Methylation of tumor suppressor gene p16 and prognosis of epithelial ovarian cancer

Gynecologic Oncology 94 (2004) 685 – 692 www.elsevier.com/locate/ygyno

Methylation of tumor suppressor gene p16 and prognosis of epithelial ovarian cancer D. Katsaros a, W. Cho b, R. Singal c, S. Fracchioli a, I.A. Rigault de la Longrais a, R. Arisio a, M. Massobrio a, M. Smith c, W. Zheng b, J. Glass c, H. Yu b,* b

a University of Turin, Turin, Italy Department of Epidemiology and Public Health, Yale Cancer Center, Yale University School of Medicine, New Haven, CT 06520-8034, USA c Louisiana State University Health Sciences Center, Shreveport, LA 71130, USA

Received 9 December 2003 Available online 31 July 2004

Abstract Objective. Methylation of p16 promoter was evaluated in ovarian cancer to determine the role of p16 methylation in ovarian cancer prognosis. Methods. Two hundred and forty-nine patients with primary epithelial ovarian cancer were selected for the study; these patients were followed for a median of 31 months. Genomic DNA extracted from fresh frozen tumor tissues were treated with sodium bisulfite and were analyzed for p16 methylation using methylation-specific PCR (MSP). Cox regression survival analysis was performed to examine the associations of p16 methylation with progression-free and overall survivals. Results. Of the 249 patients, 100 (40%) were tested positive for p16 promoter methylation. The status of p16 methylation did not change significantly with patient age, disease stage, histological grade, residual tumor size, and debulking results, although p16 methylation seemed to occur more often in patients with advanced diseases or aggressive tumors. Compared to those without p16 methylation, patients with p16 methylation had significantly higher risk for disease progression (P = 0.01). The relative risk for progression was 1.69 (95% CI: 1.12 – 2.54), and the association remained statistically significant (RR = 1.54, 95% CI: 1.01 – 2.34) after adjusting for clinical and pathological variables. The risk for death was also higher in methylation positive patients than in methylation negative patients (RR = 1.33, 95% CI: 0.88 – 2.00), but the difference was not statistically significant. Conclusion. The study suggests that promoter methylation in the p16 gene is associated with ovarian cancer progression, and evaluation of p16 methylation may have values in predicting ovarian cancer prognosis. D 2004 Elsevier Inc. All rights reserved. Keywords: Methylation; Epithelial ovarian cancer; p16

Introduction Tumor suppressor gene p16 (INK4A/MTS1/CDKN2) encodes a 15.8-kDa protein that interacts with cyclin-dependent kinase CDK4 and CDK6. CDK4 and CDK6 bind to cylcin D regulating cell cycle progression from G1 to S phases. Binding of p16 to CDKs inhibits the formation of CDK/cyclin D complex, resulting in cell cycle arrest at G1 phase. The role of p16 in cell cycle control is believed to be involved in cancer development and progression [1 – 5]. * Corresponding author. Department of Epidemiology and Public Health, Yale Cancer Center, Yale University School of Medicine, PO Box 208034, New Haven, CT 06520-8034. Fax: +1-203-785-6980. E-mail address: [email protected] (H. Yu). 0090-8258/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2004.06.018

Loss of p16 expression has been observed in many types of cancer [2]. Genetic and epigenetic alterations, including homozygous deletion, point mutation, and promoter methylation, are considered to be responsible for the loss of p16 activity [5,6]. Promoter methylation has been recognized as an important mechanism in regulation of gene expression [7,8]. Aberrant methylation of p16 promoter has been found in many forms of cancer, including ovarian cancer [5]. Clinical studies indicate that p16 expression is undetectable in about a third of ovarian cancer cases [9– 12] and that patients with low p16 expression have poor responses to chemotherapy and unfavorable survival outcome [11,13]. Cell culture experiments demonstrate that reintroducing functional p16 into p16-null ovarian cancer cells results in inhibition of cell

686

D. Katsaros et al. / Gynecologic Oncology 94 (2004) 685–692

growth and increases in apoptosis, suggesting that p16 plays a role in ovarian cancer progression [14]. Epithelial ovarian cancer is one of the most lethal malignancies; 5-year survival of patients with the disease is only 50% [15,16]. To improve our understanding of ovarian cancer progression and to search for molecular markers for prediction of ovarian cancer prognosis, we analyzed p16 promoter methylation in epithelial ovarian cancer tissue and examined its association with patient progression-free and overall survival.

Patients and methods Study subjects Between October 1991 and February 2000, 270 patients underwent surgery for ovarian cancer in the Department of Gynecology, Gynecologic Oncology Unit, at University of Turin in Italy. Of the 270 patients, 18 had borderline tumors (histological grade 0) and 2 patients had metastatic cancer to the ovary. One patient did not have sufficient tissue specimen for analysis. The final number of patients included in this study was 249, all of who had primary epithelial ovarian cancer. The age of these patients ranged between 19 and 82 years, and the mean was 57. Follow-up information was available for 212 patients, and the last follow-up update was June 2001. The overall follow-up time was between 0.6 and 114 months, and the median was 31 months. During the course of follow-up, 94 patients developed progressive disease, and among these patients, 68 died; the median time of progression-free survival was 21 months. In addition to the 68 deaths, there were 26 patients who died of the disease without remission. Disease progression was considered when either measurable or assessable disease was detected after remission. Measurable disease was defined as bidimensional lesions with clearly defined margins by physical exam or radiologic imaging and with a diameter of at least 0.5 cm. Assessable disease was defined as nonmeasurable abnormalities on CT scan or physical examination and an elevated serum CA125. Patients with an elevated CA125 as the only manifestation of disease were also eligible provided that the CA125 was above 100 units. This cutoff represents more than a doubling of the normal value and is associated with a 98% positive predictive value of tumor progression [17]. Of the 249 patients, 56 (22%) had stage 1 disease, 12 (5%) had stage 2, 137 (55%) had stage 3, and 15 (6%) had stage 4. There were 29 (12%) patients whose stage information was missing. The disease stage was determined according to the International Federation of Gynecologists and Obstetricians [18]. As required by the FIGO staging scheme, extensive surgical and cytological assessment of the disease extent was performed. These procedures included collection of ascites or peritoneal washings for cytological evaluation, total abdominal hysterectomy and bilateral salpingo-oophorectomy, infracolic omentectomy and appen-

dectomy, selective pelvic and paraaortic lymphadenectomy, and debulking of all gross tissues. If macroscopic tumor was not present, biopsy of any lesion suspected of being a tumor metastasis or any adhesion adjacent to the primary tumor, blind biopsies of bladder peritoneum and cul-de-sac, paracolic gutter, and pelvic side walls, and biopsy or smear of right hemidiaphragm were performed. After initial surgical treatment, all patients received platinum-based chemotherapy according to a standard protocol, which included platinum and cyclophosphamide before 1997 and carboplatinum and paclitaxel after 1997. Surgical specimens were also examined for histological grade and type based on World Health Organization criteria [19]. Thirty-five patients (14%) had grade 1 (well differentiated) and 40 (16%) had grade 2 tumors; over half of the patients, 141 (57%), had poorly differentiated tumors (grade 3). Thirty-three (13%) patients had missing data on histological grade. Among the patients, the most common histological type was serous papillary tumor, 86 (35%) patients having this type. The rests were undifferentiated (15%), endometrioid (17%), mucinous (7%), clear cell (6%), mullerian (6%), and others (5%). Twenty-two patients (9%) had missing data on histological type. Measurement of p16 methylation Fresh tumor samples were collected from the patients during surgery. The specimens were snap frozen in liquid nitrogen immediately after surgical resection and then stored at 80jC until analysis. Representative samples from each tissue specimen were examined in frozen section by two pathologists to confirm tumor content; the content of tumor cells in these specimens ranged between 80% and 90%. Tumor tissues were pulverized into fine power manually in liquid nitrogen. Approximately 100 mg of tissue power were used for DNA extraction, and a standard protocol of phenol – chloroform extraction and ethanol precipitation was followed. After purification, genomic DNA was treated with sodium bisulfite. The treatment converts unmethylated cytosine to thymine while keeping methylated cytosine unchanged. The bisulfite treatment method has been described elsewhere [20]. Briefly, approximately 2 Ag genomic DNA was mixed with 2 M NaOH at a 10:1 ratio, and the sample was incubated at 37jC for 15 min. The denatured DNA sample was further mixed with 5 M bisulfite solution, pH 5.0, containing 0.1 M hydroquinone, and incubated at 55jC for 4 h. After incubation, the DNA samples were purified from the bisulfite solution using QIAEX II gel extraction kit (QIAGEN Genomics, WA 98021) and were concentrated with Pellet Paint Co-Precipitant (Novagen, WI 53711). Methylation-specific PCR (MSP) was performed on the sodium bisulfite-treated DNA samples to amplify the promoter region of the p16 gene. Two pairs of PCR primers were used in the amplification, one for methylated sequences and one for unmethylated sequences. The forward and

D. Katsaros et al. / Gynecologic Oncology 94 (2004) 685–692

reverse primers for the methylated product (125 bp) were 5VTGC GTT CGG CGG TTG CGG-3V and 5V-GAC CCC GAA CCG CGA CCG-3V. For the unmethylated product (129 bp), the primers were 5V-G TGT GTT TGG TGG TTG TGG AGA-3V and 5V-CCC AAC CCC AAA CCA CAA CCA TAA-3V, respectively. The bold letters indicated C to T transition caused by bisulfite treatment; the italic letters were cytosine in the CpG islands. The MSP primers were chosen under the following conditions: (1) they should flank the promoter region of the p16 gene, (2) they should cover as many CpG sites as possible, and (3) the methylated and unmethylated PCR products should cover the same region and with slightly different size in PCR product. The PCR reaction was carried out in a 25-Al solution, which included 2.5 Al 10 AmpliTaq Gold PCR buffer, 2.5 Al 10 mM dNTP mixture, 2.0 Al 50 mM MgCl2, 1.0 Al each of primers, 0.25 Al (5 U/Al) AmpliTaq Gold polymerase (Applied Biosystems, CA 94404), 2.0 Al DNA sample containing approximately 10 ng genomic DNA, and 13.75 Al autoclaved distilled water. The PCR protocol began with sample denaturing at 95jC for 11 min, followed by 35 cycles of denaturing at 95jC for 45 s, annealing at 64jC for 45 s, and extension at 72jC for 60 s, and one cycle of elongation at 72jC for 5 min. The PCR products were examined on 2% agarose gel stained with GelStar (FMC BioProducts, ME 04841). PCR amplification of methylated and unmethylated products was carried out separately in two tubes with the same PCR conditions and reagents (except the primers). All of the samples analyzed with MSP showed unmethylated PCR products, suggesting that unmethylated DNA was present in all the specimens. The samples were considered methylation positive only when methylated PCR products were shown in gel electrophoresis regardless the status of unmethylated PCR products. A picture of a sample gel is shown in Fig. 1. To confirm PCR results, representative PCR products were purified from gel using a commercial gel extraction kit (QIAGEN Genomics, WA 98021) and were sequenced with forward primers using the ABI 377 automated sequencer. Statistical analysis MSP results were analyzed as a dichotomous variable, the presence or absence of p16 methylation. The associations of p16 methylation with clinical and pathological variables were analyzed using the chi-square test. To

Fig. 1. Results of p16 methylation-specific PCR.

687

examine the associations of p16 methylation with disease progression and death, we calculated the hazards ratio (relatively risk) and its 95% confidence interval using the Cox proportional hazards regression model. The disease progression-free survival was determined based on the time interval between the date of surgery and date of the first disease progression or end of the follow-up. The overall survival was the time interval between the date of surgery and date of death or end of the follow-up. In multivariate analysis, the associations of p16 methylation with disease progression or death were calculated after adjusting for patient age at diagnosis, disease stage, histological grade, residual tumor size, and debulking results. The Kaplan – Meier survival curves were constructed to show the survival difference between patients with and without p16 methylation.

Results Of the 249 tumor samples analyzed, 100 were found to be positive for p16 promoter methylation, which accounted for 40% of the samples. Patient age at diagnosis was not significantly different between patients with and without p16 methylation. The mean age was 57.8 years for patients with methylation and 56.9 years for those without methylation (P = 0.576). Table 1 shows the distributions of p16 methylation in relation to clinical and pathological variables. Compared to patients without p16 methylation, patients with p16 methylation were more likely to have advanced disease (stage 3 or 4), 75.6% versus 64.6%, as well as to have poorly differentiated tumors (grade 3), 71.1% versus 61.1%, although the differences were not statistically significant (P = 0.084 and 0.190, respectively). There were no significant differences in debulking results and residual tumor size between the p16 groups (P = 0.300 and 0.351, respectively). However, tumor histology was significantly different between patients with and without p16 methylation. Patients with p16 methylation were more likely to have serous tumors or undifferentiated tumors, whereas patients without 16 methylation were more prone to clear cell tumor or other types (P = 0.033). More patients with p16 methylation had disease progression, 53.6% versus 38.3% (P = 0.028), but the death rate was not significantly different between these patients, 48.8% versus 41.4% (P = 0.289). Survival analysis showed that patients with methylated p16 promoter had a 1.65-fold increase in risk for disease progression (P = 0.012), and the increased risk was sustained after clinical and pathological features, including age at diagnosis, disease stage, histological grade, residual tumor size, and debulking results, were adjusted in the analysis (Table 2). In multivariate analysis, patients with methylated p16 still had an over 50% increase in risk for disease progression compared to patients with unmethylated p16. Patients with p16 methylation also had slightly elevated risk for death, RR = 1.33 (95% CI = 0.88 – 2.00), but the

688

D. Katsaros et al. / Gynecologic Oncology 94 (2004) 685–692

Table 1 Associations of p16 methylation with clinical and pathological variables Variable

Histological grade 1 2 3 Disease stage 1–2 3–4 Debulking result Optimal Suboptimal Residual tumor 0 1 – 4 cm 5+ cm Histological type Clear cell Endometrioid Mucinous Mullerian Undifferentiated Serous papillary Others Response to postoperative treatment Complete response Partial response No response Progressive disease Disease progression No Yes Death No Yes

p16 methylation ( )

p16 methylation (+)

P value

Number

Percent

Number

Percent

25 24 77

19.8 19.1 61.1

10 16 64

11.1 17.8 71.1

0.190

46 84

35.4 64.6

22 68

24.4 75.6

0.084

71 57

55.5 44.5

42 45

48.3 51.7

0.300

61 33 34

47.7 25.8 26.5

35 30 21

40.7 34.9 24.4

0.351

13 27 13 10 17 46 11

9.5 19.7 9.5 7.3 12.4 33.6 8.0

3 15 5 4 21 40 2

3.3 16.7 5.6 4.4 23.3 44.4 2.2

0.033

91 9 3 22

72.8 2.4 7.2 17.6

61 1 6 16

72.6 1.2 7.1 19.1

0.931

79 49

61.7 38.3

39 45

46.4 53.6

0.028

75 53

58.6 41.4

43 41

51.2 48.8

0.289

grade, we also constructed Kaplan– Meier survival curves in subgroups of patients stratified by their disease stage or histological grade. Fig. 3 showed the survival curves by the p16 methylation status either in patients with low-grade tumor (upper panel) or in patients with high-grade tumor (lower panel). Although, as expected, patients with highgrade tumor had shorter survival than patients with lowgrade tumor, the survival difference between two p16 groups was quite consistent in the subgroups, suggesting that the effect of p16 methylation on survival is independent from tumor grade. Similar findings were also suggested for stage, and the survival difference between the p16 groups was more evident in patients with early stage than in patients with late stage (Fig. 4). None of the subgroup analyses were statistically significant, probably due to the small sample size involved in the analysis.

Discussion

difference was not statistically significant (P = 0.177) and was disappeared when clinical and pathological variables were adjusted in the analysis (RR = 1.15, P = 0.526). In the univariate Cox regression model, we also examined the survival outcomes in association with clinical and pathological variables. As expected, advanced disease stage, high histological grade, and large residual tumor size were all associated with bad survival outcomes (data not shown). The risk for disease progression and death was increased with these variables in a dose – response manner. Patients who had suboptimal debulking results also experienced higher risk for disease progression and death compared to those who had optimal debulking results (data not shown). Fig. 2 shows that patients without p16 methylation had better progression-free survival (upper panel) as well as overall survival (lower panel) than patients with p16 methylation. The survival rate started to separate about a year after surgery for progression-free survival and 2 years for overall survival; the difference in progression-free survival was statistically significant. To examine if p16-associated survival difference was influenced by disease stage or tumor

In this study, we found p16 promoter methylation in 40% of epithelial ovarian cancer. Compared to those without p16 methylation, patients with p16 methylation had a 50% increase in risk for disease progression. Moreover, the increased risk for disease progression was independent from clinical and pathological factors, suggesting that p16 methylation may have values in predicting ovarian cancer prognosis. The study findings seem to support the notion that p16 methylation in ovarian cancer is common and important and that the status of methylation is associated with poor prognosis. These observations are in agreement with our understanding of p16’s role in cell cycle control and the inhibitory effect of DNA methylation on gene expression. Previous studies of tumor tissues have shown that p16 expression is absent in many types of cancer [5], and the lack of expression is not entirely due to genetic changes [21]; epigenetic modification, mainly CpG island methylation, is responsible for some of the silence [22,23]. Although loss of p16 expression is seen in a third of ovarian cancer cases, several previous studies have found little evidence that methylation is responsible for this phenomenon in ovarian cancer. Shih et al. [24] reported

Table 2 Associations of p16 methylation with the risk of disease progression or death Variable

Univariate analysis Methylation ( ) Methylation (+) Multivariate analysisa Methylation ( ) Methylation (+) a

Progression

Death

RR

95% CI

RR

95% CI

1.00 1.69

1.12 – 2.54

1.00 1.33

0.88 – 2.00

1.00 1.54

1.01 – 2.34

1.00 1.15

0.75 – 1.77

Adjusted for age at surgery, disease stage, histological grade, residual tumor size, and debulking result.

D. Katsaros et al. / Gynecologic Oncology 94 (2004) 685–692

689

Fig. 2. Comparison of Kaplan – Meier survival curves between patients with and without p16 methylation. Upper panel: disease progression-free survival. Lower panel: overall survival.

no p16 hypermethylation in 45 ovarian cancer cases, but the methylation analysis was done with the use of PCR and restriction enzyme digestion, a method similar to restriction fragment length polymorphism (RFLP). Similar results were also seen in several other studies using the same method [9,25,26]. Given the fact that only a limited number of CpG islands (usually one or two) are involved in the digestion, PCR-RFLP is apparently less sensitive than methylationspecific PCR (MSP). Published studies suggest that p16 methylation is detectable in ovarian cancer if MSP is used for analysis [27 –29]. With the use of MSP, McCluskey et al. [29] found p16 methylation in 5% (2/37) ovarian cancer cases, but this finding was based on an MSP method that amplified a 224-bp product in which 140 bp was in the promoter and 84 bp was in exon 1. The investigators mentioned in the report that if they amplified only the 140-bp product in the promoter region, the number of p16 methylation was much higher, 57% (21/37), a finding that was very similar to ours [29]. Another small study also reported 56% of p16 methylation in 18 advanced ovarian cancer samples [30].

The MSP method used in our study amplified a 125-bp methylated product and a 129-bp unmethylated product, both of which resided in the same region analyzed by McCluskey et al. [29] in their study. The MSP used in the McCluskey et al.’s study, initially developed by Herman et al. [22], amplified a 150-bp methylated product and a 151-bp unmethylated product. Comparing the two methods, our PCR products were very close to those of Herman et al.’s. Our forward primer was immediately after Herman et al.’s, with 1 bp overlapping. The reason for us to choose this site was to include more CpG sites in the primer sequences; our method had four CpG sites, while Herman et al.’s had three. The reverse primers of the two PCR were completely overlapping except that ours was 2 bp longer; the reverse primers had five CpG sites. Thus, both primers together covered a total of eight CpG sites in Herman et al.’s MSP and nine CpG sites in ours. Herman et al.’s PCR products had 1 bp difference between methylated and unmethylated sequences, whereas our products had 4 bp difference between the two sequences. CpG methylation in this region has been proven to be associated with the loss of p16 expression in experi-

690

D. Katsaros et al. / Gynecologic Oncology 94 (2004) 685–692

Fig. 3. Comparison of Kaplan – Meier disease progression-free survival curves between patients with and without p16 methylation. Upper panel: patients with grade 1 or 2 tumors. Lower panel: patients with grade 3 tumors.

mental studies [22,31 –33]. Given the similarity in methodology, detecting a similar rate of p16 methylation between the two methods should be expected. To further confirm our PCR results, we performed direct DNA sequencing on two methylated and two unmethylated PCR products. The sequencing results showed that the methylated PCR products contained a sequence of ‘‘AAA GGG AAA TAG TAG CGG GCG GCG GGG AAG TAG TAT GGA GTC GGC GGC GGG GAG TAG TAT GGA GTT TTC GGT TGA TTG GTT GGT TAC GGT CGC GGT TCG GGG TCA,’’ which was part of the p16 sequence located in the region flanked by the methylated PCR primers. The sequence of the unmethylated PCR products was ‘‘AGG TAG TGG GTG GTG GGG AGT AGT ATG GAG TNG GTG GTG GGG AGT AGT ATG GAG TTT TTG GTN GAT TGG TTG GTT,’’ which corresponded to the p16 sequence amplified by the unmethylated primers. Thus, the sequencing results indicate that our PCR procedure indeed amplifies the p16 gene. Based on the findings of our study as well as literature review, hypermethylation of p16 promoter seems to occur in epithelial ovarian cancer at relatively high frequency; using

less sensitive methods or selecting different DNA sequence affects the result of methylation detection, leading to inconsistent study findings. In our study, we found that p16 methylation tended to occur more often in advanced disease or aggressive tumor and was associated with disease progression. Although no clinical studies have addressed the role of p16 methylation in ovarian cancer prognosis, the possibility of p16 methylation being a prognostic marker has been suggested for other cancer sites. Wienche et al. [34] found that p16 methylation tended to occur more often in poorly differentiated colorectal cancer; Esteller et al. [35] reported that p16 methylation was associated with shorter survival of colon cancer. Several lines of evidence suggest that p16 plays an important role in ovarian cancer progression. Ovarian cancer cell line SKOV3 does not express p16 due to deletion of the gene. Transfecting the cells with adenovirus carrying wild-type p16 can reestablish the expression of p16, and the reexpression of p16 results in reduction in cell growth and increases in cell cycle arrest and apoptosis. The transfected cells also undergo substantial reduction in capacity of tumorigenesis in nude mice [36 – 38]. Mullerian inhibiting

D. Katsaros et al. / Gynecologic Oncology 94 (2004) 685–692

691

Fig. 4. Comparison of Kaplan – Meier disease progression-free survival curves between patients with and without p16 methylation. Upper panel: patients with stage 1 or 2 diseases. Lower panel: patients with stage 3 or 4 diseases.

substance (MIS), a member of the TGF-h family, is able to inhibit the growth of ovarian cancer cells, and the inhibitory effect of MIS on ovarian cancer is mediated through the upregulation of p16 that induces cell cycle arrest at G1 phase and triggers apoptosis [39]. Kudoh et al. [13] found that ovarian cancer patients with low expression of p16 did not respond well to chemotherapy and were associated with poor prognosis. Similar findings were also reported by another study [10]. However, in our study, we did not find patients responding differently to chemotherapy by their p16 methylation status, despite the fact that patient survival outcomes, especially disease progression, were different by p16 methylation status. In the study, we also found that p16 methylation was significantly different by histological types of the tumor. Patients with methylated p16 were more likely to have serous or undifferentiated tumors; this observation appeared to be consistent with the finding made by Havrilesky et al. [12], who found that loss of p16 expression was more common in serous than in nonserous tumors. In summary, hypermethylation of p16 promoter was detected in 40% of epithelial ovarian cancer using a methylation-specific PCR. Although methylation status did not

change substantially with patient age at diagnosis, disease stage, histological grade, residual tumor size, and surgical debulking results, patients with p16 methylation compared to those without p16 methylation had a 50% increase in risk for disease progression, and the increased risk was not altered by patient clinical and pathological features, suggesting that p16 methylation may play a role in ovarian cancer progression and the status of p16 methylation in tumor tissue may be used as a biomarker to predict ovarian cancer prognosis.

Acknowledgments Drs. D. Katsaros, S. Fracchioli, and I.A. Rigault de la Longrais are partially supported by AIRC (Italian Association for Cancer Research).

References [1] Shapiro GI, Rollins BJ. p16INK4A as a human tumor suppressor. Biochim Biophys Acta 1996 (Mar. 18);1242(3):165 – 9.

692

D. Katsaros et al. / Gynecologic Oncology 94 (2004) 685–692

[2] Liggett Jr WH, Sidransky D. Role of the p16 tumor suppressor gene in cancer. J Clin Oncol 1998 (Mar.);16(3):1197 – 206. [3] Serrano M. The INK4a/ARF locus in murine tumorigenesis. Carcinogenesis 2000 (May);21(5):865 – 9. [4] Serrano M, Hannon GJ, Beach D. A new regulatory motif in cellcycle control causing specific inhibition of cyclin D/CDK4. Nature 1993 (Dec. 16);366(6456):704 – 7. [5] Rocco JW, Sidransky D. p16(MTS-1/CDKN2/INK4a) in cancer progression. Exp Cell Res 2001 (Mar. 10);264(1):42 – 55. [6] Shapiro GI, Park JE, Edwards CD, Mao L, Merlo A, Sidransky D, et al. Multiple mechanisms of p16INK4A inactivation in non-small cell lung cancer cell lines. Cancer Res 1995 Dec 15;55(24):6200 – 9. [7] Singal R, Ginder GD. DNA methylation. Blood 1999 (Jun. 15);93(12): 4059 – 70. [8] Karpf AR, Jones DA. Reactivating the expression of methylation silenced genes in human cancer. Oncogene 2002 Aug 12;21 (35): 5496 – 503. [9] Marchini S, Codegoni AM, Bonazzi C, Chiari S, Broggini M. Absence of deletions but frequent loss of expression of p16INK4 in human ovarian tumours. Br J Cancer 1997;76(2):146 – 9. [10] Kusume T, Tsuda H, Kawabata M, Inoue T, Umesaki N, Suzuki T, et al. The p16-cyclin D1/CDK4-pRb pathway and clinical outcome in epithelial ovarian cancer. Clin Cancer Res 1999 (Dec.);5(12):4152 – 7. [11] Sui L, Dong Y, Ohno M, Goto M, Inohara T, Sugimoto K, et al. Inverse expression of Cdk4 and p16 in epithelial ovarian tumors. Gynecol Oncol 2000 (Nov.);79(2):230 – 7. [12] Havrilesky LJ, Alvarez AA, Whitaker RS, Marks JR, Berchuck A. Loss of expression of the p16 tumor suppressor gene is more frequent in advanced ovarian cancers lacking p53 mutations. Gynecol Oncol 2001 (Dec.);83(3):491 – 500. [13] Kudoh K, Ichikawa Y, Yoshida S, Hirai M, Kikuchi Y, Nagata I, et al. Inactivation of p16/CDKN2 and p15/MTS2 is associated with prognosis and response to chemotherapy in ovarian cancer. Int J Cancer 2002 (Jun. 1);99(4):579 – 82. [14] Wolf JK, Mills GB, Bazzet L, Bast Jr RC, Roth JA, Gershenson DM. Adenovirus-mediated p53 growth inhibition of ovarian cancer cells is independent of endogenous p53 status. Gynecol Oncol 1999 (Nov.);75(2):261 – 6. [15] Partridge EE, Barnes MN. Epithelial ovarian cancer: prevention, diagnosis, and treatment. CA Cancer J Clin 1999;49:297 – 320. [16] Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. Cancer Statistics, 2003 CA. Cancer J Clin 2003;53:5 – 26. [17] Rustin GJ, Nelstrop AE, Tuxen MK, Lambert HE. Defining progression of ovarian carcinoma during follow-up according to CA 125: a North Thames Ovary Group Study. Ann Oncol 1996 Apr;7(4):361 – 4. [18] International Federation of Gynecology and Obstetrics. Changes in definition of clinical staging for carcinoma of the cervix and ovary. Am J Obstet Gynecol 1987;56:263 – 4. [19] Serov SF, Soully RF, Sobin LH. Histological typing of ovarian tumors. International Histological Classification of Tumors. Geneva: World Health Organization; 1973. pp. 37 – 42. [20] Singal R, Ferris R, Little JA, Wang SZ, Ginder GD. Methylation of the minimal promoter of an embryonic globin gene silences transcription in primary erythroid cells. Proc Natl Acad Sci USA 1997 (Dec. 9); 94(25):13724 – 9. [21] Cairns P, Mao L, Merlo A, Lee DJ, Schwab D, Eby Y, et al. Rates of p16 (MTS1) mutations in primary tumors with 9p loss. Science 1994 (Jul. 15);265(5170):415 – 7. [22] Herman JG, Merlo A, Mao L, Lapidus RG, Issa JP, Davidson NE, et al. Inactivation of the CDKN2/p16/MTS1 gene is frequently associated with aberrant DNA methylation in all common human cancers. Cancer Res 1995 Oct 15;55(20):4525 – 30. [23] Gonzalez-Zulueta M, Bender CM, Yang AS, Nguyen T, Beart RW,

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37] [38]

[39]

Van Tornout JM, et al. Methylation of the 5V CpG island of the p16/ CDKN2 tumor suppressor gene in normal and transformed human tissues correlates with gene silencing. Cancer Res 1995 Oct 15; 55(20):4531 – 5. Shih YC, Kerr J, Liu J, Hurst T, Khoo SK, Ward B, et al. Rare mutations and no hypermethylation at the CDKN2A locus in epithelial ovarian tumours. Int J Cancer 1997 (Mar. 4);70(5):508 – 11. Ryan A, Al-Jehani RM, Mulligan KT, Jacobs IJ. No evidence exists for methylation inactivation of the p16 tumor suppressor gene in ovarian carcinogenesis. Gynecol Oncol 1998 (Jan.);68(1):14 – 7. Niederacher D, Yan HY, An HX, Bender HG, Beckmann MW. CDKN2A gene inactivation in epithelial sporadic ovarian cancer. Br J Cancer 1999 (Aug.);80(12):1920 – 6. Milde-Langosch K, Ocon E, Becker G, Loning T. p16/MTS1 inactivation in ovarian carcinomas: high frequency of reduced protein expression associated with hyper-methylation or mutation in endometrioid and mucinous tumors. Int J Cancer 1998 (Feb. 20); 79(1):61 – 5. Wong YF, Chung TK, Cheung TH, Nobori T, Yu AL, Yu J, et al. Methylation of p16INK4A in primary gynecologic malignancy. Cancer Lett 1999 Mar 1;136(2):231 – 5. McCluskey LL, Chen C, Delgadillo E, Felix JC, Muderspach LI, Dubeau L. Differences in p16 gene methylation and expression in benign and malignant ovarian tumors. Gynecol Oncol 1999 (Jan.); 72(1):87 – 92. Jimenez E, Garcia-Foncillas J, Salgado E, Okroujnov I, Angeletti S, Huarriz M, et al. Methylation status assay of p16 and p15 promoters by methylation-specific PCR in epithelial ovarian cancer. Proc Am Assoc Cancer Res 1999;40:2913. Merlo A, Herman JG, Mao L, Lee DJ, Gabrielson E, Burger PC, et al. 5V CpG island methylation is associated with transcriptional silencing of the tumour suppressor p16/CDKN2/MTS1 in human cancers. Nat Med 1995 (Jul.);1(7):686 – 92. Gonzalez-Zulueta M, Bender CM, Yang AS, Nguyen T, Beart RW, Van Tornout JM, et al. Methylation of the 5V CpG island of the p16/ CDKN2 tumor suppressor gene in normal and transformed human tissues correlates with gene silencing. Cancer Res 1995 (Oct. 15); 55(20):4531 – 5. Otterson GA, Khleif SN, Chen W, Coxon AB, Kaye FJ. CDKN2 gene silencing in lung cancer by DNA hypermethylation and kinetics of p16INK4 protein induction by 5-aza 2’deoxycytidine. Oncogene 1995 (Sep. 21);11(6):1211 – 6. Wiencke JK, Zheng S, Lafuente A, Lafuente MJ, Grudzen C, Wrensch MR, et al. Aberrant methylation of p16INK4a in anatomic and gender-specific subtypes of sporadic colorectal cancer. Cancer Epidemiol Biomarkers Prev 1999 (Jun.);8(6):501 – 6. Esteller M, Gonzalez S, Risques RA, Marcuello E, Mangues R, Germa JR, et al. K-ras and p16 aberrations confer poor prognosis in human colorectal cancer. J Clin Oncol 2001 (Jan. 15);19 (2):299 – 304. Modesitt SC, Ramirez P, Zu Z, Bodurka-Bevers D, Gershenson D, Wolf JK. In vitro and in vivo adenovirus-mediated p53 and p16 tumor suppressor therapy in ovarian cancer. Clin Cancer Res 2001 (Jun.);7(6):1765 – 72. Wang M, Wei J, Zhang J. Replacement of the p16 gene in human ovarian cancer cells. Chin Med J (Engl.) 2001 (Aug.);114(8):857 – 9. Kawakami Y, Hama S, Hiura M, Nogawa T, Chiba T, Yokoyama T, et al. Adenovirus-mediated p16 gene transfer changes the sensitivity to taxanes and Vinca alkaloids of human ovarian cancer cells. Anticancer Res 2001 (Jul. – Aug.);21(4A):2537 – 45. Ha TU, Segev DL, Barbie D, Masiakos PT, Tran TT, Dombkowski D, et al. Mullerian inhibiting substance inhibits ovarian cell growth through an Rb-independent mechanism. J Biol Chem 2000 (Nov. 24); 275(47):37101 – 9.