Somatic Mutations of the PPP2R1B Candidate Tumor Suppressor Gene at Chromosome 11q23 are Infrequent in Ovarian Carcinomas

Abstract

Previous studies have demonstrated frequent allelic losses of distal chromosome 11q in ovarian carcinomas. The tumor suppressor gene(s) presumably targeted by these losses have not yet been identified. PPP2R1B is a candidate tumor suppressor gene at 11q23 that has recently been shown to be mutated in a subset of colorectal and lung cancers. We evaluated 5 ovarian carcinoma cell lines and 27 primary ovarian carcinomas for allelic losses of 11q23 and for mutations in the open reading frame of PPP2R1B. We also evaluated the primary tumors for allelic losses at 17p13, another chromosomal region frequently affected by losses of heterozygosity (LOH) in ovarian cancers. 11q23 and 17p13 allelic losses were identified in 25% and 74% of the carcinomas, respectively. No mutations within PPP2R1B coding sequences were found. These findings indicate that mutations of the PPP2R1B gene are infrequent in ovarian cancer and that deletions affecting the distal portion of chromosome 11q in ovarian cancer likely target inactivation of other genes.

Introduction

Tumor suppressor genes are targeted by loss-of-function mutations in cancer cells. Some tumor suppressor genes, such as p53 and p16/lnk4a, are inactivated in a wide variety of cancers, whereas others, such as APC and MEN1, are mutated only in selected tumor types [1]. Although much effort has been directed at identifying tumor suppressor genes that are frequently inactivated in ovarian cancer, success has been rather limited. To date, p53 is the only tumor suppressor gene definitively shown to be inactivated in a large percentage of ovarian carcinomas, with more than 50% harboring inactivating mutations of p53 [2–4]. Several groups have attempted to localize other tumor suppressor genes important in ovarian cancer pathogenesis by identifying specific chromosomal regions that are often deleted in ovarian carcinomas. Such deletions can be detected as allelic losses or losses of heterozygosity (LOH) in tumor DNA when compared to normal DNA from the same patient. Several previous studies evaluating allelic losses in ovarian carcinomas have shown frequent (>50%) LOH of distal chromosome 11q [5–8]. The tumor suppressor gene(s) presumably targeted by these deletions have not yet been identified.

PPP2R1B is a candidate tumor suppressor gene at 11q23 [[9],http://www.ncbi.nlm.gov/UniGene/index.html] that was recently shown to be altered in a subset of human lung and colon cancers, but not in breast or cervical cancers [10]. The PPP2R1B gene encodes the β isoform of the A subunit of the serine/threonine protein phosphatase 2A. Wang et al. [10], Lee et al. [11], and Li et al. [12] have proposed that PP2A inactivation may contribute to tumorigenesis by disrupting cell cycle regulation or regulation of telomerase activity. To determine whether PPP2R1B is the gene targeted by 11q LOH in ovarian cancers, we evaluated 27 human ovarian carcinomas and 5 ovarian carcinoma-derived cell lines for allelic losses at 11q23 and mutations in the coding region of PPP2R1B. Because LOH at 17p13 has been found in more than 50% of ovarian carcinomas [13–16], we also evaluated our set of primary carcinomas for 17p13 LOH.

Materials and Methods

Cell Lines and Tumor Specimens

Five ovarian carcinoma cell lines (CaOV3, CaOV4, NIHOVCAR3, SKOV3 and SW626) and 27 primary ovarian carcinomas were analyzed for mutations in the coding region of PPP2R1B using a reverse transcription-polymerase chain reaction (RT-PCR)-based strategy. The primary tumors included all of the major histologic subtypes of ovarian carcinoma (12 serous, 5 endometrioid, 5 clear cell and 5 mucinous).

RNA Extraction, RT-PCR and Sequencing

Total mRNA was isolated from cultured cells or microdissected frozen tissue sections using Trizol according to the manufacturer's protocol (Life Technologies, Gaithersburg, MD). All primary tumor samples analyzed contained at least 70% tumor cells. In most cases, first strand cDNA was synthesized from DNAsel-treated mRNA samples using random hexamer primers (Pharmacia Biotech, Piscataway, NJ) and Superscript II (Life Technologies). Parallel reactions lacking Superscript served as negative controls. For the remaining cases, “One-Step” RT-PCR was carried out using Superscript II-RT/Taq (Life Technologies) according to the manufacturer's protocol. Nine pairs of exon-based primers (Table 1) were designed to amplify overlapping segments of the PPP2R1B cDNA, encompassing the entire open reading frame. RT-PCR products were evaluated for size alterations on ethidium-stained agarose gels, then sequenced in their entirety using one or both primers from the original PCR.

Table 1.

PCR Primer Sequences for PPP2R1B cDNA Amplification.

Sequences (5′→3′)	Size (bp)	Exon
F1: AGGAGGAGAAAGAACATGGC	304	1,2,3
R1: AGTCAGGACCTCCCACTAGG
F2: TAGCTCTTGCTGAGCAGC	282	3,4
R2: CATTTGATGCCCTGGGATAG
F3: GTTCAGCGTTTGCTATCCCA	287	4,5,6
R3: GCATCACCAAAGTCTCAAGG
F4: GCTTGTGTCAGTATTGCCCA	269	6,7,8
R4: TCAATGGGCAAGTTCTCACC
F5: TGCTGCCCACAAAGTAAAAG	319	8,9,10
R5: GGCAGGAAGGAGAGACTGAG
F6: GAAGTGATTGGAATCCGTCG	258	10,11,12
R6: GGCCCACTCTGTACCAAAC
F7: CCACCAACAACCTCATGAAA	245	12,13
R7: AGATTTGGCCACATTGAAGC
F8: TTAAAAATGGCAGGAGACCA	320	13,14,15
R8: AGTTGTTGCCATTTGCTTGG

DNA Extraction and LOH Analyses

All of the primary ovarian carcinomas were evaluated for allelic losses at chromosome 11q23 using two microsatellite markers flanking the PPP2R1B gene (D11S1647 and D11S1987, separated by 15 centiray) [http://www.ncbi.nlm.gov/UniGene/index.html, http://www.marshmed.org/genetics/], [10]. LOH at 17p13 was evaluated using microsatellite markers flanking the p53 gene (D17S1303 and D17S1308). D17S1303 is approximately 0.5 Mb centromeric to the p53 locus, whereas D17S1308 is approximately 2 Mb telomeric to p53. DNA was extracted from microdissected tumor and matched normal frozen tissue samples using standard proteinase K digestion and phenol/chloroform extraction. Map-Pair primers (Research Genetics, Huntsville, AL) were used to amplify each 11q or 17p locus. PCR products were resolved by electrophoresis on sequencing-type polyacrylamide gels. LOH was scored when there was a relative decrease (>50%) in the intensity of the signal of one allele in the tumor compared to the allele signals in matched normal DNA. Tumors were scored as not informative when only one allele was present in DNA from matched normal tissue. DNA samples showing novel bands of the type observed in tumors with the microsatellite instability phenotype were not scored for LOH at the loci with the band shifts [17].

Results

Direct sequencing of PPP2R1B cDNAs showed that none of the five ovarian carcinoma cell lines or the 27 primary tumors contained mutations within the PPP2R1B open reading frame. The results of the LOH analyses are summarized in Table 2 and representative examples are shown in Figure 1. Allelic losses at the D11S1647 and D11S1987 loci were observed in 5/22 (23%) and 4/16 (25%) informative cases, respectively. LOH at 11q23 was observed at roughly comparable frequencies in all histologic subtypes of ovarian carcinoma. One endometrioid carcinoma showed novel bands (shifts) at both 11q loci evaluated, consistent with the microsatellite instability phenotype that has been observed with increased frequency in endometrioid ovarian carcinomas [18]. LOH at the D17S1303 and D17S1308 loci were observed in 14/19 (74%) and 12/18 (67%) of informative cases, respectively. The serous carcinomas showed the highest frequency of 17p13 allelic losses, consistent with the results of previous studies [4]. The frequent 17p allelic losses observed in our specimens confirm that our DNA samples were isolated from regions containing primarily neoplastic cells and that 11q LOH was not underestimated because of technical reasons.

Table 2.

LOH Analysis of Primary Ovarian Carcinomas.

Case No.	Tumor Type	D11S1647	D11S1987	D17S1303	D17S1308
1	Serous	-	-	LOH	LOH
2	Serous	-	-	LOH	LOH
3	Serous	-	NI	LOH	LOH
4	Serous	LOH	LOH	-	LOH
5	Serous	-	NI	LOH	LOH
6	Serous	-	NI	LOH	LOH
7	Serous	-	-	LOH	LOH
8	Serous	LOH	LOH	LOH	LOH
9	Serous	LOH	NI	LOH	NI
10	Serous	-	-	-	NI
11	Serous	-	-	NI	LOH
12	Serous	NI	NI	NI	LOH
13	Endometrioid	-	NI	LOH	NI
14	Endometrioid	NI	-	LOH	NI
15	Endometrioid	-	NI	LOH	NI
16	Endometrioid	LOH	LOH	NI	-
17	Endometrioid	Shift	Shift	NI	-
18	Clear cell	LOH	-	-	NI
19	Clear cell	-	-	LOH	LOH
20	Clear cell	-	NI	-	-
21	Clear cell	-	NI	NI	-
22	Clear cell	-	-	-	-
23	Mucinous	NI	LOH	LOH	LOH
24	Mucinous	NI	NI	NI	-
25	Mucinous	-	-	LOH	NI
26	Mucinous	-	-	NI	NI
27	Mucinous	-	-	NI	NI

LOH: loss of heterozygosity (-): Retention of heterozygosity NI: not informative. Shift: evidence for microsatellite instability phenotype.

Figure 1.

Representative tumors showing losses of heterozygosity at 11q23 and 17p13. Allelic losses are indicated by arrowheads. N: Normal DNA; T: Tumor DNA.

Discussion

Our findings provide evidence that point mutations in the transcribed portions of the PPP2R1B gene are uncommon in ovarian carcinomas. However, we are unable to exclude the possibility that other mechanisms might inactivate the gene or its protein product in some ovarian cancers. For example, homozygous deletions of PPP2R1B or mutations resulting in unstable transcripts could be missed by our RT-PCR-based mutation detection strategy. In either case, the wild-type PPP2R1B transcripts identified in our assay would then reflect the contribution of non-neoplastic cells in the primary tumor specimen. However, the absence of detectable mutations in ovarian cancer cell lines and the high percentage (>70%) of neoplastic cells in our primary tumor specimens argues against these possibilities. It is also worth noting that mutations of PPP2R1B were identified in only 15% of the primary lung and colon carcinomas evaluated by Wang et al. [10]. Hence, our study could have missed a comparably small percentage of mutations because of the limited sample size.

In the single published study of PPP2R1B mutations in human cancer, only 3 of 11 tumors with PPP2R1B mutations showed homozygosity for the mutant allele, suggesting that gross deletions are an infrequent means of inactivating PPP2R1B. Curiously, biallelic inactivation of PPP2R1B could not be demonstrated in 4 of the 11 tumors [10]. We identified 11q23 LOH in 25% of the ovarian carcinomas evaluated, and all lacked PPP2R1B gene mutations. Other studies have found distal 11q deletions in a comparably small percentage of ovarian carcinomas [16,19–21]. It is possible that the wide range in the frequencies of distal 11q LOH reported in ovarian carcinomas may reflect the specific loci examined and/or the variable distribution of histologic types amongst the groups of ovarian tumors analyzed. Several previous studies have demonstrated that specific genetic alterations are seen with variable frequency in the different ovarian carcinoma subtypes. For example, K-ras mutations have been observed more frequently in mucinous carcinomas, whereas p53 mutations and β-catenin mutations are more common in serous and endometrioid carcinomas, respectively [4,22–24].

In summary, our findings indicate that mutations of the PPP2R1B gene are infrequent in ovarian cancer and that deletions affecting the distal portion of the chromosome 11q in ovarian cancer may likely target other genes for inactivation. Additional studies of ovarian cancer and other tumor types will be needed to further define the role, if any, of PPP2R1B gene mutations in human cancer.