Survival of Lung Cancer Patients Dependent on the LOH Status for DMP1, ARF, and p53.

Lung cancer is the leading cause of cancer deaths in the world, and accounts for more solid tumor deaths than any other carcinomas. The prognostic values of DMP1, ARF, and p53-loss are unknown in lung cancer. We have conducted survival analyses of non-small cell lung cancer (NSCLC) patients from the University of Minnesota VA hospital and those from the Wake Forest University Hospital. Loss of Heterozygosity (LOH) for hDMP1 was found in 26 of 70 cases (37.1%), that of the ARF/INK4a locus was found in 33 of 70 (47.1%), and that of the p53 locus in 43 cases (61.4%) in the University of Minnesota samples. LOH for hDMP1 was associated with favorable prognosis while that of p53 predicted worse prognosis. The survival was much shorter for ARF-loss than INK4a-loss, emphasizing the importance of ARF in human NSCLC. The adverse effect of p53 LOH on NSCLC patients' survival was neutralized by simultaneous loss of the hDMP1 locus in NSCLC and breast cancer, suggesting the possible therapy of epithelial cancers with metastatic ability.

c-MYB and DMTF1 in Cancer.

The c-Myb gene encodes a transcription factor that regulates cell proliferation, differentiation, and apoptosis through protein-protein interaction and transcriptional regulation of signaling pathways. The protein is frequently overexpressed in human leukemias, breast cancers, and other solid tumors suggesting that it is a bona fide oncogene. c-MYB is often overexpressed by translocation in human tumors with t(6;7)(q23;q34) resulting in c-MYB-TCRß in T cell ALL, t(X;6)(p11;q23) with c-MYB-GATA1 in acute basophilic leukemia, and t(6;9)(q22-23;p23-24) with c-MYB-NF1B in adenoid cystic carcinoma. Antisense oligonucleotides to c-MYB were developed to purge bone marrow cells to eliminate tumor cells in leukemias. Recently, small molecules that inhibit c-MYB activity have been developed to disrupt its interaction with p300. The Dmp1 (cyclin D binding myb-like protein 1; Dmtf1) gene was isolated through its virtue for binding to cyclin D2. It is a transcription factor that has a Myb-like repeat for DNA binding. The Dmtf1 protein directly binds to the Arf promoter for transactivation and physically interacts with p53 to activate the p53 pathway. The gene is hemizygously deleted in 35-42% of human cancers and is associated with longer survival. The significances of aberrant expression of c-MYB and DMTF1 proteins in human cancers and their clinical significances are discussed.

Tumor suppression by the EGR1, DMP1, ARF, p53, and PTEN Network.

Recent studies have indicated that EGR1 is a direct regulator of tumor suppressors including TGFß1, PTEN, and p53. The Myb-like transcription factor Dmp1 is a physiological regulator of the Arf-p53 pathway through transactivation of the Arf promoter and physical interaction of p53. The Dmp1 promoter has binding sites for Egr proteins, and Egr1 is a target for Dmp1. Crosstalks between p53 and PTEN have been reported. The Egr1-Dmp1-Arf-p53-Pten pathway displays multiple modes of interaction with each other, suggesting the existence of a functional network of tumor suppressors that maintain normal cell growth and prevent the emergence of incipient cancer cells.

Oncogenic and tumor-suppressive mouse models for breast cancer engaging HER2/neu.

The human c-ErbB2 (HER2) gene is amplified in ∼20% of human breast cancers (BCs), but the protein is overexpressed in ∼30% of the cases indicating that multiple different mechanisms contribute to HER2 overexpression in tumors. It has long been used as a molecular marker of BC for subcategorization for the prediction of prognosis and determination of therapeutic strategies. In comparison to ER(+) BCs, HER2-positive BCs are more invasive, but the patients respond to monoclonal antibody therapy with trastuzumab or tyrosine kinase inhibitors at least at early stages. To understand the pathophysiology of HER2-driven carcinogenesis and test HER2-targeting therapeutic agents in vivo, numerous mouse models have been created that faithfully reproduce HER2(+) BCs in mice. They include MMTV-neu (active mutant or wild type, rat neu or HER2) models, neu promoter-driven neuNT-transgenic mice, neuNT-knock-in mice at the neu locus and doxycycline-inducible neuNT-transgenic models. HER2/neu activates the Phosphatidylinositol-3 kinase-AKT-NF-κB pathway to stimulate the mitogenic cyclin D1/Cdk4-Rb-E2F pathway. Of note, overexpression of HER2 also stimulates the cell autonomous Dmp1-Arf-p53 tumor suppressor pathway to quench oncogenic signals to prevent the emergence of cancer cells. Hence tumor development by MMTV-neu mice was dramatically accelerated in mice that lack Dmp1, Arf or p53 with invasion and metastasis. Expressions of neuNT under the endogenous promoter underwent gene amplification, closely recapitulating human HER2(+) BCs. MMTV-HER2 models have been shown to be useful to test humanized monoclonal antibodies to HER2. These mouse models will be useful for the screening of novel therapeutic agents against BCs with HER2 overexpression.

Stabilization of the p53-DNA Complex by the Nuclear Protein Dmp1α.

We recently reported the existence of a physical interaction between the Myb-like transcription factor Dmp1 (Dmtf1) and p53 in which Dmp1 antagonized polyubiquitination of p53 by Mdm2 and promoted its nuclear localization. Dmp1 significantly stabilized p53-DNA complexes on promoters that contained p53-consensus sequences, which were either supershifted or disrupted with antibodies to Dmp1. Lysates from mice injected with doxorubicin showed that Dmp1 bound to p21Cip1, Bbc3, and Thbs1 gene regulatory regions in a p53-dependent fashion. Our data suggest that acceleration of DNA-binding of p53 by Dmp1 is a critical process for Dmp1 to increase the p53 function in Arf-deficient cells.

Aberrant splicing of the DMP1-ARF-MDM2-p53 pathway in cancer.

Alternative splicing (AS) of mRNA precursors is a ubiquitous mechanism for generating numerous transcripts with different activities from one genomic locus in mammalian cells. The gene products from a single locus can thus have similar, dominant-negative or even opposing functions. Aberrant AS has been found in cancer to express proteins that promote cell growth, local invasion and metastasis. This review will focus on the aberrant splicing of tumor suppressor/oncogenes that belong to the DMP1-ARF-MDM2-p53 pathway. Our recent study shows that the DMP1 locus generates both tumor-suppressive DMP1α (p53-dependent) and oncogenic DMP1ß (p53-independent) splice variants, and the DMP1ß/α ratio increases with neoplastic transformation of breast epithelial cells. This process is associated with high DMP1ß protein expression and shorter survival of breast cancer (BC) patients. Accumulating pieces of evidence show that ARF is frequently inactivated by aberrant splicing in human cancers, demonstrating its involvement in human malignancies. Splice variants from the MDM2 locus promote cell growth in culture and accelerate tumorigenesis in vivo. Human cancers expressing these splice variants are associated with advanced stage/metastasis, and thus have negative clinical impacts. Although they lack most of the p53-binding domain, their activities are mostly dependent on p53 since they bind to wild-type MDM2. The p53 locus produces splice isoforms that have either favorable (ß/γ at the C-terminus) or negative impact (Δ40, Δ133 at the N-terminus) on patients' survival. As the oncogenic AS products from these loci are expressed only in cancer cells, they may eventually become targets for molecular therapies.

Transcription factors that interact with p53 and Mdm2.

The tumor suppressor p53 is activated upon cellular stresses such as DNA damage, oncogene activation, hypoxia, which transactivates sets of genes that induce DNA repair, cell cycle arrest, apoptosis, or autophagy, playing crucial roles in the prevention of tumor formation. The central regulator of the p53 pathway is Mdm2 which inhibits transcriptional activity, nuclear localization and protein stability. More than 30 cellular p53-binding proteins have been isolated and characterized including Mdm2, Mdm4, p300, BRCA1/2, ATM, ABL and 53BP-1/2. Most of them are nuclear proteins; however, not much is known about p53-binding transcription factors. In this review, we focus on transcription factors that directly interact with p53/Mdm2 through direct binding including Dmp1, E2F1, YB-1 and YY1. Dmp1 and YB-1 bind only to p53 while E2F1 and YY1 bind to both p53 and Mdm2. Dmp1 has been shown to bind to p53 and block all the known functions for Mdm2 on p53 inhibition, providing a secondary mechanism for tumor suppression in Arf-null cells. Although E2F1-p53 binding provides a checkpoint mechanism to silence hyperactive E2F1, YB-1 or YY1 interaction with p53 subverts the activity of p53, contributing to cell cycle progression and tumorigenesis. Thus, the modes and consequences for each protein-protein interaction vary from the viewpoint of tumor development and suppression.

DMP1ß, a splice isoform of the tumour suppressor DMP1 locus, induces proliferation and progression of breast cancer.

Our recent work has indicated that the DMP1 locus on 7q21, encoding a haplo-insufficient tumour suppressor, is hemizygously deleted at a high frequency in breast cancer. The locus encodes DMP1α protein, an activator of the p53 pathway leading to cell cycle arrest and senescence, and two other functionally undefined isoforms, DMP1ß and DMP1γ. In this study, we show that the DMP1 locus is alternatively spliced in ∼30% of breast cancer cases with relatively decreased DMP1α and increased DMP1ß expression. RNA-seq analyses of a publicly available database showed significantly increased DMP1ß mRNA in 43-55% of human breast cancers, dependent on histological subtypes. Similarly, DMP1ß protein was found to be overexpressed in ∼60% of tumours relative to their surrounding normal tissue. Importantly, alteration of DMP1 splicing and DMP1ß overexpression were associated with poor clinical outcomes of the breast cancer patients, indicating that DMP1ß may have a biological function. Indeed, DMP1ß increased proliferation of non-tumourigenic mammary epithelial cells and knockdown of endogenous DMP1 inhibited breast cancer cell growth. To determine DMP1ß's role in vivo, we established MMTV-DMP1ß transgenic mouse lines. DMP1ß overexpression was sufficient to induce mammary gland hyperplasia and multifocal tumour lesions in mice at 7-18 months of age. The tumours formed were adenosquamous carcinomas with evidence of transdifferentiation and keratinized deposits. Overall, we identify alternative splicing as a mechanism utilized by cancer cells to modulate the DMP1 locus through diminishing DMP1α tumour suppressor expression, while simultaneously up-regulating the tumour-promoting DMP1ß isoform.

Mutually exclusive inactivation of DMP1 and ARF/p53 in lung cancer.

Dmp1 (Dmtf1) is activated by oncogenic Ras-Raf signaling and induces cell-cycle arrest in an Arf, p53-dependent fashion. The survival of K-ras(LA) mice was shortened by approximately 15 weeks in both Dmp1(+/-) and Dmp1(-/-) backgrounds, the lung tumors of which showed significantly decreased frequency of p53 mutations compared to Dmp1(+/+). Approximately 40% of K-ras(LA) lung tumors from Dmp1(+/+) mice lost one allele of the Dmp1 gene, suggesting the primary involvement of Dmp1 in K-ras-induced tumorigenesis. Loss of heterozygosity (LOH) of the hDMP1 gene was detectable in approximately 35% of human lung carcinomas, which was found in mutually exclusive fashion with LOH of INK4a/ARF or that of P53. Thus, DMP1 is a pivotal tumor suppressor for both human and murine lung cancers.

Alterations of p63 and p73 in human cancers.

p53 and its related genes, p63 and p73 constitute the p53 gene family. While p53 is the most frequently mutated gene in human tumors, p63 and p73 are rarely mutated or deleted in cancers. Many studies have reported p63/p73 overexpression in human cancers while others showed that a loss of p63/p73 is associated with tumor progression and metastasis. Thus, whether p63 or p73 is a tumor suppressor gene or an oncogene has been a matter of debate. This controversy has been attributed to the existence of multiple splicing isoforms with distinct functions; the full-length TA isoform of p63 has structural and functional similarity to wild-type p53, whereas the ΔNp63 acts primarily in dominant-negative fashion against all family members of p53. Differential activities of TA and ΔN isoforms have been shown in vivo by creating isform-specific gene knockout mice. All p53, p63, p73 proteins bind to and activate target genes with p53-response elements; p63 also binds to distinct p63-response elements and regulate expression of specific target genes involved in skin, limb, and craniofacial development. Interestingly, several studies have shown that both p63 and p73 are involved in cellular response to cancer therapy and others have indicated that both of these molecules are required for p53-induced apoptosis, suggesting functional interplay among p53 family proteins. Consistent with these findings, aberrant splicing that result in ΔNp63 or ΔNp73 overexpression are frequently found in human cancers, and is associated with poor clinical outcomes of patients in the latter. Thus immunohistochemical staining of tumor specimen with ΔNp73-specific antibody might have diagnostic values in cancer clinics.

Cooperation between Dmp1 loss and cyclin D1 overexpression in breast cancer.

Cyclin D1 is a component of the core cell-cycle machinery and is frequently overexpressed in breast cancer. It physically interacts with the tumor suppressor Dmp1 that attenuates the oncogenic signals from Ras and HER2 by inducing Arf/p53-dependent cell-cycle arrest. Currently, the biological significance of Dmp1-cyclin D1 interplay in breast cancer has not been determined. Here, we show that cyclin D1 bound to Dmp1 to activate both Arf and Ink4a promoters and, consequently, induced apoptosis or G2/M cell-cycle delay in normal cells to protect them from neoplastic transformation. The cyclin D1-induced Ink4a/Arf gene expression was dependent on Dmp1 because the induction was not detected in Dmp1-deficient or DMP1-depleted cells. Arf/Ink4a expression was increased in pre-malignant mammary glands from Dmp1(+/+);MMTV-cyclin D1 and Dmp1(+/+);MMTV-D1T286A mice but significantly down-regulated in those from Dmp1-deficient mice. Selective Dmp1 deletion was found in 21% of the MMTV-D1 and D1T286A mammary carcinomas, and the Dmp1 heterozygous status significantly accelerated mouse mammary tumorigenesis with reduced apoptosis and increased metastasis. Overall, our study reveals a pivotal role of combined Dmp1 loss and cyclin D1 overexpression in breast cancer.

Mechanisms regulating DMTF1ß/γ expression and their functional interplay with DMTF1α.

The cyclin D binding myb­like transcription factor 1 (DMTF1), a haplo­insufficient tumor suppressor gene, has 3 alternatively spliced mRNA isoforms encoding DMTF1α, ß and γ proteins. Previous studies have indicated a tumor suppressive role of DMTF1α and the oncogenic activity of DMTF1ß, while the function of DMTF1γ remains largely undetermined. In the present study, the mechanisms regulating DMTF1 isoform expression were investigated and the functional interplay of DMTF1ß and γ with DMTF1α was characterized. It was found that specific regions of DMTF1ß and γ transcripts can impair their mRNA integrity or stability, and thus reduce protein expression levels. Additionally, DMTF1ß and γ proteins exhibited a reduced stability compared to DMTF1α and all 3 DMTF1 isoforms were localized in the nuclei. Two basic residues, K52 and R53, in the DMTF1 isoforms determined their nuclear localization. Importantly, both DMTF1ß and γ could associate with DMTF1α and antagonize its transactivation of the ARF promoter. Consistently, the ratios of both DMTF1ß/α and γ/α were significantly associated with a poor prognoses of breast cancer patients, suggesting oncogenic roles of DMTF1ß and γ isoforms in breast cancer development.

The oncogenicity of tumor-derived mutant p53 is enhanced by the recruitment of PLK3.

p53 mutations with single amino acid changes in cancer often lead to dominant oncogenic changes. Here, we have developed a mouse model of gain-of-function (GOF) p53-driven lung cancer utilizing conditionally active LSL p53-R172H and LSL K-Ras-G12D knock-in alleles that can be activated by Cre in lung club cells. Mutation of the p53 transactivation domain (TAD) (p53-L25Q/W26S/R172H) eliminating significant transactivation activity resulted in loss of tumorigenicity, demonstrating that transactivation mediated by or dependent on TAD is required for oncogenicity by GOF p53. GOF p53 TAD mutations significantly reduce phosphorylation of nearby p53 serine 20 (S20), which is a target for PLK3 phosphorylation. Knocking out PLK3 attenuated S20 phosphorylation along with transactivation and oncogenicity by GOF p53, indicating that GOF p53 exploits PLK3 to trigger its transactivation capability and exert oncogenic functions. Our data show a mechanistic involvement of PLK3 in mutant p53 pathway of oncogenesis.

The Arf-inducing transcription factor Dmp1 encodes a transcriptional activator of amphiregulin, thrombospondin-1, JunB and Egr1.

Dmp1 (Dmtf1) encodes a Myb-like transcription factor implicated in tumor suppression through direct activation of the Arf-p53 pathway. The human DMP1 gene is frequently deleted in non-small cell lung cancers, especially those that retain wild-type INK4a/ARF and/or p53. To identify novel genes that are regulated by Dmp1, transcriptional profiles of lung tissue from Dmp1-null and wild-type mice were generated using the GeneChip Microarray. Comparative analysis of gene expression changes between the two groups resulted in identification of numerous genes that may be regulated by Dmp1. Notably, amphiregulin (Areg), thrombospondin-1 (Tsp-1), JunB, Egr1, adrenomedullin (Adm), Bcl-3 and methyl-CpG binding domain protein 1 (Mbd1) were downregulated in the lungs from Dmp1-null mice while Gas1 and Ect2 genes were upregulated. These target genes were chosen for further analyses since they are involved in cell proliferation, transcription, angiogenesis/metastasis, apoptosis, or DNA methylation, and thus could account for the tumor suppressor phenotype of Dmp1. Dmp1 directly bound to the genomic loci of Areg, Tsp-1, JunB and Egr1. Significant upregulation or downregulation of the novel Dmp1 target genes was observed upon transient expression of Dmp1 in alveolar epithelial cells, an effect which was nullified by the inhibition of de novo mRNA synthesis. Interestingly, these genes and their protein products were significantly downregulated or upregulated in the lungs from Dmp1-heterozygous mice as well. Identification of novel Dmp1 target genes not only provides insights into the effects of Dmp1 on global gene expression, but also sheds light on the mechanism of haploid insufficiency of Dmp1 in tumor suppression.

Convergence of Itch-induced ubiquitination with MEKK1-JNK signaling in Th2 tolerance and airway inflammation.

The immune system is capable of mounting robust responses against invading pathogens but refrains from attacking self. Many studies have focused on tolerance induction of Th1 cells, whose failure results in development of autoimmune diseases. However, the molecular mechanisms governing tolerance induction in Th2 cells and its relation to allergic responses remain unclear. Here we used both in vivo and in vitro protocols to demonstrate that Th2 cells either containing a mitogen and extracellular kinase kinase 1 (MEKK1) mutant or lacking JNK1 or the E3 ubiquitin ligase Itch cannot be tolerized. In a mouse allergic model, injection of high-dose tolerizing antigen failed to block the development of airway inflammation in Itch-/- mice. This study suggests that MEKK1-JNK signaling regulates Itch E3 ligase-mediated tolerogenic process in Th2 cells. These findings have therapeutic implications for allergic diseases.

Ras-Raf-Arf signaling critically depends on the Dmp1 transcription factor.

Dmp1 prevents tumor formation by activating the Arf-p53 pathway. In cultured primary cells, the Dmp1 promoter was efficiently activated by oncogenic Ha-Ras(V12), but not by overexpressed c-Myc or E2F-1. Dmp1 promoter activation by Ras(V12) depended on Raf-MEK-ERK signaling. Induction of p19(Arf) and p21(Cip1) by oncogenic Raf was compromised in Dmp1-null cells, which were resistant to Raf-mediated premature senescence. A Ras(V12)-responsive element was mapped to the 5' leader sequence of the murine Dmp1 promoter, where endogenous Fos and Jun family proteins bind. Dmp1 promoter activation by Ras(V12) was strikingly impaired in c-Jun as well as in JunB knock-down cells, suggesting the critical role of Jun proteins in the activation of the Dmp1 promoter. A Ras(V12)-responsive element was mapped to the unique Dmp1/Ets site on the Arf promoter, where endogenous Dmp1 proteins bind upon oncogenic Raf activation. Therefore, activation of Arf by Ras/Raf signaling is indirectly mediated by Dmp1, explaining why Dmp1-null primary cells are highly susceptible to Ras-induced transformation. Our data indicate the presence of the novel Jun-Dmp1 pathway that directly links oncogenic Ras-Raf signaling and p19(Arf), independent of the classical cyclin D1/Cdk4-Rb-E2F pathway.

Aberrant expression of ETS1 and ETS2 proteins in cancer.

The ETS transcription factors regulate expression of genes involved in normal cell development, proliferation, differentiation, angiogenesis, and apoptosis, consisting of 28 family members in humans. Dysregulation of these transcription factors facilitates cell proliferation in cancers, and several members participate in invasion and metastasis by activating gene transcription. ETS1 and ETS2 are the founding members of the ETS family and regulate transcription by binding to ETS sequences. They are both involved in oncogenesis and tumor suppression depending on the biological situations used. The essential roles of ETS proteins in human telomere maintenance have been suggested, which have been linked to creation of new Ets binding sites. Recently, preferential binding of ETS2 to gain-of-function mutant p53 and ETS1 to wild type p53 (WTp53) has been suggested, raising the tumor promoting role for the former and tumor suppressive role for the latter. The oncogenic and tumor suppressive functions of ETS1 and 2 proteins have been discussed.

Translocations involving ETS family proteins in human cancer.

The ETS transcription factors regulate expression of genes involved in normal cell development, proliferation, differentiation, angiogenesis, and apoptosis, consisting of 28 family members in humans. Dysregulation of these transcription factors facilitates cell proliferation in cancers, and several members participate in invasion and metastasis by activating certain gene transcriptions. ETS1 and ETS2 are the founding members of the ETS family and regulate transcription by binding to ETS sequences. Three chimeric genes involving ETS genes have been identified in human cancers, which are EWS-FLI1 in Ewing's sarcoma, TMPRSS2-ERG in prostate cancer, and ETV6-RUNX1 in acute lymphocytic leukemia. Although these fusion transcripts definitely contribute to the pathogenesis of the disease, the impact of these fusion transcripts on patients' prognosis is highly controversial. In the present review, the roles of ETS protein translocations in human carcinogenesis are discussed.

Aberrant expression of p16INK4a in human cancers - a new biomarker?

The ARF and INK4a genes are located in the same CDKN2a locus, both showing its tumor suppressive activity. ARF has been shown to detect potentially harmful oncogenic signals, making incipient cancer cells undergo senescence or apoptosis. INK4a, on the other hand, responds to signals from aging in a variety of tissues including islets of Langerhans, neuronal cells, and cancer stem cells in general. It also detects oncogenic signals from incipient cancer cells to induce them senescent to prevent neoplastic transformation. Both of these genes are inactivated by gene deletion, promoter methylation, frame shift, and aberrant splicing although mutations changing the amino acid sequences affect only the latter. Recent studies indicated that polycomb gene products EZH2 and BMI1 repressed p16INK4a expression in primary cells, but not in cells deficient for pRB protein function. It was also reported that that p14ARF inhibits the stability of the p16INK4a protein in human cancer cell lines and mouse embryonic fibroblasts through its interaction with regenerating islet-derived protein 3γ. Overexpression of INK4a is associated with better prognosis of cancer when it is associated with human papilloma virus infection. However, it has a worse prognostic value in other tumors since it is an indicator of pRB loss. The p16INK4a tumor suppressive protein can thus be used as a biomarker to detect early stage cancer cells as well as advanced tumor cells with pRB inactivation since it is not expressed in normal cells.

Aberrant Expression of p14ARF in Human Cancers: A New Biomarker?

The ARF and INK4a genes are located on the CDKN2a locus, both showing tumor suppressive activity. ARF has been shown to monitor potentially harmful oncogenic signalings, making early stage cancer cells undergo senescence or programmed cell death to prevent cancer. Conversely, INK4a detects both aging and incipient cancer cell signals, and thus these two gene functions are different. The efficiency of detection of oncogenic signals is more efficient for the for the former than the latter in the mouse system. Both ARF and INK4a genes are inactivated by gene deletion, promoter methylation, frame shift, aberrant splicing although point mutations for the coding region affect only the latter. Recent studies show the splicing alterations that affect only ARF or both ARF and INK4a genes suggesting that ARF is inactivated in human tumors more frequently than what was previously thought. The ARF gene is activated by E2Fs and Dmp1 transcription factors while it is repressed by Bmi1, Tbx2/3, Twist1, and Pokemon nuclear proteins. It is also regulated at protein levels by Arf ubiquitin ligase named ULF, MKRN1, and Siva1. The prognostic value of ARF overexpression is controversial since it is induced in early stage cancer cells to eliminate pre-malignant cells (better prognosis); however, it may also indicate that the tumor cells have mutant p53 associated with worse prognosis. The ARF tumor suppressive protein can be used as a biomarker to detect early stage cancer cells as well as advanced stage tumors with p53 inactivation.