Background
Colorectal cancer (CRC) is the third most common cancer worldwide with an estimated 1 million cases each year contributing to over 608,000 deaths [
1‐
3]. CRCs develop from benign intraepithelial neoplasms known as adenomas, which progress to cancer after an accumulation of mutations [
4,
5]. The Wnt signalling pathway is frequently altered in CRC with ~94 % of cases possessing a mutation in a Wnt pathway gene [
6]. One of the early precipitating events for colorectal adenoma development is mutation of the
APC gene, an important tumour suppressor and regulator of β-catenin dependent Wnt signals [
5,
7,
8]. APC along with AXIN and GSK3β are responsible for degradation of cytosolic β-catenin and loss of APC leads to β-catenin accumulation, Wnt pathway hyperactivation and increased cellular proliferation and migration [
8‐
15].
In contrast, the β-catenin independent Wnt pathway affects planar cell polarity (PCP), cell adhesion and motility and is not reliant on β-catenin levels [
16‐
20]. The receptor tyrosine kinase-like orphan receptor 2 (ROR2) is a receptor tyrosine kinase which binds with WNT5A to activate the β-catenin independent Wnt pathway [
21‐
23]. In addition to activating β-catenin independent Wnt/JNK signalling, ROR2 and WNT5A interaction has been shown to antagonise downstream targets of β-catenin dependent Wnt; specifically inhibition of
AXIN2 expression and the TCF/LEF transcription factors [
16,
20,
23‐
26]. Consistent with its reported antagonism of β-catenin dependent Wnt signals, a 2010 study found
ROR2 to be silenced in colorectal cancer, resulting in increased cellular proliferation [
27]. However, other reports in colorectal cancer, melanoma and osteosarcoma have found elevated
ROR2 expression in tumours compared to normal tissue [
28‐
32]. These conflicting reports have raised questions as to the role ROR2 plays in cancer and presents the possibility that the downstream effects of ROR2 are dependent on other Wnt genes and the cellular context of the cancer itself [
33‐
35].
In this study, we investigated whether ROR2 expression is altered in colorectal cancers and adenomas. We also assessed the effects of altered ROR2 expression on β-catenin dependent Wnt signalling, proliferation, migration and invasion properties in colorectal cancer cells.
Discussion
Although
ROR2 is not normally expressed in mature adult cells, evidence from prior studies indicate that it is present in the colon epithelium as well as in parathyroid, testicular and uterine tissue [
27,
36]. Previous publications examining
ROR2 in CRC found both upregulation and downregulation of the receptor in CRC [
27,
28]. Both publications used qRT-PCR to document
ROR2 expression in 20 matching tumour and normal samples yet report different findings. The reasons for the conflicting results in these 2 publications remain unclear although differences in study cohort and methodology may explain this discrepancy.
In our study, analysis using qRT-PCR found ROR2 expression loss in the majority of both CRC cell lines (n = 23) and colorectal adenoma (n = 6) samples. In addition, analysis of 258 patient samples from the publicly available TCGA database found a significant decrease of ROR2 expression in primary tumour samples compared to the normal mucosa, providing strong evidence that ROR2 is downregulated in CRC.
Our study also uses COBRA and bisulphite sequencing to show for the first time that not only is promoter hypermethylation present in the majority of CRC cell lines but it is also present in early colorectal adenomas. Along with the methylation, there was also a corresponding loss of
ROR2 expression in the adenoma samples, leading us to hypothesise that the observed downregulation was caused by epigenetic silencing through promoter hypermethylation. This was supported by our analysis of both CRC cell lines and primary tumours samples from the publicly available TCGA database as well as data from the previous publication from Lara et al. [
27]. Our cell line experimentation also supported this hypothesis as
ROR2 expression was restored in RKO and SW620 cells following demethylation using the DNA methyltransferase inhibitor 5-aza-2′deoxycytidine. These findings of
ROR2 expression loss and promoter hypermethylation are particularly important as they have been conducted on not only cell lines but also on clinical samples from both adenomas and primary tumours. Together, the clinical data along with the analysis and experimentation of cell lines provides strong evidence that epigenetic silencing of ROR2 through promoter hypermethylation occurs early in colorectal carcinogenesis.
Although we have shown
ROR2 to be epigenetically silenced in the majority of CRC cases, the exact molecular outcomes of this loss in the colon epithelium remains unclear. Knockdown experiments confirm that
ROR2 expression loss results in a subsequent decrease of the downstream β-catenin independent Wnt genes
JNK and
NFATC1. Although previous studies have shown
ROR2 loss resulted in increased expression of the β-catenin dependent Wnt target AXIN2 [
23], we did not observe this in our in vitro cell line model. Our examination of the β-catenin dependent Wnt target
AXIN2 and
CCND1 not only revealed no apparent increase but
CCND1 expression levels were instead found to be significantly decreased. A likely explanation for this difference in findings may be the differences in the biological models used, as the previous publication which reported increased
AXIN2 expression following
ROR2 silencing used in vivo mouse models incorporating the tumour microenvironment [
24]. It is possible that in our experiments on an immortal cancer cell line, the cellular and genetic context was significantly different and that the loss of
ROR2 resulted in the activation of different signalling pathways. This is supported by recent publications which show
ROR2 and other Wnt associated genes to be capable of activating both the β-catenin dependent and β-catenin independent halves of the Wnt signalling pathway [
20,
24,
33,
34,
37].
ROR2 has been shown to interact with different co-receptors [
38,
39] and ligands [
40] as well being the target of phosphorylation by different intracellular proteins [
25,
41]. As the exact signalling consequences of these
ROR2 interactions are as yet uncertain, it is possible that
ROR2 downregulation resulted in different signalling cascades in in vivo mice and in immortal cancer cells.
Although there was no observed upregulation in the β-catenin dependent Wnt target genes following
ROR2 knockdown as reported in the literature [
23,
24], our in vitro assays on HCT116 cells still revealed an increase in proliferation and migration. There was a significant increase to cellular proliferation following
ROR2 knockdown while the observed increase to migration was close to significance with a P value of only 0.056. The effect of
ROR2 knockdown on cellular invasion was also investigated in HCT116 cells, with the results revealing a decrease in cellular invasion. These results are consistent with our findings that
ROR2 was initially lost in precancerous adenomas which possess no invasive properties. Analysis of gene expression also found no changes to the key EMT-related genes
CDH1 and
VIM following
ROR2 knockdown, suggesting that invasion capacity in CRC only occurs later during disease progression. Our combined functional analysis indicates that
ROR2 downregulation may cause increased proliferation and migration in early non-invasive adenomas, resulting in a more metastatic phenotype. The lack of observed increase in β-catenin dependent Wnt target genes indicate that these changes were not influenced by the inhibition of β-catenin dependent Wnt signals. It is possible that
ROR2 loss affected both arms of the Wnt signalling pathways as had been previously reported in breast and renal cancer, resulting in the observed phenotypic changes [
20,
33‐
35]. Another possibility is that the interaction between Wnt signalling and another signalling pathway resulted in unexpected circumstances [
42‐
44]. It is evident that
ROR2 plays a much more complex role in CRC and the Wnt signalling pathway than previously thought. Further investigations examining the direct interactions
ROR2 has with Wnt and EMT associated genes through techniques such as DNA microarrays or RNA-seq would help reveal the exact mechanism in which
ROR2 affects cellular proliferation and migration in the context of CRC progression.
As we had found loss of
ROR2 function to increase cellular proliferation and migration, we hypothesised that re-expression of the receptor may have the opposite effect. However, when
ROR2 was ectopically expressed in RKO and SW620 cells, there was no significant change observed in cellular proliferation and migration. This may have been because the level of
ROR2 expression generated by plasmid transfection was significantly higher than that of normal
ROR2 expression levels. This could have adversely affected Wnt signalling as certain pathways are sensitive to the ratios of receptors and ligands [
45,
46].
It is also possible that the RKO cell line was not functionally affected by ectopic
ROR2 expression because it did not originate from a CRC caused by aberrant Wnt signalling. Although ~94 % of CRC cases possess a mutation in a Wnt pathway gene with
APC being the gene most predominantly mutated [
6], not all CRC cases arise from a dysfunctional Wnt signalling pathway. A significant proportion of CRC cases result from other causes such as mutations and methylation in mismatch repair (MMR) genes [
47,
48]. RKO cells do not have a mutant
APC gene but they do have methylated MMR genes as well as possessing the CpG island methylator phenotype [
49]. This suggests that RKO cells originally became carcinogenic through methylation and loss of function in MMR genes rather than through aberrant Wnt signalling.
In SW620 cells, the absence of any change in proliferation and migration following ectopic ROR2 expression may have been because the cell line originated from a secondary tumour site. SW620 and SW480 cells originated from the same patient with SW620 cells obtained from a lymph node metastasis while SW480 were from the primary tumour [
50]. Having already metastasised to a secondary site, SW620 cells would possess a markedly different genetic composition than that of a primary tumour and may be resistant to any functional effects resulting from restoration of expression in an early gene target such as
ROR2.
This is a potential issue for all cell line models as they are cancer cells that are different to the colorectal adenomas in which we believe
ROR2 methylation and expression loss first occurs. To truly determine if early
ROR2 loss is involved in CRC progression in adenomas, a biological model which more closely resembles colorectal adenomas would be needed. Future research could possibly investigate functional effects of
ROR2 loss in colorectal adenomas grown in in vitro organoids [
51]. Another possibility would be to use an inducible mouse knockout model that targeted
ROR2 in the colon. Using a mouse strain that had a high prevalence for adenomas such as the APC heterozygous 57BL/6 J-ApcMin/J mouse line, would allow for the determination of whether or not early ROR2 loss potentiates adenoma growth and development.
Methods
Cell lines
All colorectal cancer cells were obtained from ATCC (American Type Culture Collection, Manassas, VA, USA). HCT116 cells were cultured in McCoy’s media (Life Technologies, Rockville, MD) supplemented with 10 % foetal bovine serum, 1× glutamine (200 mM) and penicillin/streptomycin (10 units/ml). RKO cells were cultured in RPMI media (Life Technologies, Rockville, MD) supplemented with 10 % foetal bovine serum, 1× glutamine (200 mM) and penicillin/streptomycin (10 units/ml). SW620 cells were cultured in DMEM (Life Technologies, Rockville, MD) supplemented with 10 % foetal bovine serum, 1× glutamine (200 mM) and penicillin/streptomycin (10 units/ml). Cells were grown in incubators with humidified atmosphere of 5 % CO2 at 37 °C. Cells were tested on a monthly basis to ensure there was no mycoplasma contamination.
ROR2 pFLAG plasmid construction
A ROR2 pFLAG plasmid (pROR2) was constructed by isolating the ROR2 cDNA transcript from the Addgene ROR2 plasmid using Primer 1 (CTGATATCGATGGCCCGGGGCTCGGCGCTCCCGC) and Primer 2 (TCCTCTAGATCAAGCTTCCAG CTGGACTTGG). The resulting PCR fragment then underwent restriction enzyme digestion with both EcoRV and XbaI. The DNA was then subcloned into the pFLAG-CMV™-4 plasmid containing an N-terminal epitope tag following a similar restriction enzyme digest.
ROR2 siRNA Knockdown
Cells were seeded at 1 × 106 cells into 60 mm plates (Nunc™, Thermo Fisher Scientific, Rockford, IL USA) and allowed to adhere over a 6 h period. Cells were then serum starved for 18 h before being transfected with either 60 pmoles of ROR2 siRNA or scrambled control siRNA (Life Technologies, Rockville, MD). siRNA were premixed in 250 μl of serum free McCoy’s media (Life Technologies, Rockville, MD). siRNA mixture was then combined with 6 μl of Lipofectamine® 2000 (Life Technologies, Rockville, MD) premixed in 250 μl of serum free McCoy’s media before addition to cells. After transfection, cells were incubated at 5 % CO2 at 37 °C before being used in subsequent experimentation.
Ectopic ROR2 expression
Cells were seeded at 1 × 106 cells into 60 mm plates (Nunc™, Thermo Fisher Scientific, Rockford, IL USA) and allowed to adhere over a 6 h period. Cells were then serum starved for 18 h before being transfected with either 1.4 μg of empty pFLAG-CMV™-4 plasmid or 1.4 g of pmoles of ROR2 pFLAG plasmid. Plasmid solutions were premixed in 250 μl of serum free RPMI media (Life Technologies, Rockville, MD) for RKO cells and DMEM (Life Technologies, Rockville, MD) for SW620 cells. The plasmid solutions were then combined with 6 μl of Lipofectamine® 2000 (Life Technologies, Rockville, MD) premixed in 250 μl of the appropriate serum free media before addition to cells. After transfection, cells were incubated at 5 % CO2 at 37 °C before being used in subsequent experimentation.
Quantitative real time PCR
Cell samples underwent cell lysis using 2-mercaptoethanol and RNA extraction was carried out using the RNeasy Extraction Kit (Qiagen 74106). 1 μg of RNA was quantified and treated with RNase-free DNase (Life Technologies 18068–015). The DNase treated RNA was used for cDNA synthesis using Quantitect cDNA synthesis kit (Qiagen 205313) with appropriate negative controls. The primer sequence used for
ROR2 qRT-PCR was designed to amplify a region which included all known transcript variants of
ROR2 (Forward 5′-GTCCAACGCACAGCCCAAATC-3′ & Reverse 5′-CCGGTTGCCAATGAAGCGTG-3′). qRT-PCR was performed using SYBR® Mastermix Reagent (Qiagen 204056) and the M × 5000p Thermal Cycler. Each sample was run in triplicate and the experiment was run for 40 cycles.
ROR2 results and those of Wnt & EMT related genes (AXIN2, CCND1, JNK, NFATC1, CDH1, VIM) were normalised against 3 house-keeping genes (SDHA, RPL13A, HSP90AB1). Primer sequences for additional genes can be found in Additional file
4.
ROR2 knockdown qRT-PCR experiments were repeated in triplicate and statistical significance was evaluated using unpaired
t-test.
Combined bisulphite restriction analysis (COBRA) Assay
DNA was extracted from samples before undergoing bisulphite treatment using Ez DNA Methylation™ – Gold Kit (Zymo Research, Australia). The ROR2 promoter region was amplified using ROR2 COBRA semi-nested primers which covered a 436 bp region of the 1958 bp ROR2 CpG island where MBD-Seq data indicated the greatest level of coverage. (Forward 5′-GGGTTAYGTTTATTTTAGGATTTTGTTAGGT-3′ & Forward nested 5′-GTYGTGTGTTTTTGAAGGAGGAAGATT-3′ & Reverse 5′-CTCTCAATATCCCRAACTTCAAATAAAATCTAA-3′). The PCR product was digested with TaqI restriction enzyme (Fermentas) before undergoing gel electrophoresis in a 1.5 % agarose gel. Resulting bands were visualised under UV light.
Bisulphite sequencing
DNA was extracted from samples before undergoing bisulphite treatment using Ez DNA Methylation™ – Gold Kit (Zymo Research, Australia). ROR2 COBRA semi-nested primers were used to amplify the ROR2 CpG island region. The resulting PCR product was then ligated into pCR™2.1-TOPO® plasmid (Life Technologies, Rockville, MD) before being transformed into chemically competent DH5α™ E. coli bacteria. The bacteria were utilised to clone the PCR product before being plated onto LB agar plates for blue white selection. Bacteria which contained pCR™2.1-TOPO® plasmid with ROR2 PCR inserts were sequenced using BigDye® (Life Technologies, Rockville, MD) with ROR2 Reverse and Forward nested primers before undergoing Sanger sequencing (Ramaciotti Centre, UNSW Australia).
5-aza-2-deoxycytidine treatment
Cells were seeded at 1 × 106 cells into 60 mm plates (Nunc™, Thermo Fisher Scientific, Rockford, IL USA) and allowed to adhere over a 24 h period. Cells were subsequently treated to 2.5 μM concentrations of 5-aza-2-deoxycytidine (Sigma A3656). Treatment was repeated every 24 h over a 72 h period. Control cells were treated with the vehicle control of acetic acid instead of 5-aza-2-deoxycytidine.
Data analysis of TCGA cohort
Normalised
ROR2 expression and methylation data of tumour and matched normal tissue were obtained from The Cancer Genome Atlas (
http://cancergenome.nih.gov/) and analysed by Agilent microarrays and Illumina HiSeq 2000 RNA Sequencing. Methylation values were analysed using Illumina Infinium (HumanMethylation450) arrays and the beta-value average was obtained from methylation probes that fell within the 1958 bp
ROR2 CpG island. Statistical significance of matched patient tumour and normal samples were carried out using paired
t-test. Statistical significance of expression and methylation comparison of the entire cohort was evaluated using unpaired
t-test. Statistical significance of expression in low and high methylation samples was evaluated using unpaired
t-test. The results shown in these analyses are in whole or part based upon data generated by the TCGA Research Network;
http://cancergenome.nih.gov/.
Patient samples
Forty-seven normal and 88 adenoma samples were collected from patients at Westmead Hospital using endoscopic mucosal resection (Ethics committee approval number 2008/6/4.6 and 11194, Sydney West Area Health Service Human Research and Ethics Committee) [
54]. A further six fresh colorectal adenomas and paired adjacent normal mucosa samples were taken from surgical resection specimens from 3 males and 3 females at St Vincent’s Hospital, Sydney (Ethics committee approval number H00/022 and 00113) [
55]. Informed consent was obtained from all patients participating in the study. The adenomas obtained showed no evidence of invasive malignancy (Additional file
3).
Proliferation assay
Twenty four h after ROR2 siRNA transfection, ROR2 knockdown and control HCT116 cells were lifted using 1× 0.5 % Trypsin EDTA and seeded into a clear 96-well well plate (Nunc™, Thermo Fisher Scientific, Rockford, IL USA) at 1 × 104 cells/well. Cells were allowed to adhere for 2 h before 3 wells of ROR2 knockdown cells and 3 wells of control cells were treated with 10 μl of CCK-8 reagent (Dojindo Molecular Technologies, Inc. Rockville, MD) before the plate was wrapped in foil. 10 μl of CCK-8 reagent was also added to 3 media only wells to act as control blank readings. 2 h after addition of CCK-8 reagent, the treated wells were read on Spectramax 190 plate reader at 450 nm absorbance using the media only wells as blank readings. CCK-8 reagent was applied to additional triplicate wells at 24 & 48 h after the initial seeding and their 450 nm absorbance was read to determine changes in cellular proliferation. All subsequent readings for each siRNA treatment was normalised against the initial reading 2 h after seeding. The experiment was repeated in triplicate and statistical significance was evaluated using 2 way ANOVA.
Migration assay
Seven hundred μl of media supplemented with 20 % foetal bovine serum was added to the lower chamber of transwell migration plates while 200 μl of media supplemented with 1 % foetal bovine serum was added to the insert (Corning Incorporated – Life Sciences, One Becton Circle Durham, NC 27712 USA). 24 h after ROR2 siRNA transfection, knockdown and control HCT116 cells were lifted using 1× 0.5 % Trypsin EDTA and resuspended in 1 % FBS media to a concentration of 7 × 105 cells/ml. 100 μl of cell solution was added to the inserts. The plates were incubated for 48 h at 37 °C before the inserts were removed and washed twice in PBS. Cells were then fixed with 100 % methanol for 20 min before again being washed twice in PBS. Inserts were then stained with 1 % crystal violet for 30 min before being washed twice in PBS. Non-migrated cells on the upper surface of inserts were removed using cotton swabs. The transwell membrane was then excised and mounted onto a glass slide with mounting medium (Dako CS70330-2). 4 independent field counts at 20× magnification using ImageQuant TL Software were used to assess cell numbers. The experiment was repeated in triplicate and statistical analysis was evaluated using unpaired t-test.
Invasion assay
Transwell invasion plates with pre-coated matrigel were first rehydrated using warm serum free media for 2 h at 37 °C. Media was then removed and 750 μl of media supplemented with 20 % foetal bovine serum was added to the lower chamber while 100 μl of serum free media was added to the insert (Corning Incorporated–Life Sciences, One Becton Circle Durham, NC 27712 USA). 24 h after ROR2 siRNA transfection, knockdown and control HCT116 cells were lifted using 1× 0.5 % Trypsin EDTA and resuspended in 1 % FBS media to a concentration of 7 × 105 cells/ml. 200 μl of cell solution was added to the inserts. The plates were incubated for 48 h at 37 °C before the inserts were removed and washed twice in PBS. Cells were then fixed with 100 % methanol for 20 min before again being washed twice in PBS. Inserts were then stained with 1 % crystal violet for 30 min before being washed twice in PBS. Non-migrated cells on the upper surface of inserts were removed using cotton swabs. The transwell membrane was then excised and mounted onto a glass slide with mounting medium (Dako CS70330-2). 4 independent field counts at 20× magnification using ImageQuant TL Software were used to assess cell numbers. The experiment was repeated in triplicate and statistical analysis was evaluated using unpaired t-test.