Cyclooxygenase-2 expression predicts recurrence of cervical dysplasia following loop electrosurgical excision procedure

Gynecologic Oncology 92 (2004) 596 – 602 www.elsevier.com/locate/ygyno

Cyclooxygenase-2 expression predicts recurrence of cervical dysplasia $ following loop electrosurgical excision procedure John Farley, a,* Catherine Uyehara, b Glenn Hashiro, b Christina Belnap, c Michael Birrer, d and Eric Salminen a a

Department of Obstetrics and Gynecology, Tripler Army Medical Center, TAMC, HI 96859-5000, USA b Department of Clinical Investigation, Tripler Army Medical Center, TAMC, HI 96859-5000, USA c Department of Pathology, Tripler Army Medical Center, TAMC, HI 96859-5000, USA d Cell and Cancer Biology Department, Medicine Branch, Division of Clinical Sciences, National Cancer Institute, Rockville, MD 20850, USA Received 4 April 2003

Abstract Objective. To evaluate the expression of Cox-2 protein by immunohistochemistry in cervical dysplasias, and to determine any relationship to clinical factors such as degree or recurrence of dysplasia. Methods. Immunohistochemical expression of p27 and Cox-2 was initially examined in 62 cervical LEEP specimens, which spanned the histologic spectrum from benign to severe dysplasia. Histology and cytology from colposcopic follow-up exams were reviewed for 1 year after LEEP procedure. Primary outcome variable was recurrent dysplasia, either cytologic or histologic. Statistical analysis utilizing chisquare test for trend and Fisher’s Exact tests were performed to determine relative risk of recurrent dysplasia. Results. A total of 62 LEEP specimens were examined by immunohistochemistry (IHC). This included 18 mild, 19 moderate, and 25 severely dysplastic LEEP specimens. The percentage of tumor cells in each specimen that stained for p27 protein or Cox-2 enzyme was documented. A specimen was considered positive for p27 or Cox-2 if 50% or more of the cells in a specimen were stained: 94% of mild, 89% of moderate, and 44% of severe dysplasias stained positive for p27; 50% of mild, 42% of moderate, and 68% of severe dysplasia specimens stained positive for Cox-2. The average intensity of Cox-2 staining increased with severity of dysplasia—1.6 for mild, 1.8 for moderate, and 2.1 for severe dysplasia. There was a significant increase in both Cox-2 and p27 staining when severely dysplastic specimens were compared to mild and moderate dysplasia ( P < 0.001). Of the 35 specimens that stained positive for Cox-2 protein, 59% of these specimens had positive Cox-2 staining that extended to the margins of the LEEP resection specimen. The average length of Cox-2 protein staining beyond the histologic dysplasia was 1.64 mm. Positive margin status for Cox-2 was a significant independent risk factor for persistent and recurrent dysplasia, RR 1.68 95% CI (1.07 < RR < 2.65), P < 0.027. Conclusion. Cox-2 and p27 protein expression could be involved in squamous cervical cancer carcinogenesis. Cox-2 staining is often found outside the dysplastic lesion and this factor is associated with an increased risk of persistent and recurrent dysplasia following LEEP procedure. Should the histologic margin of LEEP resection approach 2.0 mm, follow-up algorithms for these patients should include intensive surveillance to ensure adequate treatment of disease. Published by Elsevier Inc. Keywords: Cyclooxygenase-2; LEEP; Cervical dysplasia

Abbreviations: LEEP, loop electrosurgical excision procedure; Cox-2, cyclooxygenase-2; IHC, immunohistochemistry; NS398, N-(2-cyclohexyloxy-4-nitrophenyl)methanesulfonamide. $ The views expressed herein are those of the authors and do not reflect the official policy or opinion of the Department of Defense or the United States Army. * Corresponding author. Department of Obstetrics and Gynecology, Tripler Army Medical Center, 1 Jarrett White Road, TAMC, HI 968595000. Fax: +1-808-433-1552. E-mail address: [email protected] (J. Farley). 0090-8258/$ - see front matter. Published by Elsevier Inc. doi:10.1016/j.ygyno.2003.10.052

Introduction In 1985, office-based private practitioners estimated that there were 224,900 visits for condyloma acuminatum [1]. Additionally, the annual incidence of cervical dysplasia in the United States is increasing, and the range of cases might be 500,000 – 1,000,000 [2]. Currently, 1.5– 6% of all pap smears in the United States will be diagnosed with a cervical cancer precursor [3,4]. These precursors can prog-

J. Farley et al. / Gynecologic Oncology 92 (2004) 596–602

ress to invasive cervical cancer if left untreated. In 2003, we can expect 12,200 new cases of invasive cervical cancers in the United States and approximately 4100 cervical cancer-associated deaths [5]. Cervical cancer carcinogenesis has mainly focused on the role of sexually transmitted diseases [6– 8]. Specifically, human papilloma virus (HPV) is suspected to play a role in cervical cancer carcinogenesis [6– 8]. Human papilloma virus has been detected by hybrid capture technique in up to 30– 43% of women in the United States and Great Britain [9]. Fifteen percent to 25% of women who have normal cervical cytology but harbor HPV will develop cervical dysplasia over 1 –2 years [10,11]. The estimated cost to treat only one small referral base population to the University of Alabama approached 1.3 million dollars [12]. To a lesser extent, abnormal expression of certain cellular proteins or mutation of protooncogenes has been investigated in the pathogenesis of cervical cancer. Specific oncogenes and cell cycle regulators thought to play a role in cervical cancer carcinogenesis and prognosis include p53, EGFR, c-erb-2/neu, p27 and c-myc genes [8]. Cyclooxygenase-2 (Cox-2) is the rate-limiting enzyme involved in the conversion of arachidonic acid to prostaglandin H2, the precursor of a variety of molecules [13]. Recent studies have evaluated the relevance of Cox-2 in human carcinogenesis. The expression of Cox-2 protein and mRNA is increased in both esophageal and gastric precancerous dysplastic precursors [13,14]. Increased levels of Cox-2 expression have been reported in carcinomas of the colon, stomach, breast, esophagus, and cervix [13 – 19]. Cox-2 expression has been independently associated with poor outcome in a variety of cancers to include cervical and gastric cancers [14,17]. Cox-2 protein expression in cervical cancer specimens appears to ascribe to both chemotherapeutic and radiotherapeutic resistance [20 – 22]. P27 is a cyclin-dependent kinase inhibitor, which has been found to regulate neoplastic transformation of the endocervix, and may be involved in growth regulation of squamous epithelium of the cervix [23 24]. The importance of p27 in regulating the G1/S transition has made it an ideal target for evaluation of altered expression in a variety of cancers. Despite the absence of mutations in the p27 gene, expression levels of p27 have been found to correlate with prognosis in colon and breast cancer [25]. In squamous carcinomas of the head and neck, G1 growth arrest induced by Iressa was associated with an increase in both p27 and p21 [26]. When Kudo established a new oral squamous cell carcinoma cell line from clinically metastatic cancer sites, he correlated the degree of invasiveness of this cell line both to the ability of the cell line to express p27, and also the activity of p27 degradation [27]. Previous reports have found p27 protein expression to be conserved in neoplastic transformation of the cervix with 90% of preinvasive lesions overexpressing p27 by immunohistochemistry [28,29]. Upon invasion, however, the level of p27 protein expression significantly decreases [28,29].

597

Although Cox-2 is highly involved in prostaglandin synthesis, the modulating affects of COX-2 on cellular proliferation are not related to prostaglandin synthesis [30]. Data suggest that COX-2 inhibits mesangial cell proliferation by a novel mechanism that is independent of prostaglandin synthesis, but involves p53, p21, and p27 [30]. Because of the importance of Cox-2 and p27 expression in human cancer and their apparent relevance in human cancer carcinogenesis, it was the purpose of this study to elucidate the role of cyclooxygenase-2 and p27 in cervical cancer carcinogenesis. To do this, we examined the expression of Cox-2 and p27 protein by immunohistochemistry (IHC) in cervical specimens, which span the histologic spectrum from normal to severe squamous dysplasia.

Materials and methods Tissue preparation and histologic evaluation Archived, paraffin-embedded tissue blocks were obtained from the Department of Pathology at Tripler Army Medical Center according to a protocol approved by the Tripler Army Medical Center Institutional Review Board. A retrospective analysis was performed of all pathology records for LEEPs performed at our institution from January 1996 through July 1998. Routine indications for LEEP at our institution included high-grade dysplasia, positive endocervical curettage (ECC), significant discordance between cytology and histology, and persistent low-grade lesions. Specimens were identified and a representative sample of normal, mild, moderate, and severe dysplastic specimens were obtained. Serial 5-Am sections of the tumor were obtained from each tissue block, with the first slide being stained for H&E to confirm pathologic diagnosis, and the subsequent slides stained for Cox-2 or p27. Patients were examined by colposcopy every 2 – 6 months for 1 year following LEEP procedure to evaluate for recurrence. During the colposcopic exam, cervical cytology was always obtained and cervical biopsies were performed if clinically indicated. Diagnosis of recurrent dysplasia was based on cytologic or histologic diagnosis of mild, moderate, or severe dysplasia and confirmed by a second blinded observer (C.B.). Immunohistochemistry Detection of cyclooxygenase-2 (Cox-2) utilizing a mouse monoclonal antibody C22420 (Transduction Laboratories, Lexington, KY) or p27 mouse monoclonal antibody K25020 (Transduction Laboratories) was performed as previously described [31]. Briefly, before immunostaining, the deparaffinized tissue sections were microwaved in 0.01 M citrate buffer (pH 6.0) for a total of 15 min. The specimens were incubated overnight at 4jC with primary antibody, at a dilution of 1:1000 for p27, and 1:100 for Cox-

598

J. Farley et al. / Gynecologic Oncology 92 (2004) 596–602

2. Detection of primary antibody was performed using the appropriate biotin – streptavidin – peroxidase detection system (Vectastain Elite ABC kit, Vector Laboratories, Burlingame, CA), and stained with primary stain, Very Intense Purple (Vector VIP Substrate Kit, cat. no. SK-4600, Vector Laboratories), and counterstain, Methyl Green (cat. no. H3402, Vector Laboratories) as a substrate. Background staining of cervical stroma and normal epithelium was evaluated as a negative control for Cox-2, and a positive control for p27 staining, and found to be negligible for Cox2 staining and diffusely positive for p27 staining in all specimens analyzed (Fig. 1). Endocervical glandular staining for Cox-2 was used as a positive control and found to be present in all specimens evaluated. Immunohistochemical staining was evaluated by visual counting of the cells from at least 10 random high-powered fields (100) with a minimum requirement of 1000 cells counted. Specimens were considered positive staining if greater than 50% of counted cells stained positive for the respective antibody. Also, intensity of Cox-2 staining was scored from 0 to 3+ using endocervical glandular staining as a baseline. Traditional semiquantitative scoring (H score) was also performed for Cox-2 expression by multiplying the percentage of positive cells by the intensity [32]. The observer (J.F.) was blinded as to histology of specimen,

degree of dysplasia, and recurrence of dysplasia. Random specimens were selected for confirmation of analysis by a second blinded observer (C.B.). Statistical analysis Antibody staining was read as positive or negative based on immunohistochemical staining. Those samples with greater than 50% staining were considered positive, and those with less than 50% staining were considered negative. Given an initial projection of 72% incidence of positive antibody staining in the specimens, it was estimated that a sample size of 80 (20 specimens each for mild moderate, severe dysplasia, and cancer) would give 80% power to detect a 20% difference in Cox-2 protein expression ( P < 0.05). Chi-square tests of association were used to compare the occurrence of antibody staining expression and grades of cervical dysplasia.

Results One objective of our study was to evaluate p27 expression in cervical dysplasia spanning the spectrum from mild to severe dysplasia. A total of 62 LEEP specimens were

Fig. 1. Immunohistochemical expression of cyclooxygenase-2 in dysplastic cervical tissues. Photomicrograph of representative H&E and Cox-2 staining in benign cervical epithelium (A and B, 600), mild (C and D, 600), moderate (E and F, 600), severe dysplastic cervical tissues (G and H, 600) and positive Cox-2 staining at the margin of a excision specimen in benign epithelium (I and J, 600).

J. Farley et al. / Gynecologic Oncology 92 (2004) 596–602

examined by immunohistochemistry (IHC). This included 18 mild, 19 moderate, and 25 severely dysplastic LEEP specimens. In normal specimens, the mean percentage of normal tissue cells staining positive for p27 was 97%. The mean percentages of tumor cells staining positive for p27 protein in the dysplastic specimens varied from 87% in mild to 46% in severe dysplasias. Ninety-four percent of mild, 89% moderate, and 44% severe dysplasias were considered positive (>50% of cells staining) for p27 protein. There was a significant decrease in both average p27 protein expression and percentage of positive specimens staining for p27 protein when severely dysplastic specimens were compared to mild and moderate dysplasias ( P < 0.001). There was no relationship between p27 staining and recurrent dysplasia. Cox-2 protein expression and intensity of staining increased with the severity of dysplasia. The mean percentage of normal tissue cells staining for Cox-2 was 10%. The mean percentage of tumor cells staining positive for Cox-2 protein was 28%, 37%, and 67% for mild, moderate, and severe dysplasia, respectively (Fig. 1). The average intensity of Cox-2 staining also increased with severity of dysplasia. The intensity increased from 1.6 for mild to 2.1 for severe dysplasia (Table 1). Fifty percent of mild, 42% moderate, and 68% of severe dysplasia specimens were considered positive (> than 50% cells staining) for Cox-2. There was a significant increase in both percentage of cells staining for Cox-2, number of specimens considered positive for Cox-2, and H score when severely dysplastic specimens were compared to mild and moderate dysplasia ( P < 0.001; Table 1). There was also a tendency toward increased intensity of Cox-2 staining with severity of dysplasia, P = 0.06. Finally, we evaluated margin status and the presence or absence of Cox-2 protein staining at the margins of the LEEP resection specimens. Of the 35 specimens that stained positive for Cox-2 protein, 20 (59%) of these specimens had positive Cox-2 staining that extended to the margins of the LEEP resection specimen. Positive Cox-2 protein staining

Table 1 The association between p27, cyclooxygenase-2 expression, and severity of cervical dysplasia Dysplasia Cox-2 (n) Intensity Mean % H Positive (0 – 3+) positive score specimens cells

p27

Normal (62) Mild (18) Moderate (19) Severe (25) a

Mean % Positive positive specimens cells

0.7

10

0.06

20 (32%)

97

60 (97%)

1.6 1.8

28 37

0.57 0.66

9 (50%) 9 (42%)

87 74

17 (94%) 17 (89%)

2.1a

67b

1.49b 17 (68%)b 46b

11 (44%)b

P = 0.06. Difference in staining between mild and moderate dysplasias versus severe dysplasia, significant at P < 0.001. b

599

was seen in histologically normal squamous epithelium, adjacent to the histologically dysplastic epithelium (Fig. 1). The average length of Cox-2 protein staining beyond the histologic dysplasia was 1.64 mm. When Cox-2 margin status was evaluated, positive margin status for Cox-2 was a significant risk factor for persistent and recurrent dysplasia, RR 1.68 95% CI (1.07 < RR < 2.65) ( P < 0.027).

Discussion This study demonstrates an increased expression of Cox2 protein by immunohistochemistry with increasing severity of cervical dysplasias. In contrast, p27 protein expression is decreased with increasing grade of cervical dysplasia. Cox-2 protein expression extends beyond the area of histologic dysplasia to adjacent nondysplastic cells in most of the specimens analyzed, and this expression predicts persistent or recurrent dysplasia following LEEP procedure. Cyclooxygenase-2 has been associated with squamous dysplasias and carcinogenesis in a variety of organ systems, but currently has not been extensively evaluated in precancerous cervical lesions [13,14,33]. Shamma et al. [13] found significantly high Cox-2 expression by immunohistochemistry in high-grade squamous dysplasias of the esophagus. An H score, which incorporated both intensity and positivity of Cox-2 staining, was six times higher in high-grade squamous dysplasias than normal esophagus epithelium. In the evaluation of 104 gastric carcinoma tissues, Lim et al. [14] found Cox-2 protein expression not only in cancer cells, but also in precancerous lesions such as metaplastic and adenomatous cells. Saukkonen et al. [33] found 44% of gastric dysplasias that showed no histologic evidence of invasion stained positive for Cox-2. There is also some recent data suggesting that p27 expression is related to Cox-2 expression in lung cancer, renal, and human umbilical vein endothelial cells [30,34,35]. Certain Cox-2 inhibitors have been found to enhance p27 protein expression via posttranslational regulation [34]. NS398, a specific Cox-2 inhibitor, upregulated p27 protein, while other G1 acting cyclins and cyclin-dependent kinases were not changed. NS398 did not stimulate the p27 promoter, and the synthesis rate of p27 protein was not altered by NS398 [34]. Apoptosis induced in human umbilical vein endothelial cells mediated by caspase inhibitors showed upregulation of p27 and down regulation of Cox-2 expression. Finally, the induction of the cyclin-dependent kinase inhibitor p27 has been related to cyxlooxygenase-2 overexpression in mesangial cell proliferation [30]. Cox-2 expression appears to inhibit mesangial cell proliferation by a novel mechanism, independent of prostaglandin synthesis, but involving p27, p53, and p2130. Both Cox-2 and p27 protein expression have been found to be prognostic in invasive cervical carcinomas [17,19 – 24,36,37]. Cox-2 expression was evaluated in 24 patients with invasive cervical cancer and increased expression of

600

J. Farley et al. / Gynecologic Oncology 92 (2004) 596–602

Cox-2 predicted a diminished survival in these patients who were treated with radiation therapy [21]. Increased Cox-2 expression has also been associated with chemotherapy resistance in cervical cancers [20]. Evaluation of 84 cervical cancer patients found that Cox-2-positive patients had a decreased 2-year survival (38% vs. 85%) when compared to Cox-2-negative patients. In multivariate analysis, only advanced stage and Cox-2 positivity retained independent prognostic significance for overall survival [20]. High p27 expression has been shown to be a predictive factor for good prognosis in cervical squamous cell cancers treated with radiation alone [37]. Seventy-seven patients with squamous cell carcinoma of the cervix had biopsies obtained before and during radiation therapy and Oka et al. [37] found that both a decrease in mean labeling index of p27 after 27 Gray of radiation therapy and a high p27 labeling index before the start of radiation therapy were significantly associated with an increase in disease-free survival. However, the expression of Cox-2 and p27 proteins, and their impact on preinvasive squamous lesions of the cervix, has rarely been addressed in the literature and is largely limited to p27 evaluation [23,24,28,29,38]. Shiozawa et al. [23] found the expression of p27 in cervical intraepithelial neoplasias to be negligible; however, the number of p27positive cells increased with invasive squamous carcinomas. In their evaluation of p27 protein expression, they correlated staining with accelerated growth as shown by Ki-67 positivity and evaluated p27 staining only in cells that stained positive for Ki-67. They remarked that there were cells that stained positive for p27 in the superficial layer of the epithelium which were negative for Cyclin E and Ki-67, and these were not included in the final evaluation. Our findings of decreased p27 expression with increasing grade of dysplasia are consistent with other studies that evaluate the total number of positive cells staining for p27 in all layers of the epithelium, in contrast to the selective inclusion criteria of Shiozawa et al. [23] using Ki-67-positive cells. Evaluation of p27 expression in the neoplastic transformation of the endocervix found that p27 protein expression decreased with the progression to malignancy. Decreased p27 protein levels were detected in adenocarcinoma, and adenocarcinoma in situ, while normal endocervix showed the highest levels of p27 protein expression [24]. The lack of correlation with squamous precursor lesions of the cervix and p27 protein expression found by previous authors could be due to other factors such as method of calculating and quantifying p27 staining, degree of p27 protein degradation, or the influence of other cell cycle regulators such as Cyclin E [28,29,39]. Treatment of these precancerous lesions is imperative to prevent cervical cancer [40]. Approximately 66% of all dysplasia will progress to carcinoma in situ if left untreated and 2% will progress to cancer [40,41]. Treatment for these lesions involves removal of the dysplastic tissue with a variety of excisional procedures to include LEEP [42,43]. Recurrence following LEEP procedure has been quoted at

3 – 30% [44,45]. Incomplete resection has been previously shown to increase failure rate; however, controversy exists on the best way to predict residual disease [44 – 47]. In a previous analysis of this cohort of patients, looking for histologic risk factors for recurrent or persistent cervical dysplasia after LEEP, we found no single histologic factor predictive of residual or recurrent disease [44]. We did, however, identify a combination of factors such as marginal involvement of cervical dysplasia and endocervical glandular involvement with dysplasia, which doubled the risk for recurrent dysplasia [44]. In the current evaluation of 62 LEEP specimens, we found increased Cox-2 expression with increasing severity of dysplasia. Positive Cox-2 expression extended 1.6 mm on average, beyond the adjacent dysplastic epithelium into normal cells. The presence of Cox-2 staining at the LEEP margins was an independent predictor of persistent or recurrent cervical dysplasia. This single detail increased the risk by 68% for having recurrent or residual cervical dysplasia. In conclusion, it appears that Cox-2 and p27 protein expression plays a role in squamous cervical cancer carcinogenesis. Cox-2 staining is often found outside the dysplastic lesion in adjacent normal cells in a majority, 59%, of LEEP specimens. The presence of Cox-2 protein at the margins of surgical resection significantly increases the risk of persistent and recurrent dysplasia following LEEP procedure, regardless of the histologic status of the margins. Patients who are found to express Cox-2 in their cervical dysplastic lesions should have surgical resection that includes a wide margin to ensure appropriate treatment. Should the histologic margin of surgical resection approach 2.0 mm, follow-up algorithms for these patients should include intensive surveillance to ensure adequate treatment of disease, as they are at substantial risk for recurrent disease.

References [1] Becker TM, Stone KM, Alexander ER. Genital human papillomavirus infection. A growing concern. Obstet Gynecol Clin North Am 1987; 14:389 – 96. [2] Reid R, Campion MJ. The biology and significance of human papillomavirus infections in the genital tract. Yale J Biol Med 1988;61: 307 – 25. [3] Jones MH. The management of the mildly dyskaryotic smear. Br J Obstet Gynaecol 1994;101:474 – 6. [4] Kurman RJ, Henson DE, Herbst AL, Noller KL, Schiffman MH. Interim guidelines for management of abnormal cervical cytology. The 1992 National Cancer Institute Workshop. JAMA 1994;271:1866 – 9. [5] Jemal A, Thomas A, Murray T, Thun M. Cancer statistics, 2003. CA Cancer J Clin 2002;52:23 – 47. [6] van Muyden RC, ter Harmsel BW, Smedts FM, Hermans J, Kuijpers JC, Raikhlin NT, et al. Detection and typing of human papillomavirus in cervical carcinomas in Russian women: a prognostic study. Cancer 1999;85:2011 – 6. [7] Slagle BL, Kaufman RH, Reeves WC, Icenogle JP. Expression of ras, c-myc, and p53 proteins in cervical intraepithelial neoplasia. Cancer 1998;83:1401 – 18.

J. Farley et al. / Gynecologic Oncology 92 (2004) 596–602 [8] Kersemaekers AM, Fleuren GJ, Kenter GG, Van den Broek LJ, Uljee SM, Hermans J, et al. Oncogene alterations in carcinomas of the uterine cervix: overexpression of the epidermal growth factor receptor is associated with poor prognosis. Clin Cancer Res 1999;5:577 – 86. [9] Ho GY, Bierman R, Beardsley L, Chang CJ, Burk RD. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med 1998;338:423 – 8. [10] Hopman EH, Rozendaal L, Voorhorst FJ, Walboomers JM, Kenemans P, Helmerhorst TJ. High risk human papillomavirus in women with normal cervical cytology prior to the development of abnormal cytology and colposcopy. BJOG 2000;107:600 – 4. [11] Laukkanen P, Koskela P, Pukkala E, Dillner J, Laara E, Knekt P, et al. Time trends in incidence and prevalence of human papillomavirus type 6, 11 and 16 infections in Finland. J Cell Physiol 2003;84(Pt8):2105 – 9 [0022 – 1317]. [12] Chesebro MJ, Everett WD. A cost – benefit analysis of colposcopy for cervical squamous intraepithelial lesions found on Papanicolaou smear. Arch Fam Med 1996;5:576 – 81. [13] Shamma A, Yamamoto H, Doki Y, Okami J, Kondo M, Fujiwara Y, et al. Up-regulation of cyclooxygenase-2 in squamous carcinogenesis of the esophagus. Clin Cancer Res 2000;6:1229 – 38. [14] Lim HY, Joo HJ, Choi JH, Yi JW, Yang MS, Cho DY, et al. Increased expression of cyclooxygenase-2 protein in human gastric carcinoma. Clin Cancer Res 2000;6:519 – 25. [15] Eberhart CE, Coffey RJ, Radhika A, Giardiello FM, Ferrenbach S, DuBois RN. Up-regulation of cyclooxygenase 2 gene expression in human colorectal adenomas and adenocarcinomas. Gastroenterology 1994;107:1183 – 8. [16] Hwang D, Scollard D, Byrne J, Levine E. Expression of cyclooxygenase-1 and cyclooxygenase-2 in human breast cancer. J Natl Cancer Inst 1998;90:455 – 60. [17] Ryu HS, Chang KH, Yang HW, Kim MS, Kwon HC, Oh KS. High cyclooxygenase-2 expression in stage IB cervical cancer with lymph node metastasis or parametrial invasion. Gynecol Oncol 2000;76: 320 – 5. [18] Ishiko O, Sumi T, Yoshida H, Tokuyama O, Wakasa K, Haba T, et al. Cyclooxygenase-2 expression in an adenoid cystic carcinoma of the uterine cervix. Oncol Rep 2001;8:1023 – 5. [19] Kulkarni S, Rader JS, Zhang F, Liapis H, Koki AT, Masferrer JL, et al. Cyclooxygenase-2 is overexpressed in human cervical cancer. Clin Cancer Res 2001;7:429 – 34. [20] Ferrandina G, Lauriola L, Distefano MG, Zannoni GF, Gessi M, Legge F, et al. Increased cyclooxygenase-2 expression is associated with chemotherapy resistance and poor survival in cervical cancer patients. J Clin Oncol 2002;20:973 – 81. [21] Gaffney DK, Holden J, Davis M, Zempolich K, Murphy KJ, Dodson M. Elevated cyclooxygenase-2 expression correlates with diminished survival in carcinoma of the cervix treated with radiotherapy. Int J Radiat Oncol Biol Phys 2001;49:1213 – 7. [22] Gaffney DK, Holden J, Zempolich K, Murphy KJ, Dicker AP, Dodson M. Elevated COX-2 expression in cervical carcinoma: reduced cause-specific survival and pelvic control. Am J Clin Oncol 2001; 24:443 – 6. [23] Shiozawa T, Shiohara S, Kanai M, Konishi I, Fujii S, Nikaido T. Expression of the cell cycle regulator p27(Kip1) in normal squamous epithelium, cervical intraepithelial neoplasia, and invasive squamous cell carcinoma of the uterine cervix. Immunohistochemistry and functional aspects of p27(Kip1). Cancer 2001;92:3005 – 11. [24] Farley J, Gray K, Nycum L, Prentice M, Birrer MJ, Jakowlew SB. Endocervical cancer is associated with an increase in the ligands and receptors for transforming growth factor-beta and a contrasting decrease in p27(Kip1). Gynecol Oncol 2000;78:113 – 22. [25] Fredersdorf S, Burns J, Milne AM, Packham G, Fallis L, Gillett CE, et al. High level expression of p27(kip1) and cyclin D1 in some human breast cancer cells: inverse correlation between the expression of p27(kip1) and degree of malignancy in human breast and colorectal cancers. Proc Natl Acad Sci U S A 1997;94:6380 – 5.

601

[26] Di Gennaro E, Barbarino M, Bruzzese F, De Lorenzo S, Caraglia M, Abbruzzese A, et al. Critical role of both p27KIP1 and p21CIP1/WAF1 in the antiproliferative effect of ZD1839 (‘Iressa’), an epidermal growth factor receptor tyrosine kinase inhibitor, in head and neck squamous carcinoma cells. J Gen Viro 2003;195(1):139 – 50 [0021 – 9541]. [27] Kudo Y, Kitajjma S, Sato S, Miyauchi M, Ogawa I, Takata T. Establishment of an oral squamous cell carcinoma cell line with high invasive and p27 degradation activities from a lymph node metastasis. Oral 2003;39(5):515 – 20 [1368 – 8375]. [28] Goff BA, Sallin J, Garcia R, VanBlaricom A, Paley PJ, Muntz HG. Evaluation of p27 in preinvasive and invasive malignancies of the cervix. Gynecology 2003;88(1):40 – 4 [0090 – 8258]. [29] Huang LW, Chao SL, Hwang JL, Chou YY. Down-regulation of p27 is associated with malignant transformation and aggressive phenotype of cervical neoplasms. Gynecology 2002;85(3):524 – 48 [0090 – 8258]. [30] Zahner G, Wolf G, Ayoub M, Reinking R, Panzer U, Shankland SJ, et al. Cyclooxygenase-2 overexpression inhibits platelet-derived growth factor-induced mesangial cell proliferation through induction of the tumor suppressor gene p53 and the cyclin-dependent kinase inhibitors p21waf-1/cip-1 and p27kip-1. J Biol Chem 2002;277: 9763 – 71. [31] Lloyd RV, Jin L, Qian X, Kulig E. Aberrant p27kip1 expression in endocrine and other tumors. Am J Pathol 1997;150:401 – 7. [32] McCarty Jr KS, Szabo E, Flowers JL, Cox EB, Leight GS, Miller L, et al. Use of a monoclonal anti-estrogen receptor antibody in the immunohistochemical evaluation of human tumors. Cancer Res 1986; 46:4244s – 8s. [33] Saukkonen K, Nieminen O, van Rees B, Vilkki S, Harkonen M, Juhola M, et al. Expression of cyclooxygenase-2 in dysplasia of the stomach and in intestinal-type gastric adenocarcinoma. Clin Cancer Res 2001;7:1923 – 31. [34] Hung WC, Chang HC, Pan MR, Lee TH, Chuang LY. Induction of p27(KIP1) as a mechanism underlying NS398-induced growth inhibition in human lung cancer cells. Mol Pharmacol 2000;58:1398 – 403. [35] Teruyama K, Abe M, Nakano T, Iwasaka-Yagi C, Takahashi S, Yamada S, et al. Role of transcription factor Ets-1 in the apoptosis of human vascular endothelial cells. J Cell Physiol 2001;188:243 – 52. [36] Dong YL, Gangula PR, Fang L, Yallampalli C. Differential expression of cyclooxygenase-1 and -2 proteins in rat uterus and cervix during the estrous cycle, pregnancy, labor and in myometrial cells. Prostaglandins 1996;52:13 – 34. [37] Oka K, Suzuki Y, Nakano T. Expression of p27 and p53 in cervical squamous cell carcinoma patients treated with radiotherapy alone: radiotherapeutic effect and prognosis. Cancer 2000;88: 2766 – 73. [38] Troncone G, Vetrani A, de Rosa G, Gerbasio D, Palombini L. Cyclin dependent kinase inhibitor p27Kip1 expression in normal and neoplastic cervical epithelium. J Clin Pathol 1999;52:880 – 7. [39] Farley J, Smith LM, Darcy KM, Sobel E, O’Connor D, Henderson B, et al. Cyclin E expression is a significant predictor of survival in advanced, suboptimally debulked ovarian epithelial cancers: a gynecologic oncology group study. Cancer Res 2003;63:1235 – 41 [Corporate name: Gynecologic Oncology Group]. [40] Mitchell MF, Tortolero-Luna G, Wright T, Sarkar A, Richards-Kortum R, Hong WK, et al. Cervical human papillomavirus infection and intraepithelial neoplasia: a review. J Natl Cancer Inst Monogr 1996:17 – 25. [41] Barron BA, Richart RM. A statistical model of the natural history of cervical carcinoma based on a prospective study of 557 cases. J Natl Cancer Inst 1968;41:1343 – 53. [42] Luesley DM, Cullimore J, Redman CW, Lawton FG, Emens JM, Rollason TP, et al. Loop diathermy excision of the cervical transformation zone in patients with abnormal cervical smears. BMJ 1990;300:1690 – 3. [43] Kolstad P, Klem V. Long-term follow-up of 1121 cases of carcinoma in situ. Obstet Gynecol 1976;48:125 – 9.

602

J. Farley et al. / Gynecologic Oncology 92 (2004) 596–602

[44] Dietrich III CS, Yancey MK, Miyazawa K, Williams DL, Farley J. Risk factors for early cytologic abnormalities after loop electrosurgical excision procedure. Obstet Gynecol 2002;99:88 – 92. [45] Baldauf JJ, Dreyfus M, Ritter J, Cuenin C, Tissier I, Meyer P. Cytology and colposcopy after loop electrosurgical excision: implications for follow-up. Obstet Gynecol 1998;92:124 – 30. [46] Murdoch JB, Morgan PR, Lopes A, Monaghan JM. Histological in-

complete excision of CIN after large loop excision of the transformation zone (LLETZ) merits careful follow up, not retreatment. Br J Obstet Gynaecol 1992;99:990 – 3. [47] Costa S, De Nuzzo M, Infante FE, Bonavita B, Marinelli M, Rubino A, et al. Disease persistence in patients with cervical intraepithelial neoplasia undergoing electrosurgical conization. Gynecol Oncol 2002; 85:11 – 124.