CDK13 upregulation-induced formation of the positive feedback loop among circCDK13, miR-212-5p/miR-449a and E2F5 contributes to prostate carcinogenesis

Abstract

Background: Both E2F transcription factor and cyclin-dependent kinases (CDKs), which increase or decrease E2F activity by phosphorylating E2F or its partner, are involved in the control of cell proliferation, and some circRNAs and miRNAs regulate the expression of E2F and CDKs. However, little is known about whether dysregulation among E2Fs, CDKs, circRNAs and miRNAs occurs in human PCa.

Methods: The expression levels of CDK13 in PCa tissues and different cell lines were determined by quantitative real-time PCR and Western blot analysis. In vitro and in vivo assays were preformed to explore the biological effects of CDK13 in PCa cells. Co-immunoprecipitation anlysis coupled with mass spectrometry was used to identify E2F5 interaction with CDK13. A CRISPR-Cas9 complex was used to activate endogenous CDK13 and circCDK13 expression. Furthermore, the mechanism of circCDK13 was investigated by using loss-of-function and gain-of-function assays in vitro and in vivo.

Results: Here we show that CDK13 is significantly upregulated in human PCa tissues. CDK13 depletion and overexpression in PCa cells decrease and increase, respectively, cell proliferation, and the pro-proliferation effect of CDK13 is strengthened by its interaction with E2F5. Mechanistically, transcriptional activation of endogenous CDK13, but not the forced expression of CDK13 by its expression vector, remarkably promotes E2F5 protein expression by facilitating circCDK13 formation. Further, the upregulation of E2F5 enhances CDK13 transcription and promotes circCDK13 biogenesis, which in turn sponges miR-212-5p/449a and thus relieves their repression of the E2F5 expression, subsequently leading to the upregulation of E2F5 expression and PCa cell proliferation.

Conclusions: These findings suggest that CDK13 upregulation-induced formation of the positive feedback loop among circCDK13, miR-212-5p/miR-449a and E2F5 is responsible for PCa development. Targeting this newly identified regulatory axis may provide therapeutic benefit against PCa progression and drug resistance.

Keywords: CDK13; Drug resistance; E2F5; Prostate cancer; circRNAs biogenesis.

Fig. 1

CDK13 is upregulated in PCa tissues and the upregulation of CDK13 promoted proliferation in PCa cells. a, RT-qPCR detected CDK13 mRNA expression in 30 pairs of PCa and benign prostatic hyperplasia (BPH). *P < 0.05 vs. BPH. b, CDK13 protein expression was measured by Western blotting in 3 pairs of randomly selected BPH (B) and PCa (P) tissues. c, Hematoxylin and eosin (HE) staining of BPH and PCa tissues. d, Immunochemistry staining of CDK13 in BPH and PCa tissues. Scale bar = 50 mm. e, The expression levels of CDK13 mRNA in BPH and PCa tissues from the GSE13507 database (P = 0.0007). f, The expression of CDK13 was examined by RT-qPCR (up panel) or Western blotting (down panel) in human normal prostate epithelial cells (RWPE-1) and PCa cell lines (LNCaP, PC3, 22RV1 and DU145). CDK13 expression was significantly increased in PC3 and 22RV1 cell lines. *P < 0.05 vs. RWPE-1. g and h, Cell viability was measured by MTS assay (G) and colony formation assays (H) in PC3 and 22RV1 cell lines after transfected with indicated vectors. *P < 0.05, and **P < 0.01 vs. their respective empty vector. i, PC3 and 22RV1 cells were transfected with shCDK13 or control vector, cell apoptosis was detected by AnnexinV/PI flow cytometry. Right panel shows the apoptosis rate of three independent experiments. **P < 0.001 vs. pLKO. pWPI is a control vector of oeCDK13 and pLKO is the control vector of shCDK13

Fig. 2

CDK13 interacts with E2F5 in PCa cells. a, Co-immunoprecipitation coupled with mass spectrometry (CoIP-MS) was performed with an anti-CDK13 in PC3 cells to detect the proteins interacting with CDK13, which are shown on the right panel. b, Heat map showing the differential expression (fold changes) of mRNAs between PCa and BPH tissues. Red color indicates several genes that are known to be transcriptionally upregulated in PCa tissues. c, CoIP analysis was used to detect the interaction between CDK13 and E2F5, and β-actin was used as a negative control. d, in situ proximity ligation (PLA) analysis detected the interaction between CDK13 and E2F5. Red color indicates PLA-positive cells. e, Immunofluorescence staining was performed to detect the expression and location of CDK13 and E2F5 in PCa and BPH tissues. Scale bars = 100 μm. f, E2F5 mRNA level was determined by RT-qPCR in 30 pairs of PCa and normal prostate tissues. **P < 0.01 vs. normal tissues. g, The expression levels of E2F5 mRNA in BPH and PCa tissues from the GSE13507 database (P < 0.0001). h, E2F5 protein expression was detected by Western blotting in 3 pairs of randomly selected BPH (B) and PCa (P) tissues. i, Immunohistochemistry staining detected the E2F5 protein expression in PCa and BPH tissues. j and k, the correlation between CDK13 and E2F5 mRNA expression in the PCa tissues was analyzed by Pearson correlation analysis in our clinical data (R = 0.4928, P = 0.0049) or several other data published in TCGA database (Yu: R = 0.3959, P = 0.0001; Wallace: R = 0.3752, P = 0.0015; Glinsky: R = 0.238, P = 0.0347)

Fig. 3

CDK13 interacts with E2F5 to cooperatively promote PCa cell proliferation. a and b, RT-qPCR and Western blot analysis detected the CDK13 expression in PC3 cells co-transfected with dCas9–VP64, MS2–p65–HSF1 and CDK13 promoter sgRNA (hereafter abbreviated as CDK13 sgRNA) or negative control (NC) sgRNA. *P < 0.05 vs NC sgRNA. c, Western blot analysis detected the E2F5, CDK13 and p21 expression in PC3 cells transfected with shE2F5, oeE2F5 or their respective control vector. d, Western blot analysis detected the proteins described as in (C) in PC3 and 22RV1 cells transfected with CDK13 sgRNA, shE2F5 or their respective control vector. e, Cells were prepared as in (D), cell viability was measured by MTS assay. *P < 0.05 vs. their respective controls. f, PC3 and 22RV1 cells were transfected with oeE2F5 and CDK13 sgRNA either alone or together, and colony formation assay was performed. Right panel shows the quantitative analysis of colony numbers from three independent experiments. *P < 0.05 vs. control constructs or CDK13 sgRNA and oeE2F5 alone. g, PC3 cells were transfected with CDK13 sgRNA or negative control (NC) sgRNA (left panel), as well as with siPol ll or siCtl (right panel). CoIP analysis detected the interaction between E2F5, CDK13 and Pol ll. h, Colony formation assay detected the proliferation of PC3 and 22RV1 cells co-transfected with the indicated constructs. Right panel shows the quantitative analysis of colony numbers from three independent experiments. *P < 0.05 vs. control constructs or siPol ll and CDK13 sgRNA alone

Fig. 4

Transcriptional activation of endogenous CDK13 upregulates E2F5 expression by promoting circCDK13 formation. a, PC3 cells were transfected with CDK13 sgRNA, oeCDK13 or their respective control vector, and then Western blot analysis detected the CDK13 and E2F5 expression. b, PC3 cells were treated as in (A), RT-qPCR detected E2F5 mRNA expression. c and d, PC3 cells were transfected with shCDK13 or pLKO or treated with CDK13 inhibitor THZ531, and then Western blot (C) and RT-qPCR (D) analysis detected the CDK13 and E2F5 expression. e and f, PC3 cells were transfected with CDK13 sgRNA or NC sgRNA and treated with actinomycin D (Act D). Western blotting (E) detected E2F5 protein level, and RT-qPCR (F) detected the expression of circRNAs (hsa_circ_0001699, hsa_circ_0079929, hsa_circ_0079933 and hsa_circ_0079939). *P < 0.05 vs. NC sgRNA+DMSO; #p<0.05 vs. CDK13 sgRNA+DMSO. g, PCR was used to detect hsa_circ_0079929 (termed circCDK13) in PCa tissues by using convergent or divergent primers. Divergent primers amplify circCDK13 in cDNA but not in genomic DNA (gDNA). GAPDH was used as linear control. h, RT-PCR amplified full-length circCDK13 in PC3 cells, and the amplified products were confirmed by agarose gel electrophoresis. i, Sanger sequencing confirmed head-to-tail splicing of circCDK13. j, RT-qPCR detected the expression of circCDK13 in PCa and BPH tissues. **P < 0.01 vs. BPH. k, The correlation between circCDK13 and E2F5 mRNA expression in our clinical data was analyzed by Pearson correlation analysis (R = 0.3976, P = 0.0296). l, PC3 and 22RV1 cells were transfected with si-circCDK13, si-linear CDK13 or si-Ctl. RT-qPCR detected circCDK13 and CDK13 mRNA expression. Bars are mean ± SEM of triplicate samples. *P < 0.05, **P < 0.01 vs. si-Ctl. m, RT-qPCR examined the expression of CDK13 mRNA and circCDK13 in PC3 and 22RV1 cells transfected with circCDK13 or empty vector. *P < 0.05, **P < 0.01 vs. empty vector. n, Western blotting detected the E2F5 expression in PC3 and 22RV1 cells transfected with si-circCDK13, circCDK13 or their respective control. o, PC3 and 22RV1 cells were transfected with circCDK13 and shE2F5 either alone or together, and then colony formation assay was performed to detect the cell proliferation. *P < 0.05 vs. pLKO+pWPI, #p<0.05 vs. circCDK13 + pLKO

Fig. 5

circCDK13 upregulates E2F5 protein level by sequestering miR-221-5p/449a a, RT-qPCR detected the expression of E2F5 mRNA in PC3 cells transfected with circCDK13, si-circCDK13 or their respective control. b, Venn diagram displaying potential microRNAs associated with circCDK13 sequence from three online target-prediction programs. c, circCDK13 or empty vector was transfected into PC3 cells, RT-qPCR detected the pulled down efficiency of circCDK13 by biotinylated probes against circCDK13. *P < 0.05, **P < 0.01 vs. con probe or empty vector. d, RT-qPCR detected the level of indicated miRNAs pulled down by biotinylated probes against circCDK13. *P < 0.05, **P < 0.01 vs. con probe. e, Biotin-labeled E2F5 3′-UTR RNA was transfected into PC3 cells followed by a biotin pull-down assay using Streptavidin-coupled Dynabeads. The miRNAs were extracted from the sedimented beads, and the relative levels of 11 candidate miRNAs were detected by RT-qPCR. *P < 0.05, **P < 0.01 vs. control E2F5 3′-UTR. f, PC3 cells were co-transfected with miR-449a mimic, miR-212-5p mimic or both and circCDK13-directed luciferase reporter. Luciferase activity was measured by dual-luciferase reporter assays. *P < 0.05, **P < 0.01 vs. miR-NC. g, RNA in situ hybridization detected the co-localization between miR-449a or miR-212-5p with circCDK13 (arrowheads) in PC3 cells co-transfected with circCDK13 expression vector and miR-449a or miR-212-5p mimic. Nuclei were counterstained with DAPI. Scale bars = 25 μm. h, PC3 cells were co-transfected with miR-449a, miR-212-5p or control mimic (miR-NC) and wild type or mutated (mut) E2F5 3’UTR-directed luciferase reporter. Luciferase activity was measured by dual-luciferase reporter assays. Graph bars are mean ± SEM of 3 independent experiments. *P < 0.05, **P < 0.01 vs. miR-NC or E2F5 3’UTR mut. i, PC3 and 22RV1 cells were transfected with miR-212-5p, miR-449a mimic or both, and then Western blotting detected the CDK13 and E2F5 expression.

Fig. 6

E2F5 activates the transcription of CDK13 gene and upregulates circCDK13 expression. a and b, PC3 cells were transfected with oeE2F5, shE2F5 or their corresponding control vectors, and then RT-qPCR detected CDK13 mRNA (A) or circCDK13 (B). *P < 0.05, **P < 0.01 vs. control vector. c, The schematic diagram shows 3 putative E2F binding sites within the 1-kb promoter region of CDK13. d, ChIP-qPCR detected the binding of E2F5 to the CDK13 promoter region in PC3 cells. Proximal: − 263 ~ − 459; Distal: − 444 ~ − 704. Arrowheads indicate the position of primers used for qPCR. *P < 0.05 vs. IgG. e, PC3 cells were transfected with the indicated constructs. Luciferase activity was measured by dual-luciferase reporter assays. *P < 0.05 vs. CDK13-Luc or CDK13-Luc + oeE2F5, #p<0.05 vs. oeE2F5 + circCDK13. f and g, Western blot analysis examined the E2F5, CDK13 and p21 expression in PC3 and 22RV1 cells transfected with circCDK13 and oeE2F5 either alone or together (F), or in PC3 and 22RV1 cells transfected with shcircCDK13 and miR-449a/212-5p mimic either alone or together (G). h, PC3 and 22RV1 cells were transfected as in (G), and cell proliferation was measured by colony formation assay. Upper panel shows the quantitative analysis of colony numbers from three independent experiments. *P < 0.05 vs. control constructs or shcircCDK13 and miR-449a/212-5p alone

Fig. 7

CDK13-circCDK13-miR-212-5p/miR-449a-E2F5 regulatory axis participates in prostate tumorigenesis. a and b, PC3 cells engineered to stably overexpress GFP-shcircCDK13 (LV-shcircCDK13), GFP-shE2F5 (LV-shE2F5) or negative control (LV-GFP) were injected subcutaneously in 200 μl PBS/Matrigel (50:50) into the mouse forelimb (left: LV-GFP; right: LV-shcircCDK13) (left panel), or GFP-shE2F5 (left: LV-shE2F5) or both (right: LV-shcircCDK13 + LV-shE2F5) (right panel). At the final time point (21 days after injection), the tumor volumes in each group were measured both in situ by fluorescence imaging (A) and after resection of tumors (B) (n = 12 in each group). c, Tumor volume was determined by direct measurement with calipers and calculated by the formula: volume = [(length × width2) / 2]. *P < 0.05, **P < 0.01 vs. LV-GFP (n = 12 in each group), #p<0.05, ##p<0.01 vs. LV-shcircCDK13 or LV-shE2F5 (n = 10 in each group). d, Western blotting detected the expression of E2F5, CDK13 and p21 in xenograft tumors prepared as in (A). e, The xenograft models of nude mice were prepared as in (A), the TUNEL staining detected cell apoptosis in xenograft tumors. Blue staining represents the nucleus, and red staining indicates TUNEL-positive cells. Bar = 25 μm. f, PC3 and 22RV1 cells were treated with indicated pharmacological inhibitors for different signaling pathways, and Western blot analysis detected the expression of CDK13 and E2F5 protein. g-i, RT-qPCR detected the E2F5 mRNA (G), CDK13 mRNA (H), and circCDK13 (I) expression in PC3 and 22RV1 cells treated with 1-Azak or DMSO. *P < 0.05 vs. the vehicle. j, PC3 and 22RV1 cells were transfected with shE2F5 or empty vector pLKO and treated with 1-Azak for 24 h, and then cell viability was measured by MTS assay. *P < 0.05 vs. the vehicle, #p<0.05 vs. shE2F5 or 1-Azak. k, PC3 and 22RV1 cells were transfected with shcircCDK13 alone or combined with 1-Azak treatment. Cell proliferation was measured by colony formation assay. Right panel shows the quantitative analysis of colony numbers from three independent experiments. *P < 0.05 vs. the vehicle, #p<0.05 vs. shcircCDK13 or 1-Azak. l, PC3 and 22RV1 cells were treated as in (K), cell apoptosis was detected by AnnexinV/PI flow cytometry. Right panel shows the apoptosis rate of three independent experiments. *P < 0.05 vs. the vehicle, #p<0.05 vs. shcircCDK13 or 1-Azak