Tumor suppressor p53 regulates insulin receptor (INSR) gene expression via direct binding to the INSR promoter

A significant volume of clinical and epidemiological data provides support to the concept that insulin and the insulin receptor (INSR) have an important role in breast cancer. Tumor suppressor p53 is the most frequently mutated molecule in human cancer. The present study was aimed at evaluating the hypothesis that p53 governs the expression and activation of the INSR gene in breast cancer cells. In addition, the study was designed to investigate the mechanism of action of p53 in the context of INSR gene regulation. The availability of MCF7 breast cancer-derived cell lines with specific disruption of either the insulin-like growth factor-1 receptor (IGF1R) or INSR allowed us to address the impact of the IGF1R and INSR pathways on p53 expression. Data indicate that the INSR gene constitutes a target for p53 action. Wild-type p53 stimulated INSR promoter activity in control cells while disruption of endogenous IGF1R or INSR led to inhibition of promoter activity by p53. Mutant p53 strongly stimulated INSR promoter. Furthermore, p53 directly binds to the INSR promoter in cells with a disrupted IGF1R. Combined, our results identified complex functional and physical interactions between p53 and the INSR pathway. The implications of the p53-INSR interplay in breast cancer needs to be further investigated.


INTRODUCTION
The insulin/insulin-like growth factors (IGFs) create an hormonal network responsible for the regulation of important physiological events throughout life [1][2][3]. The biological information conveyed by this complex system governs multiple metabolic, endocrine, nutritional and growth processes [4][5][6][7]. In addition, this growth factor system plays key developmental roles at each stage of life, from early in utero phases until old age [8,9]. The insulin/ IGF family includes three ligands [insulin, IGF1, IGF2], three cell-surface receptors [insulin receptor (INSR), IGF1 receptor (IGF1R) and IGF2 receptor (IGF2R)], and six IGFbinding proteins (IGFBP1-6) [10][11][12][13]. Moreover, a number of non-classical insulin-like molecules have been identified in recent years, including the insulin receptor-related receptor (IRR), insulin-IGF1 hybrid receptor and a number of IGFBPrelated proteins [14,15]. The physiological roles of these new members, for the most part, are yet to be elucidated.
The classical view that emerged following the cloning and characterization of the INSR and IGF1R genes in the mid-1980s postulated that activation of INSR by insulin leads, predominantly, to metabolic activities [11,[16][17][18][19]. On the other hand, activation of IGF1R by IGF1 or IGF2 (depending on the specific organ involved and developmental stage) leads primarily to growth and differentiation events. Despite the fact that this model was based on vast experimental and clinical evidence, it is clear today that this dogmatic representation of 'insulin/ IGF1 labor division' amounts to an oversimplification of facts [20]. One of the cardinal questions still in need of a biologically plausible rationalization is why the INSR and IGF1R, even though they share the majority of their downstream cytoplasmic targets and signaling pathways, are yet responsible for mediating distinct physiological and pathological activities.
In recent years the literature has provided ample evidence of a cross-talk between the various ligands

Research Paper
Oncotarget 2425 www.oncotarget.com and receptors of the insulin/IGF family [21,22]. Both hormones are capable of activating (phosphorylating) the opposite receptor, leading to atypical mechanistic events. These cross-talk events, which are generally achieved at high doses of the ligand (usually one order of magnitude higher than the concentrations required by the high affinity-binding ligands to activate their cognate receptor) may result in activation of INSR by IGF1 (with ensuing metabolic activities). In contrast, activation of IGF1R by insulin may lead to growth events.
P53 is a transcription factor that usually accumulates in the cell in response to various insults, most notably DNA damage [23,24]. In agreement with its tumor suppressor role, hyperphosphorylated p53 arrests cell cycle progression at the G1 phase, hence enabling damaged DNA to be repaired before the replicative phase [25]. Tumor suppressor p53 is the most frequently mutated molecule in human cancer [26]. In previous studies, we provided evidence that wild-type p53 negatively regulates IGF1R gene expression [27,28]. Co-expression experiments along with in vitro transcription assays demonstrated that the mechanism of action of wild-type p53 involves transcriptional suppression of the IGF1R gene. In contrast, and consistent with their oncogenic role, tumor-derived mutant forms of p53 enhanced IGF1R gene expression [27].
Given the emerging evidence of proliferative and potentially anti-apoptotic actions of INSR, we investigated in the present paper the regulation of the INSR gene promoter by wild-type and mutant p53 in breast cancer cells. Using cells with specific disruption of the INSR (INSR-KD) or IGF1R (IGF1R-KD), we also assessed the effect of each one of these signaling pathways on p53 expression and activity. Our data indicate that: (1) activation of p53 is negatively regulated by IGF1R, as indicated by the augmented phosphorylation of p53 in IGF1R-KD cells; (2) p53 directly binds to the INSR promoter region in cells with a disrupted IGF1R; (3) wildtype p53 represses INSR promoter activity in IGF1R-KD and INSR-KD cells, while enhancing promoter activity in control cells; (4) mutant p53 stimulates INSR promoter activity in breast cancer cells. In conclusion, data is consistent with complex interactions between p53 and the IGF1/insulin signaling pathways in regulation of INSR gene transcription. The impact of these interactions in breast cancer development and the clinical ramifications of these findings merit further investigation.

Effect of IGF1R or INSR abrogation on p53 expression and phosphorylation
An increasing body of experimental and clinical evidence has identified a direct link between obesity, hyperinsulinemia and cancer risk. The role of the INSR in the initiation and progression of cancer, in general, and breast cancer in particular, has been a controversial issue for many years [29,30]. Likewise, the interplay between p53, an important player in cancer etiology, and the insulin pathway has not yet been unequivocally defined. To investigate the functional and physical interactions between INSR and p53 we employed MCF7 breast cancer-derived cells with specific disruptions of the INSR (INSR-KD) or IGF1R (IGF1R-KD) receptors [20]. The characterization of these cells has been recently reported [31]. Briefly, cells with a disrupted INSR or IGF1R did not express the abrogated receptor in neither cytoplasmic nor nuclear fractions. The existence of a compensatory mechanism was suggested by experiments showing enhanced IGF1R levels in both cytoplasmic and nuclear fractions of INSR-KD cells.
To evaluate the impact of insulin and IGF1 signaling on p53 expression and activation status, Western blots were conducted on total, cytoplasmic and nuclear fractions of control, IGF1R-KD and INSR-KD cells. No major differences in total p53 levels were seen between cell lines ( Figure 1A, 1B). However, while the nuclear fractions contain mainly a ~53-kDa band, a larger p53 molecule was detected in the cytoplasmic fractions. The nature of this band was not investigated, although we assume that it corresponds to a modified p53 molecule. Both p53 bands were detected in whole cell extracts. Western blot analysis using an antibody against phospho-p53 detected higher phospho-p53 levels in cytoplasmic and, particularly, nuclear fractions of IGF1R-KD cells. This expression pattern suggests that activation of p53 is negatively regulated by IGF1R. Hence, absence of IGF1R expression in IGF1R-KD cells leads to relaxation of inhibitory control with an ensuing rebound in p53 phosphorylation.
In addition, we assessed the impact of IGF1R or INSR abrogation on p53 sumoylation, an important modification of the p53 molecule. To this end, total, cytoplasmic and nuclear extracts of IGF1R-KD, INSR-KD and control cells were immunoprecipitated with anti-p53, electrophoresed through SDS-PAGE and immunoblotted with anti-Sumo1. Results of IP experiments revealed a minor increase in sumoylated p53 in the cytoplasmic fractions of both disrupted, compared to control, cells ( Figure 1C). These results indicate that IGF1R/INSR status does not have a major role on sumoylation (and, probably, degradation) of p53.

Regulation of the INSR promoter by wild-type and mutant p53
Next, we investigated the regulation of the INSR promoter by wild-type or mutant p53. To this end, IGF1R-KD, INSR-KD and control cells were transfected with an INSR promoter-luciferase reporter construct [pGL3(-877/-2) LUC] along with expression vectors encoding either wild-type or mutant p53 (including mutations at codons www.oncotarget.com 143, 248 or 273). Results of transfection assays indicate that basal INSR promoter activity was ~7-fold higher in IGF1R-KD and INSR-KD cells than in control cells ( Figure 2). These results suggest that abrogation of either one of these receptors leads to relaxation of inhibition of the INSR promoter. In addition, cotransfection of a wild-type p53 expression vector inhibited INSR promoter activity by 57% and 44% in IGF1R-KD and INSR-KD cells, respectively. In contrast, wild-type p53 stimulated INSR promoter activity by 2.6-fold in control cells. These results indicate that intact INSR and/or IGF1R pathways are necessary in order for p53 to stimulate the INSR promoter. Abrogation of either one of these pathways results in an inhibitory effect of p53. To assess whether the changes in INSR promoter detected above are correlated with corresponding changes in INSR protein levels, Western blots were performed using whole cell extracts of wild-type p53-transfected (or control) cells. As shown in Figure

Interactions between INSR/IGF1R and p53 in regulation of the INSR promoter
To investigate the potential functional interaction between p53 and INSR in regulation of the INSR gene, transient co-transfection assays were performed using an INSR promoter-luciferase reporter construct along with a wild-type p53 expression vector (or empty pCMV) and an INSR-A expression vector (or empty pEGFP). Results of cotransfection experiments demonstrate that INSR-A stimulated INSR promoter activity in all three cells lines but with major differences between control and disrupted cells. Thus, whereas exogenous INSR-A enhanced promoter activity by 27-fold in control cells, the stimulatory effects in IGF1R-KD and INSR-KD cells were 8.9-and 5.6-fold, respectively ( Figure 4A). Coexpression of wild-type p53 abolished the stimulatory effect of INSR-A. Again, significant differences were seen between control (40-fold decrease) and KD (3.5to 9.1-fold decreases) cells. Co-IP experiments revealed a physical interaction between INSR and p53 in nuclear fractions of IGF1R-KD cells ( Figure 4A, right inset).
We then assessed the capacity of IGF1R to activate the INSR promoter, and the interactions between IGF1R and p53. Results of cotransfection experiments indicate that the ability of IGF1R to stimulate INSR promoter activity was 40-50% lower than that of INSR-A in all three cell lines ( Figure 4B). Results of co-IP assays in INSR-KD cells identified a protein-protein interaction between IGF1R and p53 in both cytoplasmic and nuclear fractions ( Figure 4B, inset). Taken together, data indicate that both INSR and IGF1R are capable of stimulating INSR promoter activity, although the cognate receptor exhibited a stronger effect. In addition, we identified a physical interaction between p53 and INSR/IGF1R that might explain the ability of p53 to reduce the stimulatory effects of both INSR and IGF1R.

Physical interactions between p53, INSR/IGF1R and the INSR promoter
To assess the potential binding of p53 to the INSR promoter, a genomic fragment extending from nt -540 to -18 (relative to the translation start site) was labeled using a 5′-biotinylated primer. The labeled fragment was bound to streptavidin magnetic beads, incubated with nuclear extracts of IGF1R-KD, INSR-KD or control cells, and eluted with a high-salt buffer. Western blots using a p53 antibody identified a strong binding of p53 to the INSR promoter in IGF1R-KD cells ( Figure 5A). These results suggest that diminished IGF1R levels in these cells might facilitate the direct binding of p53 to the proximal INSR promoter region. In addition, the fact that p53 did not bind at all to the INSR promoter in INSR-KD cells suggests that the presence of the INSR protein is important for p53 binding to the INSR promoter.
A similar approach was employed to investigate the binding of INSR and IGF1R proteins to the INSR promoter sequence. DNA affinity chromatography experiments showed that INSR displays a strong binding to the INSR promoter in IGF1R-KD cells ( Figure 5B) whereas IGF1R exhibits a strong binding in INSR-KD cells ( Figure 5C). Combined, these results suggest that lack of IGF1R in IGF1R-KD cells facilitates INSR binding to the promoter region. On the other hand, lack of INSR in INSR-KD cells facilitates IGF1R binding to the same sequences. Of notice, while the mature IGF1R (~100-kDa band) is mainly involved in INSR promoter binding, the precursor form of INSR (~250-kDa band) is predominantly associated with the INSR promoter region.

Effect of INSR/IGF1R abrogation on the stimulatory effect of Sp1
Transcription factor Sp1 has been identified as a key activator of the INSR gene [32]. To assess the

Effect of INSR/IGF1R abrogation on the effect of p53 on cell proliferation
To investigate the impact of INSR or IGF1R abrogation on the effect of p53 on cell proliferation, IGF1R-KD, INSR-KD and control cells were transfected with a p53 expression vector in serum-containing media, and proliferation was assessed after 72 h using an XTT assay. Results obtained indicate that, under basal conditions, proliferation was enhanced in both KD cells (Figure 7). These results might imply that abrogation of a specific receptor leads to compensatory expression of the opposite receptor, with ensuing increase in proliferation. This effect was particularly evident in INSR-KD cells that express high levels of IGF1R. Expression of p53 diminished proliferation by ~50% in INSR-KD, but not IGF1R or control, cells.  were seen in the portion of cells at G2/M in both KD, in comparison to control, cells (Figure 8). p53 expression led to a 33% decrease in control cells at this stage, while the effect of p53 was attenuated in IGF1R-KD (16% decrease) and INSR-KD (22% decrease) cells. Finally, whereas p53 expression had no effect on the proportion of control cells at the S phase, a marked increase in the portion of cells at this stage was seen in IGF1R-KD (28% increase) and INSR-KD (33% increase) cells. Finally, the negative impact of p53 transfection on INSR-KD cells proliferation, described in the previous section, cannot be explained by the results of cell cycle analysis described here. Differences in the sensitivities and end-points between XTT and cell cycle assays may probably explain this discrepancy.

DISCUSSION
INSR overexpression is regarded as a common feature of many types of cancer [16,33]. In the vast majority of INSR-expressing tumors, the A-isoform (lacking exon 11) predominates over the B-isoform [30]. While the relative contribution of the INSR and IGF1R signaling pathways to the development of a malignant   Oncotarget 2432 www.oncotarget.com phenotype has been a controversial issue for many years, a mounting volume of experimental, clinical and epidemiological data provides strong support to the notion that insulin signaling is central to cancer etiology [34]. Among other lines of evidence, the connection between obesity, hyperinsulinemia and enhanced cancer risk provides a biologically plausible rationale for a key role of insulin and the INSR in breast cancer initiation and/or progression.
Accumulation of mutations constitutes an early event in cellular transformation and may, eventually, lead to the establishment of a cancerous phenotype. P53mediated cell cycle arrest enables damaged DNA to be repaired before the replicative phase of the cell cycle [25,26,35,36]. Alternatively, p53 can elicit an apoptotic program. The present study was aimed at evaluating the hypothesis that tumor suppressor p53 governs the expression and activation of the INSR gene in breast cancer (and, probably, other) cells. In addition, the study was designed to investigate the mechanism of action of p53 and the physical interactions between p53 and the INSR gene. Finally, using breast cancer-derived cell lines with disruption of either the IGF1R or INSR we assessed the specific impact of the IGF1R and INSR pathways on p53 expression and activity.
Cellular fractionation experiments revealed enhanced p53 phosphorylation in nuclear fractions of IGF1R-KD, but not INSR-KD, cells. These results suggest that activation of p53 is predominantly inhibited by IGF1R, while the INSR pathway has a minor effect on p53 phosphorylation. Furthermore, results of transient cotransfection assays indicate that: (1) INSR promoter activity was markedly increased (~7-fold) in cells with disrupted IGF1R or INSR; and (2) wild-type p53 inhibited INSR promoter activity by 57% and 44% in IGF1R-KD and INSR-KD cells, respectively, but not in cells with intact IGF1R/INSR pathways. Taken together, data indicate that the INSR promoter is under constitutive inhibition by both IGF1R and INSR. Consequently, silencing of either one of these receptors leads to relaxation of inhibition. In addition, functional INSR or IGF1R pathways are needed in order for wild-type p53 to stimulate the INSR promoter. Abrogation of either one of these pathways leads to an inhibitory effect of p53.
In terms of the mechanism of action of p53 in regulation of the INSR promoter, DNA affinity chromatography experiments showed a strong binding of p53 to the INSR promoter region in IGF1R-KD cells. This pattern is consistent with the idea that reduced IGF1R concentrations in IGF1R-KD cells might promote plates and, after 24 h, were transfected with p53-WT (or empty pCMV vector). After an additional 24 h, the cells were tripsinized, counted and plated again (in 6-well plates in triplicate, 10 5 cells/well) for 72 h. Cells were then permeabilized with Triton X-100, stained with propidium iodide and analyzed using a FacsCalibur system. * p < 0.01 versus respective pCMV-transfected cells.
Oncotarget 2433 www.oncotarget.com the direct binding of p53 to the INSR promoter region. Likewise, this approach identified a direct binding of INSR and IGF1R proteins to the INSR promoter sequence in IGF1R-KD and INSR-KD cells, respectively. In summary, the ability of p53 to govern INSR gene transcription is dictated by complex physical (protein-protein and protein-DNA) interactions involving, in addition to p53, also the INSR and IGF1R proteins. The ability of both receptors to translocate to the cell nucleus and to display DNA binding capacities has been recently reported [37][38][39][40][41].
We and others have previously reported that p53 interacts with the IGF pathway at a number of levels, including transcriptional regulation of IGF axis components (i. e., IGF1R, IGF2 and IGFBP3 genes) [27,[42][43][44]. The IGF1 and p53 signaling pathways were shown to converge in both cytoplasmic and nuclear compartments in a bidirectional fashion [45]. Hence, whereas p53 governs IGF system expression and activation, IGF1 regulates the p53/p63/p73 system [46]. Gain-of-function mutations of p53 in cancer disrupts its inhibitory activity and generates oncogenic molecules capable of transactivating the IGF1R gene. The capacity of p53 to regulate the INSR promoter was originally published by Webster et al. [47]. The authors reported a correlation between INSR and p53 expression in breast tumors. Moreover, exogenous p53 was able to repress the INSR promoter whereas a dominant-negative p53 (p53MUT248) derepressed the promoter in cells with normal p53. In addition, authors provided evidence that the effect of p53 was mediated by C/EBP and Spl transcription factors. Data presented here demonstrate that the ability of Sp1 to activate the INSR promoter was impaired in IGF1R-KD and INSR-KD cells. The stimulatory effect of Sp1 was reduced by p53 co-expression. The main difference between the Webster et al. paper and the present study is the fact that different, though overlapping, INSR promoter fragments were used. Thus, Webster et al. employed a ~1.5-kb fragment extending from nt -290 to -1823 of the hINSR promoter [47] whereas our construct extended from nt -2 to -877. As described above, we show that wildtype p53 enhanced promoter activity in control, but not IGF1R/INSR disrupted, cells.
Evidence accumulated in recent years indicate that, in addition to its capacity to govern cell cycle progression, p53 activation has also a major impact on metabolic processes, including glucose transport [48] and obesity [49]. p53 was shown to bind to more than 4,000 sites in the human genome and to regulate the expression of several hundred genes [50]. Among these genes, critical regulators of glucose, lipids and amino acids metabolism were identified. Moreover, p53 emerged as an important player in control of oxidative phosphorylation and reactive oxygen species generation. It has been suggested that induction of p53 expression may induce senescence, autophagy and apoptosis, which are tightly linked to the PI3K/Akt/mTOR pathway. Our data, linking p53 expression and action to the INSR axis, may be of major relevance in efforts to understand the role of p53 and insulin in regulation of cancer cell metabolism.
In conclusion, we have presented evidence that the INSR gene constitutes a downstream target for p53 action. Whereas wild-type p53 stimulated INSR promoter activity in control MCF7 cells, disruption of endogenous IGF1R or INSR led to inhibition of promoter activity by wild-type p53. Mutant, oncogenic versions of p53, for the most part, strongly stimulated INSR promoter. In addition, p53 exhibits direct binding to the INSR promoter region in cells with a disrupted IGF1R. Taken together, data presented here identifies complex functional and physical interactions between p53 and the INSR pathway. The clinical implications of this interplay in breast cancer needs to be critically assessed.

PCR and DNA affinity chromatography of the INSR promoter
For DNA affinity chromatography, a 523-bp human proximal INSR promoter fragment extending from nucleotides -540 to -18 was labeled using a 5′-biotinylated antisense primer, as described previously [51]. This fragment includes most of the proximal 5′-flanking region and three initiator sites located 276, 282 and 283 bp upstream of the translation site. Primer sequences were derived from genomic INSR [52] as follows: sense, 5′-GTCTCCTCGGATCAGAGCGC-3′; antisense, 5′-(Biotin)-GAGTCCCTTCCTAGGCCAGATC-3′. PCR was performed using the TermalAce™ DNA Polymerase reagent (InVitrogen, Carlsbad, CA, USA). The biotinylated PCR product was bound to streptavidin magnetic beads (Dynabeads ® M-270 Streptavidin; Dynak Biotech ASA, Oslo, Norway), and incubated with nuclear extracts of IGF1R-KD, INSR-KD or control cells. INSR promoterbinding proteins were eluted with a high salt-containing buffer and analyzed by Western blots as described below.

Protein analysis and immunoblotting
Total cell lysates and cytosolic and nuclear extracts were prepared as described [31]. Samples were electrophoresed through 10% SDS-PAGE, followed by transfer to nitrocellulose membranes. Blots were blocked with 5% skim milk and incubated overnight with antibodies listed below. Antibodies against IGF1R β-subunit (#3027), insulin receptor β-subunit (#3025), phospho-p53 (#9284) and SUMO-1 (#5718) were from Cell Signaling Technology (Danvers, MA, USA). Antibodies against p53 (mixture: DO-1 and 1801) and heat shock cognate HSC70 (#B-6) were from Santa Cruz Biotechnology (Dallas, TX, USA). An antibody against lamin B was from Abcam (Cambridge, UK). Blots were washed and incubated with the appropriate horseradish peroxidase-conjugated secondary antibody. Protein A-peroxidase conjugated was used for p53 detection (Rockland Immunochemicals Inc, Limerick, PA, USA). Proteins were detected using the enhanced chemiluminiscence reaction (Westar Supernova, Cyanagen, Bologna, Italy). HSC70 was used as a loading control and lamin B as a nuclear marker.

Plasmids and transient co-transfections
For transient co-transfection experiments, an INSR promoter-luciferase reporter construct extending from position -877 to -2 of the 5′-untranslated region of the human INSR gene [pGL3 (−877/-2) luciferase (LUC) vector] was employed. The plasmid was kindly provided by Dr. Antonino Belfiore (University of Catania, Italy). Expression vectors encoding wild-type and mutant p53 were provided by Dr. Edward Mercer (Thomas Jefferson University, Philadelphia, PA, USA). A wildtype p53 (p53WT) expression vector was constructed into the pCMV-Neo-Bam vector. p53MUT143 encodes p53 harboring a Val to Ala mutation at position 143. p53MUT248 contains an Arg to Trp mutation at position 248 and p53MUT273 is a mutant p53 in which an Arg residue at position 273 was mutated to His. All three mutants were cloned in the pCMV-Neo-Bam expression vector as described [53]. A GFP-IGF1R expression plasmid was provided by Dr. Rosemary O'Connor (University of Cork, Ireland) [54]. An INSR isoform-A expression vector was a gift from Dr. Antonino Belfiore [55]. For Sp1 expression experiments, the pEGFP-Sp1 vector was employed (Addgene, Watertown, MA, USA). For co-transfection experiments, cells were transfected with 1 µg of the INSR promoter reporter along with 1 µg of wild-type or mutant p53 (or empty pCMV vector).
In certain experiments, cells were co-transfected with 1 µg of expression vectors encoding IGF1R or INSR-A (GFP-IGF1R or GFP-INSR-A) or Sp1 expression vector. Transfections were carried out using the Jet-PEI transfection reagent (Polyplus Transfection Inc, Illkirch, France). Cells were harvested 48 h after transfection, and luciferase activity was measured as described [31]. Promoter activities are expressed as luciferase values normalized to protein concentrations. Bars represent the mean ± SEM of relative luciferase activity from triplicate samples in two separate experiments.

Proliferation assays
Cells were seeded in quadruplicate onto 96well plates (2 × 10 3 cells/well) and, after 24 h, were transfected with p53-WT or empty vector (pCMV). Cell proliferation was determined by an XTT cell proliferation kit (Biological Industries) after 72 h. The colorimetric reaction was measured using an ELISA reader at a wavelength of 450 nm and a reference absorbance of 630 nm in at least three independent assays. Cell proliferation was expressed as a percentage of optical density values obtained in p53-WT-transfected cells relative to empty vector-transfected cells. All experiments were performed in quadruplicate in three independent experiments.

Cell cycle analysis
Cells were seeded in triplicate onto 6-well plates. After 24 h the cells were transfected with p53-WT or empty vector, as described above. After 24 h, the cells were tripsinized, counted and plated again (in 6-well plates in triplicate, 10 5 cells/well) for 72 h. After incubation, the cells were washed with phosphate-buffered saline (PBS), trypsinized, centrifuged and resuspended in PBS. The cells were permeabilized with Triton X-100 before adding propidium iodide. Stained cells were analyzed using a FacsCalibur system (Cytek Development Inc, Fremont, CA, USA).

Statistical analyses
The statistical significance of the differences between groups was assessed by Student's t test (two samples, equal variance). Scanning densitometry analyses were evaluated