Epigenetic repression of ROR2 has a Wnt-mediated, pro-tumourigenic role in colon cancer

Analysis of the region 1.0 kb upstream and 0.5 kb downstream of the ROR2 transcriptional start site identified a CpG island, suggesting a potential role for CpG methylation in the regulation of ROR2 expression. To study the possible aberrant epigenetic regulation of ROR2 in colon cancer, we used bisulphite sequencing of multiple clones to determine the methylation status of a ROR2 promoter DNA region of 315 bp that spans the ROR2 transcriptional start point in healthy colon tissue, in vitro-growing colonocytes and eight colon cancer cell lines (HCT116, SW480, LOVO, HT29, HCT15, DLD1, COLO205 and RKO) (Figure 1A). This showed that the ROR2 promoter was completely unmethylated in non-tumourigenic colon primary tissues and in vitro-growing colonocytes, whilst it was densely hypermethylated in most cancer cell lines analysed (HT29, HCT15, DLD1, COLO205 and RKO). These results were confirmed using the 27K Illumina Infinium methylation platform to study the DNA methylation status of two CpG positions (located 465 bp upstream and 322 bp downstream of the ROR2 transcription start site, respectively) within the ROR2 CpG island (Figure 1B) and using methylation-specific PCR (Figure 1C).

To determine whether ROR2 promoter hypermethylation is also a frequent in vivo event, we used methylation-specific PCR to analyse 36 primary colon adenocarcinomas (Figure 1D). We detected methylation at the ROR2 promoter in 14 of the 36 tumours (38.8%), which confirmed that this is a common event in vivo.

To determine the role of the ROR2 promoter hypermethylation in gene expression, we used qRT-PCR and WB to compare ROR2 mRNA and protein levels in healthy colon epithelium and the HCT116 and SW480 cell lines, which do not show ROR2 promoter hypermethylation, with the HT29, HCT15, DLD1 and RKO lines, which present dense DNA methylation at the ROR2 promoter (Figure 2A). These experiments showed that ROR2 mRNA and protein were only detected in samples lacking ROR2 promoter hypermethylation. Although they showed dense ROR2 promoter hypermethylation, HT29 cells expressed residual levels of ROR2 mRNA, insufficient to detect ROR2 protein. To study the relationship between ROR2 promoter hypermethylation and ROR2 repression in more detail, we analysed ROR2 mRNA levels in colon cancer cell lines incubated with the demethylating drug 5-aza-2-deoxycytidine (Figure 2B). Treatment with this drug resulted in marked reactivation of ROR2 in cell lines displaying ROR2 promoter hypermethylation (HCT15, DLD1 and RKO), but not in those lacking hypermethylation (HCT116 and SW480) (Figure 2B). This implies that in vitro ROR2 promoter hypermethylation is directly associated with ROR2 repression.

To study the relationship between ROR2 promoter hypermethylation and gene expression in vivo, we analysed ROR2 mRNA levels and used tissue microarrays to determine the levels of ROR2 protein in 20 pairs of normal and tumour colon tissues obtained from the same patients (Figures 2C, D). At the mRNA level, in 18 of the 20 paired samples analysed, ROR2 was much less strongly expressed in tumour tissues than in their normal counterparts; mean ROR2 expression was seven times higher in colon epithelium than in colon tumours (Figure 2C). At the protein level, we also observed lower ROR2 levels in tumour samples in most cases; we only detected positive ROR2 staining in 35% of the colorectal tumours analysed (Figure 2D). To determine the relationship between ROR2 promoter hypermethylation with mRNA and protein expression, we selected three tumours that expressed and three that did not express ROR2 mRNA, and analysed the methylation status of the ROR2 promoter by methylation-specific PCR (Figure 2E) and ROR2 protein levels by WB. The results confirmed that tumours lacking ROR2 expression presented promoter hypermethylation and lack ROR2 protein. This corroborated a relationship between ROR2 promoter hypermethylation and gene repression in vivo. The data indicate that ROR2 promoter hypermethylation is directly associated with ROR2 repression in colon cancer, both in vitro and in vivo.

To evaluate the functional role of ROR2 repression by promoter hypermethylation in colon cancer, we transfected ROR2 in the colon cancer cell line DLD1, which shows DNA methylation-dependent ROR2 inactivation (Figure 3A). MTT analysis of the effects of ROR2 restoration on cell proliferation showed that ROR2 restoration induced a notable reduction in DLD1 cell proliferation (Figure 3B). We used colony-formation assays to further characterise the role of ROR2 in the growth of colon cancer cells, and found that restoration of ROR2 activity resulted in a 30% reduction in colony formation in DLD1 cells (Figure 3C). We also knocked down ROR2 expression by RNA interference in a ROR2-expressing colon cancer cell line unmethylated at the ROR2 CpG island, SW480. Reduced ROR2 expression (Fig. 3D) was associated with increased cell growth (Fig. 3E).

To study the in vivo role of ROR2 in colon carcinogenesis, we subcutaneously injected DLD1 cells with restored ROR2 expression and isogenic controls transfected with empty vectors into a cohort of immunodeficient nude mice (Figure 3F). In both cases, macroscopic tumours were observed 13 days after injection (Figure 3F). Restoration of ROR2 expression in DLD1 cells resulted in a marked reduction of tumour formation in vivo; at the time the mice were sacrificed, DLD1 tumours expressing ROR2 had on average half the volume of those lacking ROR2 activity (Figure 3F). This in vivo evidence links ROR2 aberrant promoter hypermethylation with colon cancer development.

Epigenetic repression of other Wnt pathway components promotes tumourigenesis by the constitutive activation of canonical Wnt [7]. In addition, ROR2 regulates TCF-mediated transcription [8]. We thus hypothesised that the role of ROR2 in colon cancer is mediated by canonical Wnt signalling. To test this possibility, we tested whether its restoration interfered with β-catenin/TCF-dependent transcription, using luciferase reporter assays to analyse the effect of exogenous ROR2 expression in ROR2-silenced DLD-1 cells (Figure 4A). Restoration of ROR2 expression reduced basal β-catenin/TCF-dependent transcription in DLD1 cells by 50%, suggesting that the role of aberrant epigenetic repression of ROR2 in colon cancer is mediated, at least in part, by canonical Wnt signalling.

ROR2 is reported to mediate the WNT5A-dependent inhibition of β-catenin-TCF-regulated genes in human embryonic cells [8, 15]. To determine whether the canonical Wnt-dependent role of aberrant epigenetic repression of ROR2 in colon cancer is WNT5A-associated, we used we used a dual approach: i) we increased extracellular levels using recombinant WNT5A protein and ii) we used anti-WNT5A antibody to greatly reduce WNT5A activity in DLD1 cells expressing ROR2 and control cells. We studied the effect of WNT5A modulation on cell growth and the Wnt signalling pathway. MTT assays showed that, independently of ROR2 expression, WNT5A removal reduced cell growth (Figure 4B, top). The decrease in cell growth was greater in DLD1 cells that did not express ROR2, although the difference was not significant (p = 0.12). Addition of recombinant WNT5A to culture medium had no relevant effect on cell growth, independently of ROR2 expression (Figure 4B, bottom).

To study the effect of WNT5A modulation on canonical Wnt signalling, we treated ROR2-expressing DLD1 cells or control DLD1 cells with anti-WNT5A antibody or with recombinant WNT5A. We determined nuclear β-catenin levels in immunofluorescence experiments (Figure 4C) and analysed β-catenin/TCF-dependent transcription in luciferase reporter assays (Figure 4D). Increasing or decreasing extracellular WNT5A levels had little impact on β-catenin nuclear location and β-catenin/TCF-dependent transcription (Figure 4C and 4D), suggesting that the canonical Wnt-dependent role of ROR2 epigenetic deregulation in DLD1 colon cancer cells does not necessarily require WNT5A.

The eight human colon cancer cell lines (HCT116, SW480, LOVO, HT29, HCT15, DLD1, COLO205 and RKO) and healthy colonocytes were obtained from the American Type Culture Collection. Cell lines were maintained in appropriate medium and treated with 2 μmol/L 5-aza-2'-deoxycytidine (Sigma) for 3 days for DNA demethylation, anti-WNT5A antibody (2 μg/ml; R&D) or recombinant WNT5A (200 ng/ml; Millipore). We obtained 36 primary colon tumours from the Spanish National Cancer Research Centre Tumour Bank and 20 pairs of same-patient healthy and colon tumour tissue from the Institute of Oncology of Asturias Tumour Bank. The study was approved by the appropriate institutional review boards.

We established ROR2 CpG island methylation status by PCR analysis of bisulphite-modified genomic DNA. Methylation status was first analysed by bisulphite genomic sequencing of the CpG island using primers: 5'-GGG GTT TTA GTT GTA GTT TTA GT-3' (sense) and 5'-CTC CTC CTT CTC CCT AAC-3' (anti-sense). Twelve independent clones were sequenced for each sample. The second analysis used methylation-specific PCR with primers specific to the methylated or modified unmethylated DNA. Primer sequences for the methylated reaction were 5'- GTT TCG TTT TGT TTA TCG GGG C - 3' (sense) and 5'-ACT AAA AAA ATT CCT TAA CGC GAA -3' (anti-sense), and for the unmethylated reaction 5'-GTT TTG TTT TGT TTA TTG GGG TGG-3' (sense) and 5'-TAA CTA AAA AAA TTC CTT AAC ACA AA-3' (anti-sense). In vitro-methylated DNA (IVD) was used as a positive control for the methylated reaction. PCR products were identified in ethidium bromide-stained 2% agarose electrophoresis gels and viewed under UV light. We designed the bisulphite genomic sequencing and methylation-specific PCR primers using Methyl Primers Express software and confirmed bisulphite sequencing and methylation-specific PCR results using Illumina Infinium Methylation Arrays (Illumina). The two CpG sites analysed were located 465 bp upstream and 322 bp downstream of the ROR2 transcriptional start point. Relative DNA methylation for each CpG site is represented by a scale from 0.0 to 1.0 (corresponding to 0% and 100% likelihood of gene promoter hypermethylation, respectively).

Total RNA (1 μg) extracted with Trizol reagent (Invitrogen) was converted to cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). The PCR reaction was performed by mixing the converted cDNA with the TaqMan Gene Expression Assay for ROR2 (Hs00171695_m1 ROR2) and GAPDH (Hs99999905_m1 GAPDH) and TaqMan Universal PCR Master Mix (Applied Biosystems). qRT-PCR was performed in the Applied Biosystems 7900 HT Fast Real-Time PCR System. Data were normalised using GAPDH as the endogenous control gene. The ROR2 protein was analysed with an anti-ROR2 antibody (from Abcam for immunohistochemistry; from Sigma for WB), following standard procedures.

The pcDNA3.1-ROR2-Flag construct was created using a pcDNA3.1-ROR2 plasmid (kindly provided by Dr J. Billiard) as a template. The construct was confirmed by DNA sequencing. Control cells were transfected using the empty plasmid (mock). DLD1 cells were transfected with the pcDNA3.1-ROR2-Flag plasmid or the empty plasmid using Lipofectamine 2000 (Invitrogen). Stable transfectants were obtained after 2 weeks selection with 1 mg/ml G418 (Calbiochem). ROR2-Flag expression was confirmed by WB using anti-Flag antibody (1:2000, Sigma).

Cell viability was determined as described [24]. Aliquots (10⁴ cells) were plated onto 96-well plates. After cell attachment, MTT was added to medium (50 μg; 100 μl/well) and incubated (3 h, 37°C, 5% CO₂). MTT was removed and MTT-formazan crystals were dissolved in DMSO (100 μl/well). Absorbance at 595 nm was determined with an automated microtitre plate reader. Optical density was directly proportional to cell number up to the maximum density measured. Results are expressed as the mean ± SD of at least five replicates and significance was assessed by standard t-test.

Colony formation was assayed on ROR2-Flag-and mock-transfected DLD1 cells. After 14 days selection, stable G418-resistant colonies were fixed, MTT-stained, and the mean number of colonies in each well was determined.

Five-week-old athymic Nude-Foxn1_nu/nu mice (Charles River) were used for tumour xenograft experiments with ROR2-Flag-and mock-transfected DLD1 cells. Twelve mice were used. Both flanks of each animal were injected subcutaneously with 2 × 10⁶ cells in 200 μl PBS; ROR2-Flag-transfected cells were injected into the right flank and empty vector cells into the left flank. Mice were weighed, and tumour width (W) and length (L) were measured every 5 days. Tumour volume was estimated according to the formula V = 0.4. L.W² (L = maximum length; W = maximum width). Mice were killed 30 days post-injection and tumours from both cell types were excised and weighed. Mean volume and tumour mass ± SEM were calculated for each group, and significant differences were assessed by a two-tailed independent-samples t-test.

Subconfluent cultures were transfected in triplicate using JetPEI (PolyPlus Transfection, Illkirch, France) following manufacturer's protocols. pTOPFLASH, and pFOPFLASH reporter plasmids have been described [25]. A Renilla luciferase plasmid (pRL-TK) was used in all experiments as an internal control. Firefly and Renilla luciferase activities were measured separately using the Dual Luciferase reagent kit (Promega, Madison, WI) and a GloMax 96 microplate luminometer (Promega).

ROR2-Flag-and mock-transfected DLD1 cells were methanol-fixed and stained using mouse anti-β-catenin (Becton Dickinson) and goat anti-α-lamin B (Santa Cruz), followed by Alexa448 donkey anti-goat IgG or Alexa594 donkey anti-mouse IgG (Molecular Probes).