD1A plays a role in prostate cancer progression. Overexpression of JMJD1A markedly stimulated prostate cancer cell migration by transwell assays (Fig. 2E) and promoted cell invasion through Matrigel (Fig. 2F). These results suggest that JMJD1A participates in the regulation of prostate cancer cell motility and invasion, which is consistent with the functions of JMJD1A that we observed.

To study JMJD1A's role in prostate cancer, we silenced JMJD1A in RV1 cells and LNCAP cells by two different JMJD1A shRNAs. Our results revealed that the shRNAs effectively knocked down JMJD1A in sh-JMJD1A cells compared with control shRNA-pLKO.1 (Fig. 3A). To determine whether JMJD1A knockdown inhibits cell proliferation, we performed an MTT assay to assess cell proliferation. Our results found that silencing of JMJD1A significantly decreased RV1 and LNCAP cell growth (Fig. 3B). We further found JMJD1A knockdown decreased the number and size of cell colonies (Fig. 3C) and reduced cell colony formation using a soft-agar assay (Fig. 3D). To further investigate the role of JMJD1A in prostate cancer metastasis, JMJD1A was silenced in two prostate cancer cell line models. As shown in Fig. 3E and F, we demonstrated that silencing of JMJD1A markedly abolished RV1 and LNCAP cell migration and invasion. Taken together, our studies reveal that JMJD1A downregulation decreases survival of prostate cancer cells and inhibits cell migration and invasion in vitro.

Many reports showed that JMJD1A affects cancer metastasis (22–26). However, the mechanism of the histone demethylase JMJD1A in prostate cancer metastasis is poorly understood. EMT is a crucial step during tumor metastasis. E-cadherin and vimentin were widely used as classic EMT markers (27–29), while Snail, ZEB1, and Slug are the key transcription factors involved in the regulation of metastasis (30–32). Here, we revealed that JMJD1A knockdown dramatically reduced Snail, but not others (Fig. 4A and B). Then, we asked whether Snail was able to reverse the effects caused by JMJD1A silencing in two prostate cancer cells. As shown in Fig. 4C,–F and Supplementary Fig. S2, depletion of JMJD1A dramatically inhibited cell proliferation, migration, and invasion, while the reexpression of Snail was able to reverse the effects. Collectively, our findings show that Snail is a key downstream effector of JMJD1A in the regulation of prostate cancer progression.

To further investigate how JMJD1A reduces Snail expression, the influence of JMJD1A on the Snail mRNA level was determined. First, we found that JMJD1A overexpression increased Snail mRNA levels (Supplementary Fig. S1A), while silencing of JMJD1A downregulated Snail mRNA levels (Supplementary Fig. S1B), suggesting that JMJD1A downregulates Snail expression at the transcriptional level, because JMJD1A demethylates H3K9me1 and H3K9Me2 to enable transcriptional activation. Thus, to reveal the mechanism of JMJD1A-mediated regulation of Snail expression, we first detected methylation levels of histone3 protein. As shown in Fig. 5A, JMJD1A overexpression significantly reduced the methylation levels of both H3K9me1 and H3K9me2. In contrast, JMJD1A dramatically increased H3K9me1 and H3K9me2 methylation (Fig. 5B). This promoted us to test whether JMJD1A protein removes mono- and di-methyl groups from H3K9 on the promoter of Snail gene and then directly interacts with Snail gene promoter. Here, using ChIP assay, we revealed that JMJD1A obviously binds to two regions of the Snail gene promoter (Fig. 5C) and the Snail gene promoter was directly immunoprecipitated with anti-JMJD1A antibodies (Fig. 5D; Supplementary Fig. S3). Moreover, ChIP assay results demonstrated that upregulated JMJD1A significantly reduced the interaction between the Snail gene promoter and both H3K9me1 and H3K9me2 compared with control (Fig. 5E and F), while JMJD1A knockdown enhanced the interaction (Fig. 5H and I). Furthermore, the luciferase reporter assays, told us that Snail gene promoter activity was activated in JMJD1A-overexpressing cells compared with control cells (Fig. 5G). In contrast, JMJD1A significantly attenuated Snail gene promoter activity compared with control cells (Fig. 5J). Taken together, these results reveal that JMJD1A transcriptionally activates Snail expression via demethylation of H3K9me1 and H3K9me2 at its special promoter region.

To further test the role of JMJD1A in prostate tumor growth in vivo, control vector or stably expressed JMJD1A RV1 cells were injected into subcutaneous tissues of Balb/c nude mice. As shown in Fig. 6A, compared with control cells, JMJD1A -overexpressing cells showed an approximately 4-fold increase in tumor volume (Fig. 6A). Moreover, the weight and size of tumors formed from control cells were decreased compared with tumors formed from JMJD1A-overexpressing cells (Fig. 6B and C). We next performed IHC staining for the proliferation marker Ki67 on xenograft tumor sections. As shown in Fig. 6D, JMJD1A overexpression significantly increased the percentage of Ki67-positive cells. Consistent with our previous finding, Western blot and qRT-PCR results also demonstrated Snail expression and mRNA levels were positively associated with JMJD1A (Fig. 6E,–G).

To further determine the functional roles of JMJD1A in prostate tumorigenesis, control cells (pLKO.1) or stably expressing shJMJD1A#1 RV1 cells were injected into the subcutaneous tissues of Balb/c nude mice, separately. As shown in Fig. 7A,–C, JMJD1A knockdown significantly retarded tumor volume and suppressed weight and size of tumors formed from stably expressing shJMJD1A#1 cells compared with tumors formed from control cells. Next, we performed IHC staining for the proliferation marker Ki67 and the apoptosis marker cleaved caspase-3 on xenograft tumor sections. JMJD1A knockdown reduced the percentage of Ki67-positive cells (Fig. 7D), concomitant with an increased number of cells positive for active caspase-3 (Fig. 7E). Furthermore, we also examined the relationship among JMJD1A and Snail in xenograft mouse models. Western blot analysis results and qRT-PCR results revealed that JMJD1A and Snail were decreased in the shJMJD1A#1 group compared with the control group (Fig. 7F,–H). In addition, we investigated whether JMJD1A knockdown decreased bone tumorigenesis of prostate cancer cells. As Supplementary Fig. S4 shows, JMJD1A knockdown inhibit bone tumorigenesis and decreased the expression of Snail. Taken together, these results suggest that JMJD1A knockdown inhibits proliferation and promotes apoptosis in xenograft prostate tumors, indicating that JMJD1A participates in prostate cancer tumorigenesis.

We previously demonstrated a tumor-promoting role of JMJD1A by regulating the activities of AR and c-Myc, and thus promoting the prostate cancer cell survival (6, 11). Herein, we found that JMJD1A protein and mRNA levels were upregulated and positively associated with more advanced stage in prostate cancer specimens. In addition, a novel molecular mechanism for JMJD1A in migration and invasion was fully elucidated in prostate cancer, an AR-driven cancer, whereby JMJD1A promotes prostate cancer cell migration and invasion through enabling Snail transcriptional activation.

Many studies have reported that histone demethylases were upregulated in human cancers (11, 33–35). However, it is still unclear whether this is a bystander event or a driver of tumorigenesis. Therefore, we first evaluated the roles of JMJD1A in prostate cancer cell proliferation, migration, and invasion. Surprisingly, we found that upregulation of JMJD1A significantly promotes prostate cancer cell proliferation, migration, and invasion, while JMJD1A knockdown exhibits the opposite effects. Our mice data also support JMJD1A exerts a tumor-driver role in prostate cancer. Targeting JMJD1A might offer an effective method to control the development of prostate cancer.

The primary cause of prostate cancer–related death is metastasis such as bone metastasis and heart metastasis, which is regulated by signaling pathways such as EMT, a dynamic process that promotes cell motility with decreased adhesive ability (36–39). Lots of studies demonstrate Snail was associated with prostate cancer cell migration (40–43) and high levels of Snail expression promoted prostate cancer cell metastasis (21). However, which histone demethylase regulates Snail in prostate cancer is still unknown. Herein, we surprisingly reveal JMJD1A promotes the prostate cancer cell metastasis partially through Snail, but does not affect E-cadherin expression, consistent with a report that Snail could promote cell survival through induction of EMT or independent of EMT (13–15).

JMJD1A regulates gene expression by modulating the methylation levels of both H3K9me1 and H3K9me2 (44). Consistent with previous reports, upregulation of JMJD1A dramatically enhanced H3K9me1 and H3K9me2 methylation levels. Interestingly, our ChIP assays showed JMJD1A interacts with the Snail special promoter region. Moreover, JMJD1A overexpression or knockdown dramatically affect the interaction between the Snail promoter and both H3K9me1 and H3K9me2, suggesting that JMJD1A induces tumor progression through regulating Snail transcription. Considering we do not perform JMJD1A ChIP-seq to show whether there are binding peak(s) of JMJD1A or at least H3K9me1 and H3K9me2 at the snail gene locus, this suggests that JMJD1A may probably regulate snail expression through indirect mechanism. Importantly, compared with the control Balb/c nude mice, the growth of JMJD1A-knockdown cells was decreased along with reducing Snail expression, while JMJD1A overexpression exhibits the opposite effects. Taken together, these results suggest a tumor-promoting role of JMJD1A in tumor metastasis. Future works are needed to determine the mechanism of JMJD1A upregulation in the setting of prostate cancer progression.

To sum up, our current studies reveal a new role of JMJD1A in promoting prostate cancer cell proliferation and migration. Even the mechanism for JMJD1A upregulation in prostate cancer cells need further study, we demonstrated that JMJD1A promoted EMT, migration, and invasion through regulating Snail expression in prostate cancer cells, and thus highlight JMJD1A as potential therapeutic target for advanced prostate cancer.

No potential conflicts of interest were disclosed.

Conception and design: S.-H. Xu

Development of methodology: X.-Y. Geng, H.-W. Jiang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D.-E. Tang, Y. Dai, D.-X. Fu

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D.-E. Tang, L.-L. Fan

Writing, review, and/or revision of the manuscript: S.-H. Xu

This work was supported by the National Natural Science Foundation of China (31700795), the Science & Technology Planning Project of Guangdong Province of China (2017B020209001), the Natural Science Foundation of Guangdong Province of China (2017A030310629), and the Science & Technology Plan of Shenzhen (JCYJ20170307095606266).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.