Lack of association between LOH in the 9p region and clinicopathologic parameters in primary breast cancer

Cancer Genetics and Cytogenetics 200 (2010) 23e27

Lack of association between LOH in the 9p region and clinicopathologic parameters in primary breast cancer Sarah Franco Vieira de Oliveiraa, Ma´rcia Maria Costa Oliveiraa, Cı´cero Andrade Urbanb, Rubens Silveira de Limab, Iglenir Joa˜o Cavallia, Enilze Maria de Souza Fonseca Ribeiroa,* a Department of Genetics, Federal University of Parana´, Centro Polite´cnico, Jardim das Ame´ricas, Curitiba, PR, 81531-970, Brazil Department of Oncology and Breast Surgery, Hospital Nossa Senhora das Grac¸as, 433 Alcides Munhoz st, Curitiba, PR, 80810-120, Brazil

b

Received 24 September 2009; received in revised form 2 March 2010; accepted 3 March 2010

Abstract

Previous studies have suggested the involvement of the 9p region in the genesis and progression of several types of cancer. To perform a more in-depth investigation of the 9p region in samples from breast carcinomas, we analyzed loss of heterozygosity (LOH) in 230 patients with primary breast cancer using five microsatellite markers spanning a genomic region of approximately 16.2 megabases. Genomic DNA was obtained from frozen tumor tissue, and peripheral blood was used as a normal reference. Among all samples, 171 (74%) were informative for at least 1 marker and 44 (25.73%) showed LOH. The LOH rates detected for all markers ranged from 10.29% (D9S169) to 15.97% (D9S1749). Among the informative cases for intragenic markers D9S1748 (CDKN2A) and D9S1749 (MTAP), we noticed a concordant loss of 90% (9/10). Associations between LOH frequencies and clinicopathologic parameters were found between marker D9S200 and tumor grade (P ! 0.05), and between marker D9S1748 and estrogen receptor (ER) status (P ! 0.05). In conclusion, our results agree with other data from the literature that point to LOH as a secondary mechanism of tumor suppressor inactivation on 9p in breast cancer, showing lower frequencies than those observed in other types of cancer. On the other hand, our results point to an interesting association between the concordant loss of genes CDKN2A and MTAP, which was not sufficiently explored in primary breast cancer. Ó 2010 Elsevier Inc. All rights reserved.

1. Introduction Breast cancer is the second most common malignancy among women and the second leading cause of cancer death in the United States [1]. In Brazil, it is the leading cause of cancer death among women, with an estimated risk of 49 cases for every 100,000 inhabitants [2]. A positive family history has been reported by 15e20% of women with breast cancer, despite the finding that less than 7% of all breast cancers are associated with known highpenetrance gene mutations [3]. Among the genetic alterations involved in the development and progression of breast cancer, allelic losses at particular chromosomal regions are common and may indicate a deletion of tumor suppressor genes. Loss of heterozygosity (LOH) is now recognized as an invaluable * Corresponding author. Tel.: þ55-413361555; fax: þ55-4133611793. E-mail address: [email protected] (E. Maria de Souza Fonseca Ribeiro). 0165-4608/$ e see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.cancergencyto.2010.03.002

tool for cancer diagnosis and prognosis, regardless of whether the corresponding target genes have been identified [4]. The short arm of chromosome 9 is involved in inversions, translocations, deletions, and LOH in several types of cancer, including those of the breast [5e9]. The best candidates in this region are closely related genes that encode CDKN2A, CDKN2B (also named p16 and p15, respectively), and MTAP. CDKN2A/2B play a critical role in cell cycle regulation encoding cyclin-dependent kinase inhibitors (CDKIs) and are located within the 9p21~p22 region [10]. CDKN2A also encodes another important protein, p14ARF, codified in an alternative reading frame and involved in apoptosis control [11]. MTAP (methylthioadenosine phosphorylase) is a housekeeping gene that encodes an essential enzyme that plays a major role in polyamine metabolism and is important for the salvage of both adenine and methionine. It is located 100 kilobases (kb) telomeric to CDKN2A; codeletion has been reported in several tumors [11] and might be exploited therapeutically

24

S.F. Vieira de Oliveira et al. / Cancer Genetics and Cytogenetics 200 (2010) 23e27

using de novo purine synthesis antimetabolites to treat a subset of tumors [12]. Somatic cancer genetics in recent years has been attempted to identify tumor-suppressor genes by mapping regions of allelic loss. To investigate LOH at 9p in breast cancer samples, we evaluated five microsatellite markers in a series of 230 sporadic breast cancer cases from Southern Brazil. In this study, we report the frequencies and their associations to clinicopathologic parameters.

2. Materials and methods 2.1. Samples Primary tumor tissues and peripheral blood specimens were obtained from 230 female patients undergoing surgery for sporadic breast cancer between 2005 and 2008 at the Hospital Nossa Senhora das Grac¸as and Hospital de Clı´nicas (Curitiba, Southern Brazil), with informed consent. The tissue samples were macroscopically dissected and stored (e80 C) until DNA extraction. The patients had no family history of breast cancer. Table 1 presents the clinicopathologic information obtained. 2.2. PCR amplification/LOH analysis DNA from matched normal and tumor tissues were amplified by polymerase chain reaction (PCR) using the following five dinucleotide repeat microsatellite markers: D9S1749, D9S1748, and D9S171 on 9p21.3; D9S169 on 9p21.2; and D9S200 on 9p13.1. The markers were chosen because of their reported high heterozygosity rate and their location near putative tumor suppressor genes that should be tested. The forward primer for each set was labeled with either of two fluorescent dyes, HEX or FAM (Applied Biosystems/PE Biosystems, Foster City, CA). In all cases, the markers were initially evaluated in the DNA from lymphocytes and the informative samples were studied in the corresponding tumor tissue. PCR was performed using 20 ng/mL of genomic DNA, 100 mmol/L Tris/HCl, 50 mmol/L MgCl2, 8 mmol/L forward primer, 8 mmol/L reverse primer, 2 mmol/L dNTPs (each), and 5 unit/ mL of Taq polymerase in a total volume of 10 mL. Reactions were cycled as follows: 96 C for 5 minutes (initial DNA denaturation), followed by 30 cycles of 96 C for 20 seconds, 55 C for 30 seconds, and 72 C for 40 seconds, with a final elongation at 72 C for 20 minutes. Allele sizes were determined by electrophoresis of PCR products in 6% denaturing polyacrylamide gels and compared with ET-ROX 400 size standards (Amersham Biosciences, Amersham, UK) using an automated sequencer (MEGABACE 1000). The fluorescent signals from the alleles with different sizes were recorded and analyzed using the software Fragment Profiler (version 1.2; Amersham Biosciences). After the visual examination

Table 1 Selected clinicopathologic features of study participants Characteristics

Value (%)

Mean age at diagnosis (mean 6 SD) <50 years O50 years Not informed Tumor size mean (mean 6 SD; n5 223) Lymph node involvement (n5216) Positive Negative Grade (n5212) I II III ER status (n5170) Positive Negative PR status (n5120) Positive Negative ERBB2 (n5126) Positive Negative Histology (n5230) IDC ILC IC MC CC TC Other subtypes*

55.84 6 13.25 years old 84 (36.53) 143 (62.17) 3 (1.30) 34.28 6 16.49 mm 123 (56.94) 93 (43.06) 30 (14.15) 127 (59.91) 55 (25.94) 127 (74.71) 43 (25.29) 84 (70.00) 36 (30.00) 41 (32.54) 85 (67.46) 186 13 5 4 2 2 18

(80.87) (5.65) (2.17) (1.74) (0.87) (0.87) (7.83)

Abbreviations: n, number of patients; SD, standard deviation; ER, estrogen receptor; PR, progesterone receptor; IDC, invasive ductal carcinoma; ILC, invasive lobular carcinoma; IC, intraductal carcinoma; MC, medullary carcinoma; CC, colloid carcinoma; TC, tubular carcinoma. * Other subtypes: carcinoma/chronic mastitis, invasive apocrine carcinoma, invasive ductolobular carcinoma, metaplastic carcinoma associate to IC, invasive papillary carcinoma, pseudo-papillary carcinoma, Paget’s disease of the nipple associated with IDC/IC, pleomorphic lobular carcinoma, lobular carcinoma not specified, tubulolobular carcinoma (3), ductal carcinoma not specified (2), lymph node with metastatic carcinoma (2), tubuloductal carcinoma (2).

of computer printouts, LOH was determined mathematically according to the following formula: ðN2=N1Þ=ðT2=T1Þ; where N1 and T1 represent the shorter peak heights in the normal and tumor samples, respectively, and N2 and T2 the higher peak in the normal and tumor samples, respectively. LOH was interpreted as present if the final quotient was at most 0.6 or at least 1.67 [13]. A ratio below 0.6 represents an allele signal reduction of 40%, which is considered indicative of allele loss. This limit was chosen because the tumor cell content after dissection is assessed to be greater than 70% and inter-assay variations of the detection system are below 10%. Homozygous and nonanalyzable peaks were designated as noninformative cases. The fractional allelic loss (FAL) was calculated as the number of LOH events per number of informative loci.

S.F. Vieira de Oliveira et al. / Cancer Genetics and Cytogenetics 200 (2010) 23e27

All cases of LOH and 10% of cases that had retention of heterozygosity (ROH) were repeated. 2.3. Statistical analysis The following statistical tests were used: chi-square (c2), G, Fisher-exact, Student’s t, analysis of variance, and regression coefficient; P ! 0.05 was considered significant. 3. Results Table 2 summarizes the LOH results. The D9S1749 marker showed the highest frequency (15.97%), and the D9S169 marker showed the lowest frequency (10.29%). There was no significant difference in the distribution of LOH frequencies among the five markers. Considering the samples informative for at least one marker (n 5 171), 44 (25.73%) showed LOH. The LOH frequencies among the chromosomal subregions were 20.99% on 9p21.3 (D9S1749, D9S1748, and D9S171); 10.29% on 9p21.2 (D9S169); and 12.90% on 9p13.1 (D9S200). Regarding the histologic subgroups, of 148 informative cases of ductal carcinomas, 40 (27.03%) showed LOH in at least one marker, with a frequency of each marker as follows: 19.32% on D9S171, 15.45% on D9S1748, 15.22% on D9S169, 19.42% on D9S200, and 19.54% on D9S1749. Of the 9 informative cases of lobular carcinoma and the 14 informative cases of carcinomas of other subtypes, 2 (22.22%) and 2 (14.29%) showed LOH in at least one marker, respectively. There was no significant difference between them (c22 5 1.15, P O 0.50). We accessed the concordant loss of the markers D9S1748 and D9S1749 (intragenic to the CDKN2A and MTAP genes, respectively), and among the 10 informative cases for both markers, 9 showed a co-deletion (90%) and one showed ROH at D9S1748 and LOH at D9S1749. The general frequency of the co-deletion was 8.50% (9/106 informative cases). 3.1. Clinicopathologic associations We found no statistical differences among patients with LOH in at least one marker and patients with ROH for clinical parameters such as mean age, time of exposure to Table 2 LOH analysis results

Marker

Patients (n)

Informative cases (%)

D9S1749 D9S1748 D9S171 D9S169 D9S200 Chi-square

215 220 224 208 224 e

55.35 71.36 69.64 65.38 69.20 e

Abbreviations: n, number of patients.

FAL (%) (LOH/informative cases) 15.97 (19/119) 11.46 (18/157) 11.54 (18/156) 10.29 (14/136) 12.90 (20/155) c252.23; PO0.50

25

endogenous estrogen (age at menopause d age at menarche: t 5 1.26; P O 0.20), mean age at first pregnancy: t 5 0.37; P O 0.70), and smoking habit (Fisher’s exact test P O 0.40). Using the regression coefficient test, we found that variation in tumor size was not dependent on the time of exposure to endogenous estrogen (b 5 e0.34, t 5 0.84; P O 0.40). Table 3 shows the mean age at the time of diagnosis and the tumor size of patients with LOH or ROH for each marker. These data were submitted for analysis of variance, and the results were not significant for any of the factors (data not shown). We found significant differences between LOH at D9S200 and the histopathologic grades (G test 5 8.46; P ! 0.05), primarily due to the absence of LOH in grade I tumors (partial result5 2.78). We also found significant differences between LOH at D9S1748 and estrogen receptor (ER) status (c21 5 5.73, P ! 0.05), primarily due to the absence of LOH in this marker in ER-negative tumors (partial result 5 3.52). We found no significant differences between LOH frequencies at each marker and the presence of metastasis in axillary lymph nodes, or the progesterone receptor (PR) or ERBB2 status (data not shown). When the markers were analyzed together, tumors from patients who presented metastasis in the axillary lymph nodes did not show an increased frequency of LOH when compared to those without metastasis (P O 0.99). Tumors with LOH were equally distributed among the histopathologic grade specimens (c22 5 1.11; P O 0.50), and no significant difference was found between LOH and the status of the ER and PR hormonal receptors or of ERBB2 (c21 5 1.96, P O 0.10; c21 5 0.60, P O 0.40; c21 5 0.39, P O 0.50, respectively). Data about the concordant loss of the markers D9S1748 and D9S1749 were insufficient to conduct a firm statistical analysis. 4. Discussion LOH and promoter hypermethylation are the main mechanisms for loss of function of a tumor suppressor gene in sporadic breast carcinomas. Several studies have been conducted using LOH analysis in breast cancer (summarized by Miller et al. [14]), but despite the hundreds of deletion studies, the number and identity of tumor suppressor genes relevant to this disease remain largely unknown. The 9p region has attracted much research interest because of the identification of important suppressor genes and the discovery that the entire region can be homozygously deleted in a larger number of tumor cell lines. In the present study, among 171 informative cases, LOH in at least one marker was found in 44 (25.73%) cases. We found low frequencies of FAL (up to 20%) in each marker with the higher frequencies on 9p21.3 at marker D9S1749 (16%), and on 9p13.1 at marker D9S200 (13%). The observed LOH frequency at marker D9S171 (the most commonly used among the published studies) did

S.F. Vieira de Oliveira et al. / Cancer Genetics and Cytogenetics 200 (2010) 23e27

26

Table 3 Mean age and tumor size (mm) between LOH/ROH for each marker D9S1749

D9S1748

D9S171

D9S169

D9S200

n

Mean 6 SD

n

Mean 6 SD

n

Mean 6 SD

n

Mean 6 SD

n

Mean 6 SD

Age

LOH ROH

19 100

62.74 6 14.05 57.4 6 12.63

19 138

57.42 6 12.24 57.14 6 13.34

18 138

57.89 6 14.09 57.33 6 13.29

14 122

58.07 6 14.77 56.93 6 12.8

20 135

59.65 6 14.97 57.06 6 13.16

Size

LOH ROH

18 79

34.89 6 16.09 33.48 6 15.17

17 105

36.18 6 17.63 32.02 6 16.36

16 84

32.87 6 16.24 33.55 6 17.19

14 91

29.71 6 15.83 33.36 6 16.87

18 94

39.33 6 14.03 33.79 6 17.01

Abbreviations: n, number of patients; SD, standard deviation.

not differ significantly from that described by Schwarzenbach et al. [13], who studied 42 patients with primary and metastatic breast carcinomas (c21 5 0.79, P O 0.30). The same observation was noted in relation to Ellsworth et al. [15], who studied 24 breast tumors with axillaries metastasis (Fisher’s exact test: P O 0.98). In relation to the data from Brenner and Aldaz [16] and Smeds et al. [17], we found a lower frequency of LOH in this marker. In the study by Brenner and Aldaz, LOH at the 9p region was analyzed using 5 polymorphic microsatellite markers in 24 breast tumor samples. LOH was observed in 14 tumors (58%), and in 12 of them at multiple markers. The majority of the losses occurred at D9S169 (58%) and D9S171 (53%). Argos et al. [18] examined the presence of LOH and DNA copy number variation in 16 microdissected invasive breast tumors using microarray analysis, and, as corroborated by our results, they did not observe a relevant frequency of LOH in the 9p region. A homozygous deletion of the 9p21~p22 region is a common event in several types of cancer, and usually it is quite large, eliminating multiple genes. A study by Cairns et al. [19] found that most deletions remove a 170-kb region, which includes the following three transcripts: MTAP, CDKN2A (p16), and p14ARF (p14). Each protein appears to affect different cellular functions: p16 affecting cell cycle, p14 affecting apoptosis, and MTAP affecting anchorage dependence [20]. The cellular selective advantage of eliminating all these genes together may explain the higher frequencies of homozygous deletions in spite of LOH in this region. In this study, we analyzed one marker intragenic to CDKN2A (D9S1748) and one intragenic to MTAP (D9S1749) that showed 12 and 16% of LOH, respectively. Accessing the concordant loss of these two markers, we noticed that among 10 informative cases for both markers, concomitant LOH was present in 9 (90%). Despite the low number of cases, these data suggest that in breast cancer as well, the co-deletion may have an important role, as described in several types of tumors in literature [11,20,21]. In this study, there were no associations between the occurrence of FAL in 9p and clinical parameters. Thus, known risk factors for breast cancer, such as longer exposure to endogenous estrogen, age at first pregnancy, and exposure to the carcinogenic effects of tobacco are not related to increasing genetic instability at 9p. We were

unable to make comparisons between these results and others in the literature due to the lack of clinical information from patients in other studies. There were no associations between FAL frequencies in at least one marker and the histopathologic parameters of tumor grade, metastasis in axillary lymph nodes, tumor size, and hormonal status, indicating that the occurrence of genetic instability in this region is not associated with development and progression of the disease. Schwarzenbach et al. [13] analyzed the LOH frequency of seven microsatellite markers, including D9S171, in primary breast tumor, peripheral blood, and bone marrow of 40 patients with nonmetastatic breast cancer and 48 patients with metastatic disease. They found an association between the marker D9S171 and LOH in highenuclear grade tumors, suggesting that the CDKN2A gene may be involved in the pathogenesis of sporadic breast cancer. This result does not agree with the results observed in our study, where the frequencies of LOH at D9S171 and in the analysis of all the markers together were distributed homogeneously among the various histologic grades. We found a small association between LOH in D9S200 (9p13.1) and the histologic grade, primarily due to the absence of LOH in tumors classified as grade I. To examine whether the low sample number (n 5 17) may have decreased the statistical power of the test, the data must be validated using a larger patient population. The same result was observed when we analyzed LOH at D9S1748. We observed a LOH frequency around 18% in ERpositive patients (n 5 65), and no LOH was observed in ER-negative patients (n 5 27). This result is in discordance with that described by Huiping et al. [22] in lobular carcinoma. They found a significant association between LOH at 9p and low content of ER and PR (P 5 0.001), suggesting that LOH on 9p might be associated with loss of ER and PR. Taking the sample size into consideration, it may be interesting to explore these data, since Huiping et al. applied their studies to a specific type of breast cancer. In conclusion, our data agree with other data from literature that point to LOH as a secondary mechanism of tumor suppressor inactivation on 9p in breast cancer, showing lower frequencies and a lack of association with clinicopathologic parameters than those observed in other types of cancer. On the other hand, our data point to an interesting association

S.F. Vieira de Oliveira et al. / Cancer Genetics and Cytogenetics 200 (2010) 23e27

between the concordant loss of the CDKN2A and MTAP genes, which was not sufficiently explored in primary breast cancer.

[11]

Acknowledgments [12]

This work was partially supported by a research grant from CNPq (Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico, Brazil).

[13]

References [14] [1] Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer Statistics 2009. Cancer J Clin 2009;59:225e49. [2] INCA Instituto Nacional do Caˆncer. Ministe´rio da Sau´de. Estimativa 2010. Available at: http://www1.inca.gov.br/estimativa/2010/). Last accessed on 26 February 2010. [3] Pasche B. Recent advances in breast cancer genetics. Cancer Treat Res 2008;141:1e10. [4] Oudin C, Bonnetain F, Boidot R, Ve´gran F, Soubeyrand M-S, Arnould L, Riedinger J-M, Lizard-Nacol S. Patterns of loss of heterozygosity in breast carcinoma during neoadjuvant chemotherapy. Int J Oncol 2007;30:1145e51. [5] Wuicik L, Cavalli LR, Corne´lio DA, Schmid Braz AT, Barbosa ML, Lima RS, Urban CA, Bleggi Torres LF, Ribeiro EMSF, Cavalli IJ. Chromosome alterations associated with positive and negative lymph node involvement in breast cancer. Cancer Genet Cytogenet 2007; 173:114e21. [6] Marsit CJ, Wiencke JK, Nelson HH, Kim DH, Hinds PW, Aldape K, Kelsey KT. Alterations of 9p in squamous cell carcinoma and adenocarcinoma of the lung: association with smoking, TP53, and survival. Cancer Genet Cytogenet 2005;162:115e21. [7] Krieger D, Moericke A, Oschlies I, Zimmermann M, Schrappe M, Reiter A, Burkhardt B. Frequency and clinical relevance of DNA microsatellite alterations of the CDKN2A/B, ATM and p53 gene loci: a comparison between pediatric precursor T-cell lymphoblastic lymphoma and T-cell lymphoblastic leukemia. Haematologica 2010;95:158e62. [8] Mistry SH, Taylor C, Randerson-Moor JA, Harland M, Turner F, Barret JH, Whitaker L, Jenkins RB, Knowles MA, Bishop JA, Bishop DT. Prevalence of 9p21 deletions in UK melanoma families. Genes Chromosomes Cancer 2005;44:292e300. [9] Zainuddin N, Jaafart H, Isa MN, Abdullah JM. Loss of heterozygosity on chromosomes 10q, 9p, 17p and 13q in Malays with malignant glioma. Neural Res 2004;26:88e92. [10] Kamb A, Shattuck-Eidens D, Eeles R, Liu Q, Gruis NA, Ding W, Hussey C, Tran T, Miki Y, Weaver-Feldhaus J, McClure M, Aitken JF, Anderson DE, Bergman W, Frants R, Goldgar DE, Green A, MacLennan R, Martin NG, Meyer LJ, Youl P, Zone JJ,

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

27

Skolnick MH, Cannon-Albright LA. Analysis of the p16 gene (CDKN2) as a candidate for the chromosome 9p melanoma susceptibility locus. Nature Genet 1994;8:23e6. Illei PB, Rusch VW, Zakowski MF, Ladanyi M. Homozygous deletion of CDKN2A and codeletion of the methylthioadenosine phosphorylase gene in the majority of pleural mesotheliomas. Clin Cancer Res 2003;9:2108e13. Lubin M, Lubin A. Selective killing of tumors deficient in methylthioadenosine phosphorylase: a novel strategy. PLoS One 2009;4: e5735. Schwarzenbach H, Muller V, Beeger C, Gottberg M, Stahmann N, Pantel C. A critical evaluation of loss of heterozygosity detected in tumor tissues, blood serum and bone marrow plasma from patients with breast cancer. Breast Cancer Res 2007;9:66e73. Miller BJ, Wang D, Krahe R, Wright FA. Pooled analysis of loss of heterozygosity in breast cancer: a genome scan provides comparative evidence for multiple tumor suppressors and identifies novel candidate regions. Am J Hum Genet 2003;73:748e67. Ellsworth RE, Ellsworth DL, Neatrour DM, Deyarmin B, Lubert SM, Sarachine MJ, Brown P, Hooke JA, Shriver CD. Allelic imbalance in primary breast carcinomas and metastatic tumors of the axillary lymph nodes. Mol Cancer Res 2005;3:71e7. Brenner AJ, Aldaz CM. Chromosome 9p allelic loss and p16/CDKN2 in breast cancer and evidence of p16 inactivation in immortal breast epithelial cells. Cancer Res 1995;55:2892e5. Smeds J, Wa¨rnberg F, Norberg T, Nordgren H, Holmberg L, Bergh J. Ductal carcinoma in situ of the breast with different histopathological grades and corresponding new breast tumour events: Analysis of loss of heterozygosity. Acta Oncol 2005;44:41e9. Argos M, Kibriya MG, Jasmine F, Olopade OI, Su T, Hibshoosh H, Ahsan H. Genome wide scan for loss of heterozygosity and chromosomal amplification in breast carcinoma using singlenucleotide polymorphism arrays. Cancer Genet Cytogenet 2008; 182:69e74. Cairns P, Polascik TJ, Eby Y, Tokino K, Califano J, Merlo A, Mao Herath J, Jenkins R, Westra W. Frequency of homozygous deletion at p16/CDKN2 in primary human tumours. Nat Genet 1995;11: 210e2. Christopher SA, Diegelman P, Porter CW, Kruger WD. Methylthioadenosine phosphorylase, a gene frequently codeleted with p16cdkN2a/ARF, acts as a tumor suppressor in a breast cancer cell line. Cancer Res 2002; 62:6639e44. Powell EL, Leoni LM, Canto MI, Forastiere AA, IocobuzioDonahue CA, Wang JS, Maltra A, Montgomery E. Concordant loss of MTAP and p16/CDKN2A expression in gastroesophageal carcinogenesis: evidence of homozygous deletion in esophageal noninvasive precursor lesions and therapeutic implications. Am J Surg Pathol 2005;29:1497e504. Huiping C, Sigurgeirsdottir JR, Jonasson JG, Eiriksdottir G, Johannsdottir JT, Egilsson V, Ingvarsson S. Chromosome alterations and E-cadherin gene mutations in human lobular breast cancer. Brit J Cancer 1999;81:1103e10.