PDE4D promotes FAK-mediated cell invasion in BRAF-mutated melanoma.

The cyclic AMP (cAMP) signaling pathway is critical in melanocyte biology for regulating differentiation. It is downregulated by phosphodiesterase (PDE) enzymes, which degrade cAMP itself. In melanoma evidence suggests that inhibition of the cAMP pathway by PDE type 4 (PDE4) favors tumor progression. For example, in melanomas harboring RAS mutations, the overexpression of PDE4 is crucial for MAPK pathway activation and proliferation induced by oncogenic RAS. Here we showed that PDE4D is overexpressed in BRAF-mutated melanoma cell lines, constitutively disrupting the cAMP pathway activation. PDE4D promoted melanoma invasion by interacting with focal adhesion kinase (FAK) through the scaffolding protein RACK1. Inhibition of PDE4 activity or inhibition of PDE4D interaction with FAK reduced invasion. PDE4D expression is increased in patients with advanced melanoma and PDE4D-FAK interaction is detectable in situ in metastatic melanoma. Our study establishes the role of PDE4D in BRAF-mutated melanoma as regulator of cell invasion, and suggests its potential as a target for preventing metastatic dissemination.

Hypoxia and MITF regulate KIT oncogenic properties in melanocytes.

KIT mutations are frequent in acral, mucosal and chronic sun-damage (CSD) melanoma, but little is known about the mechanisms driving the transformation of KIT-mutated melanocytes into melanoma cells. We showed that exposition of melanocytes harboring the (L576P)KIT mutation to a hypoxic environment induced their transformation into malignant cells. Transformed (L576P)KIT melanocytes showed downregulation of MITF expression characteristic of melanoma initiating cells (MICs). In agreement, these cells were able to form spheres in neural crest cell medium and low-adherence conditions, also a characteristic of MICs. Downregulation of MITF by RNA interference induced transformation of KIT-mutated melanocytes in normoxia and acquisition of a MIC phenotype by these cells. Hence, low level of MITF cooperates with oncogenic KIT to transform melanocytes. Activation of the cAMP pathway in transformed (L576P)KIT melanocytes stimulated MITF expression, and reduced cellular proliferation and sphere formation. These findings highlight the essential role of MITF in revealing the oncogenic activity of KIT in melanocytes and suggest that the cAMP pathway is a therapeutic target in KIT-mutated melanoma.

c-Kit mutants require hypoxia-inducible factor 1alpha to transform melanocytes.

Many studies have highlighted the critical role of c-Kit in normal melanocyte development but its role in melanoma development remains unclear. Although c-Kit expression is often lost during melanoma progression, a subset of melanoma has been found to overexpress c-Kit and mutations activating c-Kit have recently been identified in some acral and mucosal melanoma. To address the role of these c-Kit mutants in the transformation of melanocytes, we characterized the physiological responses of melanocytes expressing the most frequent c-Kit mutants found in melanoma (K642E and L576P) and a novel mutant we identified in an acral melanoma. We analysed signaling pathways activated downstream of c-Kit and showed that all three mutants led to a strong activation of the phosphatidyl-inositol-3 kinase (PI3K) pathway but only weak activation of the Ras/Raf/Mek/Erk pathway, which was not sufficient to promote uncontrolled melanocyte proliferation and transformation. However, in hypoxic conditions or coexpressed with a constitutively active form of hypoxia-inducible factor 1alpha (HIF-1alpha), c-Kit mutants activate the Ras/Raf/Mek/Erk pathway, stimulate proliferation and transform melanocytes. Proliferation of melanocytes transformed by these mutants was specifically inhibited by imatinib. These results show for the first time that melanocytes require a specific epigenetic environment to be transformed by c-Kit mutants and highlight a distinct molecular mechanism of melanocyte transformation.

Critical roles for the serine 20, but not the serine 15, phosphorylation site and for the polyproline domain in regulating p53 turnover.

The p53 tumour suppressor protein is a short-lived transcription factor that becomes stabilized in response to a wide range of cellular stresses. Ubiquitination and the targeting of p53 for degradation by the proteasome are mediated by Mdm2 (mouse double minute clone 2), a negative regulatory partner of p53. Previous studies have suggested that DNA-damage-induced phosphorylation of p53 at key N-terminal sites has a pivotal role in regulating the interaction with Mdm2 but the precise role of phosphorylation of serines 15 and 20 is still unclear. Here we show that replacement of serine 15 and a range of other key N-terminal phosphorylation sites with alanine, which cannot be phosphorylated, has little effect on the ubiquitination and degradation of full-length human p53. In contrast, replacement of serine 20 makes p53 highly sensitive to Mdm2-mediated turnover. These results define distinct roles for serines 15 and 20, two sites previously demonstrated to be dependent on phosphorylation through mechanisms mediated by DNA damage and ATM (ataxia telangiectasia mutated). We also show that the polyproline region of p53, a domain that has a key role in p53-induced apoptosis, exerts a critical influence over the Mdm2-mediated turnover of p53.

Phosphorylation of murine double minute clone 2 (MDM2) protein at serine-267 by protein kinase CK2 in vitro and in cultured cells.

Murine double minute clone 2 oncoprotein (MDM2) is a key component in the regulation of the tumour suppressor p53. MDM2 mediates the ubiqutination of p53 in the capacity of an E3 ligase and targets p53 for rapid degradation by the proteasome. Stress signals which impinge on p53, leading to its activation, promote disruption of the p53-MDM2 complex, as in the case of ionizing radiation, or block MDM2 synthesis and thereby reduce cellular MDM2 levels, as in the case of UV radiation. It is therefore likely that MDM2, which is known to be modified by ubiquitination, SUMOylation and multi-site phosphorylation, may itself be a target for stress signalling (SUMO is small ubiquitin-related modifier-1). In the present study we show that, like p53, the MDM2 protein is a substrate for phosphorylation by the protein kinase CK2 (CK2) in vitro. CK2 phosphorylates a single major site, Ser(267), which lies within the central acidic domain of MDM2. Fractionation of cellular extracts revealed the presence of a single Ser(267) protein kinase which co-purified with CK2 on ion-exchange chromatography and, like CK2, was subject to inhibition by micromolar concentrations of the CK2-specific inhibitor 5,6-dichlororibofuranosylbenzimidazole. Radiolabelling of cells expressing tagged recombinant wild-type MDM2 or a S267A (Ser(267)-->Ala) mutant, followed by phosphopeptide analysis, confirmed that Ser(267) is a cellular target for phosphorylation. Ser(267) mutants are still able to direct the degradation of p53, but in a slightly reduced capacity. These data highlight a potential route by which one of several physiological modifications occurring within the central acidic domain of the MDM2 protein can occur.

Protein kinase CK1 is a p53-threonine 18 kinase which requires prior phosphorylation of serine 15.

p53 is a potent transcription factor which is regulated by sequential multisite phosphorylation and acetylation. In this paper, we identify threonine 18 of p53, a key site in regulating the interaction between p53 and its regulatory partner MDM2, as a novel site phosphorylated in vitro by purified recombinant casein kinase 1 (CK1) delta. Strikingly, phosphorylation of threonine 18 is dependent upon prior phosphorylation of serine 15. These data highlight an additional and physiologically important target residue for CK1 in p53 and suggest a potential mechanism by which sequential modification of a pivotal N-terminal residue in p53 may occur following stress-activated modification of serine 15.

Serine15 phosphorylation stimulates p53 transactivation but does not directly influence interaction with HDM2.

The p53 tumour suppressor protein is a labile transcription factor that is activated and stabilized in response to a wide range of cellular stresses, through a mechanism involving disruption of its interaction with MDM2, a negative regulatory partner. Induction of p53 by DNA damage additionally involves a series of phosphorylation and acetylation modifications, some of which are thought to regulate MDM2 binding. Here we report the effects of introducing mutations at several known or putative N-terminal phosphorylation sites on the transactivation function of p53. These studies highlight phosphorylation of Ser15, a key phosphorylation target during the p53 activation process, as being critical for p53-dependent transactivation. Biochemical data indicate that the mechanism by which phosphorylation of Ser15 stimulates p53-dependent transactivation occurs through increased binding to the p300 coactivator protein. The data also indicate that Ser15-dependent regulation of transactivation is independent of any involvement in modulating MDM2 binding, and that Ser15 phosphorylation alone is not sufficient to block the p53-MDM2 interaction.

Phosphorylation of murine p53, but not human p53, by MAP kinase in vitro and in cultured cells highlights species-dependent variation in post-translational modification.

The p53 tumour suppressor protein is tightly regulated by protein-protein association, protein turnover and a variety of post-translational modifications. Multisite phosphorylation plays a major role in activating and in finely tuning p53 function. The proline rich domain of murine p53 is a substrate for phosphorylation, in vitro and in cultured cells, by the p42ERK2 and p44ERK1 mitogen-activated protein (MAP) kinases. However, to date there have been no reports of attempts to determine whether p53 from any other species is a substrate for MAP kinase. In this paper we confirm that murine p53 is targeted by recombinant MAP kinase and by MAP kinases in extracts of both murine and human cells. In contrast, human p53 is not a substrate for recombinant MAP kinase nor are there any detectable levels of protein kinase activity in stimulated human cell extracts which phosphorylate the proline rich domain of human p53 in vitro. Finally, although stimulation of murine fibroblasts with o-tetradecanolylphorbol 13-acetate (TPA), an indirect activator of the MAP kinase pathway, leads to site-specific phosphorylation of murine p53, similar treatment of human fibroblasts and epithelial cells showed no significant changes in the phosphorylation pattern. These data are consistent with accumulating evidence that significant species-dependent differences exist in the post-translational modification of p53.

p53 mutations in skin and internal tumors of xeroderma pigmentosum patients belonging to the complementation group C.

Fifty-eight skin biopsies and three primary internal tumors from patients affected by the rare hereditary disease xeroderma pigmentosum (XP) were studied by an improved PCR-single strand conformation polymorphism analysis to detect the mutations of the tumor suppressor gene p53. The results from cutaneous XP tumors, including 27 squamous cell carcinomas and 6 basal cell carcinomas, show a very high level (86%) of p53 mutations. The analysis of mutations found in XP skin cancers according to the complementation group of the patients shows that tandem CC-->TT transitions are a characteristic of XP-C patients with a frequency much higher in their skin tumors (85%) compared with tumors in XP patients who do not belong to group C (33%). In all XP-C biopsies, mutations were due to replication of unrepaired DNA lesions on the nontranscribed strand of the p53 gene, substantiating the preferential repair in vivo of the transcribed strand of this gene in human tissues. For the first time, we were able to analyze three primary internal tumors (a neuroendocrine tumor of the thyroid, a gastric adenocarcinoma, and a glioma of the brain) of young XP children. All of them contained one mutation on the p53 gene, which were different from the ones found in the XP skin tumors and could have resulted from unrepaired lesions caused by oxidative damage.

Recovery of the normal p53 response after UV treatment in DNA repair-deficient fibroblasts by retroviral-mediated correction with the XPD gene.

Among the major responses of human cells to DNA damage is accumulation of the p53 tumor suppressor protein, which plays a crucial role as a cell-cycle checkpoint. We have already shown that this response is different in cells from the UV-hypersensitive human syndromes xeroderma pigmentosum (XP) and trichothiodystrophy (TTD), which overlap with each other and arise from mutations in genes involved in nucleotide excision repair. In this paper we report that correction of the repair defect by retroviral-mediated transduction of the wild-type XPD gene in XP-D and TTD/XP-D untransformed primary fibroblasts leads to a normal p53 response in these cells. Thus, the complemented cells, like normal human fibroblasts, require higher UV doses (10 J/m2) for p53 induction than the parental repair-deficient XP-D or TTD/XP-D cells (both mapping at the XPD locus), which accumulate p53 protein at very low UV doses (2.5 and 5 J/m2). The p53 protein levels return to normal 24 h after irradiation when UV-induced lesions have been efficiently repaired by the restored NER activity. These data confirm our earlier results that p53 accumulation following UV treatment is directly related to the presence of unrepaired cyclobutane dimers on the transcribed strand of active genes.

Quantitative detection of ultraviolet-specific p53 mutations in normal skin from Japanese patients.

We have previously developed sensitive methods to detect UV-specific p53 mutations (CC to TT tandem mutations) and have reported that such mutations could be found in the normal skin cell populations of sun-exposed body sites, but not in those of covered sites, in Australian cancer patients. We have now further refined our allele-specific PCR method for detecting CC to TT mutations at codons 247/248 of the p53 gene to allow quantitative measurements. Using DNA containing this mutation from a tumor as a standard for calibration and 5 micrograms of genomic DNA/PCR reaction, we could detect 1 mutant allele in about 10(6) wild-type alleles. It is essential to use purified primers and 64 degrees C as the annealing temperature for PCR. Our method has been applied in a study of the correlation of sun exposure and accumulation of CC to TT mutations in normal skin biopsies from Japanese patients. There were more p53 mutations in samples taken from sites that were chronically exposed to the sun than in those from covered sites. A significant trend of increased p53 mutation frequency with increase in age of subjects was found, suggesting the cumulative nature of the mutation. On the other hand, the p53 mutation frequency was higher in patients with premalignant tumors or nonmelanocytic skin cancer than in patients with only benign tumors. These results confirm the utility of PCR-based p53 gene mutation assays for the measurement of exposure to UV as well as for predicting the risk of UV-associated skin cancer.

The role of UV-B light in skin carcinogenesis through the analysis of p53 mutations in squamous cell carcinomas of hairless mice.

Mutation spectra of the p53 gene from human skin carcinomas have been connected to solar UV radiation. For comparison we have characterized the mutation spectrum of the p53 gene in a very large sample of squamous cell carcinomas from hairless mice induced with UV of wavelength 280-320 nm (UV-B), which have substantiated the mutagenic effects of UV-B radiation in vivo. Tumors from hairless mice, random bred SKH:HR1 as well as inbred SKH:HRA strains, which are analyzed for mutations in the conserved domains of the p53 protein present a very specific mutation spectrum. The observed mutation frequency after chronic UV-B radiation in the p53 gene ranged from 54% (SKH-HRA) to 73% (SKH-HR1) among the 160 tumors analyzed. Over 95% of the mutations were found at dipyrimidine sites located in the non-transcribed strand, the majority were C-->T transitions and 5% were CC-->TT tandem double mutations. Four distinct UV-B mutation hot spots have been identified for the first time: two major ones at codons 267 (33%) and 272 (19%) and two minor ones at codons 146 (10%) and 173 (4%). The codon 267 hot spot consists of a CpG preceded by a pyrimidine, which confirms in vivo an important role for this UV-B mutable site in UV-B-induced skin tumors that is not found in other types of mouse tumors. Comparison with mutation spectra from human skin carcinomas fully validates the merits of the hairless mouse model for studying the molecular mechanisms of skin carcinogenesis. For example, the murine hot spot at codon 272 does have a full equivalent in human skin carcinomas. In contrast, the human equivalent of the murine codon 267 lacks the dipyrimidine site and therefore fails to be a pronounced hot spot in human skin carcinomas; however, this site is one of the major hot spots in human internal cancers (evidently not induced by UV radiation but probably by deamination of the 5 methyl cytosine).

Prolonged p53 protein accumulation in trichothiodystrophy fibroblasts dependent on unrepaired pyrimidine dimers on the transcribed strands of cellular genes.

Trichothiodistrophy (TTD), xeroderma pigmentosum (XP), and Cockayne's syndrome (CS) are three distinct human diseases with sensitivity to ultraviolet (UV) radiation affected by mutations in genes involved in nucleotide excision repair (NER). Among the many responses of human cells to UV irradiation, both nuclear accumulation of p53, a tumor suppressor protein, and alterations in cell-cycle checkpoints play crucial roles. The purpose of this study was to define the signals transmitted after UV-C-induced DNA damage, which activates p53 accumulation in TTD/XP-D fibroblasts, and compare this with XP-D cell lines that carry different mutations in the same gene, XPD. Our results showed that p53 was rapidly induced in the nuclei of TTD/XP-D and XP-D fibroblasts in a dose-dependent manner after UV-C irradiation, as seen in XP-A and CS-A fibroblasts, much lower doses being required for the protein accumulation than in normal human fibroblasts, XP variant cells, and XP-C cells. The kinetics of accumulation of p53 and two effector proteins involved in cell-cycle arrest, WAF1 and GADD45, were also directly related to the repair potential of the cells, as in normal human fibroblasts their levels declined after 24 h, the time required for repair of UV-induced lesions, whereas NER-deficient TTD/XP-D cells showed p53, WAF1, and GADD45 accumulation for over 72 h after irradiation. Our results indicate that p53 accumulation followed by transcriptional activation of genes implicated in growth arrest is triggered in TTD/XP-D cells by the persistence of cyclobutane pyrimidine dimers, which are known to block transcription, on the transcribed strands of active genes.

The specificity of p53 mutation spectra in sunlight induced human cancers.

Ultraviolet (UV) irradiation emitted by the sun has been clearly implicated as a major carcinogen in the formation of skin cancers in man. Indeed, the high levels of cutaneous tumors in xeroderma pigmentosum patients (XP) who are deficient in repair of UV-induced lesions have confirmed that DNA damage produced by sunlight is directly involved in the cancer development. The tumor suppressor gene, p53, very frequently found modified in human cancers, has proved to be a perfect target gene for correlating mutation spectra with different cancer causing agents as there are nearly 300 potential mutation sites available for analysis. In a comparative analysis of p53 mutations found in internal cancers with those in skin tumours we show here that clear differences exist between the types of spectra obtained. The specificity of UV induced mutations in skin cancers is confirmed when single and tandem mutations are compared. Most of the p53 point mutations found are GC to AT transitions both in skin and internal tumors where in the latter they are located mainly at CpG sequences probably due to the deamination of the unstable 5-MeC. Moreover, mutations are targeted at py-py sequences in over 90% of skin tumors whereas in internal cancers the distribution is proportional to the frequency of bipyrimidine sequences in the p53 gene. Most significantly, all mutations found in XP skin tumors are targeted at py-py sites and more than 50% are tandem CC to TT transitions considered as veritable signatures of UV-induced lesions. Tandem mutations are also relatively common (14%) in skin tumors from normal individuals compared to their very rare occurrence in internal malignancies (0.8%). Finally, nearly all mutations observed in XP skin tumors are due to unrepaired lesions remaining on the coding strand whereas no strand bias is seen in mutation location of internal or skin tumors from normal individuals. In fact the mutation spectrum analysed in XP skin cancers has permitted the first demonstration of the existence of preferential repair in man. In conclusion, using the p53 gene as a probe it is obvious that the mutation spectra from skin tumors are very similar to those observed in UV-treated gene targets in model systems but statistically different from those described in other types of human cancer. This has allowed us to demonstrate, without ambiguity, the major role of UV-induced DNA lesions in sunlight related skin carcinogenesis.

Can we predict solar ultraviolet radiation as the causal event in human tumours by analysing the mutation spectra of the p53 gene?

The tumour suppressor gene, p53, has proved to be one of the genes most often modified in human cancers. These alterations consist mainly of point mutations located in the evolutionarily conserved sequences which render the protein inactive for its normal biological functions. In fact the p53 gene presents nearly 300 potential mutation sites whose analysis should enable the correlation of specific mutation spectra with different causal agents in cancer development. In this study we have analysed the mutation spectrum of the p53 gene in skin tumours from normal individuals and repair-deficient xeroderma pigmentosum (XP) patients in comparison with mutations found in internal cancers. Point mutations are mainly GC-->AT transitions in skin tumours (74% in non-XP, 87% in XP), and also to a lesser extent in internal tumours (47%) where, however, they are mainly located at CpG (63%) sequences probably due to the deamination of the unstable 5-MeC. Moreover, mutations are targeted at py-py sequences in over 90% of skin tumours whereas the distribution of mutations in internal malignancies is proportional to the frequency of py-py sites (61%) and other sequences (39%) at mutable sites. Indeed, in XP skin tumours 100% of the mutations are targeted at py-py sequences and 55% of these are tandem CC-->TT transitions considered as a signature of UV-induced lesions. In skin tumours from normal individuals, 14% of the p53 mutations are double mutations and as in XP skin tumours all these are CC-->TT transitions. In contrast, internal tumours rarely contain tandem mutations (0.8%), and of these only 2/14 were CC-->TT transitions. Finally, nearly all (95%) of the mutations in XP are located on the non-transcribed strand while internal or non-XP skin tumours do not show this strand bias. Hence, the mutation spectrum analysed in XP skin tumours also demonstrates for the first time the existence of preferential repair in humans. In conclusion, the specificity of UV-induced p53 mutation spectra in skin tumours shows that this gene is a particularly appropriate candidate for the correlation of mutation spectra with specific damaging agents.

Specific UV-induced mutation spectrum in the p53 gene of skin tumors from DNA-repair-deficient xeroderma pigmentosum patients.

The UV component of sunlight is the major carcinogen involved in the etiology of skin cancers. We have studied the rare, hereditary syndrome xeroderma pigmentosum (XP), which is characterized by a very high incidence of cutaneous tumors on exposed skin at an early age, probably due to a deficiency in excision repair of UV-induced lesions. It is interesting to determine the UV mutation spectrum in XP skin tumors in order to correlate the absence of repair of specific DNA lesions and the initiation of skin tumors. The p53 gene is frequently mutated in human cancers and represents a good target for studying mutation spectra since there are > 100 potential sites for phenotypic mutations. Using reverse transcription-PCR and single-strand conformation polymorphism to analyze > 40 XP skin tumors (mainly basal and squamous cell carcinomas), we have found that 40% (17 out of 43) contained at least one point mutation on the p53 gene. All the mutations were located at dipyrimidine sites, essentially at CC sequences, which are hot spots for UV-induced DNA lesions. Sixty-one percent of these mutations were tandem CC-->TT mutations considered to be unique to UV-induced lesions; these mutations are not observed in internal human tumors. All the mutations, except two, must be due to translesion synthesis of unrepaired dipyrimidine lesions left on the nontranscribed strand. These results show the existence of preferential repair of UV lesions [either pyrimidine dimers or pyrimidine-pyrimidone (6-4) photoproducts] on the transcribed strand in human tissues.