Abnormalities of the APC/β-catenin pathway in endometrial cancer

Article metrics

Abstract

The activation of the APC/β-catenin signalling pathway due to β-catenin mutations has been implicated in the development of a subset of endometrial carcinomas (ECs). However, up to 25% of ECs have β-catenin nuclear accumulation without evidence of β-catenin mutations, suggesting alterations of other molecules that can modulate the Wnt pathway, such as APC, γ-catenin, AXIN1 and AXIN2. We investigated the expression pattern of β- and γ-catenin in a group of 128 endometrial carcinomas, including 95 endometrioid endometrial carcinomas (EECs) and 33 non-endometrioid endometrial carcinomas (NEECs). In addition, we evaluated the presence of loss of heterozygosity and promoter hypermethylation of the APC gene and mutations in the APC, β- and γ-catenin, AXIN1, AXIN2, and RAS genes, and phospho-Akt expression. No APC mutations were detected but LOH at the APC locus was found in 24.3% of informative cases. APC promoter 1A hypermethylation was observed in 46.6% of ECs, and was associated with the endometrioid phenotype (P=0.034) and microsatellite instability (P=0.008). Neither LOH nor promoter hypermethylation of APC was associated with nuclear catenin expression. Nuclear β-catenin expression was found in 31.2% of EECs and 3% of NEECs (P=0.002), and was significantly associated with β-catenin gene exon 3 mutations (P<0.0001). β-catenin gene exon 3 mutations were associated with the endometrioid phenotype, and were detected in 14 (14.9%) EECs, but in none of the NEECs (P=0.02). γ-catenin nuclear expression was found in 10 ECs; it was not associated with the histological type but was associated with more advanced stages (P=0.042). No mutations in γ-catenin, AXIN1 and 2 genes were detected in this series. Neither RAS mutations nor phospho-Akt expression, which were found in 16 and 27.6% of the cases, respectively, were associated with β-catenin nuclear expression. Our results demonstrated a high prevalence of alterations in molecules of the APC/β-catenin pathway, but only mutations in β-catenin gene are associated with aberrant nuclear localization of β-catenin.

Introduction

The APC/β-catenin signalling pathway plays an important role in normal and tumoral cells (Willert and Nusse, 1998; Sharpe et al., 2001). In the absence of an extracellular Wnt signal in normal cells, the free (cytoplasmic) β-catenin level is low, since the protein is targeted for destruction in the ubiquitin-proteasome system after phosphorylation by glycogen synthase kinase-3β (GSK-3β) (Aberle et al., 1997). The latter forms a complex with adenomatous polyposis coli (APC) protein (Rubinfeld et al., 1996) and other proteins, such as AXIN1 (Behrens et al., 1998), AXIN2 (Liu et al., 2000) and protein phosphatase 2A (PP2A) (Hsu et al., 1999). The most common molecular alterations in tumour cells leading to disruption of β-catenin degradation are mutations that inactivate APC or activate β-catenin itself (Morin et al., 1997). These alterations produce an accumulation of cytoplasmic β-catenin that translocates into the nucleus and, interacting with members of the lymphoid enhancer factor-1/T-cell factor (Lef-1/Tcf) (Korinek et al., 1997), activates transcription of various genes, such as cyclin D1 (Tetsu and Mccormick, 1999) and MYC (He et al., 1998). The APC/β-catenin signalling pathway has mainly been studied in the commonest type of endometrial carcinoma (EC), the endometrioid endometrial carcinoma (EEC). In these tumors, mutations of APC have not been detected (Fujino et al., 1994; Nei et al., 1999; Schosshauer et al., 2000), but β-catenin mutations occurred in around 17% of cases (Fukuchi et al., 1998; Mirabelli-Primdahl et al., 1999; Kobayashi et al., 1999; Ikeda et al., 2000; Schosshauer et al., 2000; Saegusa et al., 2001; Saegnsa and Okayasu, 2001). However, around 25% of tumours featured β-catenin nuclear accumulation without evidence of β-catenin mutations (Fukuchi et al., 1998), suggesting alterations in molecules of the APC/β-catenin pathway other than APC or β-catenin mutations. Recently, it has been demonstrated that mutations affecting AXIN1 and AXIN2 genes activate the Wnt pathway and are involved in the development of a subset of hepatic (Satoh et al., 2000), colon (Liu et al., 2000), and ovarian carcinomas (Wu et al., 2001) and medulloblastomas (Dahmen et al., 2001). In colon cancer, both β-catenin (Mirabelli-Primdahl et al., 1999) and AXIN2 (Liu et al., 2000) mutations are associated with a replication error (RER) phenotype in tumours with microsatellite instability (MSI), a phenotype that is present in about 20% of ECs (Matías-Guiu et al., 2001). In addition, nuclear expression of γ-catenin, which is closely related to β-catenin structurally and functionally, has been reported in endometrial cancer (Palacios et al., 2001) and wild-type- and mutated γ-catenin-induced transforming activity dependent on Tcf/Lef function (Kolligs et al., 2000; Caca et al., 1999). Finally, a relationship between the phosphoinositide 3-OH kinase AKT (PI3K-Akt) pathway and the levels of β-catenin has been suggested. Thus, mutated RAS, which occurs relatively frequently in ECs, produced in vitro β-catenin stablization and an increment in Tcf-Lef-1-dependent transcriptional activity by interacting with PI3K (Espada et al., 1999). Similarly, activation of the PI3-Akt pathway by different mechanisms has been associated with nuclear accumulation of β-catenin in some neoplasias (Desbois-Mouthon et al., 2001; Satyamoorthy et al., 2001).

The aim of this study was to determine the molecular alterations underlying activation of the APC/β-catenin signalling pathway in EC. We have studied the expression pattern of β- and γ-catenin in a set of 128 ECs and have analysed the association between nuclear catenin expression and some of the molecular alterations that seem to be relevant in human neoplasia in vivo. Specifically, we screened for the presence of mutations, LOH and promoter hypermethylation of the APC gene, mutations in β- and γ-catenin, AXIN1, AXIN2 and RAS and expression of phospho-Akt.

Results

β- and γ-catenin and phospho-Akt expression

A summary of clinicopathological, immunohistochemical and molecular features of this series is presented in Table 1. Nuclear β-catenin was observed in 30 ECs. The percentage of neoplastic cells with nuclear β-catenin expression varied from around 5% to more than 75% in different cases (Figure 1a,b). Cells with nuclear β-catenin expression tended to form small solid groups, disrupting the usual glandular pattern of the tumour. Squamous morules always showed β-catenin nuclear expression (Figure 1a.1), whereas mature squamous metaplasia exhibited nuclear immunostaining in six out of 30 cases. We found a statistically significant association between β-catenin nuclear expression and the endometrioid phenotype (P=0.002), whereby 29 out of 93 (31.2%) EECs expressed nuclear β-catenin, but only one out of 33 (3%) NEECs did. The only NEEC that expressed nuclear β-catenin was a mixed (endometrioid/serous) carcinoma in which nuclear immunoexpression was observed in areas of squamous metaplasia in the endometrioid component.

Table 1 Summary of clinical, pathological, immunohistochemical, and molecular features
Figure 1
figure1

Expression of β-, γ-catenin and phospho-Akt in endometrial carcinomas. (a) A case with β-catenin gene mutation shows β-catenin expression in the nucleus of morule cells but in the membrane of neoplastic glands. (b) Nuclear expression is evident in areas of a poorly differentiated endometrioid carcinoma. (c) Nuclear γ-catenin expression in a clear cell endometrial carcinoma. (d) Phospho-Akt expression in the nuclei and cytoplasm of a endometrial carcinoma

Nuclear γ-catenin was observed in 25–50% of the cells in 10 out of 89 (11.2%) ECs, five out of 62 (8.1%) EECs and five out of 27 (18.5%) NEECs (Figure 1c). There was no statistical association between nuclear γ-catenin expression and histological type (P=0.151); however, nuclear γ-catenin expression showed an association with tumour stage (P=0.042) whereby nuclear γ-catenin expression was more frequent in stages higher than I (66.6% in stage II–IV versus 28.4% in stage I). No other associations were observed between γ-catenin expression and clinicopathological or molecular features (analysis not shown).

Immunohistochemical expression of phospho-Akt was evaluated in 76 ECs, including 15 cases with nuclear β-catenin expression, six of which had β-catenin gene mutation. We observed positive staining in 21 cases (27.6%) (Figure 1d), but did not find any statistically significant association between Akt expression and clinicopathological, immunohistochemical and molecular features (data not shown).

Genetic and epigenetic analysis of APC

LOH at the APC locus was evaluated in 61 ECs, using microsatellite markers D5S346 and D5S429. Excluding 24 non-informative cases, we observed APC LOH in nine out of 37 (24.3%) tumours (five out of 19 EECs and four out of 13 NEECs) (Figure 2). No associations were observed between APC LOH and clinicopathological or molecular features (data not shown). We investigated APC promoter 1-A methylation status in 103 ECs, and observed promoter hypermethylation in 48 cases (46.6%). There was a statistically significant association (P=0.034) between APC promoter 1-A hypermethylation and the histological type, whereby the hypermethylation was more frequent in EECs (38 cases, 54.3%) than in NEECs (10 cases, 30.3%). APC promoter 1-A hypermethylation was significantly more frequent in tumours showing MSI (16 cases, 72.7%) that in MSI-negative ECs (32 cases, 39.5%) (P=0.008). This association was also significant (P=0.028) in EECs, the group in which MSI was also most frequent. APC LOH and/or promoter hypermethylation were not statistically associated with the nuclear expression of catenins or other molecular abnormalities (data not shown).

Figure 2
figure2

Examples of LOH with microsatellite marker D5S346 and promoter hypermethylation of the APC gene (U: unmethylated; M: methylated; NL: normal lymphocytes; IVD: in vitro-methylated DNA)

Mutational analysis of the APC sequence including the mutation cluster region was carried out in 44 ECs that had nuclear catenin expression and/or APC LOH and/or promoter hypermethylation. The study only revealed a single nucleotide polymorphism (SNP) (Table 2).

Table 2 Summary of genetic changes

β-catenin, γ-catenin, AXIN1, AXIN2, and RAS mutations and microsatellite instability

Single-base substitutions in exon 3 of β-catenin gene were detected in 14 out of 94 (14.9%) EECs but in none of the 33 NEECs, this difference being statistically significant (P=0.02). Changes more frequently affected serine 33 or adjacent residues (Table 2). All the other aminoacid targets for GSK3β phosphorylation (serine 37 and 45 and threonine 41) were also mutated in some cases. We observed a strong association between β-catenin mutation and β-catenin nuclear accumulation. Thus, all 14 cases (100%) with β-catenin mutations showed nuclear catenin accumulation, but only 16 (14.4%) of the tumours without mutation did so (P=0.001). On the other hand, out of 30 cases with nuclear β-catenin expression, only 14 (46.6%) were explained by gene mutation.

We carried out a mutational analysis of β-catenin on DNA extracted after laser microdissection from areas with and without nuclear β-catenin expression in 14 ECs with the aim of evaluating the possible focal nature of some mutations. In nine cases with previously detected β-catenin mutation, the same mutations were detected after microdissection in highly separated areas with and without β-catenin nuclear expression (Figure 3). On the other hand, we detected no mutations after analysing microdissected foci of β-catenin nuclear expression in five cases that did not show mutations in the genetic analysis of the complete tissue section.

Figure 3
figure3

Exon 3 β-catenin analysis in groups of cells procured by laser microdissection after immunostaining. The same mutation was detected in distant cells independently of nuclear β-catenin expression

In contrast to β-catenin, we detected no γ-catenin mutations in any ECs. We screened 35 ECs for AXIN1 and AXIN2 mutations where either β- of γ-catenin nuclear expression had been detected. We found a total of 15 base substitutions (Table 2), 13 of which were silent and two caused aminoacid substitutions. However, all nucleotide changes were also detected in the normal tissue analysed.

We detected RAS mutation in 16 out of 85 (18.8%) EECs and three out of 33 (9.1%) NEECs, although this difference was not statistically significant (P=0.312). All mutations affected codons 12 or 13 of K-RAS (Table 2). No mutations in codon 61 of K-RAS, or codons 12, 13 and 61 of H- and N-RAS were detected. No associations were observed between Ras mutations and catenin expression and clinicopathological or molecular features (data not shown).

MI was detected in 27 out of 92 (29.3%) EECs and 2 out of 33 (6.1%) NEECs, this difference being statistically significant (P=0.013). No other associations were observed between MSI and clinicopathological or molecular features (analysis not shown).

Discussion

The analysis of β-catenin nuclear expression and mutations in this large sample of NEECs and EECs demonstrated that both alterations are characteristic of the endometrioid phenotype. In the present series, 15% of EECs had β-catenin mutations in one of the serine/threonine residues targeted for phosphorylation by GSK-3β or some adjacent residues, a frequency similar to the average frequency of 17% found in previous series (Matías-Guiu et al., 2001). We found no significant relationships between β-catenin gene mutation and the clinicopathological features of age, grade and stage. However, other series have exhibited an association with early onset (Ikeda et al., 2000), low grade and absence of lymph node metastases (Saegusa et al., 2001), suggesting that β-catenin mutations might occur in a subset of less aggressive tumours, as seems to occur in ovarian carcinomas (Gamallo et al., 1999).

Nuclear β-catenin expression was observed only focally in some cases even in some that had gene mutation. In prostate cancer, β-catenin mutations are known to occur focally within the tumour, suggesting that they may appear during tumour progression in some cases (Voeller et al., 1998). To evaluate this possibility in ECs, we carried out a laser microdissection study of areas with and without nuclear immunostaining. The genetic status of β-catenin was the same in highly separated neoplastic cells whether or not they expressed nuclear β-catenin. Similar observations have been made in the colonic adenoma–carcinoma sequence. Small adenomas lack detectable nuclear β-catenin, despite the presence of APC gene mutations. However, nuclear β-catenin in large adenomas is observed in areas of tubular branching and in the invasion front in most adenocarcinomas (Kirchner and Brabletz, 2000). These data indicate that β-catenin gene mutations are not sufficient by themselves to cause accumulation of nuclear β-catenin and suggest the existence of additional modulator mechanisms of β-catenin degradation.

γ-catenin is a protein of the Armadillo family, closely related to β-catenin. γ-catenin also binds to E-cadherin, α-catenin and APC and can translocate to the nucleus, activating Tcf-dependent transcription (Caca et al., 1999; Kolligs et al., 2000). Nuclear γ-catenin has been observed in some colorectal polyps (Valizadeh et al., 1997), oesophageal adenocarcinomas (Bailey et al., 1998), and recently in 7.5% of ECs (Palacios et al., 2001). The results suggest that progression of some ECs might be mediated by Wnt activation through γ-catenin hyperexpression and nuclear translocation. Since we observed no association between γ-catenin nuclear expression and APC alterations, we examined whether γ-catenin nuclear expression was due to mutations in the phosphorylation sequence by GSK-3β, as occurs in the case of β-catenin. However, we did not find any γ-catenin mutations, as previously observed in colon (Spark et al., 1998) and ovarian (Wu et al., 2001) cancers and in cell lines originating from various tissues (Ueda et al., 2001). To date, only two γ-catenin mutations have been reported in one gastric carcinoma cell line (Caca et al., 1999) and one squamous-cell lung carcinoma cell line (Ueda et al., 2001). The absence of mutations at the NH2-regulatory region does not rule out the possibility of oncogenic potential of γ-catenin nuclear translocation. It has been demonstrated that wild-type γ-catenin functions as an oncogene when deregulated, in contrast with β-catenin, which requires amino-terminal mutations. Overexpressed γ-catenin strongly activated c-Myc expression, and c-Myc function was essential for γ-catenin transformation (Kolligs et al., 2000).

Since some ECs showed β- or γ-catenin nuclear expression without underlying gene mutation, we tried to identify other molecular alterations that could explain this finding. Unlike in colon cancer, present study indicates that APC mutations are absent or infrequent in endometrial cancer, at least in the sequence of exon 15 including the mutation cluster region. Although previous data about APC mutations in EC are limited, they confirm our observation. Thus, Schosshauer et al. (2000) analysed 32 tumours by protein truncation assay and did not demonstrate mutations between codons 686 and 1693. Nei et al. (1999) analysed APC protein by Western blot in 4 tumours and detected only wild-type protein, without evidence of additional truncated proteins. Finally, Fujino et al. (1994) stated that they failed to find any evidence of APC mutations in a thorough mutational survey of EC, although they did not present data (see their comment on page 4297 and footnote 3).

It has been reported that alleic losses at 5q21 are also infrequent in EC. In the current series, LOH at the APC locus was found in 24.3% of 37 tumours, a higher frequency than previously reported. Thus, Fujino et al. (1994) observed LOH at the APC locus in only one of 22 ECs and Jones et al. (1994) did not find LOH in seven tumours analysed. Although our sample included 13 (35%) NEECs, there was no association with the histological type or other clinicopathological features that would explain this discrepancy. Differences between studies probably reflect differences in the number of cases studied and the number of markers used, since, in contrast with previous series that used only a single marker, we analysed a larger sample with two markers.

Aberrant DNA hypermethylation of promoter region CpG island can serve as an alternative to gene mutation and allelic loss for gene silencing of tumour suppressor genes. APC promoter 1-A methylation associated with lack of expression of its transcript has been reported in several human tumours (Esteller et al., 2000; Kawakami et al., 2000; Tsuchiya et al., 2000; Dong et al., 2001; Jin et al., 2001; Virmani et al., 2001). We investigated this epigenetic alteration in 103 ECs, and observed promoter hypermethylation in 48 cases (46.6%). APC promoter 1-A hypermethylation was significantly more frequent in tumours showing MSI than in MSI-negative ECs. This association was also significant in EECs, the group in which MSI was most frequent. Very recently, Zysman et al. (2002) have also observed this association in their series of EC. Since it has been established that MSI in endometrial cancer is the result of hMLH1 promoter hypermethylation in nearly 90% of cases (Esteller et al., 1998), this association may indicate the simultaneous hypermethylation of both genes in ECs.

Although we did not find any statistically significant association between the presence of APC gene alterations and the nuclear expression of β-catenin, this does not rule out a role for APC in endometrial carcinogenesis. APC gene encodes a large protein with multiple cellular functions and interactions that include not only a role in the Wnt signalling pathway but also in intercellular adhesion, stabilising of the cytoskeleton, and probably regulation of the cell cycle and apoptosis (Fearnhead et al., 2001). In fact, it has recently been reported that a predominant fraction of APC associates with the apical membrane in a variety of epithelial cell types. This APC pool seems not to be involved in the degradation of β-catenin (Reinacher-Schick and Gumbiner, 2001).

The current study is the first to examine the possible involvement of AXIN1 and AXIN2 in endometrial carcinogenesis. AXIN1 is a multidomain protein that contains binding sites for GSK3β, β-catenin, APC, PP2A, Dishevelled, and itself (Behrens et al., 1998; Hsu et al., 1999). Axin is thought to act as a molecular scaffold for GSK3-dependent turnover of β- and γ-catenin. Recently, it has been reported that AXIN1 gene is inactivated in some hepatocellular carcinomas and medulloblastomas. Thus, Satoh et al. (2000) described six AXIN1 mutations in six tumours, three of which were point mutations and three were small deletions. These tumours did not have β-catenin mutations and there was LOH at the AXIN1 locus in three cases, implying that it plays a role as a tumour suppressor gene. In medulloblastomas, Dahmen et al. (2001) reported a single point mutation and seven large deletions (12%) in 85 tumours. In addition, a single point mutation has been identified in an endometrioid ovarian carcinoma (Wu et al., 2001) and in some colon cancer cell lines (Webster et al., 2000). We screened 35 endometrial carcinomas for AXIN1 mutations in which either β- or γ-catenin nuclear expression had been detected and did not find any mutations in AXIN1 but observed 12 single-nucleotide polymorphisms (SNPs). SNPs, with and without aminoacid changes, have been described in exons 1, 4, 5, 9 and 10 and intron 4 of the AXIN1 gene (Lin et al., 2000; Webster et al., 2000).

Liu et al. (2000) reported that AXIN2 was mutated in 11 out of 45 colorectal carcinomas with MSI, but in none out of 60 tumours without MSI. All eleven were frameshift mutations in the four mononucleotide repeat sequences located in exon 7. The authors also observed β-catenin nuclear accumulation in ten out of eleven colorectal carcinomas with AXIN2 mutations. The mutations did not occur in tumours with APC or β-catenin mutations. In vitro studies with transfection of mutant AXIN2 demonstrated that it was more stable than wild-type protein and activated TCF-dependent transcription, suggesting a dominant-negative effect. We analysed mononucleotide repeat sequences in exon 7 in all endometrial lesions and found three SNPs, one causing an aminoacid change, but we did not detect any mutations, in spite of the fact that nearly 25% of ECs had MSI. However, the genes affected by frameshift mutations might be different in colon and endometrial cancers. For example, TCF4, another component of the APC/β-catenin pathway, is frequently mutated in an (A)9 coding repeat in MSI-positive colon cancer, whereas this mutation is very infrequent in MSI-positive endometrial carcinomas (Duval et al., 1999).

A link between RAS mutations and β-catenin signalling has recently been suggested. In vitro studies have demonstrated that RAS influences the localization and activity of β-catenin by dismantling cadherin-catenin complexes, stabilizing β-catenin and promoting its nuclear accumulation. This effect is mediated PI3K and involves the formation of β-catenin-PI3K complexes and the inhibition of β-catenin-APC interaction (Espada et al., 1999). García-Rostán et al. (2001) have reported that, in tumour samples, H-, K- or N-RAS mutations are present in 66% of thyroid carcinomas with nuclear β-catenin expression without β-catenin gene mutation. In the current series, only three out of 19 (15.8%) ECs with KRAS mutations had focal nuclear β-catenin expression. Although our study cannot rule out a functional relationship between RAS and β-catenin, our observations suggest that RAS plays a major role in modulating the Wnt pathway in ECs.

It has been reported that after stimulation by insulin and/or insulin growth factor-1 the PI3K/Akt signal induces phosphorylation and inactivation of GSK3β, resulting in increased nuclear levels of β-catenin and stimulation of TCF-dependent transcription in liver cancer (Desbois-Mouthon et al., 2001) as well as in melanoma (Satyamoorthy et al., 2001). Our immunohistochemical study of ECs, evaluating the expression pf phospho-Akt, was not consistent with these in vitro results since no association was observed with nuclear β-catenin expression. Our observations are more in accordance with other studies reporting that after induction of Akt phosphorylation there was no significant effect on TCF activation (Chesire et al., 2002) and that activation of Akt alone was not sufficient to mimic to the Wnt pathway (Yuan et al., 1999). Differential regulation of GSK3b by insulin (which is mediated by Akt) and Wnt signalling has also been observed, leading to distinct downstream events (Ding et al., 2000). Finally, it must be taken into account that a decrease in GSK3β activity causes changes in β-catenin accumulation depending on the cell type studied (Staal et al., 1999).

In summary, the current study demonstrated a high prevalence of genetic, epigenetic and expression alterations of molecules of the APC/β-catenin pathway in endometrial cancer. The analysis of this large sample of NEECs and EECs demonstrates that APC promoter 1A hypermethylation, β-catenin nuclear expression and β-catenin mutations are characteristic of the endometrioid phenotype, supporting the current view of these two types of tumours as distinct entities. However, APC alterations are not functionally equivalent to β-catenin mutations in endometrial carcinogenesis. γ-catenin nuclear expression, without underlying γ-catenin mutations, was observed in carcinomas of advanced stages, suggesting a role for this molecule in the progression of EC. Mutations in AXIN1 and AXIN2 genes seem to be infrequent in endometrial lesions. Moreover, neither RAS mutations nor phosphorylate-Akt expression explains β-catenin nuclear accumulation.

Materials and methods

Tissue samples

This study comprises 128 ECs: 95 endometrioid endometrial carcinomas (EECs) and 33 non-endometrioid endometrial carcinomas (NEECs). Mean age at diagnosis was 62.7±12.3 years (range 29–89 years). Non-endometrioid cases included 5 clear cell carcinomas, 15 serous carcinomas, 7 serous-clear carcinomas and 6 endometrioid-serous carcinomas. Clinicopathological features are presented in Table 1. Some clinicopathological clinical and molecular features of 33 cases have been previously reported (Palacios et al., 2001).

Immunohistochemistry

All immunostaining was performed on paraffin-embedded tissue sections, using a heat-induced antigen retrieval step prior to exposure to the primary antibody (the slides were heated in a pressure cooker for 3 min in a 10 mM sodium citrate solution, pH 6.5). Mouse anti-human β-catenin and γ-catenin monoclonal antibodies (Transduction Laboratories, Lexington, KY, USA) were applied to sections at 1 : 1000 dilution. After incubation with the primary antibody, immunodetection was performed using biotinylated anti-mouse immunoglobulins and peroxidase-labeled streptavidin (LSAB-DAKO, Glostrup, Denmark) with the diaminobenobenzidine as the chromogen. Immunostaining was considered to be nuclear for catenins when they were detected in at least 5% of the neoplastic nuclei. The polyclonal Phospho-Akt (Ser473) antibody (IHC Specific) (Cell Signalling Technology, Beverly, MA, USA) was applied at 1 : 25 dilution and the EnVision System (Dako) for polyclonal antibodies was used for detection. Cases were considered positive for phospho-Akt expression when cytoplasmic staining was accompanied by nuclear staining in at least 5% of tumour nuclei.

Molecular studies

All molecular studies were carried out on formalin-fixed paraffin-embedded tissue. Blocks from ECs and corresponding normal tissue (myometrium) were cut at 10 μm and reviewed microscopically. When necessary, tumour tissue was manually microdissected from areas with >75% of tumour cells. DNA was extracted by proteinase K digestion and phenol/chloroform extraction. The microsatellite markers D5S346 and D5S429 were used to determine LOH at the APC locus (5q21) using previously reported primers (Esteller et al., 2000). Fluorescent-labelled PCR products were analysed with an automated sequencing System (ABI Prism 310, Applied Biosystem) using Genescan software (Applied Biosystems). The methylation status of the APC gene promoter 1A was determined on bisulphite-treated DNA by methylation-specific PCR using the previously described primers and conditions (Esteller et al., 2000). To study the presence of APC mutations, a portion of APC exon 15 (codons 1256–1551), which includes the mutation cluster region, was amplified by PCR using the following primer sets: 5′-AAGAAACAATACAGACTTATTGT-3′, 5′-ATGAGTGGGGTCTCCTGAAC-3′ (codons 1256–1384); 5′-ATCTCCCTCCAAAAGTGGTGC-3′, 5′-TATCAGCATCTGGAAGAACCT-3′ (codons 1360–1488); and 5′-AGTAAATGCTGCAGTTCAGAGG-3′, 5′-TTCTGCCTC TTTCTCTTGGTT-3′ (codons 1472–1551). In addition, we used 33 sets of primers to screen the following sequences: the phosphorylation sequence for GSK-3β in β-catenin (Palacios and Gamallo, 1998), and plakloglobin genes (Caca et al., 1999), the complete coding sequence of AXIN1 gene (Satoh et al., 2000), the four mononucleotide repeat sequences located in exon 7 of AXIN2 gene (Liu et al., 2000), and exons 12, 13 and 61 of K-, H- and N-RAS (García-Rostán et al., 2001). All PCR amplification products were denatured and subjected to SSCP-analysis by electrophoresis using 30–40% mutation detection enhancement (MDE) gel under different temperature and voltage conditions. The BigDye Sequencing Kit (Applied Biosystems, Foster City, CA, USA) was used to determine DNA sequences in cases exhibiting mobility shifts. Product reactions were run in an Applied Biosystems 3700 Genetic Analyzer (Applied Biosystems). Sequencing was performed in both directions with the primer used for PCR. All mutated cases were verified by repeated PCR-SSCP.

To determine MSI we analysed two mononucleotide repeats, BAT-26 and BAT-25. Primers, PCR amplification conditions and PCR product analysis were as previously described (Moreno-Bueno et al., 2001). We considered a phenotype to be RER+ when the tumours had deletions of more than 2 bp in BAT 26 and BAT 25, when compared with the corresponding alleles in the normal tissue. Our experience with 33 tumours included in this series, as well as in ovarian cancer (Gras et al., 2001) using the complete Bethesda panel (Boland et al., 1998), and that of other workers (Perucho, 1999; Zhou et al., 1998; Samowitz et al., 2001), indicates that the reliability of both mononucleotide repeats is so high that MSI status can be predicted in most cases by evaluating exclusively BAT-25 and BAT-26.

Laser microdissection

Small groups of neoplastic cells were procured by laser microdissection after immunostaining, following the manufacturer's recommendations (UV LASER MICRODISSECTION, SL Microtest, Germany), in nine carcinomas with nuclear β-catenin mutations and in five carcinomas with nuclear focal β-catenin expression without β-catenin mutations. Highly separated cell groups with and without nuclear β-catenin expression were selected for microdissection and DNA analysis in these cases.

Statistical analysis

The Chi-square contingency test with Yates correction, or Fisher's exact test was used to determine the statistical significance of the relationships between the immunohistochemical and clinicopathological, and genetic variables. The SPSS for Windows program (SPSS, Inc., Chicago, IL, USA) was used for this analysis.

References

  1. Aberle H, Bauer A, Stappert J, Kispert A, Kemler R . 1997 EMBO J. 16: 3797–3804

  2. Bailey T, Biddlestone L, Shepherd N, Barr H, Warner PH, Jankowski J . 1998 Am. J. Pathol. 152: 135–144

  3. Behrens J, Jerchow BA, Wurtele M, Grimm J, Asbrand C, Wirtz R, Kuhl M, Wedlich D, Birchmeier W . 1998 Science 280: 596–599

  4. Boland CR, Thibodeau SN, Hamilton SR, Sidransky DA, Eshalman JR, Burt RW, Meltzer SJ, Rodríguez-Bigas MA, Fodde R, Ranzani GN, Srivastava SA . 1998 Cancer Res. 58: 5248–5257

  5. Caca K, Kolling FT, Ji X, Hayes M, Qian J, Yahanda A, Rimm DL, Costa J, Fearon E . 1999 Cell Growth Diff. 10: 369–376

  6. Chesire DR, Ewing CM, Gage WR, Isaacs WB . 2002 Oncogene 21: 2679–2694

  7. Dahmen RP, Koch A, Denkhaus D, Tonn JC, Sörensen N, Berthold F, Behrens J, Birchmeier W, Wiestler OD, Pietsh T . 2001 Cancer Res. 61: 7039–7043

  8. Desbois-Mouthon C, Cadoret A, Blivet-Van Eggelpoël MJ, Bertrand F, Cherqui G, Perret C, Capeau J . 2001 Oncogene 20: 252–259

  9. Ding VW, Rui-Hong C, McCormick F . 2000 J. Biol. Chem. 275: 32475–32481

  10. Dong SM, Kim H-S, Rha S-H, Sidransky D . 2001 Clin. Cancer Res. 7: 1982–1986

  11. Duval A, Iacopetta B, Ranzani GN, Lothe RA, Thomas G, Hamelin R . 1999 Oncogene 18: 6806–6809

  12. Espada J, Pérez-Moreno M, Braga VMM, Rodríguez-Viciana P, Cano A . 1999 J. Cell Biol. 146: 967–980

  13. Esteller M, Levine R, Baylin SB, Hedrick Elleson L, Herman JG . 1998 Oncogene 17: 2413–2417

  14. Esteller M, Spark A, Toyota M, Sanchez-Cespedes M, Capella G, Peinado MA, Gonzalez S, Tarafa G, Sidransky D, Meltzer J, Baylin SB, Herman JG . 2000 Cancer Res. 60: 4366–4371

  15. Fearnhead NS, Britton MP, Bodmer WF . 2001 Hum. Mol. Genet. 10: 721–733

  16. Fujino T, Risinger JI, Collins NK, Liu F, Nishii H, Takahasi H, Westphal E-M, Barrett JC, Sasaki H, Kohler F, Berchuck A, Boyd J . 1994 Cancer Res. 54: 4294–4298

  17. Fukuchi T, Sakamoto M, Tsuda H, Maruyama K, Nozawa S, Hiroashi S . 1998 Cancer Res. 58: 3526–3528

  18. Gamallo C, Palacios J, Moreno-Bueno G, Calvo de Mora J, Suárez A, Armas A . 1999 Am. J. Pathol. 155: 527–536

  19. García-Rostán G, Camp RL, Herrero A, Carcangiu ML, Rimm DL, Tallini G . 2001 Am. J. Pathol. 158: 987–996

  20. Gras E, Catasus L, Argüelles R, Moreno-Bueno G, Palacios J, Gamallo C, Matías-Guiu X, Prat J . 2001 Cancer 92: 2829–2836

  21. He T-C, Spark AB, Rago C, Hemerking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B, Kinzler KW . 1998 Science 281: 1509–1512

  22. Hsu W, Zeng L, Constantini F . 1999 J. Biol. Chem. 294: 3439–3445

  23. Ikeda T, Yoshinaga K, Semba S, Kondo E, Ohmori H, Horii A . 2000 Oncol. Rep. 7: 323–326

  24. Jin Z, Tamura G, Tsuchiya T, Sakata K, Kashiwaba M, Osakabe M, Motoyama T . 2001 Br. J. Cancer 85: 69–73

  25. Jones MH, Koi S, Fujimoto I, Hasumi K, Kato K, Nakamura Y . 1994 Genes Chromosomes Cancer 9: 119–123

  26. Kawakami K, Branbender J, Lord RV, Groshen S, Greenwald BD, Krasna MJ, Yin J, Fleisher S, Abraham JM, Beer D, Sidransky D, Huss HT, Demeester DR, Eads C, Laird PW, Ilson DH, Kelsen DP, Harpole D, Moore MB, Danenberg KD, Danenberg PV, Meltzer SJ . 2000 J. Natl. Cancer Inst. 92: 1805–1811

  27. Kirchner T, Brabletz T . 2000 Am. J. Pathol. 157: 1113–1121

  28. Kobayashi K, Sagae S, Nishioka Y, Tokino T, Kudo R . 1999 Jpn. J. Cancer Res. 90: 55–59

  29. Kolligs FT, Kolligs B, Hajra KM, Hu G, Tani M, Cho KR, Fearon ER . 2000 Gen. Dev. 14: 1319–1331

  30. Korinek V, Barker N, Morin PJ, Van Wichen D, de Weger R, Kinzler KW, Vogelstein B, Clevers H . 1997 Science 275: 1784–1787

  31. Lin YM, Kato T, Satoh S, Nakamura Y, Furukawa Y . 2000 J. Hum. Genet. 45: 254–256

  32. Liu W, Dong X, Mai M, Seelan RS, Taniguchi K, Krishnadath KK, Halling KC, Cunningham JM, Quian C, Christensen E, Roche PC, Smith DI, Thibodeau SN . 2000 Nat. Genet. 26: 146–147

  33. Matias-Gviu X, Catasus L, Bussaglia E, Lagarda H, Garcia A, Pons C, Munoz J, Arguelles R, Machin P, Prat J . 2001 Human Pathol. 32: 569–577

  34. Mirabelli-Primdahl L, Gryfe R, Kim H, Millar A, Luceri D, Dale D, Holowaty E, Bapap B, Gallinger S, Redston M . 1999 Cancer 59: 3346–3351

  35. Moreno-Bueno G, Gamallo C, Perez-Gallego L, de Mora JC, Suarez A, Palacios J . 2001 Diag. Mol. Pathol. 10: 116–122

  36. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B, Kinzler KW . 1997 Science 275: 1787–1790

  37. Nei H, Saito T, Yamasaki H, Mizumoto H, Ito E, Kudo R . 1999 Mol. Carcinog. 25: 207–218

  38. Palacios J, Gamallo C . 1998 Cancer Res. 58: 1344–1347

  39. Palacios J, Catasús L, Moreno-Bueno G, Matías-Guiu X, Prat J, Gamallo C . 2001 Virchows Arch. 438: 464–469

  40. Perucho M . Correspondence re: CR Boland et al 1999 Cancer Res. 59: 249–256

  41. Reinacher-Schick A, Gumbiner BM . 2001 J. Cell Biol. 152: 491–502

  42. Rubinfeld BI, Albert I, Porfiri E, Fiol C, Munemitsu S, Polakis P . 1996 Science 272: 1023–1026

  43. Saegusa M, Okayasu I . 2001 J. Pathol. 194: 59–67

  44. Saegusa M, Hashimura M, Yoshida T, Okayasu I . 2001 Br. J. Cancer 84: 209–217

  45. Samowitz WS, Holden JA, Curtin K, Edwards SL, Walker AR, Lin HA, Robertson MA, Nichols MF, Gruenthal KM, Lynch BJ, Leppert MF, Slattery M . 2001 Am. J. Pathol. 158: 1517–1524

  46. Satoh S, Daigo Y, Furukawa Y, Kato T, Miwa N, Nishikawi T, Kawasoe T, Ishiguro H, Fujita M, Tokino T, Sasaki Y, Imaoka S, Murata M, Shimano T, Yamoka Y, Nakamura Y . 2000 Nat. Genet. 24: 245–250

  47. Satyamoorthy K, Li G, Vaidya B, Patel D, Herlyn M . 2001 Cancer Res. 61: 7318–7324

  48. Schosshauer PW, Pirog EC, Levine RL, Elleson LH . 2000 Mod. Pathol. 13: 1066–1071

  49. Sharpe C, Lawrence N, Martinez Arias A . 2001 BioEssays 23: 311–318

  50. Spark AB, Morin PJ, Vogelstein B, Kinzler KW . 1998 Cancer Res. 58: 1130–1134

  51. Staal FJ, Burgering BM, van de Wetering M, Clevers HC . 1999 Int. Immunol. 11: 317–323

  52. Tetsu O, McCormick F . 1999 Nature 398: 422–426

  53. Tsuchiya T, Tamura G, Sato K, Sakata K, Jin Z, Motoyama T, Usuba O, Kimura W, Nishizuka S, Wilson KT, James SP, Yin J, Fleisher AS, Zou T, Silverberg SG, Kong D, Meltzer SJ . 2000 Oncogene 19: 3642–3646

  54. Ueda M, Gemmill RM, West J, Winn R, Sugita M, Tanaka N, Ueki M, Drabkin HA . 2001 Br. J. Cancer 85: 64–68

  55. Valizadeh A, Karayiannakis AJ, El-Hariry I, Kmiot W, Pignatelly M . 1997 Am. J. Pathol. 150: 1977–1984

  56. Virmani AK, Rathi A, Sathyanarayama UG, Padar A, Huang CX, Cunningham HT, Farinas AJ, Milchgrub S, Euhus DM, Gilcrease M, Herman J, Minna JD, Gazdar AF . 2001 Clin. Cancer Res. 7: 1998–2004

  57. Voeller HJ, Truica CI, Gelman EP . 1998 Cancer Res. 58: 2520–2523

  58. Webster M-T, Rozycka M, Sara E, Davis E, Smalley M, Young N, Dale TC, Wooster R . 2000 Genes Chromosomes Cancer 28: 443–453

  59. Willert K, Nusse R . 1998 Curr. Opin. Gen Development 8: 95–102

  60. Wu R, Zhai Y, Fearon ER, Cho KR . 2001 Cancer Res. 61: 8247–8255

  61. Yuan H, Mao J, Li L, Wu D . 1999 J. Biol. Chem. 274: 30419–30423

  62. Zhou XP, Hoang JM, Li YJ, Seruca R, Carneiro F, Sobrinho-Simoes M, Lothe RA, Gleeson CM, Russel SEH, Muzeau F, Fléjou JF, Hoang-Zuang K, Lidereau R, Thomas G, Hamelin R . 1998 Genes Chromosomes Cancer 21: 101–107

  63. Zysman M, Saka A, Millar A, Knight J, Chapman W, Bapat B . 2002 Cancer Res. 62: 3663–3666

Download references

Acknowledgements

Supported in part by grants FIS 01/0037-01 and SAF2001-0065. G Moreno-Bueno is a recipient of a postdoctoral research grant from the Centro Nacional de Investigaciones Oncológicas, Spain, and D Sarrió is a recipient of a BEFI grant from the Fondo de Investigaciones Sanitatias (01/9132), Spain.

Author information

Correspondence to José Palacios.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • endometrial cancer
  • APC promoter hypermethylation
  • β catenin mutations
  • γ-catenin
  • AXIN1
  • AXIN2

Further reading