RIP140 regulates POLK gene expression and the response to alkylating drugs in colon cancer cells

Aim: The transcription factor RIP140 (receptor interacting protein of 140 kDa) is involved in intestinal tumorigenesis. It plays a role in the control of microsatellite instability (MSI), through the regulation of MSH2 and MSH6 gene expression. The aim of this study was to explore its effect on the expression of POLK, the gene encoding the specialized translesion synthesis (TLS) DNA polymerase κ known to perform accurate DNA synthesis at microsatellites. Methods: Different mouse models and engineered human colorectal cancer (CRC) cell lines were used to analyze by RT-qPCR, while Western blotting and luciferase assays were used to elucidate the role of RIP140 on POLK gene expression. Published DNA microarray datasets were reanalyzed. The in vitro sensitivity of CRC cells to methyl methane sulfonate and cisplatin was determined. Results: RIP140 positively regulates, at the transcriptional level, the expression of the POLK gene, and this effect involves, at least partly, the p53 tumor suppressor. In different cohorts of CRC biopsies (with or without MSI), a strong positive correlation was observed between RIP140 and POLK gene expression. In connection with its effect on POLK levels and the TLS function of this polymerase, the cellular response to methyl methane sulfonate was increased in cells lacking the Rip140 gene. Finally, the association of RIP140 expression with better overall survival of CRC patients was observed only when the corresponding tumors exhibited low levels of POLK, thus strengthening the functional link between the two genes in human CRC. Conclusion: The regulation of POLK gene expression by RIP140 could thus contribute to the maintenance of microsatellite stability, and more generally to the control of genome integrity.


INTRODUCTION
Colorectal cancer (CRC) is one of the most frequent cancers worldwide and genetic instability exerts a driving role in this malignancy [1] . The mismatch repair (MMR) system is one of the various cellular systems involved in the maintenance of genome integrity through the correction of mistakes that occur during DNA replication. Once impaired (as occurs in 2%-3% of CRC cases with an inherited component including Lynch syndrome [2,3] ), it generates microsatellite instability (MSI) and a hypermutated tumor phenotype with a high frequency of point and frameshift mutations [4,5] .
In addition to molecular machineries that cope with DNA repair, mammalian cells possess enzymes with translesion DNA synthesis (TLS) activity. The POLK gene encodes Polκ, one of the Y-family TLS polymerases, which possesses unique DNA damage bypass capability [6] . Interestingly, despite any associated proofreading exonuclease activity, Polκ was shown to display high accuracy during dinucleotide microsatellite DNA synthesis [7] , and Polκ-/-mice display a spontaneous mutator phenotype in various tissues [8] . This could be linked to its less open active site compared to other inaccurate specialized DNA polymerases from the Y family [9] . In fact, Polκ possesses a unique structural feature, an extension of its Nterminal region, called N-clasp, that protrudes from the thumb domain and encircles DNA in proximity to a primer terminus [10] . This results in a much more stable complex once Polκ binds to a mismatched DNA substrate, facilitating the extension step across various minor groove distortion or lesion. Besides its role in MSI maintenance together with MMR [7] , Polκ seems to prevent DNA damage-induced toxic effects of methylnitrosourea, another process dependent on the MMR system [11] . This link between MMR and Polκ is reinforced by the findings that POLK interacts with MSH2 [12] and partially protects human cells from the MMR-dependent cytotoxicity of O6-methylguanine lesions [11] .
Our laboratory recently reported that the RIP140 (receptor interacting protein of 140 kDa) gene was involved in the normal and tumoral development of the intestinal epithelium. RIP140, also known as NRIP1 (nuclear receptor-interacting protein 1), was first identified as a transcriptional repressor of nuclear hormone receptors [13,14] . We and others then characterized RIP140 as a coregulator of various transcription factors, including, for instance, E2F [15] or NFKB [16] . The repressive activity of RIP140 involves several inhibitory domains interacting with histone deacetylases [17] and is controlled by different post-translational modifications [18] . Using a mouse model lacking the Rip140 gene, various physiological processes were shown to be regulated by RIP140, including female fertility [19] and mammary gland morphogenesis [20] , fat metabolism [21] , pro-inflammatory cytokine response [22] , or cognition [23] .
In the intestinal epithelium, our laboratory demonstrated that RIP140 inhibits the Wnt/β-catenin signaling pathway and, as a consequence, exerts an anti-proliferative effect [24] . In line with this result, we found that this transcription coregulator inhibits the Paneth cells lineage through the regulation of SOX9 expression and activity [25] . In addition, RIP140 expression decreased in CRC samples compared to the adjacent healthy tissue. In sporadic CRC, RIP140 mRNA and protein levels significantly correlated with better overall survival of patients and were identified as good prognosis markers [24] . More recently, we demonstrated that RIP140 was acting as a transcriptional regulator of MSH2 and MSH6 gene expression and was involved in the regulation of the MSI and hypermutator phenotype in CRC cells [26] . Interestingly, a frameshift mutation in the RIP140 coding sequence identified in MSI CRC tumors exhibits a dominant negative activity and correlates with shorter overall survival of patients with advanced CRC.
In the present study, we explored whether RIP140 could regulate POLK as a factor involved in microsatellite stability. We demonstrated that, in both engineered mouse models and human CRC cells, RIP140 positively regulates the POLK gene expression at the transcriptional level, at least partly via a p53-dependent mechanism. A strong correlation was observed between the expression of the RIP140 and POLK genes in CRC biopsies. Moreover, cells knocked out for the Rip140 gene were shown to be more sensitive to methyl methane sulfonate (MMS), a drug known to induce DNA lesions activating a POLK response. Finally, our data demonstrate a significant impact of POLK gene expression on the prognosis value of RIP140 in human CRC. We propose that the modulation of POLK gene expression by RIP140 could reinforce its effect on the maintenance of genome integrity and, more particularly, on the stability of microsatellite sequences.

Cell culture and transfections
Mouse embryonic fibroblasts (MEFs) derived from wild-type or RIPKO mice previously described [24] , and the stably transfected MEFs described by Palassin et al. [26] were grown in DMEM-F12 medium supplemented with 10% FCS, 100 U/mL penicillin, 100 mg/mL streptomycin, and 100 mg/mL sodium pyruvate, with 40 µg/mL puromycine for selection of stably transfected cells. HCT116, RKO, SW480, and HT29 human colon cancer cells were grown as previously described [26] and stably transfected with the empty pEGFP vector (Clontech®) or with the same vector containing the full-length human RIP140 cDNA. The SW620 human cell line was grown identically. HCT116-GFP and HCT116-RIP140 cells were previously described [24] and grown in McCoy medium and 750 μg/mL G418. Small interfering RNA (siRNA) transfections were performed using INTERFERin® on cells seeded the day before in a 6-well plate (3 × 10 5 cells per well). Each transfection was performed in triplicate, and interference efficiencies were tested by quantitative RT-PCR.

Animals
C57BL/6J Rip140 -/-(RIPKO) mice [19] were given by Pr MG Parker (Imperial College London, London, UK). C57BL/6/129 RIP140 transgenic (RIPTg) mice were generated using the Speedy Mouse® Technology (Nucleis) by insertion of a single copy of the human RIP140 cDNA at the HPRT locus [24] . All animals were maintained under standard conditions, on a 12 h:12 h light/dark schedule and fed a chow diet ad libitum, according to European Union guidelines for the use of laboratory animals. In vivo experiments were performed in compliance with the French guidelines for experimental animal studies (agreement B34-172-27).

Luciferase assays
HCT116 cells were plated in 96-well plates (2.5 × 10 4 cells per well) 24 h prior to DNA transfection with Jet-PEI® (275 ng of total DNA). The pGL3-POLK Luc and its truncated mutant pGL3-83 vectors were previously described [29] . The pGL3-29 vector was constructed for this work and co-transfected in HCT116 cells. The pRL-CMV-renilla plasmid (Ozyme®) was used to normalize transfection efficiency. Firefly luciferase values were measured and normalized by the Renilla luciferase activity. Values were expressed as the mean ratio of luciferase activities.

Cell proliferation and cytotoxicity assays
Cells were seeded in quadruplicate at a density of 2 × 10 3 cells per well. At the indicated time, 0.5 mg/mL of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (Sigma-Aldrich®, St Louis, MO, USA) was added and incubated at 37 °C for 4 h. Formazan crystals were solubilized in DMSO and absorbance read at 560 nm on a spectrophotometer. The results were normalized to the cell density at Day 1. For cytotoxicity assays, cells were seeded in quadruplicate in a 96-well plate (2.5 × 10 3 cells per well) and exposed the day after to increasing concentrations of cytotoxic drugs orto vehicle alone. The cells were exposed to the drug during six days and cell proliferation was quantified each day using MTT assay. The results were normalized to the mean optical density of the control for each day. MMS and cisplatin were obtained from Sigma-Aldrich®.

Real-time quantitative PCR
Total RNA was extracted from cells using Quick-RNA kit (Zymo Research) according to the manufacturer's instructions. Total RNA (1 µg) was subjected to reverse-transcription using qScript cDNA SuperMix (QuantaBio, VWR). Real-time quantitative PCR (RT-qPCR) were performed with the Roche LightCycler® 480 instrument and the PerfeCTa SYBR Green FastMix (QuantaBio, VWR) and were carried out in a final volume of 10 μL using 0.25 µL of each primer (25 μM), 5 μL of the supplied enzyme mix, 2.5 μL of H 2 O, and 2 μL of the template diluted at 1:10 [see Supplementary Table 1 for primer sequences]. After pre-incubation at 95 °C, runs corresponded to 35 cycles of 15 s each at 95 °C, 5 s at 60 °C, and 15 s at 72 °C. Melting curves of the PCR products were analyzed using LightCycler® software to exclude amplification of unspecific products. The results were normalized to different housekeeping gene transcripts (mouse RS9 or human 28S) [30] .

Immunoblotting
RIPA solution was used to extract whole-cell proteins. Cell extracts were analyzed after migration of 30 µg protein extract by Western blotting using a primary polyclonal antibody against Polκ (1/1000, Abcam ab57070). Protein quantifications were normalized with the β-actin signal (1/1000, Millipore).

DNA microarray analysis
Published DNA microarray datasets, GSE39582 [31] and GSE42284 [32] , were reanalyzed for RIP140 and POLK expression using the Cancertool database [33] . A transcriptomic dataset from the TCGA-COAD RNA-seq data was also used [4] . Another published DNA microarray study obtained on a cohort encompassing 396 colon tumor samples with MSS and MSI CRC [34] was also reanalyzed for RIP140 and POLK mRNA expression. Statistical significance was assessed using a Spearman correlation analysis. Correlation between RIP140 and POLK gene expression was also studied using another cohort [35] . RNA sequencing data from the TCGA [4] were reanalyzed using Cox proportional hazard regression [36] . RNAseq data obtained from CRC samples from the TCGA dataset were analyzed as described previously [37] . The Kaplan-Meier method was used to estimate overall survival (OS) calculated from the diagnosis until death. Patients lost to follow-up were censored at the time of last contact.

Statistical analysis
All experiments were conducted independently at least three times. The results were expressed as the mean ± standard deviation (S.D.). Statistical comparisons were performed with Mann-Whitney or Spearman tests. For the Cancertool database analysis, a Pearson's test was performed for the correlation analyses. A probability level (P-value) of 0.05 was chosen for statistical significance.

RIP140 regulates POLK gene expression in mouse models
To explore the role of RIP140 on the regulation of the POLK polymerase gene expression, we first used transgenic mice in which the Rip140 gene was either knocked out (RIPKO mice) or overexpressed (RIPTg mice) [24] . As shown in Figure 1A, a significant decrease in the levels of Polκ mRNA was observed in the intestinal epithelium of RIPKO mice, whereas an increase was noted in RIPTg mice as compared to wildtype animals (WT). These regulations appeared specific since the expression of other TLS polymerase genes from the Y subfamily such as PolI did not vary between the different genotypes [ Figure 1A and data not shown]. The steady-state levels of Polκ mRNA were also significantly reduced in immortalized MEFs derived from the RIPKO mice as compared to cells isolated from WT animals [ Figure 1B]. Altogether, these results demonstrate that RIP140 positively controls the expression of the Polκ gene in mouse cells and tissues.

RIP140 regulates POLK gene expression in CRC cells
To check whether this regulation can be observed in human cells, we then analyzed the effect of RIP140 on POLK gene expression in human CRC cell lines. We confirmed the increased expression of the POLK gene at the mRNA levels in HCT116 cells either stably overexpressing RIP140 [ Figure 1C] or transiently transfected with a RIP140 expression vector [ Figure 1D]. We also unveiled that silencing the expression of the RIP140 gene in HCT116, RKO or HCT29 CRC cells significantly affected POLK mRNA abundance [ Figure 1E and F and Supplementary Figure 1] as well as the Polκ protein levels in RKO cells [ Figure 1G].

Transcriptional regulation of POLK gene transcription in CRC cells
To decipher the mechanisms underlying the positive regulation of POLK gene expression by RIP140, we set up transient transfection experiments of HCT116 cells using the POLK Luc reporter construct encompassing the proximal promoter region of the POLK gene [ Figure 2A]. As observed in Figure 2B, RIP140 significantly increased, in a dose-dependent manner, the luciferase activity driven by the POLK gene promoter, thus supporting a positive transcriptional regulation by RIP140. The same effects were observed in other CRC cells including RKO, SW480, and SW620 cells [ Figure 2C].

Mechanism of the transcriptional regulation by RIP140
Although RIP140 was first identified as a transcriptional repressor, we and others reported positive regulation of gene expression (for a review, see Ref. [38] ). In particular, we described a transcriptional activation of gene expression by RIP140 through Sp1-mediated mechanisms [39] . Since the POLK gene promoter exhibits Sp1 binding sites, we tested if the deletion of these sites affected the transcriptional response to RIP140 ectopic expression. As shown in Figure 2D, the induction of luciferase activity by RIP140 was significantly reduced when we used a short reporter construct (POLK 29 Luc), suggesting that Quantifications are expressed in AU after normalization to β-actin, used as a control of protein migration. A Mann-Whitney test was used for statistical analysis (ns = not significant, **P < 0.01, ***P < 0.001).
the regulation of POLK gene expression by RIP140 might be, at least in part, mediated by Sp1.
Interestingly, a close correlation between elevated POLK expression and p53 inactivation was previously reported in lung cancer tissues [40] . This effect is likely due to the direct role of p53 on the POLK gene since p53 strongly inhibits the POLK promoter activity in lung cancer cells, while this activity is much higher in p53 -/-MEF than in p53 +/and p53 +/+ MEFs [40] . Therefore, we checked whether the positive regulation of POLK gene expression by RIP140 was dependent on p53. By using HCT116 cells expressing or not the TP53 gene, we found that, indeed, the regulation of the POLK gene by RIP140 was abolished in HCT116 p53 -/cells when monitored at the mRNA level [ Figure 2E] or using a luciferase reporter assay [ Figure 2F], supporting that p53 is involved in the RIP140-mediated regulation of POLK levels.

Correlation of gene expression in human CRC biopsies
To validate in human CRC biopsies the expression data obtained in mouse tissues and in CRC cells in vitro, we used the Cancertool database [33] to reanalyze the published Affymetrix DNA microarray data from two cohorts, namely GSE39582 containing 585 tumors [31] and GSE42284 with 188 samples [32] . As shown in Figure 3A and B, and in perfect agreement with the data from mice, the results clearly show a significant positive correlation of RIP140 mRNA levels with those of POLK in CRC biopsies. We also reanalyzed https://dx.doi.org/10.20517/cdr.2021.133  [31] . The correlation coefficient and the p-value (Pearson's test) are indicated. (B) Same as in (A) with the cohort GSE42284 [32] . (C) Correlations between RIP140 and POLK gene expression performed with TCGA RNA-Seq cohort (n = 415) [4] showing the correlations found in the different CRC molecular subtypes described in this cohort. (D) Correlation between POLK and RIP140 gene expression in a cohort comparing the expression of "DNA replication" genes in CRC and in the adjacent normal mucosa (n = 26) [35] . another transcriptomic dataset from the TCGA-COAD RNA-seq data [4] obtained on 415 samples, which confirmed a significant correlation between RIP140 and POLK mRNA levels (r = 0.65; P < 2.2 × 10 -16 ) [ Figure 3C]. This dataset was further reanalyzed, taking into account the different molecular subtypes [31] . This allowed us to show a significant correlation of RIP140 gene expression with that of POLK in the six different subgroups, in particular in the C2 group which corresponds to the MMR deficient subgroup. Finally, we confirmed these data in another cohort comparing the expression of "DNA replication" genes in CRC and in the adjacent normal mucosa [35] . We confirmed that RIP140 gene expression strongly decreased in the tumor (data not shown) and that there was a strong correlation in the expression ratio (normal vs. tumoral) of the two genes (r = 0.68; P = 0.0001) [ Figure 3D].

Effect of the RIP MSI mutation on the regulation of POLK gene expression
We recently demonstrated that RIP140 plays an important role in controlling microsatellite instability through the regulation of MSH2 and MSH6 genes [26] . To compare the correlation between RIP140 and POLK gene expression in microsatellite stable (MSS) and instable (MSI) CRC tumors, we reanalyzed a transcriptomic dataset from 396 human CRC with both types of tumors [34] . As shown in Figure 4A, we observed a very significant positive correlation between RIP140 and POLK mRNA levels in the whole cohort (r = 0.74; P = 5.2 × 10 -69 ), thus confirming the results shown in Figure 3. Interestingly, POLK and RIP140 gene expression were significantly correlated in both MSS and MSI tumors [ Figure 4B].
In MSI CRC, we identified a frameshift mutation in the RIP140 coding sequence [26] . This mutation (called RIP MSI ) generated a truncated protein which impaired the biological activity of the RIP140 protein. As shown in Figure 4C, the RIP MSI protein was found to be less efficient than the WT protein to increase POLK gene expression. The same results were obtained on endogenous gene expression in HT29 CRC cells overexpressing wild-type or mutated RIP140 [ Figure 4D]. Interestingly, the RIP MSI mutant exhibited a dominant negative effect since its ectopic expression significantly decreased POLK mRNA accumulation only in WT MEFs, which express normal levels of RIP140, and not in RIPKO MEFs, which no longer express the Rip140 gene [ Figure 4E].

Functional consequences of the regulation of POLK by RIP140
Since the POLK gene is involved in the cellular response to cytotoxic drugs including MMS and cisplatin through a TLS replication process [41] , we next measured the importance of RIP140 on MMS and cisplatin sensitivity by comparing IC 50 ratios between RIPKO cells and their WT counterparts. As shown in Figure 5A and B, we observed a significant increase of sensitivity to MMS and cisplatin when the RIP140 gene was knocked out (IC 50 ratio WT/RIPKO = 9.7, P < 0.05 for MMS and IC 50 ratio WT/RIPKO = 1.9, P < 0.001 for cisplatin). Altogether, these data suggest that the regulation of POLK gene expression by RIP140 can affect the cellular responses to various alkylating agents, including anticancer drugs.

POLK expression impacts the prognosis value of RIP140 in CRC tumors
Using the Kaplan-Meier Plotter database, we then investigated whether POLK levels contribute to the good prognosis value of RIP140 that we previously reported in CRC patients [24] . We reanalyzed RNAseq data obtained from 452 patients with colon adenocarcinoma that we separated into two groups of 226 patients, each with low and high POLK gene expression in the corresponding tumors using the median as a cutoff value. As shown in Figure 6A and B, a statistically significant association of high expression of RIP140 with a decreased risk of death in CRC patients was observed when their tumor exhibited low POLK gene expression [ Figure 6A] but not in tumors with high POLK gene expression [ Figure 6B]. As a control, no differences were observed when the 452-patient cohort was stratified based on the median value of POLI expression [ Figure 6C and D]. Collectively, all the data obtained in this work strongly support a molecular and functional link between RIP140 and POLK in CRC.

DISCUSSION
Colorectal cancer is a frequent neoplasm with high genomic instability including MSI due to defects in the MMR system. We previously described that the transcription factor RIP140 was a key factor in the regulation of intestinal homeostasis and tumorigenesis [24] . We also showed that RIP140 could be involved in microsatellite stability through the regulation of MSH2/MSH6 gene expression [26] .
In the present study, we demonstrated that RIP140 also strongly regulates POLK gene expression, in both mouse and human cells and tissues. Data obtained using luciferase reporter assays demonstrate that this regulation took place at the transcriptional level and implicated the proximal region of the POLK gene.   Several transcription factors including p53 [40] , Sp1, and CREB [29] or AhR [42] have been shown to control the expression of the POLK gene, and some of them have their activity regulated by RIP140 [43] . The data presented herein suggest that p53 is a good candidate to mediate, at least partly, the regulation of POLK gene expression since the positive effect of RIP140 is lost in HCT116 cells no longer expressing the TP53 gene. Moreover, it has been shown that the HDAC inhibitor Trichostatin A was able to induce POLK gene expression [29] . Since RIP140 strongly interacts with HDACs [44] , it is possible that HDAC sequestration out of the POLK gene promoter also participates in the positive effect exerted by RIP140 on POLK gene transcription as already demonstrated for other transcription factors [39] . Further work will be necessary to fully decipher the mechanisms used by RIP140 to transcriptionally control POLK gene expression.
Cell survival relies on a subtle equilibrium between accurate genomic DNA replication and less stringent DNA synthesis often linked to DNA damage. POLK plays an important and critical role in such equilibrium by contributing to several important DNA transaction pathways. These include translesion synthesis, which allows stalled replication forks to bypass a lesion and restart [45] ; the replication checkpoint [46] , a crucial pathway that regulates the replication stress response to a large array of insults that cause replication arrest; and the synthesis of intrinsic non-B microsatellite DNA [7] . Consistent with these multiple and complex roles, both overexpression of POLK (associated with advanced disease stages and shorter survival in patients with glioma [47] and non-small cell lung carcinomas [48] ) and under-expression of POLK (frequently observed in colorectal, lung, stomach, and breast cancers [29,35,49,50] ) have been documented to lead to genetic instability. Indeed, excessive Polκ can interfere with fork progression and trigger a mutator phenotype and chromosome instability, while downregulation of Polκ expression can affect fork progression and induce replicative stress. Since both situations can confer a selective growth advantage during cancer cell evolution, this underlines the importance of tight regulation of POLK gene expression at the transcriptional level for the maintenance of genome integrity. Therefore, our discovery here establishing the role of RIP140 in POLK gene regulation not only may explain why POLK is abnormally expressed in cancer, notably in CRC, but can also clarify why cancers cells respond differentially to anticancer genotoxic drugs that target DNA replication forks. This last point is illustrated in the present work by cell sensitivity to MMS treatment, which can be modulated by RIP140.
In addition, it is worth mentioning that the regulation of POLK gene expression could intensify the phenotypic consequences of the regulation of MutSα gene expression by RIP140 [26] . Indeed, this TLS polymerase displays a high accuracy during dinucleotide microsatellite DNA synthesis and has been proposed to play a role in the maintenance of microsatellite stability [7] . Moreover, Polκ interacts with MSH2 [12] and could therefore participate in the control of MutSα activity. Defining the precise involvement of POLK gene induction in the effect of RIP140 on the maintenance of genome integrity in colon cancer cells (in particular on the stability of microsatellites) and on their resistance to anticancer drugs will obviously require further work. It would also be interesting to decipher whether, as it is suggested for the MSH2 protein [51] , POLK could be involved in the RIP140-mediated transcriptional regulation of gene expression.
Altogether, the present data suggest that POLK-defective cells exhibit an altered DNA replication program, thus explaining the spontaneous genetic alterations observed in POLK-deficient mice [8] . Transcriptional deregulation of POLK gene expression may therefore participate in intestinal tumorigenesis and account, at least in part, for the tumor suppressor role of RIP140 that we previously suggested.