Get 20M+ Full-Text Papers For Less Than $1.50/day. Subscribe now for You or Your Team.

Learn More →

Methylation and protein expression of DNA repair genes: association with chemotherapy exposure and survival in sporadic ovarian and peritoneal carcinomas

Methylation and protein expression of DNA repair genes: association with chemotherapy exposure... Background: DNA repair genes critically regulate the cellular response to chemotherapy and epigenetic regulation of these genes may be influenced by chemotherapy exposure. Restoration of BRCA1 and BRCA2 mediates resistance to platinum chemotherapy in recurrent BRCA1 and BRCA2 mutated hereditary ovarian carcinomas. We evaluated BRCA1, BRCA2, and MLH1 protein expression in 115 sporadic primary ovarian carcinomas, of which 31 had paired recurrent neoplasms collected after chemotherapy. Additionally, we assessed whether promoter methylation of BRCA1, MLH1 or FANCF influenced response to chemotherapy or explained alterations in protein expression after chemotherapy exposure. Results: Of 115 primary sporadic ovarian carcinomas, 39 (34%) had low BRCA1 protein and 49 (42%) had low BRCA2 expression. BRCA1 and BRCA2 protein expression were highly concordant (p < 0.0001). MLH1 protein loss occurred in 28/115 (24%) primary neoplasms. BRCA1 protein loss in primary neoplasms was associated with better survival (p = 0.02 Log Rank test) and remained significant after accounting for either stage or age in a multivariate model (p = 0.04, Cox proportional hazards). In paired specimens, BRCA1 protein expression increased in 13/21 (62%) and BRCA2 protein expression increased in 15/21 (71%) of recurrent carcinomas with low or intermediate protein in the paired primary. In contrast MLH1 expression was rarely decreased in recurrent carcinomas (1/33, 3%). Similar frequencies of MLH1, BRCA1, and FANCF promoter methylation occurred in primary carcinomas without previous chemotherapy, after neoadjuvant chemotherapy, or in recurrent neoplasms. Conclusion: Low BRCA1 expression in primary sporadic ovarian carcinoma is associated with prolonged survival. Recurrent ovarian carcinomas commonly have increased BRCA1 and/or BRCA2 protein expression post chemotherapy exposure which could mediate resistance to platinum based therapies. However, alterations in expression of these proteins after chemotherapy are not commonly mediated by promoter methylation, and other regulatory mechanisms are likely to contribute to these alterations. Page 1 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 acetylation could be a modifiable mechanism of sensitiv- Background Ovarian carcinoma is the most deadly gynecological ity or resistance to chemotherapy. Indeed, in vitro data malignancy and is the fifth leading cause of carcinoma indicate that epigenetic alterations in MLH1 and FANCF death in American women. Ovarian carcinomas are usu- modulate response to platinum agents in cell lines ally responsive to initial platinum based chemotherapy [23,28]. While loss of BRCA1, BRCA2, or FANCF confers regimens. However, even after a complete clinical sensitivity to platinum, loss of MLH1 confers resistance to response, most ovarian carcinomas do recur, with resist- platinum agents [28]. The importance of these mecha- ance to platinum therapy developing after one or more nisms in in vivo chemoresistance in ovarian carcinomas chemotherapy courses. Acquired chemotherapy resistance has not been defined. is one of the greatest clinical challenges in the treatment of women with ovarian carcinoma. We undertook this study to determine if protein expres- sion and/or promoter methylation of BRCA1, MLH1, Germline mutations in the BRCA1 and BRCA2 genes con- BRCA2 and FANCF predicted overall survival in primary fer inherited susceptibility to ovarian and breast carcino- sporadic ovarian carcinomas and whether promoter mas. Breast and ovarian carcinomas in BRCA1 or BRCA2 methylation or protein expression were altered by expo- mutation carriers usually have somatic deletions of the sure to chemotherapy in recurrent carcinomas. We also wildtype allele, rendering the neoplasm BRCA1 or BRCA2 wanted to determine the relationship of alterations of deficient. BRCA2 is identical to the Fanconi anemia (FA) proteins in the FA-BRCA pathway with alterations in p53, gene FANCD1[1]. BRCA1, BRCA2, other FA genes, and a another key sensor of DNA damage. larger number of protein partners (including the DNA mismatch repair protein MLH1) are part of a complex Results DNA damage response network (reviewed in[2]). In vitro BRCA1, MLH1 and BRCA2 protein expression in primary ovarian carcinomas studies indicate that BRCA1 and BRCA2 loss increases sensitivity to agents that cause double strand DNA breaks Protein expression for BRCA1, MLH1 and BRCA2 was and/or interstrand DNA cross-links including platinum assessed with immunohistochemistry in 115 sporadic pri- agents [3-5]. Conversely, loss of BRCA1 or BRCA2 may mary invasive neoplasms. In 31 cases, we had paired increase resistance to microtubule interfering agents such recurrent or persistent neoplasm tissue available after as taxanes and vincristine [6,7]. Most studies report chemotherapy which allowed for comparison of protein improved survival in women with ovarian carcinomas expression with the matched primary. Eight primary inva- associated with BRCA1 and BRCA2 mutations compared sive carcinomas were classified as peritoneal in origin and to women with sporadic ovarian carcinoma, consistent the remainder ovarian in origin (Table 1). Clinical and with increased sensitivity to platinum-based chemother- pathological characteristics and relationship to BRCA1 apy [8-10]. We and others have recently shown that carci- protein expression are summarized in Table 1. Peritoneal nomas from patients with inherited frameshift mutations carcinomas were less likely to have low BRCA1 protein in BRCA1 or BRCA2 exposed to chemotherapy can than ovarian carcinomas (p = 0.049, Fishers Exact, two- acquire secondary mutations that restore the reading tailed). There was no association between BRCA1 protein frame of BRCA1 or BRCA2, resulting in platinum resist- expression and grade, histology, or adequacy of surgical ance [11-13]. Thus, restoration of expression of proteins cytoreduction. Stage I carcinomas were more likely to in the FA-BRCA DNA damage response pathway may have normal BRCA1 protein expression compared to stage increase resistance to platinum and potentially other II-IV carcinomas (p = 0.03, Fisher exact, two-tailed). As agents that induce DNA damage. expected both stage I and II carcinomas were also more likely to have non serous and undifferentiated histologies While germline mutations in BRCA1 and BRCA2 predis- (8/11, 73% for stage I and 3/4 for stage II, p < 0.0001 com- pose to hereditary ovarian carcinoma, somatic mutations pared to stage III and IV, Fishers Exact, two-tailed). Stage I in these genes are rare in sporadic ovarian carcinomas [14- histologies included two serous, one undifferentiated car- 17]. However, epigenetic alterations in these and other cinoma, one small cell, two mucinous, five endometrioid, DNA repair genes may play important roles in sporadic while stage II included 3 endometrioid and one serous. ovarian and breast carcinomas and could contribute to However, normal BRCA1 protein expression was associ- responsiveness to chemotherapy. Promoter methylation ated with Stage I disease and not histology. Both serous leads to decreased expression of BRCA1, MLH1, and stage I carcinomas had normal BRCA1 protein expression FANCF protein in a subset of ovarian carcinomas [18-23]. and 2/3 of endometrioid stage II carcinomas had low BRCA1 promoter methylation occurs in 5–20% of spo- BRCA1 protein expression. radic ovarian carcinomas[20,22,24,25], while BRCA2 methylation is rare[26,27]. FANCF is a key regulator of the BRCA1, BRCA2, and MLH1 protein expression is summa- FA-BRCA pathway[23]. Epigenetic regulation of gene tran- rized in Table 2 and representative staining is shown in scription through promoter methylation or histone Figure 1. BRCA1 and BRCA2 protein levels were signifi- Page 2 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 Table 1: Clinicopathological characteristics of the sporadic neoplasms studied and BRCA1 protein expression. BRCA1 Protein Expression All Cases Low Intermediate Normal p value Primary site Ovarian 107 (93%) 39 27 41 p = 0.05 Peritoneal 8 (7%) 0 3 5 Grade Grade 1 6 (5%) 3 2 1 Grade 2 16 (16%) 5 3 8 NS Grade 3 93 (79%) 31 25 37 Histology Serous 81 (70%) 28 24 29 NS Endometrioid 11 (10%) 4 2 5 Carcinoma NOS 12 (10%) 3 3 6 MMMT 4 (3%) 1 1 2 Clear Cell 1 (1%) 0 0 1 Mucinous 3(3%) 1 0 2 Other* 3 (3%) 2 0 1 Stage I 11 (10%) 2 1 8 p = 0.03 II 4 (3%) 3 0 1 III 83 (72%) 28 26 29 IV 17 (15%) 6 3 8 Cytoreduction Optimal (< 1 cm) 73 24 18 31 NS Suboptimal 36 14 9 13 Not Available 6 1 3 2 Total 116 39 (36%) 30 (26%) 46 (38%) * Other histologies included one small cell, one transitional cell carcinoma and one carcinoma with focal giant cells. cantly and positively correlated (p < 0.0001, two-tailed). therapy influenced MLH1, BRCA1, or BRCA2 protein BRCA1 and BRCA2 protein expression were not related to levels. In 7 cases a second recurrence was available for MLH1 protein expression. expression analyses. For second recurrences, data was tab- ulated as if for a separate case when compared to the pri- BRCA1, MLH1 and BRCA2 protein expression in paired mary. Recurrent neoplasms were obtained at varying time primary and recurrent ovarian carcinomas intervals from last chemotherapy. We separated matched MLH1, BRCA1, and BRCA2 protein expression was also pairs into two groups depending on the interval since last assessed in 31 matched primary and recurrent ovarian car- chemotherapy exposure (≤ six months or >six months). cinomas from the same patient to determine if chemo- Complete data for protein and methylation analyses for each case is presented in Table S1; Additional file 1. Table 2: Protein expression in primary sporadic ovarian and Changes in protein expression in paired neoplasms are peritoneal carcinomas before chemotherapy exposure. shown schematically in Figure 2 and representative stain- ing in Figure 3. Protein Expression Low Intermediate Normal Total (≤10%) (10–30%) (>30%) Increased BRCA1 or BRCA2 protein would be expected to increase resistance to platinum therapy. In contrast, BRCA1 expression 39 (34%) 30 (26%) 46 (40%) 115 decreased MLH1 protein should be associated with plati- BRCA2 expression 49 (42%) 28 (24%) 38 (33%) 115 num resistance. In the paired sets most primary neo- MLH1 Expression* 28 (24%) 87 (76%)* 115 plasms (26/31, 84%) had normal MLH1 protein. Of the 33 recurrences occurring in the 26 cases with normal MLH1 expression was scored in two categories: ≤10% positive cells = MLH1 protein in the primary, only one had loss of MLH1 Low, >10% positivity = Normal. Page 3 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 or intermediate BRCA1 expression showed reduced BRCA1 expression in the recurrence (p = 0.01). Therefore, increases in BRCA1 expression were more common than decreases in BRCA1 expression following chemotherapy. BRCA2 protein expression also frequently varied between paired primary and recurrent neoplasms (Figures 2 and 3, Additional file 1). Among primary neoplasms with low or intermediate BRCA2 expression, the paired post-chemo- therapy specimen had increased BRCA2 expression in 15/ 21 cases (71%). In contrast to the observation for BRCA1 protein expression, primary neoplasms with normal or intermediate BRCA2 expression were equally likely to have decreased BRCA2 expression (14/23, 61%). Excluding those cases with normal BRCA1 and BRCA2 protein expression in the primary, 20 (80%) recurrent car- cinomas had increases in either BRCA1 or BRCA2 protein while five (20%) did not have an increase in expression of either protein. Likelihood of a complete response to sub- sequent chemotherapy was not related to whether increased BRCA1 or BRCA2 protein was identified in the paired recurrence, but wide heterogeneity in treatment Repr MLH1 in s Figure 1 esentativ pore ad pr ic ovarian carcino otein expression of BRCA1, mas BRCA2 and precludes assessment of response to specific agents. Representative protein expression of BRCA1, BRCA2 and MLH1 in sporadic ovarian carcinomas. BRCA1, MLH1, and FANCF promoter methylation in Protein expression is represented by brown stain. Black bars primary and recurrent ovarian carcinomas in the lower left corners represent 10 microns. A. BRCA1 Methylation was assessed for MLH1, BRCA1 and FANCF protein in a neoplasm with low expression. B. BRCA1 pro- in 104 primary and 36 recurrent sporadic carcinomas. Of tein in a neoplasm with normal expression. C. BRCA2 pro- the primary carcinomas tested for methylation, 11 were tein in a neoplasm with low expression. D. BRCA2 protein in obtained after neoadjuvant chemotherapy and 93 were a neoplasm with normal expression. E. MLH1 protein in a chemotherapy-naive. Results of methylation analysis for neoplasm with low expression. F. MLH1 protein in a neo- BRCA1, FANCF and MLH1 are summarized in Table 3. In plasm with normal expression. no cases were neoplasms methylated at more than one of these genes. Methylation did not vary between primary protein (3.0%, Figures 2 and 3, and Table S1 in Additional cases exposed to neoadjuvant chemotherapy or in primary file 1). Neither the primary nor recurrent neoplasm in this vs. recurrent cases for any gene (Table 3). Ten of the pri- pair had MLH1 promoter methylation. mary and recurrent neoplasms were matched specimens from the same patient collected at two different time In contrast to MLH1, BRCA1 and BRCA2 expression dem- points (Table S1; Additional file 1). All ten cases showed onstrated greater variability between paired primary and concordant methylation between primary and recurrence, recurrent neoplasms (Figure 2, Table S1;Additional file 1). and none were methylated at any of the three genes. In 16 carcinomas recurring after 13 primaries with normal Reduction of BRCA1 protein was significantly associated BRCA1 expression, 13(81%) maintained similar protein with BRCA1 methylation (p = 0.02). However, BRCA1 expression and three (19%) demonstrated decreased methylation only accounted for 7/52 (13.5%) of those BRCA1 expression in the recurrence. In 13 recurrences sporadic carcinomas with low or intermediate BRCA1 that followed 11 paired primaries with intermediate expression. BRCA1 expression, four had reduced, seven had increased and two had similar BRCA1 expression in the paired All cases with MLH1 methylation demonstrated microsat- recurrence. In eight recurrences that followed six primary ellite instability at BAT26 (data not shown). MLH1 meth- carcinomas with low BRCA1 expression, 2 had similar ylation was associated with loss of protein (p = 0.03), but and six (75%) had increased BRCA1 expression in the one methylated case did have apparently normal MLH1 recurrence. Overall 13 of 21 (62%) neoplastic pairs with protein expression. MLH1 methylation was more com- low or intermediate BRCA1 protein expression in the pri- mon in endometrioid ovarian carcinomas compared to all mary had increased BRCA1 protein in the recurrence. In other histologies (3/14, 21% versus 1/116, 1%, p = contrast, only eight of 33 (24%) neoplasms with normal Page 4 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 Schematic of BRCA1 Figure 2 , BRCA2, and MLH1 protein expression in paired primary and recurrent neoplasms Schematic of BRCA1, BRCA2, and MLH1 protein expression in paired primary and recurrent neoplasms. Each neoplasm is represented by a single horizontal line. A. BRCA1 protein expression in 24 primary and paired recurrent neo- plasms obtained ≤ 6 months since last chemotherapy, 8 with second recurrences. B. BRCA1 expression in 7 primary and paired recurrent neoplasms in which the recurrence was obtained more than 6 months since last chemotherapy. C. BRCA2 protein expression in 24 primary and paired recurrent neoplasms obtained ≤ 6 months since last chemotherapy. D. BRCA2 expression in 7 primary and paired recurrent neoplasm obtained more than 6 months since last chemotherapy. E. MLH1 pro- tein expression in 24 primary and paired recurrent neoplasms obtained ≤ 6 months since last chemotherapy. F. MLH1 expres- sion in 7 primary and paired recurrent neoplasm obtained more than 6 months since last chemotherapy. Page 5 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 nor FANCF methylation were associated with p53 muta- tion status. p53 mutations were more common in spo- radic ovarian carcinomas with loss or reduction of BRCA1 expression (Table 4, p = 0.01 Fishers Exact). The majority of p53 mutations were missense mutations occurring in the DNA binding domain (exons 5–8). Null mutations including frameshift, splice site, and nonsense mutations accounted for 9/33 (27%) of p53 mutations. Overall survival and protein expression or promoter methylation Individual factors influencing survival in this cohort in univariate analyses were stage (p = 0.01), age (p = 0.03) and optimal cytoreduction to a maximum neoplastic diameter less than 1 cm (p < 0.0001) and low BRCA1 expression (p = 0.02). Low BRCA1 expression in the pri- mary carcinoma was associated with longer survival com- pared to intermediate or normal BRCA1 expression (median survival 62 months vs. 45 months, p = 0.02 LogRank Test, Hazard Ratio 0.59, 95% confidence interval 0.37–0.93, Figure 4). Low BRCA1 expression remained significantly associated with improved survival in a Cox multi regression model with the covariates age (p = 0.04) or stage (p = 0.04) but was no longer significant when BRCA1, BRCA2, and MLH m Figure 3 ary and recurrent neopla1 sms protein expression in paired pri- using the co-variate optimal cytoreduction (p = 0.10). BRCA1, BRCA2, and MLH1 protein expression in Individual factors not related to survival included grade, paired primary and recurrent neoplasms. Protein MLH1 or BRCA2 protein expression, methylation of expression is represented by brown stain. Black bars in the BRCA1, MLH1 or FANCF genes, or p53 mutation. lower left corners are equal to 10 microns. A. BRCA1 expression is low in the primary neoplasm. B. BRCA1 expression is increased in the paired recurrent neoplasm. C. Response to chemotherapy BRCA2 protein expression is low in a different primary neo- The likelihood of a complete response to initial chemo- plasm. D. BRCA2 expression is increased in the paired recur- therapy was not related to BRCA1 protein levels. Likewise, rent neoplasm. E. MLH1 protein is normal in a different there was no significant difference in the complete primary neoplasm. F. In the paired recurrent neoplasm, response rate for cases with BRCA1 methylation, MLH1 MLH1 protein expression is reduced. methylation, or for cases with either BRCA1 or FANCF methylation. 0.003). FANCF and BRCA1 methylation were not associ- ated with histology. Discussion BRCA1 protein loss is common in sporadic epithelial Relationship of p53 mutations to protein expression and ovarian and peritoneal carcinomas. Low BRCA1 protein methylation of DNA repair genes expression was associated with a significantly improved p53 mutations were assessed by DNA sequencing in all overall survival. Our data suggest that somatic loss of cases tested for methylation and neither BRCA1, MLH1, BRCA1 favourably influences survival similar to the Table 3: Methylation of DNA repair genes in primary and recurrent ovarian or peritoneal carcinomas. PROPORTION METHYLATED BRCA1 MLH1 FANCF Primary, no chemotherapy 6/91 (6.6%) 3/93 (3.2%) 3/93 (3.2%) Primary, post neoadjuvant chemo 2/11 (18%) 1/11 (9.1%) 0/10 Recurrent* 2/31 (6.5%) 0/30 0/31 (0%) Total (2.2%) 10/133 (7.5%) 4/134 (3.0%) 3/134 * Ten recurrent carcinomas were paired specimens for which we also evaluated the primary neoplasm. None of these 10 paired cases were methylated at any of the three genes in either the primary or recurrent case. One of these 10 pairs did show loss of MLH1 protein in the recurrent neoplasm, but was not methylated. Page 6 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 Table 4: p53 mutations and BRCA1 protein level in primary sporadic ovarian carcinomas P53 Mutations BRCA1 protein Expression Total Tested Missense Null Total P53 mutations N (%) N (%) Low 28 10 (36%) 5 (18%) 15 (54%) Intermediate 18 8 (44%) 2 (11%) 10 (56%) Normal 33 6(18%) 2 (6%) 8 (24%)* Total 79 24 (30%) 9 (11%) 33 (42%) *p = 0.01 for difference in p53 mutation frequency for cases with normal versus reduction or loss of BRCA1 protein. improved survival in ovarian carcinomas associated with To our knowledge, this is the first study of BRCA2 protein inherited BRCA1 mutations. Two previous studies have expression in human ovarian carcinomas. Unlike BRCA1, examined the relationship between BRCA1 protein BRCA2 protein expression was not associated with overall expression and prognosis in ovarian carcinomas. Thrall survival. Our data demonstrate that BRCA1 and BRCA2 and colleagues evaluated a large number of sporadic protein expression are highly concordant in ovarian carci- advanced stage carcinomas and found that BRCA1 loss nomas. In both mouse and human, BRCA1 and BRCA2 was strongly protective for overall survival [29]. In have nearly identical patterns of message and protein another study of un-selected ovarian carcinoma, Wang expression in embryonic development, breast morpho- and colleagues failed to find an association between genesis, and in many adult tissues, suggesting that the two BRCA1 protein and prognosis [30]. However, that series genes share regulatory networks [31-34]. Similarly, in evaluated a large number of endometrioid, clear cell, breast and ovarian carcinoma cell lines, BRCA1 and grade 1 and stage I cases, with few high-grade, advanced BRCA2 mRNA are concordantly induced by adriamycin, stage, serous carcinomas, distinctly different from the ionizing radiation, and estrogen [35-37]. Our data suggest neoplasms in our study. Indeed, among the 29 stage III that ovarian carcinomas maintain the coordinate regula- and IV cases in Wang's study, there was a trend toward a tion of BRCA1 and BRCA2 seen in normal tissues. survival advantage with loss of BRCA1 protein [30]. Our data combined with these two previous studies confirm We hypothesized that BRCA1 and BRCA2 protein levels an improved prognosis for women with advanced ovarian would increase after exposure to chemotherapy, thereby carcinomas with low BRCA1 protein expression. mediating increasing platinum resistance during the dis- ease course. Since recurrent ovarian carcinomas are not routinely subjected to biopsy, our cases represent a wide variety of disease time points depending on the clinical indication for biopsy or surgery in a given patient. Conse- quently, our patients received varied amounts and types of chemotherapy before and after biopsies. An increase in BRCA1 or BRCA2 protein occurred in the majority of recurrent neoplasms that had low or intermediate protein in the paired primary. Increases in BRCA1 protein were significantly more likely than reduction of protein in paired post chemotherapy neoplasms, suggesting a selec- tion for increased BRCA1 expression. In contrast, BRCA2 protein both increased and decreased frequently post chemotherapy. An increasing tendency toward platinum resistance during the typical ovarian carcinoma disease course is a well recognized clinical challenge. Increased Overall survival in sporadi Figure 4 c ovarian carcinomas relation to BRCA1 expression in primary BRCA1 and BRCA2 expression could mediate that resist- Overall survival in relation to BRCA1 expression in ance. However, given the treatment heterogeneity of our primary sporadic ovarian carcinomas. Overall survival cases, we cannot directly relate the alterations in BRCA1 or was significantly improved in primary ovarian carcinomas (p BRCA2 expression with specific treatment responses. = 0.02, LogRank test) with low BRCA1 protein expression (median survival 62 months) compared to carcinomas with Our data are consistent with previous reports that BRCA1 intermediate or normal BRCA1 expresssion (median survival 45 months). promoter methylation occurs in 5–20% of sporadic ovar- ian carcinomas [20-22,24,25]. Interestingly, we found Page 7 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 that promoter methylation only occurs in 15% of those with treatment response, opposite to the expected associ- sporadic ovarian carcinomas with low or intermediate ation [44]. We used a traditional cut-off to define MLH1 BRCA1 expression. Thus, other regulators of BRCA1 protein deficiency at 10% of cells while the other two expression may be more important than promoter meth- studies used a continuous scoring system. However, even ylation in ovarian carcinomas for both primary and recur- when we re-evaluated our data with a continuous scoring rent neoplasms. We found no association between BRCA1 system, we still found that significant decrease in MLH1 methylation and survival in sporadic ovarian carcinomas, protein expression was rare (data not shown). The dis- in contrast to a recent small study that demonstrated a sur- crepancies between these studies and ours may stem from vival disadvantage for patients whose neoplasms were the clinical heterogeneity of cases for which paired pri- methylated at BRCA1 [38]. Our data suggest that BRCA1 mary and recurrent ovarian carcinomas are available. protein expression is a better predictor of survival than Finally, we did not find MLH1 promoter methylation BRCA1 promoter methylation, not surprising given the more frequently in either recurrent neoplasms or in pri- relatively small number of cases with BRCA1 methylation. mary neoplasms exposed to neoadjuvant chemotherapy, nor was MLH1 methylation identified in the neoplastic Low BRCA1 protein expression was associated with an pair with reduced MLH1 expression in the recurrence. increased likelihood of p53 mutation, low BRCA2 protein Consequently, our data do not support a major role for in expression, and improved survival. Only one previous vivo epigenetic alteration of MLH1 expression in the devel- study of which we are aware has examined the association opment of clinical platinum resistance in sporadic ovar- of somatic BRCA1 alterations and p53 status. In that ian carcinomas. study, BRCA1 methylation did not correlate with immu- nohistochemically detected p53 protein expression [22]. FANCF may also mediate platinum sensitivity. FANCF However, p53 immunohistochemistry correlates imper- promoter methylation confers cisplatin sensitivity to cell fectly with the presence of a somatic mutation, and lines and was hypothesized to be a mediator of in vivo BRCA1 methylation accounts for only a small fraction of platinum sensitivity[23]. In order to mediate resistance, a BRCA1 protein loss, so our studies are not equivalent. The neoplasm should be initially methylated and then lose majority of breast, ovarian, and peritoneal carcinomas methylation in recurrent disease. However, our findings associated with germline mutations in BRCA1 and BRCA2 indicate that FANCF methylation occurs rarely in primary have p53 mutations, at rates higher than found in their ovarian carcinomas (<5% of cases), and is therefore not sporadic counterparts [39-42]. Thus, many sporadic ovar- likely to be a major clinical mediator of platinum sensitiv- ian carcinomas have reduction of BRCA1 protein, muta- ity. tions in p53, and better overall survival, mimicking the phenotype of hereditary BRCA1-associated ovarian carci- Conclusion noma. Low BRCA1 protein expression in sporadic ovarian carci- nomas is associated with a favourable survival and an MLH1 is a DNA mismatch repair gene required for sensi- increased rate of p53 mutations. BRCA1 and BRCA2 pro- tivity to platinum compounds. MHL1 is commonly meth- teins are concordantly expressed in sporadic ovarian and ylated in colorectal and endometrial carcinomas and less peritoneal carcinomas. Increases in BRCA1 and BRCA2 commonly in ovarian carcinomas [19,24]. MLH1 methyl- protein expression are common in recurrent sporadic ation mediates platinum resistance in the ovarian carci- ovarian carcinomas, but the mechanism(s) responsible noma cell line A2780 [18]. We postulated that for these expression differences are unknown. Alteration chemotherapy exposure could select for neoplastic cells in methylation of the promoters of BRCA1, MLH1, and with MLH1 methylation. Indeed, cell-free methylated FANCF do not seem to commonly mediate clinical chem- MLH1 DNA is increased in the plasma of women with oresistance in women with ovarian carcinomas. ovarian carcinoma at the time of clinical relapse [43]. Increased methylation of MLH1 in recurrent ovarian car- Methods cinoma has been proposed as a rationale for clinical trials Specimens of demethylating agents alone or in combination with Patients consented to have tissue collected and to provide chemotherapy. We found no association between MLH1 blood samples using protocols approved by the Human protein and overall survival or response to chemotherapy, Subjects Division of the University of Washington Institu- and we rarely observed (3.8%) loss of MLH1 protein in tional Review Board. Paraffin blocks were obtained from paired neoplasms following chemotherapy exposure. Our pathology archives. Tissues were obtained at the time of data contrast with those of two groups who both reported surgery and snap frozen in liquid nitrogen. Tissue sections a significant decrease in MLH1 protein expression in were stained with hematoxylin and eosin and reviewed paired ovarian carcinomas following platinum chemo- for neoplastic cell percentage prior to DNA extraction. therapy [44,45]. But surprisingly, Fink and colleagues DNA was extracted using the Stratagene kit according to found that MLH1 protein expression correlated inversely the manufacturer's instructions. Normal DNA was Page 8 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 extracted from peripheral white blood cells or from nor- Controversy exists regarding the meaning and specificity mal tissue after histological review excluded contamina- of cytoplasmic staining for BRCA1 and BRCA2. Proteins tion with neoplasm. Family history was considered strong encoded by aberrantly spliced isoforms have been shown if the patient had a known BRCA1 or BRCA2 mutation, to localize to the cytoplasm for both BRCA1 and BRCA2 personal history of breast carcinoma, any relative with [47-49]. The MS110 antibody has better specificity than ovarian carcinoma, a first degree relative with premeno- other anti-BRCA1 antibodies and has an almost exclu- pausal breast carcinoma, or two relatives of any degree sively nuclear staining pattern[46]. We observed exclu- with breast carcinoma at any age. Cases without strong sively nuclear BRCA1 protein staining in both normal and family histories were considered sporadic, while those neoplastic tissues. We observed both nuclear and cyto- with suggestive family histories were considered familial plasmic BRCA2 staining in neoplastic cells, but scored and excluded. only nuclear staining as likely to be relevant to intact DNA repair. BRCA1 and BRCA2 protein was scored as previ- Immunohistochemical staining for MLH1, BRCA1 and ously described: "low" if fewer than 10% of neoplastic BRCA2 cells had nuclear staining and "intermediate" if 10% to BRCA1 protein was detected in formalin fixed paraffin 30% of neoplastic cells had nuclear staining and normal sections using the mouse monoclonal antibody MS110 if greater than 30% of neoplastic cells had nuclear staining (previously called Ab-1, Oncogene Research Products) as [46]. MLH1 protein expression was scored as loss (<10% previously described [46]. The MS110 antibody recog- of cells positive) or normal (>10% of cells positive) as is nizes an amino terminal epitope in BRCA1 (amino acid done routinely for clinical purposes. MLH1 staining was residues 89–222). BRCA2 was detected using the H-300 exclusively nuclear. For the purpose of comparing protein rabbit polyclonal antibody (Santa Cruz Biotech) that rec- expression in paired primary and recurrent neoplasms, ognizes amino acid residues 2520–2819 of human the second recurrence in those seven cases with more than BRCA2. MLH1 was detected using the mouse monoclonal one recurrence was compared with the primary neo- antibody G168-728 (BD Pharmingen). Briefly, paraffin plasms and counted as a separate case. sections were deparaffinized, re-hydrated, and treated with steam heat for 20 minutes using antigen target Methylation specific PCR Carcinomas were evaluated for BRCA1, MLH1 and FANCF retrieval solution (DAKO). Endogenous peroxidase activ- ity was quenched by treatment with 3% H O for 5 min- promoter methylation as previously described [21,23,50]. 2 2 utes. Sections were washed with PBS and non-specific All methylation specific PCR assays included a positive binding was blocked by treatment for three hours in 2% control (in vitro methylated DNA, Invitrogen) and a neg- bovine serum albumin in PBS. Primary antibody was ative (water) control. Positive assays were repeated at least diluted in PBS (MS110 at 1:250 dilution, H-300 at 1:100 once. Sodium bisulfite treated DNA was amplified with dilution, and G168-728 at 1:80) and applied to sections primers specific for either methylated or unmethylated for 14–16 hours at 4° (MS110 and H-300) or 40 minutes treated substrate. PCR products were electrophoresed on at room temperature (G168-728). Secondary antibody 3% agarose gels or 5% acrylamide gels and stained with and streptavidin biotin-peroxidase were from Universal ethidium bromide. Large Volume LSAB+, Peroxidase kit (DAKO) and were each applied for 30 minutes at room temperature. DAB p53 mutation detection DNA was amplified in separate PCR reactions for all p53 (3,3'diaminobenzidine)-nickel or DAB chromagen (DAKO) was used to visualize antibody complexes. Sec- coding exons (2–11) and flanking regulatory regions. tions were counterstained with methyl green or hematox- Primer sequences and PCR conditions are available from ylin. Staining of inflammatory cells served as a positive the authors on request. PCR products were purified and internal control for all antibodies. sequenced using Big Dye Terminator chemistry (Perkin- Elmer, Boston, MA), and run on an ABI 3100 DNA In order to demonstrate the specificity of the BRCA2 H- sequencer (Applied Biosytstems, Foster City, CA). 300 antibody we tested our protocol on a primary tumor Sequencing data were analyzed using Sequencher soft- with a BRCA2 mutation that results in protein truncation ware (Gene Codes Corporation, Ann Arbor MI). Muta- prior to the epitope recognized by H-300. This carcinoma tions were confirmed with a separate sequencing reaction has no detectable wildtype sequence. We also evaluated and were compared to the UMP TP53 mutation database BRCA2 on its paired recurrence with a known genetic at http://p53.free.fr/. reversion and presumed normal expression of BRCA2[12]. Protein expression was absent in the primary Statistics and Survival carcinoma and strongly present in all recurrent neoplastic Two by two comparisons were evaluated with the Fisher's nuclei consistent with the DNA sequence information exact tests and larger contingency tables were evaluated (data not shown). with Chi Square. Two-tailed p values were generated with InStat (Graphpad Software, San Diego, CA). All women Page 9 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 ing agents, an effect that involves the JNK pathway. Oncogene received primary therapy containing a taxane and plati- 2001, 20:6597-6606. num agent. Overall survival was calculated from the date 7. Zhou C, Smith JL, Liu J: Role of BRCA1 in cellular resistance to of diagnosis until death or last follow-up. Survival curves paclitaxel and ionizing radiation in an ovarian cancer cell line carrying a defective BRCA1. Oncogene 2003, 22:2396-2404. were generated according to the Kaplan-Meier method 8. Boyd J, Sonoda Y, Federici MG, Bogomolniy F, Rhei E, Maresco DL, using Prism (Graphpad Software). Cox regression analysis Saigo PE, Almadrones LA, Barakat RR, Brown CL, et al.: Clinico- pathologic features of BRCA-linked and sporadic ovarian was performed using JMP8 (SAS, Cary, NC). Differences cancer. Jama 2000, 283:2260-2265. between survival curves were tested with the Log-rank 9. Cass I, Baldwin RL, Varkey T, Moslehi R, Narod SA, Karlan BY: method. Multi-variate analysis was performed using Cox Improved survival in women with BRCA-associated ovarian carcinoma. Cancer 2003, 97:2187-2195. proportional hazards modelling using JMP8. 10. Chetrit A, Hirsh-Yechezkel G, Ben-David Y, Lubin F, Friedman E, Sadetzki S: Effect of BRCA1/2 mutations on long-term survival Competing interests of patients with invasive ovarian cancer: the national Israeli study of ovarian cancer. J Clin Oncol 2008, 26:20-25. The authors declare that they have no competing interests. 11. Edwards SL, Brough R, Lord CJ, Natrajan R, Vatcheva R, Levine DA, Boyd J, Reis-Filho JS, Ashworth A: Resistance to therapy caused by intragenic deletion in BRCA2. Nature 2008, 451:1111-1115. Authors' contributions 12. Sakai W, Swisher EM, Karlan BY, Agarwal MK, Higgins J, Friedman C, EMS performed molecular assays and drafted the manu- Villegas E, Jacquemont C, Farrugia DJ, Couch FJ, et al.: Secondary script. RG performed protein expression assays. RG mutations as a mechanism of cisplatin resistance in BRCA2- mutated cancers. Nature 2008, 451:1116-1120. reviewed specimen pathology. PW co-wrote the manu- 13. Swisher EM, Sakai W, Karlan BY, Wurz K, Urban N, Taniguchi T: Sec- script and contributed to the study design. TW designed ondary BRCA1 mutations in BRCA1-mutated ovarian carci- nomas with platinum resistance. Cancer Res 2008, methylation assays. TT designed the FANCF assays and 68:2581-2586. edited the manuscript. BG contributed to sample acquisi- 14. Berchuck A, Heron KA, Carney ME, Lancaster JM, Fraser EG, Vinson tion and study design. All authors read and approved the VL, Deffenbaugh AM, Miron A, Marks JR, Futreal PA, Frank TS: Fre- quency of germline and somatic BRCA1 mutations in ovar- final manuscript. ian cancer. Clin Cancer Res 1998, 4:2433-2437. 15. Foster KA, Harrington P, Kerr J, Russell P, DiCioccio RA, Scott IV, Additional material Jacobs I, Chenevix-Trench G, Ponder BA, Gayther SA: Somatic and germline mutations of the BRCA2 gene in sporadic ovarian cancer. Cancer Res 1996, 56:3622-3625. 16. Takahashi H, Chiu HC, Bandera CA, Behbakht K, Liu PC, Couch FJ, Additional file 1 Weber BL, LiVolsi VA, Furusato M, Rebane BA, et al.: Mutations of Table S1. Clinical and molecular data for all paired primary and recur- the BRCA2 gene in ovarian carcinomas. Cancer Res 1996, rent carcinomas 56:2738-2741. 17. Takahashi H, Behbakht K, McGovern PE, Chiu HC, Couch FJ, Weber Click here for file BL, Friedman LS, King MC, Furusato M, LiVolsi VA, et al.: Mutation [http://www.biomedcentral.com/content/supplementary/1476- analysis of the BRCA1 gene in ovarian cancers. Cancer Res 4598-8-48-S1.xls] 1995, 55:2998-3002. 18. Strathdee G, MacKean MJ, Illand M, Brown R: A role for methyla- tion of the hMLH1 promoter in loss of hMLH1 expression and drug resistance in ovarian cancer. Oncogene 1999, 18:2335-2341. Acknowledgements 19. Willner J, Wurz K, Allison KH, Galic V, Garcia RL, Goff BA, Swisher EM: Alternate molecular genetic pathways in ovarian carci- This work was supported by KO8 CA96610-01 (EMS) and the Florence and nomas of common histological types. Hum Pathol 2007, Marshall Schwid Award from the Gynecologic Cancer Foundation (EMS). 4:607-13. We thank Mary-Claire King, PhD for guidance and input. 20. Catteau A, Harris WH, Xu CF, Solomon E: Methylation of the BRCA1 promoter region in sporadic breast and ovarian can- cer: correlation with disease characteristics. Oncogene 1999, References 18:1957-1965. 1. Howlett NG, Taniguchi T, Olson S, Cox B, Waisfisz Q, De Die- 21. Esteller M, Silva JM, Dominguez G, Bonilla F, Matias-Guiu X, Lerma E, Smulders C, Persky N, Grompe M, Joenje H, Pals G, et al.: Biallelic Bussaglia E, Prat J, Harkes IC, Repasky EA, et al.: Promoter hyper- inactivation of BRCA2 in Fanconi anemia. Science 2002, methylation and BRCA1 inactivation in sporadic breast and 297:606-609. ovarian tumors. J Natl Cancer Inst 2000, 92:564-569. 2. Wang W: Emergence of a DNA-damage response network 22. Baldwin RL, Nemeth E, Tran H, Shvartsman H, Cass I, Narod S, Karlan consisting of Fanconi anaemia and BRCA proteins. Nat Rev BY: BRCA1 promoter region hypermethylation in ovarian Genet 2007, 8:735-748. carcinoma: a population-based study. Cancer Res 2000, 3. Husain A, He G, Venkatraman ES, Spriggs DR: BRCA1 up-regula- 60:5329-5333. tion is associated with repair-mediated resistance to cis- 23. Taniguchi T, Tischkowitz M, Ameziane N, Hodgson SV, Mathew CG, diamminedichloroplatinum(II). Cancer Res 1998, 58:1120-1123. Joenje H, Mok SC, D'Andrea AD: Disruption of the Fanconi ane- 4. Quinn JE, Kennedy RD, Mullan PB, Gilmore PM, Carty M, Johnston mia-BRCA pathway in cisplatin-sensitive ovarian tumors. PG, Harkin DP: BRCA1 functions as a differential modulator of Nat Med 2003, 9:568-574. chemotherapy-induced apoptosis. Cancer Res 2003, 24. Strathdee G, Appleton K, Illand M, Millan DW, Sargent J, Paul J, Brown 63:6221-6228. R: Primary ovarian carcinomas display multiple methylator 5. Bhattacharyya A, Ear US, Koller BH, Weichselbaum RR, Bishop DK: phenotypes involving known tumor suppressor genes. Am J The breast cancer susceptibility gene BRCA1 is required for Pathol 2001, 158:1121-1127. subnuclear assembly of Rad51 and survival following treat- 25. Teodoridis JM, Hall J, Marsh S, Kannall HD, Smyth C, Curto J, Siddiqui ment with the DNA cross-linking agent cisplatin. J Biol Chem N, Gabra H, McLeod HL, Strathdee G, Brown R: CpG island meth- 2000, 275:23899-23903. ylation of DNA damage response genes in advanced ovarian 6. Lafarge S, Sylvain V, Ferrara M, Bignon YJ: Inhibition of BRCA1 cancer. Cancer Res 2005, 65:8961-8967. leads to increased chemoresistance to microtubule-interfer- Page 10 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 26. Gras E, Cortes J, Diez O, Alonso C, Matias-Guiu X, Baiget M, Prat J: cancer before and after platinum drug-based chemotherapy. Loss of heterozygosity on chromosome 13q12-q14, BRCA-2 Clin Cancer Res 2000, 6:1415-1421. mutations and lack of BRCA-2 promoter hypermethylation 46. Wilson CA, Ramos L, Villasenor MR, Anders KH, Press MF, Clarke K, in sporadic epithelial ovarian tumors. Cancer 2001, 92:787-795. Karlan B, Chen JJ, Scully R, Livingston D, et al.: Localization of 27. Hilton JL, Geisler JP, Rathe JA, Hattermann-Zogg MA, DeYoung B, human BRCA1 and its loss in high-grade, non-inherited Buller RE: Inactivation of BRCA1 and BRCA2 in ovarian can- breast carcinomas. Nat Genet 1999, 21:236-240. cer. J Natl Cancer Inst 2002, 94:1396-1406. 47. Thakur S, Zhang HB, Peng Y, Le H, Carroll B, Ward T, Yao J, Farid 28. Brown R, Hirst GL, Gallagher WM, McIlwrath AJ, Margison GP, Zee LM, Couch FJ, Wilson RB, Weber BL: Localization of BRCA1 and AG van der, Anthoney DA: hMLH1 expression and cellular a splice variant identifies the nuclear localization signal. Mol responses of ovarian tumour cells to treatment with cyto- Cell Biol 1997, 17:444-452. toxic anticancer agents. Oncogene 1997, 15:45-52. 48. Wilson CA, Payton MN, Elliott GS, Buaas FW, Cajulis EE, Grosshans 29. Thrall M, Gallion HH, Kryscio R, Kapali M, Armstrong DK, DeLoia JA: D, Ramos L, Reese DM, Slamon DJ, Calzone FJ: Differential subcel- BRCA1 expression in a large series of sporadic ovarian carci- lular localization, expression and biological toxicity of nomas: a Gynecologic Oncology Group study. Int J Gynecol Can- BRCA1 and the splice variant BRCA1-delta11b. Oncogene cer 2006, 16(Suppl 1):166-171. 1997, 14:1-16. 30. Wang C, Horiuchi A, Imai T, Ohira S, Itoh K, Nikaido T, Katsuyama 49. Spain BH, Larson CJ, Shihabuddin LS, Gage FH, Verma IM: Trun- Y, Konishi I: Expression of BRCA1 protein in benign, border- cated BRCA2 is cytoplasmic: implications for cancer-linked line, and malignant epithelial ovarian neoplasms and its rela- mutations. Proc Natl Acad Sci USA 1999, 96:13920-13925. tionship to methylation and allelic loss of the BRCA1 gene. J 50. Herman JG, Umar A, Polyak K, Graff JR, Ahuja N, Issa JP, Markowitz Pathol 2004, 202:215-223. S, Willson JK, Hamilton SR, Kinzler KW, et al.: Incidence and func- 31. Blackshear PE, Goldsworthy SM, Foley JF, McAllister KA, Bennett LM, tional consequences of hMLH1 promoter hypermethylation Collins NK, Bunch DO, Brown P, Wiseman RW, Davis BJ: Brca1 and in colorectal carcinoma. Proc Natl Acad Sci USA 1998, Brca2 expression patterns in mitotic and meiotic cells of 95:6870-6875. mice. Oncogene 1998, 16:61-68. 32. Rajan JV, Wang M, Marquis ST, Chodosh LA: Brca2 is coordinately regulated with Brca1 during proliferation and differentiation in mammary epithelial cells. Proc Natl Acad Sci USA 1996, 93:13078-13083. 33. Bernard-Gallon DJ, Peffault de Latour M, Hizel C, Vissac C, Cure H, Pezet D, Dechelotte PJ, Chipponi J, Chassagne J, Bignon YJ: Localiza- tion of human BRCA1 and BRCA2 in non-inherited colorec- tal carcinomas and matched normal mucosas. Anticancer Res 2001, 21:2011-2020. 34. Rajan JV, Marquis ST, Gardner HP, Chodosh LA: Developmental expression of Brca2 colocalizes with Brca1 and is associated with proliferation and differentiation in multiple tissues. Dev Biol 1997, 184:385-401. 35. Andres JL, Fan S, Turkel GJ, Wang JA, Twu NF, Yuan RQ, Lamszus K, Goldberg ID, Rosen EM: Regulation of BRCA1 and BRCA2 expression in human breast cancer cells by DNA-damaging agents. Oncogene 1998, 16:2229-2241. 36. Fan S, Twu NF, Wang JA, Yuan RQ, Andres J, Goldberg ID, Rosen EM: Down-regulation of BRCA1 and BRCA2 in human ovarian cancer cells exposed to adriamycin and ultraviolet radiation. Int J Cancer 1998, 77:600-609. 37. Spillman MA, Bowcock AM: BRCA1 and BRCA2 mRNA levels are coordinately elevated in human breast cancer cells in response to estrogen. Oncogene 1996, 13:1639-1645. 38. Chiang JW, Karlan BY, Cass L, Baldwin RL: BRCA1 promoter methylation predicts adverse ovarian cancer prognosis. Gynecol Oncol 2006, 101:403-410. 39. Rhei E, Bogomolniy F, Federici MG, Maresco DL, Offit K, Robson ME, Saigo PE, Boyd J: Molecular genetic characterization of BRCA1- and BRCA2-linked hereditary ovarian cancers. Can- cer Res 1998, 58:3193-3196. 40. Ramus SJ, Bobrow LG, Pharoah PD, Finnigan DS, Fishman A, Altaras M, Harrington PA, Gayther SA, Ponder BA, Friedman LS: Increased frequency of TP53 mutations in BRCA1 and BRCA2 ovarian tumours. Genes Chromosomes Cancer 1999, 25:91-96. 41. Crook T, Crossland S, Crompton MR, Osin P, Gusterson BA: p53 mutations in BRCA1-associated familial breast cancer. Lan- Publish with Bio Med Central and every cet 1997, 350:638-639. scientist can read your work free of charge 42. Greenblatt MS, Chappuis PO, Bond JP, Hamel N, Foulkes WD: TP53 mutations in breast cancer associated with BRCA1 or "BioMed Central will be the most significant development for BRCA2 germ-line mutations: distinctive spectrum and disseminating the results of biomedical researc h in our lifetime." structural distribution. Cancer Res 2001, 61:4092-4097. Sir Paul Nurse, Cancer Research UK 43. Gifford G, Paul J, Vasey PA, Kaye SB, Brown R: The acquisition of hMLH1 methylation in plasma DNA after chemotherapy Your research papers will be: predicts poor survival for ovarian cancer patients. Clin Cancer available free of charge to the entire biomedical community Res 2004, 10:4420-4426. 44. Fink D, Nebel S, Norris PS, Baergen RN, Wilczynski SP, Costa MJ, peer reviewed and published immediately upon acceptance Haas M, Cannistra SA, Howell SB: Enrichment for DNA mis- cited in PubMed and archived on PubMed Central match repair-deficient cells during treatment with cisplatin. Int J Cancer 1998, 77:741-746. yours — you keep the copyright 45. Samimi G, Fink D, Varki NM, Husain A, Hoskins WJ, Alberts DS, BioMedcentral Howell SB: Analysis of MLH1 and MSH2 expression in ovarian Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 11 of 11 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Molecular Cancer Springer Journals

Methylation and protein expression of DNA repair genes: association with chemotherapy exposure and survival in sporadic ovarian and peritoneal carcinomas

Loading next page...
 
/lp/springer-journals/methylation-and-protein-expression-of-dna-repair-genes-association-ePl2Ugikx0

References (52)

Publisher
Springer Journals
Copyright
Copyright © 2009 by Swisher et al; licensee BioMed Central Ltd.
Subject
Biomedicine; Cancer Research; Oncology
eISSN
1476-4598
DOI
10.1186/1476-4598-8-48
pmid
19602291
Publisher site
See Article on Publisher Site

Abstract

Background: DNA repair genes critically regulate the cellular response to chemotherapy and epigenetic regulation of these genes may be influenced by chemotherapy exposure. Restoration of BRCA1 and BRCA2 mediates resistance to platinum chemotherapy in recurrent BRCA1 and BRCA2 mutated hereditary ovarian carcinomas. We evaluated BRCA1, BRCA2, and MLH1 protein expression in 115 sporadic primary ovarian carcinomas, of which 31 had paired recurrent neoplasms collected after chemotherapy. Additionally, we assessed whether promoter methylation of BRCA1, MLH1 or FANCF influenced response to chemotherapy or explained alterations in protein expression after chemotherapy exposure. Results: Of 115 primary sporadic ovarian carcinomas, 39 (34%) had low BRCA1 protein and 49 (42%) had low BRCA2 expression. BRCA1 and BRCA2 protein expression were highly concordant (p < 0.0001). MLH1 protein loss occurred in 28/115 (24%) primary neoplasms. BRCA1 protein loss in primary neoplasms was associated with better survival (p = 0.02 Log Rank test) and remained significant after accounting for either stage or age in a multivariate model (p = 0.04, Cox proportional hazards). In paired specimens, BRCA1 protein expression increased in 13/21 (62%) and BRCA2 protein expression increased in 15/21 (71%) of recurrent carcinomas with low or intermediate protein in the paired primary. In contrast MLH1 expression was rarely decreased in recurrent carcinomas (1/33, 3%). Similar frequencies of MLH1, BRCA1, and FANCF promoter methylation occurred in primary carcinomas without previous chemotherapy, after neoadjuvant chemotherapy, or in recurrent neoplasms. Conclusion: Low BRCA1 expression in primary sporadic ovarian carcinoma is associated with prolonged survival. Recurrent ovarian carcinomas commonly have increased BRCA1 and/or BRCA2 protein expression post chemotherapy exposure which could mediate resistance to platinum based therapies. However, alterations in expression of these proteins after chemotherapy are not commonly mediated by promoter methylation, and other regulatory mechanisms are likely to contribute to these alterations. Page 1 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 acetylation could be a modifiable mechanism of sensitiv- Background Ovarian carcinoma is the most deadly gynecological ity or resistance to chemotherapy. Indeed, in vitro data malignancy and is the fifth leading cause of carcinoma indicate that epigenetic alterations in MLH1 and FANCF death in American women. Ovarian carcinomas are usu- modulate response to platinum agents in cell lines ally responsive to initial platinum based chemotherapy [23,28]. While loss of BRCA1, BRCA2, or FANCF confers regimens. However, even after a complete clinical sensitivity to platinum, loss of MLH1 confers resistance to response, most ovarian carcinomas do recur, with resist- platinum agents [28]. The importance of these mecha- ance to platinum therapy developing after one or more nisms in in vivo chemoresistance in ovarian carcinomas chemotherapy courses. Acquired chemotherapy resistance has not been defined. is one of the greatest clinical challenges in the treatment of women with ovarian carcinoma. We undertook this study to determine if protein expres- sion and/or promoter methylation of BRCA1, MLH1, Germline mutations in the BRCA1 and BRCA2 genes con- BRCA2 and FANCF predicted overall survival in primary fer inherited susceptibility to ovarian and breast carcino- sporadic ovarian carcinomas and whether promoter mas. Breast and ovarian carcinomas in BRCA1 or BRCA2 methylation or protein expression were altered by expo- mutation carriers usually have somatic deletions of the sure to chemotherapy in recurrent carcinomas. We also wildtype allele, rendering the neoplasm BRCA1 or BRCA2 wanted to determine the relationship of alterations of deficient. BRCA2 is identical to the Fanconi anemia (FA) proteins in the FA-BRCA pathway with alterations in p53, gene FANCD1[1]. BRCA1, BRCA2, other FA genes, and a another key sensor of DNA damage. larger number of protein partners (including the DNA mismatch repair protein MLH1) are part of a complex Results DNA damage response network (reviewed in[2]). In vitro BRCA1, MLH1 and BRCA2 protein expression in primary ovarian carcinomas studies indicate that BRCA1 and BRCA2 loss increases sensitivity to agents that cause double strand DNA breaks Protein expression for BRCA1, MLH1 and BRCA2 was and/or interstrand DNA cross-links including platinum assessed with immunohistochemistry in 115 sporadic pri- agents [3-5]. Conversely, loss of BRCA1 or BRCA2 may mary invasive neoplasms. In 31 cases, we had paired increase resistance to microtubule interfering agents such recurrent or persistent neoplasm tissue available after as taxanes and vincristine [6,7]. Most studies report chemotherapy which allowed for comparison of protein improved survival in women with ovarian carcinomas expression with the matched primary. Eight primary inva- associated with BRCA1 and BRCA2 mutations compared sive carcinomas were classified as peritoneal in origin and to women with sporadic ovarian carcinoma, consistent the remainder ovarian in origin (Table 1). Clinical and with increased sensitivity to platinum-based chemother- pathological characteristics and relationship to BRCA1 apy [8-10]. We and others have recently shown that carci- protein expression are summarized in Table 1. Peritoneal nomas from patients with inherited frameshift mutations carcinomas were less likely to have low BRCA1 protein in BRCA1 or BRCA2 exposed to chemotherapy can than ovarian carcinomas (p = 0.049, Fishers Exact, two- acquire secondary mutations that restore the reading tailed). There was no association between BRCA1 protein frame of BRCA1 or BRCA2, resulting in platinum resist- expression and grade, histology, or adequacy of surgical ance [11-13]. Thus, restoration of expression of proteins cytoreduction. Stage I carcinomas were more likely to in the FA-BRCA DNA damage response pathway may have normal BRCA1 protein expression compared to stage increase resistance to platinum and potentially other II-IV carcinomas (p = 0.03, Fisher exact, two-tailed). As agents that induce DNA damage. expected both stage I and II carcinomas were also more likely to have non serous and undifferentiated histologies While germline mutations in BRCA1 and BRCA2 predis- (8/11, 73% for stage I and 3/4 for stage II, p < 0.0001 com- pose to hereditary ovarian carcinoma, somatic mutations pared to stage III and IV, Fishers Exact, two-tailed). Stage I in these genes are rare in sporadic ovarian carcinomas [14- histologies included two serous, one undifferentiated car- 17]. However, epigenetic alterations in these and other cinoma, one small cell, two mucinous, five endometrioid, DNA repair genes may play important roles in sporadic while stage II included 3 endometrioid and one serous. ovarian and breast carcinomas and could contribute to However, normal BRCA1 protein expression was associ- responsiveness to chemotherapy. Promoter methylation ated with Stage I disease and not histology. Both serous leads to decreased expression of BRCA1, MLH1, and stage I carcinomas had normal BRCA1 protein expression FANCF protein in a subset of ovarian carcinomas [18-23]. and 2/3 of endometrioid stage II carcinomas had low BRCA1 promoter methylation occurs in 5–20% of spo- BRCA1 protein expression. radic ovarian carcinomas[20,22,24,25], while BRCA2 methylation is rare[26,27]. FANCF is a key regulator of the BRCA1, BRCA2, and MLH1 protein expression is summa- FA-BRCA pathway[23]. Epigenetic regulation of gene tran- rized in Table 2 and representative staining is shown in scription through promoter methylation or histone Figure 1. BRCA1 and BRCA2 protein levels were signifi- Page 2 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 Table 1: Clinicopathological characteristics of the sporadic neoplasms studied and BRCA1 protein expression. BRCA1 Protein Expression All Cases Low Intermediate Normal p value Primary site Ovarian 107 (93%) 39 27 41 p = 0.05 Peritoneal 8 (7%) 0 3 5 Grade Grade 1 6 (5%) 3 2 1 Grade 2 16 (16%) 5 3 8 NS Grade 3 93 (79%) 31 25 37 Histology Serous 81 (70%) 28 24 29 NS Endometrioid 11 (10%) 4 2 5 Carcinoma NOS 12 (10%) 3 3 6 MMMT 4 (3%) 1 1 2 Clear Cell 1 (1%) 0 0 1 Mucinous 3(3%) 1 0 2 Other* 3 (3%) 2 0 1 Stage I 11 (10%) 2 1 8 p = 0.03 II 4 (3%) 3 0 1 III 83 (72%) 28 26 29 IV 17 (15%) 6 3 8 Cytoreduction Optimal (< 1 cm) 73 24 18 31 NS Suboptimal 36 14 9 13 Not Available 6 1 3 2 Total 116 39 (36%) 30 (26%) 46 (38%) * Other histologies included one small cell, one transitional cell carcinoma and one carcinoma with focal giant cells. cantly and positively correlated (p < 0.0001, two-tailed). therapy influenced MLH1, BRCA1, or BRCA2 protein BRCA1 and BRCA2 protein expression were not related to levels. In 7 cases a second recurrence was available for MLH1 protein expression. expression analyses. For second recurrences, data was tab- ulated as if for a separate case when compared to the pri- BRCA1, MLH1 and BRCA2 protein expression in paired mary. Recurrent neoplasms were obtained at varying time primary and recurrent ovarian carcinomas intervals from last chemotherapy. We separated matched MLH1, BRCA1, and BRCA2 protein expression was also pairs into two groups depending on the interval since last assessed in 31 matched primary and recurrent ovarian car- chemotherapy exposure (≤ six months or >six months). cinomas from the same patient to determine if chemo- Complete data for protein and methylation analyses for each case is presented in Table S1; Additional file 1. Table 2: Protein expression in primary sporadic ovarian and Changes in protein expression in paired neoplasms are peritoneal carcinomas before chemotherapy exposure. shown schematically in Figure 2 and representative stain- ing in Figure 3. Protein Expression Low Intermediate Normal Total (≤10%) (10–30%) (>30%) Increased BRCA1 or BRCA2 protein would be expected to increase resistance to platinum therapy. In contrast, BRCA1 expression 39 (34%) 30 (26%) 46 (40%) 115 decreased MLH1 protein should be associated with plati- BRCA2 expression 49 (42%) 28 (24%) 38 (33%) 115 num resistance. In the paired sets most primary neo- MLH1 Expression* 28 (24%) 87 (76%)* 115 plasms (26/31, 84%) had normal MLH1 protein. Of the 33 recurrences occurring in the 26 cases with normal MLH1 expression was scored in two categories: ≤10% positive cells = MLH1 protein in the primary, only one had loss of MLH1 Low, >10% positivity = Normal. Page 3 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 or intermediate BRCA1 expression showed reduced BRCA1 expression in the recurrence (p = 0.01). Therefore, increases in BRCA1 expression were more common than decreases in BRCA1 expression following chemotherapy. BRCA2 protein expression also frequently varied between paired primary and recurrent neoplasms (Figures 2 and 3, Additional file 1). Among primary neoplasms with low or intermediate BRCA2 expression, the paired post-chemo- therapy specimen had increased BRCA2 expression in 15/ 21 cases (71%). In contrast to the observation for BRCA1 protein expression, primary neoplasms with normal or intermediate BRCA2 expression were equally likely to have decreased BRCA2 expression (14/23, 61%). Excluding those cases with normal BRCA1 and BRCA2 protein expression in the primary, 20 (80%) recurrent car- cinomas had increases in either BRCA1 or BRCA2 protein while five (20%) did not have an increase in expression of either protein. Likelihood of a complete response to sub- sequent chemotherapy was not related to whether increased BRCA1 or BRCA2 protein was identified in the paired recurrence, but wide heterogeneity in treatment Repr MLH1 in s Figure 1 esentativ pore ad pr ic ovarian carcino otein expression of BRCA1, mas BRCA2 and precludes assessment of response to specific agents. Representative protein expression of BRCA1, BRCA2 and MLH1 in sporadic ovarian carcinomas. BRCA1, MLH1, and FANCF promoter methylation in Protein expression is represented by brown stain. Black bars primary and recurrent ovarian carcinomas in the lower left corners represent 10 microns. A. BRCA1 Methylation was assessed for MLH1, BRCA1 and FANCF protein in a neoplasm with low expression. B. BRCA1 pro- in 104 primary and 36 recurrent sporadic carcinomas. Of tein in a neoplasm with normal expression. C. BRCA2 pro- the primary carcinomas tested for methylation, 11 were tein in a neoplasm with low expression. D. BRCA2 protein in obtained after neoadjuvant chemotherapy and 93 were a neoplasm with normal expression. E. MLH1 protein in a chemotherapy-naive. Results of methylation analysis for neoplasm with low expression. F. MLH1 protein in a neo- BRCA1, FANCF and MLH1 are summarized in Table 3. In plasm with normal expression. no cases were neoplasms methylated at more than one of these genes. Methylation did not vary between primary protein (3.0%, Figures 2 and 3, and Table S1 in Additional cases exposed to neoadjuvant chemotherapy or in primary file 1). Neither the primary nor recurrent neoplasm in this vs. recurrent cases for any gene (Table 3). Ten of the pri- pair had MLH1 promoter methylation. mary and recurrent neoplasms were matched specimens from the same patient collected at two different time In contrast to MLH1, BRCA1 and BRCA2 expression dem- points (Table S1; Additional file 1). All ten cases showed onstrated greater variability between paired primary and concordant methylation between primary and recurrence, recurrent neoplasms (Figure 2, Table S1;Additional file 1). and none were methylated at any of the three genes. In 16 carcinomas recurring after 13 primaries with normal Reduction of BRCA1 protein was significantly associated BRCA1 expression, 13(81%) maintained similar protein with BRCA1 methylation (p = 0.02). However, BRCA1 expression and three (19%) demonstrated decreased methylation only accounted for 7/52 (13.5%) of those BRCA1 expression in the recurrence. In 13 recurrences sporadic carcinomas with low or intermediate BRCA1 that followed 11 paired primaries with intermediate expression. BRCA1 expression, four had reduced, seven had increased and two had similar BRCA1 expression in the paired All cases with MLH1 methylation demonstrated microsat- recurrence. In eight recurrences that followed six primary ellite instability at BAT26 (data not shown). MLH1 meth- carcinomas with low BRCA1 expression, 2 had similar ylation was associated with loss of protein (p = 0.03), but and six (75%) had increased BRCA1 expression in the one methylated case did have apparently normal MLH1 recurrence. Overall 13 of 21 (62%) neoplastic pairs with protein expression. MLH1 methylation was more com- low or intermediate BRCA1 protein expression in the pri- mon in endometrioid ovarian carcinomas compared to all mary had increased BRCA1 protein in the recurrence. In other histologies (3/14, 21% versus 1/116, 1%, p = contrast, only eight of 33 (24%) neoplasms with normal Page 4 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 Schematic of BRCA1 Figure 2 , BRCA2, and MLH1 protein expression in paired primary and recurrent neoplasms Schematic of BRCA1, BRCA2, and MLH1 protein expression in paired primary and recurrent neoplasms. Each neoplasm is represented by a single horizontal line. A. BRCA1 protein expression in 24 primary and paired recurrent neo- plasms obtained ≤ 6 months since last chemotherapy, 8 with second recurrences. B. BRCA1 expression in 7 primary and paired recurrent neoplasms in which the recurrence was obtained more than 6 months since last chemotherapy. C. BRCA2 protein expression in 24 primary and paired recurrent neoplasms obtained ≤ 6 months since last chemotherapy. D. BRCA2 expression in 7 primary and paired recurrent neoplasm obtained more than 6 months since last chemotherapy. E. MLH1 pro- tein expression in 24 primary and paired recurrent neoplasms obtained ≤ 6 months since last chemotherapy. F. MLH1 expres- sion in 7 primary and paired recurrent neoplasm obtained more than 6 months since last chemotherapy. Page 5 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 nor FANCF methylation were associated with p53 muta- tion status. p53 mutations were more common in spo- radic ovarian carcinomas with loss or reduction of BRCA1 expression (Table 4, p = 0.01 Fishers Exact). The majority of p53 mutations were missense mutations occurring in the DNA binding domain (exons 5–8). Null mutations including frameshift, splice site, and nonsense mutations accounted for 9/33 (27%) of p53 mutations. Overall survival and protein expression or promoter methylation Individual factors influencing survival in this cohort in univariate analyses were stage (p = 0.01), age (p = 0.03) and optimal cytoreduction to a maximum neoplastic diameter less than 1 cm (p < 0.0001) and low BRCA1 expression (p = 0.02). Low BRCA1 expression in the pri- mary carcinoma was associated with longer survival com- pared to intermediate or normal BRCA1 expression (median survival 62 months vs. 45 months, p = 0.02 LogRank Test, Hazard Ratio 0.59, 95% confidence interval 0.37–0.93, Figure 4). Low BRCA1 expression remained significantly associated with improved survival in a Cox multi regression model with the covariates age (p = 0.04) or stage (p = 0.04) but was no longer significant when BRCA1, BRCA2, and MLH m Figure 3 ary and recurrent neopla1 sms protein expression in paired pri- using the co-variate optimal cytoreduction (p = 0.10). BRCA1, BRCA2, and MLH1 protein expression in Individual factors not related to survival included grade, paired primary and recurrent neoplasms. Protein MLH1 or BRCA2 protein expression, methylation of expression is represented by brown stain. Black bars in the BRCA1, MLH1 or FANCF genes, or p53 mutation. lower left corners are equal to 10 microns. A. BRCA1 expression is low in the primary neoplasm. B. BRCA1 expression is increased in the paired recurrent neoplasm. C. Response to chemotherapy BRCA2 protein expression is low in a different primary neo- The likelihood of a complete response to initial chemo- plasm. D. BRCA2 expression is increased in the paired recur- therapy was not related to BRCA1 protein levels. Likewise, rent neoplasm. E. MLH1 protein is normal in a different there was no significant difference in the complete primary neoplasm. F. In the paired recurrent neoplasm, response rate for cases with BRCA1 methylation, MLH1 MLH1 protein expression is reduced. methylation, or for cases with either BRCA1 or FANCF methylation. 0.003). FANCF and BRCA1 methylation were not associ- ated with histology. Discussion BRCA1 protein loss is common in sporadic epithelial Relationship of p53 mutations to protein expression and ovarian and peritoneal carcinomas. Low BRCA1 protein methylation of DNA repair genes expression was associated with a significantly improved p53 mutations were assessed by DNA sequencing in all overall survival. Our data suggest that somatic loss of cases tested for methylation and neither BRCA1, MLH1, BRCA1 favourably influences survival similar to the Table 3: Methylation of DNA repair genes in primary and recurrent ovarian or peritoneal carcinomas. PROPORTION METHYLATED BRCA1 MLH1 FANCF Primary, no chemotherapy 6/91 (6.6%) 3/93 (3.2%) 3/93 (3.2%) Primary, post neoadjuvant chemo 2/11 (18%) 1/11 (9.1%) 0/10 Recurrent* 2/31 (6.5%) 0/30 0/31 (0%) Total (2.2%) 10/133 (7.5%) 4/134 (3.0%) 3/134 * Ten recurrent carcinomas were paired specimens for which we also evaluated the primary neoplasm. None of these 10 paired cases were methylated at any of the three genes in either the primary or recurrent case. One of these 10 pairs did show loss of MLH1 protein in the recurrent neoplasm, but was not methylated. Page 6 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 Table 4: p53 mutations and BRCA1 protein level in primary sporadic ovarian carcinomas P53 Mutations BRCA1 protein Expression Total Tested Missense Null Total P53 mutations N (%) N (%) Low 28 10 (36%) 5 (18%) 15 (54%) Intermediate 18 8 (44%) 2 (11%) 10 (56%) Normal 33 6(18%) 2 (6%) 8 (24%)* Total 79 24 (30%) 9 (11%) 33 (42%) *p = 0.01 for difference in p53 mutation frequency for cases with normal versus reduction or loss of BRCA1 protein. improved survival in ovarian carcinomas associated with To our knowledge, this is the first study of BRCA2 protein inherited BRCA1 mutations. Two previous studies have expression in human ovarian carcinomas. Unlike BRCA1, examined the relationship between BRCA1 protein BRCA2 protein expression was not associated with overall expression and prognosis in ovarian carcinomas. Thrall survival. Our data demonstrate that BRCA1 and BRCA2 and colleagues evaluated a large number of sporadic protein expression are highly concordant in ovarian carci- advanced stage carcinomas and found that BRCA1 loss nomas. In both mouse and human, BRCA1 and BRCA2 was strongly protective for overall survival [29]. In have nearly identical patterns of message and protein another study of un-selected ovarian carcinoma, Wang expression in embryonic development, breast morpho- and colleagues failed to find an association between genesis, and in many adult tissues, suggesting that the two BRCA1 protein and prognosis [30]. However, that series genes share regulatory networks [31-34]. Similarly, in evaluated a large number of endometrioid, clear cell, breast and ovarian carcinoma cell lines, BRCA1 and grade 1 and stage I cases, with few high-grade, advanced BRCA2 mRNA are concordantly induced by adriamycin, stage, serous carcinomas, distinctly different from the ionizing radiation, and estrogen [35-37]. Our data suggest neoplasms in our study. Indeed, among the 29 stage III that ovarian carcinomas maintain the coordinate regula- and IV cases in Wang's study, there was a trend toward a tion of BRCA1 and BRCA2 seen in normal tissues. survival advantage with loss of BRCA1 protein [30]. Our data combined with these two previous studies confirm We hypothesized that BRCA1 and BRCA2 protein levels an improved prognosis for women with advanced ovarian would increase after exposure to chemotherapy, thereby carcinomas with low BRCA1 protein expression. mediating increasing platinum resistance during the dis- ease course. Since recurrent ovarian carcinomas are not routinely subjected to biopsy, our cases represent a wide variety of disease time points depending on the clinical indication for biopsy or surgery in a given patient. Conse- quently, our patients received varied amounts and types of chemotherapy before and after biopsies. An increase in BRCA1 or BRCA2 protein occurred in the majority of recurrent neoplasms that had low or intermediate protein in the paired primary. Increases in BRCA1 protein were significantly more likely than reduction of protein in paired post chemotherapy neoplasms, suggesting a selec- tion for increased BRCA1 expression. In contrast, BRCA2 protein both increased and decreased frequently post chemotherapy. An increasing tendency toward platinum resistance during the typical ovarian carcinoma disease course is a well recognized clinical challenge. Increased Overall survival in sporadi Figure 4 c ovarian carcinomas relation to BRCA1 expression in primary BRCA1 and BRCA2 expression could mediate that resist- Overall survival in relation to BRCA1 expression in ance. However, given the treatment heterogeneity of our primary sporadic ovarian carcinomas. Overall survival cases, we cannot directly relate the alterations in BRCA1 or was significantly improved in primary ovarian carcinomas (p BRCA2 expression with specific treatment responses. = 0.02, LogRank test) with low BRCA1 protein expression (median survival 62 months) compared to carcinomas with Our data are consistent with previous reports that BRCA1 intermediate or normal BRCA1 expresssion (median survival 45 months). promoter methylation occurs in 5–20% of sporadic ovar- ian carcinomas [20-22,24,25]. Interestingly, we found Page 7 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 that promoter methylation only occurs in 15% of those with treatment response, opposite to the expected associ- sporadic ovarian carcinomas with low or intermediate ation [44]. We used a traditional cut-off to define MLH1 BRCA1 expression. Thus, other regulators of BRCA1 protein deficiency at 10% of cells while the other two expression may be more important than promoter meth- studies used a continuous scoring system. However, even ylation in ovarian carcinomas for both primary and recur- when we re-evaluated our data with a continuous scoring rent neoplasms. We found no association between BRCA1 system, we still found that significant decrease in MLH1 methylation and survival in sporadic ovarian carcinomas, protein expression was rare (data not shown). The dis- in contrast to a recent small study that demonstrated a sur- crepancies between these studies and ours may stem from vival disadvantage for patients whose neoplasms were the clinical heterogeneity of cases for which paired pri- methylated at BRCA1 [38]. Our data suggest that BRCA1 mary and recurrent ovarian carcinomas are available. protein expression is a better predictor of survival than Finally, we did not find MLH1 promoter methylation BRCA1 promoter methylation, not surprising given the more frequently in either recurrent neoplasms or in pri- relatively small number of cases with BRCA1 methylation. mary neoplasms exposed to neoadjuvant chemotherapy, nor was MLH1 methylation identified in the neoplastic Low BRCA1 protein expression was associated with an pair with reduced MLH1 expression in the recurrence. increased likelihood of p53 mutation, low BRCA2 protein Consequently, our data do not support a major role for in expression, and improved survival. Only one previous vivo epigenetic alteration of MLH1 expression in the devel- study of which we are aware has examined the association opment of clinical platinum resistance in sporadic ovar- of somatic BRCA1 alterations and p53 status. In that ian carcinomas. study, BRCA1 methylation did not correlate with immu- nohistochemically detected p53 protein expression [22]. FANCF may also mediate platinum sensitivity. FANCF However, p53 immunohistochemistry correlates imper- promoter methylation confers cisplatin sensitivity to cell fectly with the presence of a somatic mutation, and lines and was hypothesized to be a mediator of in vivo BRCA1 methylation accounts for only a small fraction of platinum sensitivity[23]. In order to mediate resistance, a BRCA1 protein loss, so our studies are not equivalent. The neoplasm should be initially methylated and then lose majority of breast, ovarian, and peritoneal carcinomas methylation in recurrent disease. However, our findings associated with germline mutations in BRCA1 and BRCA2 indicate that FANCF methylation occurs rarely in primary have p53 mutations, at rates higher than found in their ovarian carcinomas (<5% of cases), and is therefore not sporadic counterparts [39-42]. Thus, many sporadic ovar- likely to be a major clinical mediator of platinum sensitiv- ian carcinomas have reduction of BRCA1 protein, muta- ity. tions in p53, and better overall survival, mimicking the phenotype of hereditary BRCA1-associated ovarian carci- Conclusion noma. Low BRCA1 protein expression in sporadic ovarian carci- nomas is associated with a favourable survival and an MLH1 is a DNA mismatch repair gene required for sensi- increased rate of p53 mutations. BRCA1 and BRCA2 pro- tivity to platinum compounds. MHL1 is commonly meth- teins are concordantly expressed in sporadic ovarian and ylated in colorectal and endometrial carcinomas and less peritoneal carcinomas. Increases in BRCA1 and BRCA2 commonly in ovarian carcinomas [19,24]. MLH1 methyl- protein expression are common in recurrent sporadic ation mediates platinum resistance in the ovarian carci- ovarian carcinomas, but the mechanism(s) responsible noma cell line A2780 [18]. We postulated that for these expression differences are unknown. Alteration chemotherapy exposure could select for neoplastic cells in methylation of the promoters of BRCA1, MLH1, and with MLH1 methylation. Indeed, cell-free methylated FANCF do not seem to commonly mediate clinical chem- MLH1 DNA is increased in the plasma of women with oresistance in women with ovarian carcinomas. ovarian carcinoma at the time of clinical relapse [43]. Increased methylation of MLH1 in recurrent ovarian car- Methods cinoma has been proposed as a rationale for clinical trials Specimens of demethylating agents alone or in combination with Patients consented to have tissue collected and to provide chemotherapy. We found no association between MLH1 blood samples using protocols approved by the Human protein and overall survival or response to chemotherapy, Subjects Division of the University of Washington Institu- and we rarely observed (3.8%) loss of MLH1 protein in tional Review Board. Paraffin blocks were obtained from paired neoplasms following chemotherapy exposure. Our pathology archives. Tissues were obtained at the time of data contrast with those of two groups who both reported surgery and snap frozen in liquid nitrogen. Tissue sections a significant decrease in MLH1 protein expression in were stained with hematoxylin and eosin and reviewed paired ovarian carcinomas following platinum chemo- for neoplastic cell percentage prior to DNA extraction. therapy [44,45]. But surprisingly, Fink and colleagues DNA was extracted using the Stratagene kit according to found that MLH1 protein expression correlated inversely the manufacturer's instructions. Normal DNA was Page 8 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 extracted from peripheral white blood cells or from nor- Controversy exists regarding the meaning and specificity mal tissue after histological review excluded contamina- of cytoplasmic staining for BRCA1 and BRCA2. Proteins tion with neoplasm. Family history was considered strong encoded by aberrantly spliced isoforms have been shown if the patient had a known BRCA1 or BRCA2 mutation, to localize to the cytoplasm for both BRCA1 and BRCA2 personal history of breast carcinoma, any relative with [47-49]. The MS110 antibody has better specificity than ovarian carcinoma, a first degree relative with premeno- other anti-BRCA1 antibodies and has an almost exclu- pausal breast carcinoma, or two relatives of any degree sively nuclear staining pattern[46]. We observed exclu- with breast carcinoma at any age. Cases without strong sively nuclear BRCA1 protein staining in both normal and family histories were considered sporadic, while those neoplastic tissues. We observed both nuclear and cyto- with suggestive family histories were considered familial plasmic BRCA2 staining in neoplastic cells, but scored and excluded. only nuclear staining as likely to be relevant to intact DNA repair. BRCA1 and BRCA2 protein was scored as previ- Immunohistochemical staining for MLH1, BRCA1 and ously described: "low" if fewer than 10% of neoplastic BRCA2 cells had nuclear staining and "intermediate" if 10% to BRCA1 protein was detected in formalin fixed paraffin 30% of neoplastic cells had nuclear staining and normal sections using the mouse monoclonal antibody MS110 if greater than 30% of neoplastic cells had nuclear staining (previously called Ab-1, Oncogene Research Products) as [46]. MLH1 protein expression was scored as loss (<10% previously described [46]. The MS110 antibody recog- of cells positive) or normal (>10% of cells positive) as is nizes an amino terminal epitope in BRCA1 (amino acid done routinely for clinical purposes. MLH1 staining was residues 89–222). BRCA2 was detected using the H-300 exclusively nuclear. For the purpose of comparing protein rabbit polyclonal antibody (Santa Cruz Biotech) that rec- expression in paired primary and recurrent neoplasms, ognizes amino acid residues 2520–2819 of human the second recurrence in those seven cases with more than BRCA2. MLH1 was detected using the mouse monoclonal one recurrence was compared with the primary neo- antibody G168-728 (BD Pharmingen). Briefly, paraffin plasms and counted as a separate case. sections were deparaffinized, re-hydrated, and treated with steam heat for 20 minutes using antigen target Methylation specific PCR Carcinomas were evaluated for BRCA1, MLH1 and FANCF retrieval solution (DAKO). Endogenous peroxidase activ- ity was quenched by treatment with 3% H O for 5 min- promoter methylation as previously described [21,23,50]. 2 2 utes. Sections were washed with PBS and non-specific All methylation specific PCR assays included a positive binding was blocked by treatment for three hours in 2% control (in vitro methylated DNA, Invitrogen) and a neg- bovine serum albumin in PBS. Primary antibody was ative (water) control. Positive assays were repeated at least diluted in PBS (MS110 at 1:250 dilution, H-300 at 1:100 once. Sodium bisulfite treated DNA was amplified with dilution, and G168-728 at 1:80) and applied to sections primers specific for either methylated or unmethylated for 14–16 hours at 4° (MS110 and H-300) or 40 minutes treated substrate. PCR products were electrophoresed on at room temperature (G168-728). Secondary antibody 3% agarose gels or 5% acrylamide gels and stained with and streptavidin biotin-peroxidase were from Universal ethidium bromide. Large Volume LSAB+, Peroxidase kit (DAKO) and were each applied for 30 minutes at room temperature. DAB p53 mutation detection DNA was amplified in separate PCR reactions for all p53 (3,3'diaminobenzidine)-nickel or DAB chromagen (DAKO) was used to visualize antibody complexes. Sec- coding exons (2–11) and flanking regulatory regions. tions were counterstained with methyl green or hematox- Primer sequences and PCR conditions are available from ylin. Staining of inflammatory cells served as a positive the authors on request. PCR products were purified and internal control for all antibodies. sequenced using Big Dye Terminator chemistry (Perkin- Elmer, Boston, MA), and run on an ABI 3100 DNA In order to demonstrate the specificity of the BRCA2 H- sequencer (Applied Biosytstems, Foster City, CA). 300 antibody we tested our protocol on a primary tumor Sequencing data were analyzed using Sequencher soft- with a BRCA2 mutation that results in protein truncation ware (Gene Codes Corporation, Ann Arbor MI). Muta- prior to the epitope recognized by H-300. This carcinoma tions were confirmed with a separate sequencing reaction has no detectable wildtype sequence. We also evaluated and were compared to the UMP TP53 mutation database BRCA2 on its paired recurrence with a known genetic at http://p53.free.fr/. reversion and presumed normal expression of BRCA2[12]. Protein expression was absent in the primary Statistics and Survival carcinoma and strongly present in all recurrent neoplastic Two by two comparisons were evaluated with the Fisher's nuclei consistent with the DNA sequence information exact tests and larger contingency tables were evaluated (data not shown). with Chi Square. Two-tailed p values were generated with InStat (Graphpad Software, San Diego, CA). All women Page 9 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 ing agents, an effect that involves the JNK pathway. Oncogene received primary therapy containing a taxane and plati- 2001, 20:6597-6606. num agent. Overall survival was calculated from the date 7. Zhou C, Smith JL, Liu J: Role of BRCA1 in cellular resistance to of diagnosis until death or last follow-up. Survival curves paclitaxel and ionizing radiation in an ovarian cancer cell line carrying a defective BRCA1. Oncogene 2003, 22:2396-2404. were generated according to the Kaplan-Meier method 8. Boyd J, Sonoda Y, Federici MG, Bogomolniy F, Rhei E, Maresco DL, using Prism (Graphpad Software). Cox regression analysis Saigo PE, Almadrones LA, Barakat RR, Brown CL, et al.: Clinico- pathologic features of BRCA-linked and sporadic ovarian was performed using JMP8 (SAS, Cary, NC). Differences cancer. Jama 2000, 283:2260-2265. between survival curves were tested with the Log-rank 9. Cass I, Baldwin RL, Varkey T, Moslehi R, Narod SA, Karlan BY: method. Multi-variate analysis was performed using Cox Improved survival in women with BRCA-associated ovarian carcinoma. Cancer 2003, 97:2187-2195. proportional hazards modelling using JMP8. 10. Chetrit A, Hirsh-Yechezkel G, Ben-David Y, Lubin F, Friedman E, Sadetzki S: Effect of BRCA1/2 mutations on long-term survival Competing interests of patients with invasive ovarian cancer: the national Israeli study of ovarian cancer. J Clin Oncol 2008, 26:20-25. The authors declare that they have no competing interests. 11. Edwards SL, Brough R, Lord CJ, Natrajan R, Vatcheva R, Levine DA, Boyd J, Reis-Filho JS, Ashworth A: Resistance to therapy caused by intragenic deletion in BRCA2. Nature 2008, 451:1111-1115. Authors' contributions 12. Sakai W, Swisher EM, Karlan BY, Agarwal MK, Higgins J, Friedman C, EMS performed molecular assays and drafted the manu- Villegas E, Jacquemont C, Farrugia DJ, Couch FJ, et al.: Secondary script. RG performed protein expression assays. RG mutations as a mechanism of cisplatin resistance in BRCA2- mutated cancers. Nature 2008, 451:1116-1120. reviewed specimen pathology. PW co-wrote the manu- 13. Swisher EM, Sakai W, Karlan BY, Wurz K, Urban N, Taniguchi T: Sec- script and contributed to the study design. TW designed ondary BRCA1 mutations in BRCA1-mutated ovarian carci- nomas with platinum resistance. Cancer Res 2008, methylation assays. TT designed the FANCF assays and 68:2581-2586. edited the manuscript. BG contributed to sample acquisi- 14. Berchuck A, Heron KA, Carney ME, Lancaster JM, Fraser EG, Vinson tion and study design. All authors read and approved the VL, Deffenbaugh AM, Miron A, Marks JR, Futreal PA, Frank TS: Fre- quency of germline and somatic BRCA1 mutations in ovar- final manuscript. ian cancer. Clin Cancer Res 1998, 4:2433-2437. 15. Foster KA, Harrington P, Kerr J, Russell P, DiCioccio RA, Scott IV, Additional material Jacobs I, Chenevix-Trench G, Ponder BA, Gayther SA: Somatic and germline mutations of the BRCA2 gene in sporadic ovarian cancer. Cancer Res 1996, 56:3622-3625. 16. Takahashi H, Chiu HC, Bandera CA, Behbakht K, Liu PC, Couch FJ, Additional file 1 Weber BL, LiVolsi VA, Furusato M, Rebane BA, et al.: Mutations of Table S1. Clinical and molecular data for all paired primary and recur- the BRCA2 gene in ovarian carcinomas. Cancer Res 1996, rent carcinomas 56:2738-2741. 17. Takahashi H, Behbakht K, McGovern PE, Chiu HC, Couch FJ, Weber Click here for file BL, Friedman LS, King MC, Furusato M, LiVolsi VA, et al.: Mutation [http://www.biomedcentral.com/content/supplementary/1476- analysis of the BRCA1 gene in ovarian cancers. Cancer Res 4598-8-48-S1.xls] 1995, 55:2998-3002. 18. Strathdee G, MacKean MJ, Illand M, Brown R: A role for methyla- tion of the hMLH1 promoter in loss of hMLH1 expression and drug resistance in ovarian cancer. Oncogene 1999, 18:2335-2341. Acknowledgements 19. Willner J, Wurz K, Allison KH, Galic V, Garcia RL, Goff BA, Swisher EM: Alternate molecular genetic pathways in ovarian carci- This work was supported by KO8 CA96610-01 (EMS) and the Florence and nomas of common histological types. Hum Pathol 2007, Marshall Schwid Award from the Gynecologic Cancer Foundation (EMS). 4:607-13. We thank Mary-Claire King, PhD for guidance and input. 20. Catteau A, Harris WH, Xu CF, Solomon E: Methylation of the BRCA1 promoter region in sporadic breast and ovarian can- cer: correlation with disease characteristics. Oncogene 1999, References 18:1957-1965. 1. Howlett NG, Taniguchi T, Olson S, Cox B, Waisfisz Q, De Die- 21. Esteller M, Silva JM, Dominguez G, Bonilla F, Matias-Guiu X, Lerma E, Smulders C, Persky N, Grompe M, Joenje H, Pals G, et al.: Biallelic Bussaglia E, Prat J, Harkes IC, Repasky EA, et al.: Promoter hyper- inactivation of BRCA2 in Fanconi anemia. Science 2002, methylation and BRCA1 inactivation in sporadic breast and 297:606-609. ovarian tumors. J Natl Cancer Inst 2000, 92:564-569. 2. Wang W: Emergence of a DNA-damage response network 22. Baldwin RL, Nemeth E, Tran H, Shvartsman H, Cass I, Narod S, Karlan consisting of Fanconi anaemia and BRCA proteins. Nat Rev BY: BRCA1 promoter region hypermethylation in ovarian Genet 2007, 8:735-748. carcinoma: a population-based study. Cancer Res 2000, 3. Husain A, He G, Venkatraman ES, Spriggs DR: BRCA1 up-regula- 60:5329-5333. tion is associated with repair-mediated resistance to cis- 23. Taniguchi T, Tischkowitz M, Ameziane N, Hodgson SV, Mathew CG, diamminedichloroplatinum(II). Cancer Res 1998, 58:1120-1123. Joenje H, Mok SC, D'Andrea AD: Disruption of the Fanconi ane- 4. Quinn JE, Kennedy RD, Mullan PB, Gilmore PM, Carty M, Johnston mia-BRCA pathway in cisplatin-sensitive ovarian tumors. PG, Harkin DP: BRCA1 functions as a differential modulator of Nat Med 2003, 9:568-574. chemotherapy-induced apoptosis. Cancer Res 2003, 24. Strathdee G, Appleton K, Illand M, Millan DW, Sargent J, Paul J, Brown 63:6221-6228. R: Primary ovarian carcinomas display multiple methylator 5. Bhattacharyya A, Ear US, Koller BH, Weichselbaum RR, Bishop DK: phenotypes involving known tumor suppressor genes. Am J The breast cancer susceptibility gene BRCA1 is required for Pathol 2001, 158:1121-1127. subnuclear assembly of Rad51 and survival following treat- 25. Teodoridis JM, Hall J, Marsh S, Kannall HD, Smyth C, Curto J, Siddiqui ment with the DNA cross-linking agent cisplatin. J Biol Chem N, Gabra H, McLeod HL, Strathdee G, Brown R: CpG island meth- 2000, 275:23899-23903. ylation of DNA damage response genes in advanced ovarian 6. Lafarge S, Sylvain V, Ferrara M, Bignon YJ: Inhibition of BRCA1 cancer. Cancer Res 2005, 65:8961-8967. leads to increased chemoresistance to microtubule-interfer- Page 10 of 11 (page number not for citation purposes) Molecular Cancer 2009, 8:48 http://www.molecular-cancer.com/content/8/1/48 26. Gras E, Cortes J, Diez O, Alonso C, Matias-Guiu X, Baiget M, Prat J: cancer before and after platinum drug-based chemotherapy. Loss of heterozygosity on chromosome 13q12-q14, BRCA-2 Clin Cancer Res 2000, 6:1415-1421. mutations and lack of BRCA-2 promoter hypermethylation 46. Wilson CA, Ramos L, Villasenor MR, Anders KH, Press MF, Clarke K, in sporadic epithelial ovarian tumors. Cancer 2001, 92:787-795. Karlan B, Chen JJ, Scully R, Livingston D, et al.: Localization of 27. Hilton JL, Geisler JP, Rathe JA, Hattermann-Zogg MA, DeYoung B, human BRCA1 and its loss in high-grade, non-inherited Buller RE: Inactivation of BRCA1 and BRCA2 in ovarian can- breast carcinomas. Nat Genet 1999, 21:236-240. cer. J Natl Cancer Inst 2002, 94:1396-1406. 47. Thakur S, Zhang HB, Peng Y, Le H, Carroll B, Ward T, Yao J, Farid 28. Brown R, Hirst GL, Gallagher WM, McIlwrath AJ, Margison GP, Zee LM, Couch FJ, Wilson RB, Weber BL: Localization of BRCA1 and AG van der, Anthoney DA: hMLH1 expression and cellular a splice variant identifies the nuclear localization signal. Mol responses of ovarian tumour cells to treatment with cyto- Cell Biol 1997, 17:444-452. toxic anticancer agents. Oncogene 1997, 15:45-52. 48. Wilson CA, Payton MN, Elliott GS, Buaas FW, Cajulis EE, Grosshans 29. Thrall M, Gallion HH, Kryscio R, Kapali M, Armstrong DK, DeLoia JA: D, Ramos L, Reese DM, Slamon DJ, Calzone FJ: Differential subcel- BRCA1 expression in a large series of sporadic ovarian carci- lular localization, expression and biological toxicity of nomas: a Gynecologic Oncology Group study. Int J Gynecol Can- BRCA1 and the splice variant BRCA1-delta11b. Oncogene cer 2006, 16(Suppl 1):166-171. 1997, 14:1-16. 30. Wang C, Horiuchi A, Imai T, Ohira S, Itoh K, Nikaido T, Katsuyama 49. Spain BH, Larson CJ, Shihabuddin LS, Gage FH, Verma IM: Trun- Y, Konishi I: Expression of BRCA1 protein in benign, border- cated BRCA2 is cytoplasmic: implications for cancer-linked line, and malignant epithelial ovarian neoplasms and its rela- mutations. Proc Natl Acad Sci USA 1999, 96:13920-13925. tionship to methylation and allelic loss of the BRCA1 gene. J 50. Herman JG, Umar A, Polyak K, Graff JR, Ahuja N, Issa JP, Markowitz Pathol 2004, 202:215-223. S, Willson JK, Hamilton SR, Kinzler KW, et al.: Incidence and func- 31. Blackshear PE, Goldsworthy SM, Foley JF, McAllister KA, Bennett LM, tional consequences of hMLH1 promoter hypermethylation Collins NK, Bunch DO, Brown P, Wiseman RW, Davis BJ: Brca1 and in colorectal carcinoma. Proc Natl Acad Sci USA 1998, Brca2 expression patterns in mitotic and meiotic cells of 95:6870-6875. mice. Oncogene 1998, 16:61-68. 32. Rajan JV, Wang M, Marquis ST, Chodosh LA: Brca2 is coordinately regulated with Brca1 during proliferation and differentiation in mammary epithelial cells. Proc Natl Acad Sci USA 1996, 93:13078-13083. 33. Bernard-Gallon DJ, Peffault de Latour M, Hizel C, Vissac C, Cure H, Pezet D, Dechelotte PJ, Chipponi J, Chassagne J, Bignon YJ: Localiza- tion of human BRCA1 and BRCA2 in non-inherited colorec- tal carcinomas and matched normal mucosas. Anticancer Res 2001, 21:2011-2020. 34. Rajan JV, Marquis ST, Gardner HP, Chodosh LA: Developmental expression of Brca2 colocalizes with Brca1 and is associated with proliferation and differentiation in multiple tissues. Dev Biol 1997, 184:385-401. 35. Andres JL, Fan S, Turkel GJ, Wang JA, Twu NF, Yuan RQ, Lamszus K, Goldberg ID, Rosen EM: Regulation of BRCA1 and BRCA2 expression in human breast cancer cells by DNA-damaging agents. Oncogene 1998, 16:2229-2241. 36. Fan S, Twu NF, Wang JA, Yuan RQ, Andres J, Goldberg ID, Rosen EM: Down-regulation of BRCA1 and BRCA2 in human ovarian cancer cells exposed to adriamycin and ultraviolet radiation. Int J Cancer 1998, 77:600-609. 37. Spillman MA, Bowcock AM: BRCA1 and BRCA2 mRNA levels are coordinately elevated in human breast cancer cells in response to estrogen. Oncogene 1996, 13:1639-1645. 38. Chiang JW, Karlan BY, Cass L, Baldwin RL: BRCA1 promoter methylation predicts adverse ovarian cancer prognosis. Gynecol Oncol 2006, 101:403-410. 39. Rhei E, Bogomolniy F, Federici MG, Maresco DL, Offit K, Robson ME, Saigo PE, Boyd J: Molecular genetic characterization of BRCA1- and BRCA2-linked hereditary ovarian cancers. Can- cer Res 1998, 58:3193-3196. 40. Ramus SJ, Bobrow LG, Pharoah PD, Finnigan DS, Fishman A, Altaras M, Harrington PA, Gayther SA, Ponder BA, Friedman LS: Increased frequency of TP53 mutations in BRCA1 and BRCA2 ovarian tumours. Genes Chromosomes Cancer 1999, 25:91-96. 41. Crook T, Crossland S, Crompton MR, Osin P, Gusterson BA: p53 mutations in BRCA1-associated familial breast cancer. Lan- Publish with Bio Med Central and every cet 1997, 350:638-639. scientist can read your work free of charge 42. Greenblatt MS, Chappuis PO, Bond JP, Hamel N, Foulkes WD: TP53 mutations in breast cancer associated with BRCA1 or "BioMed Central will be the most significant development for BRCA2 germ-line mutations: distinctive spectrum and disseminating the results of biomedical researc h in our lifetime." structural distribution. Cancer Res 2001, 61:4092-4097. Sir Paul Nurse, Cancer Research UK 43. Gifford G, Paul J, Vasey PA, Kaye SB, Brown R: The acquisition of hMLH1 methylation in plasma DNA after chemotherapy Your research papers will be: predicts poor survival for ovarian cancer patients. Clin Cancer available free of charge to the entire biomedical community Res 2004, 10:4420-4426. 44. Fink D, Nebel S, Norris PS, Baergen RN, Wilczynski SP, Costa MJ, peer reviewed and published immediately upon acceptance Haas M, Cannistra SA, Howell SB: Enrichment for DNA mis- cited in PubMed and archived on PubMed Central match repair-deficient cells during treatment with cisplatin. Int J Cancer 1998, 77:741-746. yours — you keep the copyright 45. Samimi G, Fink D, Varki NM, Husain A, Hoskins WJ, Alberts DS, BioMedcentral Howell SB: Analysis of MLH1 and MSH2 expression in ovarian Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 11 of 11 (page number not for citation purposes)

Journal

Molecular CancerSpringer Journals

Published: Jul 14, 2009

There are no references for this article.