Identification of RECQ1-regulated transcriptome uncovers a role of RECQ1 in regulation of cancer cell migration and invasion

Abstract

The RECQ protein family of helicases has critical roles in protecting and stabilizing the genome. Three of the 5 known members of the human RecQ family are genetically linked with cancer susceptibility syndromes, but the association of the most abundant human RecQ homolog, RECQ1, with cellular transformation is yet unclear. RECQ1 is overexpressed in a variety of human cancers, indicating oncogenic functions. Here, we assessed genome-wide changes in gene expression upon knockdown of RECQ1 in HeLa and MDA-MB-231 cells. Pathway analysis suggested that RECQ1 enhances the expression of multiple genes that play key roles in cell migration, invasion, and metastasis, including EZR, ITGA2, ITGA3, ITGB4, SMAD3, and TGFBR2. Consistent with these results, silencing RECQ1 significantly reduced cell migration and invasion. In comparison to genome-wide annotated promoter regions, the promoters of genes downregulated upon RECQ1 silencing were significantly enriched for a potential G4 DNA forming sequence motif. Chromatin immunoprecipitation assays demonstrated binding of RECQ1 to the G4 motifs in the promoters of select genes downregulated upon RECQ1 silencing. In breast cancer patients, the expression of a subset of RECQ1-activated genes positively correlated with RECQ1 expression. Moreover, high RECQ1 expression was associated with poor prognosis in breast cancer. Collectively, our findings identify a novel function of RECQ1 in gene regulation and indicate that RECQ1 contributes to tumor development and progression, in part, by regulating the expression of key genes that promote cancer cell migration, invasion and metastasis.

Introduction

RecQ helicases are vital to maintain genomic stability under replication stress and germline mutations in WRN, BLM, and RECQL4 are associated with cancer predisposition syndromes. RecQ-related diseases are characterized by a variety of clinical features ranging from growth defects (Bloom Syndrome, BS; Rothmund–Thomson Syndrome, RTS) to premature aging (Werner Syndrome, WS), and these patients also have a high risk of developing cancers indicating a critical association with tumor biology.¹

RECQ1, also known as RECQL or RECQL1, is the most abundant of the five human RecQ proteins: RECQ1, WRN, BLM, RECQL4, and RECQL5.² In agreement with an earlier observation that the human RecQ helicases are more abundant in transformed and tumor cells,³ RECQ1 is upregulated in rapidly dividing cells and its expression is higher in cancer cell lines as compared with normal cells.⁴ Overexpression of RECQ1 has been reported in human glioblastoma,⁵ ovarian cancer,⁶ and hypopharyngeal cancer.⁷ Remarkably, RECQ1 silencing was shown to have a cancer cell specific effect in cell culture models and suppressed tumor growth in mouse xenograft models.^8,9 Tumor cell growth was significantly inhibited in vitro by silencing RECQ1 in hypopharyngeal carcinoma cells and the combination treatment of RECQ1 siRNA and cis-platinum (II) diammine dichloride significantly augmented the in vivo anticancer effects of the drug.⁷ A meta-analyses of gene expression pattern included RECQ1 among the common signature genes for cancer as defined by the Gene Ontology Consortium.¹⁰ However, the molecular mechanisms through which RECQ1 might promote tumorigenesis are not fully understood. RECQ1 is a DNA helicase that acts to restore productive DNA replication following stress and hence prevent subsequent genomic instability.^2,11,12 RECQ1 is specifically enriched at the replication origins and common fragile sites which are genomic hotspots for instability and mutagenesis in cancer.¹³ The ability of RECQ1 to facilitate replication stress could be especially important for cancer cells. We hypothesized that RECQ1 plays a role in regulation of gene expression as a component of the constitutive replication stress response in tumor cells that undergo rapid proliferation. Indeed, several proteins involved in DNA repair response are involved in transcription regulation either by binding directly to DNA or through their interaction with specific transcription factors.^14-18 In fact, members of the RecQ family including WRN, BLM, and RECQL5 have been shown to modulate transcription and gene expression.^19-27 RECQ1 was recently reported to be important for accurate transcription directed by the human HomoID box²⁸ but a direct role of RECQ1 on global gene expression has not been examined.

To investigate a potential role of RECQ1 in gene regulation, we have analyzed genome-wide changes in gene expression upon RECQ1 knockdown in HeLa (cervical adenocarcinoma) cells that have been previously used to investigate genome stability functions of RECQ1,^29-32 and highly invasive breast cancer MDA-MB-231 cells which are widely used for molecular and functional analyses of cell migration, invasion and metastasis.³³ We report a subset of genes involved in cell migration and invasion as consistently regulated by RECQ1 and demonstrate that RECQ1 knockdown significantly inhibits cell migration and invasion. Our results provide novel evidence for the involvement of RECQ1 in regulation of gene expression that is likely to be independent of its role in DNA damage repair.

Results

RECQ1 regulates gene expression

To examine a potential role for RECQ1 in regulation of gene expression, we transfected MDA-MB-231 cells with a control (CTL) siRNA or a pool of 4 siRNAs (smart pool, 20 nM) targeting RECQ1. Forty-eight hours (48 h) later, total RNA was isolated and whole cell protein lysates were prepared. We observed significant knockdown of RECQ1 mRNA (>80%) and protein (>75%), as measured by reverse transcription followed by quantitative real-time PCR (RT-qPCR) (Fig. 1A) and immunoblotting (Fig. 1B). To determine whether knockdown of RECQ1 resulted in significant changes in gene expression, we performed cDNA microarrays in triplicate from MDA-MB-231 cells transfected with CTL or RECQ1 siRNAs for 48 h. As shown in the heat map (Fig. 1C), knockdown of RECQ1 significantly altered the expression of a large number of genes. As expected, RECQ1 (designated as RECQL in microarrays) was among the top 10 downregulated genes (Table S1). Moreover, the mRNA levels of other members of the RECQ family including BLM, RECQL4, RECQL5, and WRN didn’t change significantly in the microarrays, demonstrating the specificity of the RECQ1 siRNAs (Table S1).

Figure 1. Depletion of RECQ1 in MDA-MB-231 cells alters gene expression. (A) MDA-MB-231 cells were transfected with a CTL siRNA or RECQ1 siRNAs for 48 h and the extent of RECQ1 silencing was assessed by RT-qPCR normalized to GAPDH. The housekeeping gene UBC was used as negative control. (B) MDA-MB-231 cells were transfected with CTL or RECQ1 siRNAs for 48 h and RECQ1 protein levels were measured by immunoblotting. GAPDH was used as loading control. (C) MDA-MB-231 cells were transfected with CTL or RECQ1 siRNAs for 48 h and cDNA microarrays were performed. Heat map generated from microarrays shows the genes up- or downregulated (≥1.4-fold, adjusted P < 0.05). Upregulated genes are shown in red and downregulated genes are shown in green. (D) Table shows the number of genes up- or downregulated in the microarrays. (E) Table shows fold change in mRNA expression of a subset of RECQ1-regulated genes selected for validation. (F and G) MDA-MB-231 cells were transfected with CTL or RECQ1 siRNAs for 48 h and changes in mRNA expression of a subset of RECQ1-regulated genes was validated by RT-qPCR normalized to GAPDH.

To identify genes differentially expressed upon knockdown of RECQ1 in MDA-MB-231 cells, we used an arbitrary cut-off of 1.42-fold (>30% downregulation). With this cut-off, 60 genes were upregulated and 187 were downregulated (Fig. 1D). We also analyzed the microarray data based on false discovery rate (adjusted P < 0.05) and found 285 upregulated and 502 downregulated genes (Fig. 1D). This analysis indicates that knockdown of RECQ1 in MDA-MB-231 cells results in many more downregulated genes as compared with upregulated genes. To validate the microarray results we selected 3 upregulated (CCNB1, CD55, and FST) and 6 downregulated genes (CDK6, CENPA, EBNA1BP2, PTK2, SSR1, and TGFBI) (Fig. 1E). Consistent with the microarray results, the levels of CCNB1, CD55, and FST mRNAs increased significantly (~1.5–3-fold) whereas CDK6, CENPA, EBNA1BP2, PTK2, SSR1, and TGFBI mRNAs were significantly downregulated (~30%–80%) upon knockdown of RECQ1 (Fig. 1F and G) as measured by RT-qPCR. These results suggest that RECQ1 plays a role in regulation of gene expression and it may primarily act as a positive regulator of gene expression in MDA-MB-231 cells.

Gene expression changes upon RECQ1 knockdown are not cell-type specific

In order to ascertain that changes in gene expression downstream of RECQ1 were not cell-type specific, we decided to look at changes in the transcriptome after knocking down RECQ1 in HeLa cells. HeLa cells were transfected with a CTL siRNA or RECQ1 siRNAs (20 nM) for 48 h and significant knockdown (>80%) of RECQ1 mRNA and protein (>90%) was confirmed by RT-qPCR (Fig. 2A) and immunoblotting (Fig. 2B). We next performed cDNA microarrays in triplicate from CTL or RECQ1 knockdown cells. Similar to our results from MDA-MB-231 cells, knockdown of RECQ1 did not significantly alter the mRNA levels of other members of the RecQ family including BLM, RECQL4, RECQL5 and WRN, further confirming that RECQ1 does not regulate the expression of these RecQ proteins and that the siRNAs specifically targeted RECQ1 mRNA (Table S2). With an arbitrary cut-off of 1.42-fold that was used for analysis in MDA-MB-231 cells, 217 genes were upregulated and 282 were downregulated (Fig. 2C). Using a false discovery rate cut-off (adjusted P < 0.05), 301 genes were upregulated and 276 were downregulated (Fig. 2C). Therefore, in contrast to MDA-MB-231 cells where we observed significantly more downregulated genes as compared with upregulated genes, in HeLa cells, the number of upregulated and downregulated genes was comparable.

Figure 2. Regulation of gene expression by RECQ1 is not cell-type specific. (A) HeLa cells were transfected with CTL or RECQ1 siRNAs for 48 h and RECQ1 silencing was assessed by RT-qPCR normalized to GAPDH. The housekeeping gene UBC was used as negative control. (B) HeLa cells were transfected with CTL or RECQ1 siRNAs for 48 h and RECQ1 protein levels were measured by immunoblotting. GAPDH was used as loading control. (C) HeLa cells were transfected with CTL or RECQ1 siRNAs for 48 h and cDNA microarrays were performed. Table shows the number of genes up- or downregulated in the microarrays (≥1.4-fold, adjusted P < 0.05). (D) Table shows fold change in mRNA expression of a subset of RECQ1-regulated genes that were also regulated by RECQ1 in MDA-MB-231 cells (Fig. 1E). (E and F) HeLa cells were transfected with CTL or RECQ1 siRNAs for 48 h and changes in mRNA expression of a subset of RECQ1-regulated genes was validated by RT-qPCR normalized to GAPDH. Venn diagram shows the overlap between the genes downregulated (G) or upregulated (H) in the microarrays performed after RECQ1 knockdown in MDA-MB-231 (Fig. 1) and HeLa cells (≥1.4-fold, adjusted P < 0.05). (I) Binding of RECQ1 to the predicted G4 motifs in the promoters of select RECQ1-regulated genes was assessed by ChIP assays from MDA-MB-231 cells. A genomic locus from human chromosome 3 and 17 was used as negative controls, negative CTL #1 and #2, respectively.

To determine genes commonly regulated by RECQ1 in MDA-MB-231 and HeLa cells, we first compared the fold change in microarrays for 8 differentially expressed genes (CCNB1, CD55, FST, CDK6, CENPA, EBNA1BP2, SSR1, and TGFBI) between MDA-MB-231 and HeLa cells. Among these 8 genes, 2 out of 3 (CD55 and FST) were significantly upregulated and 5 out of 5 (CDK6, CENPA, EBNA1BP2, SSR1, and TGFBI) were significantly downregulated in both cell lines (Fig. 2D). These results were validated by RT-qPCR from HeLa cells transfected with CTL or RECQ1 siRNAs for 48 h (Fig. 2E and F). To further examine the overlap between RECQ1-regulated genes in the 2 cell lines, we compared the list of differentially expressed genes (adjusted P < 0.05). Ninety one genes were downregulated and 36 were upregulated in both MDA-MB-231 and HeLa cells (Fig. 2G and H). These results indicate that majority of gene expression changes resulting from RECQ1 knockdown could be cell-type specific, however, a subset of these changes are common between HeLa and MDA-MB-231 cells. When we compared the abundance of RECQ1 protein between MDA-MB-231 and HeLa cells, we found that RECQ1 was more abundant in MDA-MB-231 cells (Fig. S1). We next analyzed the microarray data from HeLa and MDA-MB-231 cells transfected with CTL siRNA (Table S3) and looked at the mRNA expression of a subset of RECQ1-regulated genes that we validated by RT-qPCR. Within this subset, we found that some RECQ1-activated genes such as CENPA, PTK2, ITGA2, ITGA3, and TGFBR2 were expressed at higher levels in MDA-MB-231 cells (Table S4). Interestingly, although we found elevated RECQ1 protein levels in MDA-MB-231 cells, RECQ1 mRNA levels did not change significantly in the microarrays between HeLa and MDA-MB-231 cells (Table S3), indicating that post-transcriptional regulation may contribute to the higher RECQ1 protein levels in MDA-MB-231 cells. Other RECQ1 targets such as EBNA1BP2 and CDK6 were not differentially expressed between HeLa and MDA-MB-231 cells suggesting that although MDA-MB-231 cells have higher RECQ1 protein levels as compared with HeLa, the regulation of RECQ1 targets genes in these 2 diverse cell lines may be mediated by complex mechanisms.

Among the DNA substrates that have been shown to be bound by RECQ1, are the four-stranded G4 DNA structures which are formed in vivo by the sequence motif G _{≥ 3}N_xG _{≥ 3}N_xG _{≥ 3}N_xG _{≥ 3} (G4 motifs) and function in essential cellular processes including DNA replication and transcription.^34,35 Since G4 motifs are enriched in the promoter regions of many genes, we searched for the G4 motif in the promoters of RECQ1 regulated genes.³⁶ Our microarray results from MDA-MB-231 cells indicated that more genes were downregulated than upregulated upon RECQ1 knockdown. We therefore searched for G4 motifs in the promoters (−1 kb to +300 bp from transcription start site) of 502 genes downregulated upon knockdown of RECQ1 in MDA-MB-231 cells (Fig. 1D). Of the 454 unique promoter regions, 243 (Tables S5 and S6) had a G4 motif (P = 0.065). Although more than half (53.5%) of the RECQ1-downregulated genes had a G4 motif in their promoter, among the 19 395 promoters of all the protein-coding genes in the human genome, the promoters of 9682 genes (49.9%) had a G4 motif (Tables S5 and S6). Next, we tested the binding of RECQ1 to the G4 motifs in the promoter of 6 genes (Table S7) that were downregulated upon RECQ1 knockdown (CDK6, EBNA1BP2, EZR, ITGA2, ITGA3, and RRAS). Significant enrichment of RECQ1 to the G4 motifs in the promoters of each of these genes was observed (Fig. 2I). Strikingly, the strongest enrichment (7.7-fold in RECQ1 IP as compared with IgG IP) was observed for EBNA1BP2, a gene that was among the top 3 downregulated genes after RECQ1 knockdown in MDA-MB-231 and HeLa cells. Although this result points toward a direct role of RECQ1 in regulating gene expression, more studies are needed to further establish this finding.

RECQ1 regulates a network of genes involved in cell proliferation and cellular movement

Our microarray results indicated that 187 genes were downregulated whereas only 60 genes were upregulated upon RECQ1 knockdown in MDA-MB-231 cells (Fig. 1D), indicating that RECQ1 may function as a positive regulator of gene expression in these cells. Therefore, we decided to focus on the genes downregulated upon RECQ1 knockdown. To identify the biological processes associated with these genes we used Ingenuity Pathways. Consistent with the reported reduced proliferation upon knockdown of RECQ1,^4,29 cellular proliferation was the most significantly over-represented process in the RECQ1-downregulated gene set (Fig. 3A). The top 10 over-represented processes also included cellular movement and cell morphology suggesting that RECQ1 may play a role in regulating cellular movement and cell morphology by enhancing the expression of multiple genes involved in these biological processes. To confirm that this result was not restricted to MDA-MB-231 cells, we used ingenuity pathways for the HeLa downregulated gene set. Indeed, cellular movement, cell morphology and cellular proliferation were among the top 10 over-represented processes in the genes downregulated upon knockdown of RECQ1 in HeLa cells (Fig. S2A). Moreover, we observed significant enrichment of these biological processes for the 91 genes that were downregulated upon knockdown of RECQ1 in both MDA-MB-231 and HeLa cells (Fig. S2B). These analyses suggest that RECQ1 plays a role in enhancing the expression of multiple genes involved in regulation of cellular proliferation, cellular movement and cell morphology, and this effect is not cell-type specific.

Figure 3. RECQ1 increases the expression of a network of genes involved in diverse biological processes including proliferation and cellular movement. (A) Gene Ontology (GO) analysis (Ingenuity pathways) for the 411 genes downregulated upon knockdown of RECQ1 in MDA-MB-231 cells. (B) Direct gene interaction network (Ingenuity pathways) for a subset of the 411 RECQ1-downregulated genes.

The cellular proliferation genes downregulated upon RECQ1 silencing included several key cell proliferation genes such as CDK6 (cyclin dependent kinase 6, a protein important for cell cycle G₁ phase progression and G₁/S transition through regulation of Rb phosphorylation), CENPA (centromere protein A, a protein required for chromosome segregation during cell division), DICER1 (a ribonuclease required for microRNA biogenesis and for cell cycle progression), GAS6 (growth arrest-specific 6, a gamma-carboxyglutamic acid [Gl]-containing protein involved in the stimulation of cell proliferation), SMAD3 (smad family member 3, a transcription factor that regulates the anti-mitogenic activity of TGF-β), IRS1 (insulin receptor substrate 1, a protein that regulates cell size and malignant transformation), BMPR2 (bone morphogenetic protein receptor, type II, a transmembrane serine/threonine kinase involved in TGF-β signaling), TGFBR2 (transforming growth factor, β receptor II, a TGF-β receptor that phosphorylates proteins which then enter the nucleus and regulate the transcription of a subset of genes related to cell proliferation) and YES1 (v-yes-1 Yamaguchi sarcoma viral oncogene homolog 1, a SRC family oncogene that has tyrosine kinase activity and promotes cell cycle at G₁). Our analyses suggest that RECQ1 may promote cell cycle progression by enhancing the expression of these genes.

Knockdown of RECQ1 also resulted in downregulation of mRNAs encoding several proteins that are important for regulation of cellular movement and cell morphology. In particular, multiple components involved in integrin signaling such as ITGA2 (integrin α 2), ITGA3 (integrin α 3), ITGB4 (integrin β 4) were significantly downregulated. Integrins are heterodimers comprised of α and β subunits that are noncovalently associated transmembrane glycoprotein receptors. Integrins regulate cell-cell adhesion, and transduce signals that regulate gene expression and cell growth. RECQ1 may therefore enhance the expression of these integrin components to modulate gene expression, cell growth and cell-cell adhesion. In addition to regulation of integrin signaling, RECQ1 also enhanced the expression of genes critical for regulation of cellular movement including RHOB (ras homolog family member B, an endosomal small GTPase that regulates actin organization and vesicle trafficking), EZR (Ezrin, an oncogene that plays a role in cell surface structure adhesion, migration and organization), VIM (Vimentin, a member of the intermediate filament family that makes up the cytoskeleton and is involved in maintaining cell shape and cell migration), PTK2 (protein tyrosine kinase 2, a protein tyrosine kinase which is found concentrated in the focal adhesions that form between cells and plays a role in cell growth and cell interactions). In addition to these genes, some RECQ1-regulated genes involved in cellular proliferation such as BMPR2, DICER, IRS1, SMAD3, and TGFBR2 also modulate cellular movement. Direct gene network analysis (Ingenuity Pathways) for the mRNAs downregulated upon RECQ1 silencing in MDA-MB-231 cells suggested many of these above-mentioned genes involved in proliferation and cellular movement form a highly connected gene network through protein-protein or protein-DNA interactions (Fig. 3B). These analyses suggest that RECQ1 may enhance the expression of multiple pathways to promote proliferation and cellular movement.

To validate the role of RECQ1 in regulating the expression of genes involved in cellular movement, we transfected MDA-MB-231 and HeLa cells with CTL or RECQ1 siRNAs for 48 h and performed RT-qPCR for 8 genes (EZR, ITGA2, ITGA3, ITGB4, RHOB, RRAS, SMAD3, and TGFBR2) involved in this process (Fig. 4A). With the exception of RHOB mRNA, we consistently observed significant decrease in the expression of 7 out of 8 genes in both MDA-MB-231 and HeLa cells (Fig. 4B and C). To examine the effect on RECQ1 knockdown at the protein levels for a subset of these genes we transfected the cells with CTL or RECQ1 siRNAs for 48 h and performed immunoblotting from whole cell lysates. Downregulation at the protein level was observed for 3 genes (CENPA, ITGB4, and EZR) in MDA-MB-231 cells (Fig. 4D) and 4 genes (CENPA, ITGB4, CDK6, and EZR) in HeLa cells (Fig. 4E). We also validated increased protein expression of CCNB1, a gene that was upregulated upon RECQ1 knockdown in both MDA-MB-231 and HeLa cells (Figs. 1F and 2E). As expected, RECQ1 protein was dramatically downregulated upon knockdown of RECQ1 in both cell lines (Fig. 4D and E). To determine if changes in gene expression upon RECQ1 knockdown were also sustained over a longer period of time, we transduced HeLa and MDA-MB-231 cells with a RECQ1-specific shRNA.¹² Stable knockdown of RECQ1 significantly reduced RECQ1 protein levels (Fig. 4F). We next performed RT-qPCR for 13 genes that were downregulated and 3 that were upregulated when RECQ1 was transiently knocked down in these cell lines by siRNAs. Strikingly, in both cell lines, as compared with control shRNA transduced cells, the RECQ1 shRNA-transduced cells displayed substantially decreased mRNA levels of more than 10 out of 13 genes that were also downregulated by RECQ1 siRNA transfection (Fig. 4G and H). Moreover, the mRNA levels of the 3 genes that were upregulated upon RECQ1 siRNA transfection were upregulated upon stable knockdown of RECQ1 in both cell lines (Fig. 4I and J).

Figure 4. RECQ1 silencing alters the expression of many genes involved in regulation of cell migration, invasion and metastasis. (A) Table shows fold change in expression in the microarrays performed from MDA-MB-231 cells for 8 RECQ1-regulated genes involved in cell migration, invasion and metastasis. Changes in mRNA levels of the 8 RECQ1-regulated genes was assessed by RT-qPCR from MDA-MB-231 (B) or HeLa (C) cells transfected for 48 h with CTL or RECQ1 siRNAs. Changes in protein levels of select RECQ1-regulated genes in MDA-MB-231 (D) or HeLa cells (E) was determined by immunoblotting following transfection with CTL or RECQ1 siRNAs. Changes in protein levels were quantitated by densitometry. (F) RECQ1 protein level was measured by immunoblotting after transducing HeLa or MDA-MB-231 cells with a CTL or RECQ1 shRNA. GAPDH was used as loading control. (G and H) The mRNA expression of RECQ1-activated genes was measured by RT-qPCR after stable knockdown of RECQ1 in MDA-MB-231 (G) and HeLa (H) cells. (I and J) The effect of stable knockdown on RECQ1-repressed genes was assessed by RT-qPCR from MDA-MB-231 (I) and HeLa cells (J).

We next asked the question whether ectopically overexpressed RECQ1 would also regulate gene expression. To test this, we transfected MDA-MB-231 and HeLa cells for 48 h with pCDNA3 or a FLAG-RECQ1 overexpressing plasmid and performed RT-qPCR. Although we observed increased expression of some genes that were downregulated when RECQ1 was knocked down in these cells, for most genes the magnitude of the change was not substantial (Fig. S3A and B). Similar results were also observed in the non-tumorigenic MCF10A cells (Fig. S3C). We reasoned that this result could be because these cells have very high expression of RECQ1 protein. To test this, we overexpressed a RECQ1-shRNA-resistant RECQ1 in HeLa cells that were transduced with a CTL or RECQ1 shRNA (Fig. 5A). We next performed RT-qPCR to determine if rescue of RECQ1 expression in RECQ1-shRNA transduced cells would elevate the expression of RECQ1-activated genes. Indeed, we observed upregulation of 10 out of 10 genes that were downregulated in RECQ1 knockdown cells (Fig. 5B). Moreover, in this RECQ1 rescue experiment we observed downregulation of 3 out of 3 genes that were upregulated in RECQ1 knockdown cells (Fig. 5C). Importantly, we were able to rescue the expression of RECQ1-activated genes and suppress the expression of RECQ1-repressed genes upon re-introduction of RECQ1 in MDA-MB-231 cells (Fig. 5D and E). However, overexpression of RECQ1 in the CTL shRNA transduced cells did not significantly alter the expression of these RECQ-regulated genes (data not shown). These results further suggest a role of RECQ1 in regulating the expression of these genes and also suggest that high RECQ1 protein levels in these 2 cell lines prevents regulation of these genes by ectopically expressed RECQ1. To further establish the role of RECQ1 in gene regulation we decided to determine if mouse RECQ1 may also regulate these genes. To do this, we measured their expression in RECQ1+/+ MEFs (mouse embryonic fibroblasts) and RECQ−/− MEFs.³⁷ As expected, RECQ1 protein was detected only in RECQ1+/+ MEFs (Fig. 5F). We found increased mRNA expression of 3 out of 3 genes that were repressed by RECQ1 in HeLa or MDA-MB-231 cells (Fig. 5G). In addition, in RECQ1−/− MEFs, we observed significantly reduced expression of 7 out of 11 genes that were downregulated when RECQ1 was knocked down in HeLa or MDA-MB-231 cells (Fig. 5H). Taken together, these results confirmed a role of RECQ1 in regulating the expression of multiple genes including specific genes involved in cellular movement.

Figure 5. Re-introducing RECQ1 in RECQ1-depleted cells rescues the expression of RECQ1-regulated genes. (A) RECQ1 protein levels were measured by immunoblotting from HeLa cells transduced with a CTL or RECQ1 shRNA and mock transfected or transfected with a RECQ1-shRNA resistant RECQ1 expressing plasmid (RECQ1 rescue). GAPDH was used as loading control. RECQ1 protein levels were quantitated by densitometry. (B and C) The effect of RECQ1 rescue on the mRNA expression of RECQ1-activated genes (B) and RECQ1-repressed genes (C) was measured by RT-qPCR in HeLa cells. (D and E) The effect of RECQ1 rescue on the mRNA expression of RECQ1-activated genes (D) and RECQ1-repressed genes (E) was measured by RT-qPCR in MDA-MB-231 cells. (F) RECQ1 protein levels were measured in whole cell lysates from RECQ1+/+ and RECQ1−/− MEFs. GAPDH was used as loading control. (G and H) The mRNA expression of RECQ1-repressed genes (G) and RECQ1-activated genes (H) were measured by RT-qPCR from RECQ1+/+ and RECQ1−/− MEFs.

RECQ1 promotes cell migration and invasion

To further investigate the role of RECQ1 in regulation of cellular movement, we assessed the effect of RECQ1 knockdown on cell migration and invasion as previously described.³⁸ Briefly, MDA-MB-231 cells were transfected with CTL or RECQ1 siRNAs for 48 h and the effect on cell migration and invasion was determined by performing transwell migration and matrigel invasion assays. Silencing RECQ1 significantly reduced cell migration (~60%) and invasion (~65%) of MDA-MB-231 cells (Fig. 6A and B). Reduction in invasion upon RECQ1 knockdown was also observed in HeLa cells, further confirming a role of RECQ1 in regulation of invasion (Fig. 6C). To determine if the decrease in migration and invasion was due to impaired proliferation we knocked down RECQ1 for 48 h and performed MTT assays after seeding the cells in 96-well plates. Although we observed some decrease in cell proliferation in RECQ1 knockdown cells starting 5 d post-transfection, knocking down RECQ1 did not significantly alter cell proliferation before 5 d (Fig. 6D). Since the cell migration and invasion assays were performed 4 d post transfection, we conclude that the reduction in cell migration and invasion after RECQ1 knockdown was not due to reduced proliferation. To further confirm this, we transfected MDA-MB-231 cells with CTL or RECQ1 siRNAs and determined the effect on cell cycle progression by flow cytometry (FACS) analysis. Although we observed a slight increase in G₁ compartment after knocking down RECQ1 for 48 h, this effect was reduced at the 72 h time point (Fig. 6E; Fig. S4). We did not observe any apoptosis when RECQ1 was knocked down in MDA-MB-231 cells (Fig. S4). These results therefore suggest that in MDA-MB-231 cells, RECQ1 plays a role in enhancing cell migration and invasion without significantly influencing cell proliferation.

Figure 6. RECQ1 promotes cell migration and invasion in MDA-MB-231 cells. (A) Representative picture of migrating cells at the 48 h time point in transwell migration assays of MDA-MB-231 cells transfected with CTL or RECQ1 siRNAs for 48 h (top). (Bottom) Fold change in cell migration in from 3 independent experiments was quantified by ImageJ. (B and C) Representative picture of invading cells at 48 h time point in matrigel invasion assays of MDA-MB-231 (B) or HeLa (C) cells transfected with CTL or RECQ1 siRNAs for 48 h (top). (Bottom) Fold change in invasion from 3 independent experiments was quantified by ImageJ. (D) MDA-MB-231 cells transfected with CTL or RECQ1 siRNAs for 48 h and seeded into 96-well plates. The effect of RECQ1 silencing on proliferation of MDA-MB-231 cells was assessed using Cell Counting Kit8. (E) MDA-MB-231 cells were transfected with CTL or RECQ1 siRNAs for 48 or 72 h and the effect on cell cycle was determined by FACS analysis after Propidium iodide (PI) staining. (F) RECQ1 expression correlates with prognosis in breast cancer patients. Kaplan-Meier analyses (log rank test) are shown from publicly available microarray data dichotomized at the median value into high and low expressing groups of breast cancer patients.

We next decided to determine if our results from the breast cancer MDA-MB-231 cell line could be extended to breast cancer patient samples. Using the online Kaplan–Meier plotter for 3455 breast cancer patient samples,³⁹ we found that high RECQ1 expression was significantly associated (P = 4.2E-06) with poor survival (Fig. 6F). Remarkably, expression levels of WRN, BLM, and RECQL5 did not correlate with survival in breast cancer whereas high expression of RECQL4 was also associated with poor survival (http://kmplot.com/analysis). Next, we analyzed the TCGA database (595 breast cancer patient samples) for correlation between the expression of RECQ1 mRNA and the mRNAs encoding 10 RECQ1 downregulated genes including CDK6, CENPA, EBNA1BP2, SSR1, TGFBI, TGFBR2, ITGA2, ITGA3, ITGB4, and EZR. TCGA breast cancer gene expression data (level 3) was downloaded from TCGA Data Portal (http://tcga-data.nci.nih.gov/tcga/tcgaHome2.jsp). The mRNA expression of 6 out of 10 genes (CDK6, CENPA, EBNA1BP2, SSR1, TGFBI, and TGFBR2) showed significant positive correlation with RECQ1 mRNA expression (Fig. S5). Among these 6 genes, CDK6, CENPA, and TGFBI showed the strongest positive correlation with RECQ1 mRNA. The expression of the remaining 4 genes did not correlate or showed slight negative correlation with RECQ1 expression. This analysis indicates that RECQ1 and a subset of its regulated genes may play a role in breast cancer. Taken together, our results suggest a role of RECQ1 in enhancing the expression of multiple genes that play important functions in cell proliferation, migration and invasion.

Discussion

Genomic instability contributes significantly to cancer development.⁴⁰ Recent findings indicate that, in both precancerous lesions and cancers, activated oncogenes stall replication forks and cause DNA double strand breakage; the inability of cancer cells to induce an appropriate DNA damage response leads to genetic instability.⁴¹ Suppression of RECQ1 expression in human cells leads to accumulation of DNA damage, increased sensitivity to DNA damaging agents that stall replication forks and induce double strand breaks, and chromosomal instability.^11,13,29,31 Our data indicates for the first time that in addition to the reported genomic instability, knockdown of RECQ1 has significant effect on gene expression.

Significantly reduced cell migration and invasion observed in RECQ1-knockdown breast cancer cells suggests that RECQ1 expression supports tumorigenesis. Functionally, our results are consistent with a recent report where RECQL4 knockdown in metastatic prostate cancer cells drastically reduced their cell invasiveness in vitro and tumorigenicity in vivo, showing that RECQL4 is essential for prostate cancer promotion.⁴² Uncontrollable cell growth and local tissue invasion are the common features of almost all cancer types. DNA repair pathways are the most represented among the overexpressed genes in primary tumors that will metastasize.⁴³ Increased expression of DNA repair factors such as RAD51 and XRCC3 correlate with cellular invasiveness in breast cancer.^44,45 The DNA repair and checkpoint protein RAD9 promotes cell migration and anoikis resistance⁴⁶ and also functions as a transcriptional activator.⁴⁷ Alterations in expression of RecQ homologs have been reported in cancer patients and associated with response to genotoxic chemotherapeutic drugs.^48,49 We noticed that RECQ1 is overexpressed and amplified in many clinical cancer samples (http://www.cbioportal.org/public-portal).¹³ RECQ1 is also differentially expressed in response to anticancer drugs carboplatin⁵⁰ and decitabine.⁵¹ A polymorphism in RECQ1, A159C, has been associated with faster tumor progression and significantly reduced survival of pancreatic adenocarcinoma patients that received gemcitabine and radiotherapy.^52,53 Though it remains to be experimentally tested, the A159C SNP is located in the 3′UTR and, thus, may functionally alter RECQ1 expression and affect clinical outcome. Identification of genes that correlate with RECQ1 expression may also have clinical implications. For instance, positive correlation of RECQ1 expression with CDK6 in breast cancer may be relevant in stratifying patients for chemotherapy as aberrant CDK6 expression is frequently observed in human tumors and inhibitors of CDK6 are in clinical trials.^54,55

RecQ family proteins have been implicated in aspects of transcription control and regulation of gene expression. Yeast RecQ homolog Sgs1 functions in RNA polymerase I transcription of the rRNA (rRNA) genes⁵⁶ subsequent processing of which are fundamental control steps in the synthesis of functional ribosomes. Studies in WS and BS cells suggest that WRN and BLM have similar functions in facilitating RNA polymerase I-mediated rRNA transcription.^19,57 WRN and BLM helicase disrupt RNA-DNA hybrids in vitro and R-loop structures that are formed during RNA polymerase I transcription and known to inhibit progression of the transcription complex.⁵⁷ Similar to WRN and BLM, RECQ1 is enriched in nucleolus where ribosome biogenesis from rRNA transcription and ribosome subunit assembly takes place, however, RECQ1 is unable to unwind RNA-DNA hybrids and its activity on R-loop has not been tested yet.^29,58 We note that EBNA1BP2, one of the most robustly downregulated genes in RECQ1 knockdown in both HeLa and MDA-MB-231 cells, is critical for ribosomal biogenesis.⁵⁹ There is also evidence for the involvement of WRN protein in transcription mediated by RNA polymerase II, the enzyme responsible for mRNA synthesis in eukaryotes.⁶⁰ Transcriptional profile of WS cells closely resemble that of normally aged individuals.²⁶ Remarkably, transient knockdown of WRN in normal human diploid fibroblasts resulted in an expression profile similar to fibroblasts from old donor patients and revealed significant gene expression changes associated with adipogenesis and inflammation that might contribute to the WS pathology.²⁴ Gene expression changes relating to oxidative stress response have been noted in microarray analysis of RTS patient cells bearing mutations in RECLQ4.⁶¹ The mechanism of gene expression changes in WRN or RECQL4 deficiency is not known but a direct mechanistic role of RECQL5 in suppressing RNA polymerase II-dependent transcription has been demonstrated by elegant biochemical analyses.^62,63

Our observation with altered gene expression in RECQ1 knockdown is similar to what has been previously reported for Saccharomyces cerevisiae cells lacking Sgs1 RecQ helicase.⁶⁴ Sgs1Δ mutant cells display downregulation of genes having G4 forming potential within their transcriptional unit⁶⁴ whereas WS and BS cells exhibit increased expression of genes with potential G4 forming sequences.³⁴ G4 motifs are abundant in the human genome and G4 DNA structures within the promoter regions of genes, such as c-MYC⁶⁵ and KRAS,⁶⁶ function as a transcriptional repressor element.³⁴ Moreover, binding of protein factors to G4 DNA near the transcriptional start sites has also been shown to regulate gene expression.^67-69 Recently demonstrated genomewide enrichment of XPB and XPD helicases at G4 motifs suggest that binding to quadruplex structures may recruit XPB and/or XPD to elements that regulate expression of genes in specific signaling and regulatory pathways associated with distinct cancers.⁷⁰ Our ChIP results demonstrating binding of RECQ1 to the potential G4 DNA forming sequences in select differentially expressed genes support a role for RECQ1 in G4 DNA mediated transcription. However, our observation that RECQ1 is also enriched at G4 motif in the promoter region of c-MYC whose expression was not altered by RECQ1 knockdown indicates that the G4 DNA per se may not govern RECQ1 for modulating transcription. Similar to XPB,⁷⁰ RECQ1 binds but does not unwind G4 DNA in vitro.^58,71 Changes in transcriptome upon RECQ1 silencing may be through direct regulation by RECQ1 or through yet unknown indirect mechanism. Our ChIP data indicates binding of RECQ1 at G4 motifs; however whether RECQ1 binds to G4 DNA directly or through other factors remains to be determined. This notion is supported by the fact that at least 2 of the known RECQ1 protein partners, PARP-1 and Ku70, bind to the critical G4 DNA sequence element of KRAS and regulate transcription.⁶⁹ In addition to transcription, G4 DNA is implicated in recruitment of host factors for origin recognition by SV40 and Epstein-Barr virus²¹ and the majority of the human replication origins correspond to G4 motif.³⁴ We note that RECQ1 is constitutively present at human replication origin^13,32 and binds to viral replication origin to regulate Epstein-Barr virus lytic-cycle replication.⁷² Future studies will determine mechanisms of RECQ1 in recognition and resolution of G4 DNA structures critical for essential cellular processes including replication and transcription.

RECQ1 is a very abundant nuclear protein that associates with chromatin. Recent mass spectrometry studies indicate its association with core histones in addition to the DNA repair proteins PARP-1 and Ku70/80 that have also been shown to function as regulators of transcription,^30,31,73 however, it remains to be determined whether these interactions with accessory factors are critical in affecting gene expression changes. PARP-1 regulates gene transcription by modulating structure and function of chromatin and chromatin-associated proteins.⁷⁴ Regulated initiation of gene transcription involves recruitment of DNA topoisomerase IIβ/PARP-1 complex to specific promoter regions leading to the formation of transient double strand DNA breaks and induction of PARP-1 enzymatic activity to enable local changes of chromatin architecture.⁷⁵ Double strand breaks are also shown to induce transcriptional silencing⁷⁶ and the Ku70/80 complex is involved in transcriptional inhibition after double strand break induction.⁷⁷ Although a chromatin remodeling function for RECQ1 has not been shown, functional interactions with PARP-1 and Ku70/80 have been described.^30,31,73

Given the known roles of RECQ1 in cellular response to genotoxic stress, our results support a potential dual function of RECQ1 in DNA damage repair and transcription regulation. It is conceivable that RECQ1 may function as a context dependent regulator of transcription depending on the cell type, DNA damage, and promoter context. Results presented here are from experiments performed in cancer lines following acute depletion of RECQ1 in the absence of exogenous DNA damage. Careful investigation of RECQ1-dependent gene expression changes in normal, non-tumorigenic vs. tumorigenic cells may identify RECQ1 regulated genes critical for cellular transformation. Investigation of RECQ1 interactions with cellular proteins and chromatin will help elucidate the mechanism of observed gene expression changes.

Materials and Methods

Cell culture and transfections

HeLa, MDA-MB-231, MCF10A cells, and RECQ1 knockout and wild-type MEFs³⁷ were maintained in Dulbecco Modified Eagle Medium (DMEM) containing 10% fetal bovine serum (FBS). Stable knockdown of RECQ1 in HeLa and MDA-MB-231 cells was achieved as described.¹² All cells were cultured in a humidified atmosphere containing 5% CO₂ at 37 °C and routinely checked for mycoplasma contamination. On-Target plus SMARTpool small interfering RNAs (siRNAs) against RECQ1 and control siRNAs were purchased from Dharmacon. All siRNA transfections were performed by reverse transfection at a final concentration of 20 nM using Lipofectamine RNAiMAX (Invitrogen) as instructed by the manufacturer. Plasmid transfections for pcDNA3, FLAG-RECQ1,³⁰ or shRNA-resistant RECQ1,¹² were performed using Lipofectamine 2000 (Invitrogen) as instructed by the manufacturer.

RNA isolation and RT-qPCR

Total RNA from cultured cells was isolated by using mini RNeasy isolation kit (Qiagen) as directed by the manufacturer. For quantitative reverse transcription-PCR (RT-qPCR) analysis, 500 µg of total RNA was reverse transcribed using the iScript RT kit (Bio-Rad), and qPCR was performed using SYBR green (Applied Biosystems) as directed by manufacturer.

Immunoblotting

Whole-cell lysates were prepared by using radioimmunoprecipitation assay (RIPA) buffer containing protease inhibitor cocktail (Roche), and protein was quantified using bicinchoninic acid (BCA) protein quantification kit (Thermo Scientific). Ten microgram of total protein per lane was used for immunoblotting. The following primary antibodies were used: anti-CENPA, anti-ITGB4, anti-CDK6, anti-CCNB1, and anti-GAPDH (all Cell Signaling) at 1:1000 dilution; anti-EZR (Abcam) at 1:2000 dilution; anti-RECQ1 (Bethyl lab) at 1:1000 dilution.

Microarray analysis

HeLa or MDA-MB-231 cells were reversed transfected with a CTL siRNA or siRNA against RECQ1 at 20 nM final concentration; 48 h later, total RNA was isolated by using RNeasy minikit (Qiagen). Samples for microarray were labeled using Illumina TotalPrep RNA amplification kit (Applied Biosystems) as recommended by the manufacture. Microarrays were performed using a HumanHT-12 v4 Expression Bead Chip kit (Illumina) and analyzed with the R/Bioconductor packages (Lumi.limma).

Cell proliferation, migration, and invasion assays

MDA-MB-231 cells were reversed transfected with a CTL siRNA or RECQ1 siRNAs. Forty-eight hours (48 h) later, cells were reseeded onto 96-well plates at 1500 cells/well and cell growth was analyzed using Cell counting kit 8 (Fluka). For transwell migration assays, cells were grown in serum-reduced medium (1% FBS) for 5 h, after which 5 × 10⁴ cells/well were plated onto the top chamber with a noncoated membrane (24-well insert, 8-μm pore size; BD Biosciences). For matrigel invasion assay, matrigel (BD Biosciences) was coated onto Transwell insert for 5 h and incubated at 37 °C. A total of 1 × 10⁵ cells/well were plated onto the top chamber in medium containing 1% FBS. In both assays, the lower chamber was filled with 0.7 ml of growth medium containing 10% FBS used as a chemoattractant. After 48 h incubation, cells that did not migrate or invade through the pores were removed by using a cotton swab. Cells on the lower surface of the membrane were fixed and stained with crystal violet, pictures were taken using a digital camera, and migrated or invaded cells were counted and quantified with software ImageJ.

Chromatin Immunoprecipitation (ChIP)

MDA-MB-231 cells grown at a density of 1 × 10⁷ per 15 cm diameter dish were used for ChIP experiments following our previously optimized protocol.¹³ Following phenol/chloroform extraction and ethanol precipitation, sheared DNA fragments served as template in quantitative real-time PCR (qPCR) analysis. qPCR was performed using Taq Universal SYBR Green Supermix (Bio-Rad) with technical triplicates and threshold cycle numbers (C_t) were determined with an iQ5 thermal cycler (Bio-Rad). Fold enrichment of the targeted genomic sequences were calculated over IgG as: fold enrichment = 2^-(Ct_IP^{− Ct}_IgG⁾, where Ct_IP and Ct_IgG are mean threshold cycles of PCR done in triplicates on DNA samples immunoprecipitated with specific antibody and control IgG, respectively. All qPCR reactions were also checked by melt curve analyses and agarose gel electrophoresis to confirm the presence of a single specific product. The sequences of the qPCR primers are listed in Table S6.