The E2F1–Mir-520/372/373–SPOP Axis Modulates Progression Of

Published OnlineFirst October 22, 2018; DOI: 10.1158/0008-5472.CAN-18-1662 Cancer Molecular Cell Biology Research

The E2F1–miR-520/372/373–SPOP Axis Modulates Progression of Renal Carcinoma Meng Ding1,2, Xiaolan Lu1,2, Cheng Wang1,2, Quan Zhao2, Jingping Ge3, Qiuyuan Xia4, Junjun Wang1,KeZen2, Chen-Yu Zhang2, and Chunni Zhang1,2

Abstract

Although renal cell carcinoma (RCC) is the most malignant nificantly decreased both tumor size and lung metastasis urologic cancer, its pathogenesis remains unclear, and effective ratio in mice bearing orthotopic xenograft tumors. Decreased treatments for advanced RCC are still lacking. Here, we report expression of miR-520/372/373 family was mediated by tran- that a novel E2F1–miR-520/372/373–SPOP axis controls RCC scription factor E2F1. In conclusion, our results demonstrate carcinogenesis. Speckle-type POZ protein (SPOP) was upre- that the E2F1-miR-520/372/373–SPOP axis functions as a gulated in over 90% of RCC tissues, whereas the miR-520/372/ key signaling pathway in RCC progression and metastasis and 373 family was downregulated and correlated inversely with represents a promising opportunity for targeted therapies. SPOP protein levels in RCC tissues. The miR-520/372/373 family targeted the SPOP 30-UTR and suppressed SPOP Significance: These findings show that the E2F1-miR-520/ protein expression, leading to elevation of PTEN and DUSP7 372/373 family–SPOP axis promotes RCC progression, levels and, consequently, decreased proliferation, invasion/ thereby contributing to our understanding of RCC patho- migration, and metastasis of RCC cells in vitro and in vivo. genesis and unveiling new avenues for more effective Tail-vein delivery of therapeutic miR-520/372/373 family sig- targeted therapies. Cancer Res; 78(24); 1–14. 2018 AACR.

Introduction required (5, 6). Speckle-type POZ protein (SPOP) is a BTB/POZ domain protein that encodes an E3 ubiquitin ligase component Renal cell carcinoma (RCC) accounts for approximately 4% of (7). SPOP has been reported to participate in many biological all cancers and is the third most common urologic malignancy, functions, especially in regulating tumor growth including RCC which has the highest fatality rate among all urologic cancers (1, (6, 8–13). SPOP is highly expressed in RCC, in particular, in 2). RCC is notorious for its resistance to routine chemotherapy ccRCC (6, 8, 11, 12). It has been demonstrated that SPOP is and radiotherapy (3). When metastasis occurs, it is mostly incur- transcriptionally activated by hypoxia-inducible factor (HIF; able with a very low 5-year survival rate (3). The most commonly refs. 11, 14, 15) and induces RCC tumorigenesis through observed RCC subtype is clear-cell renal cell carcinoma (ccRCC), degradation of some tumor suppressors, such as PTEN, ERK which accounts for approximately 75%–85% of all cases of RCC, phosphatase, and DUSP7 in RCC (6, 11, 16, 17). Therefore, with the highest rates of local invasion, metastasis, mortality, and SPOP is regarded as an attractive target speciﬁctoRCCthatmay refraction to current treatments (4, 5). Although intensive effort yield novel drug discovery efforts (11, 12). Recently, it has been has been put toward exploring the pathogenesis of this disease, shown that a structure-based design and subsequent hit opti- the mechanisms involved in the development and progression of mization yield small molecules that can inhibit the SPOP– RCC and more effective therapeutic approaches are urgently substrate protein interaction and thus suppress oncogenic SPOP signaling pathways in kidney cancer (12). However,

1 despite recent exploration of biological functions, studies on Department of Clinical Laboratory, Jinling Hospital, State Key Laboratory of ﬁ Analytical Chemistry for Life Science, NJU Advanced Institute for Life Sciences the mechanism of SPOP expression in RCC are still insuf cient (NAILS), School of Life Sciences, Nanjing University, Nanjing, China. 2State Key (6, 18). Further investigation of the biological function, clinical Laboratory of Pharmaceutical Biotechnology and Jiangsu Engineering Research signiﬁcance, and mechanism of SPOP is critical for an under- Center for MicroRNA Biology and Biotechnology, NJU Advanced Institute for standing of RCC pathogenesis. Life Sciences (NAILS), School of Life Science, Nanjing University, Nanjing, China. The discovery of miRNAs has added another layer of complex- 3Department of Urology, Jinling Hospital, Nanjing University School of Medicine, 4 ity to the control of gene expression. miRNAs are recently dis- Nanjing University, Nanjing, China. Department of Pathology, Jinling Hospital, – Nanjing University School of Medicine, Nanjing University, Nanjing, China. covered small (19 22 nucleotides) noncoding RNAs that function as repressors of gene activity by binding to complementary Note: Supplementary data for this article are available at Cancer Research sequences in the 30 untranslated region (30-UTR) of target gene Online (http://cancerres.aacrjournals.org/). transcripts, leading to mRNA degradation and/or translational M. Ding, X. Lu, and C. Wang contributed equally to this article. repression (19, 20). Conservative estimates show that approxi- Corresponding Authors: Chunni Zhang, Department of Clinical Laboratory, mately 60% of all human mRNAs are regulated by miRNAs, Jinling Hospital, Nanjing University, 305 East Zhongshan Rd., Nanjing covering virtually all cellular functions (21). We and others have 210002, China. Phone: 8625-8086-3082; Fax: 8625-8368-6234; E-mail: shown that miRNAs are frequently aberrantly expressed in cancer [email protected]; and Chen-Yu Zhang, [email protected] tissues, have been referred to as "onco(genic)-miRNAs" or "tumor doi: 10.1158/0008-5472.CAN-18-1662 suppressor miRNAs," and play crucial roles in diverse regulation 2018 American Association for Cancer Research. pathways in multiple types of tumors, including RCC (22–24)

www.aacrjournals.org OF1

Ding et al.

(18)In this study, we found that SPOP was specifically targeted by (Gibco) and incubated in 5% CO2 at 37 C in a water-saturated the miR-520/372/373 family, which controlled the development atmosphere. All cell lines were confirmed to be Mycoplasma-free and progression of RCC both in vitro and in vivo, indicating a and authenticated by short tandem repeat DNA profiling by tumor suppressor role of this miRNA family in RCC. Furthermore, Genewiz Inc. we demonstrated the mechanism by which this miRNA family was regulated by the transcription factor E2F1. The newly iden- miRNA-related reagents, plasmid, and transfection tified E2F1-miR-520/372/373 family–SPOP axis functions as a miRNA mimics and negative controls were purchased from key pathway regulating tumorigenesis of RCC. GenePharma (GenePharma), and lentiviruses to overexpress miR- 520c-3p, miR-372-3p, and miR-373-3p were purchased from Materials and Methods Invitrogen. Cholesterol-conjugated miRNA mimics and siRNA sequences targeting different sites of gene open reading frames Tissue samples, IHC staining, and H&E staining (ORF) of SPOP and E2F1 were designed and synthesized by Tissue microarray (TMA) chips containing a total of 90 pairs of RiboBio Co., Ltd., and the sequences of the siRNAs were listed ccRCC samples and matched normal adjacent tissues (NAT) were in Supplementary Table S2. obtained from Shanghai Biochip Co., Ltd (Shanghai, China). For the protein overexpression assay, a mammalian expression Fresh human surgical specimens (paired normal and cancerous plasmid encoding the human SPOP ORF without the 30-UTR was tissues) were obtained from 30 patients undergoing surgery for purchased from GeneCopoeia. An empty plasmid (pReceiver- ccRCC at Jinling Hospital of Nanjing University (Nanjing, China). M02) served as the negative control. Transfection was performed All patients underwent a tumor resection prior to the receipt of any with Lipofectamine 2000 reagents (Invitrogen) according to the adjunctive therapy. The tissue collection and analyses were manufacturer's protocol. Cells were used for the experiments after approved by the Ethics Committee of Jinling Hospital (Nanjing, transfection for 24 or 48 hours. China), and all participants provided written informed consent. The studies were conducted in accordance with the International RNA isolation and qRT-PCR Ethical Guidelines for Biomedical Research Involving Human Total RNA was extracted from tissues and cells with TRIzol Subjects (CIOMS). All fresh tissues were immediately snap-frozen reagent (Invitrogen) following the manufacturer's protocol. in liquid nitrogen and stored at 80 C until analysis. IHC and Reverse transcription and qRT-PCR for miRNAs were performed hematoxylin and eosin (H&E) staining were performed using as described previously (25). qRT-PCR for SPOP and E2F1 paraffin-embedded sections of biopsies from ccRCC samples and mRNA were performed using a SYBR Premix E Taq Reverse matched NATs according to standard protocols by Cell Signaling Transcription PCR kit (Takara), and the expression levels were Technology. The antibodies used for IHC were as follows: anti- normalized to GAPDH. The sequences of the primers were SPOP antibody (ProteinTech; 16750-1-AP), anti-E2F1 antibody listed in Supplementary Table S3. The reactions were performed (Cell Signaling Technology; #3742), anti-Ki67 antibody (Cell using the LightCycler 480II Detection System (Roche). All Signaling Technology; #9027), anti-vimentin antibody (BD Bio- reactions, including no template controls, were performed in sciences; 550513), and anti-E-cadherin antibody (BD Biosciences; triplicate. 610181). Finally, each sample was scored as 0–3 (0, negative; 1, weak; 2, moderate; 3, strong) according to staining intensities. The Western blotting and antibodies IHC staining was scored by multiple experienced pathologists, Protein was extracted from tissues and cells by RIPA lysis including Q. Xia and Dr. B. Wu at the Department of Pathology, buffer, and protein levels were analyzed via Western blotting Jinling Hospital (Nanjing, China). For survival analyses, the and performed as described previously (26). The antibodies follow-up data of 81 patients were collected, and the optimal used for Western blotting were as follows: anti-SPOP antibody cutoffs of high- and low-expression levels of SPOP protein were (ProteinTech; 16750-1-AP, 1:500), anti-E2F1 antibody (Cell identified as the IHC score 2 as high expression and <2 as low Signaling Technology; #3742, 1:1,000), anti-PTEN antibody expression. The clinical features of the patients are shown in (Cell Signaling Technology; #9559, 1:1,000), anti-DUSP7 anti- Supplementary Table S1. body (ABGENT; AP8450a, 1:500), anti-SP1 antibody (Abcam; ab227383,1:1,000), anti-c-Myc antibody (Cell Signaling Tech- In situ hybridization nology; #13987, 1:1,000), and anti-GAPDH antibody (Santa In situ hybridizations for miR-520c-3p, miR-372-3p, and Cruz Biotechnology; sc-25778, 1:2,000). Protein levels were fi fi miR-373-3p were performed on xed paraf n-embedded normalized to GAPDH. Protein bands were analyzed using sections of biopsies from ccRCC samples and NATs according ImageJ software. to standard protocols by Roche Technology. The oligonucle- otide probes complementary to miR-520c-3p, miR-372-3p, Luciferase reporter assay and miR-373-3p were purchased from the Exonbio Lab Luciferase reporter assay was performed to test the direct (Guangzhou, China). binding of miR-520/372/373 family to the target gene SPOP, as well as the direct binding of E2F1 to the promoter of miR-520 and Cell culture miR-372/373 cluster. The details can be seen in Supplementary The human renal carcinoma cell lines A498 and ACHN were Information. purchased from the Shanghai Institute of Cell Biology, Chinese Academy of Sciences (Shanghai, China). Cells were cultured in CCK-8 assay, EdU incorporation assay, migration, and invasion MEM-NEAA medium (Gibco). The mouse renal carcinoma cell assays line Renca was purchased fromATCCandculturedinDMEM The cell viability in vitro was assessed using the Cell Counting (Gibco). All cell lines were supplemented with 10% FBS Kit-8 (CCK-8; Dojindo). Cell proliferation was further

OF2 Cancer Res; 78(24) December 15, 2018 Cancer Research

E2F1-miR-520/372/373-SPOP Axis in Renal Cancer

A H&E SPOP

×40

×200

NAT ccRCC NAT ccRCC

BCn = 120 Low STOP (n = 20) 2.0 High STOP (n = 61)

100 1.5 90

1.0 80 P = 0.001 70 0.5 60

0.0 of survival (%) Percent 50 IHC Score of SPOP in cytoplasm NAT ccRCC 020406080 100 Time (months)

D E P < 0.001 SPOP 5 GAPDH 4 N1 T1 N2 T2 N3 T3 N4 T4 N5 T5 N6 T6 N7 T7 N8 T8 N9 T9 N10 T10 SPOP 3 GAPDH 2

N11 T11 N12 T12 N13 T13 N14 T14 N15 T15 N16 T16 N17 T17 N18 T18 N19 T19 N20 T20 SPOP 1

GAPDH SPOP protein level Relative 0 NAT ccRCC N21 T21 N22 T22 N23 T23 N24 T24 N25 T25 N26 T26 N27 T27 N28 T28 N29 T29 N30 T30

Figure 1. Upregulation of SPOP in ccRCC tissues is associated with poor prognosis. A, Representative images of H&E staining and IHC staining of SPOP protein in ccRCC tissues and matched NAT (n ¼ 120). Micrographs are shown at original magnifications (40 and 200) as indicated. B, Total IHC score of SPOP in ccRCC tissues and matched NATs; the data are shown as the mean SD (n ¼ 120). C, Kaplan–Meier survival analysis of 81 patients with ccRCC stratified by SPOP expression. D, Western blotting analysis of SPOP expression level in 30 pairs of ccRCC tissues and NATs. GAPDH served as a loading control. E, Quantitative analysis of the density of protein bands in D. , P < 0.001. determined by incorporation of 5-ethynyl-20-deoxyuridine Luciferase-transduced cell generation and in vivo assay (EdU) with the EdU Cell Proliferation Assay Kit (RiboBio), Stable luciferase-expressing ACHN (Luc-ACHN) and Renca which was performed according to the manufacturer's protocol. (Luc-Renca) cells were developed by transduction with a recom- Cell migration ability was assessed by a wound-healing assay, binant lentiviral vector encoding puromycin resistance and firefly and the details were provided in Supplementary Information. luciferase, and the details were described in Supplementary The 8 mm pore polycarbonate membrane inserts (Corning) Information. coated with Matrigel were used to analyze the invasive ability Six-week-old male nude mice and BALB/c were purchased of cells according to the manufacturer's protocol, and the from the Model Animal Research Center of Nanjing University details were provided in Supplementary Information. All assays (Nanjing, China) and maintained under specific pathogen-free were independently repeated at least three times. conditions at Nanjing University (Nanjing, China). ACHN,

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 OF3

Ding et al.

AB3′ UTR of SPOP mRNA 3 NAT ccRCC miR-373-3p miR-146b-3p miR-204 miR-372-3p miR-204 2 miR-211 miR-520c-3p miR-101-3p miR-211

0 0.2 k 0.4 k 0.6 k 0.8 k 1.0 k 1.2 k 1.4 k 1

miR-302b miR-365-5p miR-125a-3p miR-98 Fold change of miRNA level Fold miR-145-5p miR-216b-5p 0

miR-98 miR-365 miR-211miR-302bmiR-101miR-145miR-146b miR-372-3pmiR-373-3p miR-520c-3p C 3.0 3.0 2.0 P < 0.001 P < 0.001 P < 0.001 D 2.5 2.5 1.5 2.0 2.0

1.5 1.5 1.0 miR-520c-3p 1.0 1.0 0.5 0.5 0.5 Relative miR-372-3p level Relative miR-373-3p level Relative

Relative miR-520c-3p level Relative miR-372-3p 0.0 0.0 0.0 NATccRCC NAT ccRCC NAT ccRCC

miR-373-3p

E 3 P < 0.001 1.5 P < 0.001 1.5 P < 0.001 R = −0.752 R = −0.661 R = −0.647 2 1.0 miR-520c-3p 1.0

1 0.5 ccRCC NAT miR-372-3p 0.5 0 0.0

− −0.5 1 change of miR-372-3p Fold change of miR-373-3p Fold 0.0 miR-373-3p Fold change of miR-520c-3p Fold 0246810 0 246810 0 246810 miRNA DAPI Merge Fold change of SPOP protein Fold change of SPOP protein Fold change of SPOP protein

F G SPOP miRNA Control miRNA Control miRNA Mimics miRNA Mimics 1.5 1.5

1.0 1.0 Human Chimp Rhesus Squirrel 0.5 0.5 Mouse Rat 0.0 0.0 hsa-miR-520c-3p: activity (%) luciferase Relative activity (%) luciferase Relative hsa-miR-372-3p: miR-372-3pmiR-373-3p miR-372-3pmiR-373-3p miR-372-3pmiR-373-3p miR-372-3pmiR-373-3p hsa-miR-373-3p: miR-520c-3p miR-520c-3p miR-520c-3p miR-520c-3p MUT-SPOP 3′-UTR: SPOP 3′UTR MUT-SPOP 3′UTR SPOP 3′UTR MUT-SPOP 3′UTR A498 ACHN

H A498 ACHN I A498 ACHN

SPOP SPOP SPOP SPOP

PTEN PTEN PTEN PTEN

DUSP7 DUSP7 DUSP7 DUSP7

GAPDH GAPDH GAPDH GAPDH

SPOP SPOP miR-NC miR-NC Control Control

miR-520c-3pmiR-372-3pmiR-373-3p miR-520c-3pmiR-372-3pmiR-373-3p

miR-372-3pmiR-373-3p + SPOP + SPOP miR-520c-3pmiR-372-3p + SPOPmiR-373-3p + SPOP + SPOP miR-520c-3p + SPOP

OF4 Cancer Res; 78(24) December 15, 2018 Cancer Research

E2F1-miR-520/372/373-SPOP Axis in Renal Cancer

Luc-ACHN, and Luc-Renca cells were infected with the miR-520c- SPOP promotes proliferation, invasion, and migration of renal 3p, miR-372-3p, miR-373-3p, or negative control overexpression carcinoma cells in vitro lentivirus, transfected with the SPOP overexpression plasmid, or We subsequently assessed the effect of SPOP on proliferation, cotransfected the SPOP overexpression plasmid with the miR- invasion, and migration in RCC cells in this study. Two human 520c-3p, miR-372-3p, or miR-373-3p overexpression lentiviruses renal carcinoma cell lines A498 and ACHN had SPOP knocked separately. The details of animal experiments can be seen in down by RNAi (si-SPOP-1 and si-SPOP-2) in which SPOP mRNA Supplementary Information. All animal care and handling pro- expression was reduced to less than 40% of the scrambled control cedures were performed in accordance with the NIH's Guide for (Supplementary Fig. S1A). Western blotting showed that si-SPOP- the Care and Use of Laboratory Animals and were approved by the 1 and si-SPOP-2 resulted in markedly increased protein levels of Institutional Review Board of Nanjing University (Nanjing, PTEN and DUSP7 (Supplementary Fig. S1B and S1C), the down- China). stream genes shown to be degraded by SPOP (11). Knocking down SPOP significantly reduced the cell growth rate in both Chromatin immunoprecipitation assay A498 and ACHN cell lines by CCK-8 assay (Supplementary ChIP assay was performed with Chromatin Immunoprecipita- Fig. S1D). EdU assay, staining to detect nucleotide analogue fi tion (ChIP) Assay Kit (Millipore) according to the manufacturer's incorporation into replicated DNA, further con rmed the cell protocol. ChIP-grade anti-E2F1 antibody (Cell Signaling Tech- proliferative role of SPOP (Supplementary Fig. S1E). Moreover, nology; #3742) or control IgG (Santa Cruz Biotechnology) was transwell invasion and wound-healing assays revealed that knock- used. The primer pairs used for qRT-PCR analysis of ChIP results down of SPOP in cells indeed caused decreased cell invasion and were listed in Supplementary Table S2. migration ability (Supplementary Fig. S1F and S1G). Thus, these results further confirm that SPOP promotes renal carcinoma cell proliferation, invasion, and migration in vitro. Statistical analysis Each experiment was conducted at least three times. The results are presented as the means SD. Spearman rank correlation Prediction of SPOP as a common target of the miR-520/372/373 analysis was used to examine the correlation between two groups. family Kaplan–Meier survival analysis was used to show the overall miRNAs have been demonstrated to play crucial roles in cancer survival of patients with ccRCC with high or low SPOP expression. development by directing the posttranscriptional regulation of The differences were considered statistically significant at P < 0.05 target gene expression (22). To identify potential miRNAs target- using Student t test. ing SPOP, we performed computational predictions with a com- bination of three algorithms: TargetScan (27), miRanda (28), and Results PicTar (29). These candidate miRNAs targeting the SPOP 30-UTR Upregulation of SPOP correlates with poor prognosis in ccRCC are shown in Fig. 2A. We then examined the expression of these To demonstrate the clinical significance of SPOP expression in miRNAs in 30 pairs of ccRCC tissue samples and NATs using qRT- ccRCC, we firstly examined SPOP expression in ccRCC tissue PCR, and found that 6 miRNAs were downregulated in ccRCC specimens from a cohort of 120 patients with ccRCC by IHC tissues (miR-211, miR-302b, miR-145, miR-520c-3p, miR-372- staining using a specific anti-SPOP antibody. We observed that 3p, and miR-373-3p). Notably, miR-520c-3p, miR-372-3p, and 95% (114/120) of ccRCC tissue samples showed significantly miR-373-3p were identified to be the most significantly down- elevated SPOP levels compared with matched normal adjacent regulated among these miRNAs (lower in 86.7%, 86.7%, and tissues (NAT; Fig. 1A and B). Notably, the Kaplan–Meier survival 93.3% of ccRCC tissues compared with NATs, respectively; Fig. 2B analysis showed that patients with ccRCC with high SPOP expres- and C). Furthermore, miR-520c-3p, miR-372-3p, and miR-373- sion had shorter overall survival (Fig. 1C). We further confirmed 3p share the same seed sequences, and belong to a miRNA family. the expression level of SPOP protein in 30 paired tissue specimens Several reports have showed that miRNAs from miR-520/372/373 by Western blotting and found coherently higher expression of family can target same genes and play crucial roles in the pro- SPOP in ccRCC tissues than in pair-matched NATs (Fig. 1D and E). gression of hepatic carcinoma and breast cancer (30, 31). Thus, in These results demonstrate that SPOP expression levels are this study, we chose to focus on the miRNA family to explore markedly upregulated in ccRCC tissues and correlate with poor whether it was involved in the regulation of SPOP during RCC prognosis in ccRCC, suggesting that SPOP may play a key role in progression. We further performed in situ hybridization to con- ccRCC tumorigenesis. firm the downregulated expression of the miR-520/372/373

Figure 2. Prediction and validation of SPOP as a common target of the miR-520/372/373 family. A, Schematic overview regarding these 10 candidate miRNAs targeting SPOP 30-UTR. B, Expression of these candidate miRNAs in 30 pairs of ccRCC tissue samples and NATs using qRT-PCR; U6 served as the internal reference. The DD relative miRNA expression was calculated with the equation 2 Ct, where DC ¼ ( C C )andDDC ¼ DC DC . C, Relative expression of the t tmiRNA tU6 t tccRCC tNAT miR-520/372/373 family in 30 pairs of ccRCC tissue samples and NATs using qRT-PCR. D, Representative images of the expression of miR-520/372/373 family in paired tissues using in situ hybridization (n ¼ 5). E, Spearman rank correlation scatter plot of the fold change in miR-520/372/373 family levels and SPOP protein levels in ccRCC tissues (n ¼ 30). F, Schematic depicting the highly conserved binding site for the miR-520/372/373 family in the 30-UTR of SPOP mRNA and mutation of the indicated nucleotides in seed sequences. G, Luciferase reporter assays of the effect of the miR-520/372/373 family on wild-type and mutant reporter genes in A498 and ACHN cells. H, Western blotting analysis of SPOP protein levels, as well as its downstream protein PTEN and DUSP7, in miR-NC, miR-520c-3p-, miR-372-3p-, and miR-373-3p–transfected A498 and ACHN cells. I, Western blotting analysis of SPOP protein levels, as well as its downstream protein PTEN and DUSP7, in miR-NC plus vector control (control), miR-NC plus SPOP vector (SPOP), SPOP vector plus miR-520c-3p, SPOP vector plus miR-372-3p, and SPOP vector plus miR-373-3p–transfected A498 and ACHN cells. In H and I, blots are representative of three independent experiments, and GAPDH served as the loading control. , P < 0.05; , P < 0.01; , P < 0.001.

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 OF5

Ding et al.

ADmiR-NC Control miR-520c-3p SPOP miR-372-3p miR-520c-3p + SPOP A498 miR-373-3p A498 miR-372-3p + SPOP miR-373-3p + SPOP 1.5 4 ACHN ACHN 3 1.0

miR-NC Control SPOP 2 miR-372-3p miR-373-3p miR-520c-3p 0.5 miR-520c-3p+ SPOPmiR-372-3p+ SPOPmiR-373-3p+ SPOP B E 1

0.0 0 A498 Relative EdU positive cells EdU positive Relative A498 ACHN cells EdU-positive Relative A498 ACHN

1.5 4 ACHN ACHN 3 1.0 miR-NC Control SPOP miR-520c-3p miR-372-3p miR-373-3p miR-520c-3p+ SPOPmiR-372-3p+ SPOPmiR-373-3p+ SPOP 2 CF0.5 1 Relative invasion Relative Relative invasion Relative

0.0 0 A498 ACHN A498 A498 ACHN A498 80 100 60

40 ACHN A498 ACHN 50 20 wound healed (%) wound wound healed (%) wound Relative percentage of Relative miR-NC 0 SPOP percentage of Relative 0 Control A498 ACHN miR-520c-3p miR-372-3p miR-373-3p A498 ACHN miR-520c-3p+ SPOPmiR-372-3p+ SPOPmiR-373-3p+ SPOP

Figure 3. The miR-520/372/373 family inhibits renal carcinoma cell growth, invasion, and migration by suppressing SPOP in vitro. A–C, Representative images and histogram statistics from EdU, invasion, and wound-healing assays of miR-NC-, miR-520c-3p-, miR-372-3p-, and miR-373-3p–transfected A498 and ACHN cells. D–F, Representative images and histogram statistics from EdU, invasion, and wound-healing assays of control, SPOP, SPOP plus miR-520c-3p, SPOP plus miR-372-3p, and SPOP vector plus miR-373-3p–transfected A498 and ACHN cells. The results are shown as the mean SD from three independent experiments. , P < 0.05; , P < 0.01.

family in ccRCC tissues (Fig. 2D). Importantly, Spearman rank fected by the overexpression of miR-520c-3p, miR-372-3p, or correlation analysis revealed that there was a significantly inverse miR-373-3p (Fig. 2G). correlation between miR-520/372/373 family expression levels We then assessed the effects of the miR-520/372/373 family on and SPOP protein levels in ccRCC tissue samples (R ¼0.752, P < SPOP protein expression levels in A498 and ACHN cells after 0.001; R ¼0.661, P < 0.001; R ¼0.647, P < 0.001; Fig. 2E). On transient transfection with miRNA mimics. The transient trans- the basis of these findings, we reasoned that miR-520/372/373 fection efficiencies of miRNA mimics in A498 and ACHN cell lines may be a potential regulator of SPOP. were detected by qRT-PCR (Supplementary Fig. S2A and S2B). The expression of SPOP protein was significantly inhibited, whereas Validation that SPOP is posttranscriptionally regulated by the PTEN and DUSP7 proteins were markedly increased in cancer cells miR-520/372/373 family by the transfection with mimics of miR-520c-3p, miR-372-3p, or To further validate that SPOP is regulated by the miR-520/372/ miR-373-3p (Fig. 2H; Supplementary Fig. S2C). However, SPOP 373 family through the presumed binding sites in the 30-UTR of mRNA levels were unchanged after miR-520/372/373 family SPOP mRNA, a luciferase reporter assay was performed. The 200- transfection (Supplementary Fig. S2D and S2E). We next sought bp fragment in the 30-UTR of SPOP mRNA containing the pre- to determine whether miR-520/372/373 family can attenuate the sumed miR-520/372/373 family binding sites was inset down- inhibitory effects of SPOP on PTEN and DUSP7. We performed stream of the firefly luciferase gene in a reporter plasmid (Fig. 2F). rescue experiments by transfecting A498 and ACHN cells with a The resulting plasmid was transfected into A498 and ACHN plasmid overexpressing SPOP or cotransfected with an SPOP cells along with miR-NC (a scrambled control), miR-520c-3p, overexpression plasmid plus each of the miRNA mimics simul- miR-372-3p, or miR-373-3p individually. As anticipated, miR- taneously. Efficient overexpression of SPOP in these cells was 520c-3p, miR-372-3p, and miR-373-3p significantly reduced achieved, and the total SPOP protein level was markedly the luciferase reporter activity compared with that of miR-NC reduced when cotransfected with miRNA mimics (Fig. 2I; (Fig.2G).Furthermore,weengineered point mutations into the Supplementary Fig. S2F). As expected, ectopic expression of corresponding binding sequence in the 30-UTR of SPOP to these three miRNAs in renal carcinoma cells dramatically eliminate the putative binding sites (Fig. 2F). The luciferase counteracted the inhibitory effects of SPOP on PTEN and activity of this mutation luciferase reporter plasmid was unaf- DUSP7 (Fig. 2I; Supplementary Fig. S2F). Overall, we conclude

OF6 Cancer Res; 78(24) December 15, 2018 Cancer Research

E2F1-miR-520/372/373-SPOP Axis in Renal Cancer

ABC 0.6 Implanted with Control ACHN cells

ACHN cells overexpressing miR-NC miR-520c-3p 0.4 ACHN cells overexpressing Sample collection miR-372-3p and analysis 0.2 ACHN cells overexpressing mass (g) Tumor miR-373-3p miR-520c-3p 0.0 miR-372-3p

0 612182430 miR-NC (Days) miR-520c-3pmiR-372-3pmiR-373-3p miR-373-3p

DEmiR-NC F )

3 miR-520c-3p 1,500 miR-372-3p SPOP miR-373-3p H&E 1,250 1,000 PTEN 750 Ki67 DUSP7 500 250 GAPDH 0 SPOP 0 6 12 18 24 30 Tumor mean volume (mm mean volume Tumor Days after injection miR-NC miR-NC miR-520c-3p miR-372-3p miR-373-3p miR-520c-3pmiR-372-3pmiR-373-3p

H I 1.5 G Control ACHN cells

ACHN cells overexpressing SPOP Control Implanted with 1.0 ACHN cells overexpressing miR-520c-3p and SPOP SPOP

ACHN cells overexpressing 0.5

miR-372-3p and SPOP mass (g) Tumor ACHN cells overexpressing miR-373-3p and SPOP miR-520c-3p 0.0 + SPOP

ControlSPOP miR-372-3p 0 6 12 18 24 30 +SPOP (Days) miR-372-3pmiR-373-3p + SPOP + SPOP miR-373-3p miR-520c-3p + SPOP + SPOP

Control JKSPOP L

) miR-520c-3p + SPOP SPOP 3 1,500 miR-372-3p + SPOP miR-373-3p + SPOP H&E 1,250 PTEN 1,000 750 Ki67 DUSP7 500 GAPDH 250 0 SPOP 0 6 12 18 24 30 Control SPOP Tumor mean volume (mm mean volume Tumor Control SPOP miR-372-3p miR-373-3p miR-520c-3pmiR-372-3pmiR-373-3p Days after injection miR-520c-3p +SPOP + SPOP + SPOP +SPOP + SPOP + SPOP

Figure 4. The miR-520/372/373 family suppresses renal tumor growth by inhibiting SPOP in a subcutaneous xenograft model. A, Flowchart of the experimental design. ACHN cells infected with miR-NC, miR-520c-3p, miR-372-3p, or miR-373-3p overexpression lentivirus were implanted subcutaneously into nude mice (6 mice per group), and tumor growth was evaluated on day 30 after implantation. B, Representative images of tumors from mice of each group. C, Quantitative analysis of the tumor weights in B. D, Time course of tumor growth in implanted mice and tumor volume were measured every 6 days for 30 days after implantation. E, Representative images of H&E staining and IHC stainingforSPOPandKi67oftumorsfromtheimplantedmice.F, Western blotting analysis of SPOP, PTEN, and DUSP7 protein levels in tumors from the implanted mice. G, Flowchart of the rescue experimental design. ACHN cells transfected with a plasmid for overexpressing SPOP or cotransfected with SPOP overexpression plasmid plus each of three miRNA lentiviral expression vectors were implanted subcutaneously into nude mice (6 mice per group), and tumor growth was evaluated on day 30 after implantation. H, Representative images of tumors from mice of each group. I, Quantitative analysis of the tumor weights in H. J, Time course of tumor growth in implanted mice, and tumor volume was measured every 6 days for 30 days after implantation. K, Representative images of H&E staining and IHC staining for SPOP and Ki67 of tumors from the implanted mice. L, Western blotting analysis of SPOP, PTEN, and DUSP7 protein levels in tumors from the implanted mice. , P < 0.05; , P < 0.01.

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 OF7

Ding et al.

miR-NC AC DmiR-520c-3p 0.5 miR-373-3p 3 miR-373-3p 0.4 H&E

0.3 2 0.2

Vimentin Tumor mass (g) Tumor 0.1 1

0.0 cells positive Relative 0 E-cadherin miR-NC miR-NC miR-372-3pmiR-373-3p miR-520c-3p miR-372-3p miR-373-3p Vimentin miR-520c-3p E-cadherin

Control BE0.5 FSPOP miR-520c-3p + SPOP 0.4 3 miR-373-3p + SPOP H&E 0.3 miR-373-3p + SPOP

0.2 2

Tumor mass (g) Tumor 0.1 Vimentin 1 0.0 Relative protein cells Relative

ControlSPOP 0 E-cadherin

Control SPOP Vimentin miR-520c-3p miR-372-3p miR-373-3p E-cadherin miR-520c-3pmiR-372-3p miR-373-3p+ SPOP + SPOP + SPOP + SPOP + SPOP + SPOP GIK

miR-NC H&E 2.5 3.0 2.0 miR-NC 2.5 ×109 miR-NC 1.5 2.0 miR-520c-3p miR-520c-3p ×107 miR-520c-3p 1.5 1.0 1.0 miR-372-3p 0.5 0.5 miR-372-3p Radiance 2 Radiance (p/sec/m /sr) (p/sec/m2/sr) miR-373-3p miR-372-3p miR-373-3p miR-373-3p 0 1234 (Week)

HJL

Control

H&E Control SPOP

SPOP 2.0 Control SPOP

1.5 miR-520c-3p miR-520c-3p + SPOP + SPOP ×108

1.0 miR-372-3p + SPOP (p/sec/m Radiance miR-520c-3p miR-372-3p miR-373-3p miR-372-3p 0.5 0.2 0.4 0.6 0.8 1.0 + SPOP + SPOP + SPOP + SPOP miR-373-3p

+ SPOP × 10 2 /sr) Radiance 8 miR-373-3p 0 1234(p/sec/m2/sr) + SPOP (Week)

Figure 5. The miR-520/372/373 family suppresses renal tumor invasion and migration by inhibiting SPOP in vivo. A, Quantitative analysis of the tumor weights of mice implanted with miR-NC, miR-520c-3p, miR-372-3p, or miR-373-3p–overexpressing lentivirus–infected ACHN cells. Constructed ACHN cells were injected into the subrenal capsule of the left kidney of nude mice (5 mice/group). B, Quantitative analysis of the tumor weights of mice implanted with a plasmid for overexpressing SPOP or cotransfected with SPOP overexpression plasmid plus each of three miRNA lentiviral expression vector-infected ACHN cells. C, Representative images of H&E and IHC staining for vimentin and E-cadherin of tumors from the implanted mice. (Continued on the following page.)

OF8 Cancer Res; 78(24) December 15, 2018 Cancer Research

E2F1-miR-520/372/373-SPOP Axis in Renal Cancer

that the miR-520/372/373 family specifically regulates SPOP Supplementary Fig. S5A). Furthermore, H&E staining of the protein expression posttranscriptionally. xenograft tissues showed less cell invasion and more confluent necrotic areas in tumors of each miRNA group compared with The miR-520/372/373 family inhibits renal carcinoma cell those of the control group (Fig. 4E). The levels of the three growth, invasion, and migration by suppressing SPOP in vitro miRNAs in tumors were confirmed by qRT-PCR (Supplementary To explore the biologic effects of the miR-520/372/373 family Fig. S5B), Moreover, the SPOP protein level was greatly reduced, on renal carcinoma cells, miR-520/372/373 family expression and PTEN and DUSP7 protein levels were markedly elevated in was manipulated in vitro, and function analyses were conducted in the miRNA-transfected tumors compared with the miRNA control both A498 and ACHN cell lines by transient transfection of group (Fig. 4F; Supplementary Fig. S5C). miRNA mimics. CCK-8 assay revealed that the proliferation In addition, a rescue experiment in vivo was performed. Figure rate of cells was markedly decreased upon overexpression of 4G shows the design flow chart. In the rescue experiment, ACHN miR-520c-3p, miR-372-3p, and miR-373-3p (Supplementary cells were transfected with a plasmid for overexpressing SPOP or Fig. S3A). Consistent with the results from the CCK-8 assay, the cotransfected with an SPOP overexpression plasmid plus each of percentages of EdU-positive cells were significantly lower in three miRNA lentiviral expression vectors individually. Then, the cells transfected with any of these three miRNAs in both cell lines constructed ACHN cells were implanted subcutaneously into compared with the control cells, which is in contrast to SPOP nude mice, and 30 days after implantation, tumor growth was (Fig. 3A). Furthermore, treatment of cells with these miRNAs evaluated. The tumor weights and volumes were dramatically significantly suppressed the migration and invasion abilities of increased upon overexpressing SPOP compared with the controls both cell lines (Fig. 3B and C). Moreover, functional restoration but significantly attenuated in the coexpressed miRNA and SPOP assays showed that the miR-520/372/373 family significantly groups compared with the SPOP-alone overexpression group reversed SPOP-led promotion of A498 and ACHN cell prolifer- (Fig. 4H–J). Consistently, the percentage of Ki67-positive cells ation, invasion, and migration (Supplementary Fig. S3B; Fig. 3D– significantly increased in the SPOP-overexpressed tumors, and F). In addition to transient transfection of miRNA mimics, func- these miRNAs effectively reversed the stimulating effect of SPOP tion analyses were also conducted in cells stably transfecting with on tumor growth (Fig. 4K; Supplementary Fig. S5D). H&E stain- lentiviral expression vectors. The stable transfection efficiencies of ing in coexpressed miRNA and SPOP groups also showed less cell lentiviral expression vectors of these three miRNAs in renal invasion and more confluent necrotic areas than those of the carcinoma cell lines were detected by qRT-PCR (Supplementary SPOP-overexpressing group (Fig. 4K). The levels of three miRNAs Fig. S4A). Function analyses showed that stably transfecting cells in tumors were quantified by qRT-PCR (Supplementary Fig. S5E), with these miRNAs significantly suppressed the proliferation, and Western blotting further confirmed that the miRNA family migration, and invasion abilities of both cell lines (Supplemen- could reduce the total SPOP protein level, thus restoring the tary Fig. S4B–S4E), which was consistent with the results of cells downstream proteins PTEN and DUSP7 in vivo (Fig. 4L; Supple- after transient transfection of miRNA mimics. Collectively, our mentary Fig. S5F). data strongly suggest that the miR-520/372/373 family suppresses renal carcinoma cell proliferation and migration by directly The miR-520/372/373 family suppresses renal tumor invasion targeting SPOP. and metastasis by inhibiting SPOP in vivo To investigate the effect of the miR-520/372/373 family on The miR-520/372/373 family suppresses renal tumor growth by renal carcinoma cell invasion in vivo, an orthotopic xenograft inhibiting SPOP in vivo modelofhumanrenalcellcarcinomainnudemicewas To confirm the growth-inhibitory effect of the miR-520/372/ introduced. The nine constructed ACHN cell groups as 373 family on renal carcinoma cells in vivo, we established a described above were implanted into the left kidney of nude subcutaneous xenografted model of human renal carcinoma in mice separately. The left kidney and paired right kidney in each nude mice. Figure 4A illuminates the experimental design. In this mouse were compared after 30 days of implantation. As shown experiment, ACHN cells infected with miR-NC, miR-520c-3p, in Fig. 5A and B, tumor weights were significantly reduced in miR-372-3p, or miR-373-3p overexpression lentiviruses were miRNA-overexpressing groups compared with those of the implanted subcutaneously into nude mice, and tumor growth miRNA control group regardless of co-xpression of SPOP, was evaluated on day 30 after implantation. As shown in Fig. 4B– which was consistent with the results of the subcutaneous D, tumor weights and volumes were greatly reduced in miR-520c- xenografted model. Then, the tumors were histologically exam- 3p, miR-372-3p, and miR-373-3p stably transfected tumors. ined by H&E staining, and markedly reduced invasiveness of Consistently, the percentage of Ki67-positive cells was also the miR-520c-3p-, miR-372-3p-, and miR-373-3p–expressing decreased in these three miRNA-transfected tumors (Fig. 4E; tumors was observed (Fig. 5C). Tumors in the control group

(Continued.) D, Relative protein levels of vimentin and E-cadherin in tumor regions by IHC score analysis. E, Representative images of H&E and IHC staining for vimentin and E-cadherin of tumors from the implanted mice. F, Relative protein levels of vimentin and E-cadherin in tumor regions by IHC score analysis. G, Luc-Renca cells infected with miR-NC, miR-520c-3p, miR-372-3p or miR-373-3p overexpression lentivirus were injected into the subrenal capsule of the left kidney of BALB/c mice (5 mice/group). The tumor development in each group was evaluated every week by bioluminescence imaging system. Representative bioluminescent images acquired at the indicated time points are shown. H, Luc-Renca cells transfected with a plasmid for overexpressing SPOP or cotransfected with SPOP overexpression plasmid plus each of the three miRNA lentiviral expression vectors were injected into the subrenal capsule of the left kidney of BALB/c mice (5 mice/group). The tumor development in each group was evaluated every week by bioluminescence imaging system. Representative bioluminescent images acquired at the indicated time points are shown. I, Bioluminescent images of lungs of BALB/c mice were acquired on day 28 after implantation. J, Bioluminescent images of lungs of BALB/c mice were acquired on day 28 after implantation. K, Representative H&E staining images of lungs of BALB/c mice as in I. L, Representative H&E staining images of lungs of BALB/c mice as in J. , P < 0.05; , P < 0.01; , P < 0.001.

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 OF9

Ding et al.

had widely spread invasive fronts characterized by an infiltrated miR-372/373 using the MAPPER Search Engine (32). As a result, tumor–stroma interface and the presence of tumor cells inter- three candidate transcription factors were predicted (Fig. 6A). mingled in the stromal side, whereas in contrast, tumors in the To examine the role of these transcription factors in regu- miRNA groups were well-contained with an intact tumor–stroma lating the miR-520/372/373 family, their siRNAs were synthe- interface (Fig. 5C). Furthermore, we examined changes in expres- sized and then transfected into ACHN cells (Supplementary sion levels of key EMT-related markers in tumors of each Fig. S7A–S7C). qRT-PCR indicated that the expression levels of group. IHC staining experiments showed that miR-520/372/ miR-520c-3p, miR-372-3p, and miR-373-3p were markedly upre- 373 family overexpression reduced the expression of the mesen- gulated simultaneously only when E2F1 was knocked down, chymal marker vimentin but increased the expression of epithelial suggesting that E2F1 is a potentially common upstream regulator marker E-cadherin (Fig. 5C and D). Moreover, coexpression of of the three miRNAs (Fig. 6A). each of the three miRNAs with SPOP could reverse the expression To investigate whether E2F1, as a sequence-specific transcrip- of vimentin and E-cadherin and attenuate the proinvasive effect tional suppressor, can recognize and bind to the E2F1-responsive of SPOP to some degree (Fig. 5E and F). elements and negatively regulate the expression of these miRNAs, To further assess the suppressive role of the miR-520/372/373 we performed ChIP assays using an antibody against E2F1 in family in tumor metastasis in vivo, a new orthotopic xenograft ACHN cells. As shown in Fig. 6B and C; Supplementary Fig. S7D, model with high efficiency in metastasis was employed. First, robust qRT-PCR product enrichment of the region was around the luciferase-labeled Renca cells (Luc-Renca) were forced to express predicted binding sites (A-site-4 for miR-520c-3p and B-site-1 miRNAs and/or SPOP as the foregoing. Thereafter, these Luc- for miR-372/373). Subsequently, we cloned the sequences of Renca cells were implanted into the left kidney of BALB/c mice. E2F1-binding sites A-site-4 andB-site-1intotheupstream The tumor development in each group was evaluated every region of a firefly luciferase reporter gene and then transfected week by a bioluminescence imaging system. Mice receiving the constructed plasmids into ACHN cells. Luciferase reporter miRNA-overexpressing cells showed significantly reduced kidney assays revealed that knockdown of the expression of E2F1 cancer growth signals (photon radiance) compared with the accelerated the transcription of fireflyluciferaseinA-site-4- mice receiving control cells (Fig. 5G; Supplementary Fig. S5G). and B-site-1–containing plasmids, whereas fireflyluciferase When comparing with SPOP-overexpressing group, coexpressing activity was unaffected after knockdown of the expression of miRNA with SPOP led to a markedly declining growth signal E2F1 when the binding sites were mutated (Fig. 6D). These (Fig. 5H; Supplementary Fig. S5H), further confirming that the results suggest that E2F1 binds to specific transcription factor miR-520/372/373 family could reverse the stimulate effect of binding sites of the miR-520/372/373 family and suppresses SPOP on tumor growth. Twenty-eight days after implantation, we transcription in renal carcinoma cells. measured the lung metastatic lesions using the bioluminescence Subsequently, we sought to examine the effect of E2F1 on SPOP imaging system. As shown in Fig. 5I and J and Supplementary and its downstream protein PTEN and DUSP7. siRNA-mediated Fig. S5I and S5J, the mean bioluminescence intensity in the lung suppression of ectopic E2F1 significantly downregulated SPOP was significantly decreased in miRNA-overexpressing groups com- but upregulated PTEN and DUSP7 (Fig. 6E), whereas si-E2F1 did pared with that of the control miRNA group, whereas it was not change the expression of SPOP, PTEN, or DUSP7 when markedly increased in the SPOP-overexpressing group in contrast cotransfected with anti-miRs of the miR-520/372/373 family (Fig. with that of the control vector group. Notably, there was a signif- 6E, Supplementary Fig. S7E), suggesting that E2F1 could indi- icantly reversal in the metastatic ratio when coexpressing each of rectly regulate the expression of the SPOP signaling pathway the three miRNAs with SPOP in tumors. In addition, H&E staining through the miR-520/372/373 family. in lungs further confirmed these results (Fig. 5K and L). Taken We then further investigated whether E2F1 is involved in the together, these results indicate that miR-520/372/373 family act as pathogenesis of human ccRCC through the miR-520/372/373 a tumor suppressor, restraining tumor invasion and metastasis via family based on clinical samples. We examined E2F1 protein level inhibiting the expression of the oncogene SPOP in RCC in vivo. in 30 pairs of ccRCC tissues and their NATs by IHC and Western We next tested the feasibility of using the miR-520/372/373 blotting analysis. We observed that E2F1 protein was significantly family therapeutically to treat RCC. First, luciferase-labeled ACHN elevated in the ccRCC tissues (Fig. 6F–H). Notably, Spearman cells (Luc-ACHN) were injected into the subrenal capsule of the rank correlation analysis showed a good reverse correlation left kidney of nude mice. Two weeks after the implantation, between levels of E2F1 protein and the miR-520/372/373 family cholesterol-conjugated miRNA mimics (agomir) were then deliv- in ccRCC tissues (R ¼0.505, P ¼ 0.004; R ¼0.508, P ¼ 0.004; ered via tail vein every 4 days for 16 days (Supplementary Fig. S6A). R ¼0.476, P ¼ 0.008; Fig. 6I). These results indicated that the Four treatment groups were involved: agomir-miR-NC, agomir- decreased levels of the miR-520/372/373 family are, at least in miR-520c-3p, agomir-miR-372-3p, and agomir-miR-373-3p. As part, attributed to the overexpression of E2F1 in ccRCC. Thus, we shown in Supplementary Fig. S6B and S6C, the tumor sizes and conclude that E2F1 specifically regulates miR-520/372/373 fam- lung metastasis ratio were dramatically decreased after agomir- ily expression transcriptionally and indirectly promotes SPOP miR-520c-3p, agomir-miR-372-3p, or agomir-miR-373-3p treat- expression. ment compared with those of the agomir-miR-NC treatment group. It has been demonstrated that let-7a-d, let-7i, miR-15b-16-2, Combining all of these results, the miR-520/372/373 family could miR-106b-25, and miR-17-20a clusters are direct targets of be a potential target for therapy of human renal cell carcinoma. E2F1/3 (33, 34). Therefore, we examined the expression levels of 10 E2F1/3 induced miRNAs including let-7a, let-7d, let-7i, Expression of miR-520/372/373 is regulated by E2F1 miR-106b, miR-93, miR-25, miR-15b, miR-16-2, miR-17 and To test the transcriptional regulatory mechanisms of the miR- miR-20a, in 20 pairs of ccRCC tissues and their NATs by qRT- 520/372/373 family, we predicted common putative transcrip- PCR analysis. We observed that 7 miRNAs including let-7d, tion factors binding to the promoter regions of miR-520c and miR-106b, miR-93, miR-15b, miR-16-2, miR-17 and miR-20a,

OF10 Cancer Res; 78(24) December 15, 2018 Cancer Research

E2F1-miR-520/372/373-SPOP Axis in Renal Cancer

si-NC si-cMyc A B 0 3040 4398 7340 8210 4 si-SP1 si-E2F1

3 miR-520c A-Site-1 A-Site-2 A-Site-3 A-Site-4 2 (TTCGCCC) (TTGCCTC) (ATGCCGC) (TTGCCTC)

1 0 2650 7661 9772 Relative expression level expression Relative 0 miR-372/373 B-Site-1 B-Site-2 B-Site-3 (TTGGAGC) (TTGCAGC) (TTGCCGC) miR-520c-3pmiR-372-3pmiR-373-3p

CDE IgG 5 si-NC 40 E2F1 si-E2F1 E2F1 4 30 3 SPOP 20 2 PTEN

10 1 DUSP7 Relative of input (%) Relative Relative luciferase activity luciferase Relative 0 0 GAPDH WT WT MUT MUT A-Site-1A-Site-2A-Site-3A-Site-4B-Site-1B-Site-2B-Site-3 si-NC A-Site-4 B-Site-1 si-E2F1si-E2F1 F +anti-miRs E2F1 G 4 P < 0.001 GAPDH

N1 T1 N2 T2 N3 T3 N4 T4 N5 T5 N6 T6 3 E2F1

GAPDH 2

N7 T7 N8 T8 N9 T9 N10 T10 N11 T11 N12 T12 N13 T13 N14 T14 N15 T15 N16 T16 N17 T17 N18 T18 1 E2F1 Relative E2F1 protein level Relative GAPDH 0 NAT ccRCC N19 T19 N20 T20 N21 T21 N22 T22 N23 T23 N24 T24 N15 T25 N26 T26 N27 T27 N28 T28 N29 T29 N30 T30 H

×200 I

3 P = 0.004 1.5 P = 0.004 1.5 P = 0.008 R = −0.505 R = −0.508 R = −0.476 2 1.0 1.0

1 0.5 0.5

×400 0 0.0 0.0

−1 −0.5 −0.5 Fold change of miR-372-3p Fold change of miR-373-3p Fold Fold change of miR-520c-3p Fold 0 5 10 15 0 5 10 15 0 5 10 15 Fold change of E2F1 protein Fold change of E2F1 protein Fold change of E2F1 protein NAT ccRCC

Figure 6. The miR-520/372/373 family was regulated by E2F1. A, qRT-PCR analysis of the expression levels of miR-520c-3p, miR-372-3p, and miR-373-3p in ACHN cells transfected with si-NC, si-E2F1, si-SP1, or si-cMyc. B, Schematic illustrating the four putative E2F1-binding motifs (A-site-1, 2, 3, and 4) in the miR-520c promoter region and three putative E2F1-binding motifs (B-site-1, 2, and 3) in the miR-372/373 cluster promoter region. C, Direct binding of E2F1 to promoter regions of miR-520c and the miR-372/373 cluster indicated by qRT-PCR–based ChIP assays. D, The wild-type or mutated sequence of E2F1-binding sites (A-site-4 and B-site-1) was inserted into the upstream region of a firefly luciferase reporter gene. Luciferase reporter assays revealed that knockdown of the expression of E2F1 accelerated the transcription of firefly luciferase in A-site-4- and B-site-1–containing plasmids, whereas they showed no effect when the binding sites were mutated. E, Western blotting analysis of the protein levels of E2F1, SPOP, PTEN, and DUSP7 in ACHN cells transfected with si-E2F1 or cotransfected with si-E2F1 and anti-miRNAs of the miR-520/372/373 family. F, Western blotting analysis of the E2F1 expression level in 30 pairs of ccRCC tissues and NATs. GAPDH served as the loading control. G, Quantitative analysis of the density of protein bands in F. H, Representative images of IHC staining of E2F1 protein in ccRCC tissues and NATs (n ¼ 30). Micrographs are shown at original magnifications (200 and 400) as indicated. I, Spearman rank correlation scatter plot of the fold change in miR-520/372/373 family levels and E2F1 protein levels in ccRCC tissues (n ¼ 30). , P < 0.05; , P < 0.01; , P < 0.001. www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 OF11

Ding et al.

were significantly upregulated in ccRCC tissues (Supplementary E2F1 is the first-described and most-studied member of the E2F Fig. S7F). These results indicate that E2F1 may also regulate family of transcription factors containing one or more conserved other miRNAs in RCC progression; however, the function and DNA binding sites that bind target promoters to regulate their underlying mechanism remain to be further elucidated. expression and are involved in numerous important biological processes (46, 47). Moreover, the von Hippel-Lindau (VHL) Discussion protein, a member of the E3 ubiquitin ligase complex, could suppress E2F1 expression at both the mRNA and protein level in In this study, we firstly demonstrated the oncoprotein role of human ccRCC tissue and RCC cells (48). However, inactivating SPOP in RCC carcinogenesis. Moreover, we revealed that SPOP mutations or silencing of the VHL tumor suppressor gene was posttranscriptionally regulated by the miR-520/372/373 occurred in 80% of ccRCC cases, resulting in enrichment of the family, which restrained the proliferation, invasion, and metas- E2F1 protein. Numerous studies reveal that E2F1 is equipped with tasis of RCC cells, and effectively attenuated the promotive effect an ability to regulate tumor development in many types of human of SPOP on tumorigenesis of RCC in vitro and multiple in vivo cancer (49, 50). It has been shown that aberrant expression of models. Furthermore, we demonstrated that the decreased expres- E2F1 is a key regulator to drive human RCC invasion and sion of the miRNA family was mediated by transcription factor metastasis (51, 52). Nevertheless, the molecular mechanism by E2F1. Our data demonstrate the regulation and mechanism of the which E2F1 promotes the development of RCC is still unclear. In miR-520/372/373 family on SPOP and the possibility of the this study, we found that E2F1 transcriptionally suppressed the E2F1–miR-520/372/373 family–SPOP axis as a potential thera- miR-520/372/373 family by binding to the promoters of peutic avenue for the treatment of RCC. miR-520c and the miR-372/373 cluster, thus indirectly regulated The miR-520/372/373 family was initially reported to be spe- SPOP signaling expression. Furthermore, we observed that the cifically expressed in human embryonic stem cells (35, 36), and E2F1 level was dramatically overexpressed in ccRCC tissues com- emerging evidence links the miRNA family with cancer progres- pared with NATs and negatively correlated with the expression of sion. In particular, the miR-520/372/373 family seems to play a the miR-520/372/373 family. These data demonstrated that E2F1 dual role in tumor development depending on the different cancer was a crucial regulator of the miR-520/372/373 family. Never- types and contexts. miR-372 and miR-373 have been identified as theless, we cannot exclude the possibility that E2F1 may also oncogenes promoting the proliferation and tumorigenesis in regulate other miRNAs in RCC progression. In this study, human testicular germ cell tumors (37). The oncogenic roles of we observed that 7 E2F1/3-induced miRNAs, including let-7d, the miRNA family were also observed in human breast cancer (38), miR-106b, miR-93, miR-15b, miR-16-2, miR-17, and miR-20a, gastric cancer (39), hepatocellular carcinoma (HCC; ref. 40) and were significantly upregulated in ccRCC tissues when compared colorectal carcinoma (41). However, contradictory findings were with NATs. However, the function and underlying mechanism present in breast cancer and HCC. It was also demonstrated that remain to be further illuminated. the miR-520/373 family has a tumor-suppressive role attributed to As summarized in our working model in Fig. 7, our data inhibition of the NFkB and TGFb signaling pathways (30). Fur- demonstrated a critical tumor suppressor role of the miR-520/ thermore, low miR-372 expression was correlated with poor 372/373 family in RCC by regulating the SPOP pathways. More- prognosis and tumor metastasis in HCC (42). Moreover, the miRNA family was also shown to be a tumor suppressor gene in cervical carcinoma (43), endometrial adenocarcinoma (44) and ovarian carcinoma (45). Nevertheless, the role of the miR-520/ 372/373 family in RCC was largely unknown. In this study, we identified the miR-520/372/373 family as a tumor suppressor by miR-520c miR-372/373 directly inhibiting the expression of SPOP protein in RCC. In clinical samples, the miR-520/372/373 family was statistically lower in ccRCC tissues and negatively correlated with SPOP. Furthermore, the miR-520/372/373 family was found to target 0 the 3 -UTR of SPOP and suppress the expression of SPOP, thus 5′ AAAAA 3′ promoting the expression of PTEN and DUSP7, two downstream genes demonstrated to be degraded by SPOP, resulting in the decreased proliferation, migration, and invasion of RCC cell lines. Moreover, orthotopic delivery of miR-520c-3p, miR-372-3p, and miR-373-3p markedly blocked SPOP expression and reversed the promotive effect of SPOP on tumor growth and metastasis, as shown by significantly reduced weights and volumes, declined percentage of Ki67-positive cells, decreased expression of vimentin, and increased expression of E-cadherin as well as attenuated lung metastatic lesions when coexpressing each of the three miRNAs with SPOP in tumors. Taken together, these results demonstrated that inducing miR-520/372/373 family expression RCC Development may provide an effective therapy in RCC. To our knowledge, the role and therapeutic potential of the miR-520/372/373 family in RCC, and its regulating mechanism on SPOP, have not been Figure 7. previously documented. Schematic for the E2F1/miR-520/372/373 family-SPOP axis in RCC.

OF12 Cancer Res; 78(24) December 15, 2018 Cancer Research

E2F1-miR-520/372/373-SPOP Axis in Renal Cancer

over, we illuminated that the miR-520/372/373–SPOP axis was Administrative, technical, or material support (i.e., reporting or organizing modulated by the transcription factor E2F1. In conclusion, our data, constructing databases): M. Ding, C. Wang, J. Ge results suggest that the E2F1–miR-520/372/373 family–SPOP Study supervision: J. Wang, C.-Y. Zhang axis may represent a potential therapeutic avenue to control Acknowledgments human renal cancer and impede metastasis. This work was supported by grants from National Natural Science Foundation of China (no. 81472021 and 81672102) and Fund of fl State Key Laboratory of Analytical Chemistry for Life Science (no. Disclosure of Potential Con icts of Interest 5431ZZXM1601 to C. Zhang); National Basic Research Program of China fl No potential con icts of interest were disclosed. (no. 2014CB542300 to C.-Y. Zhang); National Natural Science Foundation of China (81772282 and 81401257) and Foundation of Jiangsu Provincial Authors' Contributions Medical Youth Talent (QNRC2016893 to C. Wang). Conception and design: M. Ding, J. Wang, K. Zen, C.-Y. Zhang, C. Zhang Development of methodology: M. Ding, X. Lu The costs of publication of this article were defrayed in part by the Acquisition of data (provided animals, acquired and managed patients, payment of page charges. This article must therefore be hereby marked advertisement provided facilities, etc.): M. Ding, X. Lu, C. Wang, Q. Zhao, J. Ge, C. Zhang in accordance with 18 U.S.C. Section 1734 solely to indicate Analysis and interpretation of data (e.g., statistical analysis, biostatistics, this fact. computational analysis): M. Ding, C. Wang, Q. Zhao, Q. Xia, C. Zhang Writing, review, and/or revision of the manuscript: M. Ding, C. Wang, K. Zen, Received May 31, 2018; revised September 11, 2018; accepted October 15, C.-Y. Zhang, C. Zhang 2018; published first October 22, 2018.

References 1. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2017. CA Cancer J Clin for ubiquitination by Cul3-based ubiquitin ligase. J Biol Chem 2006; 2017;67:7–30. 281:12664–72. 2. Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer 17. Hernandez-Munoz I, Lund AH, van der Stoop P, Boutsma E, Muijrers I, statistics, 2012. CA Cancer J Clin 2015;65:87–108. Verhoeven E, et al. Stable X chromosome inactivation involves the PRC1 3. Hollingsworth JM, Miller DC, Daignault S, Hollenbeck BK. Five-year Polycomb complex and requires histone MACROH2A1 and the CUL- survival after surgical treatment for kidney cancer: a population-based LIN3/SPOP ubiquitin E3 ligase. Proc Natl Acad Sci U S A 2005; competing risk analysis. Cancer 2007;109:1763–8. 102:7635–40. 4. Lopez-Beltran A, Scarpelli M, Montironi R, Kirkali Z. 2004 WHO 18. Huang CJ, Chen HY, Lin WY, Choo KB. Differential expression of classification of the renal tumors of the adults. Eur Urol 2006;49: speckled POZ protein, SPOP: putative regulation by miR-145. J Biosci 798–805. 2014;39:401–13. 5. Chen X, Wang X, Ruan A, Han W, Zhao Y, Lu X, et al. miR-141 is a key 19. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. regulator of renal cell carcinoma proliferation and metastasis by control- Cell 2004;116:281–97. ling EphA2 expression. Clin Cancer Res 2014;20:2617–30. 20. Wu L, Belasco JG. Let me count the ways: mechanisms of gene regulation by 6. Zhao W, Zhou J, Deng Z, Gao Y, Cheng Y. SPOP promotes tumor miRNAs and siRNAs. Mol Cell 2008;29:1–7. progression via activation of beta-catenin/TCF4 complex in clear cell renal 21. Zhang Y, Huang B, Wang HY, Chang A, Zheng XFS. Emerging cell carcinoma. Int J Oncol 2016;49:1001–8. role of MicroRNAs in mTOR signaling. Cell Mol Life Sci 2017;74: 7. Huang CJ, Chen CY, Chen HH, Tsai SF, Choo KB. TDPOZ, a family of 2613–25. bipartite animal and plant proteins that contain the TRAF (TD) and POZ/ 22. Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell BTB domains. Gene 2004;324:117–27. 2009;136:215–33. 8. Liu J, Ghanim M, Xue L, Brown CD, Iossifov I, Angeletti C, et al. Analysis of 23. Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, Peck D, et al. Drosophila segmentation network identifies a JNK pathway factor over- MicroRNA expression profiles classify human cancers. Nature 2005; expressed in kidney cancer. Science 2009;323:1218–22. 435:834–8. 9. Kan Z, Jaiswal BS, Stinson J, Janakiraman V, Bhatt D, Stern HM, et al. 24. Esquela-Kerscher A, Slack FJ. Oncomirs - microRNAs with a role in cancer. Diverse somatic mutation patterns and pathway alterations in human Nat Rev Cancer 2006;6:259–69. cancers. Nature 2010;466:869–73. 25. Wang C, Hu J, Lu M, Gu H, Zhou X, Chen X, et al. A panel of five serum 10. Berger MF, Lawrence MS, Demichelis F, Drier Y, Cibulskis K, Sivachenko miRNAs as a potential diagnostic tool for early-stage renal cell carcinoma. AY, et al. The genomic complexity of primary human prostate cancer. Sci Rep 2015;5:7610. Nature 2011;470:214–20. 26. Ding M, Wang X, Wang C, Liu X, Zen K, Wang W, et al. Distinct expression 11. Li G, Ci W, Karmakar S, Chen K, Dhar R, Fan Z, et al. SPOP promotes profile of HCMV encoded miRNAs in plasma from oral lichen planus tumorigenesis by acting as a key regulatory hub in kidney cancer. Cancer patients. J Transl Med 2017;15:133. Cell 2014;25:455–68. 27. Lewis BP, Shih IH, Jones-Rhoades MW, Bartel DP, Burge CB. Prediction of 12. Guo ZQ, Zheng T, Chen B, Luo C, Ouyang S, Gong S, et al. Small-molecule mammalian microRNA targets. Cell 2003;115:787–98. targeting of E3 ligase adaptor SPOP in kidney cancer. Cancer Cell 28. John B, Enright AJ, Aravin A, Tuschl T, Sander C, Marks DS. Human 2016;30:474–84. MicroRNA targets. PLoS Biol 2004;2:e363. 13. Liu X, Sun G, Sun X. RNA interference-mediated silencing of speckle-type 29. Krek A, Grun D, Poy MN, Wolf R, Rosenberg L, Epstein EJ, et al. POZ protein promotes apoptosis of renal cell cancer cells. Onco Targets Combinatorial microRNA target predictions. Nat Genet 2005;37: Ther 2016;9:2393–402. 495–500. 14. Kim WY, Kaelin WG. Role of VHL gene mutation in human cancer. J Clin 30. Keklikoglou I, Koerner C, Schmidt C, Zhang JD, Heckmann D, Oncol 2004;22:4991–5004. Shavinskaya A, et al. MicroRNA-520/373 family functions as a tumor 15. Coppin C, Kollmannsberger C, Le L, Porzsolt F, Wilt TJ. Targeted suppressor in estrogen receptor negative breast cancer by targeting therapy for advanced renal cell cancer (RCC): a Cochrane sys- NF-kappaB and TGF-beta signaling pathways. Oncogene 2012;31: tematic review of published randomised trials. BJU Int 2011; 4150–63. 108:1556–63. 31. Guo H, Liu H, Mitchelson K, Rao H, Luo M, Xie L, et al. MicroRNAs-372/ 16. Kwon JE, La M, Oh KH, Oh YM, Kim GR, Seol JH, et al. BTB domain- 373 promote the expression of hepatitis B virus through the targeting of containing speckle-type POZ protein (SPOP) serves as an adaptor of Daxx nuclear factor I/B. Hepatology 2011;54:808–819.

www.aacrjournals.org Cancer Res; 78(24) December 15, 2018 OF13

Ding et al.

32. Marinescu VD, Kohane IS, Riva A. MAPPER: a search engine for 42. Wu G, Wang Y, Lu X, He H, Liu H, Meng X, et al. Low mir-372 expression the computational identiﬁcation of putative transcription factor correlates with poor prognosis and tumor metastasis in hepatocellular binding sites in multiple genomes. BMC Bioinformatics 2005; carcinoma. BMC Cancer 2015;15:182. 6:79. 43. Tian RQ, Wang XH, Hou LJ, Jia WH, Yang Q, Li YX, et al. MicroRNA-372 is 33. Bueno MJ, Gomez de Cedron M, Laresgoiti U, Fernandez-Piqueras J, down-regulated and targets cyclin-dependent kinase 2 (CDK2) and cyclin Zubiaga AM, Malumbres M. Multiple E2F-induced microRNAs prevent A1 in human cervical cancer, which may contribute to tumorigenesis. J Biol replicative stress in response to mitogenic signaling. Mol Cell Biol Chem 2011;286:25556–63. 2010;30:2983–2995. 44. Liu BL, Sun KX, Zong ZH, Chen S, Zhao Y. MicroRNA-372 inhibits endo- 34. Aguda BD, Kim Y, Piper-Hunter MG, Friedman A, Marsh CB. MicroRNA metrial carcinoma development by targeting the expression of the Ras regulation of a cancer network: consequences of the feedback loops homolog gene family member C (RhoC). Oncotarget 2016;7:6649–64. involving miR-17-92, E2F, and Myc. Proc Natl Acad Sci U S A 2008; 45. Guan X, Zong ZH, Chen S, Sang XB, Wu DD, Wang LL, et al. The role 105:19678–83. of miR-372 in ovarian carcinoma cell proliferation. Gene 2017;624: 35. Ren J, Jin P, Wang E, Marincola FM, Stroncek DF. MicroRNA and gene 14–20. expression patterns in the differentiation of human embryonic stem cells. 46. Chen HZ, Tsai SY, Leone G. Emerging roles of E2Fs in cancer: an exit from J Transl Med 2009;7:20. cell cycle control. Nat Rev Cancer 2009;9:785–97. 36. Laurent LC. MicroRNAs in embryonic stem cells and early embryonic 47. Attwooll C, Lazzerini Denchi E, Helin K. The E2F family: speciﬁc functions development. J Cell Mol Med 2008;12:2181–8. and overlapping interests. EMBO J 2004;23:4709–16. 37. Voorhoeve PM, le Sage C, Schrier M, Gillis AJ, Stoop H, Nagel R, et al. A 48. Mans DA, Vermaat JS, Weijts BG, van Rooijen E, van Reeuwijk J, Boldt K, genetic screen implicates miRNA-372 and miRNA-373 as oncogenes in et al. Regulation of E2F1 by the von Hippel-Lindau tumour suppressor testicular germ cell tumors. Cell 2006;124:1169–81. protein predicts survival in renal cell cancer patients. J Pathol 2013; 38. Huang Q, Gumireddy K, Schrier M, le Sage C, Nagel R, Nair S, et al. The 231:117–29. microRNAs miR-373 and miR-520c promote tumour invasion and metas- 49. Laine A, Westermarck J. Molecular pathways: harnessing E2F1 regulation tasis. Nat Cell Biol 2008;10:202–10. for prosenescence therapy in p53-defective cancer cells. Clin Cancer Res 39. Cho WJ, Shin JM, Kim JS, Lee MR, Hong KS, Lee JH, et al. miR-372 regulates 2014;20:3644–50. cell cycle and apoptosis of ags human gastric cancer cell line through direct 50. Tian W, Cui F, Esteban MA. E2F1 in renal cancer: Mr Hyde disguised as Dr regulation of LATS2. Mol Cells 2009;28:521–7. Jekyll? J Pathol 2013;231:143–6. 40. Gu H, Guo X, Zou L, Zhu H, Zhang J. Upregulation of microRNA-372 51. Ma X, Gao Y, Fan Y, Ni D, Zhang Y, Chen W, et al. Overexpression of associates with tumor progression and prognosis in hepatocellular carci- E2F1 promotes tumor malignancy and correlates with TNM stages in noma. Mol Cell Biochem 2013;375:23–30. clear cell renal cell carcinoma. PLoS One 2013;8:e73436. 41. Yamashita S, Yamamoto H, Mimori K, Nishida N, Takahashi H, Haraguchi 52. Gao Y, Li H, Ma X, Fan Y, Ni D, Zhang Y, et al. KLF6 suppresses metastasis of N, et al. MicroRNA-372 is associated with poor prognosis in colorectal clear cell renal cell carcinoma via transcriptional repression of E2F1. Cancer cancer. Oncology 2012;82:205–12. Res 2017;77:330–42.

OF14 Cancer Res; 78(24) December 15, 2018 Cancer Research

The E2F1−miR-520/372/373−SPOP Axis Modulates Progression of Renal Carcinoma

Meng Ding, Xiaolan Lu, Cheng Wang, et al.

Cancer Res Published OnlineFirst October 22, 2018.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-18-1662

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2018/10/20/0008-5472.CAN-18-1662.DC1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2018/11/28/0008-5472.CAN-18-1662. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.