Cyclin D1 gene (CCND1) mutations in endometrial cancer

Oncogene (2003) 22, 6115–6118

& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

Cyclin D1 gene (CCND1) mutations in endometrial cancer

1 Laboratory of Breast and Gynaecological Cancer, Molecular Pathology Programme, Centro Nacional de Investigaciones Oncolo´gicas (CNIO), Melchor Fernańdez Almagro 3, 28029 Madrid, Spain; 2Cytogenetics Unit, Biotechnology Programme, Centro Nacional de Investigaciones Oncolo´gicas (CNIO), Melchor Fernańdez Almagro 3, 28029 Madrid, Spain; 3Department of Pathology, Hospital Universitario La Paz, Paseo de la Castellana 246, 28029 Madrid, Spain; 4Department of Pathology, Hospital de la Santa Creu i Sant Pau, Sant Antoni Ma Claret 167, 08025 Barcelona, Spain; 5Department of Pathology and Molecular Genetics, Hospital Universitari Arnau de Vilanova, Universitat de Lleida, Av Rovira Roure 80, 25198 Lleida, Spain

Cyclin D1 is frequently overexpressed in human neoplasias by gene rearrangement and amplification, but no mutations in the CCND1 gene have so far been reported. However, in vitro mutagenesis of CCND1 has shown that substitutions affecting threonine 286 residue produced cyclin D1 nuclear accumulation, by interfering with protein degradation and induced neoplastic transformation in murine fibroblasts. To test whether similar genetic changes may occur in vivo, we analysed a series of 60 endometrioid endometrial carcinomas (EECs) for cyclin D1 expression and gene amplification by immunohistochemistry and FISH, respectively. Two of 17 carcinomas showing cyclin D1 expression in more than 5% of neoplastic cells, but without gene amplification, were found to harbor single-base substitutions in CCND1 that changed proline 287 into threonine and serine, respectively. Both cases expressed cyclin D1 in more than 50% of neoplastic cells. Additionally, seven tumors with cyclin D1 overexpression of an independent series of 59 EECs were also analysed, and a 12-bp in-frame deletion that eliminated amino acids 289–292 was detected in one case with cylin D1 expression in more than 50% of neoplastic cells. In contrast, no mutations of the CCND1 gene were detected in a set of breast carcinomas with cyclin D1 overexpression without gene amplification. In summary, our data indicate that mutations of CCND1, which probably render the protein insensitive to degradation, represent a previously unreported mechanism of cyclin D1 overexpression in human tumors in vivo. Oncogene (2003) 22, 6115–6118. doi:10.1038/sj.onc.1206868 Keywords: endometrial carcinoma; cyclin D1 expression; cyclin D1 mutation

Cyclin D1, encoded by the CCND1 gene located on 11q13, plays an important role in the progression of the cell cycle. It associates with cyclin-dependent kinases (CDKs) CDK4 and CDK6 to phosphorylate the *Correspondence: J Palacios; E-mail: [email protected] Received 5 December 2002; revised 9 June 2003; accepted 13 June 2003

retinoblastoma protein (Rb) during the G1 phase (Sherr and Roberts, 1999). The control of cyclin D1 degradation by ubiquitination in the 26S proteasome is important for maintaining appropriate cyclin D1 levels during the cell cycle. Phosphorylation of cyclin D1 at threonine 286 is required for its ubiquitination, nuclear export and degradation in the cytoplasm (Sherr, 1996; Diehl et al., 1996, 1998; Alt et al., 2000; Germain et al., 2000). This phosphorylation is mediated by glycogen synthase kinase 3-b (GSK3-b) and is greatly enhanced by the binding of cyclin D1 to CDK4 (Diehl et al., 1997, 1998). Cyclin D1 acts as an oncogene in different human neoplasias when overexpressed (Weinstat-Saslow et al., 1995; Sherr et al., 1997; Steeg and Zhou, 1998). To date, no CCND1 mutations have been reported, but cyclin D1 overexpression may be the result of CCND1 rearrangement or amplification. For example, mantle cell lymphoma is characterized by 11q13 translocation and CCND1 rearrangement (Campo et al., 1999). CCND1 amplifications have been reported in several human neoplasias, such as in situ and infiltrating ductal breast carcinoma, and in bladder, head and neck, lung and prostate cancer (Simpson et al., 1997; Schraml et al., 1999; Hoechtlen-Vollmar et al., 2000; Kaltz-Wittmer et al., 2000). In addition, cyclin D1 overexpression without gene amplification occurs, for instance, in most colorectal and 15% of breast carcinomas overexpressing cyclin D1 (Russell et al., 1999). Defective degradation of cyclin D1 has been proposed as a mechanism of protein overexpression in some of these cases, although specific alterations in the degradation machinery have not been found in vivo. However, in vitro studies have shown that threonine 286 mutations render the protein insensitive to degradation, and that this cyclin D1 mutant induces transformation in murine fibroblasts (Alt et al., 2000). Endometrial carcinoma (EC) is the most common invasive malignancy of the female genital tract in most Western countries. Cyclin D1 overexpression has been reported in ECs (Nikaido et al., 1996; Schmitz et al., 2000; Soslow et al., 2000; Tsuda et al., 2000; Cao et al., 2002; Quddus et al., 2002; Machin et al., 2002), but the underlying molecular abnormalities remain to be

ONCOGENOMICS

Gema Moreno-Bueno1, Sandra Rodrı´ guez-Perales2, Carolina Sańchez-Este´vez1, David Hardisson3,4, David Sarrio´1, Jaime Prat5, Juan C Cigudosa2, Xavier Matias-Guiu5 and Jose´ Palacios*,1

Cyclin D1 gene mutation in endometrial cancer G Moreno-Bueno et al

6116

identified. Taking into account the in vitro evidence that substitution at threonine 286 is oncogenic, we have analysed the 30 -end coding sequence of the CCND1 gene in order to establish whether alteration of this sequence occurs in ECs, a tumor in which CCND1 amplification seems to be absent or infrequent (Schraml et al., 1999). In addition, a series of breast carcinomas, a tumor in which CCND1 amplification is common, was also evaluated for this possible alteration. To carry out this study, we used 60 endometrioid ECS. All tumors were diagnosed in the Pathology Department of La Paz Hospital, Madrid, Spain. Some clinical and molecular features of this series have been previously reported (Moreno-Bueno et al., 2002). Mean age at diagnosis was 62.47712.04 years (range 31–89). The series included 29 grade 1, 20 grade 2 and 11 grade 3 tumors. Tumor stage was recorded in 59 tumors: 45 were in stage I, seven in stage II and seven in stage III/ IV at diagnosis. The tumors were fixed in 10% formalin and embedded in paraffin. All immunohistochemical, FISH and molecular analyses were carried out on paraffin-embedded samples. Cyclin D1 expression was evaluated by immunohistochemistry using the mouse anti-human cyclin D1 antibody (DCS-6 clone, DAKO, Glostrup, Denmark), which was applied at 1 : 100 dilution, after a heatinduced antigen retrieval step. Immunodetection was performed using biotinylated anti-mouse immunoglobulins and peroxidase-labeled streptavidin (LSABDAKO) with diaminobenzidine as the chromogen. In negative controls, the primary antibody was omitted or replaced with an irrelevant antibody. Nuclear cyclin D1 expression was graded from 1 to 3 (1 ¼ fewer than 5%, 2 ¼ 5–25%, 3 ¼ more than 25% of stained tumor nuclei). Cyclin D1 expression in 0–5%, 6–25% and more than 25% of neoplastic cells was observed in 43 (71.7%), 8 (13.3%) and nine (15%) tumors, respectively. There was a statistically significant relationship between cyclin D1 expression and tumor grade (P ¼ 0.03), whereby cyclin D1 overexpression (cyclin D1 expression in more than 25% of neoplastic cells) was observed in 36.3% of grade 3 carcinomas, but only in 13 and 5% of grade 1 and 2 carcinomas, respectively. No statistical association was observed between cyclin D1 expression and tumor stage. We analysed CCND1 gene amplification in 30 EECs: the 17 cases with cyclin D1 expression in more than 5% of cells and 13 additional cases with cyclin D1 expression in fewer than 5% of cells. FISH analysis was performed using Spectrum Orange-labelled CCND1 probe, with the FITC-labelled centromeric chromosome 11 probe as an internal control (Vysis, Downer’s Grove, IL, USA). Amplification was defined as the presence (in X5% of tumor cells) of either >10 gene signals or more than three times as many gene signals than centromere signals of the chromosome (Schraml et al., 1999). Only one EEC, which expressed cyclin D1 in more than 25% of cell nuclei, showed gene amplification. In addition, another carcinoma with 6–25% cyclin D1-stained cells showed four CCND1 gene copies with only two signals of the centromeric probe. Although this is the first report of CCND1 amplification in endometrial cancer, Oncogene

our results indicate a low frequency of this alteration in the endometrioid histotype. To study CCDN1 mutations, we designed two primers to amplify the sequence comprising the region between nucleotides 990 and 1133 of the CCND1 gene (GenBank Nucleotide Sequence Database Accession number NM_053056): 50 CAGGCCCAGCAGAACAT 30 and 50 CTGCGGGTGGCGGTG 30 in forward and reverse directions, respectively. The amplicon includes threonine 286, the amino-acid target for phosphorylation by GSK3-b, which is required for cyclin D1 ubiquitination (Diehl et al., 1997; Germain et al., 2000). Formalin-fixed paraffin-embedded blocks from 50 endometrial carcinomas (including all cases expressing cyclin D1 in more than 5% of tumor cells) and corresponding normal tissue (myometrium) were cut at 10 mm and reviewed microscopically. When necessary, tumor tissue was manually microdissected from areas with 475% of tumor cells. DNA was extracted from these samples and PCR amplification was performed with 300 nm of each oligonucleotide, 200 mm deoxinucleotide triphosphate, 0.25 U of Taq polymerase (Biotools B&M Labs, Madrid, Spain) and 50–100 ng DNA in a 20 ml volume. The PCR conditions were 951C for 5 min followed by 35 cycles (denaturation: 951C for 30 s, annealing: 651C for 15 s and extension: 721C for 30 s), and a final elongation cycle at 701C for 10 min. Amplified PCR products were denatured and subjected to single-strand conformational polymorphism (SSCP) analysis using 30–40% MDE. Cases exhibiting mobility shifts were submitted to direct sequencing (tumors T47 and T76). In addittion, in tumor T38, the abnormal band was excised from SSCP gel and, after reamplification and purification, was sequenced in both strands. All mutated cases were verified by repeating three times the above-described procedure. We found that two EECs with cyclin D1 expression in more than 50% of the cells had single-base substitutions affecting codon 287, one CCC to ACC and one CCC to TCC change leading to a change from proline to threonine and serine, respectively (Figure 1). Since the percentage of mutation was low, all cases were subjected to heteroduplex analysis by conformation-sensitive gel electrophoresis (CSGE), in order to confirm these results. Amplified samples were held at 951C for 5 min and at 651C for 1 h to generate heteroduplexes. The products were diluted 1 : 2 in sucrose buffer and loaded in a partially denaturing MDE gel at constant 7 W power. Gels were silver stained and dried on a vacuum gel dryer. Only tumors T47 and T76 showed heteroduplexes formation. In order to assess the sensitivity of these techniques to detect CCND1 mutations, the DNA from the two mutated cases, extracted from manually microdissected tumor cells (purity about 90%), was diluted 1 : 5, 1 : 10, 1 : 15 and 1 : 20 with DNA from normal tissue. We detected abnormal shifts by SSCP in samples diluted 1 : 5, and heteroduplex formation by CSGE in 1 : 10 dilutions, indicating detection of at least 5% of mutated alleles in a mixed sample by this last technique.


6117

Figure 1 CCND1 mutations in three EECs. (a) Tumor number 47 (T47) shows cyclin D1 expression in all tumor cells. (b) Conformation-sensitive gel electrophoresis analysis of endometrial carcinomas. In T47 heteroduplex formation was detected, indicating the presence of a mutation in the tumor DNA. The mutation was still detected after 1 : 5 and 1 : 10 dilution of the tumor DNA with normal tissue DNA. Tumors T21, T18, T54 and T62 show a wild-type pattern. (c) T47 shows a single-base substitution in codon 287 changing proline to threonine, compared with the wild-type sequence (WT). Single-base substitution in tumur 76 (T76) changes proline to serine. In-frame deletion of 12 nucleotides eliminates amino acids 289–293 (T38)

To determine whether similar CCND1 mutations occurred in one independent series of EECs, we analysed seven tumors with cyclin D1 expression in more than 25% of the cells of a previously reported series of 59 EECs (Machin et al., 2002). One tumor with more than 50% of cyclin D1-expressing cells had a 12-bp in-frame deletion that eliminates amino acids 289–292 (Figure 1). The mutations described here seem to be important for different reasons. First, they were tumor-specific, given that they were found only in the tumor samples but not in the corresponding normal tissues in independent DNA extractions and PCR amplifications. Second, they were found in samples with extensive nuclear cyclin D1 overexpression in which other molecular alterations responsible for cyclin D1 overexpression were excluded. Thus, none of the cases had CCND1 gene amplification or b-catenin mutations (Machin et al., 2002; MorenoBueno et al., 2002), and only one case showed p53 expression in about 15% of neoplastic cells. Third, these in vivo genetic alterations fit well with previous experi-

mental studies that showed the importance of threonine 286 in the degradation of cyclin D1. Diehl et al. (1997) demonstrated that cyclin D1 turnover is governed by ubiquitination and proteasomal degradation, which are positively regulated by cyclin D1 phosphorylation of threonine 286. Although CDK4bound cyclin D1 molecules are intrinsically unstable (half-life o 30 min), a cyclin D1 mutant (T286A) containing an alanine for threonine 286 substitution failed to be efficiently polyubiquitinated, and was markedly stabilized (half-life approximately 3.5 h) when inducibly expressed (Diehl et al., 1997). These authors also demonstrated that GSK-3b phosphorylates cyclin D1 specifically at threonine 286, thereby triggering rapid cyclin D1 turnover, while the cyclin D1 mutant T286A was refractory to phosphorylation by GSK-3 and remained in the nucleus throughout the cell cycle (Diehl et al., 1998). Recently, it has been observed that GSK3b-dependent phosphorylation promotes cyclin D1 nuclear export by facilitating the association of cyclin Oncogene


6118

D1 with the nuclear-exporting CRM1 (Alt et al., 2000). Cyclin D1 mutant T286A remains in the nucleus throughout the cell cycle due to its reduced binding to CRM1. In addition, constitutive overexpression of the nuclear cyclin D1-T286A in murine fibroblasts results in cellular transformation and promotes tumor growth in immune-compromised mice (Alt et al., 2000). In contrast, overexpression of wild-type cyclin D1 is not sufficient to promote neoplastic growth without cooperating additional mutations. Mutations in proline 287, as observed in present study, would have the same effects as those described in vitro for threonine 286 mutations since GSK-3b is a proline-directed enzyme that requires proline 287 for its proper function (Diehl et al., 1997). The absence of three residues close to threonine 286 and proline 287, as observed in an additional case, probably caused important conformational changes that would impact on enzyme function. The situation described here for cyclin D1 is similar to that observed in b-catenin, which bears mutations in amino-acid targets for GSK-3b phosphorylation or adjacent residues. These mutations produce accumulation of b-catenin by impaired ubiquitination and subsequent degradation (Morin et al., 1997). To evaluate whether or not similar CCND1 mutations also occur in other tumor types, we analysed a group of breast cancer. Cyclin D1 overexpression and CCND1 amplification have been reported in about 30 and 20% of breast cancer, respectively. We constructed a tissue microarray (TMA) containing 190 cores from 95 breast carcinomas. Cyclin D1 expression and CCND1 ampli-

fication were assessed by immunohistochemistry and FISH on TMA sections. TMA FISH analysis yielded valuable information in 80% of cases (77 out of 96 tumors), and CCND1 amplification was observed in 15 (19.5%) tumors. These values are similar to those from previous studies in breast cancer using TMA (Schraml et al., 1999). CNND1 mutations were screened in 30 tumors included in the TMA that had more than 5% of cyclin D1-expressing cells (including 15 cases with more than 25% of positive cells) without gene amplification. However, we did not find any mutations in these tumors. These results could suggest that CCND1 mutations may be less frequent in those tumor types in which gene amplification is a common source of protein overexpression. In summary, this study provides the first evidence of a previously unreported mechanism of cyclin D1 overexpression in human tumors in vivo: CCND1 mutations affecting critical residues involved in protein degradation. Further studies are needed to assess the frequency of this alteration in other tumor types, mainly in those human neoplasias in which genetic alterations of the CCND1 gene are infrequent. Acknowledgements We are grateful to Rau´l Cassia and Alicia Barroso for excellent technical assistance. The study is supported in part by Grants FIS PI020355, FIS 01/1656 and SAF2001-0065. GM-B is the recipient of a postdoctoral research grant for de CNIO Spain. SR-P is a recipient of a research grant from the CNIO, Spain. DS is a recipient of a research grant from FIS (BEFI, 1/9132).

References Alt JR, Cleveland JL, Hannink M and Diehl JA. (2000). Genes Dev., 24, 3102–3114. Campo E, Raffeld M and Jaffe ES. (1999). Semin. Hemathol., 36, 115–127. Cao QJ, Einstein MH, Anderson PS, Runowicz CD, Balan R and Jones JG. (2002). Int. J. Gynecol. Pathol., 21, 147– 154. Diehl JA, Cheng M, Roussel MF and Sherr CJ. (1998). Genes Dev., 12, 3499–3511. Diehl JA, Zindy F and Sherr CJ. (1997). Genes Dev., 11, 957– 972. Germain D, Russell A, Thompsonm A and Hendley J. (2000). J. Biol. Chem., 275, 12074–12079. Hoechtlen-Vollmar W, Menzelm G, Bartl R, Lamerz R, Wick M and Seidel D. (2000). Br. J. Haematol., 109, 30– 38. Kaltz-Wittmer C, Klenk U, Glaessgen A, Aust DE, Diebold J, Lohrs U and Baretton GB. (2000). Lab. Invest., 80, 1455– 1464. Machin P, Catasus L, Pons C, Munozn J, Matias-Guiu X and Prat J. (2002). Hum. Pathol., 33, 206–212. Moreno-Bueno G, Hardisson D, Sańchez C, Sarrio´ D, Cassia R, Garcı´ a-Rostań G and Palacios J. (2002). Oncogene, 21, 7981–7991. Morin PJ, Sparks AB, Korinek V, Barker N, Clevers H, Vogelstein B and Kinzler KW. (1997). Science, 275, 1787– 1790.

Oncogene

Nikaido T, Li SF, Shiozawa T and Fujii S. (1996). Cancer, 78, 1248–1253. Quddus M, Latkovich P, Castellani WJ, James Sung C Steinhoff MM, Briggs RC and Miranda RN. (2002). Arch. Pathol. Lab. Med., 126, 459–463. Russell A, Thompson MA, Hendley J, Trute L, Armes J and Germain D. (1999). Oncogene, 18, 1983–1991. Schmitz MJ, Hendricks DT, Farley J, Taylor RR, Geradts J, Rose GS and Birrer MJ. (2000). Gynecol. Oncol., 77, 439– 445. Schraml P, Kononenm J, Bubendorf L, Moch H, Bissig H, Nocito A, Mihatsch MJ, Kallioniemi OP and Sauter G. (1999). Clin. Cancer Res., 5, 1966–1975. Sherr CJ. (1996). Science, 274, 1672–1677 (review). Sherr CJ and Roberts JM. (1999). Genes Dev., 13, 501–512 (review). Simpson JF, Quan DE, O’Malley F, Odom-Maryon T and Clarke PE. (1997). Am. J. Pathol., 151, 161–168. Soslow RA, Shen PU, Chung MH, Isacson C and Baergen RN. (2000). Int. J. Gynecol. Pathol., 19, 329–334. Steeg PS and Zhou Q. (1998). Breast Cancer Res. Treat., 52, 17–28 (review). Tsuda H, Yamamoto K, Inoue T, Uchiyama I and Umesaki N. (2000). Br. J. Cancer, 82, 675–682. Weinstat-Saslow D, Merino MJ, Manrow RE, Lawrence JA, Bluth RF, Wittenbel KD, Simpson JF, Page DL and Steeg PS. (1995). Nat. Med., 1, 1257–1260.