Alternative and aberrant splicing of MDM2 mRNA in human cancer.

MDM2 has been characterized as a protein that binds to and facilitates degradation of the tumor suppressor p53. Interestingly, more than 40 different splice variants of MDM2 transcripts have been identified both in tumors and normal tissues, and the majority of these variants do not contain sequence encoding the p53 binding site. This review describes the different splice forms, the tissues in which they have been identified, and their association with tumor progression and prognosis. In addition, we discuss the potential functions of these variants and how they interact with full-length MDM2 protein.

MDM2 splice variants predominantly localize to the nucleoplasm mediated by a COOH-terminal nuclear localization signal.

Of the >40 alternative and aberrant splice variants of MDM2 that have been described to date, the majority has lost both the well-characterized nuclear localization signal (NLS1) and the nuclear export signal (NES) sequence. Because cellular localization of proteins provides insight regarding their potential function, we determined the localization of three different MDM2 splice variants. The splice variants chosen were the common variants MDM2-A and MDM2-B. In addition, MDM2-FB26 was chosen because it is one of the few variants described that contains the complete p53-binding site. All three splice variants predominantly localized to the nucleus. Nuclear localization of MDM2-A and MDM2-B was controlled by a previously uncharacterized nuclear localization signal (NLS2), whereas nucleoplasmic localization of MDM2-FB26 was mediated by NLS1. p53 and full-length MDM2 colocalized with the splice variants in the nucleus. MDM2-A and MDM2-B both contain a COOH-terminal RING finger domain, and interaction with full-length MDM2 through this domain was confirmed. MDM2-FB26 was the only splice variant evaluated that contained a p53-binding domain; however, interaction between MDM2-FB26 and p53 could not be shown. p14(ARF) did not colocalize with the splice variants and was predominantly expressed within the nucleoli. In summary, nuclear localization signals responsible for the nucleoplasmic distribution of MDM2 splice variants have been characterized. Colocalization and interaction of MDM2-A and MDM2-B with full-length MDM2 in the nucleus have important physiologic consequences, for example, deregulation of p53 activity.

Differential cooperation of oncogenes with p53 and Bax to induce apoptosis in rhabdomyosarcoma.

BACKGROUND: Deregulated expression of oncogenes such as MYC and PAX3-FKHR often occurs in rhabdomyosarcomas. MYC can enhance cell proliferation and apoptosis under specific conditions, whereas PAX3-FKHR has only been described as anti-apoptotic. RESULTS: In order to evaluate how MYC and PAX3-FKHR oncogenes influenced p53-mediated apoptosis, rhabdomyosarcoma cells were developed to independently express MYC and PAX3-FKHR cDNAs. Exogenous wild-type p53 expression in MYC transfected cells resulted in apoptosis, whereas there was only a slight effect in those transfected with PAX3-FKHR. Both oncoproteins induced BAX, but BAX induction alone without expression of wild-type p53 was insufficient to induce apoptosis. Data generated from genetically modified MEFs suggested that expression of all three proteins; MYC, BAX and p53, was required for maximal cell death to occur. CONCLUSION: We conclude that cooperation between p53 and oncoproteins to induce apoptosis is dependent upon the specific oncoprotein expressed and that oncogene-mediated induction of BAX is necessary but insufficient to enhance p53-mediated apoptosis. These data demonstrate a novel relationship between MYC and p53-dependent apoptosis, independent of the ability of MYC to induce p53 that may be important in transformed cells other than rhabdomyosarcoma.

P21Waf1/Cip1 dysfunction in neuroblastoma: a novel mechanism of attenuating G0-G1 cell cycle arrest.

In normal cells in which DNA has been damaged, p53 induces the expression of p21(Waf1/Cip1); p21, in turn, binds to cyclin-dependent kinase 2 (cdk2) and inhibits its function. Inhibition of cdk2 results in cell cycle arrest in G(0)-G(1). Although p53 is transcriptionally active and induces p21 expression in neuroblastoma (NB) cells, the G(0)-G(1) checkpoint is attenuated. Here we report that the mechanism that mediates this defect in NB cells is the inability of p21 to bind to, or inhibit the activity of cdk2. However, when recombinant p21 protein was added to NB cell extracts in vitro, the protein inhibited the activity of cdk2. This finding suggests that endogenous p21 protein in NB cells is inactive and may be bound either to a protein complex or in a conformation that precludes its binding to cdk2. The dysfunction of p21 in NB cells represents a novel mechanism by which the G(0)-G(1) cell cycle checkpoint can be inactivated. This mechanism may be important in regulating the growth of NB and potentially other types of tumors. Cdk inhibitors currently being developed for clinical use may be useful therapy for tumors such as NB in which endogenous cdk inhibitors are defective.

Efficacy and toxicity of a virus-directed enzyme prodrug therapy purging method: preclinical assessment and application to bone marrow samples from neuroblastoma patients.

Autologous stem cell transplantation is used to rescue cancer patients from myelosuppression caused by high-dose chemotherapy. However, autologous grafts often contain tumor cells that can contribute directly to relapse. Current purging methods are useful when fewer than 1% tumor cells contaminate the bone marrow, and patients with tumor burdens of >1% are considered ineligible for chemotherapy that necessitates stem cell rescue. Using neuroblastoma (NB) as a model system, we developed a method that is effective even with tumor burdens of 10-25%. Mixtures of NB-1691 NB cells and CD34(+) hematopoietic cells purged by this method showed no evidence of viable tumor cells as assessed by clonogenic assays or reverse transcription-PCR for the NB cell markers tyrosine hydroxylase and N-MYC. The efficacy and lack of toxicity of the method were verified using in vivo mouse models. Severe combined immunodeficient mice that received purged cell preparations containing 10% NB-1691 cells survived without evidence of disease for the observation period (>1 year), whereas mice that received unpurged cells developed disseminated disease requiring euthanasia 73-96 days after injection of cells. No evidence of toxicity to the mice was detected by numerous laboratory values for bone marrow, liver, and kidney function, and no difference was seen in the ability of purged cell mixtures versus unmanipulated CD34(+) cells to reconstitute the marrow of non-obese diabetic severe combined immunodeficient mice. In a pilot study, marrow was obtained from eight patients who had >/=1% metastatic tumor burden. All eight samples were purged to the level of detection by reverse transcription-PCR (samples from seven patients) or clonogenic potential (sample from one patient), whichever assay was used. The described adenovirus/rabbit carboxylesterase/CPT-11 (irinotecan, 7-ethyl-10[4-(1-piperidino)-1-piperidino]carbonyloxycamptothecin) virus-directed enzyme prodrug method may be useful for patients whose tumor burdens exceed 1% at the time of stem cell harvest. Assessment of purging efficacy with additional samples from NB patients is ongoing.

Epigenetic regulation of O6-methylguanine-DNA methyltransferase gene expression by histone acetylation and methyl-CpG binding proteins.

Transcriptional silencing of the DNA repair gene, O6-methylguanine-DNA methyltransferase (MGMT) in a proportion of transformed cell lines is associated with methylated CpG hotspots in the MGMT 5' flank. The goal of the study was to evaluate the mechanism by which CpG methylation of theMGMT promoter region influenced silencing of the gene. Analysis of histone acetylation status in two regions of the promoter using chromatin immunoprecipitation assay showed that a higher level of histone acetylation was associated with expression in three MGMT-expressing cell lines (HeLa CCL2, HT29, and Raji) compared with three MGMT-silenced cell lines (HeLa S3, BE, and TK6). To determine how the modulation of CpG methylation and histone acetylation influenced MGMT expression, we exposed the cells to 5-aza-2'deoxycytidine (5-Aza-dC), inhibitor of DNA methylation, which strongly up-regulated MGMT expression in three MGMT-silenced cell lines whereas trichostatin A, inhibitor of histone deacetylase, weakly induced MGMT. However, combined treatment with 5-Aza-dC and trichostatin A significantly up-regulated MGMT RNA expression to a greater extent than in cells treated with either agent alone suggesting that histone deacetylation plays a role in MGMT silencing but that CpG methylation has a dominant effect. Consistent with enhanced MGMT expression, 5-Aza-dC increased the association of acetylated histone H3 and H4 bound to the MGMT promoter. Chromatin immunoprecipitation analysis of methyl-CpG binding domain containing proteins detected a greater amount of MeCP2, MBD1, and CAF-1 bound to the MGMT promoter in MGMT-silenced cells. Our findings implicate specific MBD proteins in methylation-mediated transcriptional silencing of MGMT.

MDM2 and its splice variant messenger RNAs: expression in tumors and down-regulation using antisense oligonucleotides.

Alternative splicing has an important role in expanding protein diversity. An example of a gene with more than one transcript is the MDM2 oncogene. To date, more than 40 different splice variants have been isolated from both tumor and normal tissues. Here, we review what is known about the alteration of MDM2 mRNA expression, focusing on alternative splicing and potential functions of different MDM2 isoforms. We also discuss the progress that has been made in the development of antisense oligonucleotides targeted to MDM2 for use as a potential cancer therapy.

MDM2 does not influence p53-mediated sensitivity to DNA-damaging drugs.

MDM2 inhibits transactivation properties of the tumor suppressor protein p53 by binding to and facilitating proteasomal degradation of p53. Because MDM2 targets p53 for degradation, it was anticipated that cells that overexpress MDM2 would not contain functional wild-type p53 (wtp53). However, p53 and MDM2 in cells with damaged DNA can become phosphorylated, and their binding to each other can become inhibited. Thus, p53 remains functional and induces apoptosis of damaged cells. Here we report the results of experiments designed to investigate whether MDM2 amplification and overexpression can inhibit p53-mediated chemosensitivity to DNA-damaging drugs. Two cell lines in which MDM2 is amplified, NB-1691 and Rh18, were transduced with an adenoviral expression vector for p53 (Ad.p53). Although functional wtp53 was detected, no change in chemosensitivity was observed, suggesting that endogenous wtp53 may have been active in the MDM2-amplified cells. The adenoviral vector Ad.MDM2 was used to generate MDM2 expression in a rhabdomyosarcoma cell line, Rh30-CI.27, engineered to express inducible wtp53. When p53 expression was induced, cells became chemosensitive to actinomycin D in the presence or absence of MDM2 expression; this result suggests that MDM2 cannot inhibit p53-mediated chemosensitivity. There was no evidence of a reduced amount of MDM2-p53 binding after drug exposure, but the remaining unbound wtp53 may be functional and capable of potentiating cytotoxicity. In conclusion, MDM2 expression is important in inhibiting p53 function during tumor development but not during the DNA damage-mediated cytotoxic response.

MDM2-A, a common Mdm2 splice variant, causes perinatal lethality, reduced longevity and enhanced senescence.

MDM2 is the predominant negative regulator of p53 that functions to maintain the appropriate level of expression and activity of this central tumor suppressor. Mdm2-a is a commonly identified splice variant of Mdm2; however, its physiological function is unclear. To gain insight into the activity of MDM2-A and its potential impact on p53, an Mdm2-a transgenic mouse model was generated. Mdm2-a transgenic mice displayed a homozygous-lethal phenotype that could be rescued by a reduction in p53 expression, demonstrating a dependence upon p53. Mdm2-a hemizygous mice exhibited reduced longevity, and enhanced senescence was observed in their salivary glands. In addition, the transgenic mice lacked typical, accelerated aging phenotypes. Growth of transgenic mouse embryonic fibroblasts (MEFs) was inhibited relative to wild-type MEFs, and MDM2-A was shown to bind to full-length MDM2 in an interaction that could increase p53 activity via reduced MDM2 inhibition. Evidence of p53 activation was shown in the Mdm2-a transgenic MEFs, including p53-dependent growth inhibition and elevated expression of the p53 target protein p21. In addition, MDM2-A increased senescence in a p21-independent manner. In conclusion, unexpected roles for MDM2-A in longevity and senescence were identified in a transgenic mouse model, suggesting that Mdm2 splice variants might be determinants of these phenotypes in vivo.

The MDM2-a splice variant of MDM2 alters transformation in vitro and the tumor spectrum in both Arf- and p53-null models of tumorigenesis.

MDM2-A is a common splice variant of murine double minute 2 (MDM2) that is frequently detected in many tumor types. Our previous work has characterized MDM2-A as an activator of p53, and therefore, in a wild-type p53 background, this splice variant would be predicted to confer p53-dependent tumor protection. To test this hypothesis, we used Mdm2-a transgenic mice to assess transformation and tumorigenesis in tumor susceptible murine models. A MDM2-A-dependent decrease in transformation was observed in Arf-null mouse embryonic fibroblasts (MEF) or when wild-type MEFs were exposed to the carcinogen ethylnitrosourea. However, this reduced transformation did not confer tumor protection in vivo; Mdm2-a/Arf-null mice and ethylnitrosourea-treated MDM2-expressing mice developed similar tumor types with equivalent latency compared with their respective controls. Interestingly, when p53 was deleted, MDM2-A expression enhanced transformation of p53-null MEFs and altered tumor spectrum in vivo. In addition, p53-heterozygous mice that expressed MDM2-A developed aggressive mammary tumors that were not observed in p53-heterozygous controls. In conclusion, we found that although MDM2-A expression enhances p53 activity and decreases transformation in vitro, it cannot confer tumor protection. In contrast, MDM2-A seems to exhibit a novel transforming potential in cells where p53 function is compromised. These data show that MDM2 splice variants, such as MDM2-A, may provide protection against transformation of normal tissues having intact p53. However, when such splice variants are expressed in tumors that have defects in the p53 pathway, these isoforms may contribute to tumor progression, which could explain why their expression is often associated with aggressive tumor types.

MDM2 displays differential activities dependent upon the activation status of NFkappaB.

MDM2 is an oncoprotein best characterized for its role in the inactivation and degradation of the p53 tumor suppressor. However, MDM2 has many other binding partners and its p53-independent role in the regulation of cell growth and survival appears to be extremely complex. This report describes the expression of MDM2 in two rhabdomyosarcoma cell lines, both expressing a mutant p53 gene. Expression of MDM2 in Rh30 cells enhanced cell growth whereas expression of MDM2 in RD cells suppressed their growth and enhanced the rate of spontaneous apoptosis. The mechanism for these opposite phenotypes was demonstrated to be due to differential effects on the NFkappaB pathway. Previously MDM2 has been shown to activate NFkappaB through activation of transcription of the p65RelA subunit. In Rh30 cells MDM2 acted similarly to previously described, thereby promoting growth of Rh30 cells. In untreated RD cells p65RelA was constitutively overexpressed resulting in activation of the NFkappaB pathway. Expression of MDM2 in RD cells transcriptionally repressed p65RelA and suppressed NFkappaB activity, resulting in a reduced growth rate and enhanced apoptosis. The MDM2-sensitive region of the p65 promoter was localized to a 225 bp fragment to which MDM2 protein was shown to bind. The observation that MDM2 induces apoptosis under certain circumstances may help to explain the apparently surprising clinical studies that have shown that MDM2 expression in tumors is often associated with a favorable prognosis.

Selection for mutations in the cDNAs of transgenic mice upon expression of an embryonic lethal protein.

The generation of transgenic mouse models to study in vivo functions of specific proteins has become common practice. In addition, PCR technology allows efficient and rapid identification of founder mice by the analysis of tail tip DNA. Whilst the DNA construct used in the microinjection of one-cell-stage embryos is usually sequenced it is not common practice to sequence the PCR product once the transgene has been inserted into the mouse genome. In this report we describe why sequencing of inserted transgenes is important. Upon generation of transgenic mice expressing a splice variant of MDM2, MDM2-A, three of four founders contained mutations within the Mdm2-a cDNA sequence. The observation that selection against expression of wild-type MDM2-A resulted in the generation of mice expressing mutant transgenes highlights the importance of sequencing the transgenes of founder mice.

Selective chemosensitization of rhabdomyosarcoma cell lines following wild-type p53 adenoviral transduction.

Rhabdomyosarcoma (RMS) cell lines were transduced with an adenoviral vector containing the wild-type p53 (wtp53) cDNA (Ad-p53) and then exposed to four cytotoxic agents: actinomycin D, vincristine, 5-fluorouracil and bleomycin. Potentiation of cytotoxicity following wild-type p53 expression varied from 0- to 20-fold for different drugs and between cell lines. It appeared that alveolar RMS cells (n = 2) were more susceptible to p53-mediated chemosensitization than embryonal RMS cells (n = 3), although this was independent of pax3-FKHR expression. Overall, cells that were most chemosensitive prior to Ad-p53 exposure were those that were most susceptible to p53 potentiation of cytotoxicity. The different results obtained with these RMS cell lines does not appear to be related to expression of pax3-FKHR, p21, Bax or Bcl-2 but may in part be due to differential regulation of p53 target genes, such as MDM2. In conclusion, exogenous wild-type expression selectively chemosensitizes RMS cells to cytotoxic agents. However, expression of transcriptionally active wtp53 does not predict a chemosensitive phenotype.

p53-mediated regulation of expression of a rabbit liver carboxylesterase confers sensitivity to 7-ethyl-10-[4-(1-piperidino)-1-piperidino]carbonyloxycamptothecin (CPT-11).

We have exploited the ability of wild-type (wt) p53 to repress gene expression and produce tumor-selective cytotoxicity using viral-directed enzyme prodrug therapy. Vectors containing either the cytomegalovirus or Rous sarcoma virus promoter regulating transcription of a rabbit liver carboxylesterase (CE) have been constructed. Upon transfection of these plasmids into cells expressing either wt or mutant p53, differential expression of the CE has been observed, resulting in sensitization of the cells expressing the latter protein to the anticancer prodrug irinotecan, 7-ethyl-10-[4-(1-piperidino)-1-piperidino] carb- onyloxycamptothecin (CPT-11). Transduction of isogenic cell lines with adenovirus containing CE under control of the Rous sarcoma virus promoter confirmed the decreased sensitization of cells expressing wtp53 to CPT-11. These studies indicate that the inactivation of wtp53 by mutant p53 in human tumor cells may be sufficient enough to generate a therapeutic window for enhanced cytotoxicity with CPT-11.