CBX7 gene expression plays a negative role in adipocyte cell growth and differentiation

CBX7 is a chromobox family protein component of a polycomb group (PcG) multiprotein PRC1-like complex, which is required to maintain the transcriptionally repressive state of many genes, including Hox genes, throughout development. PcG PRC1 complex acts via chromatin remodeling and modification of histones: it mediates monoubiquitination of histone H2A ‘Lys-119’, rendering chromatin heritably changed in its expressibility; promotes histone H3 trimethylation at ‘Lys-9’ (H3K9me3); binds to trimethylated lysine residues in histones, and possibly also other proteins (Schuettengruber et al., 2007; Wu et al., 2009; Scott et al., 2007). CBX7 has a critical role in the progression step of carcinogenesis. Indeed, a drastic downregulation of CBX7 expression has been reported by several groups in malignant neoplasias including thyroid (Pallante et al., 2008), pancreatic (Karamitopoulou et al., 2010), colon (Pallante et al., 2010), lung (Forzati et al., 2012), gastric (Jiang et al., 2010), bladder (Hinz et al., 2008), breast (Mansueto et al., 2010) carcinomas and a reduced growth rate is achieved by the restoration of CBX7 expression in carcinoma cells of different origin (Pallante et al., 2008; Pallante et al., 2010; Mansueto et al., 2010; Li et al., 2010). Moreover, a direct correlation between the loss of CBX7 expression and a more advanced stage of neoplastic disease and a poor survival has been frequently reported (Karamitopoulou et al., 2010; Pallante et al., 2010). The ability of CBX7 to positively or negatively regulate the expression of genes with a critical role in the epithelial–mesenchymal transition such as CDH1 (E-cadherin) (Federico et al., 2009), SPINK1, SPP1 (osteopontin, a gene involved in the metastatic process), STEAP1 and others (Pallante et al., 2014) likely accounts for this correlation.

The tumor suppressor role of CBX7 has been very recently confirmed by the phenotype of Cbx7-null mice (Forzati et al., 2012). Indeed, these mice develop liver and lung adenomas and carcinomas, and accordingly, mouse embryonic fibroblasts (MEFs) derived from Cbx7-knockout (KO) embryos have a higher growth rate and a reduced susceptibility to senescence than their wild-type (WT) counterparts. Interestingly, we have also recently reported that CBX7 interacts with the High Mobility Group A1 (HMGA1) proteins, and that the two proteins have a competitory activity on the CCNE1 promoter: HMGA1 activating and CBX7 inhibiting CCNE1 promoter activity. We also suggest that overexpression of CCNE1, due to the lack of the negative control by CBX7, may be the main mechanism responsible for the neoplastic phenotype of the Cbx7-KO mice (Forzati et al., 2012).

The analysis of Cbx7-KO mice also showed a significant increase in fat tissue. This observation led us to investigate the role of CBX7 in the process of adipogenesis.

The Cbx7-null mice were indistinguishable from their WT and heterozygous littermates until the fifth month of age, after which the body weight of both heterozygous and homozygous Cbx7-null mice (females and males) began to increase, raising up to 36% more than WT at 10 months (Fig. 1A).

Interestingly, Cbx7^−/− mutant mice challenged with a high-fat diet (HFD), shortly after weaning, gained weight more rapidly than WT control mice and earlier than standard diet (SD)-fed Cbx7^−/− mice (Fig. 1B). In fact, although the background strain is susceptible to weight gain starting from 4 weeks of HFD exposure, the body weight of all the HFD-fed Cbx7^−/− mice was already markedly increased even after only two weeks of diet with an average increase of 45%. Analysis of the food intake revealed that the Cbx7^−/− mice had a tendency to consume less food than their WT littermates (data not shown). Autoptic analysis of 10-month-old heterozygous and homozygous null mice, fed with a SD diet, revealed a significant increase in fat tissue, mainly at the abdominal region, which probably accounts for the increased weight gain. To better evaluate differences in body weight, fat and lean body mass (LBM), we examined cohorts of age- and sex-matched mice by dual-energy X-ray absorptiometry (DEXA). DEXA analysis showed significant differences (P<0.05) in the body weight of both female and male 10-month-old Cbx7^+/+, Cbx7^+/− and Cbx7^−/− mice (Fig. 2A,B and data not shown). Similarly, the analysis of fat tissue confirmed an increased fat tissue in 10-month-old Cbx7^−/− and Cbx7^+/− mice than in their corresponding WT controls (Fig. 2C and data not shown). The overall difference in total fat mice between Cbx7^+/+, Cbx7^+/− and Cbx7^−/− correlated with proportional differences in the amounts of intra abdominal fat (Fig. 2D and data not shown).

The LBM analysis was also significantly higher in Cbx7^−/− and Cbx7^+/− mice than in Cbx7^+/+ mice at 10 months of age (Fig. 2E and data not shown).

The results shown above suggested a critical role of Cbx7 expression in adipocyte cell growth and/or differentiation. Therefore, we investigated whether the expression of Cbx7 is regulated during adipocyte differentiation. As a model system we used the 3T3-L1 preadipocyte cells, which have been extensively characterized (Student et al., 1980). These cells undergo adipocyte conversion upon exposure to fetal bovine serum and the differentiating agents dexamethasone, methylisobutylxanthine and insulin (Student et al., 1980). Cells were harvested at various times during differentiation and both RNA and proteins were extracted.

Semiquantitative and quantitative RT-PCR analysis showed that endogenous Cbx7 levels first decreased, between 6 h and 1 day after treatment with differentiating agents, and then returned to levels comparable to pretreatment (time zero) after 48 h (Fig. 3A, upper middle panel). Very similar results were obtained when Cbx7 protein expression was analyzed (Fig. 3A, lower panel).These findings indicate that the expression of Cbx7 is regulated during adipocyte differentiation. Cell differentiation was verified by measuring the expression of various adipogenic markers, including CCAAT/enhancer binding protein beta (Cebpb), peroxisome proliferator-activated receptor gamma (Pparg) and adipocyte protein (Ap2) (data not shown).

To obtain additional evidence that Cbx7 is involved in adipocyte differentiation, we isolated MEFs from Cbx7^+/+, Cbx7^+/− and Cbx7^−/− embryos at 12.5 days post coitum (dpc), induced them to differentiate into adipocytes, as previously described (Hansen et al., 2002), and then stained them with Oil Red O, which is a marker of adipocyte differentiation. As shown in Fig. 3B,C, the number of cell clusters filled with lipid droplets (stained in red) was much higher in the Cbx7^−/− MEFs than in their WT counterparts, whereas Cbx7^+/− MEFs showed an intermediate behaviour. Accordingly, during the differentiating treatment, the expression level of Cebpb and Pparg, two adipocyte markers, differed in Cbx7^−/− and Cbx7^+/− MEFs versus WT controls (Fig. 3D–F). Indeed, the expression level of Cebpb, a transcription factor expressed during the early stages of adipogenesis (Farmer, 2006), increased after 3 h of differentiating treatment and, then, rapidly decreased after 6 h in WT MEFs. Conversely, Cebpb mRNA and proteins levels were slightly lower in Cbx7^−/− and Cbx7^+/− than in Cbx7^+/+ MEFs, but remained stable even after 48 h (Fig. 3D,F).

Similarly, the expression of Pparg, a transcription factor involved in later stages of adipogenesis (Farmer, 2006), was much higher in Cbx7^−/− and Cbx7^+/− than in Cbx7^+/+ MEFs (Fig. 3E).

To better define the role of the Cbx7 protein in adipocyte differentiation, we generated Cbx7^−/− embryonic stem (ES) cells, starting from an already described (Forzati et al., 2012) Cbx7^+/− clone (Fig. 4A) and investigated their ability to undergo adipocyte differentiation (see Materials and Methods) compared to WT controls. After 20 days of treatment, embryoid bodies (EBs) were stained with Oil Red O to evaluate adipocyte differentiation. We observed a drastic increase in the ability to achieve adipogenesis in Cbx7^−/− compared to WT EBs, as determined by a drastic increase in the percentage of Oil Red O-positive Cbx7^−/− EBs and in the number of adipocytes per EB in Cbx7^−/− compared with WT EBs (Fig. 4B,C). Several Cbx7-null clones were analyzed for morphological differentiation and expression of adipocyte-specific markers by qRT-PCR assay. In particular, Cd34, an adipocyte precursor marker (Rodeheffer et al., 2008) and Cebpb, which is expressed early during adipocyte differentiation, were analyzed in undifferentiated ES cells and after 9 days of differentiation, whereas Pparg, which is expressed later, was analyzed after 9 and 15 days of differentiation. Finally, Ap2 and leptin, which are the latest markers of the adipocyte differentiation, were analyzed only after 15 days of differentiation (Fig. 4D). The expression of all these genes is not detectable in WT undifferentiated ES cells, and induced after differentiation. Conversely, Cd34 and Cebpb were already expressed in Cbx7^+/− and Cbx7^−/− untreated cells (Fig. 4D, left panel) and were significantly upregulated in differentiating EBs compared with WT controls (Fig. 4D, middle panel). Similarly, Pparg was strongly upregulated in Cbx7^−/− (and to a lesser extent also in Cbx7^+/−) EBs at 9 days of differentiation compared to WT controls, and similar expression levels were observed in both Cbx7^+/− and Cbx7^−/− EBs also at 15 days of differentiation. Finally, the increase of Ap2 and leptin expression at 15 days of differentiation was significantly higher in Cbx7^−/− EBs compared to controls (Fig. 4D, right panel).

Therefore, these results confirm that the loss of Cbx7 gene expression enhances adipocyte differentiation.

As a further control of the negative role of CBX7 in adipocyte differentiation, we isolated MEFs from two mouse transgenic lines (TG-Cbx7-1 and TG-Cbx7-2), overexpressing the Cbx7 gene under the transcriptional control of a cytomegalovirus promoter (Fig. 5A) (Forzati et al., 2012), and evaluated their ability to differentiate into adipocytes. Oil Red O staining showed that the TG-Cbx7 MEFs have a lower rate of adipocyte differentiation with respect to WT MEFs (Fig. 5B,C). Consistently, significant reduction in Cebpb mRNA and protein levels was observed in MEFs overexpressing Cbx7 after 7 h of the differentiating treatment (Fig. 5D,F). Also Pparg expression, evaluated by qRT-PCR, was much lower in the TG-Cbx7 MEFs at 24 and 34 hours in comparison with the WT controls (Fig. 5E).

To verify whether CBX7 has a role in adipocyte differentiation also in humans, we focused our studies on two human adipose-derived stem (ADS) cells, named ADS1 and ADS3 cells. ADS cells have been demonstrated to be able to differentiate in adipocyte (Folgiero et al., 2010) (Fig. 6A; supplementary material Fig. S1A). We first evaluated the expression of CBX7 during the adipocyte differentiation of these cells. As shown in Fig. 6B (and supplementary material Fig. S1B), CBX7 downregulation is observed in both ADS1 and ADS3 cells after few hours of the differentiating treatment, as it occurs for the murine 3T3L1 cells (Fig. 3A). Then, the CBX7 mRNA and protein levels return to those comparable to pretreatment levels in terminally differentiated adipocytes (Fig. 6B; supplementary material Fig. S1B).

Subsequently, to verify that CBX7 is a negative regulator of differentiation we transfected the adipose-derived stem cells with CBX7 (Fig. 6C; supplementary material Fig. S1C). The overexpression of CBX7 resulted in altered adipocyte differentiation. In fact, the Oil Red O staining showed that the cells overexpressing CBX7 have a lower rate of adipocyte differentiation with respect to control cells (Fig. 6D). Consistenly, after the differentiating treatment the CEBPB and PPARG levels are lower in the cells overexpressing CBX7 compared to the control cells at the same time points (Fig. 6E,F; supplementary material Fig. S1D).

The polycomb group (PcG) proteins form chromatin-modifying complexes that are essential for embryonic development and stem cell renewal and are commonly deregulated in cancer (Bracken et al., 2006). CBX7 is a chromobox family protein and a member of the polycomb repressive complex 1 (PRC1). We have recently generated Cbx7-KO mice validating a tumor suppressor role of the Cbx7 gene since these mice developed liver and lung adenomas and carcinomas that were associated with an increased expression of the cyclin E gene. However, Cbx7-KO mice also showed a significantly higher amount of fat tissue with respect to WT. These findings suggested a negative role of the Cbx7 gene in the control of adipocyte cell proliferation and differentiation prompting us to investigate the role of Cbx7 in adipogenesis. Several experiments reported here seem to confirm this suggestion. Indeed, MEFs and ES cells null for Cbx7 go towards adipocyte differentiation much more efficiently than their WT counterparts, and this effect is Cbx7 dose-dependent since MEFs and ES cells carrying only one impaired Cbx7 allele show an intermediate behaviour. Moreover, MEFs overexpressing Cbx7 showed a lower ability to differentiate to adipocytes with respect to the WT MEFs. Interestingly, we have been able to confirm the negative role of CBX7 in adipogenic differentiation also in human adipose-derived stem cells. Indeed, we show that the differentiation of these cells results in an immediate downregulation of CBX7 expression, whereas CBX7 overexpression has an inhibitory effect on their adipogenic differentiation.

Since we have previously shown that CBX7 interacts with HMGA1 (Forzati et al., 2012) we hypothesize that the negative role of CBX7 in adipocyte differentiation occurs via its antagonistic effect on HMGA1. Therefore, the mechanism would be very similar to that already described for the regulation of the CCNE1 gene, where the HMGA1 positive regulation is antagonized by the opposite effect exerted by CBX7 (Forzati et al., 2012). Consistently, our preliminary results demonstrate that CBX7 inhibits the HMGA1-induced positive regulation of the OB gene, and chromatin immunoprecipitation (ChIP) assays and functional analyses demonstrate that both proteins are able to bind the OB gene promoter with opposite competitive effects on its activity. Interestingly, HMGA1 and CBX7 act in an opposite manner in human malignancies, HMGA1 overexpression and CBX7 downregulation correlating with a poor prognosis (Pallante et al., 2008, Pallante et al., 2010; Karamitopoulou et al., 2010; Fusco and Fedele, 2007; Fedele and Fusco, 2010). Since we have previously shown that HMGA1 is able to negatively regulate CBX7 expression (Mansueto et al., 2010) and that HMGA1 positively regulates miR-181 that has CBX7 as target, we can envisage a HMGA1-CBX7 network that operates in the regulation of tumour progression and adipocyte cell growth and differentiation. Interestingly, a drastic increase in miR-181b expression has been observed in 3T3-L1 cells after induction of adipocyte differentiation (R.F., personal communication).

A role of a member of the polycomb gene protein family in adipogenesis is not surprising since previous studies have shown that Ezh2 and its H3K27 methyltransferase activity are required for adipogenesis. Ezh2 directly represses Wnt1, -6, -10a, and -10b genes in preadipocytes and during adipogenesis. Deletion of Ezh2 eliminates H3K27me3 on Wnt promoters and derepresses Wnt expression, which leads to activation of Wnt/β-catenin signaling and inhibition of adipogenesis. Ectopic expression of the WT Ezh2, but not the enzymatically inactive F667I mutant, prevents the loss of H3K27me3 and the defective adipogenesis in Ezh2^−/− preadipocytes (Wang et al., 2010). Moreover, the involvement of the polycomb family protein in adipogenic differentiation seems also supported by data reporting histone H3 modifications associated with differentiation and long-term culture of mesenchymal adipose stem cells. Indeed, it has been shown that adipogenic differentiation at early passage results in H3K27 demethylation and H3K9 acetylation specifically on adipogenic promoters, whereas at late passages transcriptional upregulation is impaired, H3K27 trimethylation is maintained and H3K9 acetylation is inhibited on the adipogenic promoters implicating a polycomb-mediated epigenetic program regulating adipogenesis (Noer et al., 2009).

In this context, the results reported here would suggest the participation of also the polycomb family member CBX7, in the control of adipocyte cell growth and differentiation likely acting with an opposite role with respect to Ezh2 (Noer et al., 2009).

The generation and genotyping of the Cbx7^+/−, Cbx7^−/− and transgenic (TG-Cbx7) mice has been already described (Forzati et al., 2012). All mice were maintained under standardized nonbarrier conditions in the Laboratory Animal Facility of Istituto dei Tumori di Napoli (Naples, Italy) and all studies were conducted in accordance with Italian regulations for experimentations on animals.

Primary MEFs from Cbx7^+/+, Cbx7^+/−, Cbx7^−/− and TG-Cbx7 mice, were obtained from 12.5 dpc embryos following standard procedures, as previously described (Forzati et al., 2012). They were grown in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies, Grand Island, NY) containing 10% fetal bovine serum (FBS) (Hyclone, South Logan, Utah), 1% glutamine (Life Technologies), 1% penicillin/streptomycin and 1% gentamicin (Life Technologies) (Forzati et al., 2012).

The mouse NIH 3T3-L1 cells used in this study were generously donated by Dr. E. Santos (National Cancer Institute, NIH, Bethesda MD). Cells were grown in DMEM supplemented with 10% calf serum (CS) (Life Technologies).

ES cells were grown on a layer of mitomycin C-inactivated primary fibroblasts in DMEM, 15% FBS (Euroclone, Pero (MI) Italy), 2 mM glutamine (Life Technologies), 1 mM sodium pyruvate (Life Technologies), 1× non-essential amino acids (Life Technologies), 50 ug/ml penicillin/streptomycin (Life Technologies), 1000 units/ml leukemia inhibitory factor (LIF, Chemicon), and 0.1 mM β-mercaptoethanol. Medium was changed daily, and ES cells were split every 2–3 days (Dani et al., 1997). ES cells homozygous for the Cbx7-null allele (Cbx7 double-KO) were obtained by selecting Cbx7^+/− ES cells, previously described (Forzati et al., 2012) with high geneticin concentration (Life Technologies) (12 mg/ml). Among 20 clones obtained and screened by Southern blot as previously described (Forzati et al., 2012), two resulted Cbx7 double-KO.

Human embryonic kidney HEK 293 cells were grown in DMEM containing 10% FBS, 1% glutamine and 1% penicillin/streptomycin.

The ADS cells were grown in DMEM/F12 supplemented with 20% FBS (Hyclone), 1% antibiotics and 1% fungizone (Life Technologies) (Folgiero e al., 2010).

Adipocyte differentiation was induced in Cbx7^+/+, Cbx7^+/−, Cbx7^−/− and TG-Cbx7 MEFs, as well as in 3T3-L1 pre-adypocytic cells. Induction of adipocyte differentiation was performed essentially as described elsewhere (Student et al., 1980). Briefly, cells were grown in their medium until confluence. Two days later, the medium was supplemented with 0.5 mM 1-methyl-3-isobutylxanthine, 10⁻⁶ M dexamethasone and 10 mg/ml insulin for 48 h. Cells were cultured for a further 6 days more in the same medium but without dexamethasone and methylisobutylxanthine. After 20 days of differentiation treatment, the culture medium was removed, and cells were washed twice with phosphate buffer saline (PBS) and fixed with 4% formalin/PBS. Cells were then washed three times with PBS, incubated in 60% isopropyl alcohol for 10 min and stained with 1.8% Oil Red O (Sigma St Louis, MO) in 60% isopropyl alcohol for 10 min. After three more washings with deionized water, cells were scored for adipocyte differentiation and photographed.

ES cells were induced to differentiate as described (Dani et al., 1997). Briefly, hanging drops containing 10³ cells in 20 µl culture medium were maintained for 2 days under the lid of bacteriological dishes filled with PBS. The EBs formed were then transferred into bacteriological plates and maintained for 3 days in suspension in cultivation medium supplemented either with 0.1% DMSO or with all-trans-retinoic acid (ATRA) (10⁻⁸ µM). Medium was changed every day. EBs were maintained 2 days more in suspension in cultivation medium and then were allowed to settle onto gelatin-coated plates in the presence of differentiation medium. This medium consists of cultivation medium supplemented with 85 nM insulin, 2 nM triiodothyronine, and 10% selected fetal calf serum (Life Technologies). Medium was changed every 2 days (Dani et al., 1997). EBs were harvested for RNA extraction after 0, 9, and 15 days of differentiation. Alternatively, after 16 days, EBs were stained with Oil Red O to assess adipocyte differentiation. In order to exactly quantify the Oil Red O staining, samples were de-stained with 100% isopropanol for 10 minutes. The absorbance of the eluates was measured at 500 nm in microplate reader (LX 800, Universal Microplate Reader, BioTek, Winooski, VT). 100% isopropanol was used as blanc.

Human ADS cells were induced to differentiate as described (Folgiero et al., 2010). Briefly, cells were grown in their medium until confluence. Then, 0.5 µM 1-methyl-3-isobutylxanthine, 1 µM dexamethasone, 10 µg/ml insulin, and 100 µM indomethacin were added to the culture medium. The Oil Red O staining was performed as described above.

Total RNA was extracted using TRI-reagent solution (Invitrogen, Carlsbad, CA,) according to the manufacturer's protocol. cDNA was synthesized from total RNA using random hexamers (100 mM) and MuLV reverse transcriptase (Perkin Elmer, Santa Clara, CA). Semi-quantitative PCRs were run with a 25 µl volume for 25 cycles (30 sec at 94°C, 30 sec at 55°C, and 30 sec at 72°C) using the Gene Amp PCR System 9700 (Applied Biosystems, Foster City, CA). RNA samples that had not been reverse transcribed before PCR served as negative control. Quantitative RT-PCR was performed with the SYBR Green PCR Master Mix (Applied Biosystems) under the following conditions: 10 min at 95°C followed by 40 cycles (15 sec at 95°C and 1 min at 60°C). Subsequently, a dissociation curve was run to verify amplification specificity. Each reaction was performed in triplicate. We used the 2^−ΔCT method to calculate the relative expression levels (Livak and Schmittgen, 2001). The following primer sequences were used to amplify the indicated mouse genes: Cbx7: 5′-ttgtcatggcctacgagga-3′ (forward) and 5′-tgggtttcggacctctctt-3′ (reverse); Pparg: 5′-tgctgttatgggtgaaactctg-3′ (forward) and 5′-ctgtgtcaaccatggtaatttctt-3′ (reverse); Cebpb: 5′-tctactacgagcccgactgc-3′ (forward) and 5′-ggtaggggctgaagtcgat-3′ (reverse); G6pd: 5′-gaaagcagagtgagcccttc-3′ (forward) and 5′-cataggaattacgggcaaaga-3′ (reverse) and the human ones CBX7: 5′-cgagtatctggtgaagtggaaa-3′ (forward) and 5′-gggggtccaagatgtgct-3′ (reverse); PPARG: 5′-aaggccattttctcaaacga-3′(forward) and 5′-tcaaaggagtgggagtggtc3′ (reverse); CEBPB: 5′-gcaagagccgcgacaag-3′ (forward) and 5′-ggctcgggcagctgctt-3′ (reverse); G6PD: 5′-gatctaccgcatcgaccact-3′ (forward) and 5′-agatcctgttggcaaatctca-3′ (reverse).

The mice were generally kept and bred with ad libitum access to water and pelleted standard food (Mucedola, Milano, Italy). The body weight was measured monthly, beginning at 4 weeks of age, for ten months. Body length was measured by manual immobilization, extension and measurement of the nasal–anal length. Analysis of body fat and LBM in Cbx7^−/−, Cbx7^+/−, and WT controls was performed by DEXA using the Lunar PIXImus Mouse Densitometer (Wipro GE Healthcare, Madison, WI) (Nagy and Clair, 2000). Before analysis, mice were anesthesized with isoflurane (isoflurane 1.5% plus 2 L/min oxygen).

Soon after weaning, cohorts of 8 Cbx7^+/+ and Cbx7^−/− mice were provided with either a low-fat diet (LFD) containing 10% fat or a high-fat diet (HFD) containing 45% fat (Mucedola) for up to 45 days. Body weight and food intake were monitored every 15 days.

Expression vectors coding for CBX7 (Federico et al., 2009), or the empty vectors were transfected into human ADS cells using Neon Electroporation System (Invitrogen) according to manufacturer's instructions, and neomycin resistant mass cell populations were obtained.

Protein extraction and Western blotting were carried out as reported elsewhere (Forzati et al., 2012). The antibodies used were: anti-GAPDH (sc-32233, Santa Cruz, Biotechnology, Inc., Santa Cruz, CA), anti-CEBPB (sc-150, Santa Cruz), anti-PPARG (2443-Cell Signaling Technology, Inc., Danvers, MA)), anti-CBX7 (sc-481, Santa Cruz).

We used the Student's t-test for intergroup comparisons. The Kruskall–Wallis test, and Dunn and Bonferroni post hoc analyses were used to compare DEXA results. The statistical analyses were performed using GraphPad Prism v.6.0 (La Jolla, CA, USA). A P value<0.05 was considered statistically significant.

This work was supported by grants from Associazione Italiana per la Ricerca sul Cancro AIRC (IG11477), the Ministero dell'Università e della Ricerca Scientifica e Tecnologica–MIUR (PRIN2010), Filas Lazio and AIRC 5×1000 to R.F.