The first 2 authors contributed equally to this article.
Article first published online: 14 JUL 2011
Copyright © 2011 American Cancer Society
Volume 118, Issue 4, pages 1110–1118, 15 February 2012
How to Cite
Yu, Z., Zhang, B., Cui, B., Wang, Y., Han, P. and Wang, X. (2012), Identification of spliced variants of the proto-oncogene hdm2 in colorectal cancer. Cancer, 118: 1110–1118. doi: 10.1002/cncr.26330
We thank the members of the Department of Colorectal Surgery, Third Affiliated Hospital of Harbin Medical University for their stimulating discussion.
- Issue published online: 3 FEB 2012
- Article first published online: 14 JUL 2011
- Manuscript Accepted: 13 MAY 2011
- Manuscript Revised: 8 APR 2011
- Manuscript Received: 23 FEB 2011
- alternative splicing;
- tumor progression;
- colorectal cancer
The human double minute 2 (hdm2) oncogene is a negative regulator of the p53 gene. Expression and alternative splicing of the hdm2 gene may contribute to colorectal cancer development or progression. This study aimed to determine the presence and identification of aberrant mRNA transcripts of hdm2 in colorectal cancer tissues and cell lines, and determine the nature of their association with clinicopathological characteristics and survival of patients.
A total of 69 colorectal cancer and corresponding normal tissue specimens and 10 colon cancer cell lines were recruited for polymerase chain reaction and DNA sequencing analyses of hdm2 mRNA. Genomic DNA from these tissues and cells was also extracted for p53 gene mutation analysis. The association of hdm2 fragmented transcripts and p53 gene mutation with clinicopathological data was then statistically analyzed.
In 62 cases (89.9%; 62 of 69) of colorectal cancer tissues the full-length hdm2 was amplified, whereas 7 cases had no hdm2 transcripts. Thirty-two of 62 cases (51.6%) and 6 of 10 cell lines (60%) showed at least 1 hdm2 spliced variant. A total of 4 hdm2 splicing variants were found in colorectal cancer tissues and cells, that is, lack of nucleotides between 157 and 292 bp in hdm2/1338, 81 to 901 bp in hdm2/707, 157 to 292, 407 to 505, and 668 to 901 bp in hdm2/1007, and 610 to 883 in hdm2/1200. Of these, hdm2/1338 is a novel hdm2 variant in colorectal cancer. Mutation in p53 was detected in 21 cases (33.8%; 21 of 62). Although there was no association found between expression of hdm2 splicing variants and p53 gene mutations, expression of hdm2 splicing variants was associated with advanced tumor stage (P = .022) and distant metastasis (P = .004) in wild-type p53 cases, and with poor survival of patients (P = .039).
The data from the current study provide the first evidence that hdm2 mRNA is frequently mutated by alternative splicing in colorectal cancer, and may play a role in colorectal tumorigenesis or cancer progression. Cancer 2012; . © 2011 American Cancer Society.
The murine double minute 2 (Mdm2) oncogene (also known as the Mdm2 p53-binding protein homolog [mouse]) was originally cloned as an amplified gene on double minute chromosomes in a spontaneously transformed 3T3 murine cell line.1 Mdm2 was later found to be an important negative regulator of the tumor suppressor gene p53 (or tumor protein p53) via either the ubiquitin-proteasome pathway or as an inhibitor of p53 transcriptional activation.2 As the human counterpart, hdm2 encodes a 90-kDa nuclear phosphoprotein, and has been shown to be overexpressed in several types of human tumors, including soft tissue sarcoma, glioma, and breast cancer.3 Identical to the Mdm2 oncogene, hdm2 functions as a negative regulator that interferes with p53 in an autoregulatory feedback loop,4, 5 inhibiting p53 function either by binding to the transactivation domain of the p53 protein6 or promoting rapid p53 degradation using a proteasome-dependent mechanism.7, 8 Previous studies demonstrated that alterations of p53 protein were detected in 50% to 80% of colorectal cancers and therefore may contribute to colorectal cancer development.9 As a negative regulator, hdm2 may play a role in p53 protein alteration and colorectal carcinogenesis, although this and the underlying mechanism remain to be defined. Moreover, another recent study demonstrated that >40 different splice variants of hdm2 transcripts have been detected in human cancers, and some of these variants had a function independent of changes in p53. Thus hdm2 may also function in a p53-independent fashion, and this needs to be further explored.3
Colorectal cancer is among the most common malignancies in the world. In the United States alone, the lifetime risk of developing colon cancer is about 7%. The risk factors that contribute to colorectal carcinogenesis include old age, a diet high in fat but low in fruits and vegetables, physical inactivity, a family history of colorectal cancer, adenomatous polyps, inflammatory bowel disease, cigarette smoking, alcohol consumption, and obesity.10 These risk factors contribute to colorectal carcinogenesis through multiple gene alterations; little is known about the role of hdm2 and its spliced variants in colon cancer. Therefore, in this study we detected and analyzed transcripts of hdm2, noting instances of aberrant splicing, in 69 cases of primary colorectal cancer tissue specimens to determine their distinct patterns or occurrence and oncogenic activities. We then sought associations of hdm2 mRNA alterations with the clinicopathological characteristics and clinical outcomes in patients with colorectal cancer.
MATERIALS AND METHODS
A protocol to use human tissue specimens and clinicopathological data in this study was reviewed and approved by the University of Harbin Institutional Medical Research Committee. The patients or their guardians agreed and signed an informed written consent form for participation in the current study. We collected 69 tumor samples from patients diagnosed with colorectal cancer. These patients underwent surgical treatment in the Department of Colorectal Surgery at the Third Affiliated Hospital of Harbin Medical University between 1998 and 2003. The tissue specimens were stored in liquid nitrogen until used. Frozen tissue sections stained with hematoxylin & eosin were reviewed and confirmed for the diagnosis. These and the corresponding tissue specimens were used for this study. In addition, clinicopathological data were also collected from the patients' medical history (Table 1).
|Case||hdm2 Splice Variants, No.||Name of hdm2 Splice Variants||p53 Gene Mutation|
|1T||2||hdm2/1338 and hdm2/1007|
|2T||2||hdm2/1338 and hdm2/1007||R213stop; P278Q|
|5T||2||hdm2/1338 and hdm2/1007||S240I|
|9T||2||hdm2/1338 and hdm2/1200|
|32T||2||hdm2/1338 and hdm2/1200||R248E|
|35T||2||hdm2/1200 and hdm2/707|
|38T||3||hdm2/1338, hdm2/1007, and hdm2/707|
|63T||2||hdm2/1338 and hdm2/1007|
|67T||2||hdm2/1338 and hdm2/1007|
|70T||3||hdm2/1338, hdm2/1200, and hdm2/1007|
|SW48||2||hdm2/1338 and hdm2/1007|
Cell Lines and Culture
The human colorectal cancer cell lines KM12C, KM12SM, and KM1214 were obtained from The University of Texas MD Anderson Cancer Center at Houston, Texas, and other cell lines were purchased from the American Type Culture Collection (Manassas, Va). Human colorectal cancer cell lines KM12C, KM125M, KM1214, and RKO were cultured in Eagle minimum essential medium containing 10% fetal bovine serum (FBS) and antibiotics. Human colon cancer cell lines SW480, SW48, and SW620 were cultured in Leibovitz L-15 medium containing 10% FBS. The HT29 and HCT116 cell lines were cultured in McCoy 5a medium containing 5% FBS. LoVo cells were cultured in F-12K medium containing 10% FBS. All the cell lines were maintained at 37°C in 95% air and 5% CO2.
RNA Isolation and Reverse Transcription
Total RNA was isolated from the cultured colon cancer cell lines or 10 to 20 (depending on the size of the tissues) consecutive 10-μm-thick frozen tissue sections using TRIzol (Life Technologies, Gaithersburg, Md) according to the manufacturer's instructions. The RNA was then precipitated with isopropanol, washed with 70% ethanol, allowed to dry at room temperature, and dissolved in 30 μL of RNase-free H2O. The concentration of RNA was measured using a spectrophotometer (SmartSpec TM 3000 from Bio-Rad Laboratories, Hercules, Calif) at a ratio of 260/280. The integrity of the isolated RNA was analyzed for 28S, 18S, and 5S rRNA bands using Northern blotting. After that, the RNA was reverse transcribed into cDNA using mouse mammary leukemia virus reverse transcriptase (Life Technologies) according to the manufacturer's protocol. The newly synthesized cDNA was then stored at −80°C until used.
Design of hdm2 Primers for Polymerase Chain Reaction
A nested polymerase chain reaction (PCR) protocol was used to amplify the full-length hdm2 cDNA as described by Matsumoto et al,11 with some modifications. In brief, the entire open reading frame of the hdm2 (1473 bp) cDNA was amplified using the following nested primer sets: external primer pair 5′-CTGGGGAGTCTTGAGG GACC-3′ (sense) and 5′-CAGGTTGTCTAAATTCCTAG-3′ (antisense), and internal primer pair 5′-CGCGAAAACCCCGGATGGTGAG-3′ (sense) and 5′-CTCTTATAGACAGGTCAACTAG-3′ (antisense). PCR was performed with a 20-μL reaction mixture containing 1× reaction buffer (Clontech, Palo Alto, Calif), 1.1 mM Mg(Oac)2, 0.4 μM of each primer, 0.2 mM deoxynucleotide triphosphates (dNTPs), 0.4 U Tth DNA polymerase, and 1.0 μL template cDNA. The first PCR was performed for 25 cycles of 95°C for 30 seconds, 55°C for 30 seconds, and 68°C for 4.5 minutes, using the external primers in a PTC-200 Peltier Thermal Cycler (Bio-Rad Laboratories). The PCR tubes were sealed with a mineral oil overlay. One microliter of the first-round PCR product was used as the template for the second round, in which conditions were similar to the first-round PCR, but using the internal primers for 35 cycles. The PCR products were separated in a 2% agarose gel and then stained with SYBR Gold and observed with an ultraviolet (UV) transilluminator (Atto, Tokyo, Japan).
In addition, PCR was also performed to assess the hdm2 gene for potential deletions or mutations using genomic DNA from colorectal cancer tissue specimens. The primers used for genomic DNA analysis were: 5′-GCAGGCAAATGTGCAATACCAAC-3′ (sense) and 5′-CTTCTTTAGATACACTTAACTCCG-3′ (antisense).
Sequencing of hdm2 PCR Products
PCR products were then subjected to DNA sequencing analysis to detect any deletion or mutation of the hdm2 gene. Briefly, PCR products were first separated in a 2% agarose gel. The corresponding bands were excised from the gel and purified using a QIAquick Gel Extraction kit (Qiagen, Valencia, Calif). These PCR products were cloned into the TOPO pCRII vector. After transformation of Escherichia coli and bacterial growth in lysogeny broth medium, the plasmid was isolated using a Wizard plasmid miniprep kit from Promega (Madison, Wis) and sequenced using a BigDye Terminator v1.1 cycle sequencing kit (Applied Biosystems, Foster City, Calif) according to the manufacturer's instructions. The data were analyzed with an ABI PRISM 310 Genetic Analyzer and accompanying software (Applied Biosystems).
Detection of p53 Gene Mutations
Genomic DNA from colorectal cancer and corresponding normal tissue specimens was extracted by protease K digestion and purified by serial treatments with phenol and chloroform. The genomic DNA was then subjected to PCR amplification for p53 gene and sequence analysis. The primers for detection of p53 tumor suppressor gene mutations were designed for amplification of the exons 5, 6, 7, and 8 with ordinary upstream and downstream primers as described,12 that is, exon 5: 5′-TTCCTCTTCCTGCAGTACTC-3′ (sense) and 5′-CAGCTGCTCACCATCGCTAT-3′ (antisense); exon 6: 5′-CACTGATTGCTTAGGTTCT-3′ (sense) and 5′-AGTTGCAAACCAGACCTCAG-3′ (antisense); exon 7: 5′-GTGTTATCTCCTAGGTTGGC-3′ (sense) and 5′-CAAGTGGCTCCTGACCTGGA-3′ (antisense); and exon 8: 5′-CCTATCCTGAGTAGTGGTAA-3′ (sense) and 5′-TCCTGCTTGCTTACCTCGCT-3′ (antisense). PCR amplification was performed in 50-μL reaction mixtures containing 1.5 mM MgCl2, 0.2 μM each primer, 0.2 mM dNTPs, 2.5 U of Taq polymerase (Promega), and 200 ng of DNA. PCR conditions were 35 cycles of 94°C for 1 minute, 53°C for 1 minute, and 72°C for 1 minute with a PTC-200 Peltier Thermal Cycler. The PCR tubes were sealed with a mineral oil overlay. PCR products were resolved by electrophoresis in 10% polyacrylamide gel with SYBR Gold staining and observed under UV transillumination to check for quality before sequencing. The PCR products were purified in microcentrifuge tubes to eliminate unused nucleotides and primers at 1500 × g × 8 × 3 (SUPREC-02; Takara Bio, Otsu, Japan) and repurified with a Spin column (Centri-Sep, Applied Biosystems). After that, DNA concentration was measured using a spectrophotometer (SmartSpec 3000, Bio-Rad Laboratories). For DNA sequencing, 10 ng DNA was added to a 20-μL mixture: 4 μL 5× BigDye Sequencing buffer, 8 μL 2.5× Ready Reaction Premix, 3.2 pmol sequence primer, and distilled water. DNA was then amplified and sequenced with M13 forward and reverse primers by using an ABI PRISM 310 Genetic Analyzer. The data were analyzed with the software attached to the analyzer.
The experimental data were analyzed according to the data types and distributions using Fisher exact test or the chi-square test. Data with P < .05 were considered statistically significant. In addition, the association between hdm2 expression and overall survival was analyzed using the log-rank test.
Altered Expression of hdm2 mRNA in Colorectal Cancer Tissues and Cell Lines
We first detected hdm2 mRNA expression in 69 cases of colorectal cancer and corresponding normal tissue specimens and 10 colon cancer cell lines (KM12C, KM1214, KM125M, SW480, SW620, SW48, HT29, LoVo, HCT116, and RKO) using reverse transcription PCR. Our data showed that a 1473-bp fragment corresponding to the full-length hdm2 mRNA was evident in 62 tumor tissues and all the cell lines, whereas there were 7 specimens that did not produce any PCR products. In addition, 32 of these 62 colorectal cancer tissue specimens and 6 of 10 colon cancer cell lines (ie, SW480, SW48, SW620, HT29, LoVo, and RKO) showed additional smaller size bands in the PCR gel (Table 1). In contrast, the corresponding normal tissue samples had only the normal-sized hdm2 products (Fig. 1A). The subsequent direct sequencing of the PCR products revealed several hdm2 transcript variants in colorectal cancers (see below). However, there were no deletions or mutations of the hdm2 gene detected in any of the samples analyzed (data not shown).
Aberrant Alternative hdm2 mRNA Splicing in Colorectal Cancer Tissues and Cells
Next, we sequenced all PCR products, that is, both normal-sized and abnormal-sized hdm2 PCR products. Full-length hdm2 cDNAs were present as well as 4 smaller fragments of 1338, 1200, 1007, and 707 bp (Fig. 1). After DNA sequencing, the data showed that 1 or more hdm2 splicing variants were in 51.6% (32 of 62) of the colorectal cancer tissue specimens and 60% (6 of 10) of the colorectal cancer cell lines (Table 1). Specifically, of the patient cases, 1 had 1338, 1200, and 1007 bp fragments, 2 had 1338 and 1200 bp, 5 had 1338 and 1007 bp, 1 had 1200 and 1007 bp, and another had 1200 and 707 bp. In addition, a 1007-bp PCR product was detected in the SW480, SW48, SW620, HT29, and RKO cell lines, and a 707-bp product was detected in the LoVo cell line (Fig. 1B).
Furthermore, as shown in Figure 2, the 707-bp fragment was an alternatively spliced hdm2 mRNA. It encodes for sequences present in hdm2 exons 3 and 12. RNA was spliced between exon 3 and 12, occurring at the precise exon splicing donor acceptor motifs, resulting in a missing large central portion of hdm2 mRNA between nucleotides 82 and 901 that codes 90% of the amino acids in the hdm2 p53-binding domain and the entire nuclear localization signal and acidic domain. The zinc and RING finger domains remained intact. This alternative splice was identical to that described previously by Evans et al13 and referred to there as hdm2-B, and has been proven to promote p53-independent cell growth, inhibit apoptosis, and induce tumor formation in transgenic mice.14 hdm2-B is the most frequently detected hdm2 spliced variant found in human tumors.3 But in our study, it was detected in only 4 cases of colorectal cancer and the LoVo cell line (Table 1).
In addition, we also found 3 other alternative splices of hdm2 that have never been reported before in colorectal cancers. The 1338-bp fragment, which lacks nucleotides 157 to 292 (hdm2 of 1338), results in loss of 40% of the p53-binding domain of hdm2 protein. This was the most prevalent transcript observed in our study, detected in 38.7% (24 of 62) of colorectal cancer cases. The 1007-bp fragment (hdm2 of 1007) has nucleotides 157 to 292, 407 to 505, and 668 to 901 deleted, resulting in loss of part of the p53-binding and acidic domains. The 1200-bp fragment has a deletion between nucleotides 610 and 883. Here, the p53-binding domain was intact, but the acidic domain was missing.
Association of Aberrant hdm2 mRNA Expression With Clinicopathological Characteristics
We further sought for associations between abnormalities in hdm2 transcriptional splicing and the clinicopathological data collected from colorectal cancer patients. By using Fisher exact test, we found that the presence of abnormal hdm2 mRNA fragments was significantly associated with higher tumor stage (P = .022) and distant metastasis (P = .004; Table 2). Other clinicopathological data were not found to be associated with any abnormalities of the hdm2 gene. Furthermore, we also found a statistically significant correlation between expression of hdm2 splicing variants and poor overall survival (P = .039; Fig. 3).
|Characteristic||Patients, n = 62||Abnormal hdm2||Positive Rate, %||P|
|Present, n = 32||Absent, n = 30|
Association of Altered hdm2 mRNA Expression With p53 Gene Mutation
Because hdm2 protein can bind to p53 protein and is a negative regulator of p53 function and expression, we analyzed these colorectal cancer tissue specimens for p53 gene mutations or deletions by sequencing the p53 gene exons 5, 6, 7, and 8 as described previously.12 The data showed that mutations in p53 were present in 33.8% (21 of 62) of colorectal cancer cases (Table 1), which is lower relative to previous reports.9 However, our data did not demonstrate any association between abnormalities of hdm2 gene transcription and p53 gene mutations in exons 5 to 8. Particularly, our data showed that the presence of hdm2 splicing variants of the hdm2 gene was statistically associated with distant metastasis (P = .01). hdm2 splicing variants of the hdm2 gene were marginally associated with advanced tumor stage (P = .08) in those colorectal cancer cases with wild-type p53. There was no association between hdm2 splicing variants of the hdm2 gene and cases with mutated p53 (Table 3). Moreover, of 9 cases of colorectal cancer with distant metastasis, 8 cases were positive for hdm2 splicing variants of the hdm2 gene, suggesting that hdm2 splicing variant-associated distant metastasis depended on p53 gene inactivation.
|Stage/Metastasis||hdm2 Spliced Variant (Wild-Type p53)||P||hdm2 Spliced Variant (Mutant p53)||P|
Gene transcriptional variants resulting from alternative mRNA splicing have been found in many tumor suppressors and oncogenes, such as BRCA2,15 Wilms tumor suppressor gene,16 and Hugl-1,17 all of which are associated with human carcinogenesis. Identification of such aberrant variants of gene transcripts will help us understand gene functions and the mechanisms of cancer development.18, 19 To this end, alternative splicing of hdm2 transcripts has been described in glioblastomas,20 astrocytomas,11 ovarian and bladder carcinomas,21 and breast cancer.22 The current study demonstrated for the first time aberrant splicing of hdm2 mRNA in colorectal cancer. We found that 51.6% of cases (32 of 62) and 6 of 10 cell lines had at least 1 hdm2 spliced variant. A total of 4 hdm2 splicing variants were discovered in these cancer tissues and cells. Of these 4 variants, 3 transcripts are novel, and 1 (the 707-bp transcript) is identical to hdm2-B. These hdm2-splicing variants resulted in the deletion of the p53-binding domain in the hdm2 protein. Clinically, the presence of alternatively spliced hdm2 transcripts was significantly associated with advanced Dukes tumor stage (P = .022) and distant metastasis (P = .004) of colorectal cancer as well as poor prognosis. In addition, these alternatively spliced hdm2 transcripts were not associated with p53 gene mutations. The data from the current study clearly demonstrate that alternatively spliced hdm2 transcripts may play a role in colorectal tumorigenesis or cancer progression.
In the present study, the 157- to 292-bp deletion in the hdm2/1338 variant was found in 75% (24 of 32) of colorectal cancer tissues and 5 of 6 of the colorectal cancer cell lines. Expression of this variant was associated with progression of colorectal cancer and poor prognosis. However, the molecular mechanisms underlying these results remain unknown. Theoretically, deletion of the p53-binding domain of the hdm2 protein would have a positive impact on the function of p53 protein, and should have led to a better clinical outcome for patients. Nevertheless, our current data showed just the opposite. The underlying mechanism by which expression of the hdm2/1338 variant led to poor survival and aggressive tumor behavior needs to be further studied.
Furthermore, the variant Hdm2/707 (hdm2-B) is the most frequently expressed hdm2 variant in numerous types of cancer, including ovarian and bladder cancers,21 breast cancer,11 soft-tissue sarcoma,23 and giant cell tumors of the bone.24 In our current study, it was detected in only 4 cases of colorectal cancer tissues and 1 cell line (LoVo). These data suggest that hdm2-B may be a common hdm2 splicing variant of hdm2 in various tumors. In addition, hdm2/1200 with an intact p53-binding domain was detected in 10 cases of colorectal cancer. Sequence analysis revealed the partial lack of the acidic domain in this spliced variant of hdm2. The 610- to 883-bp deletion in the hdm2/1200 variant had similar characteristics to those of the site described and characterized by Brown et al25 as the growth inhibitory domain ID1. Deletion of this region, which contains part of the acidic domain, allowed stable expression of hdm2 in NIH3T3 cells and enhanced the tumorigenic potential of these cells.25 Collectively, the high frequency of aberrant splicing of hdm2 mRNA suggests that aberrant splicing could be an important mechanism by which hdm2 promotes its oncogenic activity in various types of human cancers.
In addition, although we detected p53 gene mutations in these colorectal cancer tissues, there was no statistical association between aberrant hdm2 transcripts and p53 gene mutations. However, we did find that the presence of hdm2 splicing variants was associated with distant metastases of colon cancer (P = .01) and marginally associated with advanced tumor stage (P = .08). The latter occurred in colorectal cancer with wild-type p53, but not in those with mutant p53. Moreover, of 9 cases with distant metastases, 8 cases had both hdm2 splicing variants and wild-type p53. These data indicate that hdm2 splicing variants may function as a negative regulator of p53 even if their p53-binding domain is deleted, because hdm2 has 2 ways to inactivate p53: degradation of p53 protein and inhibition of p53 activity. It remains to be determined which of these is applicable to hdm2 splicing variants. A previous study reported that p53 mutations may contribute to liver metastasis of colon cancer, but no direct confirmation was shown.26 Our findings suggest that p53 accumulation may have been a consequence of aberrant hdm2 splicing, independent of p53 mutation. The data support the hypothesis that hdm2 spliced variants may result in hdm2 protein capable of degrading p53 in the cells. However, mutant p53 is unable to act as a transcription factor or induce apoptosis regardless of the presence of either full-length or splicing variants of hdm2 mRNA.3 In other words, our results suggest that hdm2 or spliced variants of hdm2 act through the suppression of p53 activity to induce cancer development or its progression.
In any event, the data presented in the current study indicate that aberrant spliced hdm2 could be further evaluated as a tumor marker for diagnosis and prognosis of colorectal cancer in clinical practice. To do so, a larger series of cases will be required to verify our current data. Furthermore, down-regulation of the aberrantly spliced variants in tumor cells using antisense oligonucleotides or other techniques may be developed as a novel antitumor strategy for potential chemotherapy of human cancer.27 However, we will first explore the effects of these splicing variants in colorectal cancer.
In summary, the data from the current study clearly demonstrate the presence of aberrant hdm2 splicing variants in colorectal cancer, and their correlation with advanced tumor stages (P = .022), distant metastases (P = .004), and poor survival of patients (P = .039). The hdm2 splicing variants may be further evaluated as tumor markers for diagnosis and prognosis of colorectal cancer as well as for a novel anticancer therapeutic strategy. Further studies will explore the mechanism and pathologic consequences of aberrant hdm2 splicing in colorectal cancer.
This study was supported by the Special Foundation of the President of the Third Affiliated Hospital of Harbin Medical University.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
- 10American Cancer Society. Global Facts and Figures 2007. Available at: http://www.cancer.org/acs/groups/content/@nho/documents/document/globalfactsandfigures2007rev2p.pdf/. Accessed on July 7, 2011.
- 14An alternative splice form of Mdm2 induces p53-independent cell growth and tumorigenesis. J Biol Chem. 2004; 6: 4877-4886., , , et al.
- 16The Wilms tumor gene, WT1, in normal and abnormal nephrogenesis. Pediatr Nephrol. 1999; 1: 620-625..