The CSB repair factor is overexpressed in cancer cells, increases apoptotic resistance, and promotes tumor growth.

In the present study we show that a number of cancer cell lines from different tissues display dramatically increased expression of the Cockayne Syndrome group B (CSB) protein, a DNA repair factor, that has recently been shown to be involved in cell robustness. Furthermore, we demonstrated that ablation of this protein by antisense technology causes devastating effects on tumor cells through a drastic reduction of cell proliferation and massive induction of apoptosis, while non-transformed cells remain unaffected. Finally, suppression of CSB in cancer cells makes these cells hypersensitive to a variety of commonly used cancer chemotherapeutic agents. Based on these results, we conclude that cancer cells overexpress CSB protein in order to enhance their anti-apoptotic capacity. The fact that CSB suppression specifically affects only cancerous cells, without harming healthy cells, suggests that CSB may be a very attractive target for the development of new anticancer therapies.

Introduction

Resistance to apoptosis is a fundamental characteristic of cancer cells and the primary cause of treatment failure against this devastating disease [1]. Activation of programmed cell death in cancer cells offers novel and potentially useful approaches for improving anticancer therapy and provides alternative tools to conventional chemotherapy. CSB is a SWI/SNF-like DNA-dependent ATPase that can wind DNA and remodel chromatin [2-5]. Mutations in the csb gene give rise to Cockayne syndrome (CS), an autosomal recessive disorder characterized by premature aging and affecting growth, development and maintenance of a wide range of tissues and organs [6,7]. In the context of cell metabolism, CSB plays a number of different functions. This protein participates in the transcription-coupled repair (TCR) sub-pathway of nucleotide excision repair (NER). TCR rapidly removes bulky DNA lesions located on the transcribed strand of active genes [8]. In addition, CSB plays a role during transcription by stimulating all three classes of nuclear RNA polymerases [3,9,10]. Finally, we recently demonstrated that CSB plays a critical role in cell robustness by negatively modulating p53 activity after cellular stress, including DNA damage and hypoxia [11]. CSB performs two main functions by counteracting p53 activity: first, by interacting with p53, CSB releases and redistributes the limiting transcriptional co-factor p300 acetyl-transferase to gene expression programs with opposite purposes (pro-survival pathways) [12]; second, CSB down-regulates the cellular levels of p53, by stimulating its ubiquitination and degradation [13]. Accordingly, the deregulation of p53 and the subsequent enhanced apoptotic response in the absence of the CSB protein gives rise to the pronounced cell fragility observed in CS patients upon exposure to stressors of a broad nature. Of interest-, we have previously shown that CSB also counteracts p53-independent apoptosis [14].

Therefore, it seems that CSB functions as an anti-apoptotic factor that re-equilibrates the physiological response toward cell proliferation and survival rather than cell cycle arrest and cell death upon stress. Based on these findings, we believe that CSB represents a strategic target for anticancer therapy. Our hypothesis suggests that the inhibition or down regulation of CSB in cancer cells may result in the down regulation of pro-survival programs aimed to allow cancer cells to evade apoptosis. In the present study we showed that CSB is overexpressed in a variety of cancer cell lines and tissues. Importantly, the down regulation of CSB in these cancer cells resulted in a marked increase of apoptosis. Furthermore, down regulation of CSB also made these cells hypersensitive to anti-cancer chemotherapeutic drugs.

Materials and methods

Cell lines

Tumor cell lines HeLa, MGH and USB were grown in DMEM containing 10% FCS and antibiotics. Prostate tumor cells (PC3) were cultured in RPMI containing 10% FCS and antibiotics. Normal prostate epithelium cells (RWPE1) were cultured in Keratinocyte medium (Invitrogen), with EGF (5 ng/ml) and BPE (0.05 mg/ml). Normal primary human fibroblasts (C3PV) were cultured in MEM containing 15% fetal bovine serum, essential and non-essential amino acids, vitamins and antibiotics. Breast tumor cells MCF7 were cultured in Eagle's MEM containing 0.6 μg/ml bovine insulin and 10% FBS. Breast tumor cells T47D were cultured in RPMI-1640 Medium containing 0.6 μg/ml bovine insulin and 10% FBS. Non-tumorigenic breast epithelial cell line (MCF10A) was cultured in DMEM containing EGF (20 ng/ml), Cholera toxin (100 ng/ml), hydrocortisone (500 μg/ml), 0.01 mg/ml bovine insulin and 5% horse serum.

Oligonucleotides transfection

The day before transfection, cells (1 × 105 for 6-well dishes and 2 × 104 for 24 well dishes) were plated using medium without antibiotics. Immediately before transfection the medium was replaced with Optimem and oligonucleotides (200 nM final concentration) were delivered using Oligofectamine (Invitrogen, USA) following manufacturer's instructions. At four hours Optimem was replaced with complete medium.

Nine different antisense oligonucleotides targeting CSB mRNA (accession number NM_000124) were used. The control oligonucleotides were the reverse of the antisense sequence. Oligonucleotide sequences are available on request.

Retrotranscription and real-time quantitative PCR

RNA was isolated using the NucleoSpin RNA II kit (Macherey-Nagel). cDNA synthesis was performed using the First Strand cDNA Synthesis kit (Fermentas). Real-time quantitative PCR was carried out with SYBR green master mixture (Promega) using Mx3005P Real-Time PCR system (Agilent). Results were normalized to beta-actin. Primers sequences are available on request. qPCR arrays containing cDNA synthesized from RNA of cancer and normal tissues were bought at Origene (TissueScan™ cDNA Array).

Site direct mutagenesis

Site direct mutagenesis at three independent sites was performed, one site at a time, using QuikChange XL Site-Directed Mutagenesis Kit (Stratagene). As visualized in Fig. 2E (bottom panel) the cDNA coding sequence of the CSB wt expression vector has been changed at nucleotide position 3249 (A to G), 3255 (T to G) and 3262 (A to G).

Fig. 2

(A) qRT-PCR to analyze CSB mRNA expression 12 h after transfection. The results, normalized to b-actin, were averaged from values obtained by performing three different experiments. Values are means ± SD. (B–F) MTT assay: (B) 96 well plate after an MTT assay performed in HeLa cells using either sense, rows (1–3) and antisense, rows (4–6) oligonucleotides. HeLa cells were seeded in 96 well plates in quadruplicate, 18 h later cells were transfected with sense or antisense oligonucleotides (200 nM final concentration). 48 h after the transfection, MTT absorbance was measured as described in Section 2. (C) Graph illustrating MTT absorbance in HeLa cell 48 h later the transfection, cells have been treated with 200 nM oligonucleotides. In (D) HeLa cells have been treated with decreasing concentration of oligonucleotides (combination 7 + 9 + 11) and MTT absorbance was read 48 h later the transfection. (E) CSB relative mRNA expression (12 h after the transfection; left panel) and viability (48 h later the transfection; right panel) in HeLa cells treated with sense or antisense oligonucleotide 11 (200 nM final concentration). Purple columns concern HeLa cells that had been previously transfected with a three points-mutagenized CSB coding cDNA which express a transcript resistant to ASO11 degradation. (F) Graph illustrating MTT absorbance in normal (C3PV, RWPE1 and MCF 10A) and tumoral (MGH, USB, PC3, MCF7 and T47D) cell lines 48 h later the transfection, cells have been treated with 200 nM oligonucleotides of either sense or antisense oligonucleotides (combination 7 + 9 + 11). The results were averaged from values obtained by performing three MTT arrays. Values are means ± SD.

Cell viability assay

Cell viability was evaluated using MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide] cell proliferation assay. Cells were plated in 96-well plates one day before oligonucleotides transfection as described above. MTT was added to each well (0.5 mg/ml) at the indicated times. After incubation for 3 h at 37 °C, the supernatant was replaced with 150 μl of solution (10% SDS, 0.6% acetic acid in DMSO) to dissolve the formazan crystals and produce a purple solution. Optical density measurements were obtained using a scanning spectrophotometer DTX 880 Multimode Detector (Beckman Coulter). The readings were made using a 630 nm (background) and a 570 nm filter. The assays were conducted in quadruplicate for each condition.

Proliferation assay

The day before transfection, 1 × 105 cells were plated in 6-well plates. The cell proliferation assay was performed using 0.05% trypan blue solution to distinguish live and dead cells within a Burker chamber. Three replicate counts were determined by the same operator at each time point (at the moment of transfection and 24, 48 and 72 h later).

Apoptosis analysis

One day before transfection, 1 × 105 cells were plated in 6-well plates. 48 h later the oligonucleotides transfection a combination of fluorescein diacetate (FDA; 15 μg/ml), propidium iodide (PI, 5 μg/ml) and Hoechst (HO, 2 μg/ml), were used to differentiate apoptotic and necrotic cells from viable cells. FDA and HO are vital dyes that stain the cytoplasm and nucleus of the viable cells, respectively. The necrotic and the late stage of apoptotic cells are readily identified by PI staining. Cells in the early phase (viable––HO stained) and late phase (dead––PI stained) of apoptosis displayed the characteristic pattern of chromatin fragmentation. Approximately 2000 randomly chosen cells were microscopically analyzed to determine apoptosis levels.

Results

CSB gene is overexpressed in cancer cells

In order to assess whether CSB is overexpressed in cancer cells, we measured CSB protein levels in a variety of tumors. We first screened human cancer cell lines of various origins: bladder (MGH and USB), cervix (HeLa), prostate (PC3) and breast (T47D and MCF7) for CSB protein expression. These data were compared with those obtained using normal primary fibroblast (C3PV), normal immortalized prostate epithelial cells (RWPE1) and normal non-tumorigenic breast epithelial cells (MCF 10A). As shown in Fig. 1A and B, western blot analysis using total cellular extracts and relative quantification revealed an increase in CSB protein levels in cancer cells (MGH, USB, HeLa, PC3, T47D and MCF7), compared to normal cells (C3PV, RWPE1 and MCF 10A). Among the cancer cell lines that we analyzed, only two breast cancer cell lines (ZR-75-1 and MDA-MB-231) did not display a significant increase in CSB protein expression (data not shown). The similar increase in protein levels, as detected by western blot, for both full-length CSB and the CSB/PGBD3 isoforms, in which the first 5 exons of CSB are alternatively spliced with the PGBD3 transposase [15], suggested a transcriptional mechanism at the basis of their overexpression, since both are under the control of the same promoter. Therefore, we performed a quantification of CSB mRNA using quantitative real-time PCR and confirmed that the elevated protein levels were due to an increase in mRNA; all tumor cell lines analyzed displayed an overall increased CSB mRNA level ranging from 3 to 5 fold (Fig. 1C). We also screened the expression of CSB mRNA in a large variety of cancer tissues. Comparative RT-PCR gene expression analysis of array from a large variety of cancer and normal tissues, such as breast, ovarian, lung and kidney, showed high levels of CSB mRNA expression in several tumor tissues.

Fig. 1

(A) Western blot analysis showing CSB full-length, CSB-PGBD3 and actin protein levels. Actin is used as loading control. (B) Relative quantification of CSB protein expression obtained by normalizing CSB by b-actin protein amount. (C) qRT-PCR analysis of CSB mRNA expression. The results, normalized to b-actin, were averaged from values obtained by performing three different experiments. Values are means ± SD. (D) Tissue Scan Cancer qPCR Arrays (Origene) containing cDNAs from normal and different stage cancer tissues were subjected to qRT-PCR with CSB specific primers. CSB levels were normalized to b-actin. The results were averaged from values obtained by performing three PCR arrays. Values are means ± SD. **p < 0.01, ***p < 0.001. Green: normal tissues. Gray: Cancer tissues. Clinical information including disease stage for each sample are available in supplementary Fig. 1. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

As shown in Fig. 1D, we observed a significant overexpression of CSB mRNA in four of five kidney and lung cancer tissue samples, while only two of five cancer tissue samples from breast and ovarian displayed CSB mRNA overexpression compared to the normal tissue sample. The heterogeneous expression of CSB among the cancer tissue samples that we analyzed is in line with the extended genetic and epigenetic heterogeneity in cancer.

Inhibition of CSB mRNA expression using antisense oligonucleotides strategy

We have shown that CSB protein negatively modulates both p53-dependent and independent apoptosis (14). Therefore we asked if the overexpression of CSB in cancer cells rendered these cells resistant to apoptosis. We decided to suppress CSB expression and test if the absence of CSB would render cells more sensitive to apoptotic stimuli. To inhibit CSB expression we designed antisense oligonucleotides (ASOs) to target and induce the degradation of its messenger RNA (mRNA). Sense oligonucleotides (SOs) for each target sequence were used as control. Fig. 2A shows CSB mRNA expression in HeLa cells, 12 h after transfecting either antisense or sense oligonucleotides (final concentration 200 nM). We selected three ASOs (numbers 7, 9 and 11) that resulted in the most efficient reduction of CSB mRNA expression when used in combination.

CSB suppression reduces viability and proliferation of cancer cells

Next we examined the impact of CSB suppression on cell viability and proliferation in HeLa cells. Either individual or combinations of two or three individual ASOs (7, 9 and 11) were used. Viability at 48 h post transfection was measured by MTT assay. Results in Fig. 2B and C showed a dramatic reduction in viability of HeLa cells when CSB expression was suppressed. In particular the 7 + 9 + 11 combination caused a reduction around 90% of the viability compared to the control samples that were transfected with the corresponding SOs. Transfection of decreasing concentration of oligonucleotides (7 + 9 + 11) showed that reduction of cell viability was proportional with the amount of ASOs used (Fig. 2D). In order to exclude that the reduction of cell viability related to any off-target effects instead of CSB suppression, we performed rescue experiments with a CSB cDNA that was resistant to ASO. We modified the CSB cDNA sequence of a CSB expressing vector (pcDNA-CSB*) to perturb ASO annealing. To minimize the mutagenesis of the cDNA sequence we focused on the ASO11 target sequence, because it has a profound effect on known-down and viability even when transfected alone. As shown in Fig. 2E, re-expression of CSB* rescued cell viability illustrating that CSB protein is fundamental for the viability of cancer cells overexpressing it.

Next, we performed the MTT assay in the other cancer cell lines that we had found to overexpress CSB (MGH, USB, PC3, MCF7 and T47D), substantially validating our findings that CSB suppression dramatically reduced the viability of cancer cell lines characterized by its overexpression. Importantly, the MTT assay performed on normal cells (C3PV, RWPE1 and MCF 10A) showed that knockdown of CSB did not substantially affect the viability of normal cells that were also found not to overexpress CSB (Fig. 2F).

Furthermore, inhibition of proliferation in HeLa cells, as a function of CSB suppression, was measured by directly counting viable cells at different times. In particular, cell counts were performed at the time of transfection and 24, 48 and 72 h after the oligonucleotides transfection (Fig. 3A). Inhibition of CSB protein expression, using ASO 7 + 9 + 11, completely arrested cell proliferation of HeLa cells throughout the entire time course, while the respective SOs combination did not have any effect (Fig. 3B). The slight increase in cell proliferation as observed in the untreated sample (k) could be due to a minimal toxicity of the transfection reagent.

Fig. 3

Cell proliferation assay. 1 × 105 cells have been seeded in 6-well plates and transfected 18 h after seeding. Cells have been counted at the time of transfection and 24, 48, 72 h after transfection (A). The analysis was performed in triplicate. Results have been graphed in panels B and C.

Again, using the same assay to analyze cell proliferation, we confirmed that suppression of CSB strongly reduced the levels of proliferation in all the cancer lines characterized by its overexpression but not in the non tumorigenic normal cell lines (Fig. 3C).

To quantitatively assess the rates of apoptosis induced by the suppression of CSB expression, we used a combination of fluorescent dyes to analyze the morphological alterations that cells undergo during apoptosis. Fluorescein di-acetate (FDA) and Hoechst (HO) stain the cytoplasm and the nucleus, respectively, of viable cells. The necrotic and the late stage apoptotic cells were stained by propidium iodide (PI), which requires the loss of the cytoplasm and nuclear membrane integrity. As shown in Fig. 4A, apoptotic cells were not detectable when HeLa cells were transfected with SOs. In contrast, apoptotic cells were readily visible when the CSB expression was suppressed by ASOs in HeLa cells. Further morphological analysis by microscopy (Fig. 4B) showed that approximately 55% of HeLa cells transfected with ASOs had undergone apoptosis, while only 5% of cells transfected with SOs had undergone apoptosis. Again, we extended this analysis to the other cell lines, confirming that CSB suppression dramatically enhanced apoptosis in tumor but not in non-tumoral cell lines (Fig. 4C).

Fig. 4

(A) Images show that suppression of CSB massively induced apoptosis in HeLa cells. Apoptosis was analyzed morphologically by staining cells with fluoresceine diacetate, Hoechst and Propidium iodide. Apoptotic cells displayed the fragmented chromatin characteristic of the apoptotic process. (B–C) Percentage of apoptotic cells as measured 48 h later the transfection of sense and antisense oligonucleotides (combination 7 + 9 + 11) in normal and tumoral cell lines is graphed in panel. (D) MTT cell viability assay for the HeLa cell line treated with either antisense (ASO) or sense (SO) CSB oligonucleotide (increasing concentration of combination 7 + 9 + 11) followed by an IC50 dose of oxaliplatin (0.55 μM), mytomicin-C (0.50 μg/ml) or 5-fluorouracil (3 μg/ml). The half-inhibition dose IC50 was previously calculated. The results are expressed as percent viability relative to the untreated control. (E) MTT cell viability assay for the C3PV normal fibroblasts treated with either antisense (ASO) or sense (SO) CSB oligonucleotide (increasing concentration of combination 7 + 9 + 11) followed by oxaliplatin (0.55 μM).

To evaluate whether down-regulation of CSB sensitizes cancer cells to conventional chemotherapy agents we combined ASOs administration to oxaliplatin, mytomicin-C and 5-fluorouracil treatments. Cellular viability was assayed with increasing concentrations of CSB ASOs (50 and 100 nM) followed by the addition of a fixed concentration of conventional chemotherapeutic agents that we had previously found to correspond to the respective IC50 dose. The combination of ASOs and chemotherapy agents led to a significant reduction of cell viability (additive effect) when compared to the effect displayed when cells were treated either with the chemotherapy agent or ASO (compare sample 2 and 10 with 6, for instance), while SOs did not increase cancer cell death either alone or in combination with chemotherapy agents (Fig. 4D). Normal cells C3PV were also found not to be sensitive to single treatment (ASO or oxaliplatin) or combined treatment, at least at the doses that we used in Hela cells (Fig. 4E).

Discussion

We have shown that a number of cancers display overexpression of CSB and that its suppression induces massive cell death via apoptosis. It is well known that the expression of oncogenes, such as Ras or Myc, elicits the cell suicide machinery, a mechanism today widely accepted as an innate tumor-suppressive program. However, cancer cells evolved a mechanism to buffer this pro-apoptotic response and to bring the cell back under the apoptotic activation threshold [16]. We propose that CSB may play an important role in this context. First, it is reasonable to believe that the function of CSB is to control p53 activity before cancer cells acquire mutations that permanently inactivate this tumor suppressor. Moreover, we have previously shown that CSB also plays a role in the evasion of the p53-independent apoptosis [14]. This suggests that overexpression of CSB may remain fundamental to escape apoptosis even later, when cells have eventually inactivated the p53 pathway.

Furthermore, adaptation to hypoxia is a fundamental step during cancer progression when the tumor mass becomes larger than 2 mm, thus it is no longer supported by the pre-existing vasculature and new vessels need to be generated [17]. We have previously shown that CSB plays a fundamental role in the HIF-1 controlled transcriptional programs [12]. CSB governs the adaptation to hypoxia that also involves the triggering of extravasion and angiogenesis, as well as the potentiation of glucose metabolism, known as the Warburg effect [18,19]. All these processes are fundamental steps in cancer progression; therefore we believe that by boosting the CSB/HIF-1 axis, cancer cells gain a growth advantage.

Moreover, the overexpression of CSB would also confer cancer cells an increased DNA repair capacity, which could explain the emergence of DNA damage (and thus anti-cancer therapy) drug resistance developed by cancer cells. Recent work has demonstrated, for instance, that ablation of transcription coupled repair, increased cisplatin sensitivity of several prostate and colorectal carcinoma cell lines with specific defects in p53 and/or DNA mismatch repair [20]. Accordingly, our data showed that silencing of CSB sensitized cervix cancer cells (HeLa) to DNA damaging agents such as oxaliplatin, Mitomicin-C and 5-fluouracil.

Taken together these results shed light on the oncogenic role of CSB and open the door for a novel direction in cancer therapy. Further studies might be envisaged to understand whether the overexpression of CSB is confined to certain subset of tumors and/or specific combinatorial patterns of gene mutations, in addition to the design and screening of new anticancer drugs that inhibit CSB protein. Additionally, it would be of interest to determine the mechanism by which CSB becomes up regulated in cancer cells. Genomic rearrangements, including aneuploidy and polyploidy, may account for the increased levels of CSB. Skygrams depicting additional copies of 10q11 locus in the tumoral cells used in our studies appear to confirm this hypothesis (http://www.ncbi.nlm.nih.gov/sky). Further studies will be necessary to elucidate the molecular basis for the up regulation of CSB.

Importantly, the inhibition of CSB for cancer therapy also represents a novel and potentially beneficial avenue of research because while most anti-cancer therapies that damage DNA result in the generation of secondary tumors due to the DNA damage inflicted by the therapeutic agent to healthy cells, the suppression of CSB avoids this complication. Alternatively, we can hypothesize that transient suppression of CSB during a chemotherapy regiment might allow a reduced dose of chemotherapy agents in order to be less harmful for healthy cells.

Conflict of interest

None declared.