Significance of signal transducer and activator of transcription 3 during multistage mouse skin carcinogenesis

Signal transducer and activator of transcription 3 (Stat3) is a signaling molecule that transduces signal from cell surface receptors, itself translocates into the nucleus, binds to consensus promoter sequences and activates gene transcription. Here, we showed that Stat3 is constitutively activated in both premalignant tumors (papillomas) and squamous cell carcinomas of mouse skin that is induced by topical treatment with an initiator 7,12-dimethylbenz[a]anthracene (DMBA) followed by a tumor promoter 12-O-tetradecanoyl-phorbol-13-acetate (TPA). Additional data demonstrated that epidermal growth factor signaling contributes to the activation of Stat3 in this model. Using mice where Stat3 function is abrogated in keratinocytes via the Cre-LoxP system (K5Cre.Stat3 flox/flox), we demonstrated that Stat3 is required for de novo carcinogenesis since Stat3 deficiency leads to a complete abrogation of skin tumor development induced by DMBA and TPA. We subsequently showed that Stat3 plays a role in both the initiation and promotion stages of carcinogenesis. During initiation, Stat3 functions as an anti-apoptotic molecule for maintaining the survival of DNA-damaged keratinocyte stem cells. During promotion, Stat3 functions as a critical regulator for G1 to S phase cell cycle progression to confer selective clonal expansion of initiated cells into papillomas. On the other hand, using transgenic mice over-expressing a constitutively dimerized form of Stat3 (Stat3C) in keratinocytes (K5.Stat3C), we revealed a role for Stat3 in tumor progression. After treatment with DMBA and TPA, K5.Stat3C transgenic mice developed skin tumors with a shorter latency when 100% bypassed the premalignant stage and became carcinoma in situ. Histological and immunohistochemical analysis revealed these tumors as highly vascularized and poorly differentiated. More strikingly, these tumors exhibited invasion into surrounding mesenchymal tissue, some of which metastasized into lung. The tumor-mesenchymal front was characterized by partial loss of E-cadherin and elevation of vimentin, markers characterizing epithelial-mesenchymal transition. On the other hand, inhibition of Stat3 via a decoy oligonucleotide led to a significant reduction of tumor size in approximately 50% of all papillomas tested. In conclusion, we demonstrated that Stat3 plays a critical in all three stages (initiation, promotion and progression) of skin carcinogenesis, and it may potentially become a good target for cancer prevention and anti-cancer therapy. ^

Roles of IkappaB kinase alpha in centrosome duplication and skin carcinogenesis

IκB kinase α (IKKα) is one kinase subunit of the IKK complex that is responsible for NF-κB activation. Previous studies have shown that IKKα determines mouse keratinocyte terminal differentiation independent of the NF-κB pathway. Accumulating evidence suggests that IKKα functions as a tumor suppressor in skin carcinogenesis; however, the downstream pathways mediating this function are largely unknown. By using primary cultured keratinocytes, we found that Ikkα-/- cells developed aneuploidy and underwent spontaneous immortalization and transformation while wild type cells underwent terminal differentiation in the same culture condition. Using proteomic analysis we identified nucleophosmin (NPM), a centrosome duplication regulator, as an IKKα substrate. We further demonstrated that IKKα interacted with NPM and colocalized with NPM on the centrosome, suggesting that NPM is a physiological substrate of IKKα. Loss of IKKα reduced centrosome-bound NPM and promoted abnormal centrosome amplification, which contributed to aneuploidy development. Detailed analysis revealed that ablation of IKKα target site serine-125 of NPM induced destabilization of NPM hexamers, disrupted NPM association with centrosomes, and resulted in abnormal centrosome amplification. Re-introduction of IKKα rescued the defect in Ikkα-/- keratinocytes. Thus, IKKα is required for maintaining proper centrosome duplication by phosphorylating NPM. ^ UV is the major etiological agent for human skin cancer and UV-induced mouse skin carcinogenesis is one of the most relevant experimental models for human skin carcinogenesis. Thus, we further evaluated IKKα function in UV-induced skin carcinogenesis in Ikkα+/- mice. We demonstrated that IKKα is also critical in UV skin carcinogenesis, as evidenced by increased tumor multiplicity and reduced tumor latency in Ikkα+/- mice after chronic UVB treatment. Reduced expression of IKKα decreased UV-induced apoptosis and promoted accumulation of P53 mutations in the epidermis. This indicates that IKKα is critical for UV-induced apoptosis in vivo and thus prevents mutation accumulation that is important for tumor development. ^ Together, these findings uncover previously unknown in vivo functions of IKKα in centrosome duplication and apoptosis, thus providing a possible mechanism of how loss of IKKα may contribute to skin carcinogenesis. ^

The effect of DNA methylation in carcinogenesis

CpG island methylation within single gene promoters can silence expression of associated genes. We first extended these studies to bidirectional gene pairs controlled by single promoters. We showed that hypermethylation of bidirectional promoter-associated CpG island silences gene pairs (WNT9A/CD558500, CTDSPL/BC040563, and KCNK15/BF 195580) simultaneously. Hypomethylation of these promoters by 5-aza-2'-deoxycytidine treatment reactivated or enhanced gene expression bidirectionally. These results were further confirmed by luciferase assays. Methylation of WNT9A/CD558500 and CTDSPL/BC040563 promoters occurs frequently in primary colon cancers and acute lymphoid leukemia, respectively. ^ Next we sought to understand the origins of hypermethylation in cancer. CpG islands associated with tumor suppressor genes are normally free from methylation, but can be hypermethylated in cancer. It remains poorly understood how these genes are protected from methylation in normal tissues. In our studies, we aimed to determine if cis-acting elements in these genes are responsible for this protection, using the tumor suppressor gene p16 as a model. We found that Alu repeats located both upstream and downstream of the p16 promoter become hypermethylated with age. In colon cancer samples, the methylation level is particularly high, and the promoter can also be affected. Therefore, the protection in the promoter against methylation spreading could fail during tumorigenesis. This methylation pattern in p16 was also observed in cell lines of different tissue origins, and their methylation levels were found to be inversely correlated with that of active histone modification markers (H3K4-3me and H3K9-Ac). To identify the mechanism of protection against methylation spreading, we constructed serial deletions of the p16 protected region and used silencing of a neomycin reporter gene to evaluate the protective effects of these fragments. A 126 bp element was identified within the region which exerts bidirectional protection against DNA methylation, independently of its transcriptional activity. The protective strength of this element is comparable to that of the HS4 insulator. During long-term culture, the presence of this element significantly slowed methylation spreading. In conclusion, we have found that an element located in the p16 promoter is responsible for protection against DNA methylation spreading in normal tissues. The failure of protective cis-elements may be a general feature of tumor-suppressor gene silencing during tumorigenesis. ^

AN INVESTIGATION OF THE MODIFYING EFFECTS OF VITAMIN A AND HYPOXIC CELL SENSITIZERS IN RADIATION CARCINOGENESIS IN MICE

The effect of vitamin A (retinyl acetate) and three hypoxic cell sensitizers (metronidazole, misonidazole and desmethylmisonidazole) on lung tumor development in strain A mice exposed to radiation was assessed.^ In experiments involving vitamin A, two groups of mice were fed a low vitamin A diet (< 100 IU/100g diet) while the two other groups were fed a high vitamin A diet (800 IU/100g diet). After two weeks one group maintained on the high vitamin A diet and one group maintained on the low vitamin A diet were given an acute dose of 500 rad of gamma radiation to the thoracic region. The circulating level of plasma vitamin A in all four groups of mice was monitored. A difference in circulating vitamin A in the mice maintained on high and low vitamin A diet became evident by 20 weeks and continued for the duration of the experiment. Mice were killed 18, 26, and 40 weeks post irradiation, their lungs were removed and the number of surface adenomas were counted. There was a significant increase in the number of mice bearing lung tumors and the mean number of lung tumors per mouse in the irradiated group maintained on the high vitamin A diet at 40 weeks post irradiation as compared to the irradiated group maintained on a low vitamin A diet (p < 0.05). Under the conditions of this experiment the development of pulmonary adenomas in irradiated strain A mice appears to relate directly to circulating levels of vitamin A.^ In the other experiment two dose levels of the hypoxic cell sensitizers, 0.2mg/g and 0.6mg/g, were used either alone or in combination with 900 rad of gamma radiation in a fractionated dose schedule of twice a week for three weeks. In the groups of mice which received hypoxic cell sensitizers only, the prevalence and the mean number of lung tumors per mouse were somewhat increased (p < 0.10) in the higher dose group (0.6mg/g) of misonidazole but was not significantly different from the control animals in the other two sensitizer groups. The combination of hypoxic cell sensitizer and radiation did not show any significant enhancement of lung tumor response when compared with the group which received radiation only. The dose of radiation used in this study significantly enhanced lung tumor formation in mice when compared with the control group. Thus, under the experimental exposure conditions used in this investigation, which were very similar to the exposure conditions occurring in clinical treatment, all three hypoxic cell sensitizers did not sensitize the mouse to the carcinogenic effects of gamma radiation.^

EVIDENCE OF HUMAN ENDOGENOUS RETROVIRUS K INVOLVEMENT IN HUMAN CANCER

Many lines of clinical and experimental evidence indicate a viral role in carcinogenesis (1-6). Our access to patient plasma, serum, and tissue samples from invasive breast cancer (N=19), ductal carcinoma in situ (N=13), malignant ovarian cancer (N=12), and benign ovarian tumors (N=9), via IRB-approved and informed consent protocols through M.D. Anderson Cancer Center, as well as normal donor plasmas purchased from Gulf Coast Regional Blood Center (N=6), has allowed us to survey primary patient blood and tissue samples, healthy donor blood from the general population, as well as commercially available human cell lines for the presence of human endogenous retrovirus K (HERV-K) Env viral RNA (vRNA), protein, and viral particles. We hypothesize that HERV-K proteins are tumor-associated antigens and as such can be profiled and targeted in patients for diagnostic and therapeutic purposes. To test this hypothesis, we employed isopycnic ultracentrifugation, a microplate-based reverse transcriptase enzyme activity assay, reverse transcription – polymerase chain reaction (RT-PCR), cDNA sequencing, SDS-PAGE and western blotting, immunofluorescent staining, confocal microscopy, and transmission electron microscopy to evaluate v HERV-K activation in cancer. Data from large numbers of patients tested by reverse transcriptase activity assay were analyzed statistically by t-test to determine the potential use of this assay as a diagnostic tool for cancer. Significant reverse transcriptase enzyme activity was detected in 75% of ovarian cancer patients, 53.8% of ductal carcinoma in situ patient, and 42.1% of invasive breast cancer patient samples. Only 11.1% of benign ovarian patient and 16.7% of normal donor samples tested positive. HERV-K Env vRNA, or Env SU were detected in the majority of cancer types screened, as demonstrated by the results shown herein, and were largely absent in normal controls. These findings support our hypothesis that the presence of HERV-K in patient blood circulation is an indicator of cancer or pre-malignancy in vivo, that the presence of HERV-K Env on tumor cell surfaces is indicative of malignant phenotype, and that HERV-K Env is a tumor-associated antigen useful not only as a diagnostic screening tool to predict patient disease status, but also as an exploitable therapeutic target for various novel antibody-based immunotherapies.

Alterations in the ras oncogene and the p53 tumor suppressor gene in ultraviolet radiation-induced skin carcinogenesis

A rapid increase of the ultraviolet radiation (UVR)-related skin cancer incidence has attracted more and more public attention during the last few decades. Prevention and treatment of UVR-related skin cancer has become an important public health issue in the United States. Recent studies indicate that mutations in ras and/or p53 genes may be involved in UVR-induced skin tumor development but the precise molecular mechanism remains unclear. In this study, alterations of H-ras and p53 genes were investigated in different stages of carcinogenesis in a chronic UVR (solar simulator) exposure-induced Sencar mouse skin carcinogenesis model in order to clarify the role of the alterations of these genes during the skin carcinogenesis process and to further understand the mechanisms by which UVR causes skin cancer.^ Positive ras-p21 staining in cell membranes and cytosol were detected in 18/33 (55%) of squamous cell carcinomas (SCCs), but were not detected in UV-exposed skin, papillomas, or spindle cell tumors (SCTs). Positive staining of the malignant progression marker K13 was found in 17/33 (52%) of SCCs only. A significant positive correlation was observed between the K13 and the ras-p21 expression. Polymerase chain reaction (PCR)-based single strand conformation polymorphism (SSCP) analysis and gene sequencing analysis revealed three point mutations, one (codon 56) in UV-exposed non-tumor bearing skin and the other two (codons 21 and 13) in SCCs. No UV-specific mutation patterns were found.^ Positive p53 nuclear staining was found in 10/37 (27%) of SCCs and 12/24 (50%) of SCTs, but was not detected in normal skin or papillomas. PCR-based SSCP and sequencing analysis revealed eight point mutations in exons 5 and 6 (four in SCTs, two in SCCs, and two in UV-exposed skin) including six C-T or C-A transitions. Four of the mutations occurred at a dipyrimidine (CC) sequence. The pattern of the mutations indicated that the mutagenic lesions were induced by UVR.^ These results indicate that overexpression of ras-p21 in conjunction with aberrant expression of K13 occurred frequently in UVR-induced SCCs in Sencar mouse skin. The point mutation in the H-ras gene appeared to be a rare event in UVR skin carcinogenesis and may not be responsible for overexpression of ras-p21. UVR-induced P53 gene alteration is a frequent event in UVR-induced SCCs and later stage SCT tumors in Sencar mice skin, suggesting the p53 gene mutation plays an important role in skin tumor malignant progression. ^

Modulation of cytochrome P450 enzyme activity and PAH -induced skin carcinogenesis by naturally occurring coumarins

The present study was designed to determine the potential anticarcinogenic activity of naturally occurring coumarins and their mechanism of action. The results indicated that several naturally occurring coumarins including bergamottin, coriandrin, imperatorin, isopimpinellin, and ostruthin, to which humans are routinely exposed in the diet, were effective inhibitors and/or inactivators of CYP1A1-mediated ethoxyresorufin-O-dealkylase (EROD) or CYP2B1-mediated pentoxyresorufin-O-dealkylase (PROD) in mouse liver microsomes. In addition, bergamottin and corandrin were also found to be inhibitors of purified human P450 1A1 in vitro. Further studies with coriandrin revealed that this compound was a mechanism-based inactivator of P450 1A1 and covalently bound to the P450 1A1 apoprotein. In cultured mouse keratinocytes, bergamottin and coriandrin effectively inhibited the B(a) P metabolism and significantly decreased covalent binding of B(a) P and DMBA to keratinocyte DNA and anti-diol-epoxide-DNA adducts derived from both B(a) P and DMBA in keratinocytes. The data from in vivo experiments showed that bergamottin and coriandrin were potent inhibitors of covalent binding of B (a) P to epidermal DNA and the formation of (+) anti BPDE-DNA adduct, whereas imperatorin and isopimpinellin were more potent inhibitors of covalent binding of DMBA to epidermal DNA. The ability of coumarins to inhibit covalent binding of B (a) P to DNA in mouse epidermis was positively correlated with their inhibitory effect P450 1A1 in vitro, while the inhibitory effect of coumarins on covalent binding of DMBA to epidermal DNA was positively correlated with their inhibitory effects on P450 2B1 and negatively to their inhibitory activity toward P450 1A1. The data from tumor experiments indicated that bergamottin, ostruthin, and coriandrin inhibited tumor initiation by B (a) P in a two-stage carcinogenesis protocol. Bergamottin was most effective in this regard and produced a dose dependent inhibition of papilloma formation in these experiments. In addition, imperatorin was an effective inhibitor of skin tumorigenesis induced by DMBA in SENCAR mouse skin using both a two-stage and a complete carcinogenesis protocol. At dose levels higher than those effective against DMBA, imperatorin also inhibited tumor initiation by B (a) P. The results to date demonstrate that several naturally occurring coumarins possess the ability to block tumor initiation and tumorigenesis by PAHs such as B (a) P and DMBA through inhibition of the P450s involved in the metabolic activation of these hydrocarbons. A working model for the involvement of specific P450s in the metabolic activation of these two PAHs was proposed. ^

Implication of bFGF signaling in lung carcinogenesis

A Western Array Screening system in conjunction with an in vitro lung carcinogenesis model, which consists of human bronchial epithelial (HBE) cells representing normal (NHBE), immortalized (BEAS-2B and 1799), transformed (1198), and tumorigenic (1170-I) was used to test the hypothesis that lung carcinogenesis involves specific changes in signaling proteins. Forty six proteins whose expression was upregulated by >2 fold and 23 proteins whose expression was downregulated by >2 fold in 1170-I compared to NHBE cells were identified. The levels of six proteins including bFGF (both intracellular and secreted), Akt and p70s6K in the PI3KJp70s6K pathway and the bFGF receptor (FGFR1) were upregulated in different stages of lung carcinogenesis. Akt activity and phospho-p70s6K were also increased in 1170-I compared to NHBE cells, suggesting that PI3K/p70s6K pathway is activated during lung carcinogenesis. bFGF treatment stimulated the growth of the 1170-I cells. Both tyrosine phosphorylation of FGFR1 and cell growth were inhibited in 1170-I cells after overexpression of dominant-negative(DN) FGFR1. Growth inhibition involved a G2 arrest related to decreased cdc2 activity, cdc25C downregulation, Wee1, p21(WAF1) and p27(Kip1) upregulation. Apoptosis was observed in tumorigenic but not in normal cells after overexpression of DNFGFR1. Confluent NHBE cells, were much less sensitive to the growth inhibition by DNFGFR1 compared to other cell lines analyzed. bFGF increased phospho-Akt and phospho-p70s6K in 1170-I cells. The Akt inhibitor LY294002 and the p70s6K inhibitor rapamycin inhibited bFGF-stimulated cell growth in 1170-I cells. Both agents downregulated the bFGF-induced increase in S phase by inducing G1 arrest. Also, LY294002 inhibited bFGF increased phospho-Akt, while both LY294002 and rapamycin inhibited bFGF increased phospho-p70s6K. Thus, cell proliferation stimulated by bFGF in 1170-I cells was at least partially mediated by PI3K/p70s6K pathway. Hsp90 was upregulated by bFGF in 1170-I cells. Its inhibitor geldanamycin inhibited the bFGF-stimulated growth via inducing apoptosis and G2 arrest through decreases in cdc2 expression/activity and p21 upregulation, and decreased Akt/phospho-Akt, p70s6K/phospho-p70s6K and Bad. Hsp90, p70s6K and Bad were found in the same complex, which may be important for signaling cell survival. Taken together, our study suggests that bFGF signaling, especially PI3K/p70s6K pathway, is important for lung carcinogenesis. ^

Functional studies of *Ras and Bcl-2 oncoproteins in keratinocyte homeostasis and multistep skin carcinogenesis

To assess the effect of deregulated Ha-ras and bcl-2, individually and in combination on epidermal keratinocyte homeostasis and during multistep skin carcinogenesis, we generated skin-specific transgenic mice and keratinocyte transfectants constitutively expressing oncogenic Ha-ras and bcl-2 proteins. The deregulated Ha-ras and bcl-2 expression contributing to homeostatic imbalances in the skin had an additive effect on the probability of tumor development. They were also cooperative in incidence, growth, and latency of tumor formation, and they exhibited synergistic cooperation in malignant transformation of benign papillomas. To explain the homeostatic imbalances by Ha-ras and bcl-2 overexpression in the skin, we investigated the three major cellular processes of proliferation, cell death, and differentiation. Epidermal expression of Bcl-2 retarded keratinocyte proliferation in the epidermis of neonatal mice compared with results for control littermates. Constitutive expression of Ha-ras increased keratinocyte proliferation, and co-expression of bcl-2 modestly suppressed the ras-mediated abnormal proliferation of neonatal keratinocytes. Bcl-2 proteins in keratinocytes protected UV-treated cells from apoptotic cell death regardless of oncogenic ras expression in both non-neoplastic neonatal epidermis and human keratinocyte cell lines. The spontaneous apoptotic index (AI) was also lower in papillomas constitutively expressing bcl-2 compared with the ones that developed in control mice. Ras-overexpressing epidermis, including that in ras/bcl-2 double transgenic mice, had abnormal differentiation patterns compared with controls. The oncogenic ras protein had alterations in both epidermal distribution and the extent of cytokeratin 14 and involucrin expression. Abnormal expression of the hyperproliferation marker cytokeratin 6 and modest down regulation of cytokeratin 1 were also detected. Late appearance of filaggrin was another abnormal phenotype of the ras-expressing epidermis. Overexpression of bcl-2 had no effect on epidermal differentiation. Together, these findings suggest that constitutive expression of oncogenic Ha-ras and bcl-2 are important determinants of epidermal proliferation, viability and differentiation. In summary, our results demonstrated that the disruption of epidermal homeostasis by overexpressed ras and bcl-2 predisposes to hyperplastic growth of the epidermis and to papilloma development and that these proteins with distinct mechanisms for oncogenesis are functionally synergistic for malignant transformation of chemically induced skin carcinogenesis. ^