Migration and invasion is inhibited by silencing ROR1 and ROR2 in chemoresistant ovarian cancer

Ovarian cancer survival remains poor despite recent advances in our understanding of genetic profiles. Unfortunately, the majority of ovarian cancer patients have recurrent disease after chemotherapy and lack other treatment options. Wnt signalling has been extensively implicated in cancer progression and chemoresistance. Therefore, we investigated the previously described Wnt receptors ROR1 and ROR2 as regulators of epithelial-to-mesenchymal transition (EMT) in a clinically relevant cell line model. The parental A2780- and cisplatin-resistant A2780-cis cell lines were used as a model of ovarian cancer chemoresistance. Proliferation, adhesion, migration and invasion were measured after transient overexpression of ROR1 and ROR2 in the parental A2780 cell line, and silencing of ROR1 and ROR2 in the A2780-cis cell line. Here we show that ROR1 and ROR2 expression is increased in A2780-cis cells, alongside β-catenin-independent Wnt targets. Knockdown of ROR1 and ROR2 significantly inhibited cell migration and invasion and simultaneous knockdown of ROR1 and ROR2 significantly sensitised cells to cisplatin, whilereas ROR overexpression in the parental cell line increased cell invasion. Therefore, ROR1 and ROR2 have the potential as novel drug targets in metastatic and recurrent ovarian cancer patients.


INTRODUCTION
The low survival rates for ovarian cancer patients are due, in part, to the late diagnosis and the development of resistance to traditional chemotherapy in recurrent disease. Despite seminal studies in recent years investigating the genetic and molecular aetiology of epithelial ovarian cancer, particularly the high-grade ovarian cancer serous subtype, 1-3 our understanding of ovarian cancer progression and chemoresistance is still very limited.
Used for over 30 years, cisplatin remains a common chemotherapy treatment option following cytoreductive surgery for patients with advanced ovarian cancer. 4 Cisplatin is a platinum analogue that prevents normal DNA function through reactions that cause intrastrand or interstand crosslinks. 5 Patients with ovarian cancer generally respond well to chemotherapy but the tumours often recur, which is a clinical challenge.
Ovarian cancer is a heterogeneous disease and each subtype has different propensities to metastasise. Extensive genomic analysis of high-grade ovarian cancer serous subtype has allowed for the classification of four major molecular subtypes: mesenchymal, immunoreactive, differentiated and proliferative. 1,[6][7][8][9] The mesenchymal phenotype exhibits a shorter disease-specific survival and increased Wnt signalling. 9 In individual patient case studies, a switch to this mesenchymal phenotype has been shown in samples taken from metastatic sites, such as omental implants in comparison with primary tumour. 7 Epithelial-to-mesenchymal transition (EMT) has also been implicated in high-grade ovarian cancer serous subtype invasiveness and chemoresistance through in vivo studies. [10][11][12][13] The EMT profile of ovarian cancer cell lines can predict responses to cisplatin, which suggests that the more aggressive, mesenchymal-type cells are more resistant to chemotherapy. 13 An important signalling cascade involved in EMT regulation is the Wnt signalling pathway, with increasing evidence suggesting that β-catenin-independent signalling has a critical role in this dynamic process. 10,[14][15][16][17][18][19][20][21][22] Recent studies demonstrate that the novel Wnt receptors ROR1 and ROR2 correlate with worse prognosis and drive EMT in a variety of tumour types including breast cancer, cervical cancer and melanoma. 15,[23][24][25][26] Our laboratory reported that upregulation of Wnt5a in epithelial ovarian cancer regulates EMT 10 and confirmed that the novel Wnt5a receptor, ROR2, is also upregulated in a patient cohort of ovarian cancer. We recently demonstrated that ROR2 and its sister receptor, ROR1, regulate ovarian cancer migration and invasion. 27 Patients who develop recurrent and chemoresistant disease generally do not have additional surgery. Therefore, owing to the challenge of patient sample collection, we used a well-established clinically relevant cell line model for continuing the investigation into RORs in ovarian cancer. Here we report that ROR1 and ROR2 expression is increased in this chemoresistant cell line and support our initial findings that ROR1 and ROR2 regulate ovarian cancer cell migration and invasion. ROR1 and ROR2 present as possible targets for novel therapies for the treatment of ovarian cancer.

RESULTS
ROR1 and ROR2 expression is increased in the A2780-cis chemoresistant cell line model Our previous study 27 identified that ROR1 and ROR2 are upregulated in epithelial ovarian cancer and regulate cell migration and invasion. Clinically, ovarian cancer patients present with aggressive disease, which often recurs after chemotherapy or radiation therapy. 28 Therefore, we chose to continue our investigation of RORs in epithelial ovarian cancer using a wellestablished chemoresistant cell line, the A2780-cis model. A2780-cis cells have an increased expression of both ROR1 and ROR2 at the transcriptional level compared with the parental A2780 cell line (Supplementary Figure 1A). The A2780 cell line had no mRNA expression of ROR2; however, both ROR1 and ROR2 protein were present in each cell line (Supplementary Figure 1B).
In addition to overexpression of both ROR receptors, the A2780cis cells had increased β-catenin-independent Wnt signalling, as seen by significant upregulation of downstream targets CJUN, RHOA and NFAT (Supplementary Figures 1C; P o0.05), and a possible decrease in β-catenin-dependent signalling, as measured by a decrease in AXIN2 (Supplementary Figure 1D).
Previously, it has been shown that the A2780-cis cells exhibit a mesenchymal phenotype with increased migration and invasion capacity compared with the parental A2780 cells. 11 Therefore, we confirmed the EMT profile of the A2780-cis cells and similarly found significant upregulation of important EMT genes such as MMP3 (600-fold increase) and BMP2 (50-fold increase) (Supplementary Figure 1E) and significant downregulation of CDH1 (50-fold decrease) and CDH2 (9000-fold decrease) (Supplementary Figure 1F).
ROR2, but not ROR1, silencing in A2780-cis decreases proliferation and has no effect on cell adhesion Successful small interfering RNA (siRNA) transfection of A2780-cis significantly decreased ROR expression at the mRNA (Figures 1a, d    Simultaneous ROR1 and ROR2 silencing sensitises A2780-cis cells to cisplatin A2780-cis cells were treated with increased concentration of cisplatin to determine cell viability. ROR1 and ROR2 knockdown separately had no effect on chemoresistance (Figures 6a and b). However, combined ROR1 and ROR2 knockdown had a minor but statistically significant chemosensitising effect on the A2780-cis cells to cisplatin when compared with the control cells (Figure 6c, P o0.05). This result was replicated using an additional set of ROR silencers ( Figure 6d).
Neither ROR1 nor ROR2 overexpression affects A2780 cell proliferation or migration We then investigated the role of RORs in this model through ectopic expression in the parental A2780 cell line. Successful plasmid transfection of A2780 was confirmed by quantitative PCR with an increased ROR expression at the mRNA level (Figures 7a-c). ROR overexpression does not change cisplatin sensitivity in the parental A2780 cell line A2780 cells were treated with increased concentration of cisplatin to determine cell viability. ROR1 and ROR2 overexpression separately had no effect on chemoresistance (Figures 12a and b). Double ROR1 and ROR2 transfection was also not sufficient to significantly increase cell chemoresistance (Figure 12c).

DISCUSSION
Here we have shown that silencing ROR1 and ROR2 decreases cell migration and invasion in the cisplatin-resistant cell line A2780-cis. We believe this is a more clinically appropriate model for the investigation of ovarian cancer as patients often develop recurrent and resistant disease. This paper supports our previous study, 27 which found that ROR2 was upregulated in ovarian cancer patients and that knockdown of ROR1 and ROR2 together significantly inhibited cell migration and invasion. Additionally, other groups have also reported the overexpression of ROR1 in ovarian cancer patients and linked to poor clinical outcome and increased capacity of spheroid formation. 29,30 Here we have additionally shown that these receptors may enhance chemoresistance, and that knockdown may sensitise these cells to cisplatin.
The ability of ovarian cancer cells to adhere to a secondary site is extremely important in ovarian cancer metastasis because of the unique dissemination process of shedding into peritoneal fluid. 31 During the adhesion assays, we noted that the A2780-cis cell line did not adhere to collagen but only fibronectin, whereas the A2780 adhered to both extracellular matrix components. This is an interesting observation and supports the profile array results (Supplementary Figure 1E and F), suggesting that these A2780-cis cells are in a process of upregulated EMT with very low adhesion components. 12 Recently, β-catenin-dependent Wnt signalling has been implicated in the A2780-and cisplatin-resistant cell lines. 32 It was shown that Wnt gatekeepers such as DKK1 and SFRP4 were significantly downregulated in the A2780-cis cells compared with A2780. Additionally, downstream targets such as JUN were upregulated, supporting our current and previous results. 17,27 The study, however, also saw lower β-catenin protein levels in the A2780-cis cells, which may indicate some β-catenin-independent inhibitory action via the ROR receptors; unfortunately, no other β-catenin-independent Wnt targets were investigated.
Overcoming chemoresistance is a major challenge in epithelial ovarian cancer treatment and although there is initial remission in 75% of patients, subsequent recurrence occurs o2 years posttreatment. 33 Attempts to identify characteristics of chemoresistant cells have shown markers of EMT in vitro. 34 Silencing of EMT transcription factors Snail and Slug induces chemosensitivity in A2780 cisplatin-resistant cells, while upregulation leads to radiotherapy resistance and chemoresistance in A4 ovarian cancer cells. 11,35 In our study, we found that simultaneous silencing of ROR1 and ROR2 significantly sensitised cells to cisplatin (Figure 6c), a commonly used chemotherapy agent for ovarian cancer patients. We confirmed these results using an additional set of siRNA silencers (Figure 6d). We hypothesise that the A2780cis cells undergo a reversion of EMT, that is, gain epithelial characteristics after double ROR knockdown, which causes their sensitivity to cisplatin. Indeed, it would be important to investigate these changes further through profiling stable knockdown and patient cell models.
During the course of the viability assays, we observed a shift in the chemoresistance of the A2780-cis cells. The viability assay with siRNA set B was completed~4 months after set A. Therefore, it is highly likely that these cells would have changed under consistent pressure from cisplatin treatments and continual passaging. However, even though this was seen, we still noted some reduction in chemoresistance after double ROR knockdown using silencer set B (Figure 6d). The chemosensitising results, although significant at some concentration points, are minor and investigations are ongoing.
In this study, we demonstrated that ROR2 silencing alone decreased cell proliferation (Figure 2c), yet ROR1 or double ROR1 and ROR2 silencing did not (Figures 2a and e). We also found that ROR1 silencing alone had no effect on invasion (Figure 5a), yet ROR2 and double ROR1 and ROR2 silencing significantly perturbed invasion (Figures 5c and e). We also performed ROR1, ROR2 and simultaneous ROR plasmid overexpression in the parental A2780 cell line. Overexpressing the ROR receptors was not sufficient to increase migration ( Figure 10) or chemoresistance of the A2780 cells ( Figure 12). Interestingly, we did observe an increase in invasion after ROR1, ROR2 and double overexpression ( Figure 11). ROR1 knockdown in the chemoresistant A2780-cis cell line did not significantly alter cell invasion ( Figure 5a); however, ROR1 overexpression in the parental A2780 cells did. This may be because the A2780 cell line does not normally express ROR1; therefore, overexpressing the protein at very high levels could markedly change their behaviour.
These differing results between ROR1 and ROR2 after silencing and overexpression suggests that they may facilitate alternative signalling pathways as discussed previously. 27 Recently, it has been found that ROR1 is critical in the structure and formation of caveolae 36 and is a target for the frequently amplified NKX2-1 gene in lung adenocarcinoma. 37 Interestingly, its significance as a scaffold protein for a number of other RTKs was found in sustaining caveolae formation and prosurvival signalling. ROR2 may be more involved in planar cell polarity signalling through GTPases RhoA and Rac, a branch of the β-catenin-independent Wnt signalling pathway. 38 CONCLUSION ROR1 and ROR2 are upregulated in a chemoresistant model of ovarian cancer and regulate cell migration and invasion through EMT. Targeting the ROR receptors provides a potential avenue for novel therapeutics to overcome the current inevitable development of chemoresistance in ovarian cancer patients.

Cell culture
The epithelial ovarian cancer cell line A2780 was kindly donated by Dr Michelle Henderson (Children's Cancer Institute, UNSW, Sydney, Australia) and was originally purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Its daughter chemoresistant cell line A2780-cis was obtained from Sigma-Aldrich (no. 93112517, St Louis, MO, USA). Both cell lines were cultured as per the supplier's recommendations (RPMI-1640 containing 10% foetal bovine serum). Media were supplemented with penicillin/streptomycin and GlutaMAX (Life Technologies, Carlsbad, CA, USA). A2780-cis cells were maintained in 1μM cisplatin every 2-3 passages. Cells were grown in 5% CO 2 at 37°C and were routinely tested negative for mycoplasma contamination.

ROR silencing
Cells were transfected with siRNA targeting ROR1 and ROR2 as described previously. 27 ROR silencing efficiency was confirmed by western blotting and quantitative real-time polymerase chain reaction (qRT-PCR) as described previously. 27

ROR overexpression
A plasmid encoding human ROR2 with pFLAG tagged at the N-terminal end was constructed by subcloning the ROR2 cDNA transcript into pFLAG-CMV-4 plasmid (Sigma-Aldrich). ROR1 mouse expression plasmid (pcDNA3-mRor1-flag) was generously donated by Prof Yasuhiro Minami (Kobe University, Kobe, Japan). Plasmid transfections were conducted using polyethylenimine (Polysciences, Warwrington, PA, USA) according to the manufacturer's specifications. In all transfection experiments, 1 × 10 6 cells were seeded into 6-well plates and serum starved overnight. Cells were transfected the following day with 700 ng of ROR1 or ROR2 expression vector, or an empty vector pFLAG control (labelled in figures as negative control). Transfection mixture was removed 24 h later and cells were washed with serum-free media and replaced with complete media containing 10% foetal bovine serum. The efficiency of ROR overexpression was confirmed by western blotting and qRT-PCR. In addition to the primers previously used to detect ROR1 and ROR2, 27 additional ROR1 primers were used to detect the ROR1 expression plasmid (F: 5′-GGGCAACCAACTATGGCTCT-3′; R: 5′-TGTTGCCACACACT GGAAGT-3′).

EMT PCR Array
Human EMT RT 2 Profiler PCR Array (PAS0907; SABiosciences Qiagen, Valencia, CA, USA) and RT 2 Real-Timer SyBR Green/ROX PCR Mix (SABioscience Qiagen) was used to measure mRNA expression levels of 93 EMT genes in the A2780 and A2780-cis cell lines according to the manufacturer's protocol. Each array was repeated in triplicate.

Proliferation assay
As described previously, 27 cell proliferation was measured using the CCK8 Kit (Dojindo, Rockville, MD, USA) according to the manufacturer's protocol. Briefly, 7 h after ROR siRNA or plasmid transfection, 100 μl of cells were seeded in triplicate into a 96-well plate at a concentration of 4 × 10 4 cells per ml in triplicate. At 3 h after the addition of CCK8, absorbance was read

Viability assay
Cell viability was measured using the CCK8 Kit (Dojindo) according to the manufacturer's protocol. At 24 h after ROR siRNA or plasmid transfection, 100 μl of cells were seeded in triplicate into a 96-well plate at a concentration of 4x10 4 cells per ml. They were left to adhere for 24 h, treated with cisplatin in serial dilutions up to a concentration of 1000 mM and left to incubate at 37°C for another 24 h. At 3 h after the addition of CCK8, absorbance was read at 450nm using the SpectraMax 190 Microplate Reader (Molecular Devices). A2780 cell viability was repeated in duplicate and used as a comparison for A2780-cis, which was repeated in triplicate. A second set of ROR silencers were used to validate the results, as described previously. 27 Adhesion assay As described previously, 27 cell adhesion was measured against collagen type I and fibronectin. After 2 h of incubation on precoated collagen or fibronectin plates, cells were stained with crystal violet and lysed with acetic acid. Absorbance was measured at 595 nm using the SpectraMax Plus190 Microplate Reader (Molecular Devices). Each assay was repeated in triplicate.

Wound healing assay
Wound healing was analysed using IBIDI Culture-Inserts (IBIDI GmbH, Martinsried, Germany) as described previously. 27 Cells were plated at a concentration of 1 × 10 5 cells per ml, and after 24 h of incubation, culture inserts were removed. Photographs of the movement of cells into the scratch area were taken every 6-12 h until the scratch area had closed using a Leica DMIL microscope (Leica Microsystems, North Ryde, NSW, Australia). Wound healing was then analysed using TScratch software (ETH Zurich, Zurich, Switzerland). 39 Each assay was repeated in triplicate.

Transwell migration assay
Cell migration was measured using Transwell inserts (Corning Life Sciences, Tewksbury, MA, USA) as described previously, 27 according to the manufacturer's instructions. The transwells for the A2780 cells required overnight precoating of collagen (10 μg/ml in H 2 O) before seeding. Cells   were seeded in transwell inserts at a concentration of 1 × 10 6 cells per ml (200 μl) and incubated overnight. ImageJ (Java) software (Rasband, NIH, Bethesda, MD, USA) was used to obtain an average cell count of the four stained membrane images. Each assay was repeated in triplicate.

Transwell invasion assay
Cell invasion was measured using Matrigel precoated transwell inserts (BioCoat Matrigel Invasion Chambers, Corning Life Sciences, Tewksbury, MA, USA) according to the manufacturer's instructions as described previously 27 and were optimised using untreated cells. Cells were seeded in transwell inserts at a concentration of 5 × 10 5 cells per ml (100 μl) and incubated for 48 h. ImageJ (Java) software was used to obtain an average cell count of the four stained membrane images. Each ROR silencing assay was repeated in triplicate and each ROR overexpression assay was completed in duplicate.

Statistical analysis
As described previously, 27 all in vitro experimental results are expressed as mean ± s.d. An F test was first used to determine equal or unequal data variance (s.d.). A Student's t-test type 2 was used to determine significance if equal variance. If unequal variance, a Student's t-test type 3 was used to determine significance. T-test values below Po 0.05 were considered statistically significant. *Po 0.05, **Po0.01 and ***P o0.001.