Wild-Type p53 Overexpression and Its Correlation With MDM2 and p14ARF Alterations: An Alternative Pathway to Non–Small-Cell Lung Cancer
http://www.100md.com
《临床肿瘤学》
the Department of Life Sciences, National Taiwan Normal University, Taipei
Department of Pathology, and Division of Thoracic Surgery Taichung Veterans General Hospital, Taichung, Taiwan, Republic of China
ABSTRACT
PATIENTS AND METHODS: We performed gene and protein alteration studies on p53 and its upstream effectors, MDM2 and p14ARF, in tumors from 94 non–small-cell lung cancer (NSCLC) patients.
RESULTS: Immunohistochemical and sequencing analyses indicated that 37 tumors showed overexpression of wild-type p53. An absence of nuclear staining of MDM2 protein was found in 95% of these tumors (35 of 37; P < .001). The tumors with negative MDM2 staining showed a significantly high concordance of loss of Akt activity and low MDM2 mRNA expression (P < .001). Sequencing analysis revealed five distinct MDM2 splicing variants disrupting the conserved p53 binding domain. Corresponding variant proteins were detected in three lung cancer cell lines using the Western blot analysis. Our results also indicated that among the tumors with overexpression of the wild-type p53, 92% (34 of 37) showed immunoreactivity to p14ARF (P = .001). In addition, the deregulation of p53 and MDM2 genes was significantly associated with squamous lung cancer (P < .05) and was correlated with advanced stages (P < .05) and poor prognosis (P < .05).
CONCLUSION: Our data suggest that immunopositivity of p14ARF together with a low expression of MDM2 contributes to accumulation of the wild-type p53, and that deregulation of the p53-MDM2-p14ARF pathway is important in the pathogenesis and outcome of a subset of NSCLC.
INTRODUCTION
MDM2 was found to be amplified in a portion of human cancers,13 but amplification in malignant epithelial-derived cancers, including non–small-cell lung cancer (NSCLC), occurs at lower frequencies (approximately 5% to 15%).13-15 MDM2 physically interacts through its amino-terminal domain with p53,9 and through its E3 ubiquitin ligase activity MDM2 transfers monoubiquitin tags onto the lysine residues in p53,9 which leads to the degradation of the p53 protein. Interestingly, experimental evidence also indicates that MDM2 overexpression causes G1 arrest in normal cells, and two cell-cycle inhibitory domains were identified in the middle of MDM2 protein.16,17 These data suggest that MDM2 possibly has a dual role (ie, tumorigenesis and growth arrest) in regulating cell growth. In addition, alternative splicing of MDM2 and the generation of short proteins occurs in many human and mouse tumors.18,19 However, reports on MDM2-aberrant splicing in NSCLC are scarce.20 More recently, it was also shown that mitogen-induced Akt physically associates with and phosphorylates MDM2, leading to an enhanced activity of MDM2 and increased p53 degradation.21,22
The INK4a/ARF locus on human chromosome region 9p21 encodes two distinct proteins, which are translated in different reading frames from alternatively spliced transcripts. p14ARF is an upstream regulator in the p53 tumor suppressor pathways. The p14ARF tumor suppressor, in response to oncogenic and hyperproliferative signals as well as signals of DNA damage, binds to MDM2 and consequently activates p53 by blocking p53 and MDM2 nuclear export and p53 degradation.5,6,23,24 Thus, p14ARF overexpression can lead to the stabilization of p53. It has also been reported that the p14ARF-mediated G1 and G2 arrests are abolished in mouse embryonic fibroblasts lacking functional p53, indicating that p53 also acts downstream of p14ARF in the cell cycle regulatory pathway.25,26
To investigate the role of the p53 tumor suppressor gene in lung tumorigenesis, we examined the frequency and mutation spectra of p53 tumor suppressor gene in 60 lung cancer patients. We found that a relatively reduced frequency of p53 mutation with a much greater frequency of p53 protein overexpression reflected that the p53 protein stabilized in the absence of p53 gene mutation.25 To confirm this p53 immunohistochemical abnormality, specimens of resected NSCLC from another set of 94 NSCLC patients were collected and were examined for p53 mutation spectrum and p53 protein expression in this study. In addition, to test the possibility that p53 protein accumulation may result from alternative mechanisms leading to p53 protein stabilization, we performed alteration analyses of the p53 upstream proteins, p14ARF and MDM2, in these patients. The data indicated that in 92% of patients (35 of 38) with wild-type p53 overexpression, MDM2 expression was decreased and p14ARF expression was correspondingly prominent. In addition, the alteration of p53 and MDM2 genes was associated with lung cancers of more advanced stages and poorer prognosis.
PATIENTS AND METHODS
Surgically resected tumor samples were immediately snap-frozen, and subsequently stored in liquid nitrogen. For the RNA expression assay and cDNA sequencing, the total RNA was prepared from matched pairs of primary tumors and nearby normal lung tissues using Trizol reagent (Invitrogen, Carlsbad, CA). cDNA was synthesized using SuperScript reverse transcriptase (Invitrogen) with the protocols provided by the manufacturer.
Analysis of Protein Expression: Immunohistochemistry Assay
The monoclonal antibody, DO-7 (1:150; DAKO, Glostrup, Denmark), which recognizes both wild-type and mutant-type of p53 proteins, was used as the primary antibody to detect p53 protein expression. Another monoclonal antibody, p53/Ab-5 (1:20; Oncogene, Boston, MA), which recognizes only the wild-type of p53 protein, was later used to identify the status of overexpressed p53. The monoclonal antibodies IF-2 (1:100; Zymed, San Francisco, CA) and N-20 (1:50; Santa Cruz Biotechnology, Santa Cruz, CA) were used as the primary antibodies to detect the central domain and amino-terminal domain of MDM2, respectively. The polyclonal antibody, phospho-Akt (Ser 473; 1:50; Cell Signaling Technology, Beverly, MA), was used to detect the active form of Akt kinase. The evaluation of the immunohistochemistry was conducted blindly, without knowledge of the clinical and pathologic characteristics of the patients. If the positive immunostaining was observed in the positive control cells, the patient case was judged for the immunostaining of the tumor cells. Staining for p53 was scored 3 if more than 50% tumor cells were immunostained positive; 2 if 26% to 50% of cells were positive; 1 if 11% to 25% of cells were positive; and 0 if less than 10% of cells were positive. The stains were graded p53 overexpression if the score was 2 or 3.26 Staining for MDM2 was 3 if more than 50% tumor cells were positive; 2 if 11% to 50% of cells were positive; 1 if 1% to 10% of cells were positive; and 0 if 0% of cells were positive. The stains were graded as exhibiting overexpression if the score was 3 and graded as negative when there was complete absence of staining in the tumor cells.14 Staining for Akt was scored 3 if more than 30% of the tumor cells were immunostaining positive; 2 if 11% to 30% of cells were positive; 1 if 1% to 10% of cells were positive; and 0 if 0% cells were positive. Staining of 3 was evaluated as overexpression.27 The antibody and evaluation used for p14ARF were described previously.28
Mutation Spectrum Analysis for p53 Gene
Tumors were analyzed for DNA sequence mutations in exons 5 to 11 of p53 tumor suppressor gene as describe previously.25 Polymerase chain reaction (PCR) fragments were purified with Qiaex II Gel Extraction Kit (Qiagen, Valencia, CA), then sequenced using ABI 377 Automatic Sequencer (PE Applied Biosystems, Foster City, CA). Sequencing analysis of the opposite strand for all of the mutations were repeated at least once using independent PCR products.
Analysis of mRNA Expression of MDM2 and p14ARF Genes: Multiplex Reverse Transcriptase Polymerase Chain Reaction and cDNA Sequencing
mRNA expression was assayed in a multiplex reverse transcriptase polymerase chain reaction (RT-PCR) analysis using the -actin gene as an internal control. The exons 1 to 2 of the p14ARF gene and the -actin gene were amplified using primers described by Gazzeri et al.29 The exons 3 to 12 of the MDM2 gene were amplified using a nested PCR described by Sigalas et al.19 To quantify the relative levels of gene expression in the multiplex RT-PCR assay, the value for the internal standard (-actin) in each test tube was used as the baseline value for gene expression in that sample, and a relative value was calculated for each target transcript amplified from each tumor and matched normal sample. Tumor cells that exhibited mRNA expression below 50% of that of normal cells were deemed to have a low expression pattern.28 Abnormal cDNA fragments detected in RT-PCR assays of MDM2 gene were eluted, purified, and sequenced as described above.
Analysis of Genomic Sequence of MDM2 Gene: Multiplex PCR
The copy number of genomic sequence of MDM2 gene was assayed in a multiplex PCR analysis using the promoter sequence of GAPDH gene as an internal control. The exons 5 and 9 of MDM2 gene were amplified using the primers to their intron sequences. The primer nucleotide sequences were as follows: for the exon 5 of MDM2 gene, sense-5'-ATCTCCTGACCTCGTGATCCAT-3', antisense-5'-AGATGCCAGAGCTCAG GTTCTCA-3'; for the exon 9 of MDM2 gene, sense-5'-TTCTGCTGTAACAGTTGGACA-3', antisense-5'-TTTGACTGTACTATTGGTGCAGT-3'; for the GAPDH gene, sense-5'-AATGAAAGGCACACTGTCTCTCTC-3', antisense-5'-GTTTCTGCACGGAAGGTCAC-3'. PCR was performed for 35 cycles with the annealing temperature of 65°C. The PCR products were then electrophoresed, stained with ethidium bromide, and quantified for the levels relative to the control sequence.
Western Blot Analysis
Total cell protein (50 μg) was subjected to electrophoresis using an 8% sodium dodecyl sulfate polyacrylamide gel. Subsequently, proteins were transferred to a nitrocellulose membrane (Millipore, Eschborn, Germany) at 200 mA for 1.5 hours using a semidry blotting chamber. Nonspecific binding sites were blocked by 4% low-fat dry milk in Tris-buffered saline and 0.1% Tween-20 for 1 hour at room temperature. Immunodetection of MDM2 was performed using N-20 antibodies (dilution 1:200) at room temperature for 2 hours. After incubation with a horseradish peroxidase–conjugated secondary antibody, immunoreactive proteins were visualized by Western blotting luminol reagent (Santa Cruz Biotechnology). Signal detection was performed using the SuperSignal West Pico Chemiluminescent Substrate (Pierce Biotechnology, Rockford, IL) according to the manufacturer's instructions. Control -actin antibody was purchased from Santa Cruz Biotechnology, Inc.
Statistical Analysis
Pearson's {chi}2 test was used to compare the frequency of p53, MDM2, and p14ARF alterations in NSCLC patients with different characteristics, including sex and smoking status, and various clinicopathologic parameters, such as tumor type and tumor stage. The comparison of age distributions between patients with and without alterations was analyzed with a two-sample t test. Censoring was performed on patients who were still alive or had died as a result of some other cause at the end of the study for the CSS analysis. Censored patients were those who had not yet experienced the recurrence by the time of the follow-up or patients who died without relapse and were censored at the date of death for the DFS analysis. The Kaplan-Meier method was used to estimate the probability of survival as a function of time and the median survival. The log-rank test was used to assess the significance of the difference between pairs of survival probabilities. In addition, we analyzed prognosis significance of gene alterations by stratifying the various clinicopathologic parameters of patients. Multivariate Cox regression analysis using the likelihood ratio test was done to evaluate the influence of clinical stage on the effect of gene alterations in the CSS and DFS. Statistical analysis was performed using Statistical Package for Social Sciences software (SPSS, Chicago, IL). P < .05 was considered to be statistically significant.
RESULTS
MDM2 Alterations and Akt Inactivation in Primary NSCLC Tumors
We therefore tested the hypothesis that p53 stabilization in the wild-type tumors may be through deregulation of the upstream effectors of p53 (ie, MDM2 and/or p14ARF). Immunohistochemical staining of MDM2 protein was performed on 94 tumor samples using the IF-2 antibody directed to its central domain. Fifty-two lung tumors showed staining of moderate to strong intensity within the nucleus of tumor cells (Fig 2A). The remaining 42 tumors (45%) showed an absence of nuclear staining of MDM2 protein (Fig 2B; Table 1). Among the tumors with overexpression of wild-type p53, 95% (35 of 37) showed a negative expression of MDM2.
Given that expression of active Akt kinase promotes activity of MDM2, we hypothesized that diminishing levels of MDM2 may reflect the decrease of Akt activity, and we then immunostained the active form of Akt using an antibody specific to the phosphorylated Akt. The data indicated that 48 tumors that showed positive MDM2 staining also showed positive staining for active Akt (Fig 2D). There were 36 tumors that had loss of expression of both proteins, suggesting that low MDM2 expression correlated with low Akt activity (Fig 2E). A high concordance (84 of 94; 89%) was observed between MDM2 and Akt expression (P = .001).
Analysis of a variety of human tumors has revealed the presence of multiple, alternatively spliced forms of MDM2 messages, which leads to the inactivation of MDM2. Therefore, we analyzed mRNA expression level and isoforms of MDM2 transcripts with semiquantitative multiplex RT-PCR and cDNA sequencing, respectively (Fig 3). Decreased MDM2 full-length transcripts were shown by semiquantitative multiplex RT-PCR to occur in 39 tumors (41%; Fig 3A; Table 1). The correlation between MDM2 protein and mRNA expressions showed a 91% concordance (86 of 94). Negative MDM2 protein expression was significantly associated with low MDM2 mRNA expression (P = .001). The RT-PCR also revealed that 25 of 39 tumors with low levels of full-length MDM2 transcript had shown multiple smaller transcripts. A total of seven distinct variants were found: Del.a, 1,016 bp, 10 cases; Del.b, 941 bp, five cases; Del.c, 707 bp, 11 cases; Del.d, 542 bp, five cases; Del.e, 449 bp, three cases; Del.f, 303 bp, two cases; and Del.g, 274 bp, two cases (Fig 3B). Sequence analysis was carried out to determine the composition of the abnormally sized MDM2 transcripts. The 542-bp product has not been reported in the literature and its sequencing data showed a lack of exons 4 to 11, as well as a lack of the first 165 base pairs of exon 12, leaving a residual exon 12 starting from nucleotide 1066 (Fig 3C). The sequence of nucleotides 1064 and 1065 is AG, which may serve as a cryptic 3' intron junction during splicing.
To examine whether aberrant transcripts resulted from deletions in the genomic MDM2 gene, we analyzed genomic DNA samples from a subset of 16 patients with aberrantly sized transcripts for the deletion of exons 5 and 9, which are the exons frequently missing in the mRNA, using a multiplex PCR assay (Fig 3D). None of the genomic DNA samples showed a deletion of either exon 5 or 9, thus providing evidence that the aberrant MDM2 transcripts may not be caused by a deletion of the genomic sequence of MDM2. The data suggested that the frequent lack of exons 5 to 9 in MDM2 cDNA is predominantly due to splicing abnormalities.
To determine whether the splice variants express the truncated proteins, three lung cancer cell lines that showed aberrant mRNA expression were analyzed for protein expression using the Western blot assay with the N-20 antibody directed to the amino-terminal of the MDM2 protein that should detect most described MDM2 protein variants. As shown in Figure 4, the MDM2 proteins (p75, p57, and p40) were found to be expressed more abundantly than the wild-type p90 protein in these lung cancer cell lines. These proteins probably correspond with MDM2 variants from transcripts of 1016, 941, and 707 bp, respectively. These variant proteins with a loss of central regions, including the epitope of IF-2 antibody, showed no signal detecting by IF-2 antibody with same exposure time (data not shown).
p14ARF Protein and mRNA Expression in Primary NSCLC Tumors
It has been shown that p14ARF inhibits p53 and MDM2 binding, and thereby maintains p53 at a high level. Immunohistochemical staining was therefore performed for p14ARF protein on 94 tumor samples. Sixty-one lung tumors showed staining of moderate to strong intensity. The remaining 33 lung cancers (35%) showed aberrant expression of p14ARF protein, characterized by a low level or complete absence of nuclear staining (Table 1). Among the tumors with overexpression of wild-type p53, 92% (34 of 37) showed a positive expression of p14ARF.
Decreased p14ARF transcripts were shown by semiquantitative multiplex RT-PCR to occur in 29 tumors (31%; Table 1). The correlation between p14ARF protein and mRNA expressions showed an 87% concordance (82 of 94). Negative protein expression was significantly associated with low mRNA expression (P = .001).
Concordance Analysis of p53, MDM2, and p14ARF Alterations
The data on the protein analysis of p53, MDM2, and p14ARF were cross-referenced to investigate the correlations between these three genes (Table 1). A striking discovery was made about the association between the overexpression of p53 protein with low expression of MDM2 protein and positive expression of the p14ARF protein. p53 overexpression inversely correlated with low MDM2 expression (P < .001) and frequently was accompanied by the expression of the p14ARF protein (P = .001). In addition, positive p14ARF protein expression was significantly associated with negative MDM2 expression (P = .001). Overall, 34 of 37 patients with p53 overexpression showed absence or low expression of MDM2 protein and moderate to strong expression of p14ARF protein.
Alterations of p53, MDM2, and p14ARF Genes and Their Correlation With Clinical Parameters and Prognoses of NSCLC Patients
To examine whether there were associations of alterations at the p53, MDM2, and p14ARF genes with the clinical characteristics of patients, the occurrence of alteration at each gene was compared with the patients' clinicopathologic parameters, including sex, age, smoking habit, tumor type, and tumor stage (Table 1). {chi}2 analysis showed that p53 mutation and protein overexpression occurred primarily in patients suffering from SQ types of lung cancer (P = .003 and P = .038, respectively). p53 overexpression was significantly restricted to patients suffering from stage III or IV lung cancers (P = .040). Low MDM2 gene expression also appeared to be more frequent in SQ than in AD, and the statistical significance of this difference was found for the protein expression data (P = .033). There was a significant trend for low expression of MDM2 in patients with advanced stages; the incidence of low mRNA expression of MDM2 gene was statistically associated with stage III and IV patients (P = .013). In addition, a highly significant correlation between low p14ARF expression and AD patients was found (P = .016; Table 1).
The relationship between postoperative survival and the altered protein expression of the p53 and MDM2 was analyzed in 94 patients. p53 protein overexpression was statistically associated with a poorer prognosis for CSS (P = .042; Fig 5A) and for DFS (P = .029; Fig 5B) even after data were adjusted for tumor staging factor. There was a trend for shorter DFS in patients with low MDM2 protein expression when compared with those with positive protein expression, and in patients with altered MDM2 splicing variants when compared with those without splicing variants. However, this did not reach statistical significance. Nevertheless, we found that patients with altered MDM2 expression had worse prognoses compared with those without, especially in AD lung cancer (P = .037 for CSS; P = .010 for DFS; Figs 5C and 5D, respectively). In the patients with early-stage lung cancer, in whom identifying unfavorable prognostic factors is important, patients with alternative splicing of MDM2 mRNA had significantly worse prognoses than those without (P = .020 for CSS; P = .029 for DFS; Figs 5E and 5F, respectively). To test the hypothesis of an increased correlation of cancer prognosis when more than one disease-related gene was considered, we analyzed the postoperative survival time of patients characterized with p53-negative expression and MDM2-positive expression. These patients with well-regulated p53-MDM2 expression had significantly better prognoses compared with other patients (P = .034 for CSS; P = .012 for DFS; Figs 5G and 5H, respectively).
DISCUSSION
Consistent with previous studies that found a substantial amount of samples with completely negative staining of MDM2 in oral cancer30 and lung cancer,14,15 we found a decreased expression of MDM2 in 45% of NSCLC tumors. However, reports of the presence of multiple splicing variants of the MDM2 gene are limited regarding primary lung cancers.20 To our knowledge, this study is the first report to show the expression of splicing variants of MDM2 at both the RNA and protein levels in lung cancer. Our sequencing data of all the MDM2 variants demonstrated mRNA splicing that disrupted not only the conserved p53 binding domain, but also further toward the carboxy-terminus, the conserved nuclear localization sequence, the cell-cycle inhibitory domains, and/or the acidic and RING domains. Given that the N-20 antibody should detect both the full-length and most-described MDM2 protein variants, we are able to distinguish these forms of proteins using the N-20 antibody in the immunohistochemistry analysis for the 23 tumors that showed an absence of nuclear staining of MDM2 protein with the IF-2 antibody. N-20 antibody exhibited distinct signals within the cytoplasm in addition to the positive nuclear staining in all tumors examined (Fig 2C). The result suggested that a certain amount of MDM2 variant proteins is located within the cytoplasm and the splice variant may be the predominant form of MDM2 expressed in the lung tumors with p53 wild-type overexpressed protein. Some of these short MDM2 proteins have been shown to function as dominant negative inhibitors to the activity of full-length MDM2 and thus to the overexpression of p53,19,20 and can be oncogenic when assessed in an in vivo context.31 MDM2 variants lacking of the p53-binding domain may lead to instability of the p53-MDM2 feedback loop system,18,19 inhibit the function of the wild-type p53, and result in tumor overgrowth. In addition, the disruption of the cell-cycle inhibitory domain, which occurs in all of the splice MDM2 variants, may induce tumorigenesis.16,17
These observations, in conjunction with our results and other reports in patients with breast, brain, and bone cancers32 that correlated MDM2 splicing variants with advanced tumor stage, suggest that at least some MDM2 splice variants might be important in determining tumor behavior. In addition, our observation in tumors with low MDM2 expression that expression of the active form of Akt was also decreased indicated that the deregulation of Akt pathway might also contribute to the inactivation of MDM2.
In conclusion, our observation of decreased MDM2 expression associated with prominent p14ARF expression in tumors with wild-type p53 overexpression indicates that MDM2 inhibition and p14ARF activation contribute to the stabilization of p53 in the absence of gene mutation in a subgroup of NSCLC patients. This subgroup of patients was also characterized by a more aggressive disease and poorer prognosis. We hypothesize that differentially spliced tumorigenic MDM2 transcripts are first generated and their protein products then interfere with the wild-type MDM2 when interacting and degrading p53. This, together with the inactivation of Akt phosphorylation and the reactivation of p14ARF, results in the abnormal accumulation of p53 and ultimately in its functional inactivation. The p53 overexpression could result in growth arrest or apoptosis, but these tumors are cycling cells. Other mechanisms must inactivate the p53 pathway.
It has been shown recently that p53 appears to repress transcription from the DNA cytosine methyltransferase 1 (DNMT1) promoter by direct, sequence-specific binding to a p53 consensus site in exon 1 of DNMT1 gene.33 In addition, wild-type p53 can downregulate proliferating cell nuclear antigen (PCNA) mRNA and expression of PCNA protein in a concentration-dependent manner.34,35 Our preliminary data indicated that 47% and 74% of the tumors with wild-type p53 stabilization showed overexpression of the DNMT and PCNA protein, respectively. These data, along with our result of the low expression of full-length MDM2 transcript in these tumors, suggests an inactivation of p53 transcriptional regulation in these tumors. The deregulation of p53-mdm2-p14ARF pathway eventually leads to disease progression and poorer prognosis of lung cancer. The expression of MDM2 splice variants may be a mechanism of a secondary phenomenon caused by the differential expression of splice regulatory factors. In keeping with this aberrant transcript result, multiple FHIT aberrant transcripts were also found in these NSCLC tumors (data not shown). Our observations add an additional dimension to the remarkable array of alternative transformation mechanisms associated with the p53-MDM2-p14ARF pathway and promote the re-evaluation of the role of the MDM2 oncogene in the initiation and progression of cancer, and its relation to the p53 tumor suppressor protein.
Authors' Disclosures of Potential Conflicts of Interest
NOTES
Supported in part by Grants NHRI93A1-NSCLC06-5 and NSC92-2320-B-003-003 from the National Science Council (The Executive Yuan, Republic of China).
REFERENCES
1. Oren M: Decision making by p53: Life, death and cancer. Cell Death Differ 10:431-442, 2003
2. Bennett WP, Hussain SP, Vahakangas KH, et al: Molecular epidemiology of human cancer risk: Gene-environment interactions and p53 mutation spectrum in human lung cancer. J Pathol 187:8-18, 1999
3. Midgley CA, Lane DP: p53 protein stability in tumour cells is not determined by mutation but is dependent on Mdm2 binding. Oncogene 15:1179-1189, 1997
4. Costa A, Marasca R, Valentinis B, et al: p53 gene point mutations in relation to p53 nuclear protein accumulation in colorectal cancers. J Pathol 176:45-53, 1995
5. Lowe SW, Sherr CJ: Tumor suppression by Ink4a-Arf: Progress and puzzles. Curr Opin Genet Dev 13:77-83, 2003
6. Sharpless NE, DePinho RA. The INK4a/ARF locus and its two gene products. Curr Opin Genet Dev 9:22-30, 1999
7. Vonlanthen S, Heighway J, Tschan MP, et al: Expression of p16INK4a/p16-alpha and p19ARF/p16-beta is frequently altered in non-small cell lung cancer and correlates with p53 overexpression. Oncogene 17:2779-2785, 1998
8. Xue Q, Sano T, Kashiwabara K, et al: Aberrant expression of pRb, p16, p14ARF, MDM2, p21 and p53 in stage I adenocarcinomas of the lung. Pathol Int 52:103-109, 2002
9. Moll UM, Petrenko O: The MDM2-p53 interaction. Mol Cancer Res 1:1001-1008, 2003
10. Li M, Brooks CL, Wu-Baer F, et al: Mono-versus polyubiquitination: Differential control of p53 fate by Mdm2. Science 302:1972-1975, 2003
11. Koga T, Hashimoto S, Sugio K, et al: Heterogeneous distribution of P53 immunoreactivity in human lung adenocarcinoma correlates with MDM2 protein expression, rather than with P53 gene mutation. Int J Cancer 95:232-239, 2001
12. Eymin B, Gazzeri S, Brambilla C, et al: Mdm2 overexpression and p14(ARF) inactivation are two mutually exclusive events in primary human lung tumors. Oncogene 21:2750-2761, 2002
13. Momand J, Jung D, Wilczynski S, et al: The MDM2 gene amplification database. Nucleic Acids Res 26:3453-3459, 1998
14. Ko JL, Cheng YW, Chang SL, et al: MDM2 mRNA expression is a favorable prognostic factor in non-small-cell lung cancer. Int J Cancer 89:265-270, 2000
15. Higashiyama M, Doi O, Kodama K, et al: MDM2 gene amplification and expression in non-small-cell lung cancer: Immunohistochemical expression of its protein is a favourable prognostic marker in patients without p53 protein accumulation. Br J Cancer 75:1302-1308, 1997
16. Dang J, Kuo ML, Eischen CM, et al: The RING domain of Mdm2 can inhibit cell proliferation. Cancer Res 62:1222-1230, 2002
17. Brown DR, Thomas CA, Deb SP: The human oncoprotein MDM2 arrests the cell cycle: Elimination of its cell-cycle-inhibitory function induces tumorigenesis. EMBO J 17:2513-2525, 1998
18. Bartel F, Taubert H, Harris LC: Alternative and aberrant splicing of MDM2 mRNA in human cancer. Cancer Cell 2:9-15, 2002
19. Sigalas I, Calvert AH, Anderson JJ, et al: Alternatively spliced mdm2 transcripts with loss of p53 binding domain sequences: Transforming ability and frequent detection in human cancer. Nat Med 2:912-917, 1996
20. Evans SC, Viswanathan M, Grier JD, et al: An alternatively spliced HDM2 product increases p53 activity by inhibiting HDM2. Oncogene 20:4041-4049, 2001
21. Gottlieb TM, Leal JF, Seger R, et al: Cross-talk between Akt, p53 and Mdm2: Possible implications for the regulation of apoptosis. Oncogene 21:1299-1303, 2002
22. Zhou BP, Liao Y, Xia W, et al: HER-2/neu induces p53 ubiquitination via Akt-mediated MDM2 phosphorylation. Nat Cell Biol 3:973-982, 2001
23. Pomerantz J, Schreiber-Agus N, Liegeois NJ, et al: The INK4a tumor suppressor gene product, p19Arf, interacts with MDM2 and neutralizes MDM2's inhibition of p53. Cell 92:713-723, 1998
24. Zhang Y, Xiong Y, Yarbrough WG: ARF promotes MDM2 degradation and stabilizes p53: ARF-INK4a locus deletion impairs both the Rb and p53 tumor suppression pathways. Cell 92:725-734, 1998
25. Wang YC, Chen CY, Chen SK, et al: High frequency of deletion mutations in p53 gene from squamous cell lung cancer patients in Taiwan. Cancer Res 58:328-333, 1998
26. Pinyol M, Hernandez L, Martinez A, et al: INK4a/ARF locus alterations in human non-Hodgkin's lymphomas mainly occur in tumors with wild-type p53 gene. Am J Pathol 156:1987-1996, 2000
27. Lee SH, Kim HS, Park WS, et al: Non-small cell lung cancers frequently express phosphorylated Akt; an immunohistochemical study. APMIS 110:587-592, 2002
28. Hsu HS, Wang YC, Tseng RC, et al: CpG island methylation is responsible for the p14ARF inactivation and inversely correlates with p53 overexpression in resected non-small cell lung cancer. Clin Cancer Res 10:4734-4741, 2004
29. Gazzeri S, Della Valle V, Chaussade L, et al: The human p19ARF protein encoded by the beta transcript of the p16INK4a gene is frequently lost in small cell lung cancer. Cancer Res 58:3926-3931, 1998
30. Sano T, Hikino T, Xue Q, et al: Immunohistochemical inactivation of p14ARF concomitant with MDM2 overexpression inversely correlates with p53 overexpression in oral squamous cell carcinoma. Pathol Int 50:709-716, 2000
31. Fridman JS, Hernando E, Hemann MT, et al: Tumor promotion by Mdm2 splice variants unable to bind p53. Cancer Res 63:5703-5706, 2003
32. Onel K, Cordon-Cardo C: MDM2 and prognosis. Mol Cancer Res 2:1-8, 2004
33. Peterson EJ, Bogler O, Taylor SM: p53-mediated repression of DNA methyltransferase 1 expression by specific DNA binding. Cancer Res 63:6579-6582, 2003
34. Mercer WE, Shields MT, Lin D, et al: Growth suppression induced by wild type p53 protein is accompanied by selective down regulation of proliferating cell nuclear antigen expression. Proc Natl Acad Sci U S A 88:1958-1962, 1991
35. Shivakumar CV, Brown DR, Deb S, et al: Wild-type human p53 transactivates the human proliferating cell nuclear antigen promoter. Mol Cell Biol 15:6785-6793, 1995(Yu-Chien Wang, Ruo-Kai Li)
Department of Pathology, and Division of Thoracic Surgery Taichung Veterans General Hospital, Taichung, Taiwan, Republic of China
ABSTRACT
PATIENTS AND METHODS: We performed gene and protein alteration studies on p53 and its upstream effectors, MDM2 and p14ARF, in tumors from 94 non–small-cell lung cancer (NSCLC) patients.
RESULTS: Immunohistochemical and sequencing analyses indicated that 37 tumors showed overexpression of wild-type p53. An absence of nuclear staining of MDM2 protein was found in 95% of these tumors (35 of 37; P < .001). The tumors with negative MDM2 staining showed a significantly high concordance of loss of Akt activity and low MDM2 mRNA expression (P < .001). Sequencing analysis revealed five distinct MDM2 splicing variants disrupting the conserved p53 binding domain. Corresponding variant proteins were detected in three lung cancer cell lines using the Western blot analysis. Our results also indicated that among the tumors with overexpression of the wild-type p53, 92% (34 of 37) showed immunoreactivity to p14ARF (P = .001). In addition, the deregulation of p53 and MDM2 genes was significantly associated with squamous lung cancer (P < .05) and was correlated with advanced stages (P < .05) and poor prognosis (P < .05).
CONCLUSION: Our data suggest that immunopositivity of p14ARF together with a low expression of MDM2 contributes to accumulation of the wild-type p53, and that deregulation of the p53-MDM2-p14ARF pathway is important in the pathogenesis and outcome of a subset of NSCLC.
INTRODUCTION
MDM2 was found to be amplified in a portion of human cancers,13 but amplification in malignant epithelial-derived cancers, including non–small-cell lung cancer (NSCLC), occurs at lower frequencies (approximately 5% to 15%).13-15 MDM2 physically interacts through its amino-terminal domain with p53,9 and through its E3 ubiquitin ligase activity MDM2 transfers monoubiquitin tags onto the lysine residues in p53,9 which leads to the degradation of the p53 protein. Interestingly, experimental evidence also indicates that MDM2 overexpression causes G1 arrest in normal cells, and two cell-cycle inhibitory domains were identified in the middle of MDM2 protein.16,17 These data suggest that MDM2 possibly has a dual role (ie, tumorigenesis and growth arrest) in regulating cell growth. In addition, alternative splicing of MDM2 and the generation of short proteins occurs in many human and mouse tumors.18,19 However, reports on MDM2-aberrant splicing in NSCLC are scarce.20 More recently, it was also shown that mitogen-induced Akt physically associates with and phosphorylates MDM2, leading to an enhanced activity of MDM2 and increased p53 degradation.21,22
The INK4a/ARF locus on human chromosome region 9p21 encodes two distinct proteins, which are translated in different reading frames from alternatively spliced transcripts. p14ARF is an upstream regulator in the p53 tumor suppressor pathways. The p14ARF tumor suppressor, in response to oncogenic and hyperproliferative signals as well as signals of DNA damage, binds to MDM2 and consequently activates p53 by blocking p53 and MDM2 nuclear export and p53 degradation.5,6,23,24 Thus, p14ARF overexpression can lead to the stabilization of p53. It has also been reported that the p14ARF-mediated G1 and G2 arrests are abolished in mouse embryonic fibroblasts lacking functional p53, indicating that p53 also acts downstream of p14ARF in the cell cycle regulatory pathway.25,26
To investigate the role of the p53 tumor suppressor gene in lung tumorigenesis, we examined the frequency and mutation spectra of p53 tumor suppressor gene in 60 lung cancer patients. We found that a relatively reduced frequency of p53 mutation with a much greater frequency of p53 protein overexpression reflected that the p53 protein stabilized in the absence of p53 gene mutation.25 To confirm this p53 immunohistochemical abnormality, specimens of resected NSCLC from another set of 94 NSCLC patients were collected and were examined for p53 mutation spectrum and p53 protein expression in this study. In addition, to test the possibility that p53 protein accumulation may result from alternative mechanisms leading to p53 protein stabilization, we performed alteration analyses of the p53 upstream proteins, p14ARF and MDM2, in these patients. The data indicated that in 92% of patients (35 of 38) with wild-type p53 overexpression, MDM2 expression was decreased and p14ARF expression was correspondingly prominent. In addition, the alteration of p53 and MDM2 genes was associated with lung cancers of more advanced stages and poorer prognosis.
PATIENTS AND METHODS
Surgically resected tumor samples were immediately snap-frozen, and subsequently stored in liquid nitrogen. For the RNA expression assay and cDNA sequencing, the total RNA was prepared from matched pairs of primary tumors and nearby normal lung tissues using Trizol reagent (Invitrogen, Carlsbad, CA). cDNA was synthesized using SuperScript reverse transcriptase (Invitrogen) with the protocols provided by the manufacturer.
Analysis of Protein Expression: Immunohistochemistry Assay
The monoclonal antibody, DO-7 (1:150; DAKO, Glostrup, Denmark), which recognizes both wild-type and mutant-type of p53 proteins, was used as the primary antibody to detect p53 protein expression. Another monoclonal antibody, p53/Ab-5 (1:20; Oncogene, Boston, MA), which recognizes only the wild-type of p53 protein, was later used to identify the status of overexpressed p53. The monoclonal antibodies IF-2 (1:100; Zymed, San Francisco, CA) and N-20 (1:50; Santa Cruz Biotechnology, Santa Cruz, CA) were used as the primary antibodies to detect the central domain and amino-terminal domain of MDM2, respectively. The polyclonal antibody, phospho-Akt (Ser 473; 1:50; Cell Signaling Technology, Beverly, MA), was used to detect the active form of Akt kinase. The evaluation of the immunohistochemistry was conducted blindly, without knowledge of the clinical and pathologic characteristics of the patients. If the positive immunostaining was observed in the positive control cells, the patient case was judged for the immunostaining of the tumor cells. Staining for p53 was scored 3 if more than 50% tumor cells were immunostained positive; 2 if 26% to 50% of cells were positive; 1 if 11% to 25% of cells were positive; and 0 if less than 10% of cells were positive. The stains were graded p53 overexpression if the score was 2 or 3.26 Staining for MDM2 was 3 if more than 50% tumor cells were positive; 2 if 11% to 50% of cells were positive; 1 if 1% to 10% of cells were positive; and 0 if 0% of cells were positive. The stains were graded as exhibiting overexpression if the score was 3 and graded as negative when there was complete absence of staining in the tumor cells.14 Staining for Akt was scored 3 if more than 30% of the tumor cells were immunostaining positive; 2 if 11% to 30% of cells were positive; 1 if 1% to 10% of cells were positive; and 0 if 0% cells were positive. Staining of 3 was evaluated as overexpression.27 The antibody and evaluation used for p14ARF were described previously.28
Mutation Spectrum Analysis for p53 Gene
Tumors were analyzed for DNA sequence mutations in exons 5 to 11 of p53 tumor suppressor gene as describe previously.25 Polymerase chain reaction (PCR) fragments were purified with Qiaex II Gel Extraction Kit (Qiagen, Valencia, CA), then sequenced using ABI 377 Automatic Sequencer (PE Applied Biosystems, Foster City, CA). Sequencing analysis of the opposite strand for all of the mutations were repeated at least once using independent PCR products.
Analysis of mRNA Expression of MDM2 and p14ARF Genes: Multiplex Reverse Transcriptase Polymerase Chain Reaction and cDNA Sequencing
mRNA expression was assayed in a multiplex reverse transcriptase polymerase chain reaction (RT-PCR) analysis using the -actin gene as an internal control. The exons 1 to 2 of the p14ARF gene and the -actin gene were amplified using primers described by Gazzeri et al.29 The exons 3 to 12 of the MDM2 gene were amplified using a nested PCR described by Sigalas et al.19 To quantify the relative levels of gene expression in the multiplex RT-PCR assay, the value for the internal standard (-actin) in each test tube was used as the baseline value for gene expression in that sample, and a relative value was calculated for each target transcript amplified from each tumor and matched normal sample. Tumor cells that exhibited mRNA expression below 50% of that of normal cells were deemed to have a low expression pattern.28 Abnormal cDNA fragments detected in RT-PCR assays of MDM2 gene were eluted, purified, and sequenced as described above.
Analysis of Genomic Sequence of MDM2 Gene: Multiplex PCR
The copy number of genomic sequence of MDM2 gene was assayed in a multiplex PCR analysis using the promoter sequence of GAPDH gene as an internal control. The exons 5 and 9 of MDM2 gene were amplified using the primers to their intron sequences. The primer nucleotide sequences were as follows: for the exon 5 of MDM2 gene, sense-5'-ATCTCCTGACCTCGTGATCCAT-3', antisense-5'-AGATGCCAGAGCTCAG GTTCTCA-3'; for the exon 9 of MDM2 gene, sense-5'-TTCTGCTGTAACAGTTGGACA-3', antisense-5'-TTTGACTGTACTATTGGTGCAGT-3'; for the GAPDH gene, sense-5'-AATGAAAGGCACACTGTCTCTCTC-3', antisense-5'-GTTTCTGCACGGAAGGTCAC-3'. PCR was performed for 35 cycles with the annealing temperature of 65°C. The PCR products were then electrophoresed, stained with ethidium bromide, and quantified for the levels relative to the control sequence.
Western Blot Analysis
Total cell protein (50 μg) was subjected to electrophoresis using an 8% sodium dodecyl sulfate polyacrylamide gel. Subsequently, proteins were transferred to a nitrocellulose membrane (Millipore, Eschborn, Germany) at 200 mA for 1.5 hours using a semidry blotting chamber. Nonspecific binding sites were blocked by 4% low-fat dry milk in Tris-buffered saline and 0.1% Tween-20 for 1 hour at room temperature. Immunodetection of MDM2 was performed using N-20 antibodies (dilution 1:200) at room temperature for 2 hours. After incubation with a horseradish peroxidase–conjugated secondary antibody, immunoreactive proteins were visualized by Western blotting luminol reagent (Santa Cruz Biotechnology). Signal detection was performed using the SuperSignal West Pico Chemiluminescent Substrate (Pierce Biotechnology, Rockford, IL) according to the manufacturer's instructions. Control -actin antibody was purchased from Santa Cruz Biotechnology, Inc.
Statistical Analysis
Pearson's {chi}2 test was used to compare the frequency of p53, MDM2, and p14ARF alterations in NSCLC patients with different characteristics, including sex and smoking status, and various clinicopathologic parameters, such as tumor type and tumor stage. The comparison of age distributions between patients with and without alterations was analyzed with a two-sample t test. Censoring was performed on patients who were still alive or had died as a result of some other cause at the end of the study for the CSS analysis. Censored patients were those who had not yet experienced the recurrence by the time of the follow-up or patients who died without relapse and were censored at the date of death for the DFS analysis. The Kaplan-Meier method was used to estimate the probability of survival as a function of time and the median survival. The log-rank test was used to assess the significance of the difference between pairs of survival probabilities. In addition, we analyzed prognosis significance of gene alterations by stratifying the various clinicopathologic parameters of patients. Multivariate Cox regression analysis using the likelihood ratio test was done to evaluate the influence of clinical stage on the effect of gene alterations in the CSS and DFS. Statistical analysis was performed using Statistical Package for Social Sciences software (SPSS, Chicago, IL). P < .05 was considered to be statistically significant.
RESULTS
MDM2 Alterations and Akt Inactivation in Primary NSCLC Tumors
We therefore tested the hypothesis that p53 stabilization in the wild-type tumors may be through deregulation of the upstream effectors of p53 (ie, MDM2 and/or p14ARF). Immunohistochemical staining of MDM2 protein was performed on 94 tumor samples using the IF-2 antibody directed to its central domain. Fifty-two lung tumors showed staining of moderate to strong intensity within the nucleus of tumor cells (Fig 2A). The remaining 42 tumors (45%) showed an absence of nuclear staining of MDM2 protein (Fig 2B; Table 1). Among the tumors with overexpression of wild-type p53, 95% (35 of 37) showed a negative expression of MDM2.
Given that expression of active Akt kinase promotes activity of MDM2, we hypothesized that diminishing levels of MDM2 may reflect the decrease of Akt activity, and we then immunostained the active form of Akt using an antibody specific to the phosphorylated Akt. The data indicated that 48 tumors that showed positive MDM2 staining also showed positive staining for active Akt (Fig 2D). There were 36 tumors that had loss of expression of both proteins, suggesting that low MDM2 expression correlated with low Akt activity (Fig 2E). A high concordance (84 of 94; 89%) was observed between MDM2 and Akt expression (P = .001).
Analysis of a variety of human tumors has revealed the presence of multiple, alternatively spliced forms of MDM2 messages, which leads to the inactivation of MDM2. Therefore, we analyzed mRNA expression level and isoforms of MDM2 transcripts with semiquantitative multiplex RT-PCR and cDNA sequencing, respectively (Fig 3). Decreased MDM2 full-length transcripts were shown by semiquantitative multiplex RT-PCR to occur in 39 tumors (41%; Fig 3A; Table 1). The correlation between MDM2 protein and mRNA expressions showed a 91% concordance (86 of 94). Negative MDM2 protein expression was significantly associated with low MDM2 mRNA expression (P = .001). The RT-PCR also revealed that 25 of 39 tumors with low levels of full-length MDM2 transcript had shown multiple smaller transcripts. A total of seven distinct variants were found: Del.a, 1,016 bp, 10 cases; Del.b, 941 bp, five cases; Del.c, 707 bp, 11 cases; Del.d, 542 bp, five cases; Del.e, 449 bp, three cases; Del.f, 303 bp, two cases; and Del.g, 274 bp, two cases (Fig 3B). Sequence analysis was carried out to determine the composition of the abnormally sized MDM2 transcripts. The 542-bp product has not been reported in the literature and its sequencing data showed a lack of exons 4 to 11, as well as a lack of the first 165 base pairs of exon 12, leaving a residual exon 12 starting from nucleotide 1066 (Fig 3C). The sequence of nucleotides 1064 and 1065 is AG, which may serve as a cryptic 3' intron junction during splicing.
To examine whether aberrant transcripts resulted from deletions in the genomic MDM2 gene, we analyzed genomic DNA samples from a subset of 16 patients with aberrantly sized transcripts for the deletion of exons 5 and 9, which are the exons frequently missing in the mRNA, using a multiplex PCR assay (Fig 3D). None of the genomic DNA samples showed a deletion of either exon 5 or 9, thus providing evidence that the aberrant MDM2 transcripts may not be caused by a deletion of the genomic sequence of MDM2. The data suggested that the frequent lack of exons 5 to 9 in MDM2 cDNA is predominantly due to splicing abnormalities.
To determine whether the splice variants express the truncated proteins, three lung cancer cell lines that showed aberrant mRNA expression were analyzed for protein expression using the Western blot assay with the N-20 antibody directed to the amino-terminal of the MDM2 protein that should detect most described MDM2 protein variants. As shown in Figure 4, the MDM2 proteins (p75, p57, and p40) were found to be expressed more abundantly than the wild-type p90 protein in these lung cancer cell lines. These proteins probably correspond with MDM2 variants from transcripts of 1016, 941, and 707 bp, respectively. These variant proteins with a loss of central regions, including the epitope of IF-2 antibody, showed no signal detecting by IF-2 antibody with same exposure time (data not shown).
p14ARF Protein and mRNA Expression in Primary NSCLC Tumors
It has been shown that p14ARF inhibits p53 and MDM2 binding, and thereby maintains p53 at a high level. Immunohistochemical staining was therefore performed for p14ARF protein on 94 tumor samples. Sixty-one lung tumors showed staining of moderate to strong intensity. The remaining 33 lung cancers (35%) showed aberrant expression of p14ARF protein, characterized by a low level or complete absence of nuclear staining (Table 1). Among the tumors with overexpression of wild-type p53, 92% (34 of 37) showed a positive expression of p14ARF.
Decreased p14ARF transcripts were shown by semiquantitative multiplex RT-PCR to occur in 29 tumors (31%; Table 1). The correlation between p14ARF protein and mRNA expressions showed an 87% concordance (82 of 94). Negative protein expression was significantly associated with low mRNA expression (P = .001).
Concordance Analysis of p53, MDM2, and p14ARF Alterations
The data on the protein analysis of p53, MDM2, and p14ARF were cross-referenced to investigate the correlations between these three genes (Table 1). A striking discovery was made about the association between the overexpression of p53 protein with low expression of MDM2 protein and positive expression of the p14ARF protein. p53 overexpression inversely correlated with low MDM2 expression (P < .001) and frequently was accompanied by the expression of the p14ARF protein (P = .001). In addition, positive p14ARF protein expression was significantly associated with negative MDM2 expression (P = .001). Overall, 34 of 37 patients with p53 overexpression showed absence or low expression of MDM2 protein and moderate to strong expression of p14ARF protein.
Alterations of p53, MDM2, and p14ARF Genes and Their Correlation With Clinical Parameters and Prognoses of NSCLC Patients
To examine whether there were associations of alterations at the p53, MDM2, and p14ARF genes with the clinical characteristics of patients, the occurrence of alteration at each gene was compared with the patients' clinicopathologic parameters, including sex, age, smoking habit, tumor type, and tumor stage (Table 1). {chi}2 analysis showed that p53 mutation and protein overexpression occurred primarily in patients suffering from SQ types of lung cancer (P = .003 and P = .038, respectively). p53 overexpression was significantly restricted to patients suffering from stage III or IV lung cancers (P = .040). Low MDM2 gene expression also appeared to be more frequent in SQ than in AD, and the statistical significance of this difference was found for the protein expression data (P = .033). There was a significant trend for low expression of MDM2 in patients with advanced stages; the incidence of low mRNA expression of MDM2 gene was statistically associated with stage III and IV patients (P = .013). In addition, a highly significant correlation between low p14ARF expression and AD patients was found (P = .016; Table 1).
The relationship between postoperative survival and the altered protein expression of the p53 and MDM2 was analyzed in 94 patients. p53 protein overexpression was statistically associated with a poorer prognosis for CSS (P = .042; Fig 5A) and for DFS (P = .029; Fig 5B) even after data were adjusted for tumor staging factor. There was a trend for shorter DFS in patients with low MDM2 protein expression when compared with those with positive protein expression, and in patients with altered MDM2 splicing variants when compared with those without splicing variants. However, this did not reach statistical significance. Nevertheless, we found that patients with altered MDM2 expression had worse prognoses compared with those without, especially in AD lung cancer (P = .037 for CSS; P = .010 for DFS; Figs 5C and 5D, respectively). In the patients with early-stage lung cancer, in whom identifying unfavorable prognostic factors is important, patients with alternative splicing of MDM2 mRNA had significantly worse prognoses than those without (P = .020 for CSS; P = .029 for DFS; Figs 5E and 5F, respectively). To test the hypothesis of an increased correlation of cancer prognosis when more than one disease-related gene was considered, we analyzed the postoperative survival time of patients characterized with p53-negative expression and MDM2-positive expression. These patients with well-regulated p53-MDM2 expression had significantly better prognoses compared with other patients (P = .034 for CSS; P = .012 for DFS; Figs 5G and 5H, respectively).
DISCUSSION
Consistent with previous studies that found a substantial amount of samples with completely negative staining of MDM2 in oral cancer30 and lung cancer,14,15 we found a decreased expression of MDM2 in 45% of NSCLC tumors. However, reports of the presence of multiple splicing variants of the MDM2 gene are limited regarding primary lung cancers.20 To our knowledge, this study is the first report to show the expression of splicing variants of MDM2 at both the RNA and protein levels in lung cancer. Our sequencing data of all the MDM2 variants demonstrated mRNA splicing that disrupted not only the conserved p53 binding domain, but also further toward the carboxy-terminus, the conserved nuclear localization sequence, the cell-cycle inhibitory domains, and/or the acidic and RING domains. Given that the N-20 antibody should detect both the full-length and most-described MDM2 protein variants, we are able to distinguish these forms of proteins using the N-20 antibody in the immunohistochemistry analysis for the 23 tumors that showed an absence of nuclear staining of MDM2 protein with the IF-2 antibody. N-20 antibody exhibited distinct signals within the cytoplasm in addition to the positive nuclear staining in all tumors examined (Fig 2C). The result suggested that a certain amount of MDM2 variant proteins is located within the cytoplasm and the splice variant may be the predominant form of MDM2 expressed in the lung tumors with p53 wild-type overexpressed protein. Some of these short MDM2 proteins have been shown to function as dominant negative inhibitors to the activity of full-length MDM2 and thus to the overexpression of p53,19,20 and can be oncogenic when assessed in an in vivo context.31 MDM2 variants lacking of the p53-binding domain may lead to instability of the p53-MDM2 feedback loop system,18,19 inhibit the function of the wild-type p53, and result in tumor overgrowth. In addition, the disruption of the cell-cycle inhibitory domain, which occurs in all of the splice MDM2 variants, may induce tumorigenesis.16,17
These observations, in conjunction with our results and other reports in patients with breast, brain, and bone cancers32 that correlated MDM2 splicing variants with advanced tumor stage, suggest that at least some MDM2 splice variants might be important in determining tumor behavior. In addition, our observation in tumors with low MDM2 expression that expression of the active form of Akt was also decreased indicated that the deregulation of Akt pathway might also contribute to the inactivation of MDM2.
In conclusion, our observation of decreased MDM2 expression associated with prominent p14ARF expression in tumors with wild-type p53 overexpression indicates that MDM2 inhibition and p14ARF activation contribute to the stabilization of p53 in the absence of gene mutation in a subgroup of NSCLC patients. This subgroup of patients was also characterized by a more aggressive disease and poorer prognosis. We hypothesize that differentially spliced tumorigenic MDM2 transcripts are first generated and their protein products then interfere with the wild-type MDM2 when interacting and degrading p53. This, together with the inactivation of Akt phosphorylation and the reactivation of p14ARF, results in the abnormal accumulation of p53 and ultimately in its functional inactivation. The p53 overexpression could result in growth arrest or apoptosis, but these tumors are cycling cells. Other mechanisms must inactivate the p53 pathway.
It has been shown recently that p53 appears to repress transcription from the DNA cytosine methyltransferase 1 (DNMT1) promoter by direct, sequence-specific binding to a p53 consensus site in exon 1 of DNMT1 gene.33 In addition, wild-type p53 can downregulate proliferating cell nuclear antigen (PCNA) mRNA and expression of PCNA protein in a concentration-dependent manner.34,35 Our preliminary data indicated that 47% and 74% of the tumors with wild-type p53 stabilization showed overexpression of the DNMT and PCNA protein, respectively. These data, along with our result of the low expression of full-length MDM2 transcript in these tumors, suggests an inactivation of p53 transcriptional regulation in these tumors. The deregulation of p53-mdm2-p14ARF pathway eventually leads to disease progression and poorer prognosis of lung cancer. The expression of MDM2 splice variants may be a mechanism of a secondary phenomenon caused by the differential expression of splice regulatory factors. In keeping with this aberrant transcript result, multiple FHIT aberrant transcripts were also found in these NSCLC tumors (data not shown). Our observations add an additional dimension to the remarkable array of alternative transformation mechanisms associated with the p53-MDM2-p14ARF pathway and promote the re-evaluation of the role of the MDM2 oncogene in the initiation and progression of cancer, and its relation to the p53 tumor suppressor protein.
Authors' Disclosures of Potential Conflicts of Interest
NOTES
Supported in part by Grants NHRI93A1-NSCLC06-5 and NSC92-2320-B-003-003 from the National Science Council (The Executive Yuan, Republic of China).
REFERENCES
1. Oren M: Decision making by p53: Life, death and cancer. Cell Death Differ 10:431-442, 2003
2. Bennett WP, Hussain SP, Vahakangas KH, et al: Molecular epidemiology of human cancer risk: Gene-environment interactions and p53 mutation spectrum in human lung cancer. J Pathol 187:8-18, 1999
3. Midgley CA, Lane DP: p53 protein stability in tumour cells is not determined by mutation but is dependent on Mdm2 binding. Oncogene 15:1179-1189, 1997
4. Costa A, Marasca R, Valentinis B, et al: p53 gene point mutations in relation to p53 nuclear protein accumulation in colorectal cancers. J Pathol 176:45-53, 1995
5. Lowe SW, Sherr CJ: Tumor suppression by Ink4a-Arf: Progress and puzzles. Curr Opin Genet Dev 13:77-83, 2003
6. Sharpless NE, DePinho RA. The INK4a/ARF locus and its two gene products. Curr Opin Genet Dev 9:22-30, 1999
7. Vonlanthen S, Heighway J, Tschan MP, et al: Expression of p16INK4a/p16-alpha and p19ARF/p16-beta is frequently altered in non-small cell lung cancer and correlates with p53 overexpression. Oncogene 17:2779-2785, 1998
8. Xue Q, Sano T, Kashiwabara K, et al: Aberrant expression of pRb, p16, p14ARF, MDM2, p21 and p53 in stage I adenocarcinomas of the lung. Pathol Int 52:103-109, 2002
9. Moll UM, Petrenko O: The MDM2-p53 interaction. Mol Cancer Res 1:1001-1008, 2003
10. Li M, Brooks CL, Wu-Baer F, et al: Mono-versus polyubiquitination: Differential control of p53 fate by Mdm2. Science 302:1972-1975, 2003
11. Koga T, Hashimoto S, Sugio K, et al: Heterogeneous distribution of P53 immunoreactivity in human lung adenocarcinoma correlates with MDM2 protein expression, rather than with P53 gene mutation. Int J Cancer 95:232-239, 2001
12. Eymin B, Gazzeri S, Brambilla C, et al: Mdm2 overexpression and p14(ARF) inactivation are two mutually exclusive events in primary human lung tumors. Oncogene 21:2750-2761, 2002
13. Momand J, Jung D, Wilczynski S, et al: The MDM2 gene amplification database. Nucleic Acids Res 26:3453-3459, 1998
14. Ko JL, Cheng YW, Chang SL, et al: MDM2 mRNA expression is a favorable prognostic factor in non-small-cell lung cancer. Int J Cancer 89:265-270, 2000
15. Higashiyama M, Doi O, Kodama K, et al: MDM2 gene amplification and expression in non-small-cell lung cancer: Immunohistochemical expression of its protein is a favourable prognostic marker in patients without p53 protein accumulation. Br J Cancer 75:1302-1308, 1997
16. Dang J, Kuo ML, Eischen CM, et al: The RING domain of Mdm2 can inhibit cell proliferation. Cancer Res 62:1222-1230, 2002
17. Brown DR, Thomas CA, Deb SP: The human oncoprotein MDM2 arrests the cell cycle: Elimination of its cell-cycle-inhibitory function induces tumorigenesis. EMBO J 17:2513-2525, 1998
18. Bartel F, Taubert H, Harris LC: Alternative and aberrant splicing of MDM2 mRNA in human cancer. Cancer Cell 2:9-15, 2002
19. Sigalas I, Calvert AH, Anderson JJ, et al: Alternatively spliced mdm2 transcripts with loss of p53 binding domain sequences: Transforming ability and frequent detection in human cancer. Nat Med 2:912-917, 1996
20. Evans SC, Viswanathan M, Grier JD, et al: An alternatively spliced HDM2 product increases p53 activity by inhibiting HDM2. Oncogene 20:4041-4049, 2001
21. Gottlieb TM, Leal JF, Seger R, et al: Cross-talk between Akt, p53 and Mdm2: Possible implications for the regulation of apoptosis. Oncogene 21:1299-1303, 2002
22. Zhou BP, Liao Y, Xia W, et al: HER-2/neu induces p53 ubiquitination via Akt-mediated MDM2 phosphorylation. Nat Cell Biol 3:973-982, 2001
23. Pomerantz J, Schreiber-Agus N, Liegeois NJ, et al: The INK4a tumor suppressor gene product, p19Arf, interacts with MDM2 and neutralizes MDM2's inhibition of p53. Cell 92:713-723, 1998
24. Zhang Y, Xiong Y, Yarbrough WG: ARF promotes MDM2 degradation and stabilizes p53: ARF-INK4a locus deletion impairs both the Rb and p53 tumor suppression pathways. Cell 92:725-734, 1998
25. Wang YC, Chen CY, Chen SK, et al: High frequency of deletion mutations in p53 gene from squamous cell lung cancer patients in Taiwan. Cancer Res 58:328-333, 1998
26. Pinyol M, Hernandez L, Martinez A, et al: INK4a/ARF locus alterations in human non-Hodgkin's lymphomas mainly occur in tumors with wild-type p53 gene. Am J Pathol 156:1987-1996, 2000
27. Lee SH, Kim HS, Park WS, et al: Non-small cell lung cancers frequently express phosphorylated Akt; an immunohistochemical study. APMIS 110:587-592, 2002
28. Hsu HS, Wang YC, Tseng RC, et al: CpG island methylation is responsible for the p14ARF inactivation and inversely correlates with p53 overexpression in resected non-small cell lung cancer. Clin Cancer Res 10:4734-4741, 2004
29. Gazzeri S, Della Valle V, Chaussade L, et al: The human p19ARF protein encoded by the beta transcript of the p16INK4a gene is frequently lost in small cell lung cancer. Cancer Res 58:3926-3931, 1998
30. Sano T, Hikino T, Xue Q, et al: Immunohistochemical inactivation of p14ARF concomitant with MDM2 overexpression inversely correlates with p53 overexpression in oral squamous cell carcinoma. Pathol Int 50:709-716, 2000
31. Fridman JS, Hernando E, Hemann MT, et al: Tumor promotion by Mdm2 splice variants unable to bind p53. Cancer Res 63:5703-5706, 2003
32. Onel K, Cordon-Cardo C: MDM2 and prognosis. Mol Cancer Res 2:1-8, 2004
33. Peterson EJ, Bogler O, Taylor SM: p53-mediated repression of DNA methyltransferase 1 expression by specific DNA binding. Cancer Res 63:6579-6582, 2003
34. Mercer WE, Shields MT, Lin D, et al: Growth suppression induced by wild type p53 protein is accompanied by selective down regulation of proliferating cell nuclear antigen expression. Proc Natl Acad Sci U S A 88:1958-1962, 1991
35. Shivakumar CV, Brown DR, Deb S, et al: Wild-type human p53 transactivates the human proliferating cell nuclear antigen promoter. Mol Cell Biol 15:6785-6793, 1995(Yu-Chien Wang, Ruo-Kai Li)