[Targeting the spliceosome: A new therapeutic strategy to counteract chemotherapy resistance in lung cancer?] / Cibler le splicéosome : une nouvelle stratégie thérapeutique pour contrecarrer la résistance à la chimiothérapie dans les cancers du poumon ?

Lung cancer is the first cancer-related cause of death worldwide. This is in partially due to therapeutic resistance, which occurs in around 70% of patients, especially those receiving platinum salts, the gold-standard chemotherapy. The massive deregulation of alternative transcript splicing processes observed in many cancers has led to the development of a new class of pharmacological agents aimed at inhibiting the activity of the splicing machinery (spliceosome). The molecular mechanisms by which these inhibitors act remain largely unknown, as do the benefits of using them in combination with other therapies. In this context, our work is focused on an inhibitor of the SRPK1 kinase, a major regulator of the spliceosome.

Nuclear trafficking of EGFR by Vps34 represses Arf expression to promote lung tumor cell survival.

Epidermal growth factor receptor (EGFR) is a cell surface receptor that has an essential role in cell proliferation and survival, and overexpression of EGFR is a common feature of human cancers. In Non-small-cell lung cancer (NSCLC), activating mutations of EGFR have also been described. We recently showed that mutant EGFR-L858R inhibits the expression of the p14ARF tumor-suppressor protein to promote cell survival. In this study, we defined the molecular bases by which EGFR controls Arf expression. Using various lung tumor models, we showed that EGF stimulation inhibits Arf transcription by a mechanism involving the nuclear transport and recruitment of EGFR to the Arf promoter. We unraveled the vesicular trafficking protein Vps34 as a mediator of EGFR nuclear trafficking and showed that its neutralization prevents the accumulation of EGFR to the Arf promoter in response to ligand activation. Finally, in lung tumor cells that carry mutant EGFR-L858R, we demonstrated that inhibition of Vps34 using small interfering RNA restrains nuclear EGFR location and restores Arf expression leading to apoptosis. These findings identify the Arf tumor suppressor as a new transcriptional target of nuclear EGFR and highlight Vps34 as an important regulator of the nuclear EGFR/Arf survival pathway. As a whole, they provide a mechanistic explanation to the inverse correlation between nuclear expression of EGFR and overall survival in NSCLC patients.

p14(ARF) inhibits the growth of lung adenocarcinoma cells harbouring an EGFR L858R mutation by activating a STAT3-dependent pro-apoptotic signalling pathway.

Epidermal growth factor receptor (EGFR) stimulates proliferative and survival signals. Activating mutations of EGFR are involved in the aetiology and maintenance of the malignant phenotype of lung tumours. We previously described the frequent association of these mutations with the decreased expression of the p14(ARF) tumour suppressor, another common feature of lung cancer. Based on these data, we postulated that p14(ARF) could protect cells against untimely or excessive mitotic signals induced by mutant EGFR. In this study, we demonstrate that p14(ARF) promotes apoptosis in lung tumour cells harbouring the EGFR L858R mutation through the accumulation of phosphorylated signal transducer and activator of transcription 3 (STAT3) on Tyr 705 residue, which leads to Bcl-2 downregulation. Using siRNA against PTP-RT, the phosphatase that specifically targets Tyr 705 residue, we show that accumulation of pSTAT3-Tyr705 promotes EGFR L858R mutant cell death, thereby confirming the existence of a STAT3-dependent pro-apoptotic pathway in these cells. Finally, we show that the expression of the EGFR L858R mutant represses p14(ARF) expression and inhibits STAT3/Bcl-2 signalling. These results identify a novel link between the p14(ARF) and EGFR pathways and suggest that EGFR L858R counteracts the pro-apoptotic function of p14(ARF) by downregulating its expression to promote carcinogenesis.

The transcription factor E2F1 and the SR protein SC35 control the ratio of pro-angiogenic versus antiangiogenic isoforms of vascular endothelial growth factor-A to inhibit neovascularization in vivo.

The transcription factor E2F1 has a crucial role in the control of cell growth and has been shown to regulate neoangiogenesis in a p53-dependent manner through inhibition of activity of the VEGF-A (vascular endothelial growth factor) promoter. Besides being regulated by transcription, VEGF-A is also highly regulated by pre-mRNA alternative splicing, resulting in the expression of several VEGF isoforms with either pro-(VEGF(xxx)) or anti-(VEGF(xxx)b) angiogenic properties. Recently, we identified the SR (Ser-Rich/Arg) protein SC35, a splicing factor, as a new transcriptional target of E2F1. Here, we show that E2F1 downregulates the activity of the VEGF-A promoter in tumour cells independently of p53, leading to a strong decrease in VEGF(xxx) mRNA levels. We further show that, strikingly, E2F1 alters the ratio of pro-VEGF(xxx) versus anti-VEGF(xxx)b angiogenic isoforms, favouring the antiangiogenic isoforms, by a mechanism involving the induction of SC35 expression. Finally, using lung tumour xenografts in nude mice, we provide evidence that E2F1 and SC35 proteins increase the VEGF(165)b/VEGF ratio and decrease tumour neovascularization in vivo. Overall, these findings highlight E2F1 and SC35 as two regulators of the VEGF(xxx)/VEGF(xxx)b angiogenic switch in human cancer cells, a role that could be crucial during tumour progression, as well as in tumour response to antiangiogenic therapies.

E2F1 controls alternative splicing pattern of genes involved in apoptosis through upregulation of the splicing factor SC35.

The transcription factor E2F1 has a key function during S phase progression and apoptosis. It has been well-demonstrated that the apoptotic function of E2F1 involves its ability to transactivate pro-apoptotic target genes. Alternative splicing of pre-mRNAs also has an important function in the regulation of apoptosis. In this study, we identify the splicing factor SC35, a member of the Ser-Rich Arg (SR) proteins family, as a new transcriptional target of E2F1. We demonstrate that E2F1 requires SC35 to switch the alternative splicing profile of various apoptotic genes such as c-flip, caspases-8 and -9 and Bcl-x, towards the expression of pro-apoptotic splice variants. Finally, we provide evidence that E2F1 upregulates SC35 in response to DNA-damaging agents and show that SC35 is required for apoptosis in response to these drugs. Taken together, these results demonstrate that E2F1 controls pre-mRNA processing events to induce apoptosis and identify the SC35 SR protein as a key direct E2F1-target in this setting.

Altered pattern of Cul-1 protein expression and neddylation in human lung tumours: relationships with CAND1 and cyclin E protein levels.

The Cul-1 protein is the scaffold element of SCF complexes that are involved in the proteasomal degradation of numerous proteins regulating cell cycle progression. Owing to this central role in cell growth control, aberrant expression of the components of SCF is thought to play a role during tumourigenesis. Nothing is known about Cul-1 expression in human tumours. In this study, we have analysed its status in a series of 128 human lung carcinomas, comprising 50 non-small cell lung cancers (NSCLCs; 29 squamous cell carcinomas and 21 adenocarcinomas) and 78 neuroendocrine (NE) lung tumours (24 typical and atypical carcinoids, 19 large cell NE carcinomas and 35 small cell lung carcinomas), using immunohistochemistry. We report for the first time an altered pattern of Cul-1 expression in human tumours; indeed, we show that Cul-1 expression is up-regulated in 40% (51/128) of all lung tumours as compared to normal lung tissues, including 34% (17/50), 75% (18/24) and 30% (16/54) of NSCLCs, carcinoids and high grade neuroendocrine lung carcinomas, respectively. Furthermore, we demonstrate that high levels of Cul-1 protein are associated with a low KI67 proliferative index (p = 0.005) and with a decrease in the cyclin E oncoprotein (p = 0.0003), one of the major targets of SCF complexes. These data suggest that up-regulation of Cul-1 could protect cells from hyperproliferative signals through cyclin E down-regulation. Cul-1 is modified by neddylation, a post-translational modification that grafts ubiquitin-like Nedd8/Rub1 residues and controls Cul-1 activity. We also provide evidence that neddylated forms of Cul-1 are specifically expressed in high-grade NE lung tumours and are associated with down-regulation of the Cul-1 inhibitor CAND1 (p = 0.03) and a high level of cyclin E (p = 0.0002). These data support the notion that alterations in the Cul-1 neddylation/deneddylation pathway could contribute to the development of these highly aggressive lung tumours.

E2F-1, Skp2 and cyclin E oncoproteins are upregulated and directly correlated in high-grade neuroendocrine lung tumors.

The transcription factor E2F-1 plays a crucial role in the control of cellular growth. We previously reported its differential pattern of expression in human lung tumors. In this study, we have investigated the relationships linking the status of E2F-1 and a mediator of its proteasomal degradation, the S-phase kinase-associated protein 2 (Skp2) F-box protein. Using immunohistochemistry in a series of 129 lung tumors of all histological types, we demonstrate that Skp2 accumulates preferentially in high-grade neuroendocrine (HGNE) lung carcinomas (86%, P<0.0001), and show that Skp2 overexpression is associated with advanced stages (P<0.0001) and nodal metastasis (P<0.0001) in neuroendocrine (NE) lung tumors. Unexpectedly, we observe that Skp2 and E2F-1 expression directly correlates in NE lung tumors (P<0.0001). Moreover, using cellular models, we identify Skp2 as a new E2F-1 transcriptional target. Furthermore, we also provide evidence that Skp2 interacts physiologically with E2F-1 and stimulates its transcriptional activity toward the cyclin E promoter. Consistently, we demonstrate that cyclin E expression directly correlates with Skp2 (P<0.0001) and E2F-1 (P=0.0001) status in NE lung tumors. Overall, our data provide the first evidence of a direct and functional interconnection between the E2F-1, Skp2 and cyclin E oncoproteins in HGNE lung carcinomas.

Human tumor suppressor p14ARF negatively regulates rRNA transcription and inhibits UBF1 transcription factor phosphorylation.

The nucleolar Arf protein has been shown to regulate cell cycle through both p53-dependent and -independent pathways. In addition to the well-characterized Arf-mdm2-p53 pathway, several partners of Arf have recently been described that could participate in alternative regulation process. Among those is the nucleolar protein B23/NPM, involved in the sequential maturation of rRNA. p19ARF can interact with B23/NPM in high molecular complexes and partially inhibit the cleavage of the 32S rRNA, whereas the human p14ARF protein has been shown to participate in the degradation of NPM/B23 by the proteasome. These data led to define Arf as a negative regulator of ribosomal RNA maturation. Our recent finding that the human p14ARF protein was able to specifically interact with the rRNA promoter in a p53-independent context, led us to analyse in vitro and in vivo the consequences of this interaction. Luciferase assay and pulse-chase experiments demonstrated that the rRNA transcription was strongly reduced upon p14ARF overexpression. Investigations on potential interactions between p14ARF and the transcription machinery proteins demonstrated that the upstream binding factor (UBF), required for the initiation of the transcriptional complex, was a new partner of the p14ARF protein. We next examined the phosphorylation status of UBF as UBF phosphorylation is required to recruit on the promoter factors involved in the transcriptional complex. Upon p14ARF overexpression, UBF was found hypophosphorylated, thus unable to efficiently recruit the transcription complex. Taken together, these data define a new p53-independent pathway that could regulate cell cycle through the negative control of rRNA transcription.

p14ARF promotes RB accumulation through inhibition of its Tip60-dependent acetylation.

p14ARF is a tumour suppressor which plays a critical role in p53-dependent or -independent cell growth control. Several studies have recently provided evidence that p14ARF can also interfere either directly or indirectly with some components of the RB signalling pathway to mediate its antiproliferative activity. The aim of this study was to explore the existence of direct relationships between p14ARF and RB proteins. We show that p14ARF promotes the accumulation of a hypoacetylated RB protein, when it is upregulated in a model of stable-inducible clones or physiologically induced following cell exposure to cytotoxic agents. Looking for the mechanisms involved in this process, we demonstrate that the histone acetyl transferase Tip60 directly interacts with RB and stimulates its degradation by the proteasome through acetylation of its C-terminus. Furthermore, and consistent with p14ARF-induced RB accumulation, we provide evidence that p14ARF prevents Tip60-mediated RB acetylation, therefore precluding its proteasomal degradation. Overall, our results identify a novel mechanism by which p14ARF controls the RB pathway to trigger its antiproliferative function.

E2F1 induces apoptosis and sensitizes human lung adenocarcinoma cells to death-receptor-mediated apoptosis through specific downregulation of c-FLIP(short).

E2F1 is a transcription factor that plays a well-documented role during S phase progression and apoptosis. We had previously postulated that the low level of E2F1 in primary lung adenocarcinoma contributes to their carcinogenesis. Here, we show that E2F1 triggers apoptosis in various lung adenocarcinoma cell lines by a mechanism involving the specific downregulation of the cellular FLICE-inhibitory protein short, leading to caspase-8 activation at the death-inducing signaling complex. Importantly, we also provide evidence that E2F1 sensitizes tumor as well as primary cells to apoptosis mediated by FAS ligand or tumor necrosis factor-related apoptosis-inducing ligand, and enhances the cytotoxic effect of T lymphocytes against tumor cells. Finally, we describe the specific overexpression of c-FLIP(S) in human lung adenocarcinomas with low level of E2F1. Overall, our data identify E2F1 as a critical determinant of the cellular response to death-receptor-mediated apoptosis, and suggest that its downregulation contributes to the immune escape of lung adenocarcinoma tumor cells.

Intercellular trafficking and enhanced in vivo antitumour activity of a non-virally delivered P27-VP22 fusion protein.

VP22, a structural protein from herpes simplex virus type I, exhibits the unique property of intercellular trafficking. This protein is exported from primary expressing cells and subsequently imported into neighbouring cells. This property is conserved when VP22 is genetically fused to a protein, making it a promising tool to enhance the delivery of a gene product. We chose to study the intercellular transport and biological effect of a fusion protein between the putative tumour suppressor gene p27(Kip1) and VP22. We show that in vitro, P27VP22 is able to spread as efficiently as VP22. Functionality of the P27VP22 protein was demonstrated by its ability to inhibit cyclin/CDK2 complexes activity. In proliferation and clonogenicity assays, transfection with the P27VP22 plasmid resulted in a stronger cell growth inhibition when compared to transfection with the p27(Kip1) vector. In vivo, sub cutaneous tumours established in nude mice were injected with naked DNA encoding P27 or P27VP22. Our results show that P27VP22 can spread in vivo and that injections of the P27VP22 plasmid resulted in a significantly greater antitumour activity than injections of the P27 plasmid. This study confirms the usefulness of VP22-mediated delivery and suggests that P27VP22 may have applications in cancer gene therapy.

Expression of p15 and p15.5 products in neuroendocrine lung tumours: relationship with p15(INK4b) methylation status.

The cell cycle inhibitor p15(INK4B) is frequently inactivated by homozygous deletions together with p16(INK4a)/p14(ARF) in many tumour types. Although it is now well established that p16(INK4a) and p14(ARF) act as tumour suppressor genes, the role of p15(INK4b) remains to be well defined. In order to explore the possibility of a selective deregulation of p15(INK4b) in human lung carcinogenesis, we studied p15(INK4b) status in neuroendocrine (NE) lung tumours where homozygous deletions of the p16(INK4a)/p14(ARF) locus are rarely observed. Expressions of p15 and p15.5 protein isoforms were analysed in a series of eight control normal lung, 12 tumour-associated normal lung, five low grade and 15 high grade neuroendocrine (NE) lung tumours and relationship with a specific p15(INK4b) methylation status was studied. Using Western blot analysis, we showed that p15 and p15.5 isoforms displayed a high heterogeneous pattern of expression in both normal and tumour tissues. P15 and p15.5 expressions were correlated in control normal lung (P<0.04) whereas they were not in tumours and associated normal lung. The level of p15.5 was significantly higher in associated normal lung and in tumours (P<0.02 respectively), specially in low grade tumours (P<0.01), than in control normal lung. Furthermore, p15.5 expression was more variable in tumours than in normal lung (P<0.01) and in low grade than in high grade NE lung tumours (P<0.02). Levels of p15 and p15.5 were distinct (up- or downregulated) from those observed in paired normal lung in 4/12 (33%) and 10/12 (83%) tumours respectively. Aberrant methylation at the 5' end of p15(INK4b) gene was observed in 15% of NE lung tumours using PCR-based assay, in a region proximal to the translation start where methylation did not occur in control and associated normal lung. However, no correlation could be assessed with protein status. MSP analysis of CpG islands proximal to the transcription start revealed methylation in all normal and tumour samples. No correlation was found between p15(INK4b) and p16(INK4a) or p14(ARF) status. These data suggest that complex deregulation of p15.5 is implicated in the carcinogenesis of human NE lung tumours independently of p16(INK4a) and p14(ARF) status.

Distinct pattern of E2F1 expression in human lung tumours: E2F1 is upregulated in small cell lung carcinoma.

The transcription factor E2F1 is a key component of cell cycle that acts to transactivate genes required for S phase entry. Thus, it plays an important role in cellular proliferation, oncogenesis and differentiation. In order to investigate its potential implication in human lung carcinogenesis, we studied E2F1 protein expression by Western blotting and immunohistochemistry in a series of 58 human lung tumours of all histological types. We showed that E2F1 product was overexpressed in 92% (24/26) of small cell lung carcinoma (SCLC) and in 50% (5/10) of large cell neuroendocrine carcinoma (LCNEC) whereas it was undetectable in 90% (10/11) of adenocarcinoma and 82% (9/11) of squamous carcinoma when compared to corresponding normal lung. No amplification was found but an increase in E2F1 mRNA expression was detected in 75% (18/24) of SCLC overexpressing E2F1 product. In these tumours and in contrast with NSCLC, upregulation of E2F1 product was associated with its nuclear accumulation and with overexpression of several of its target-genes. Moreover, E2F1 overexpression in NE lung tumours was significantly associated with a high KI67 index (P<0.0001) as well as a Bcl-2:Bax ratio >1 (P<0.001). Overall, these results demonstrate a distinct pattern of E2F1 expression in human lung tumours and suggest that its deregulation could be involved in the carcinogenesis of SCLC.

Human ARF binds E2F1 and inhibits its transcriptional activity.

The INK4a/ARF locus which is frequently inactivated in human tumours encodes two different tumour suppressive proteins, p16(INK4a) and ARF. p16(INK4a) is a major component of the RB pathway. ARF is part of an ARF-mdm2-p53 network that exerts a negative control on hyperproliferative signals emanating from oncogenic stimuli. Among these is the transcription factor E2F1, a final effector of the RB pathway, that induces ARF expression. Recent data suggest that ARF function is not restricted to the p53 pathway. However, ARF target(s) implicated in this p53-independent function remains to be identified. We show that ARF is able to inhibit the proliferation of human cell lines independently of their p53 status. In this context, we demonstrate that ARF interacts physically with E2F1 and inhibits its transcriptional activity. Moreover, we show that mdm2 is required for the modulation of E2F1 activity by ARF. Beside the well-known p53 and mdm2 partners, these results identify E2F1 as a new ARF target. Thus, ARF can be viewed as a dual-acting tumour suppressor protein in both the p53 and RB pathways, further emphasizing its role in tumour surveillance.

Caspase-induced proteolysis of the cyclin-dependent kinase inhibitor p27Kip1 mediates its anti-apoptotic activity.

The caspase-mediated cleavage of a limited number of cellular proteins is a common feature of apoptotic cell death. This cleavage usually inhibits the function of the target protein or generates peptides that actively contribute to the death process. In the present study, we demonstrate that the cyclin-dependent kinase inhibitor p27Kip1 is cleaved by caspases in human leukemic cells exposed to apoptotic stimuli. We have shown recently that p27Kip1 overexpression delayed leukemic cell death in response to cytotoxic drugs. In transient transfection experiments, the p23 and the p15 N-terminal peptides generated by p27Kip1 proteolysis demonstrate an anti-apoptotic effect similar to that induced by the wild-type protein, whereas cleavage-resistant mutants have lost their protective effect. Moreover, stable transfection of a cleavage-resistant mutant of p27Kip1 sensitizes leukemic cells to drug-induced cell death. Altogether, these results indicate that proteolysis of p27Kip1 triggered by caspases mediates the anti-apoptotic activity of the protein.

Selective inhibition of apoptosis by TPA-induced differentiation of U937 leukemic cells.

U937 leukemic cells treated for 24 h with 16 nM 12-O-tetradecanoylphorbol 13-acetate (TPA), that induces their macrophagic terminal differentiation, become resistant to etoposide-induced apoptosis. Exposure of undifferentiated U937 cells to 50 microM etoposide for 6 h, that triggers apoptosis in 80% cells, activates procaspase-2L, -3 and -8, induces the mitochondrial release of cytochrome c and decreases Mcl-1 expression without modifying Bcl-2, Bcl-xL and Bax protein levels. All these events are inhibited in TPA-differentiated U937 cells that are also resistant to vinblastine-induced and Fas-mediated cell death. Interestingly, these cells are not inherently resistant to apoptosis induction. Exposure of TPA-differentiated U937 cells to 0.8 microg/ml cycloheximide for 24 h, that triggers apoptosis in 50% cells, activates procaspase-2L, -3 and -8, induces the mitochondrial release of cytochrome c and decreases Bcl-xL expression without modifying Bcl-2, Mcl-1 and Bax protein levels. All these events are not observed in undifferentiated cells treated in similar conditions. These results indicate that the apoptotic pathway that involves the release of cytochrome c from mitochondria and the cleavage of procaspases remains functional in TPA-differentiated cells.

p27Kip1 induces drug resistance by preventing apoptosis upstream of cytochrome c release and procaspase-3 activation in leukemic cells.

The cyclin-dependent kinase inhibitor p27Kip1 has been implicated as a drug resistance factor in tumor cells grown as spheroids or confluent monolayers. Here, we show that p27Kip1 overexpression also induces resistance to drug-induced apoptosis and cytotoxicity in human leukemic cells growing in suspension. The anti-apoptotic effect of p27Kip1 is not restricted to DNA-damaging agents but extends to the tubulin poison vinblastin, agonistic anti-Fas antibodies and macromolecule synthesis inhibitors. To further identify at which level this protein interferes with the cell death pathway, we investigated its influence on caspase activation and mitochondrial changes. Exposure of mock-transfected U937 cells to 50 microm etoposide activates procaspase-3 and the long isoform of procaspase-2 and induces mitochondrial potential decrease and cytochrome c release from mitochondria to the cytosol. All these events are prevented by p27Kip1 overexpression. p27Kip1 does not modulate Bcl-2, Bcl-X(L), Mcl-1 and Bax protein level in leukemic cells but suppresses Mcl-1 expression decrease observed in mock-transfected U937 cells undergoing etoposide-induced cell death. We conclude that p27Kip1 prevents cell death upstream of the final pathway common to many apoptotic stimuli that involves cytochrome c release from mitochondria and activation of downstream caspases.