Epigenetic silencing and CRISPR-mediated reactivation of tight junction protein claudin10b (CLDN10B) in renal cancer

Abstract

Background: The kidney's tubular system relies on cell polarity and tight junctions to maintain structure and function and disruptions contribute to diseases like cancer. Loss of tight junction proteins such as Claudins can actively contribute to tumorigenesis.

Results: We aimed to identify biomarkers for renal carcinoma, after kidney transplantation and conventional kidney tumors. We identified the epigenetic silencing of the Claudin 10 gene isoform B (CLDN10B) through DNA hypermethylation in renal cancers, including clear cell (ccRCC), papillary (pRCC) and post-transplantation renal carcinoma (PT-ccRCC). In contrast, CLDN10A was hypomethylated in ccRCC and pRCC. Differential methylation of the isoforms discriminates RCC from other malignancies. The epigenetic alteration of CLDN10B significantly correlated with reduced patient survival and advanced tumor staging. CLDN10B overexpression or induction significantly inhibited migration, cell cycle progression, and cellular growth. Using a CRISPR-based epigenetic editing tool reactivated CLDN10B to its endogenous level using VP160 and TET1 by promoter demethylation and significantly demonstrated its tumor-suppressive effects in 2D and 3D cell models.

Conclusion: Our findings suggest that CLDN10B acts as a tumor suppressor, and its epigenetic regulation may represent a therapeutic target for RCC. Ultimately, understanding CLDN10B's regulation and function could provide new insights into renal cancer treatment.

Keywords: CLDN10; CRISPR-Cas9; DNA (hyper)methylation; Epigenetic editing; Renal cell carcinoma (RCC); Tumor suppressor.

Fig. 1

Identification CLDN10 of differentially methylated candidates in PTM by methylation array EPIC1 of PT-ccRCC patient samples. a Differential methylation in PT-ccRCC samples (groups tumors and normal controls) depicted as heatmap (transform zscore, Anova 0.01 fdr). b Methylation changes as fold plot of n = 100 most significantly changed genes, CLDN10 marked with asterisk. Blanks are not gene associated reporters. c Hypermethylation of CpG island in tumor samples for CLDN10B is shown. The candidate is depicted across its genomic region with hypermethylated CpG island in tumor vs. matching control. d Methylation changes across CGIs of CLDN10B and RASSF1A in tumor vs. control. RASSF1A serves as a positive control. e CoBRA methylation analysis for CLDN10B CpG island reveals specific hypermethylation of tumor samples vs. control in five patient samples. Digest with Taq1 ( +) of PCR product shows specific tumor methylation vs. unmethylated normal samples. Additionally mock treatment (-) and 100 bp marker. CLDN10B specific PCR product is 172 bp

Fig. 2

CLDN10 and clinical parameters of ccRCC. a Reduced survival probability with high methylation shown for CLDN10B (cg16275739 representative CpG 200 bp upstream of isoform B TSS, positioned in CLDN10B isoform CGI in TCGA ccRCC KIRC dataset, analyzed by KM MethSurv). b Decreased expression correlates with tumor grade for CLDN10 across all stages (TCGA data ccRCC KIRC grouped by gene; separated by tumor grade; no distinction between isoforms possible, analyzed by R2) and correlation of clinical parameters c vital status relative to CLDN10 expression and d lymph nodes examined (no distinction between isoforms possible, analyzed by R2)

Fig. 3

Differential methylation of CLDN10 isoform A and isoform B in PTM ccRCC. a Methylome analysis (450 k array) of kidney cancer cell lines (Esteller dataset, gsm No, probes for CLDN10) depicts CGIs (green) for Isoform A with hypomethylation in yellow and isoform B with hypermethylation in blue (analysis R2). b Methylation profile across CLDN10 promoters for both isoforms (ß-value) for kidney cancer (data TCGA Renal Clear Cell Carcinoma and Renal Papillary Cell Carcinoma, analysis Wanderer). c Significant demethylation of CLDN10A and significant hypermethylation of CLDN10B in ccRCC and in pRCC in red vs control in gray (data TCGA by Smart as ß-value, aggregated by mean). d Distinct expression pattern of CLDN10 isoforms in normal tissues and loss of expression renal cell lines by qRT-PCR (normalized, normal tissues: heart, breast, liver, lung, kidney; cell lines: HEK293T transformed, cancer cell lines MZ1257/1973 are ccRCC). e Epigenetic reexpression of CLDN10 isoforms upon pharmacological demethylation treatment by Aza (conc. 0 µM to 20 µM, signif. p < 0.001, t-test) in kidney cancer cell lines (ccRCC). f CLDN10B expression via Western Blotting in Aza treated ccRCC cell line MZ1973 (4d and 7d, 7,5 µM Aza) confirms CLDN10 (affinity AF0133) induction with DNMT3A (64B814 Thermo) reduction (Vinculin V9131 Sigma, GAPDH 14C10 cell signaling). g Equivalent result for MZ1257 are shown

Fig. 4

CLDN10 reexpression hinders cellular fitness 2D and 3D culture. a CLDN10 overexpression as EGFP fusion in fluorescence microscopy in HEK cells (24 h) and CLDN10 induction in HEK Trex inducible cell line (24 h), together with CLDN10 counterstaining (red). b Induction of CLDN10 in 2D culture after reveals diminished wound healing capacity. c Cell cycle arrest upon CLDN10 induction after 24 h in HEK Trex measured by flow cytometry (propidium iodide stained) using BDCantoII. d-f 3D cell line growth was established for CLDN10 inducible cell line HEK Trex and control cell line under and dynamic cell culture conditions. Reduced spheroid size after 48 h upon CLDN10 induction (doxycycline) in CLDN10-GFP inducible cell line HEK Trex (vs. GFP control cells), f) microscopic growth reduction (48 h) and d according CLDN10 expression induction (RNA level, 72 h). g-h Transcriptome analysis of 3D cell model upon CLDN10 induction reveals an association with GO term cell death. Cells were seeded for 3D formation in dynamic cell culture and dox induced. Spheroids were harvested after 3 days for RNA isolation, reverse transcription and RNA-seq. g K-means clustering (3 groups, log2 zscore, passes/rounds set 10) shows separation of CLDN10 from control cell line with heatmap of top-deregulated genes. h) Upon CLDN10 induction co-deregulated genes is significantly correlated with several GO-terms

Fig. 5

CLDN10B signaling pathway induction. Whole proteome analysis upon CLDN10B induction reveals reexpression association with GO-term keratinization and intermediate filaments. HEK293T cells were transfected with Cas9 under guidance with CLDN10B guide #2, cultured for three days and harvested. Experiment was performed in quadruplicates. a Significant network appears upon CLDN10B induction with data being analyzed by string, filtering by effect > 0.5, with number of nodes n = 26 and protein–protein interaction enrichment p-value 4.57e-09 and b associated GO term of biological process, molecular function and cellular component. Likewise filtered proteome by effect < -0.5 gave no significant results (String)

Fig. 6

Epigenetic editing with gene reexpression, promoter demethylation and reactivation of tumor-suppressive function is achieved by VP160 and TET1 effectors on CLDN10B. For epigenetic editing of CLDN10B, HEK cells were transfected with according CRISPR-dCas vectors (either VP160 or TET1), RNA isolated, reversely transcribed and expression analyzed by RT-PCR. DNA was isolated, BS treated and CLDN10B CGI pyrosequenced. a dCas9-TET1 increases the endogenous background level (non-guided) of CLDN10B expression, compared to untreated HEK cells. b CLDN10B was demethylated by VP160 and TET1, that reactivated CLDN10B expression (c) and its tumor-suppressive function (d). d Wound healing assay in 2D cell culture reveals reduced cell migration after EpiEdit treatment. Scratch was placed 72 h post transfection, wound closure was examined by measuring the cell-free area at 0 h and 48 h (120 h post transfection) (Suppl. Figure 18 a)). e–g d Cas9-VP160 treatment of HEK cells in 2D and dynamic 3D cell culture. 3 Mio. cells were transferred 72 h post transfection (2D) into dynamic 3D cell culture (analyzed 120 h post transfection) (Suppl. Figure 18 b)). Demethylation of CLDN10B by dCas9-VP160 (e) leads to reexpression f and consequently to a reduced microscopic spheroid size (g + h)

Fig. 7

CLDN10B is silenced in kidney cancer, and its therapeutic reactivation is achievable. During carcinogenesis, hypermethylation of the CLDN10B promoter, particularly at its CpG island, occurs. This leads to the epigenetic inactivation and silencing of CLDN10B in kidney cancer and a loss of its tumor suppressor function. Using a CRISPR-dCas9 system, it is possible to reverse this epigenetic modification. By fusing dCas9 to effectors like TET1 (for demethylation) or VP160 (for both demethylation and transcriptional activation), targeted reactivation of CLDN10B can be achieved. Restoring CLDN10B expression suppresses cell growth