• ovary;
  • immunohistochemistry;
  • prognostic marker;
  • cyclins;
  • microarray analysis


  1. Top of page
  2. Abstract


Cyclins, cyclin dependent kinases (cdks), and their inhibitors act in combination to regulate progression through the cell cycle and often are dysregulated in carcinoma. The authors hypothesized that cyclin E plays an important role in ovarian carcinogenesis and that its overexpression may be an indicator of a poor prognosis.


Immunohistochemical analysis of cyclin E expression was performed by image analysis in normal ovaries, cystadenomas, tumors of low malignant potential, and 405 primary ovarian carcinomas by using tissue microarray technology.


Overexpression of cyclin E was found in 63.2% of the samples and was associated with clear cell, poorly differentiated, and serous carcinoma (P ≤ .001), high-grade tumors (P ≤ .001), late-stage disease (P = .002), age older than 60 years at the time of diagnosis (P = .04), and suboptimal cytoreduction (P = .001). A high percentage of cyclin E-expressing cells was associated with a poor outcome in univariate and in multivariate analyses. In addition, cyclin E levels also reduced survival in the late-stage disease group and in patients who underwent suboptimal debulking.


Cyclin E was identified as an independent prognostic factor in patients with ovarian carcinoma. The accumulation of cyclin E protein may be a late event in tumorigenesis and may contribute to disease progression in these patients. Cancer 2006. © 2006 American Cancer Society.

It is estimated that ovarian carcinoma is the fourth leading cause of cancer deaths among women in the United States, that 16,210 women in the United States will die from ovarian carcinoma in 2005, and that 22,220 women will have a diagnosis of the disease in 2005.1 Currently, > 50% of women who are diagnosed with ovarian carcinoma die within 5 years, and only 25% of ovarian carcinomas in the U.S. are diagnosed at an early stage. When the disease is diagnosed in advanced stages, the 5-year survival rate is only approximately 25%.1 Nevertheless, there is a broad variability of clinical behaviors, ranging from an excellent response to treatment and a high likelihood of cure to rapid progression, reflecting the heterogeneity of the tumor. Thus, the identification of new markers that can distinguish patients who are likely to have poor outcomes may provide patients with better therapeutic opportunities and a more customized antitumor therapy. In addition, these novel markers potentially may be useful as novel therapeutic targets.

Cell cycle deregulation is one of the hallmarks of carcinoma.2 Progression through the cell cycle is governed by cyclin-dependent kinases (cdks), which are regulated by phosphorylation, activated by binding to cyclins, and inhibited by cdk inhibitors. The coordinated expression of cyclins, cdks, and cdk inhibitors often is deregulated in carcinoma.3 The cyclin E/Cdk2 complex is among the main regulators of the G1/S transition. Several studies consistently have demonstrated that cyclin E is associated with disease progression in various malignancies and is associated with a poor prognosis in patients with breast, bladder, and colorectal carcinoma.4–6 Most of those studies analyzed ovarian carcinoma in only small numbers of patients.7–10

We hypothesized that cyclin E may play a critical role in ovarian carcinoma development and that its overexpression may be correlated with a poor prognosis. To test this hypothesis, we undertook the current study to assess the levels of cyclin E during ovarian carcinoma progression and to evaluate its prognostic significance by using tissue microarrays in a large cohort of patients with ovarian carcinoma.


  1. Top of page
  2. Abstract


Samples from women with primary epithelial ovarian carcinoma who had undergone initial surgery at The University of Texas M. D. Anderson Cancer Center between 1990 and 2001 were included in this study. In total, 441 patients were identified. Follow-up information was updated through June 2005 by reviewing medical records and the United States Social Security Index. Demographic and survival data were entered into a comprehensive data base that was created with Microsoft Access (version 97). Histopathologic diagnoses were based on World Health Organization criteria, tumor grading was based on Gynecologic Oncology Group criteria, and disease staging was assigned according to the International Federation of Gynecology and Obstetrics staging system.11–15 Grading of serous carcinomas was done by using a two-tier system (low-grade and high-grade) according to the new criteria proposed by Malpica et al.16 Disease-specific survival (overall survival) rates were calculated as the percentage of patients who survived with disease for a defined period and is reported as the time since diagnosis, and only deaths from the disease were counted. The time to progression (disease-free survival) was defined as the interval from the date of initiation of treatment until disease progression, as evidenced clinically by an increase ≥ 25% in tumor size, the appearance of new lesions, or an increase in the serum CA125 level to more than twice the upper limit of normal.17, 18 The mean time to progression was calculated as the average time from diagnosis until the date of recurrence. To analyze response to primary therapy, patients were grouped as responders or nonresponders to the primary therapy. The later group also was subdivided according to whether patients had progressive disease or recurrent disease. Progressive disease was defined as no disease remission observed after the initiation of treatment. Recurrent disease was defined as disease remission that was documented clinically but was not sustained. In addition, normal ovarian tissues from 5 women without ovarian carcinoma, 5 serous cysts, 5 mucinous cysts, 10 mucinous low-malignant-potential (LMP) tumors, and 10 serous LMP tumors also were studied. The extent of cytoreduction was defined as optimal if the greatest dimension of any residual lesion after surgery was < 1 cm or as suboptimal if residual disease was ≥ 1 cm.19, 20 The use of tissue blocks and chart reviews were approved by the Institutional Review Board of The University of Texas M. D. Anderson Cancer Center.

Construction of the Tissue Microarrays

Tissue blocks were stored under ambient conditions at approximately 24 °C. Hematoxylin and eosin-stained sections were reviewed by a pathologist who selected representative areas of the tumors from which to acquire cores for microarray analysis. Tissue microarray blocks were constructed by taking core samples from morphologically representative areas of paraffin embedded tumor tissues and assembling them on a recipient paraffin block.21 This was done with a precision instrument (Beecher Instruments, Silver Spring, MD) that uses two separate core needles for punching the donor and recipient blocks and a micrometer-precise coordinate system for assembling tissue samples on a block. For each sample, 2 replicate, 1-mm core-diameter samples were collected, and each was placed on a separate recipient block. The final tissue microarray consisted of 6 blocks: The first pair (Blocks 1a and 1b) contained duplicates of 158 spots, the second pair (Blocks 2a and 2b) contained duplicates of 164 spots, and the third pair (Blocks 3a and 3b) contained duplicates of 119 spots. All samples were spaced 0.5 mm apart. Five-micrometer sections were obtained from the microarray and stained with hematoxylin and eosin to confirm the presence of tumor and to assess tumor histology. Tumor samples were arranged randomly on the blocks. An additional tissue microarray block was constructed to array the 5 normal ovarian tissue samples, the 5 serous cysts, the 5 mucinous cysts, the 10 serous LMP tumors, the 10 mucinous LMP tumors, the 6 low-grade serous carcinomas, and the 10 high-grade serous carcinomas.

Sample tracking was based on coordinate positions for each tissue spot in the tissue microarray block: The spots were transferred onto tissue microarray slides for staining. This sample tracking system was linked to a Microsoft Access data base that contained demographic, clinicopathologic, and survival data on the patients who provided the samples, thereby allowing rapid links between histologic data and clinical features. The array was read according to the given tissue microarray map; each core was scored individually, and the results were presented as the mean of the two replicate core samples. Samples in which no tumor was found or no cores were available were excluded from the final data analysis.

Immunohistochemical Analysis

The tissue microarray slides were subjected to immunohistochemical staining as follows: After initial deparaffinization, endogenous peroxidase activity was blocked by using 0.3% hydrogen peroxide. Deparaffinized sections were microwaved in 10 mM citrate buffer, pH 6.0, to unmask the epitopes. Then, the slides were incubated for 1 hour at room temperature with a monoclonal antibody against cyclin E (1:200, clone HE12; Santa Cruz Biotechnologies, Santa Cruz, CA), next with a biotin-labeled secondary antibody for 20 minutes, and finally with a 1:40 solution of streptavidin:peroxidase for 20 minutes. Tissues were then stained for 5 minutes with 0.05% 3′,3-diaminobenzidine tetrahydrochloride that had been prepared freshly in 0.05 M Tris buffer, pH 7.6, containing 0.024% hydrogen peroxide. Then, tissues were counterstained with hematoxylin, dehydrated, and mounted. All dilutions of the antibody, biotin-labeled secondary antibody, and streptavidin-peroxidase were made in phosphate buffered saline, pH 7.4, that contained 1% bovine serum albumin. Breast carcinoma was used as a positive control. Negative controls were made by replacing the primary antibody with phosphate buffered saline. All controls yielded satisfactory results.

Immunostainings for cyclin E were analyzed by computerized image analysis (Ariol SL-50; Applied Imaging, San Jose, CA). Samples with > 10% stained nuclei were considered positive.8, 22 The mean of the results from the two replicate core samples from each tumor specimen was considered for each patient.

Statistical Analysis

Differences in proportions were evaluated by chi-square analysis. Kruskal-Wallis and Mann-Whitney tests were used to compare multiple independent samples on the tissue microarray block that contained normal ovarian tissues and the various ovarian tumors. Disease survival rates were calculated by using the Kaplan–Meier method and were compared by using the log-rank test. Cox proportional hazards regression models were used for multivariate analysis of survival, and the level of cyclin was considered as a continuous variable. Results were considered statistically significant at the P < .05 level. Statistica software (Statsoft, Inc., Tulsa, OK) was used for the statistical analysis.


  1. Top of page
  2. Abstract

Patient Characteristics

Of the 441 samples that were arrayed originally, 405 samples were scorable after applying the selection criteria. The mean age of the 405 patients was 59 years (range, 20-96 yrs). The mean follow-up interval was 64 months (range, 1-120 mos), and the overall survival rate at 5 years was 38%.

Comparison of Cyclin E Expression in Ovarian Tumors

The average number of cells analyzed for each tumor was 1733 cells (range, 644-4765 cells). Higher levels of cyclin E expression were found in high-grade than in low-grade serous carcinomas (P ≤ .001; Mann-Whitney test) and in serous LMP compared with benign cysts or normal ovarian surface epithelial cells (P = .004; Mann-Whitney test). No difference in expression was observed between benign cysts and normal ovarian surface epithelium. Similar findings were observed for endometrioid and mucinous lesions. Clear cell carcinomas showed more frequent overexpression of cyclin E (Figs. 1, 2).

thumbnail image

Figure 1. Negative expression of cyclin E is observed in normal surface epithelial cells (A) and in a mucinous cystadenoma (B). Positive nuclear expression is observed in a mucinous tumor of low malignant potential (C), in a serous tumor of low malignant potential (D), in a low-grade serous carcinoma (E), and in a high-grade serous carcinoma (F). (G) Negative immunostaining is observed in a high-grade serous carcinoma. Original magnification, × 40 (F); × 10 (G).

Download figure to PowerPoint

thumbnail image

Figure 2. This chart illustrates the results from a quantitative analysis of cyclin E expression in normal ovarian surface epithelial cells (OSE), cystadenomas, tumors with low malignancy potential (LMP), and low-grade and high-grade tumors (Kruskal-Wallis test; P ≤ 0.001). Min: minimum; max; maximum; NS: not significant.

Download figure to PowerPoint

Association between Expression of Cyclin E and Clinicopathologic Variables

The results of immunostaining of the microarrays, organized according to clinicopathologic characteristics of the patients, are shown in Table 1. The expression of cyclin E was completely negative in 62 patients (15.3%), 1% to 10% in 93 patients (23%), 11% to 50% in 213 patients (52.6%), and > 50% in 37 patients (9.1%). A high percentage of nuclear expression (> 10%) was associated with clear cell, poorly differentiated, and serous carcinoma (P ≤ .001); high-grade tumors (P ≤ 0.001); late-stage disease (P = .002); age older than 60 years at the time of diagnosis (P = .04); and suboptimal level of cytoreduction (P = .001).

Table 1. Correlations between the Expression of Cyclin E and Clinicopathologic Parameters
CharacteristicCyclin E Expression: No. of Patients (%)
< 10% No. (%)> 10% No. (%)Total No.Mean Expression, %P*
  • MMMT: malignant mixed Mullerian tumor; FIGO: International Federation of Gynecology and Obstetrics.

  • *

    P values were calculated by using a chi-square test of independence.

Histologic type    ≤ .001
 Serous carcinoma97 (30)218 (70)31522.9 
 Endometrioid carcinoma32 (82)7 (18)398.9a 
 Mucinous carcinoma7 (88)1 (12)87.2 
 Clear-cell carcinoma4 (24)13 (76)1719.9 
 MMMT4 (44)5 (56)917.7 
 Poorly differentiated carcinoma2 (20)8 (80)1022.0 
 Transitional cell carcinoma34727.8 
Tumor grade    ≤ .001
 Low10 (45)21 (55)2216.5 
 Intermediate16 (94)1 (6)175.0 
 High123 (32)243 (68)36622.0 
FIGO disease stage    .002
 Stage I19 (56)15 (44)3413.0 
 Stage II18 (60)12 (40)3015.0 
 Stage III89 (33)180 (67)26923.2 
 Stage IV23 (32)49 (68)7220.5 
Age    .04
 < 59 y67 (43)88 (57)15519.4 
 ≥ 60 y82 (33)168 (67)25022.0 
Extent of cytoreduction    .001
 Optimal90 (45)111 (55)20118.2 
 Suboptimal59 (29)145 (71)20424.0

The correlation of cyclin E with response to primary therapy is shown in Table 2. In total, 316 patients (78%) received postsurgical cisplatin-based treatment, either alone or in combination with other adjuvant drugs. In 37 patients (9%), cisplatin-based treatment was administered before surgical debulking surgery. Nine patients (2%) received other forms of treatment (topotecan or paclitaxel alone). In 43 patients, the treatment protocol was unknown. Overall, higher levels of expression of cyclin E (> 10%) were observed in the nonresponding group than in responders to primary therapy (P < .001). Similar proportions of cyclin E expression were observed in the postsurgical cisplatin-based treatment subgroup (P < .001), but not in the presurgical cisplatin-based treatment subgroup (P = .7).

Table 2. Correlation of Cyclin E Expression and Response to Primary Therapy
Response to primary therapy*Cyclin E expression: No. of patients
< 10%> 10%Total
  • *

    P values were calculated by using a chi-square test of independence (response to primary therapy, P < .001).

  • Cisplatin-based postsurgery subgroup (P < .001).

  • Cisplatin-based presurgery subgroup (P = .7).

Unknown response31114
 Cisplatin-based regimens   
 Other regimens123
 Unknown regimen11314
Nonresponders-progressive disease   
 Cisplatin-based regimens   
 Other regimens112
 Unknown regimen246
Nonresponders-recurrent disease   
 Cisplatin-based regimens   
 Other regimens134
 Unknown regimen639

Association of Cyclin E with Overall Survival and Progression-Free Survival

Overall ovarian carcinoma-specific survival rates at 2 years and 5 years are shown in relation to the expression of cyclin E in Table 3. At the time of this report, of 405 analyzed patients, 65 patients were alive without clinical evidence of ovarian carcinoma, 71 were alive with ovarian carcinoma, 235 had died from ovarian carcinoma, 8 had died of unrelated causes, and 26 had been lost follow-up. A significant association between the expression of cyclin E and disease-specific survival (P ≤ .001) was observed (Fig. 3A). Patients who had tumors with < 10% positive nuclei had better survival in terms of both rate and duration compared with patients who had tumors with > 10% positive nuclei (P ≤ .001). The expression of cyclin E also reduced overall survival (P ≤ .001) when only patients with late-stage disease were analyzed (Fig. 3B) and among the patients who underwent suboptimal debulking (P ≤ .001) (Fig. 3C). In a univariate analysis of the time to disease progression (Fig. 3D), patients who had tumors that expressed lower levels of cyclin E (< 10%) lived longer without recurrence of the disease compared with patients who had higher expression levels (P < .001). In the multivariate analysis, the expression of cyclin E was a statistically significant variable when it was assessed with all tumor stages (P ≤ .001) and in patients with late-stage disease (P < .001) (Table 4).

Table 3. Cyclin E Expression and Overall Disease-Specific Survival
Cyclin E expressionNo. of PatientsMean Survival, mosSurvival rate, %P*
  • *

    P values were derived from a Kaplan–Meier analysis.

< 10%14956.284.267.0 
> 10%25636.564.925.0≤ .001
thumbnail image

Figure 3. These graphs illustrate the association of cyclin E expression with overall survival in all patients with ovarian carcinoma (n = 405 women) (A), in patients with late-stage disease (n = 341 women) (B), and in patients who underwent suboptimal debulking (n = 204 women) (C). (D) Progression-free survival is illustrated for all patients with ovarian carcinoma (n = 405 women). P values were derived from a Kaplan–Meier analysis.

Download figure to PowerPoint

Table 4. Univariate and Multivariate Analyses of Factors Predictive of Death from Disease and in Patients with Late-Stage Ovarian Carcinoma
Outcomes and Variables*No. of EventsUnivariate AnalysisMultivariate Analysis
HR (95% CI)PHR (95% CI)P
  • HR: hazard ratio; 95% CI: 95% confidence interval.

  • *

    Patient age and cyclin E expression levels were considered continuous variables. All others were evaluated as dichotomous variables, as follows: tumor grade: Grade 1 vs. Grade 2-3; disease stage: early-stage disease vs. late-stage disease; and extent of cytoreduction: optimal vs. suboptimal.

  • P values were derived from a Cox proportional-hazards model.

All patients     
 Tumor grade4052.3 (1.1-4.6).011.2 (0.6-2.5).54
 Disease stage4054.2 (2.6-6.9)≤ .0013.1 (1.8-5.2)≤ .001
 Age4051.5 (1.1-1.9).0011.3 (1.0-1.7).04
 Extent of cytoreduction3682.06 (0.7-1.4)< .0011.8 (1.4-2.4)≤.001
 Cyclin E expression4053.1 (2.2-4.6)≤ .0012.4 (1.8-3.3)≤.001
Patients with late-stage disease     
 Tumor grade3412.2 (0.9-4.9).051.3 (0.5-2.9).53
 Age3411.3 (1.0-1.7).021.2 (0.9-1.6).09
 Extent of cytoreduction3282.1 (1.5-2.8)≤ .0011.8 (1.4-2.4)≤ .001
 Cyclin E expression3412.7 (1.9-3.8)≤ .0012.5 (1.8-3.5)≤ .001


  1. Top of page
  2. Abstract

In the current study of 405 well characterized patients with long-term follow up, high levels of cyclin E expression were observed in 63.2% of the ovarian carcinoma samples, which correlated with serous, clear cell, and poorly differentiated carcinoma; higher tumor grade; late-stage disease; patient age older than 60 years at the time of diagnosis; and suboptimal cytoreduction. High percentages of cells that expressed cyclin E were associated with poor survival in both univariate and multivariate analyses of all factors that influenced survival. Furthermore, we showed that patients who had high levels of cyclin E (> 10%) had earlier recurrences compared with patients who had low levels of cyclin E.

Abnormal expression of cyclin E has been reported for different types of carcinoma and is associated with a poor prognosis in patients with breast,4 stomach,23 and colorectal carcinoma.6 A few studies also have evaluated the prognostic significance of cyclin E in patients with ovarian carcinoma. In a recent study of 139 suboptimally debulked patients who had advanced ovarian carcinoma by the Gynecologic Oncology Group,8 high levels of cyclin E were an independent factor that predicted a poor prognosis. Sui et al.9 observed that increased cyclin E expression was correlated with tumor progression and a poor prognosis in patients with ovarian carcinoma. In another study that involved 93 patients with primary ovarian carcinoma, only a trend toward an association with a poor prognosis was observed.7 In contrast, other studies have found no correlation between cyclin E expression and survival in patients with ovarian carcinoma.24, 25 The selected nature of the population, the small number of patients, and differences in quantification methods may account for these differences.

Six splice variants were described recently in addition to the typical, full-length cyclin E protein (50 kDa). These 6 low-molecular-weight (LMW) isoforms (ranging in size from 34 kDa to 49 kDa) are generated only in tumor cells and not in normal tissue by proteolysis and further posttranslational modifications.26–28 These novel LMW isoforms are hyperactive both biochemically and biologically compared with the full-length cyclin E, inducing cells to enter the cell cycle.29 In ovarian carcinoma, cyclin E is overexpressed primarily in the LMW forms.30 These isoforms of cyclin E are detected by most antibodies directed against cyclin E, because the antibodies bind to the carboxy terminus of cyclin E, which remains unaltered during the posttranscriptional modification.28, 31 Therefore, the expression of the LMW isoforms can be detected with most immunoperoxidase techniques, such as the technique used in the current study.

Cyclin E gene amplification and protein accumulation are late events according to some studies32, 33 but have been observed early in the progression to malignancy by other researchers.34, 35 In addition, early reports showed that prolonged cyclin E expression produced chromosome instability.36 Whether cyclin E is an early or late event in ovarian carcinogenesis remains a matter of controversy. A clear progression model has not been described, although a dualistic model has been proposed recently.37, 38 In this study, we found that nuclear expression of cyclin E increased progressively in ovarian tumors compared with normal ovarian surface epithelial cells. Moreover, the high correlation with high tumor grade and late disease stage suggests that an increase in cyclin E levels may be either a late event in tumorigenesis or a contributing factor to disease progression in patients with ovarian carcinoma. In summary, the current results demonstrate that the level of cyclin E protein expression is increased progressively during tumor progression and clearly is correlated with aggressive tumor behavior and that cyclin E expression is an independent predictor of a poor prognosis in patients with ovarian carcinoma.


  1. Top of page
  2. Abstract