- Email: [email protected]
Mutation and expression of the TP53 gene in early stage epithelial ovarian carcinoma
Gynecologic Oncology 93 (2004) 301 – 306 www.elsevier.com/locate/ygyno
Mutation and expression of the TP53 gene in early stage epithelial ovarian carcinoma Mario M. Leitao, a Robert A. Soslow, b Rebecca N. Baergen, c Narciso Olvera, a Crispinita Arroyo, a and Jeff Boyd a,d,* a Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA c Department of Pathology, New York Presbyterian Hospital-Weill Medical College of Cornell University, New York, NY 10021, USA d Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA b
Received 26 August 2003
Abstract Objective. The early natural history of epithelial ovarian carcinoma remains poorly understood. Mutation of the TP53 gene is common in advanced-stage (III – IV) ovarian cancers, but less well described in early stage (I – II) tumors. The purpose of this study was to perform a comprehensive analysis of TP53 mutation and p53 expression status in early stage ovarian carcinomas. Methods. Seventy-three cases of various histologic types, including 46 stage I and 27 stage II tumors, were subjected to direct sequence analysis of the entire TP53 coding region and exon – intron junctions as well as immunohistochemical assessment of p53 expression. Results. Overall, mutations were identified in 24 of 73 (34%) cases. However, a significant difference in the distribution of mutations among histologic types was observed; TP53 mutations were present in 14 of 21 (67%) serous cancers and 11 of 52 (21%) non-serous cancers ( P = 0.0002). Mutations were equally common between stage I and stage II tumors of serous histology. With respect to the correlation between TP53 mutation and p53 immunopositivity, the sensitivity (58%), specificity (71%), positive predictive value (64%), and negative predictive value (83%) were not sufficiently robust to justify use of p53 expression as a surrogate or screen for mutation. Conclusions. These data indicate that TP53 mutation is common in early stage ovarian carcinomas of serous histology, with a mutation frequency comparable to that reported for advanced-stage tumors, and is therefore likely to occur early in the progression of the most common histologic variant of ovarian carcinoma. D 2004 Elsevier Inc. All rights reserved. Keywords: TP53; p53; Tumor suppressor gene; Early stage; Ovarian cancer
Introduction Cancer of the ovary is the second most common gynecologic malignancy but accounts for more deaths than all other gynecologic cancers combined, with 25,400 new cases and 14,300 deaths expected for 2003 in the United States [1]. Epithelial ovarian carcinoma accounts for the great majority of both cases and deaths at this tumor site. Stage for stage, survival rates are similar to those for other solid tumors and exceed 80– 90% for early stage (FIGO stage I –II) disease; however, 75% of cases are diagnosed after spread from the * Corresponding author. Department of Surgery, Memorial SloanKettering Cancer Center, Box 201, 1275 York Avenue, New York, NY 10021. Fax: +1-212-717-3538. E-mail address: [email protected] (J. Boyd). 0090-8258/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2004.01.043
ovary and more than half are of advanced stage (FIGO stage III – IV) at diagnosis [2]. These statistics reflect the fact that there are no effective screening strategies currently available for ovarian carcinoma [3]. Furthermore, the early natural history of this tumor, at both histologic and molecular genetic levels, remains obscure. Identification and characterization of early events in ovarian tumorigenesis should facilitate the development of effective screening programs. The TP53 tumor suppressor gene is commonly mutated in a broad spectrum of human cancers [4]. The encoded p53 protein normally functions to sense DNA damage or other types of oncogenic stress, and in response, regulates a wide variety of downstream molecular events that culminate in cell cycle arrest or apoptosis [5– 7]. Loss of p53 function eliminates a critical source of cell cycle checkpoint control, contributing to inappropriate cell cycle progression and
302
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
tumorigenesis. Many of the common missense mutations that occur in TP53 are associated with posttranslational stabilization and relative overexpression of p53, allowing for its immunohistochemical detection. The correlation between mutation and expression is not perfect, however, as wild-type p53 protein is also stabilized in response to several types of oncogenic stress [6,7], and some TP53 mutations are associated with immunohistochemically undetectable p53 protein [8]. Somatic mutation of the TP53 gene represents the most common molecular genetic alteration known to occur in epithelial ovarian carcinoma [9]. A recent review summarizes the literature on TP53 mutation (41 studies) and p53 expression (98 studies) in various types of epithelial ovarian tumors [10]. Pooled prevalence estimates indicate that TP53 mutations are rare in benign (1%) and low malignant potential (5%) tumors, and common in invasive ovarian carcinomas (45%). Pooled prevalence estimates for p53 expression follow the same trend, but are somewhat higher for each of the three groups. For invasive carcinomas, the prevalence of mutation and expression increases with increasing tumor grade and stage, and is more common in tumors of serous histology. However, this literature is dominated by studies of small sample size, reliance on convenience samples, cases poorly characterized with respect to clinicopathologic features, use of a broad range of mutation screening techniques that fail to identify all mutations, and frequent use of p53 expression as a surrogate for TP53 mutation. Furthermore, the predominance of advanced-stage diagnoses in most series has further limited the molecular genetic characterization of early stage ovarian cancers. The prevailing consensus that TP53 mutation is more common in advanced stage than in early stage ovarian cancers, and thus represents a ‘‘late’’ event in the progression of ovarian carcinoma, has important implications for understanding the early natural history of ovarian tumorigenesis. In light of the limitations of the existing literature on this topic, we sought to perform a comprehensive analysis of TP53 mutation and p53 expression on a large sample of consecutive, early stage epithelial ovarian carcinoma cases from a single institution, well-characterized surgically and pathologically. The findings of this study proved instructive with respect to the role of TP53 mutation in ovarian tumorigenesis.
Materials and methods Case selection All cases of early stage (FIGO stage I or II), invasive, epithelial ovarian carcinoma that underwent primary surgical staging at Memorial Sloan-Kettering Cancer Center over the 21-year period from January 1, 1980, to December 31, 2000 were identified. After exclusion of 24 cases associated with a germline BRCA mutation (N = 9) or a strong family
history of breast or ovarian cancer (N = 15), 145 cases were eligible for study. Diagnostic pathology slides were not available for five cases. The remaining 140 cases were then reviewed independently by one or two gynecologic pathologists using current diagnostic criteria, and 46 cases were eliminated based on reclassification of tumors as other than invasive epithelial ovarian carcinoma. Archival formalinfixed and paraffin-embedded tissue specimens were available for 73 of the remaining 94 cases, and these were analyzed for TP53 mutation and p53 expression in this study. All tumor tissues analyzed were primary site specimens obtained before any chemotherapy or radiation, and noncancerous tissues from the reproductive tract were also accessioned. This study was approved by the Institutional Review Board of Memorial Sloan-Kettering Cancer Center. Tissue processing and DNA isolation A hematoxylin and eosin-stained slide was prepared from each tumor tissue block and reviewed to confirm the original pathologic diagnosis. An additional tissue section of 10-Am thickness was then obtained from each block and placed on Superfrost non-plus slides (Fisher Scientific) for microdissection. Tissue sections were deparaffinized with xylenes and rehydrated with a graded series of ethanol– water washes. Areas of tissue corresponding to invasive carcinoma were manually microdissected using an 18 G 1 1/ 2 PrecisionGlide needle (Becton, Dickinson and Co.) attached to a 5-cm3 syringe. Microdissected cells were suspended in 50 Al of digestion buffer consisting of 10 mM Tris – HCl, pH 8.0, 1 mM EDTA, 1% Tween, and 0.2 mg/ml of proteinase K. Samples were incubated at 54jC for 24 h, and then at 37jC until digestion was complete as determined by clarity of the solution. Following inactivation of proteinase K by heating at 95jC for 10 min, samples were diluted 1:20 with dH20 and stored at 4jC. TP53 mutation analysis Intron-based PCR primers flanking exons 2– 11 of the TP53 gene were designed according to the genomic sequence in GenBank accession number U94788. Primer sequences are available upon request. Individual TP53 exons and flanking exon– intron junctions were amplified in a two-step PCR protocol, the first accomplished in a 20 Al PCR containing 2 Al of template, 1.5 mM MgCl2, 50 mM KCl, 10 mM Tris – HCl, pH 8.3, each dNTP at 200 AM, forward and reverse primers at 0.8 AM each, and 0.75 U Taq polymerase (Roche). Thirty-five cycles of PCR were performed, each consisting of 20 s at 95jC, 20 s at 55jC, and 30 s at 72jC, followed by a 7 min extension at 72jC. Four microliters from this reaction was then used as template in a second PCR, under identical conditions, using nested PCR primers. Products from the second PCR were subjected to electrophoresis in 2% agarose gels, visualized with ethidium bromide, and eluted into 30 Al of dH2O using the QIAquick gel extraction kit (Qiagen).
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
303
Fig. 1. TP53 mutation detection and p53 expression analysis. Forward (A) and reverse (B) strand sequence of case #120 showing D259Y (GAC > TAC) mutation. Immunohistochemical analysis of p53 expression in case #39 showing +++ immunoreactivity (C), and in case #134 showing negative immunoreactivity (D).
All PCR experiments were performed using additional reactions containing water in place of DNA template to monitor for potential DNA contamination. Sequence analysis of PCR products was accomplished using the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems) in 20-Al reactions containing 6 Al of template, 4 Al of 5 BigDye Terminator Ready Reaction Mix, and primer at 160 nM. The primers used for sequencing reactions were the same as nested primers used for second round PCR amplification. Sequence analysis was performed using an ABI PRISM 377 automated sequencer (Applied Biosystems). All PCR products were sequenced in both directions using forward and reverse primers, and potential sequence variants were confirmed in independent PCR and sequencing reactions using original DNA samples as template. All exons and splice junctions were subjected to compete sequence analysis for all tumors. Immunohistochemical analysis of p53 expression Using the same fixed and paraffin-embedded tumor specimens that were analyzed for TP53 mutations, 5-Am tissue sections were prepared for immunohistochemistry using standard techniques. The mouse antihuman monoclonal DO-7 antibody (DakoCytomation) was used at a dilution of 1:500. This antibody is highly discriminant for both mutant and wild-type forms of p53 in human tissues at this concentration [11]. Following deparaffinization and rehydration, tissue sections were heated to boiling in citrate buffer, then incubated overnight at 4jC with primary antibody. Visualization was accomplished using the Mouse IgG Vectastain ABC Kit (Vector Laboratories), with secondary antibody used at a 1:500 dilution and diaminobenzidine used as
chromogen. Tissue sections were lightly counterstained with hematoxylin. An additional tissue section from each case was subjected to analysis without primary antibody as a negative control. Scoring for p53 expression was based on the proportion of cells in a given tumor specimen exhibiting nuclear immunopositivity as well as intensity of staining. Scoring for proportion of positive cells was 0 for no positive cells, 1 for 1 –15% positive cells, 2 for 16 – 30% positive cells, 3 for 31– 50% positive cells, and 4 for greater than 50% positive cells.
Table 1 Distribution of TP53 mutations in early stage ovarian carcinomas
Total cases Stage I II Histology Serous Endometrioid Clear cell Mucinous Adenocarcinoma, NOS Malignant Brenner Mixed Grade 1 2 3 a
Number of cases
Number with mutation (%)
73
25 (34)
46 27
9 (20) 16 (59)
0.0006a
21 20 17 5 1
14 4 2 1 1
0.0002b
(67) (20) (12) (20) (100)
2
1 (50)
7
2 (29)
19 30 24
2 (11) 11 (37) 12 (50)
Stage I compared to stage II. Serous compared to all other histologic types combined. c Grade 1 compared to grades 2 – 3. b
P
0.01c
304
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
Table 2 Clinicopathologic features and TP53 – p53 status for individual tumors Case no.
Mutation
1 3 4 7 16 27 36 39
Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type R273H
40
R273C
41 42 43 44
Wild-type IVS6 + 1G > A Wild-type R342X
45 46
F338fsX346 R248Q
48 50
P142del E298X
53 54 56 57 58 59 60 62
Wild-type Wild-type S241fsX246 Wild-type Wild-type F212fsX246 Wild-type Y327X
63 64
Wild-type R248Q
65
C238Y
66
R175H
71 72 73 74 80
Wild-type Wild-type Wild-type Wild-type C238Y
81 83 84 85
Wild-type Wild-type Wild-type R213X
86 87
Wild-type W146X G154S
88
K120E
89 90
Wild-type R175H
91 93
Wild-type Wild-type
Nucleotide change
IHC
Histology
Grade
Stage
+++
Endometrioid Serous Clear cell Mixed Endometrioid Clear cell Mixed Serous
1 2 2 3 1 3 3 2
IC IIC IC IIC IA IC IIC IIC
++
Serous
2
IIC
++ + ++ + CGT > CAT CGT > TGT
delC
CGG > CAG TGT > TAT CGC > CAC
CGC > CAC
Wild-type Wild-type Wild-type Wild-type Wild-type R181H
101 102
H233fsX238 Y220C
103 104 105 106 107 109 112 113 114 115 116 117 118 119 120
Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type IVS6 + 1G > A Wild-type Wild-type Wild-type Wild-type Wild-type D259Y
131 134 138 141 144
Wild-type A83fsX117 Wild-type Wild-type Wild-type
CGC > CAC del25 TAT > TGT
IHC
Histology
Grade
Stage
+
+
Endometrioid Endometrioid Endometrioid Endometrioid Endometrioid Clear cell
1 1 3 1 1 2
IC IA IC IC IA IC
+++
Serous Endometrioid
3 2
IIC IC
Serous Endometrioid Clear cell Clear cell Endometrioid Clear cell Clear cell Endometrioid Endometrioid
2 3 2 3 1 2 3 2 3
IC IC IA IA IA IA IIA IA IIC
Brenner Serous Serous Mixed Mixed Serous
2 3 2 2 1 3
IC IIC IIC IC IIA IIC
Mucinous Mucinous Mucinous Mucinous Mucinous
1 1 1 2 1
IC IA IA IA IC
+
Endometrioid Serous
1 2
IIB IC
+ +++
Serous Serous
3 3
IIB IIA
+++ +
Serous Serous
3 2
IA IIC
+++
Mixed Serous Serous Clear cell Clear cell Clear cell Clear cell Serous
1 2 3 3 3 2 2 3
IA IB IIB IC IC IC IA IIC
+ ++
Endometrioid Brenner
1 2
IIA IA
+++
Adenoca
3
IIC
+++
Serous
3
IC
Clear cell Serous Clear cell Clear cell Endometrioid
2 2 2 2 2
IC IIB IC IC IIB
+
Clear cell Endometrioid Endometrioid Endometrioid
1 2 1 1
IA IA IC IA
Comparisons among groups were assessed with chisquare and Fisher’s exact tests, as appropriate.
++ +++
Serous Serous
3 2
IC IIC
Results
+++ +
++ +
CGA > TGA TGG > TAG GGC > AGC AAG > GAG
94 95 96 98 99 100
Nucleotide change
IC IIC
+++
TGT > TAT
Mutation
2 2
delT TAT > TAG
Case no.
Clear cell Mixed
G>A
CGA > TGA insC CGG > CAG del3C CAG > TAG
Table 2 (continued)
G>A +++ +++ +
GAC > TAC del16
+++
Scoring for intensity of positive cells was based on a scale of 0 – 3. A final score for p53 expression was derived by multiplying proportion and intensity scores, with 0 – 1 considered negative, 2– 4 considered +, 5 – 8 considered ++, and 9 – 12 considered +++. Correlative analyses for TP53 mutation and p53 expression were performed with any level of positivity (+, ++, or +++) considered as positive. Statistical analyses
TP53 mutations +++
Mixed
3
IIC
++
Clear cell Serous
3 3
IC IIC
Endometrioid Endometrioid
1 2
IC IIB
Of the 73 early stage ovarian carcinomas analyzed, 25 (34%) were found to contain a TP53 mutation (Fig. 1). Data on the distribution of mutations by stage, histologic type, and grade are summarized in Table 1. Mutations were significantly more common in stage II than in stage I tumors, in serous than in nonserous tumors, and in moderate
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
to high grade compared to low grade tumors. In tumors of serous histology, the proportion with a TP53 mutation was not significantly different in stage I (3/6, 50%) compared to stage II (11/15, 73%) tumors ( P = 0.9). Differences with respect to grade were not assessed as there were no lowgrade serous tumors in this series. Significant differences were not noted within other histologic types, although this may well have reflected inadequate power associated with relatively small sample sizes. Data on the nature of individual TP53 mutations are presented in Table 2. A total of 26 mutations were identified in 25 tumors. The majority of mutations (20/26, 77%) were single nucleotide substitutions, 85% of which were transitions and 15% of which were transversions. Of the 20 single nucleotide substitution mutations, 13 (65%) led to missense alterations, 5 (25%) led to nonsense alterations, and 2 (10%) affected intronic splice junctions. Of the six frameshift mutations, five (83%) were deletions and one (17%) was an insertion, all occurring within the coding region. One deletion mutation was predicted to cause an in-frame amino acid deletion, while the remainder was predicted to result in premature protein translation. The majority of mutations (21/26, 81%) occurred within exons 5 – 8; no mutations were detected in exons 2, 3, or 11. Five of the mutations (K120E, P142del, A83fsX117, H233fsX238, and F338fsX346) have not previously been reported in human cancers, while five additional mutations (G154S, W146X, R213X, E298X, and Y327X) have not previously been reported in ovarian cancers [12,13]. p53 expression Immunohistochemical analysis of p53 expression revealed that 33/73 (45%) tumors were immunopositive to some degree. Of tumors with a TP53 mutation, p53 expression was positive in 18/25 (72%) cases, including 12/12 (100%) tumors with a missense mutation, 6/11 (55%) tumors with a frameshift or nonsense mutation, and 0/2 tumors with an intronic splice site mutation (Table 2 and Fig. 1). Tumor #87, which contained both frameshift and missense mutations, was considered in the frameshift mutation category because this occurred 5V to the missense mutation. Of tumors with no TP53 mutation, 15/48 (31%) cases were also immunopositive for p53 expression. In terms of p53 expression status, 18/33 (55%) immunopositive tumors contained a mutation, while 33/40 (83%) immunonegative tumors did not contain a mutation. The sensitivity, specificity, positive predictive value, and negative predictive value of p53 expression for the presence of a TP53 mutation were 72%, 69%, 55%, and 83%, respectively.
Discussion These data indicate that TP53 mutations occur in approximately one-third (34%) of all stage I or II invasive
305
epithelial ovarian carcinomas. Mutations are significantly more common in stage II than in stage I cancers, and in moderately to poorly differentiated than in well-differentiated tumors. However, when stratified by histologic subtype, the majority (67%) of early stage serous tumors harbors TP53 mutations, and the fraction is not significantly different in stage I compared to stage II tumors. In contrast, only 12– 20% of early stage endometrioid, clear cell, and mucinous tumors have mutated TP53. These findings are of clinical relevance as we have recently shown that TP53 mutation in early stage epithelial ovarian carcinoma is an adverse prognostic indicator [14]. Consideration of these findings in light of the existing literature on TP53 mutation in ovarian cancer is instructive with respect to the temporal nature of mutation in disease progression. The fraction of early stage serous ovarian cancers with mutant TP53 (67%) reported here is somewhat higher than that reported (51%) for all stages of invasive serous ovarian carcinoma combined in a pooled analysis [10]. The higher fraction reported in this study likely reflects the comprehensive nature of the mutational analysis performed; the great majority of published data on TP53 mutation is derived from screening techniques with incomplete sensitivity and specificity, such as immunohistochemical or single-strand conformation polymorphism analyses. This finding is consistent with the conclusion that TP53 mutation occurs before metastatic dissemination of serous ovarian carcinomas, which account for the majority of all invasive epithelial ovarian cancers [15]. Similarly, the fractions of early stage endometrioid (20%), clear cell (12%), and mucinous (20%) tumors with TP53 mutations, although derived from relatively small samples, are similar to pooled literature values for these histologic types considered without respect to stage (33%, 9%, and 30%, respectively), suggesting that when TP53 mutation occurs in these less common histologic variants, it also occurs early in the natural history of the disease. Several conclusions may also be drawn from the data on p53 expression as assessed by immunohistochemical staining. From a practical perspective, the sensitivity, specificity, and predictive values of p53 immunoreactivity as a surrogate for TP53 mutation are modest. While this has been noted before for epithelial ovarian carcinoma [8], there is a large literature based solely in p53 expression and its clinical implications; however, it is clear that ovarian cancers classified solely from p53 expression status represent heterogenous groups of tumors. The mechanism through which p53 protein accumulation occurs in the presence of missense mutations is well described [7], but it remains less clear what molecular determinants drive p53 accumulation in cancers with wild-type p53. It is noteworthy that in this series of early stage ovarian cancers, nearly all (95%) serous tumors were shown to have either TP53 mutation or p53 expression, while only a minor fraction of endometrioid (35%), clear cell (35%), and mucinous (20%) tumors were affected by this combination of TP53 – p53
306
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
aberrations. This finding provides additional support to the hypothesis that serous ovarian carcinomas arise through a distinct molecular pathway(s) compared to ovarian carcinomas of nonserous histologic subtypes [16]. Acknowledgments This work was supported by grants from the W.M. Keck Foundation and the National Institutes of Health (R01 CA71840 and U01 CA88175). References [1] Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. Cancer statistics, 2003. CA Cancer J Clin 2003;53:5 – 26. [2] Ozols RF, Rubin SC, Thomas GM, Robboy SJ. Epithelial ovarian cancer. In: Hoskins WJ, Perez CA, Young RC, editors. Principles and Practice of Gynecologic Oncology. Third ed. Philadelphia: Lippincott, Williams and Wilkins; 2000. p. 981 – 1058. [3] Rosenthal A, Jacobs I. Ovarian cancer screening. Semin Oncol 1998; 25:315 – 25. [4] Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 1994;54:4855 – 78. [5] Levine AJ. p53, the cellular gatekeeper for growth and division. Cell 1997;88:323 – 31. [6] Oren M. Regulation of the p53 tumor suppressor protein. J Biol Chem 1999;274:36031 – 4. [7] Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature 2000;408:307 – 10.
[8] Casey G, Lopez ME, Ramos JC, Plummer SJ, Arboleda MJ, Shaughnessy M, et al. DNA sequence analysis of exons 2 through 11 and immunohistochemical staining are required to detect all known p53 alterations in human malignancies. Oncogene 1996;13:1971 – 81. [9] Boyd J, Hamilton TC, Berchuck A. Oncogenes and tumor-suppressor genes. In: Hoskins WJ, Perez CA, Young RC, editors. Principles and Practice of Gynecologic Oncology. Third ed. Philadelphia: Lippincott, Williams and Wilkins; 2000. p. 103 – 28. [10] Kmet LM, Cook LS, Magliocco AM. A review of p53 expression and mutation in human benign, low malignant potential, and invasive epithelial ovarian tumors. Cancer 2003;97:389 – 404. [11] Bonsing BA, Corver WE, Gorsira MC, van Vliet M, Oud PS, Cornelisse CJ, et al. Specificity of seven monoclonal antibodies against p53 evaluated with Western blotting, immunohistochemistry, confocal laser scanning microscopy, and flow cytometry. Cytometry 1997;28: 11 – 24. [12] Hernandez-Boussard T, Rodriguez-Tome P, Montesano R, Hainaut P. IARC p53 mutation database: a relational database to compile and analyze p53 mutations in human tumors and cell lines. International Agency for Research on Cancer. Hum Mutat 1999;14:1 – 8. [13] Soussi T, Dehouche K, Beroud C. p53 website and analysis of p53 gene mutations in human cancer: forging a link between epidemiology and carcinogenesis. Hum Mutat 2000;15:105 – 13. [14] Leitao MM, Boyd J, Hummer A, Olvera N, Arroyo CD, Venkatraman E, et al. Clinicopathologic analysis of early-stage sporadic epithelial ovarian carcinoma. Am J Surg Pathol 2004;28:147 – 59. [15] Russell P. Surface epithelial-stromal tumors of the ovary. In: Kurman RJ, editor. Blaustein’s Pathology of the Female Genital Tract. Fourth ed. New York: Springer-Verlag; 1994. p. 705 – 82. [16] Schwartz DR, Kardia SL, Shedden KA, Kuick R, Michailidis G, Taylor JM, et al. Gene expression in ovarian cancer reflects both morphology and biological behavior, distinguishing clear cell from other poor-prognosis ovarian carcinomas. Cancer Res 2002;62: 4722 – 9.
Mutation and expression of the TP53 gene in early stage epithelial ovarian carcinoma Mario M. Leitao, a Robert A. Soslow, b Rebecca N. Baergen, c Narciso Olvera, a Crispinita Arroyo, a and Jeff Boyd a,d,* a Department of Surgery, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA c Department of Pathology, New York Presbyterian Hospital-Weill Medical College of Cornell University, New York, NY 10021, USA d Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA b
Received 26 August 2003
Abstract Objective. The early natural history of epithelial ovarian carcinoma remains poorly understood. Mutation of the TP53 gene is common in advanced-stage (III – IV) ovarian cancers, but less well described in early stage (I – II) tumors. The purpose of this study was to perform a comprehensive analysis of TP53 mutation and p53 expression status in early stage ovarian carcinomas. Methods. Seventy-three cases of various histologic types, including 46 stage I and 27 stage II tumors, were subjected to direct sequence analysis of the entire TP53 coding region and exon – intron junctions as well as immunohistochemical assessment of p53 expression. Results. Overall, mutations were identified in 24 of 73 (34%) cases. However, a significant difference in the distribution of mutations among histologic types was observed; TP53 mutations were present in 14 of 21 (67%) serous cancers and 11 of 52 (21%) non-serous cancers ( P = 0.0002). Mutations were equally common between stage I and stage II tumors of serous histology. With respect to the correlation between TP53 mutation and p53 immunopositivity, the sensitivity (58%), specificity (71%), positive predictive value (64%), and negative predictive value (83%) were not sufficiently robust to justify use of p53 expression as a surrogate or screen for mutation. Conclusions. These data indicate that TP53 mutation is common in early stage ovarian carcinomas of serous histology, with a mutation frequency comparable to that reported for advanced-stage tumors, and is therefore likely to occur early in the progression of the most common histologic variant of ovarian carcinoma. D 2004 Elsevier Inc. All rights reserved. Keywords: TP53; p53; Tumor suppressor gene; Early stage; Ovarian cancer
Introduction Cancer of the ovary is the second most common gynecologic malignancy but accounts for more deaths than all other gynecologic cancers combined, with 25,400 new cases and 14,300 deaths expected for 2003 in the United States [1]. Epithelial ovarian carcinoma accounts for the great majority of both cases and deaths at this tumor site. Stage for stage, survival rates are similar to those for other solid tumors and exceed 80– 90% for early stage (FIGO stage I –II) disease; however, 75% of cases are diagnosed after spread from the * Corresponding author. Department of Surgery, Memorial SloanKettering Cancer Center, Box 201, 1275 York Avenue, New York, NY 10021. Fax: +1-212-717-3538. E-mail address: [email protected] (J. Boyd). 0090-8258/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2004.01.043
ovary and more than half are of advanced stage (FIGO stage III – IV) at diagnosis [2]. These statistics reflect the fact that there are no effective screening strategies currently available for ovarian carcinoma [3]. Furthermore, the early natural history of this tumor, at both histologic and molecular genetic levels, remains obscure. Identification and characterization of early events in ovarian tumorigenesis should facilitate the development of effective screening programs. The TP53 tumor suppressor gene is commonly mutated in a broad spectrum of human cancers [4]. The encoded p53 protein normally functions to sense DNA damage or other types of oncogenic stress, and in response, regulates a wide variety of downstream molecular events that culminate in cell cycle arrest or apoptosis [5– 7]. Loss of p53 function eliminates a critical source of cell cycle checkpoint control, contributing to inappropriate cell cycle progression and
302
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
tumorigenesis. Many of the common missense mutations that occur in TP53 are associated with posttranslational stabilization and relative overexpression of p53, allowing for its immunohistochemical detection. The correlation between mutation and expression is not perfect, however, as wild-type p53 protein is also stabilized in response to several types of oncogenic stress [6,7], and some TP53 mutations are associated with immunohistochemically undetectable p53 protein [8]. Somatic mutation of the TP53 gene represents the most common molecular genetic alteration known to occur in epithelial ovarian carcinoma [9]. A recent review summarizes the literature on TP53 mutation (41 studies) and p53 expression (98 studies) in various types of epithelial ovarian tumors [10]. Pooled prevalence estimates indicate that TP53 mutations are rare in benign (1%) and low malignant potential (5%) tumors, and common in invasive ovarian carcinomas (45%). Pooled prevalence estimates for p53 expression follow the same trend, but are somewhat higher for each of the three groups. For invasive carcinomas, the prevalence of mutation and expression increases with increasing tumor grade and stage, and is more common in tumors of serous histology. However, this literature is dominated by studies of small sample size, reliance on convenience samples, cases poorly characterized with respect to clinicopathologic features, use of a broad range of mutation screening techniques that fail to identify all mutations, and frequent use of p53 expression as a surrogate for TP53 mutation. Furthermore, the predominance of advanced-stage diagnoses in most series has further limited the molecular genetic characterization of early stage ovarian cancers. The prevailing consensus that TP53 mutation is more common in advanced stage than in early stage ovarian cancers, and thus represents a ‘‘late’’ event in the progression of ovarian carcinoma, has important implications for understanding the early natural history of ovarian tumorigenesis. In light of the limitations of the existing literature on this topic, we sought to perform a comprehensive analysis of TP53 mutation and p53 expression on a large sample of consecutive, early stage epithelial ovarian carcinoma cases from a single institution, well-characterized surgically and pathologically. The findings of this study proved instructive with respect to the role of TP53 mutation in ovarian tumorigenesis.
Materials and methods Case selection All cases of early stage (FIGO stage I or II), invasive, epithelial ovarian carcinoma that underwent primary surgical staging at Memorial Sloan-Kettering Cancer Center over the 21-year period from January 1, 1980, to December 31, 2000 were identified. After exclusion of 24 cases associated with a germline BRCA mutation (N = 9) or a strong family
history of breast or ovarian cancer (N = 15), 145 cases were eligible for study. Diagnostic pathology slides were not available for five cases. The remaining 140 cases were then reviewed independently by one or two gynecologic pathologists using current diagnostic criteria, and 46 cases were eliminated based on reclassification of tumors as other than invasive epithelial ovarian carcinoma. Archival formalinfixed and paraffin-embedded tissue specimens were available for 73 of the remaining 94 cases, and these were analyzed for TP53 mutation and p53 expression in this study. All tumor tissues analyzed were primary site specimens obtained before any chemotherapy or radiation, and noncancerous tissues from the reproductive tract were also accessioned. This study was approved by the Institutional Review Board of Memorial Sloan-Kettering Cancer Center. Tissue processing and DNA isolation A hematoxylin and eosin-stained slide was prepared from each tumor tissue block and reviewed to confirm the original pathologic diagnosis. An additional tissue section of 10-Am thickness was then obtained from each block and placed on Superfrost non-plus slides (Fisher Scientific) for microdissection. Tissue sections were deparaffinized with xylenes and rehydrated with a graded series of ethanol– water washes. Areas of tissue corresponding to invasive carcinoma were manually microdissected using an 18 G 1 1/ 2 PrecisionGlide needle (Becton, Dickinson and Co.) attached to a 5-cm3 syringe. Microdissected cells were suspended in 50 Al of digestion buffer consisting of 10 mM Tris – HCl, pH 8.0, 1 mM EDTA, 1% Tween, and 0.2 mg/ml of proteinase K. Samples were incubated at 54jC for 24 h, and then at 37jC until digestion was complete as determined by clarity of the solution. Following inactivation of proteinase K by heating at 95jC for 10 min, samples were diluted 1:20 with dH20 and stored at 4jC. TP53 mutation analysis Intron-based PCR primers flanking exons 2– 11 of the TP53 gene were designed according to the genomic sequence in GenBank accession number U94788. Primer sequences are available upon request. Individual TP53 exons and flanking exon– intron junctions were amplified in a two-step PCR protocol, the first accomplished in a 20 Al PCR containing 2 Al of template, 1.5 mM MgCl2, 50 mM KCl, 10 mM Tris – HCl, pH 8.3, each dNTP at 200 AM, forward and reverse primers at 0.8 AM each, and 0.75 U Taq polymerase (Roche). Thirty-five cycles of PCR were performed, each consisting of 20 s at 95jC, 20 s at 55jC, and 30 s at 72jC, followed by a 7 min extension at 72jC. Four microliters from this reaction was then used as template in a second PCR, under identical conditions, using nested PCR primers. Products from the second PCR were subjected to electrophoresis in 2% agarose gels, visualized with ethidium bromide, and eluted into 30 Al of dH2O using the QIAquick gel extraction kit (Qiagen).
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
303
Fig. 1. TP53 mutation detection and p53 expression analysis. Forward (A) and reverse (B) strand sequence of case #120 showing D259Y (GAC > TAC) mutation. Immunohistochemical analysis of p53 expression in case #39 showing +++ immunoreactivity (C), and in case #134 showing negative immunoreactivity (D).
All PCR experiments were performed using additional reactions containing water in place of DNA template to monitor for potential DNA contamination. Sequence analysis of PCR products was accomplished using the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems) in 20-Al reactions containing 6 Al of template, 4 Al of 5 BigDye Terminator Ready Reaction Mix, and primer at 160 nM. The primers used for sequencing reactions were the same as nested primers used for second round PCR amplification. Sequence analysis was performed using an ABI PRISM 377 automated sequencer (Applied Biosystems). All PCR products were sequenced in both directions using forward and reverse primers, and potential sequence variants were confirmed in independent PCR and sequencing reactions using original DNA samples as template. All exons and splice junctions were subjected to compete sequence analysis for all tumors. Immunohistochemical analysis of p53 expression Using the same fixed and paraffin-embedded tumor specimens that were analyzed for TP53 mutations, 5-Am tissue sections were prepared for immunohistochemistry using standard techniques. The mouse antihuman monoclonal DO-7 antibody (DakoCytomation) was used at a dilution of 1:500. This antibody is highly discriminant for both mutant and wild-type forms of p53 in human tissues at this concentration [11]. Following deparaffinization and rehydration, tissue sections were heated to boiling in citrate buffer, then incubated overnight at 4jC with primary antibody. Visualization was accomplished using the Mouse IgG Vectastain ABC Kit (Vector Laboratories), with secondary antibody used at a 1:500 dilution and diaminobenzidine used as
chromogen. Tissue sections were lightly counterstained with hematoxylin. An additional tissue section from each case was subjected to analysis without primary antibody as a negative control. Scoring for p53 expression was based on the proportion of cells in a given tumor specimen exhibiting nuclear immunopositivity as well as intensity of staining. Scoring for proportion of positive cells was 0 for no positive cells, 1 for 1 –15% positive cells, 2 for 16 – 30% positive cells, 3 for 31– 50% positive cells, and 4 for greater than 50% positive cells.
Table 1 Distribution of TP53 mutations in early stage ovarian carcinomas
Total cases Stage I II Histology Serous Endometrioid Clear cell Mucinous Adenocarcinoma, NOS Malignant Brenner Mixed Grade 1 2 3 a
Number of cases
Number with mutation (%)
73
25 (34)
46 27
9 (20) 16 (59)
0.0006a
21 20 17 5 1
14 4 2 1 1
0.0002b
(67) (20) (12) (20) (100)
2
1 (50)
7
2 (29)
19 30 24
2 (11) 11 (37) 12 (50)
Stage I compared to stage II. Serous compared to all other histologic types combined. c Grade 1 compared to grades 2 – 3. b
P
0.01c
304
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
Table 2 Clinicopathologic features and TP53 – p53 status for individual tumors Case no.
Mutation
1 3 4 7 16 27 36 39
Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type R273H
40
R273C
41 42 43 44
Wild-type IVS6 + 1G > A Wild-type R342X
45 46
F338fsX346 R248Q
48 50
P142del E298X
53 54 56 57 58 59 60 62
Wild-type Wild-type S241fsX246 Wild-type Wild-type F212fsX246 Wild-type Y327X
63 64
Wild-type R248Q
65
C238Y
66
R175H
71 72 73 74 80
Wild-type Wild-type Wild-type Wild-type C238Y
81 83 84 85
Wild-type Wild-type Wild-type R213X
86 87
Wild-type W146X G154S
88
K120E
89 90
Wild-type R175H
91 93
Wild-type Wild-type
Nucleotide change
IHC
Histology
Grade
Stage
+++
Endometrioid Serous Clear cell Mixed Endometrioid Clear cell Mixed Serous
1 2 2 3 1 3 3 2
IC IIC IC IIC IA IC IIC IIC
++
Serous
2
IIC
++ + ++ + CGT > CAT CGT > TGT
delC
CGG > CAG TGT > TAT CGC > CAC
CGC > CAC
Wild-type Wild-type Wild-type Wild-type Wild-type R181H
101 102
H233fsX238 Y220C
103 104 105 106 107 109 112 113 114 115 116 117 118 119 120
Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type Wild-type IVS6 + 1G > A Wild-type Wild-type Wild-type Wild-type Wild-type D259Y
131 134 138 141 144
Wild-type A83fsX117 Wild-type Wild-type Wild-type
CGC > CAC del25 TAT > TGT
IHC
Histology
Grade
Stage
+
+
Endometrioid Endometrioid Endometrioid Endometrioid Endometrioid Clear cell
1 1 3 1 1 2
IC IA IC IC IA IC
+++
Serous Endometrioid
3 2
IIC IC
Serous Endometrioid Clear cell Clear cell Endometrioid Clear cell Clear cell Endometrioid Endometrioid
2 3 2 3 1 2 3 2 3
IC IC IA IA IA IA IIA IA IIC
Brenner Serous Serous Mixed Mixed Serous
2 3 2 2 1 3
IC IIC IIC IC IIA IIC
Mucinous Mucinous Mucinous Mucinous Mucinous
1 1 1 2 1
IC IA IA IA IC
+
Endometrioid Serous
1 2
IIB IC
+ +++
Serous Serous
3 3
IIB IIA
+++ +
Serous Serous
3 2
IA IIC
+++
Mixed Serous Serous Clear cell Clear cell Clear cell Clear cell Serous
1 2 3 3 3 2 2 3
IA IB IIB IC IC IC IA IIC
+ ++
Endometrioid Brenner
1 2
IIA IA
+++
Adenoca
3
IIC
+++
Serous
3
IC
Clear cell Serous Clear cell Clear cell Endometrioid
2 2 2 2 2
IC IIB IC IC IIB
+
Clear cell Endometrioid Endometrioid Endometrioid
1 2 1 1
IA IA IC IA
Comparisons among groups were assessed with chisquare and Fisher’s exact tests, as appropriate.
++ +++
Serous Serous
3 2
IC IIC
Results
+++ +
++ +
CGA > TGA TGG > TAG GGC > AGC AAG > GAG
94 95 96 98 99 100
Nucleotide change
IC IIC
+++
TGT > TAT
Mutation
2 2
delT TAT > TAG
Case no.
Clear cell Mixed
G>A
CGA > TGA insC CGG > CAG del3C CAG > TAG
Table 2 (continued)
G>A +++ +++ +
GAC > TAC del16
+++
Scoring for intensity of positive cells was based on a scale of 0 – 3. A final score for p53 expression was derived by multiplying proportion and intensity scores, with 0 – 1 considered negative, 2– 4 considered +, 5 – 8 considered ++, and 9 – 12 considered +++. Correlative analyses for TP53 mutation and p53 expression were performed with any level of positivity (+, ++, or +++) considered as positive. Statistical analyses
TP53 mutations +++
Mixed
3
IIC
++
Clear cell Serous
3 3
IC IIC
Endometrioid Endometrioid
1 2
IC IIB
Of the 73 early stage ovarian carcinomas analyzed, 25 (34%) were found to contain a TP53 mutation (Fig. 1). Data on the distribution of mutations by stage, histologic type, and grade are summarized in Table 1. Mutations were significantly more common in stage II than in stage I tumors, in serous than in nonserous tumors, and in moderate
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
to high grade compared to low grade tumors. In tumors of serous histology, the proportion with a TP53 mutation was not significantly different in stage I (3/6, 50%) compared to stage II (11/15, 73%) tumors ( P = 0.9). Differences with respect to grade were not assessed as there were no lowgrade serous tumors in this series. Significant differences were not noted within other histologic types, although this may well have reflected inadequate power associated with relatively small sample sizes. Data on the nature of individual TP53 mutations are presented in Table 2. A total of 26 mutations were identified in 25 tumors. The majority of mutations (20/26, 77%) were single nucleotide substitutions, 85% of which were transitions and 15% of which were transversions. Of the 20 single nucleotide substitution mutations, 13 (65%) led to missense alterations, 5 (25%) led to nonsense alterations, and 2 (10%) affected intronic splice junctions. Of the six frameshift mutations, five (83%) were deletions and one (17%) was an insertion, all occurring within the coding region. One deletion mutation was predicted to cause an in-frame amino acid deletion, while the remainder was predicted to result in premature protein translation. The majority of mutations (21/26, 81%) occurred within exons 5 – 8; no mutations were detected in exons 2, 3, or 11. Five of the mutations (K120E, P142del, A83fsX117, H233fsX238, and F338fsX346) have not previously been reported in human cancers, while five additional mutations (G154S, W146X, R213X, E298X, and Y327X) have not previously been reported in ovarian cancers [12,13]. p53 expression Immunohistochemical analysis of p53 expression revealed that 33/73 (45%) tumors were immunopositive to some degree. Of tumors with a TP53 mutation, p53 expression was positive in 18/25 (72%) cases, including 12/12 (100%) tumors with a missense mutation, 6/11 (55%) tumors with a frameshift or nonsense mutation, and 0/2 tumors with an intronic splice site mutation (Table 2 and Fig. 1). Tumor #87, which contained both frameshift and missense mutations, was considered in the frameshift mutation category because this occurred 5V to the missense mutation. Of tumors with no TP53 mutation, 15/48 (31%) cases were also immunopositive for p53 expression. In terms of p53 expression status, 18/33 (55%) immunopositive tumors contained a mutation, while 33/40 (83%) immunonegative tumors did not contain a mutation. The sensitivity, specificity, positive predictive value, and negative predictive value of p53 expression for the presence of a TP53 mutation were 72%, 69%, 55%, and 83%, respectively.
Discussion These data indicate that TP53 mutations occur in approximately one-third (34%) of all stage I or II invasive
305
epithelial ovarian carcinomas. Mutations are significantly more common in stage II than in stage I cancers, and in moderately to poorly differentiated than in well-differentiated tumors. However, when stratified by histologic subtype, the majority (67%) of early stage serous tumors harbors TP53 mutations, and the fraction is not significantly different in stage I compared to stage II tumors. In contrast, only 12– 20% of early stage endometrioid, clear cell, and mucinous tumors have mutated TP53. These findings are of clinical relevance as we have recently shown that TP53 mutation in early stage epithelial ovarian carcinoma is an adverse prognostic indicator [14]. Consideration of these findings in light of the existing literature on TP53 mutation in ovarian cancer is instructive with respect to the temporal nature of mutation in disease progression. The fraction of early stage serous ovarian cancers with mutant TP53 (67%) reported here is somewhat higher than that reported (51%) for all stages of invasive serous ovarian carcinoma combined in a pooled analysis [10]. The higher fraction reported in this study likely reflects the comprehensive nature of the mutational analysis performed; the great majority of published data on TP53 mutation is derived from screening techniques with incomplete sensitivity and specificity, such as immunohistochemical or single-strand conformation polymorphism analyses. This finding is consistent with the conclusion that TP53 mutation occurs before metastatic dissemination of serous ovarian carcinomas, which account for the majority of all invasive epithelial ovarian cancers [15]. Similarly, the fractions of early stage endometrioid (20%), clear cell (12%), and mucinous (20%) tumors with TP53 mutations, although derived from relatively small samples, are similar to pooled literature values for these histologic types considered without respect to stage (33%, 9%, and 30%, respectively), suggesting that when TP53 mutation occurs in these less common histologic variants, it also occurs early in the natural history of the disease. Several conclusions may also be drawn from the data on p53 expression as assessed by immunohistochemical staining. From a practical perspective, the sensitivity, specificity, and predictive values of p53 immunoreactivity as a surrogate for TP53 mutation are modest. While this has been noted before for epithelial ovarian carcinoma [8], there is a large literature based solely in p53 expression and its clinical implications; however, it is clear that ovarian cancers classified solely from p53 expression status represent heterogenous groups of tumors. The mechanism through which p53 protein accumulation occurs in the presence of missense mutations is well described [7], but it remains less clear what molecular determinants drive p53 accumulation in cancers with wild-type p53. It is noteworthy that in this series of early stage ovarian cancers, nearly all (95%) serous tumors were shown to have either TP53 mutation or p53 expression, while only a minor fraction of endometrioid (35%), clear cell (35%), and mucinous (20%) tumors were affected by this combination of TP53 – p53
306
M.M. Leitao et al. / Gynecologic Oncology 93 (2004) 301–306
aberrations. This finding provides additional support to the hypothesis that serous ovarian carcinomas arise through a distinct molecular pathway(s) compared to ovarian carcinomas of nonserous histologic subtypes [16]. Acknowledgments This work was supported by grants from the W.M. Keck Foundation and the National Institutes of Health (R01 CA71840 and U01 CA88175). References [1] Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. Cancer statistics, 2003. CA Cancer J Clin 2003;53:5 – 26. [2] Ozols RF, Rubin SC, Thomas GM, Robboy SJ. Epithelial ovarian cancer. In: Hoskins WJ, Perez CA, Young RC, editors. Principles and Practice of Gynecologic Oncology. Third ed. Philadelphia: Lippincott, Williams and Wilkins; 2000. p. 981 – 1058. [3] Rosenthal A, Jacobs I. Ovarian cancer screening. Semin Oncol 1998; 25:315 – 25. [4] Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 1994;54:4855 – 78. [5] Levine AJ. p53, the cellular gatekeeper for growth and division. Cell 1997;88:323 – 31. [6] Oren M. Regulation of the p53 tumor suppressor protein. J Biol Chem 1999;274:36031 – 4. [7] Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature 2000;408:307 – 10.
[8] Casey G, Lopez ME, Ramos JC, Plummer SJ, Arboleda MJ, Shaughnessy M, et al. DNA sequence analysis of exons 2 through 11 and immunohistochemical staining are required to detect all known p53 alterations in human malignancies. Oncogene 1996;13:1971 – 81. [9] Boyd J, Hamilton TC, Berchuck A. Oncogenes and tumor-suppressor genes. In: Hoskins WJ, Perez CA, Young RC, editors. Principles and Practice of Gynecologic Oncology. Third ed. Philadelphia: Lippincott, Williams and Wilkins; 2000. p. 103 – 28. [10] Kmet LM, Cook LS, Magliocco AM. A review of p53 expression and mutation in human benign, low malignant potential, and invasive epithelial ovarian tumors. Cancer 2003;97:389 – 404. [11] Bonsing BA, Corver WE, Gorsira MC, van Vliet M, Oud PS, Cornelisse CJ, et al. Specificity of seven monoclonal antibodies against p53 evaluated with Western blotting, immunohistochemistry, confocal laser scanning microscopy, and flow cytometry. Cytometry 1997;28: 11 – 24. [12] Hernandez-Boussard T, Rodriguez-Tome P, Montesano R, Hainaut P. IARC p53 mutation database: a relational database to compile and analyze p53 mutations in human tumors and cell lines. International Agency for Research on Cancer. Hum Mutat 1999;14:1 – 8. [13] Soussi T, Dehouche K, Beroud C. p53 website and analysis of p53 gene mutations in human cancer: forging a link between epidemiology and carcinogenesis. Hum Mutat 2000;15:105 – 13. [14] Leitao MM, Boyd J, Hummer A, Olvera N, Arroyo CD, Venkatraman E, et al. Clinicopathologic analysis of early-stage sporadic epithelial ovarian carcinoma. Am J Surg Pathol 2004;28:147 – 59. [15] Russell P. Surface epithelial-stromal tumors of the ovary. In: Kurman RJ, editor. Blaustein’s Pathology of the Female Genital Tract. Fourth ed. New York: Springer-Verlag; 1994. p. 705 – 82. [16] Schwartz DR, Kardia SL, Shedden KA, Kuick R, Michailidis G, Taylor JM, et al. Gene expression in ovarian cancer reflects both morphology and biological behavior, distinguishing clear cell from other poor-prognosis ovarian carcinomas. Cancer Res 2002;62: 4722 – 9.