Cancer Cell

Article TRPM3 and miR-204 Establish a Regulatory Circuit that Controls Oncogenic Autophagy in Clear Cell Renal Cell Carcinoma Daniel P. Hall,1,8 Nicholas G. Cost,1,5,8 Shailaja Hegde,1,8 Emily Kellner,1 Olga Mikhaylova,1 Yiwen Stratton,1 Birgit Ehmer,1 William A. Abplanalp,1 Raghav Pandey,1 Jacek Biesiada,4 Christian Harteneck,6 David R. Plas,1 Jarek Meller,3,4,7 and Maria F. Czyzyk-Krzeska1,2,* 1Department

of Cancer Biology, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0521, USA of Veterans Affairs, VA Research Service, Cincinnati, OH 45220, USA 3Department of Environmental Health, University of Cincinnati College of Medicine, Cincinnati, OH 45267-0056, USA 4Division of Biomedical Informatics, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH 45229, USA 5Division of Pediatric Urology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH 45229, USA 6Department of Pharmacology and Experimental Therapy, Institute of Experimental and Clinical Pharmacology and Toxicology, Eberhard Karls University Hospitals and Clinics, and Interfaculty Center of Pharmacogenomics and Drug Research, University of Tu¨bingen, 72074 Tu¨bingen, Germany 7Department of Informatics, Nicolas Copernicus University, 87-100 Torun, Poland 8Co-first author *Correspondence: [email protected] http://dx.doi.org/10.1016/j.ccell.2014.09.015 2Department

SUMMARY

Autophagy promotes tumor growth by generating nutrients from the degradation of intracellular structures. Here we establish, using shRNAs, a dominant-negative mutant, and a pharmacologic inhibitor, mefenamic acid (MFA), that the Transient Receptor Potential Melastatin 3 (TRPM3) channel promotes the growth of clear cell renal cell carcinoma (ccRCC) and stimulates MAP1LC3A (LC3A) and MAP1LC3B (LC3B) autophagy. Increased expression of TRPM3 in RCC leads to Ca2+ influx, activation of CAMKK2, AMPK, and ULK1, and phagophore formation. In addition, TRPM3 Ca2+ and Zn2+ fluxes inhibit miR-214, which directly targets LC3A and LC3B. The von Hippel-Lindau tumor suppressor (VHL) represses TRPM3 directly through miR204 and indirectly through another miR-204 target, Caveolin 1 (CAV1). INTRODUCTION Clear cell renal cell carcinoma (ccRCC) is a malignant kidney cancer distinguished by early loss of the von Hippel-Lindau tumor suppressor protein (VHL), leading to accumulation of the hypoxia-inducible transcription factor (HIF). Induction of HIF-responsive genes provides access to extracellular nutrients and oxygen through an increased blood supply and a shift toward glycolytic metabolic pathways, both promoting tumor growth. Antiangiogenic therapies are currently the approach of choice in treating metastatic ccRCC (Coppin et al., 2011). One of the most important factors diminishing the success of antiangiogenic

therapies is that cancer cells adapt to exogenous nutrient deprivation by activating mechanisms of autophagy (Hu et al., 2012). Autophagy is a homeostatic, tightly regulated process that provides organelle quality control and generates intracellular nutrients from lysosomal processing of cellular structures. It is generally believed that autophagy is tumor-suppressing in the early stages of cancer, protecting cells from oxidative stress and resulting in genomic instability (White and DiPaola, 2009). However, during tumor growth, many cancers become addicted to autophagy as a source of nutrients. Moreover, autophagy is activated by different cancer therapies, causing resistance to treatment and promoting the use of autophagic inhibitors such

Significance Metastatic ccRCC is a largely incurable disease, and the existing treatments targeting angiogenesis, tyrosine kinase receptors, and the mTOR pathway are only partially effective. Here we discovered a network in the regulation of autophagy and ccRCC oncogenesis that can be subject to experimental therapeutics. This pathway, downstream from the VHL and miR204 tumor suppressors, involves oncogenic activity of the TRPM3 channel and its role in controlling autophagic programs through Ca2+ and Zn2+ influx. We showed that a small molecule inhibitor of TRPM3, MFA, recapitulates the effects of genetic manipulations of TRPM3. This is an important proof of concept showing that TRPM3 represents an actionable target for the treatment of ccRCC. 738 Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc.

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

as chloroquine and hydrochloroquine in combination therapies, including those for ccRCC (Lotze et al., 2013). Our laboratory demonstrated previously that the growth of ccRCC depends on active autophagy and correlates with the expression of an autophagy regulator, LC3B (Mikhaylova et al., 2012). We showed that VHL regulates oncogenic autophagic pathways by inducing miR-204, which represses the expression of LC3B. Loss of VHL and the resulting loss of miR-204 contribute to the active LC3B-mediated autophagic pathway. In parallel, VHL, through repression of HIF, stimulates the expression and activity of LC3C, an LC3B paralog, inducing tumor-suppressing autophagy. MiR-204 is expressed from intron 6 of the gene encoding a transient receptor potential nonselective cation channel, subfamily M, member 3 (TRPM3), which conducts Ca2+ and Zn2+ ions (Harteneck, 2005; Oberwinkler and Phillipp, 2007; Wagner et al., 2010). Both TRPM3 and miR-204 show enriched expression in the human kidney (Grimm et al., 2003). It is activated by a decrease in osmolality (Grimm et al., 2003), endogenous sphingosine and dihydrosphingosine (Grimm et al., 2005), and extracellular exposure to the hormone pregnenolone sulfate (PS) (Wagner et al., 2008; Naylor et al., 2008). TRPM3 is specifically inhibited by mefenamic acid (MFA) (Klose et al., 2011). VHLmediated induction of miR-204 expression correlates with VHL-induced expression of two short TRPM3 transcripts but not TRPM3 transcripts encoding the full-size functional channel protein (Mikhaylova et al., 2012). Because of the genomic localization of miR-204 within the TRPM3 gene and because several members of the TRPM family are implicated in various malignancies, we investigated the expression, activity, and mechanism of TRPM3 activity in ccRCC. RESULTS TRPM3 Promotes Tumor Growth in ccRCC Analysis of the ratio of TRPM3 protein in ccRCC samples to matched kidney samples, using semiquantitative western blotting of total protein lysates, revealed that the levels of TRPM3 were increased in 64% (42 of 66) tumors with VHL-inactive (mutated, deleted, or hypermethylated) but only in 33% (8 of 24) tumors with wild-type VHL (Figures 1A and 1B). Immunostaining with an anti-TRPM3 antibody showed intense membranous and some diffuse cytoplasmic staining in ccRCC, whereas normal kidney showed staining in the basolateral membrane (Figure 1C). We then generated stable knockdowns of the TRPM3 channel (TRPM3KD) using lentiviral small hairpin RNAs (shRNAs) in 786-O or A498 VHL() cells (Figure S1A available online). TRPM3KD decreased PS-induced (Figure S1B) and constitutive (Figure S1C) intracellular calcium concentration ([Ca2+]i), indicating the functional effectiveness of the knockdown. TRPM3KD inhibited the formation and growth of tumors from these cells in orthotopic xenografts (Figures 1D and 1E). The expression of dominant-negative TRPM3 (TRPM3DN) mutated in the pore-forming domain (PYW > AAA), which inhibits the activity of the endogenous channel by interfering with the formation of homotetramers necessary for channel activity (Hoffmann et al., 2010; Figure S1D), diminished the growth of orthotopic xenograft tumors (Figures 1D and 1E). Growth inhibition was re-

flected by a decrease in staining for the proliferation marker Ki67 (Figure 1F). The role of TRPM3 in tumor growth was further established by reconstitution of TRPM3KD (shRNA I) cells with shRNA-resistant TRPM3 (TRPM3R) (Figure S1E). TRPM3R-expressing cells formed tumors that were comparable in size and incidence with tumor formed by cells expressing endogenous TRPM3 (Figures 1G–1K). Note that TRPM3KD cells formed only three small tumors (19%) compared with control cells (100%) and TRPM3R cells (88%) (Figure 1H), and they were filled with acellular proteinous material (Figure 1J) with very few Ki67-stained nuclei (Figure 1K). Because these tumors were too small to measure with calipers, they are not shown in Figure 1G. TRPM3 rescue was accompanied by recovery of Ki67 staining (Figure 1K). The tumor-suppressive effect of TRPM3KD was not accompanied by an increase in apoptosis, shown by PARP cleavage and cleaved caspase 3 staining (Figures S1F and S1G). These data demonstrate that increased TRPM3 expression is a frequent event supporting the growth of ccRCC. TRPM3 Expression Is Regulated by miR-204 Consistent with the increased expression of TRPM3 in tumors with inactive VHL (Figure 1A), human RCC cell lines with inactive VHL (786-O and A498) showed higher levels of TRPM3 expression compared with RCC cell lines with wild-type endogenous VHL (Caki-1) (Figure 2A). Reconstitution of VHL in 786-O and A498 cells (VHL(+)) decreased the expression of TRPM3 in these cell lines (Figure 2B). This regulation was not mediated at the level of mRNA (Figure S2A) or proteasomal or lysosomal degradation (data not shown). While investigating the molecular mechanism of VHL-regulated TRPM3 expression, we observed that the 30 UTR of the TRPM3 transcript has a putative miR-204-binding site (TargetScan, Figure S2B), suggesting that miR-204 may inhibit the translation of its host gene product. Consistent with this hypothesis, transduction of an anti-miR-204 ZIP lentivirus into three different VHL(+) RCC cell lines with high levels of endogenous miR-204 (Mikhaylova et al., 2012) induced TRPM3 protein accumulation (Figure 2C). Further expression of pre-miR-204 was without effect in these cells (Figure S2C). However, transduction of a pre-miR-204 lentivirus reduced the expression of TRPM3 protein in VHL() RCC cell lines that express very low levels of endogenous miR-204 (Figure 2D). Transduction of the anti-miR-204 ZIP lentivirus was without further effect (Figure S2D). Cotransfection of the luciferase-TRPM3 30 UTR reporter construct with pre-miR-204 significantly repressed the activity of the reporter construct containing a wild-type but not a mutated miR-204 binding site (Figure 2E). The significant inverse correlation between steady-state levels of TRPM3 protein and miR-204 in extracts from normal kidney further supports these findings (Figures 2F and 2G). These results indicate that TRPM3 is a direct target of its own intronic miR-204. Expression of TRPM3 Requires Another Direct Target of miR-204, Caveolin 1 Caveolin 1 (CAV1) is a protein highly expressed in ccRCC. Its expression levels correlate with disease stage (Horiguchi et al., 2004; Tamaskar et al., 2007; Campbell et al., 2008; Steffens et al., 2011). We found higher levels of CAV1 in ccRCC compared with normal kidney in matched pairs by semiquantitative

Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc. 739

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

Figure 1. The TRPM3 Channel Is Overexpressed in ccRCC and Regulates the Growth of Xenograft Tumors (A) Box-and-whisker plot showing the quantification of the tumor/kidney ratio of TRPM3 protein levels in tumors with wild-type (WT) or inactive VHL. The boxes represent lower and upper quartiles separated by the median (thick horizontal line), and the whiskers extend to the minimum and maximum values. Means ± SD of each distribution are indicated by closed dots and crosses on the whiskers, respectively. (B) Representative western blot showing the expression of TRPM3 protein in lysates of ccRCCs (T) with deleted VHL compared with adjacent kidneys (K). Different bands detected by the antibody may correspond to different splice variants of TRPM3 (Ensembl) or posttranslational modifications. (C) Representative images of TRPM3 immunocytochemistry in sections of two different ccRCCs and normal kidney. (legend continued on next page)

740 Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc.

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

Figure 2. VHL Inhibits TRPM3 Expression through Direct Targeting by miR-204 (A) Western blot showing levels of TRPM3 protein in VHL() cells and in the Caki-1 cell line with endogenous VHL. (B) Western blot showing TRPM3 protein levels in 786-O and A498 cells with reconstituted VHL (VHL(+)) compared with isogenic VHL() cells. (C) Western blot showing the effects of anti-miR-204 on the expression of TRPM3 protein in the indicated VHL(+) RCC cells. (D) Western blot showing the effects of pre-miR-204 on the expression of TRPM3 protein in the indicated VHL() RCC cells. (E) Effects of miR-204 on the activity of the luciferase (Luc) reporter construct containing the WT or mutated miR-204 binding site (mut) 30 UTR of TRPM3 in 786-O VHL() cells. The ratio of luciferase activity in cells transduced with wild-type miR-204 to the activity in cells transduced with a nontargeting lentivirus control is shown. Error bars indicate SEM. (F) Representative western blot showing the expression of TRPM3 protein (top) in nine kidneys with low and high levels of miR-204 (bottom). (G) Regression analysis showing the negative correlation between the normalized TRPM3 protein level and the normalized miR-204 level in human kidneys. Each dot represents a sample, and the solid line represents the linear regression fit, with the Pearson correlation coefficient (r) as well as the slope and intercept of the fitted line shown in the upper left corner of the box. See also Figure S2.

immunoblotting of CAV1 protein (Figure S3A). This resulted from higher levels of CAV1 in more advanced tumors, grades G3-G4, but not G1-G2 (Figure S3B). The levels of CAV1 were significantly higher in ccRCC with inactive VHL compared with tumors with wild-type VHL (Figure S3C). We found higher levels of CAV1 in RCC cell lines with inactive VHL compared to Caki-1 cells with

endogenous wild-type VHL (Figure 3A) and saw a significant reduction in CAV1 protein expression in VHL() cells upon reconstitution of VHL (Figure 3B). Although this decrease was accompanied by a decrease in the levels of mRNA (Figure 3B), in agreement with a previous report of CAV1 regulation by HIF (Wang et al., 2012), the decrease was not sufficient to account

(D) Bar graph showing the incidence and weight of tumors formed in orthotopic xenografts by VHL() cells with endogenous (endog.) TRPM3, TRPM3KD (shRNA I and III), and TRPM3DN. The tumor weight includes the kidney in which the tumor was growing. NK, average weight of normal kidney. Control cells with endogenous TRPM3 were stably transfected with scrambled pLKO.1 vector. (E) Hematoxylin and eosin (H&E) staining of representative sections of xenograft tumors formed by 786-O VHL() cells shown in (D). The clear separation of tumors from kidneys in the case of tumors formed by cells with TRPM3KD(shRNA III) or TRPM3DN is indicated by a black dashed line. (F) Representative sections from tumors formed by 786-O VHL() cells shown in (D) stained for Ki67 and bar graph quantification of Ki67-positive nuclei compared with the total number of nuclei. (G) Box-and-whisker plot showing the subcutaneous growth of tumors formed by 786-O VHL() cells with endogenous TRPM3 or TRPM3R cells. Open circles represent points that are outside of the 1.5 interquartile range from the box. Day 1 corresponds to the time point 2.5 weeks after injections. (H) Examples and incidence of tumors formed by 786-O VHL() cell lines with the indicated status of TRPM3. Scale bar, 1 cm. (I) Weight of the collected tumors formed by 786-O VHL() cell lines with the indicated status of TRPM3. (J) Representative H&E staining of sections from tumors shown in (H). (K) Representative sections from subcutaneous tumors in (H) stained for Ki67 and bar graph quantification of Ki67-positive nuclei compared with the total number of nuclei. Unless indicated otherwise indicated, all scale bars are 50 mm, and error bars indicate SEM. See also Figure S1.

Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc. 741

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

Figure 3. CAV1 Is a miR-204 Target and Regulates Expression of TRPM3 (A) Western blot showing levels of CAV1 protein in the indicated cell lines. (B) Expression of CAV1 protein and mRNA using western blot (left) and qRT-PCR (right) in 786-O and A498 VHL(+) cells compared with the isogenic VHL() cells. Error bars indicate SD. IB, immunoblot. (C) Western blot showing the effects of pre-miR-204 on the expression of CAV1 protein in VHL() cells. (D) Western blot showing the effects of anti-miR-204 on the expression of CAV1 protein in VHL(+) cells. (E) Effects of miR-204 on the activity of a luciferase reporter containing the WT or mutated miR-204 binding site chimeric 30 UTR of CAV1 in 786-O VHL() cells. Error bars indicate SEM. (F) Effects of CAV1KD on the expression levels of TRPM3 protein in 786-O VHL() cells. (G) Effects of stable overexpression of FLAG-tagged CAV1 on TRPM3 protein levels in two different pools (P1 and P2) of 786-O VHL(+) cells. Endog., endogenous. (H) Regression analysis demonstrating the positive correlation between normalized TRPM3 and normalized CAV1 protein levels in human ccRCCs. The analysis was performed as in (G). (I) Representative western blot showing protein levels of TRPM3 and CAV1 in 14 human ccRCCs. See also Figure S3.

for the decrease in CAV1 protein. The 30 UTR of CAV1 contains a conserved miR-204 site (Figure S3D). We found that exogenous pre-miR-204 substantially repressed levels of CAV1 protein (Figure 3C) with minimal effects on CAV1 mRNA (Figure S3E) in VHL() cells. Consistent with the fact that these cells express little endogenous miR-204, anti-miR-204 did not affect the expression of CAV1 protein (Figure S3F). In contrast, expression of CAV1 protein, but not mRNA, was induced by the expression

of anti-miR-204 (Figure 3D; Figure S3E) and was not affected by pre-miR-204 (Figure S3G) in VHL(+) cells that express endogenous miR-204. Wild-type but not a mutated 30 UTR of CAV1 conferred miR-204 regulation to a luciferase reporter (Figure 3E). This indicates that the loss of VHL contributes to high levels of CAV1 in ccRCC through the loss of miR-204. Knockdown of CAV1 (CAV1KD) reduced the expression of TRPM3 in VHL() cells to levels similar to those in VHL(+) cells

742 Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc.

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

(Figure 3F). Overexpression of FLAG-tagged CAV1 without the 30 UTR in VHL(+) cells induced expression of TRPM3 (Figure 3G). These effects were only partially mediated by changes in TRPM3 mRNA (Figure S3H). We also found a significant correlation between levels of CAV1 and TRPM3 in human ccRCC (Figures 3H and 3I). Consistent with the effects of TRPM3 on tumor formation, CAV1KD diminished tumor formation by RCC VHL() cells (Figure S3I), whereas overexpression of CAV1 in VHL(+) cells increased the incidence of tumors formed by these cells to the level formed by VHL() cells, although these tumors were small (Figure S3J). Importantly, we found colocalization of CAV1 and TRPM3 intracellularly and in the plasma membrane, indicating that CAV1 plays a role in the trafficking or localization of TRPM3 to the membrane (Figure S3K), as described similarly for another TRP channel (Lockwich et al., 2000). These data establish CAV1 as an upstream regulator of TRPM3 expression. TRPM3 Regulates Autophagy Our previous identification of miR-204 as a critical regulator of autophagy prompted the hypothesis that TRPM3 may contribute to autophagy regulation as well. We investigated autophagic activity in VHL() control and TRPM3KD cells after 48 hr of growth under low-serum (0.1%) conditions. The number of autophagic vesicles counted from transmission electron microscopy (TEM) sections (Figure S4A) was decreased significantly in TRPM3KD cells (Figure 4A). Moreover, 786-O and A498 TRPM3KD cells showed a decreased accumulation of lipidated LC3A and LC3B (LC3A-II and LC3B-II) in the presence of chloroquine (CQ) (Figure 4B; Figure S4B). A similar effect was achieved by expression of a TRPM3DN mutant (Figure 4C; Figure S4C). The TRPM3KD effects were reversed by expression of exogenous TRPM3R (Figure 4D). A decrease in LC3A/B-II was also produced by knockdown of CAV1 (Figure S4D). Although autophagic activity was also lower in TRPM3KD cells compared with VHL() control cells under culture conditions of 10% serum, the accumulation of LC3A/B-II in response to thapsigargin under normal serum conditions was similar for the two cell lines (Figure S4E). Because thapsigargin increases [Ca2+]i, it can rescue diminished extracellular Ca2+ flux through TRPM3. This evidence points to the role of TRPM3-dependent cytosolic [Ca2+] in the regulation of autophagy. At the same time, the differences in autophagy between control and TRPM3KD cells do not involve mTOR because we did not detect increased mTOR activity (Figure S4F), and inhibition of mTOR induced autophagy in both cell types to the same degree (Figure S4E). Autophagy dependence on TRPM3 was specific for VHL() cells. TRPM3KD in 786-O and A498 VHL(+) cells that express little TRPM3 did not inhibit autophagy (Figure 4E). However, expression of TRPM3WT, but not TRPM3DN, in human embryonic kidney 293 (HEK293) cells induced the accumulation of LC3A/B-II (Figure S4G), suggesting that the increased expression of TRPM3 may force a connection with the autophagic pathway. Consistent with immunoblotting experiments, immunostaining for LC3B puncta in sections from orthotopic xenografts was substantially stronger in tumors formed by VHL() cells compared with tumors formed by TRPM3KD cells or cells expressing TRPM3DN (Figure 4F). Reconstitution of TRPM3R in cells with TRPM3KD rescued the accumulation of LC3B puncta in tumor cells (Figure 4G). Importantly, reconstitution of LC3B

(Figure S4H) restored the incidence of tumor formation by TRPM3KD cells that failed to grow tumors in xenografts models. However, these tumors were smaller than the tumors formed by control cells (Figure 4H). The simultaneous regulation of LC3A and LC3B suggests that both paralogs partner in the formation of similar autophagosomes. We observed that endogenous LC3B coimmunoprecipitated endogenous LC3A (Figure S4I) and that they exhibited overlapping localization in 786-O and A498 VHL() cells (Figure S4J). Together these data indicate that TRPM3 is necessary for oncogenic autophagy under starvation conditions in VHL() RCC cells. TRPM3 Regulates Autophagy through the Upstream CAMKK2-ULK1 Cascade The activity of upstream and downstream regulatory steps in the autophagic pathway is determined from the overall number of puncta immunolabelled for ATG16L and ATG5–ATG12, which correspond to the phagophore formation sites (Mikhaylova et al., 2012). We saw a significant decrease in the number of ATG16L (Figure 5A; Figure S5A) and ATG5–ATG12 (Figure 5B; Figure S5B) puncta in TRPM3KD cells compared with VHL() control cells, but we did not measure changes in the steadystate levels of ATG16L or ATG5–ATG12 conjugate by western blot analysis (Figure 5C). This indicates that TRPM3 regulates autophagy upstream of phagophore formation. The role of TRPM3 in Ca2+ influx regulation is well recognized (Grimm et al., 2003, 2005; Wagner et al., 2008; Naylor et al., 2008, 2010), and its activity is linked to calmodulin-regulated signaling pathways (Holakovska et al., 2012). [Ca2+]i regulates autophagy through the calmodulin-dependent activation of CAMKK2 and phosphorylation of AMPK (Høyer-Hansen et al., 2007; Ghislat et al., 2012), which, in turn, phosphorylates ULKs (Egan et al., 2011). We anticipated that the TRPM3KD-caused decrease in [Ca2+]i (Figures S1B and S1C) would inhibit CAMKK2 and AMPK activity and, as a result, decrease ULK1 phosphorylation and autophagic activity. We did measure a decrease in the phosphorylation of Thr172 of AMPKa and a decrease in ULK1 phosphorylation on Ser317 in TRPM3KD cells compared with control cells, indicating decreased activity of both kinases (Figure 5D). Phosphorylation of AMPKa Thr172 and ULK1 Ser317 were also diminished in VHL() cells by expression of the TRPM3DN mutant (Figure S5C) and by CAV1KD (Figure S5D). Phosphorylation of AMPKa Thr172 and ULK1 Ser317 was recovered in TRPM3KD cells by reconstitution of TRPM3R (Figure 5E). TRPM3KD in VHL(+) cells did not repress either phosphorylation, consistent with the low level of TRPM3 and a lack of connection between TRPM3 and autophagy in these cells (Figure 5F). The effects of TRPM3 on AMPK/ULK1 phosphorylation could be forced in HEK293 cells transiently transfected with exogenous wild-type TRPM3 but not TRPM3DN (Figure S5E), consistent with the effects of exogenous TRPM3 on LC3A/B autophagy (Figure S4G). Next we determined the role of CAMKK2 and AMPK in the pathway. The expression of constitutively active CAMKK2 (Green et al., 2011) in TRPM3KD cells restored autophagy, including ULK1 and AMPK phosphorylation, to levels comparable with control VHL() cells (Figure 5G). Knockdown of CAMKK2 inhibited autophagy and AMPK and ULK1

Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc. 743

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

Figure 4. The TRPM3 Channel Regulates Autophagy (A) Bar graph quantification of the number of autophagic vesicles (top) and the cytoplasmic area (bottom) in TEM sections of 20 786-O VHL() control (stably transfected with scrambled pLKO.1 vector [Scr]) or TRPM3KD cells. In this and all subsequent figures, if not indicated otherwise, ‘‘TRPM3KD’’ refers to knockdowns using shRNA I. (legend continued on next page)

744 Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc.

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

phosphorylation in VHL() cells, indicating that autophagic flux requires CAMKK2 activity (Figure 5H). A similar effect was achieved by treatment of VHL() cells with the CAMKK2 inhibitor STO-609 (Figure S5F). Loss of autophagy and decreased phosphorylation of ULK1 and AMPK in TRPM3KD cells were rescued by expression of constitutively active AMPK (Figure 5I), whereas AMPKa1/AMPKa2 knockdown inhibited autophagy and ULK phosphorylation in VHL() cells (Figure 5J). These data indicate that high-level expression of TRPM3 in VHL() cells contributes to oncogenic autophagy through a CAMKK2-AMPK-ULK1 regulatory pathway. TRPM3 Regulates Autophagy through miR-214 A close analysis of the LC3A/LC3B-I and LC3A/LC3B-II forms by western blot analysis showed a decrease in both forms in TRPM3KD cells in the presence of CQ rather than a shift of the LC3A/LC3B-II to the LC3A/LC3B-I form, as would be expected based only on the decreased upstream input to LC3 lipidation (Figure 4B; Figure S4B). This suggests an additional inhibition of LC3A/B at the protein accumulation level, even if this is not reflected in an increased number of puncta representing early phagophores (Figures 5A and 5B). We determined that the effects of TRPM3 activity on the accumulation of LC3A/B were not mediated through mRNAs levels (Figure S6A) or proteasome degradation (Figure S6B). We hypothesized that TRPM3 regulates the expression of LC3A/B at translation, possibly by microRNAs targeting both proteins. Subsequent analysis of the 30 UTRs of LC3A and LC3B revealed the presence of putative sites for miR-214 (TargetScanHuman 6.2; Figure S6C), which is of particular interest because its expression is decreased in ccRCC compared with normal kidney (Chow et al., 2010; Osanto et al., 2012), supporting a role of miR-214 in tumor suppression. We found that the expression of miR-214 was induced in TRPM3KD cells and that this effect was reversed by the expression of TRPM3R (Figure 6A). Reconstitution of VHL in VHL() cells, which leads to lower levels of TRPM3 (Figure 2), also induced the expression of miR-214 (Figure 6B). The level of miR-214 was higher in the Caki-1 cell line with wild-type VHL compared with cell lines lacking VHL (Figure 6C). The expression of exogenous TRPM3WT in 786-O VHL(+) or HEK293 cells resulted in a 2-fold repression of miR-214, whereas TRPM3DN did not have this effect (Figure 6D). Consistent with low TRPM3 activity in VHL(+) cells, TRPM3KD in these cells had no effect on miR214 levels (Figure 6E). Treatment of VHL() cells with pre-miR214 repressed the expression of LC3A/B, whereas treatment with anti-miR-214 had no effect (Figure 6F). In contrast, treatment of TRPM3KD cells (Figure 6G) or VHL(+) cells (Figure 6H)

with anti-miR-214 resulted in an increase in both protein levels. The wild-type 30 UTRs of LC3A and LC3B conferred regulation by miR-214 to the respective luciferase reporters. This activity was abolished with a mutated miR-214 binding site (Figure 6I). This establishes that LC3A and LC3B are direct targets of endogenous miR-214. Next we investigated the signaling pathway regulating the expression of miR-214 by TRPM3. Consistent with the effects of TRPM3 on ULK1 phosphorylation, the expression of miR-214 was inhibited significantly in TRPM3KD cells by constitutively active CAMKK2 or AMPKa2 (Figure 6J), whereas its expression was induced in VHL() cells by knockdown of CAMKK2 or AMPKa1/AMPKa2 (Figure 6K). The CAMKK2 inhibitor STO-609 also induced the expression of miR-214 (Figure S6D). In contrast, expression of miR-214 was not affected by ULK1 or PIK3C3 KD (Figure 6K; Figure S6E) or inhibition of CaMKII, protein kinase C, or calcineurin with specific inhibitors (Figure S6D). A decrease in miR-214 activity using an anti-miR only partially rescued the effects of CAMKK2 inhibition (Figure 6L), supporting evidence that the TRPM3-CAMKK2 pathway regulates autophagy at two levels: upstream at phagophore formation and downstream through miR-214 at LC3s protein translation. Because the TRPM3 channel also conducts Zn2+, we tested the role of Zn2+ in autophagy regulation. First we found that N,N,N0 ,N0 -tetrakis(2-pyridinylmethyl)-1,2-ethanediamine (TPEN), a specific chelator of intracellular Zn2+, induced miR-214 expression in VHL() but not VHL(+) cells (Figure S6F). TPEN treatment partially recapitulated the inhibition of LC3A/LC3B by TRPM3KD in VHL() cells (Figure S6G). TPEN did not affect ULK1 and AMPK phosphorylation (Figure S6H), an indication that Zn2+ does not act upstream of phagophore formation. The effects of TPEN on LC3A/LC3B accumulation in VHL() cells were completely reversed by the expression of anti-miR-214 (Figure S6I), indicating that miR-214 levels account for the effects of Zn2+ on autophagy. These data demonstrate two parallel mechanisms regulating miR-214 by TRPM3, one through the activity of the CAMKK2-AMPK pathway and the other through Zn2+ in a manner independent from the CAMKK2-AMPK pathway. MFA Mimics the Effects of TRPM3KD on LC3A and LC3B Autophagy MFA is a nonsteroidal, anti-inflammatory, Food and Drug Administration-approved drug for the treatment of pain and a specific inhibitor of TRPM3 (Klose et al., 2011). MFA treatment mimicked the effects of TRPM3KD on tumor formation. We treated nude mice bearing subcutaneous xenograft tumors of 786-O or

(B) Western blot and bar graph quantification showing effects of TRPM3KD on the accumulation of LC3A/LC3B-II in VHL() cells. The dashed line shows the level of LC3-II accumulation in control cells, which is accepted as 1. (C) Western blots showing the effects of TRPM3DN on the accumulation of LC3A/B-II in 786-O VHL() cells. P1 and P2 are two separate pools of cells. (D) Western blot showing the rescue of LC3A/LC3B-II accumulation upon reconstitution of TRPM3R in 786-0 VHL() TRPM3KD cells. (E) Western blot showing the effects of TRPM3KD on the accumulation of LC3A/LC3B-II in VHL(+) cells. (F) Immunostaining for LC3B puncta in sections from respective orthotopic xenografts of 786-O VHL() cells with the indicated status of TRPM3. Scale bar, 20 mm. (G) Immunostaining for LC3B puncta in sections from subcutaneous xenografts of 786-O VHL() cells with the indicated status of TRPM3. Scale bar, 20 mm. (H) Incidence and weight of tumors formed in orthotopic xenografts by 786-0 VHL() TRPM3KD cells with re-expressed FLAG-LC3B. Note that TRPM3KD(shI) cells did not form tumors, similar as shown in Figure 1D. In all panels, error bars indicate SEM. See also Figure S4.

Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc. 745

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

Figure 5. TRPM3 Regulates Autophagy through the CAMKK2/AMPK/ULK1 Pathway (A) Bar graph quantification of ATG-16L puncta detected by immunofluorescence staining in 786-O VHL() control and TRPM3KD cells. Error bars indicate SEM. (B) Bar graph quantification of ATG5-ATG12 puncta detected by immunofluorescence staining in 786-O VHL() control and TRPM3KD cells. Error bars indicate SEM. (legend continued on next page)

746 Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc.

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

A498 VHL() cells of approximately 50 mm3 with daily injections of MFA (100 mg/kg, Somchit et al., 2004) or indomethacin (IDM, 3 mg/kg, Connolly et al., 2002), a control nonsteroidal anti-inflammatory drug, for 9 days. The doses of both drugs were chosen based on the highest doses described in the literature as used in mice. MFA treatment resulted in a significant reduction of tumor growth and, in some cases, tumor regression (Figure 7A). In contrast, IDM treatment did not affect tumor growth compared with the controls (Figure 7A). Tumor growth was reflected by the proliferation marker Ki67. Significantly fewer stained nuclei were seen in sections of tumors treated with MFA compared with tumors from control or IDM-treated mice (Figure 7B). Treatment of 786-O and A498 VHL() cells for 48 hr with concentrations of MFA known to inhibit channel activity as measured by calcium imaging (Klose et al., 2011) resulted in the reduction of TRPM3 at the protein and mRNA levels (Figure 7C). Consistent with the TRPM3KD experiments, MFA treatment reduced the accumulation of LC3A/LC3B-II and ULK1 and AMPK phosphorylations (Figure 7D; Figure S7A). MFA treatment diminished the number of LC3B puncta in sections from xenograft tumors (Figure S7B). MFA also induced an increase in miR-214 levels in VHL() cells (Figure 7E). In contrast, MFA was ineffective in regulating autophagy and miR-214 levels in VHL(+) cells (Figures S7C and S7D). Overall, these data provide a proof of concept showing that specific targeting of TRPM3 by small molecule inhibitors may represent a treatment for patients with ccRCC with increased TRPM3. DISCUSSION In this work, we identified VHL as an upstream master regulator of the robust homeostatic network that tightly regulates starvation autophagy at multiple levels (Figure 8). Because either too much or not enough autophagy can lead to cell death, there is a physiological necessity for robust and fine-tuned control of this process. Several steps of this pathway are inhibited by converging mechanisms. For example, miR-204 not only directly represses TRPM3 but also represses it indirectly through CAV1 and directly inhibits accumulation of LC3B. Similarly, protein levels of LC3B are downregulated by at least two microRNAs, the expressions of which are regulated differently upstream (miR-204) and downstream (miR-214) from TRPM3 activity. In addition, phagophore initiation is controlled by Ca2+ regulation of AMPK-ULK1 axis. The robustness of the networks controlling autophagy has several implications for cancer. Cancer cells develop specific mechanisms that maintain autophagy and are not present in noncancer cells. We have demonstrated a connection between

TRPM3 and CAV1, two molecules upregulated in ccRCC, but additional mechanisms are likely to exist. Disruption of the controlling network at the level of one regulator, particularly downstream, may be insufficient to augment autophagy. For example, loss of miR-214 may be compensated for by miR-204 or inhibition of phagophore formation. This may explain why pro-oncogenic, augmented autophagy appears to take place in more advanced cancers after disruptions at multiple regulatory levels. On the other hand, cancers such as ccRCC that are associated with an early loss of VHL, a master regulator of autophagy control, may become dependent on autophagy much earlier. This also indicates that the inhibition of oncogenic autophagy may require directing therapeutics toward multiple targets. Effective inhibition of autophagy may require a combination of inhibitors against the TRPM3 and LC3B adaptor protein interaction. Currently, targeting TRP channels with small molecule agonists and antagonists is a recognized and clinically used therapeutic approach for the treatment of chronic, inflammatory, neuropathic, and visceral pain (Brederson et al., 2013). Our study provides a proof of concept showing that the TRPM3 channel may be a useful target for the treatment of renal cancer. However, expression of TRPM3 on pancreatic b cells as a source of Ca2+ and Zn2+ necessary for insulin secretion represents a potential drawback. This study also provides an important validated case of an intronic miR-mediated self-loop with negative feedback because miR-204 inhibits the expression of its host gene product, TRPM3. This minimal regulatory circuit can be predicted by algorithms (Bosia et al., 2012), but there are few well validated examples, such as miR-126 and EGFL7 (Sun et al., 2010; Nikolic et al., 2010). The discovery of the miR-204-TRPM3 relationship is particularly interesting because the expression of miR-204 results from the VHL-stimulated generation of short transcripts that do not encode the full-length channel protein (Mikhaylova et al., 2012). Some stimuli may induce the expression of both miR-204 and TRPM3.The activity of miR-204 would then contribute to the return of TRPM3 activity to the prestimulus level. With the role of TRPM3 in cancer growth, this feedback loop may be crucial to prevent the progression of oncogenic transformation. The closest homolog of TRPM3, TRPM1, encodes miR-211, a paralog of miR-204, in its intron, but TRPM1 does not have a predicted miR-204/211 site in its 30 UTR. This indicates a specific role of the TRPM3-miR-204 minimal regulatory loop. We reported previously that miR-204 is lost in the majority of ccRCCs compared with adjacent kidney (Mikhaylova et al., 2012). An independent cohort of ccRCC published by The Cancer Genome Atlas Research Network (2013) confirmed this

(C) Western blot showing steady-state expression levels of ATG16L protein and ATG5-ATG12 conjugate in 786-O VHL() control and TRPM3KD cells. (D) Western blot showing the effects of TRPM3KD on the protein expression and phosphorylation of AMPKa and ULK1 in VHL() cells. (E) Western blot showing the effects of TRPM3R on the protein expression and phosphorylation of AMPKa and ULK1 in 786-O VHL() TRPM3KD cells. (F) Western blot showing the effects of TRPM3KD on the protein expression and phosphorylation of AMPKa and ULK1 in VHL(+) cells. (G) Western blot showing the effects of the constitutively active CAMKK2 on the accumulation of LC3A/LC3B-II and phosphorylation of AMPKa and ULK1 in 786-O VHL() TRPM3KD cells. EV, empty virus. (H) Western blot showing the effects of CAMKK2 shRNA on the accumulation of LC3A/LC3B-II and phosphorylation of AMPK and ULK1 in 786-O VHL() cells. (I) Western blot showing the effects of the constitutively active AMPKa2 on the accumulation of LC3A/LC3B-II and phosphorylation of AMPK and ULK1 in 786-O TRPM3KD cells. EV, empty virus. (J) Western blot showing the effects of AMPKa1/a2 shRNAs on the accumulation of LC3A/LC3B-II and phosphorylation of AMPK and ULK1 in 786-O VHL() cells. See also Figure S5.

Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc. 747

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

Figure 6. TRPM3 Regulates Autophagy through miR-214 (A) Normalized levels of miR-214 in 786-O VHL() control, TRPM3KD, and TRPM3R cells. (B) Normalized levels of miR-214 in 786-O VHL() control, TRPM3KD, and VHL(+) cells. (C) Normalized levels of miR-214 in VHL() RCC cell lines and the VHL(+) Caki-1 cell line. (D) Effects of overexpression of TRPM3WT or TRPM3DN on the steady-state levels of miR-214 in the indicated cells. (E) Effects of TRPM3KD on normalized levels of miR-214 in VHL(+) cells. (legend continued on next page)

748 Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc.

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

result. Other examples of miR-204 as a tumor suppressor or enhancing responsiveness to treatment in cancer have been reported (Imam et al., 2012), including in pancreatic cancer (Ohuchida et al., 2011), gastric cancer (Sacconi et al., 2012), head and neck cancer (Lee et al., 2010), neuroblastoma (Ryan et al., 2012), and glioma (Ying et al., 2013). However, the expression of miR-204 is induced in insulinomas (Roldo et al., 2006). The regulation of insulin folding requires Zn2+ entering b cells through TRPM3 (Wagner et al., 2008, 2010). A high expression of miR204, in this case, may be a consequence of increased TRPM3. The miR-214 gene is expressed, together with miR-199a, from a single noncoding 7.9 kb intronic region located on the opposite strand of the gene encoding dynamin 3 (DNM3os) (Lee et al., 2009). This gene is regulated by its own promoter under the control of the Twist-1 transcription factor (Lee et al., 2009). However, other transcriptional regulatory mechanisms are not known. MiR-214 is lost in human ccRCC (Chow et al., 2010, Osanto et al., 2012). In other cancers, it has tumor-promoting or tumor-suppressing activities, depending on the regulated targets. One of its targets is the PTEN tumor suppressor (Yang et al., 2008). MiR-214 also targets components of the ubiquitin proteasome system, and its decrease contributes to dilated cardiomyopathy (Baumgarten et al., 2013). Our identification of miR-214 in the Ca2+/Zn2+-regulated pathways controlling autophagy is potentially related to its previously described role as a regulator of cardiomyocyte Ca2+ homeostasis and survival during cardiac injury (Aurora et al., 2012). An important finding of our work is that extracellular Ca2+, and potentially Zn2+, contributes to the growth of ccRCC by stimulating autophagy. Hypercalcemia is a well documented paraneoplastic syndrome that also occurs in ccRCC. Parathyroid hormone-related protein (PTHrP) produced by cancer cells is considered to be the predominant cause (Pepper et al., 2007). PTHrP levels are higher in ccRCC compared with other histological types, and ccRCCs express PTHrP in 95% of cases (Gotoh et al., 1993). PTHrP acts as a survival factor for ccRCC, and its expression is negatively regulated by VHL at the level of mRNA stability (Danilin et al., 2009). An antibody to the first 34 amino acids of the PTHrP protein inhibits growth of a SKRC-1 RCC cell line, indicating that PTHrP is an autocrine growth factor in RCC (Burton et al., 1990). Multivariate analyses show that PTHrP expression is comparable with the tumor stage as a significant prognostic indicator of outcome in advanced ccRCC (Iwamura et al., 1999). Hypercalcemia in ccRCC patients has been correlated with VEGF expression. VEGF expression is then correlated to PTHrP levels (Feng et al., 2013). In view of our data, we propose that hypercalcemia may support the growth of ccRCC through the promotion of oncogenic autophagy.

In summary, we have identified a regulatory network under the VHL tumor suppressor that, when released from VHL control, contributes to the growth of ccRCC. Importantly, this network has several potential targets that can be investigated for therapeutic interventions. EXPERIMENTAL PROCEDURES Cell Culture and Treatments Human VHL() and VHL(+) 786-O, A498, and Caki-1 cells have been described previously (Mikhaylova et al., 2008; Yi et al., 2010). All cell lines are authenticated periodically (Genetica). Unless indicated otherwise, experiments were performed with the following time line: cells (10–20 3 104 cells/ml) were plated in medium with 10% serum. After 24 hr, the medium was changed to contain 0.1% serum, and cells were collected 48 hr later. Specific treatments were applied during the 48 hr period. For rescue experiments, cells were transfected with plasmids or transduced with viruses and, 24 hr later, starved for 48 hr. For autophagosome and LC3 experiments, cells were treated with 100 mM CQ for 1 hr before collection. In all relevant experiments, controls were treated with empty or scrambled viruses or plasmids. Cells were collected in radioimmunoprecipitation assay buffer. Five to thirty micrograms of extracts were separated on 10% or 4%–12% polyacrylamide gels and transferred onto a polyvinylidene fluoride membrane. Blots were probed with relevant antibodies. RNA extractions and quantitative RT-PCR (qRT-PCR) have been described before (Mikhaylova et al., 2012). The primer sequences, antibodies, and chemicals used are listed in the Supplemental Experimental Procedures. Human RCC Tumors Fresh-frozen samples or paraffin-embedded samples of ccRCC and matched normal kidney were obtained and analyzed for the VHL gene sequence and protein expression as described previously (Yi et al., 2010). All human samples were from anonymous donors and exempt from the Institutional Review Board protocol. Lysates were analyzed for proteins of interest using semiquantitative immunoblotting (Yi et al., 2010; Mikhaylova et al., 2012). Constructs and Plasmids The following miR reagents were used. Pre-miR-204 and pre-miR-214 precursors and pre-miR negative controls were purchased from Ambion. AntimiR-204 and 214 and negative controls were purchased from Dharmacon. Lentivector-based microRNA precursor construct miR-204 and miR-214, miRZip lentivector-based anti-microRNA-204 and anti-microRNA-214, and negative controls were obtained from System Biosciences. The GoClone luciferase 30 UTR chimeric constructs of LC3s and TRPM3 were purchased from Switchgear Genomics. To alter the miR-214 binding sequence, the core CCTGCTG sequence in the 30 UTR of LC3A was mutated to AGTTGTC, and the core CTGCTGA sequence in the 30 UTR of LC3B was mutated to GTTGTC (GenScript). To alter the miR-204 binding site, the core GGG sequence in the 30 UTR of TRPM3 was replaced with TTT. The 30 UTR of CAV1 was cloned into SpeI/MluI sites of the pMIR-REPORT luciferase construct (Ambion), and the core GGG sequence was mutated to CCC. All shRNA constructs were purchased from Open Biosystems. A list of shRNAs used is provided in the Supplemental Experimental Procedures. Lentiviral DNA constructs were VSV-G envelope-packaged (Cincinnati Children’s Hospital Medical Center (CCHMC), Viral Vector Core). For stable pooled cells, transduction was followed by

(F) Effects of treatment with the indicated concentrations of pre-miR-214 and anti-miR-214 on the accumulation of LC3A/LC3B in starved VHL() cells. All cells were treated with CQ. (G) Effects of anti-miR-214 (lentivirus, lenti) treatments on the accumulation of LC3A/LC3B in starved 786-O VHL() TRPM3KD cells. In the case of siRNA, only CQ-treated cells are shown. (H) Effects of lentiviral anti-miR-214 and pre-miR-214 treatments on the accumulation of LC3A/B in starved 786-O VHL(+) cells. (I) Effects of pre-miR-214 on the activity of luciferase reporter constructs containing LC3A or LC3B 30 UTRs with either WT or mutated miR-214 binding sites in 786-O VHL() cells. (J) Effects of constitutively active CAMKK2 and AMPKa2 on the accumulation of miR-214 in starved 786-O VHL() TRPM3KD cells. (K) Effects of CAMKK2, AMPKa1/2, ULK1, and PIK3C3 knockdowns on the accumulation of miR-214 in starved 786-O VHL() cells. (L) Effects of anti-miR214 on the accumulation of LC3A and LC3B in 786-O VHL() cells treated with STO-609. In all panels, n represents a number of individual experiments in triplicate. If n < 3, error bars indicate SD. If n R 3, error bars indicate SEM. See also Figure S6.

Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc. 749

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

Figure 7. The Effects of TRPM3KD Are Reproduced by MFA, a Small Molecule Blocker of TRPM3 (A) Box-and-whisker plot showing the growth of subcutaneous xenograft tumors formed by the indicated VHL() cells in mice treated with MFA, IDM, or vehicle. Open circles represent points that are outside of the 1.5 interquartile range from the box. (B) Representative sections of 786-O tumors from each treatment group stained for Ki67 and bar graph quantification performed as in Figure 1F. Scale bar, 100 mM. (C) Effects of treatment with MFA on the expression of endogenous TRPM3 measured by western blot and by qRT-PCR in VHL() cells. (D) Effects of MFA treatments of 786-O and A498 VHL() cells on the accumulation of LC3A/LC3B-II and phosphorylation of ULK1-Ser317. R, ratio of p-ULK1Ser317 to total ULK1. (E) Effects of MFA on accumulation of miR-214 in 786-O and A498 VHL() cells. In all panels, error bars indicate SEM. See also Figure S7.

plasmid-appropriate selection. TRPM3 wild-type and mutated constructs had yellow fluorescent protein fused to the C terminus and were expressed from the pcDNA3.1 plasmid (Hoffmann et al., 2010). The TRPM3 shRNA-insensitive mutant was obtained by synonymous mutation of the shRNA binding site and replacement of the neomycin selection marker with a zeocin cassette in pcDNA3.1 (GenScript). The construct of constitutively active FLAG-tagged CAMKK2 was obtained from Addgene (plasmid no. 33324), and the cDNA was subcloned into the pSMPUW-Hygro lentiviral construct (Cell Biolabs). The lentiviral construct of constitutively active myc-tagged AMPKa2 was obtained from Dr. B. Dasgupta (CCHMC). Transfection and Transduction For plasmid transfections, we used Fugene or Lipofectamine. For microRNA constructs or pooled siRNA transfections, we used Lipofectamine 2000. All were performed according to the manufacturer’s protocol. Transductions

with lentiviral particles were performed using 2 mg/ml Polybrene. The luciferase activity of the GoClone 30 UTR reporter constructs was measured using a LightSwitch luciferase assay kit (Switchgear Genomics) according to the manufacturer’s protocol. The luciferase activity of the pMIR-REPORT construct containing the CAV1 30 UTR was measured using a standard luciferase assay kit (Promega) using cotransfection of Renilla luciferase for normalization. All results are shown as a ratio of the effect of the miR mimetic to the effect of the nontargeting control. Mouse Xenografts Experiments on mice were performed in accordance with University of Cincinnati Institutional Animal Care and Use Committee-approved protocols. For intrakidney injections, 5 3 104 cells were resuspended in 50% Matrigel to a final volume of 30 ml and then injected slowly into the subcapsular renal parenchyma of 4- to 5-week-old athymic nude mice. Tumors were collected after

750 Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc.

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

tions between variables were assessed by means of regression analysis and by Pearson correlation coefficient. All data were log2-transformed to facilitate robust and outlier-insensitive analyses. Differences of p < 0.05 were considered to be statistically significant. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures and seven figures and can be found with this article online at http://dx.doi. org/10.1016/j.ccell.2014.09.015. AUTHOR CONTRIBUTION

Figure 8. Model Summarizing the Proposed Regulatory Network Controlling Oncogenic LC3A/LC3B Autophagy Black lines, oncogenic pathways; gray lines, tumor-suppressing pathways; asterisk, work reported by Mikhaylova et al. (2012).

10–12 weeks. For subcutaneous injections, 1 3 106 cells in cell culture medium containing 50% Matrigel were injected into the flanks of athymic nude mice. Three standard dimensions of the tumor were measured, and the volume of the tumor was calculated using the equation d1 3 d2 3 d3 3 p / 6. After 23 weeks, these tumors reached an approximate average volume of 50 mm3, and mice were randomly assigned to treatment groups. At the end of the experiments, mice were sacrificed, and tumors were fixed in 4% paraformaldehyde and processed for immunocytochemistry using anti-Ki67, anti-LC3B, and anti-caspase 3 antibodies. Ki-67-stained nuclei were counted as described before (Mikhaylova et al., 2012). Immunocytochemistry, Immunofluorescence, and TEM For immunocytochemistry, sections of fixed and paraffin-embedded tumors were processed in the Pathology Research Core at CCHMC. For immunofluorescence, cells were plated and treated as described before (Mikhaylova et al., 2012), and images were acquired on a Zeiss LSM710 confocal microscope or on a Zeiss Axioplan 2 microscope with the appropriate filter cubes and a Zeiss Axiocam MRm to record the images. The number of puncta per cell and the number of cells expressing puncta versus the total number of cells were counted manually in five independent fields containing between 50–100 cells, or three to five sections were counted for each type of tumor. TEM samples were processed in the Electron Microscopy Core at CCHMC as described before (Mikhaylova et al., 2012). Calcium Imaging Cell suspensions were incubated with RPMI-1640 medium containing 10 mM Indo-1 (Life Technologies) for 30 min at 37 C in the dark with repeated mixing. Cells were then washed twice with RPMI-1640 medium and resuspended at 2 3 106 cells/ml. Prewarmed samples were first analyzed for 2 min at about 2,000 events/s to establish the baseline. Then PS or PS and EGTA were quickly added to the suspension, and cells were further acquired for another 200 ms before ionomycin was added as the positive control for Ca2+ influx. Measurements were performed using a BD-LSR fluorescence-activated cell sorting machine equipped with UV lasers in the Flow Cytometry Core at Shriners Hospital for Children in Cincinnati. Ratiometric analyses of Indo-1 violet to Indo-1 blue were performed using FlowJo software. Statistical Analysis For descriptive statistics, data are expressed as mean ± SEM unless indicated otherwise. Analysis of differential expression was performed using ANOVA followed by Tukey-Kramer multiple comparison tests. For the analysis of tumor growth, repeated-measures ANOVA (parametric and nonparametric) were performed. Standard box-and-whisker plots were used to compare distributions of the normalized abundance measure for individual proteins. Correla-

D.P.H., E.K., O.M., Y.S., W.A.A., and R.P. performed experiments. N.G.C. contributed to the design and performed experiments using drug treatments. S.H. contributed to the design and performed all rescue experiments. B.E. performed confocal microscopy analysis, and J.B. performed statistical analyses. C.H. contributed to the design of the experiments and provided all TRPM3 constructs. D.R.P., J. M., and M.F.C.K guided the project and contributed experimental design and data interpretation. M.F.C.K. provided overall direction and wrote the manuscript. ACKNOWLEDGMENTS This work was supported by NCI Grant CA122346 and a BLR&D VA Merit Award (to M.C.K.), by NCI Grants CA133164 and CA168815 (to D.R.P.), and by UC Grant P30-ES006096 CEG. S.H. was supported by Grant T32 CA117846. We thank C. Wilson and undergraduate students P.J. Krzeski and A. Price for technical assistance, B. Dispasquale for processing sections for immunocytochemistry, G. Ciraolo for processing tissues for TEM, G. Doerman for preparing the figures, and B. Peace for editing the manuscript. This work is dedicated to the memory of George Clerc, who died of ccRCC on August 9, 1990. Received: October 29, 2013 Revised: August 6, 2014 Accepted: September 25, 2014 Published: November 10, 2014 REFERENCES Aurora, A.B., Mahmoud, A.I., Luo, X., Johnson, B.A., van Rooij, E., Matsuzaki, S., Humphries, K.M., Hill, J.A., Bassel-Duby, R., Sadek, H.A., and Olson, E.N. (2012). MicroRNA-214 protects the mouse heart from ischemic injury by controlling Ca2⁺ overload and cell death. J. Clin. Invest. 122, 1222–1232. Baumgarten, A., Bang, C., Tschirner, A., Engelmann, A., Adams, V., von Haehling, S., Doehner, W., Pregla, R., Anker, M.S., Blecharz, K., et al. (2013). TWIST1 regulates the activity of ubiquitin proteasome system via the miR-199/214 cluster in human end-stage dilated cardiomyopathy. Int. J. Cardiol. 168, 1447–1452. Bosia, C., Osella, M., Baroudi, M.E., Cora`, D., and Caselle, M. (2012). Gene autoregulation via intronic microRNAs and its functions. BMC Syst. Biol. 6, 131. Brederson, J.-D., Kym, P.R., and Szallasi, A. (2013). Targeting TRP channels for pain relief. Eur. J. Pharmacol. 716, 61–76. Burton, P.B., Moniz, C., and Knight, D.E. (1990). Parathyroid hormone related peptide can function as an autocrine growth factor in human renal cell carcinoma. Biochem. Biophys. Res. Commun. 167, 1134–1138. Campbell, L., Jasani, B., Edwards, K., Gumbleton, M., and Griffiths, D.F.R. (2008). Combined expression of caveolin-1 and an activated AKT/mTOR pathway predicts reduced disease-free survival in clinically confined renal cell carcinoma. Br. J. Cancer 98, 931–940. Cancer Genome Atlas Research Network (2013). Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature 499, 43–49. Chow, T.F., Youssef, Y.M., Lianidou, E., Romaschin, A.D., Honey, R.J., Stewart, R., Pace, K.T., and Yousef, G.M. (2010). Differential expression

Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc. 751

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

profiling of microRNAs and their potential involvement in renal cell carcinoma pathogenesis. Clin. Biochem. 43, 150–158. Connolly, E.M., Harmey, J.H., O’Grady, T., Foley, D., Roche-Nagle, G., Kay, E., and Bouchier-Hayes, D.J. (2002). Cyclo-oxygenase inhibition reduces tumour growth and metastasis in an orthotopic model of breast cancer. Br. J. Cancer 87, 231–237.

Iwamura, M., Wu, W., Muramoto, M., Ohori, M., Egawa, S., Uchida, T., and Baba, S. (1999). Parathyroid hormone-related protein is an independent prognostic factor for renal cell carcinoma. Cancer 86, 1028–1034. Klose, C., Straub, I., Riehle, M., Ranta, F., Krautwurst, D., Ullrich, S., Meyerhof, W., and Harteneck, C. (2011). Fenamates as TRP channel blockers: mefenamic acid selectively blocks TRPM3. Br. J. Pharmacol. 162, 1757–1769.

Coppin, C., Kollmannsberger, C., Le, L., Porzsolt, F., and Wilt, T.J. (2011). Targeted therapy for advanced renal cell cancer (RCC): a Cochrane systematic review of published randomised trials. BJU Int. 108, 1556–1563.

Lee, Y.-B., Bantounas, I., Lee, D.-Y., Phylactou, L., Caldwell, M.A., and Uney, J.B. (2009). Twist-1 regulates the miR-199a/214 cluster during development. Nucleic Acids Res. 37, 123–128.

Danilin, S., Sourbier, C., Thomas, L., Rothhut, S., Lindner, V., Helwig, J.J., Jacqmin, D., Lang, H., and Massfelder, T. (2009). von Hippel-Lindau tumor suppressor gene-dependent mRNA stabilization of the survival factor parathyroid hormone-related protein in human renal cell carcinoma by the RNA-binding protein HuR. Carcinogenesis 30, 387–396.

Lee, Y., Yang, X., Huang, Y., Fan, H., Zhang, Q., Wu, Y., Li, J., Hasina, R., Cheng, C., Lingen, M.W., et al. (2010). Network modeling identifies molecular functions targeted by miR-204 to suppress head and neck tumor metastasis. PLoS Comput. Biol. 6, e1000730.

Egan, D., Kim, J., Shaw, R.J., and Guan, K.-L. (2011). The autophagy initiating kinase ULK1 is regulated via opposing phosphorylation by AMPK and mTOR. Autophagy 7, 643–644. Feng, C.C., Ding, G.X., Song, N.H., Li, X., Wu, Z., Jiang, H.W., and Ding, Q. (2013). Paraneoplastic hormones: parathyroid hormone-related protein (PTHrP) and erythropoietin (EPO) are related to vascular endothelial growth factor (VEGF) expression in clear cell renal cell carcinoma. Tumour Biol. 34, 3471–3476. Ghislat, G., Patron, M., Rizzuto, R., and Knecht, E. (2012). Withdrawal of essential amino acids increases autophagy by a pathway involving Ca2+/ calmodulin-dependent kinase kinase-b (CaMKK-b). J. Biol. Chem. 287, 38625–38636. Gotoh, A., Kitazawa, S., Mizuno, Y., Takenaka, A., Arakawa, S., Matsumoto, O., Kitazawa, R., Fujimori, T., Maeda, S., and Kamidono, S. (1993). Common expression of parathyroid hormone-related protein and no correlation of calcium level in renal cell carcinomas. Cancer 71, 2803–2806. Green, M.F., Anderson, K.A., and Means, A.R. (2011). Characterization of the CaMKKb-AMPK signaling complex. Cell. Signal. 23, 2005–2012.

Lockwich, T.P., Liu, X., Singh, B.B., Jadlowiec, J., Weiland, S., and Ambudkar, I.S. (2000). Assembly of Trp1 in a signaling complex associated with caveolinscaffolding lipid raft domains. J. Biol. Chem. 275, 11934–11942. Lotze, M.T., Maranchie, J., and Appleman, L. (2013). Inhibiting autophagy: a novel approach for the treatment of renal cell carcinoma. Cancer J. 19, 341–347. Mikhaylova, O., Ignacak, M.L., Barankiewicz, T.J., Harbaugh, S.V., Yi, Y., Maxwell, P.H., Schneider, M., Van Geyte, K., Carmeliet, P., Revelo, M.P., et al. (2008). The von Hippel-Lindau tumor suppressor protein and Egl-9Type proline hydroxylases regulate the large subunit of RNA polymerase II in response to oxidative stress. Mol. Cell. Biol. 28, 2701–2717. Mikhaylova, O., Stratton, Y., Hall, D., Kellner, E., Ehmer, B., Drew, A.F., Gallo, C.A., Plas, D.R., Biesiada, J., Meller, J., and Czyzyk-Krzeska, M.F. (2012). VHL-regulated MiR-204 suppresses tumor growth through inhibition of LC3B-mediated autophagy in renal clear cell carcinoma. Cancer Cell 21, 532–546. Naylor, J., Milligan, C.J., Zeng, F., Jones, C., and Beech, D.J. (2008). Production of a specific extracellular inhibitor of TRPM3 channels. Br. J. Pharmacol. 155, 567–573.

Grimm, C., Kraft, R., Sauerbruch, S., Schultz, G., and Harteneck, C. (2003). Molecular and functional characterization of the melastatin-related cation channel TRPM3. J. Biol. Chem. 278, 21493–21501.

Naylor, J., Li, J., Milligan, C.J., Zeng, F., Sukumar, P., Hou, B., Sedo, A., Yuldasheva, N., Majeed, Y., Beri, D., et al. (2010). Pregnenolone sulphateand cholesterol-regulated TRPM3 channels coupled to vascular smooth muscle secretion and contraction. Circ. Res. 106, 1507–1515.

Grimm, C., Kraft, R., Schultz, G., and Harteneck, C. (2005). Activation of the melastatin-related cation channel TRPM3 by D-erythro-sphingosine [corrected]. Mol. Pharmacol. 67, 798–805.

Nikolic, I., Plate, K.H., and Schmidt, M.H. (2010). EGFL7 meets miRNA-126: an angiogenesis alliance. J. Angiogenes Res. 2, 9.

Harteneck, C. (2005). Function and pharmacology of TRPM cation channels. Naunyn Schmiedebergs Arch. Pharmacol. 371, 307–314. Hoffmann, A., Grimm, C., Kraft, R., Goldbaum, O., Wrede, A., Nolte, C., Hanisch, U.K., Richter-Landsberg, C., Bru¨ck, W., Kettenmann, H., and Harteneck, C. (2010). TRPM3 is expressed in sphingosine-responsive myelinating oligodendrocytes. J. Neurochem. 114, 654–665. Holakovska, B., Grycova, L., Jirku, M., Sulc, M., Bumba, L., and Teisinger, J. (2012). Calmodulin and S100A1 protein interact with N terminus of TRPM3 channel. J. Biol. Chem. 287, 16645–16655. Horiguchi, A., Asano, T., Asakuma, J., Asano, T., Sumitomo, M., and Hayakawa, M. (2004). Impact of caveolin-1 expression on clinicopathological parameters in renal cell carcinoma. J. Urol. 172, 718–722. Høyer-Hansen, M., Bastholm, L., Szyniarowski, P., Campanella, M., Szabadkai, G., Farkas, T., Bianchi, K., Fehrenbacher, N., Elling, F., Rizzuto, R., et al. (2007). Control of macroautophagy by calcium, calmodulin-dependent kinase kinase-beta, and Bcl-2. Mol. Cell 25, 193–205. Hu, Y.-L., Jahangiri, A., Delay, M., and Aghi, M.K. (2012). Tumor cell autophagy as an adaptive response mediating resistance to treatments such as antiangiogenic therapy. Cancer Res. 72, 4294–4299. Imam, J.S., Plyler, J.R., Bansal, H., Prajapati, S., Bansal, S., Rebeles, J., Chen, H.I., Chang, Y.F., Panneerdoss, S., Zoghi, B., et al. (2012). Genomic loss of tumor suppressor miRNA-204 promotes cancer cell migration and invasion by activating AKT/mTOR/Rac1 signaling and actin reorganization. PLoS ONE 7, e52397.

Oberwinkler, J., and Phillipp, S.E. (2007). TRPM3. Handbook Exp. Pharmacol. 179, 253–267. Ohuchida, K., Mizumoto, K., Kayashima, T., Fujita, H., Moriyama, T., Ohtsuka, T., Ueda, J., Nagai, E., Hashizume, M., and Tanaka, M. (2011). MicroRNA expression as a predictive marker for gemcitabine response after surgical resection of pancreatic cancer. Ann. Surg. Oncol. 18, 2381–2387. Osanto, S., Qin, Y., Buermans, H.P., Berkers, J., Lerut, E., Goeman, J.J., and van Poppel, H. (2012). Genome-wide microRNA expression analysis of clear cell renal cell carcinoma by next generation deep sequencing. PLoS ONE 7, e38298. Pepper, K., Jaowattana, U., Starsiak, M.D., Halkar, R., Hornaman, K., Wang, W., Dayamani, P., and Tangpricha, V. (2007). Renal cell carcinoma presenting with paraneoplastic hypercalcemic coma: a case report and review of the literature. J. Gen. Intern. Med. 22, 1042–1046. Roldo, C., Missiaglia, E., Hagan, J.P., Falconi, M., Capelli, P., Bersani, S., Calin, G.A., Volinia, S., Liu, C.-G., Scarpa, A., and Croce, C.M. (2006). MicroRNA expression abnormalities in pancreatic endocrine and acinar tumors are associated with distinctive pathologic features and clinical behavior. J. Clin. Oncol. 24, 4677–4684. Ryan, J., Tivnan, A., Fay, J., Bryan, K., Meehan, M., Creevey, L., Lynch, J., Bray, I.M., O’Meara, A., Tracey, L., et al. (2012). MicroRNA-204 increases sensitivity of neuroblastoma cells to cisplatin and is associated with a favourable clinical outcome. Br. J. Cancer 107, 967–976. Sacconi, A., Biagioni, F., Canu, V., Mori, F., Di Benedetto, A., Lorenzon, L., Ercolani, C., Di Agostino, S., Cambria, A.M., Germoni, S., et al. (2012). miR-

752 Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc.

Cancer Cell TRPM3 Channel in Clear Cell Renal Cell Carcinoma

204 targets Bcl-2 expression and enhances responsiveness of gastric cancer. Cell Death Dis. 3, e423. Somchit, N., Sanat, F., Gan, E.H., Shahrin, I.A., and Zuraini, A. (2004). Liver injury induced by the non-steroidal anti-inflammatory drug mefenamic acid. Singapore Med. J. 45, 530–532. Steffens, S., Schrader, A.J., Blasig, H., Vetter, G., Eggers, H., Tra¨nkenschuh, W., Kuczyk, M.A., and Serth, J. (2011). Caveolin 1 protein expression in renal cell carcinoma predicts survival. BMC Urol. 11, 25. Sun, Y., Bai, Y., Zhang, F., Wang, Y., Guo, Y., and Guo, L. (2010). miR-126 inhibits non-small cell lung cancer cells proliferation by targeting EGFL7. Biochem. Biophys. Res. Commun. 391, 1483–1489. Tamaskar, I., Choueiri, T.K., Sercia, L., Rini, B., Bukowski, R., and Zhou, M. (2007). Differential expression of caveolin-1 in renal neoplasms. Cancer 110, 776–782. Wagner, T.F., Loch, S., Lambert, S., Straub, I., Mannebach, S., Mathar, I., Du¨fer, M., Lis, A., Flockerzi, V., Philipp, S.E., and Oberwinkler, J. (2008). Transient receptor potential M3 channels are ionotropic steroid receptors in pancreatic beta cells. Nat. Cell Biol. 10, 1421–1430. Wagner, T.F., Drews, A., Loch, S., Mohr, F., Philipp, S.E., Lambert, S., and Oberwinkler, J. (2010). TRPM3 channels provide a regulated influx pathway for zinc in pancreatic beta cells. Pflugers Arch. 460, 755–765.

Wang, Y., Roche, O., Xu, C., Moriyama, E.H., Heir, P., Chung, J., Roos, F.C., Chen, Y., Finak, G., Milosevic, M., et al. (2012). Hypoxia promotes ligand-independent EGF receptor signaling via hypoxia-inducible factor-mediated upregulation of caveolin-1. Proc. Natl. Acad. Sci. USA 109, 4892–4897. White, E., and DiPaola, R.S. (2009). The double-edged sword of autophagy modulation in cancer. Clin. Cancer Res. 15, 5308–5316. Yang, H., Kong, W., He, L., Zhao, J.J., O’Donnell, J.D., Wang, J., Wenham, R.M., Coppola, D., Kruk, P.A., Nicosia, S.V., and Cheng, J.Q. (2008). MicroRNA expression profiling in human ovarian cancer: miR-214 induces cell survival and cisplatin resistance by targeting PTEN. Cancer Res. 68, 425–433. Yi, Y., Mikhaylova, O., Mamedova, A., Bastola, P., Biesiada, J., Alshaikh, E., Levin, L., Sheridan, R.M., Meller, J., and Czyzyk-Krzeska, M.F. (2010). von Hippel-Lindau-dependent patterns of RNA polymerase II hydroxylation in human renal clear cell carcinomas. Clin. Cancer Res. 16, 5142–5152. Ying, Z., Li, Y., Wu, J., Zhu, X., Yang, Y., Tian, H., Li, W., Hu, B., Cheng, S.Y., and Li, M. (2013). Loss of miR-204 expression enhances glioma migration and stem cell-like phenotype. Cancer Res. 73, 990–999.

Cancer Cell 26, 738–753, November 10, 2014 ª2014 Elsevier Inc. 753

TRPM3 and miR-204 establish a regulatory circuit that controls oncogenic autophagy in clear cell renal cell carcinoma.

Autophagy promotes tumor growth by generating nutrients from the degradation of intracellular structures. Here we establish, using shRNAs, a dominant-...
5MB Sizes 0 Downloads 8 Views