Inhibition of MUS81 by siRNA Reverses Resistance to Cisplatin in Human Ovarian Cancer Cell Lines

Purpose: Drugs that induce DNA interstrand crosslinks form the mainstay of anticancer treatments for different cancers. These drugs are used to treat ovarian cancer which is the most prevalent gynaecological cancer. Five-year survival rates are approximately 40% and the development of drug resistant disease is an important factor in treatment failure. Methods: In this study a comprehensive evaluation of the expression and function of the site-specific endonuclease MUS81 was conducted. Using quantitative real time PCR analysis and imaging flow cytometry we determined the mRNA and protein expression of MUS81 in three ovarian cancer cell lines and two immortalised human fibroblast cell lines which had been made resistant to cisplatin by chronic exposure. siRNA knockdown of MUS81 was employed to determine the effect on overall cell survival which was assessed using clonogenic assays. Results: In the five cisplatin-resistant cell lines we observed increased MUS81 mRNA expression. In addition MUS81 protein expression in the form of discrete nuclear foci in cells was observed in all cell lines following cisplatin exposure, there being significantly more foci in cisplatin resistant cell lines. siRNA knockdown of MUS81 significantly reduced both mRNA and protein levels in two cell lines (SK-OV-3 and MRC5-SV1 – wild-type and resistant) and critically re-sensitised cisplatin resistant cells to wild-type level, determined by clonogenic assay. Conclusion: MUS81 is central to the development of cisplatin resistance in ovarian cancer cell lines. Inhibition of MUS81 restored drug sensitivity to the cells. MUS81 may be a useful therapeutic target to overcome drug resistance in ovarian and other cancers. range primer mix; Fast This one well of and cDNA


Introduction
Chemotherapeutic drugs that introduce DNA interstrand crosslinks (ICLs) are central to the cytotoxicity induced in many chemotherapeutic regimens for the treatment of cancer. In fact this class of bifunctional alkylating agents are the mainstay of many successful anticancer treatments (Huang and Li, 2013). Indeed, the persistence of ICLs in the DNA of cancer cells following drug exposure is directly related to the cytotoxicity (cell killing) efficiency of ICL-mediated chemotherapy drugs (Palom et al., 2002, Clingen et al., 2007. ICLs are lesions that are composed of a covalent linkage between opposite strands of the DNA double helix. They result in DNA strand distortion whereby the ICL prevents strand dissociation thus inhibiting DNA synthesis, replication and transcription ultimately leading to cell death (Deans and West, 2011). ICLs are moreover a challenging lesion for the cellular DNA repair machinery to resolve and require the coordinated interaction of distinct DNA repair mechanisms (Hashimoto et al., 2016). While the precise mechanism of ICL repair is still to be fully elucidated, a fundamental understanding of the mechanism together with essential molecular components has been well-described. In brief, following the creation of an ICL by drug exposure and the resultant DNA strand distortion, proteins of the Fanconi Anaemia (FA) pathway are recruited to the lesion to conduct the initial incision events leading to ICL removal. Here the Fanconi-associated nuclease 1 (FAN1) with its inherent 5'-3' exonuclease activity and 5'FLAP endonuclease activity cleaves the ICL in a process called unhooking.
As a result, the stalled replication fork caused by the initial ICL is converted into a DNA double strand break. Following the initial incision activity of the FAN1 nuclease other endonuclease which include MUS81-EME1 and the XPF-ERCC1 protein cut DNA on the 3' side of the ICL. The result here is a double strand break with potential loss of genetic information which if repaired by non-homologous end joining would potentially lead to frameshift insertion and deletion (INDEL) mutations. However, in cycling cells in the G2 and S phases, the lesion becomes substrate for repair by homologous recombination where a Holliday junction is created mediated by the Rad 51 protein. Following strand resolution by the action of resolvases, the DNA is repaired in an error free manner (Deans and West, 2011;Huang andLi, 2013, Hashimoto et al., 2016).
Inherited defects in ICL repair factors are linked to human disorders associated with hypersensitivity to DNA damaging agents including chemotherapeutic drugs which induce ICLs, and also increased cancer incidence such as FA and Xeroderma Pigmentosum (XP). There are some 17 components of the FA group of proteins where clinical and cellular hypersensitivity to DNA damaging agents may be displayed to a greater of lesser extent (Palovcak et al., 2017). In addition, there are seven complementation groups of XP whose proteins mediate nucleotide excision repair of large bulky adducts such as ultraviolet radiation induced dimer photoproducts and also ICLs (Knoch et al., 2012).
The development of resistance to anticancer chemotherapy can limit the effectiveness of treatment for many human adult solid cancers (e.g., Carvalho et al., 2010). In many cases tumours intially respond well to chemotherapy treatment but can subsequently relapse with drug resistant disease. The mechanisms whereby the cancer cells acquire resistance can occur by multiple mechanisms and include, for example, increased drug efflux due to overexperssion of p-glycoprotein leading to multiple drug resistance ; elevated damage tolerance following drug exposure (Johnson et al., 1997)  repair, specifically enhanced activity of the XPF-ERCC1 site specific endonuclease (Ferry et al., 2000).
Moreover, in vivo studies have demonstrated alterations in HR activity with increased expression of key HR proteins such as Rad51 in human ovarian cancer (e.g., Helleday, 2010).
In this present study we have examined the role of the MUS81 3' structure-specific endonuclease in the repair of ICLs in drug resistant ovarian cancer cell lines. As stated earlier, one of the integral proteins in the ICL repair process is MUS81. It associates with either EME1 or EME2, forming a 5' to 3'structure-specific endonuclease. MUS81 was first identified in fission yeast through its association with Cds1, a replication checkpoint kinase and in budding yeast where it associates with Rad54, a recombination repair protein (Boddy et al., 2000). It is involved in vital cellular processes such as meiosis (Agostinho et al., 2013) and resolving recombination intermediates by cleaving nicked Holliday junctions (Chen et al, 2001). Also, Hanada et al., 2007 found that stalled replication forks were unable to be recovered in the absence of MUS81; ultimately leading to the formation of chromosomal aberrations.
There is evidence that the function of MUS81 can mediate the response of cancer cell lines to anticancer drugs. For example, siRNA inhibition of MUS81 enhanced the sensitivity of breast cancer cell lines to the anticancer drug 5-fluorouracil (Qian et al., 2014). Also, MUS81 has been shown to be associated with alterations in cell proliferation and cisplatin sensitivity in serous ovarian cancer in vivo (Xie et al., 2016). Finally, resistance to cisplatin in human hepatocellular carcinoma cell lines is increased by SiRNA knockdown of MUS81 (Wu et al., 2016) Here we examine the role of MUS81 in resistance to cisplatin in five pairs of cell lines. Three pairs are derived from epithelial ovarian cancer in which a derivative of each wild-type cell line has been made resistant to cisplatin by chronic exposure to cisplatin (SK-OV-3, A2780 and PEA-1). Two pairs of the cell lines were derived from SV40 large-T antigen immortalised human diploid fibroblasts (MRC5-SV1 and NB1-hTERT) where again cisplatin resistance was developed by chronic drug exposure (Adam-Zahir et al., 2014). Increased MUS81 protein expression was observed in the cisplatin resistant derivative when compared to the sensitive wild-type counterpart. Elevated protein expression was detected in the form of MUS81-containing DNA damage foci during imaging flow cytometry (e.g., Parris et al., 2015). Moreover, this finding was corroborated by examination of mRNA levels using quantitative PCR. Here elevated levels of MUS81-specific mRNA were detected in the drug resistant cell lines compared to wild-type. Most importantly however, in two cell lines examined (SK-OV-3 and MRC5-SV1) siRNA inhibition of MUS81 resulted in a reversal of cisplatin resistance to wild-type levels.
It is concluded that elevated repair of ICLs during HR repair is an important mechanism whereby cancer cells (and in particular ovarian cancer) acquire resistance to anticancer crosslinking agents. This present study and together with previous investigations point to elevated expression of the MUS81 endonuclease contributing to drug resistance. We submit that MUS81 could provide a useful biochemical marker of induced drug resistance in ovarian (and other) cancers and may provide a target where selective inhibition of the endonuclease could reverse drug resistance in recurrent disease.

Cell Lines
Cell lines were purchased from the European Collection of Cell Cultures (ECACC) (Porton Down, Salisbury, Wiltshire, UK) or generated at Brunel University London (Uxbridge, Middlesex, UK).
Cisplatin resistant cell lines were created by continuous exposure of the wild-type cell line to increasing doses of the cis-Diamine-dichloro-platinum (cisplatin). Details of the cell lines are listed in Table 1. Cell number and viability were determined as required using a Countess™ automated cell counter based upon the method of trypan blue exclusion (Invitrogen, Renfrewshire, UK). Cell lines were used over a restricted range of ten passages, during which cell viability was not less than 90% for any experiment.
Resistance to either cisplatin in the resistant derivative cell lines was maintained by exposing the cells while growing as monolayers at approximately 80% confluence, to the relevant drug for one passage every three passages.

Exposure of Cells to Chemotherapeutic Drugs
Cisplatin solutions (Sigma Aldrich Ltd, Gillingham, Dorset, UK) were prepared initially in dimethyl sulfoxide (Fisher Scientific, Loughborough, Leicestershire, UK) and final concentrations of drug solutions were prepared initially in the appropriate serum-free cell culture medium for the cell lines (Labtech International Ltd). Cisplatin was then further diluted in the relevant complete medium to the required concentrations.

Clonogenic Assay after Exposure of Cells to Chemotherapeutic Drugs
Approximately 500 000 cells were seeded into 5 cm dishes in 5 ml complete medium and incubated overnight. Cells were exposed to the drug solutions for 1 hour, after which the cells were Cells were re-suspended in the appropriate complete medium, counted using a "Countess™"

MTT Assay after Exposure of Cells to Chemotherapeutic Drugs
The A2780 wild-type and cisplatin resistance derivative did not form discrete colonies that were easy to count in an accurate manner following a clonogenic assay. To determine cell viability after cisplatin exposure MTT assay was conducted on these cell lines instead. Cells were plated out into 96-well microplates at a density of 25 000 cells in 100 µL per well and exposed to cisplatin for 1 hour, then incubated (incubation conditions described previously) for 48 hours. The 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) reagent (50 mg) (Merck-Millipore Ltd, Watford, UK) was dissolved in 10 ml PBS and sterilised by filtration. 10 µL of this solution was added to each well containing cell suspension, and the contents mixed by gentle agitation of the plate, followed by incubation for 3 hours (incubation conditions detailed previously). was added to each sample, except the AlexaFluor®488 compensation samples. The samples were submitted for analysis by imaging flow cytometry.

Imaging Flow Cytometry after Exposure of Cells to Chemotherapeutic Drugs
Imaging flow cytometry was undertaken using the Imagestream X Mark II system equipped with INSPIRE™ data acquisition software (Luminex Instruments, Thermofisher Ltd, UK) as previously

Analysis of Cell Images -Calculation of Foci Number
Foci indicating the presence of the MUS81 protein were quantified in approximately 10,000 (1000 for the compensation samples) images of cells per time point captured using the INSPIRE™ imaging flow cytometry software. Foci were quantified in a similar manner as described previously (Parris et al., 2015), with the spot counting wizard in the IDEAS™ software permitting simplified foci quantitation. The wizard consists of sequential steps that identified all captured images of single cells that were in optical focus. Next, images of fifty cells from this population with varying numbers of clearly defined foci were selected to create a truth population. Images from the truth population were used to produce a mask that identified only those foci located within the cell nuclei of each image. The finalised truth population was then saved as a template and applied to quantitate foci present in all acquired cell images.

Exposure of Cells to Chemotherapeutic Drugs and Cell Fixation for Quantitative Real-time PCR Analysis
Approximately 1 × 10 6 cells were seeded into 10 cm dishes in 10 ml complete medium and incubated overnight (incubation conditions described previously). Cells were exposed to cisplatin for 1 hour, after which the cells were trypsinised and counted as previously described.

Verification of RNA Quality using Spectrophotometry & Agarose Gel Electrophoresis (qPCR)
The quality and quantity of the RNA was measured using a NanoDrop™ 2000c spectrophotometer (ThermoFisher Scientific, Hemel Hempstead, Hertfordshire, UK). Quality was determined using the 260/280 value, and quantity was measured in nanograms/microlitre.
The integrity of the RNA was verified by gel electrophoresis using a 2% agarose gel.

Production of cDNA from RNA (q-PCR)
All the following reagents were supplied by Invitrogen: random hexamer primers; dNTP mix; first strand buffer; DTT; SuperScript® III reverse transcriptase. DEPC-treated water was supplied by Sigma-Aldrich and RNase Away® reagent by Life Technologies.
The RNA solutions were diluted with DEPC-treated water so that the reaction mixture contained at

Verification of cDNA Quality (q-PCR)
The quality of the cDNA was verified by PCR followed by agarose gel electrophoresis. The reagents used were: ReadyMix™ REDTaq® PCR reaction mix with MgCl2 (Sigma-Aldrich); nucleasefree water (ThermoFisher Scientific); primer pair for the GAPDH gene at 5 µM concentration with the following base sequences: forward primer GAAGGTGAAGGTCGGAGTC; reverse primer GAAGATGGTGATGGGATTTC (Sigma-Aldrich). The integrity of the PCR products (and hence the quality of the cDNA) was verified by gel electrophoresis using a 2% agarose gel.

Determination of the Optimum Housekeeping gene for qPCR using the geNorm™ Kit
Two reference genes were identified using the geNorm™ Human Reference 12-Gene Selection Kit (Primerdesign Ltd, Eastleigh, Hampshire, UK) as per the manufacturer's instructions.
The kit included primer sets for the following reference genes: The resulting data was analysed with the qbase PLUS software, which gave the average expression stability value for each reference gene. The two most stable reference genes with the least variation between cDNA samples were found to be YWHAZ and SDHA. Therefore, these two genes were used as internal reference standards for all subsequent q-PCR analyses.

Establishing the Percentage Efficiency of the MUS81, YWHAZ & SDHA Primers (QPCR)
The percentage efficiency of the MUS81, YWHAZ and SDHA primer pairs from the geNorm™ Human Reference 12-Gene Selection Kit (details as above) was determined as follows. The reagents

Reduction of MUS81 Gene Expression using the siRNA Knockdown Technique
Wild-type SK-OV-3, cisplatin resistant SK-OV-3 R , wild-type MRC-5 SV1 and cisplatin resistant MRC-5 SV1 R cells were plated out into 6-well plates at optimum densities (established previously).
The transfection reagents (all supplied by Dharmacon) were prepared as follows. The The antibiotic-free complete medium was removed from the cells by aspiration and the siRNA/transfection reagent/antibiotic-free complete medium solution added. The cells were incubated (conditions described previously) for 3 days if a single knockdown was being performed, before being processed for either MUS81 protein expression or for a clonogenic assay. A double knockdown was achieved by incubating the cells for 2 days in the first batch of siRNA/transfection reagent/antibiotic-free complete medium solution, then removing and replenishing it with a fresh batch of siRNA/transfection reagent/antibiotic-free complete medium solution, before processing the cells for either MUS81 protein expression or for a clonogenic assay after a further 2 days incubation.

Determination of MUS81 mRNA and Protein Expression Post (Double) siRNA Knockdown after Cisplatin Exposure using qPCR and/or Imaging Flow Cytometry
The aim of these experiments was to determine the MUS81 expression using either qPCR or imaging flow cytometry on cells that had undergone a double knockdown event before being exposed to cisplatin for 1 hour. In both cases, the cells underwent the double knockdown of the MUS81 gene (method described above). Two days after the addition of the second batch of siRNA/transfection reagent/antibiotic-free complete medium solution, the cells were incubated for 1 hour in 312 µg/ml cisplatin. Cisplatin-containing medium was removed and replaced with the appropriate complete medium, the cells were incubated for 24 or 48 hours before being fixed for either qPCR or imaging flow cytometry to determine MUS81 foci number (MUS81 protein expression).

Determination of Cell Survival using Clonogenic Assays Post (Double) siRNA Knockdown of MUS81 Gene with Cisplatin Exposure
The aim of these experiments was to determine cell survival using a clonogenic assay on cells that had undergone a double knockdown event before being exposed to cisplatin for 1 hour.
Cells were plated out for the clonogenic assay (as described above) 24 hours after being exposed to 12 µg/ml cisplatin for 1 hour, at a density of 2000 cells/10 cm plate. The cells were incubated for two or three weeks, before being fixed, stained and counted as described above.

Determination of Cisplatin-Resistance in cell lines.
In  Table 2.

q-PCR Determination of MUS81 mRNA Levels in Wild-Type and Cisplatin resistant ovarian cancer
and immortalised human fibroblast cell lines.

Expression of MUS81 Foci in Cell Nuclei Following siRNA Knockdown after Cisplatin Exposure
To determine the expression of MUS81 protein in the form of nuclear foci, cells were stained for MUS81 using the immnocytochemical method described above, following siRNA 'double' knockdown and exposure to cisplatin. As proof principle of the effect of a reduction in MUS81 expression on foci formation, analysis was carried out in two of the five pairs of cell lines described above, namely wild-

Cell Survival of SK-OV-3 and MRC5-SV1 Cells after MUS81 Knockdown
To determine the effect on clonogenic cell survival of the SK-OV-3 and MRC5-SV1 cisplatin resistant cells following siRNA MUS81 knockdown, cells were subjected to either a scrambled or on-target MUS81 double knockdown (over a period of four days) after which cells were exposed to an IC50 concentration of cisplatin (12 µg/ml) for one hr. Survival in the siRNA on target and scrambled knockdown cells was compared to that of untreated control cells and the data can be seen in Figure   7. For SK-OV-3, in scrambled cells survival was significantly reduced from 100% in untreated to 91.82% +/-1.52 SE (P = 0.0057, Students unpaired T-test). However, in on target siRNA cells, survival was further and significantly reduced to 47.55% +/-5.16 SE (P = 0.00053, Students unpaired T-test).
In the MRC5-SV1 cell a similar pattern of cell survival was observed after exposure to an IC50 concentration of cisplatin. In cells exposed to scrambled siRNA survival was non-significantly reduced

Funding
This study was made possible by a grant from the Bart's Charity, UK https://www.bartscharity.org.uk/

Conflicts of Interest/Competing interests
There are no conflicts of interest or competing interests declared by any author.