- Email: [email protected]
Levels of the DNA adduct, N7-methyldeoxyguanosine, are associated with increased risk of failure of treatment of cervical intraepithelial neoplasia
Gynecologic Oncology 93 (2004) 605 – 609 www.elsevier.com/locate/ygyno
Levels of the DNA adduct, N7-methyldeoxyguanosine, are associated with increased risk of failure of treatment of cervical intraepithelial neoplasia N.N. Acladious, a K.L. Harrison, b C.J. Sutton, c A.C. Povey, b,d,* D. Mandal, a and H. Kitchener e a
b
Department of Genito-Urinary Medicine, Manchester Royal Infirmary, Manchester, UK Centre for Occupational and Environmental Health, University of Manchester, Manchester M13 9PL, UK c Statistics Group, Faculty of Science, University of Central Lancashire, Preston PR1 2HE, UK d Cancer Research UK Carcinogenesis Group, Paterson Institute for Cancer Research, Manchester, UK e Academic Unit of Obstetrics and Gynaecology, University of Manchester, UK Received 29 August 2003 Available online 12 April 2004
Abstract Objective. To determine whether exposure to methylating agents was a risk factor for treatment failure in women undergoing colposcopic examination. Methods. Nine hundred fifty-eight women attending for colposcopic examination after abnormal cervical smear test results were recruited into the study cohort. Information on demographic factors, smoking and other risk factors was obtained and a pre-treatment biopsy was taken and stored at 70jC. After follow-up, cases who had treatment failure of cervical intraepithelial neoplasia (CIN) within 2 years following treatment were identified (n = 77) and matched to women with no treatment failure of CIN in this time period (controls, n = 154). DNA was extracted from the pre-treatment biopsies and levels of N7-methyl-deoxyguanosine (N7-MedG), a marker of exposure to methylating agents, were quantified as the ring-opened form of the base damage by a validated immunoslotblot assay. Results. Sufficient DNA for N7-MedG analysis was extracted from 61 subjects corresponding to 20 matched case control pairs. N7-MedG was detected in cervical DNA with levels ranging from non-detected (<0.1 Amol/mol dG) to 4.83 Amol/mol dG. N7-MedG levels were significantly higher in cases (geometric mean 0.99 Amol/mol dG) than controls (0.33 Amol/mol dG; P = 0.01). There were no associations between N7-MedG levels and HPV or smoking status. Log N7-MedG content, after adjustment for HPV status at time of treatment, was found to be significantly associated with increased risk of treatment failure (OR 5.74, 95% CI 1.05 – 31.23). Conclusions. The association between pre-treatment levels of DNA damage induced by methylating agents and subsequent treatment failure implicates methylating agent exposure as a causative factor in treatment failure. D 2004 Elsevier Inc. All rights reserved. Keywords: N7-methyldeoxyguanosine; Cervical intraepithelial neoplasia; Methylating agent
Introduction After conservative treatment of cervical intraepithelial neoplasia (CIN), the risk of invasive cervical cancer among these women is about five times greater than that among the general population of women [1]. Treatment failure has been associated with the size of the lesion, partially involved excision margins and the degree of CIN [2 – 4]. In a recent prospective study, we also found that persistent * Corresponding author. Centre for Occupational and Environmental Health, University of Manchester, Fourth Floor, Humanities Building, Oxford Road, Manchester M13 9PL, UK. Fax: +44-161-275-5595. E-mail address: [email protected] (A.C. Povey). 0090-8258/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2004.03.005
HPV infection and smoking increased the risk of treatment failure [5]. The epidemiological evidence linking cigarette smoking to cervical cancer risk is controversial but exposure to chemical carcinogens is necessary to achieve malignant transformations in HPV-infected cells in vitro [6 –8]. Carcinogen-DNA adducts are present in the cervix [9 –13]. The cervix contains enzymes that can metabolise carcinogens [14,15] and expression of these enzymes may be modified by HPV infection [15] and may help to determine cervical cancer risk [16]. Furthermore, increased cancer risk following exposure to both a virus (HBV) and a chemical carcinogen (aflatoxin) has been described previously [17]. As the tobacco-specific nitrosamine (4-(methylnitrosoamine)-1-(3-pyridyl)-1-butanone (NNK) was more
606
N.N. Acladious et al. / Gynecologic Oncology 93 (2004) 605–609
prevalent in cervical mucus samples from smokers than non-smokers [18], we have investigated whether exposure to alkylating agents, as measured by N7-methyldeoxyguanosine (N7-MedG) levels in cervical DNA, is higher in those women who experience treatment failure than those who do not.
Materials and methods A prospective multi-centre study was conducted on a cohort of women living in the North West of England who were treated for CIN after abnormal cervical smear test results. This study has been described in full elsewhere [5]. Briefly, women who attended one of three colposcopy clinics at three hospitals (Royal Bolton Hospital, Saint Mary’s Hospital, Rochdale Royal Infirmary) for treatment of CIN between August 1, 1995 and July 31, 1999 were recruited. Pregnant patients, cervical cancer patients and those with conditions causing diminished immune competence (such as corticosteriod treatment, human immune deficiency virus infection, organ transplantation, chemotherapy, radiotherapy) were excluded from the study. At recruitment, information on demographic factors, smoking and other risk factors was obtained and a pretreatment biopsy was taken and stored at 70jC until analysed. Women were treated for CIN with all the centres having the facilities for treatment by loop diathermy, cone biopsy, CO2 laser and cold coagulation. Upon follow-up, women who had treatment failure of CIN were then identified. For the purposes of this nested case-control study, cases were defined as those patients who had treatment failure of CIN within 2 years following treatment, the diagnosis of which had to be confirmed histologically. Controls were selected from patients with no treatment failure of CIN in this time period, which was confirmed by two negative smears, the second of which was taken at least 24 months after treatment. Controls were matched for lesional and treatment factors (the size of the lesion, the method of treatment, the degree of CIN and the state of the excision margins) known to affect the risk of treatment failure. Those women with other smear results during follow-up (e.g., borderline, inadequate, abnormal results without histological confirmation or reverted to normal without treatment) were excluded as cases or controls. Nine hundred and fifty-eight patients were recruited into the study cohort; 77 cases were identified and 154 controls were selected based on the matching criteria. DNA was extracted from the pre-treatment cervical biopsy using the Qiagen genomic DNA extraction kit following the standard protocol (Qiagen, Crawley, UK) except that proteinase K and ribonuclease A digestion was carried out overnight at 4jC, followed by 1 h at 37jC, to reduce loss of N7-MedG [19]. N7-MedG was quantified as the ring-opened form of the adduct by an immunoslotblot assay [20]. The detection limit
of the assay is 0.1 Amol N7-MedG/mol dG; those samples with values below this limit were entered in the analysis as 0.05 Amol/mol dG. Transformation of N7-MedG levels into log N7-MeG resulted in a more normal distribution and analysis was then performed using parametric procedures: results for N7-MedG levels are summarised as the geometric mean (95% CI). HPV testing was performed by PCR as previously described [5]. Characteristics of matched cases and controls were summarised in terms of medians and inter-quartile ranges (IQRs) for quantitative variables and in terms of frequency and percentage for the categorical variable. Treatment failure of CIN was modelled using multiple conditional logistic regression. The significance of variables in the model was assessed using the Wald test and using a 5% significance level. Categorical predictor variables considered were coded using the standard indicator (dummy) variables approach and odds ratio estimates with 95% confidence intervals (CIs) obtained. As there were relatively small numbers of samples in this matched comparison (41 out of a total of 61 analysed), associations between N7MedG levels and smoking and HPV status were also explored in all the analysed samples. No significant difference was found in the results when data from all the 61 subjects were analysed.
Results and discussion DNA extractions from single cervical biopsies yielded 2.9 F 2.14 Ag DNA (mean F SD; range 0.1 –26.7 Ag). Sufficient DNA for analysis of N7-MedG (>2.5 Ag DNA) was extracted from the biopsies of 61 patients (20 cases and 41 controls). Of these 41 controls, 21 were matched to the Table 1 Characteristics of sub-set of cases and matched controls from nested casecontrol study Characteristic
Cases (n = 20)
Controls (n = 21)
Median age (IQR) (years) Number HPV-positive at treatment/total sample (%) Number HPV-positive at follow-up/total sample (%) Smoking status: current/ex/never (%) Median time smoked (IQR) (years)a Median age started smoking (IQR) (years)a Median number of cigarettes smoked per day (IQR)a Median number of cigarettes smoked per day now (IQR)b Median time since last smoked (IQR) (years)c
27 (5.5) 10/20 (50.0)
26 (8) 17/21 (81.0)
6/15 (40.0)
1/10 (10.0)
17/1/2 (85/5/10)
11/3/7 (52/14/33)
10 (3)
10 (10.5)
15 (3.5)
15 (6)
10 (12.5)
5 (5)
10 (12.5)
8 (5)
3 (–)
8 (–)
a
Current and ex-smokers only. Current smokers only. c Ex-smokers only. b
N.N. Acladious et al. / Gynecologic Oncology 93 (2004) 605–609
607
Fig. 1. Determination of N7-MedG in cervical DNA. Panel A, immunoslot blot analysis standards and cervical biopsy DNA. All slots contain 1 Ag of DNA. Columns A and B rows 1 – 8 contain methylated standards ranging from 0 to 5.4 fmol N7-MedG/Ag DNA. The remaining slots are all human DNA samples. Panel B, standard curve for N7-MedG quantitation.
20 cases resulting in 20 matched case-control groups; the remaining 20 controls could not be matched to any cases. Of these 61 DNA samples, 33 were analysed more than twice (with a mean CV 9.0 F 7.6). The population characteristics of this matched sub-set can be seen in Table 1. No differences between cases and controls were statistically significant, although there was weak evidence that controls were more likely than cases to be HPV-positive at the time of diagnosis ( P = 0.081, OR 0.31; 95% CI 0.08 –1.16). The lack of HPV follow-up data led to a considerable lack of power to detect any difference between cases and controls ( P = 0.21, OR 4.00; 95% CI 0.45 –35.8). N7-MedG was detected in cervical DNA with levels ranging from non-detected (<0.1 Amol/mol dG) to 4.83 Amol/mol dG. A typical example of an immunoslotblot with a standard curve is shown in Fig. 1. The N7-MedG levels for the matched samples are summarised in Table 2. N7MedG levels in cases (geometric mean 0.99 Amol/mol dG) were significantly higher than those found in controls (0.33 Amol/mol dG; P = 0.01). The difference in N7-MedG levels between cases and the matched controls was 1.24 Amol/mol
dG, (95% CI 0.40 –2.08; P = 0.006). Cases were also more likely than controls to have N7-MedG levels greater than the median level (14 out of 20 vs. 6 out of 20, respectively; OR 5.44, 95% CI 1.18 – 27.12). There was no detectable association between N7-MedG levels and HPV infection with the geometric mean N7MedG level (95% CI) in HPV-positive women being 0.44 (0.28 –0.69) Amol/mol dG and 0.91 (0.34 –2.44) Amol/mol dG in women without HPV infection ( P = 0.12). The effect of log N7-MedG on treatment failure was assessed both independently and jointly with HPV status at the time of
Table 2 N7-MedG levels (Amol/mol dG) by treatment failure status with respect to smoking status Cases
All Smokers Non/ex smokers
Controls
n
Geometric mean (95% CI)
n
Geometric mean (95% CI)
20 17 3
0.99 (0.52 – 1.90) 1.04 (0.54 – 2.00) 0.74 (0.002 – 261.8)
21 11 10
0.32 (0.19 – 0.54) 0.27 (0.10 – 0.72) 0.42 (0.23 – 0.64)
608
N.N. Acladious et al. / Gynecologic Oncology 93 (2004) 605–609
diagnosis of CIN using multiple conditional logistic regression modelling. Log N7-MedG content, after adjustment for HPV status at time of treatment, was significantly associated with increased risk of treatment failure (OR 5.74, 95% CI 1.05 – 31.23). For a 10-fold increase in N7-MedG content of cervical DNA, there was an estimated 5- to 6-fold increase in the odds of CIN treatment failure. N7-MedG is a specific marker of exposure to methylating agents, including those present in cigarette smoke [21]. Current smokers had similar N7-MedG levels to those of current non-smokers with a geometric mean (95% CI) of 0.61 (0.34 – 1.09) and 0.48 (0.24 – 0.97) Amol/mol dG, respectively ( P = 0.62). There was no smoking-related difference in adduct levels in case or control populations analysed individually (Table 2). In current smokers, there was no association between N7-MedG levels and numbers of cigarettes smoked per day (r = 0.24; P = 0.22). In addition, there was no significant difference in the distribution of adduct levels as 16/28 (57.1%) smokers and 4/12 (33.3%) non-smokers had adduct levels above the median ( P = 0.17). The lack of an association between smoking and N7-MedG levels may reflect the small sample size, or alternatively, it could suggest that women in this study were misreporting their own smoking habit or that other sources of exposure, for example, dietary may be important [22,23]. A more specific exposure marker, such as cotinine in urine or serum, would help to identify the contribution of tobacco-specific nitrosamines [24] to cervical DNA alkylation. Alkylating agents are known mutagens and carcinogens that can cause cancer in humans at high doses [25]. Exposures, at lower doses, will lead to DNA damage but evidence linking such exposures to human cancers is more limited [26]. Pre-treatment levels of N7-MedG in cervical DNA were associated with an increased risk of treatment failure, which given the prospective nature of the study, provides strong evidence of a causal relationship between exposure and outcome. That such an association was detected within 2 years of treatment may highlight the synergistic nature of the interaction between a chemical carcinogen and a virus. However, the main limitation of this study is the small numbers of biological samples that were analysed which, for example, precluded a detailed investigation of the joint effect of smoking and the level of N7MedG on the risk of treatment failure. Whilst it is possible that the results may have occurred by chance or by some other (unaccounted for) confounding factor, the data are consistent with the known carcinogenic nature of methylating agents. Further work is thus warranted to confirm this association.
Acknowledgment Work in the laboratory was supported by Cancer Research UK.
References [1] Soutter WP, de Barros Lopes A, Fletcher A, Monaghan JM, Duncan ID, Paraskevaidis E, et al. Invasive cervical cancer after conservative therapy for cervical intraepithelial neoplasia. Lancet 1997; 349:978 – 80. [2] Andersen ES, Pedersen B, Nielsen K. Laser conization: the results of treatment of cervical intraepithelial neoplasia. Gynecol Oncol 1994; 54:201 – 4. [3] Shafi MI, Dunn JA, Buxton EJ, Finn CB, Jordan JA, Luesley DM. Abnormal cervical cytology following large loop excision of the transformation zone: a case controlled study. Br J Obstet Gynaecol 1993;100:145 – 8. [4] Cox T. Management of cervical intraepithelial neoplasia. Lancet 1999;53:857 – 9. [5] Acladious NN, Sutton C, Mandal D, Hopkins R, Zaklama M, Kitchener H. Persistent human papillomavirus infection and smoking increase risk of failure of treatment of cervical intraepithelial neoplasia (CIN). Int J Cancer 2002;98:435 – 9. [6] Nakao Y, Yang X, Yokoyama M, Pater MM, Pater A. Malignant transformation of human ectocervical cells immortalized by HPV 18: in vitro model of carcinogenesis by cigarette smoke. Carcinogenesis 1996;17:577 – 83. [7] Li SL, Kim MS, Cherrick HM, Doniger J, Park NH. Sequential combined tumorigenic effect of HPV-16 and chemical carcinogens. Carcinogenesis 1992;13:1981 – 7. [8] Garrett LR, Perez-Reyes N, Smith PP, McDougall JK. Interaction of HPV-18 and nitrosomethylurea in the induction of squamous cell carcinoma. Carcinogenesis 1993;14:329 – 32. [9] Ali S, Astley SB, Sheldon TA, Peel KR, Wells M. Detection and measurement of DNA adducts in the cervix of smokers and nonsmokers. Int J Gynecol Cancer 1994;4:188 – 93. [10] Simons AM, Mugica van Herckenrode C, Rodriguez JA, Maitland N, Anderson M, Phillips DH, et al. Demonstration of smoking-related DNA damage in cervical epithelium and correlation with human papillomavirus type 16, using exfoliated cervical cells. Br J Cancer 1995;71:246 – 9. [11] King MM, Hollingsworth A, Cuzick J, Garner RC. The detection of adducts in human cervix tissue DNA using 32P-postlabelling: a study of the relationship with smoking history and oral contraceptive use. Carcinogenesis 1994;15:1097 – 100. [12] Phillips DH, She MN. DNA adducts in cervical tissue of smokers and non-smokers. Mutat Res 1994;313:277 – 84. [13] Melikian AA, Sun P, Prokopczyk B, El-Bayoumy K, Hoffmann D, Wang X, et al. Identification of benzo[a]pyrene metabolites in cervical mucus and DNA adducts in cervical tissues in humans by gas chromatography-mass spectrometry. Cancer Lett 1999; 146:127 – 34. [14] Patel KR, Astley S, Adams DJ, Lacey CJ, Ali SW, Wells M. Expression of cytochrome P450 enzymes in the cervix. An immunohistochemical study. Int J Gynecol Cancer 1993;3:159 – 63. [15] Farin FM, Bigler LG, Oda D, McDougall JK, Omiecinski CJ. Expression of cytochrome P450 and microsomal expoxide hydrolase in cervical and oral epithelial cells immortalized by human papillomavirus type 16 E6/E7 genes. Carcinogenesis 1995;16: 1670 – 4. [16] Warwick A, Sarhanis P, Redman C, Pemble S, Taylor JB, Ketterer B, et al. Theta class glutathione S-transferase GSTT1 genotypes and susceptibility to cervical neoplasia: interactions with GSTM1, CYP2D6 and smoking. Carcinogenesis 1994;15:2841 – 5. [17] Ross RK, Yuan J-M, Yu MC, Wogan GN, Qian G-S, Tu J-T, et al. Urinary aflatoxin biomarkers and risk of hepatocellular carcinoma. Lancet 1992;339:943 – 6. [18] Prokopczyk B, Cox JE, Hoffmann D, Waggoner SE. Identification of tobacco-specific carcinogen in the cervical mucus of smokers and nonsmokers. J Natl Cancer Inst 1997;89:868 – 73.
N.N. Acladious et al. / Gynecologic Oncology 93 (2004) 605–609 [19] Haque K, Cooper DP, van Delft JHM, Lee SM, Povey AC. Accurate and sensitive quantitation of N7-methyldeoxyguanosine-3V-monophosphate by 32P-postlabelling and storage-phosphor imaging. Chem Res Toxicol 1997;10:660 – 6. [20] Harrison KL, Wood M, Lees NP, Hall CN, Margison GP, Povey AC. Development and application of a sensitive and rapid immunoassay for the quantitation of N7-methyldeoxyguanosine in DNA samples. Chem Res Toxicol 2001;14:295 – 301. [21] Bianchini F, Wild CP. 7-methyldeoxyguanosine as a marker of exposure to environmental methylating agents. Toxicol Lett 1994; 72:175 – 84. [22] Simons AM, Phillips DH, Coleman DV. Damage to DNA in cer-
[23] [24] [25]
[26]
609
vical epithelium related to smoking tobacco. Br Med J 1993; 306:1444 – 8. Sillett RW, Wilson MB, Malcolm RE, Ball KP. Deception among smokers. Br Med J 1978;2:1185 – 6. Hecht SS. DNA adduct formation from tobacco-specific N-nitrosamines. Mutat Res 1999;424:127 – 42. Leone G, Voso MT, Sica S, Morosetti R, Pagano L. Therapy related leukemias: susceptibility, prevention and treatment. Leuk Lymphoma 2001;41:255 – 76. Hecht SS. Approaches to cancer prevention based on an understanding of N-nitrosamine carcinogenesis. Proc Soc Exp Biol Med 1997;216:181 – 91.
Levels of the DNA adduct, N7-methyldeoxyguanosine, are associated with increased risk of failure of treatment of cervical intraepithelial neoplasia N.N. Acladious, a K.L. Harrison, b C.J. Sutton, c A.C. Povey, b,d,* D. Mandal, a and H. Kitchener e a
b
Department of Genito-Urinary Medicine, Manchester Royal Infirmary, Manchester, UK Centre for Occupational and Environmental Health, University of Manchester, Manchester M13 9PL, UK c Statistics Group, Faculty of Science, University of Central Lancashire, Preston PR1 2HE, UK d Cancer Research UK Carcinogenesis Group, Paterson Institute for Cancer Research, Manchester, UK e Academic Unit of Obstetrics and Gynaecology, University of Manchester, UK Received 29 August 2003 Available online 12 April 2004
Abstract Objective. To determine whether exposure to methylating agents was a risk factor for treatment failure in women undergoing colposcopic examination. Methods. Nine hundred fifty-eight women attending for colposcopic examination after abnormal cervical smear test results were recruited into the study cohort. Information on demographic factors, smoking and other risk factors was obtained and a pre-treatment biopsy was taken and stored at 70jC. After follow-up, cases who had treatment failure of cervical intraepithelial neoplasia (CIN) within 2 years following treatment were identified (n = 77) and matched to women with no treatment failure of CIN in this time period (controls, n = 154). DNA was extracted from the pre-treatment biopsies and levels of N7-methyl-deoxyguanosine (N7-MedG), a marker of exposure to methylating agents, were quantified as the ring-opened form of the base damage by a validated immunoslotblot assay. Results. Sufficient DNA for N7-MedG analysis was extracted from 61 subjects corresponding to 20 matched case control pairs. N7-MedG was detected in cervical DNA with levels ranging from non-detected (<0.1 Amol/mol dG) to 4.83 Amol/mol dG. N7-MedG levels were significantly higher in cases (geometric mean 0.99 Amol/mol dG) than controls (0.33 Amol/mol dG; P = 0.01). There were no associations between N7-MedG levels and HPV or smoking status. Log N7-MedG content, after adjustment for HPV status at time of treatment, was found to be significantly associated with increased risk of treatment failure (OR 5.74, 95% CI 1.05 – 31.23). Conclusions. The association between pre-treatment levels of DNA damage induced by methylating agents and subsequent treatment failure implicates methylating agent exposure as a causative factor in treatment failure. D 2004 Elsevier Inc. All rights reserved. Keywords: N7-methyldeoxyguanosine; Cervical intraepithelial neoplasia; Methylating agent
Introduction After conservative treatment of cervical intraepithelial neoplasia (CIN), the risk of invasive cervical cancer among these women is about five times greater than that among the general population of women [1]. Treatment failure has been associated with the size of the lesion, partially involved excision margins and the degree of CIN [2 – 4]. In a recent prospective study, we also found that persistent * Corresponding author. Centre for Occupational and Environmental Health, University of Manchester, Fourth Floor, Humanities Building, Oxford Road, Manchester M13 9PL, UK. Fax: +44-161-275-5595. E-mail address: [email protected] (A.C. Povey). 0090-8258/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2004.03.005
HPV infection and smoking increased the risk of treatment failure [5]. The epidemiological evidence linking cigarette smoking to cervical cancer risk is controversial but exposure to chemical carcinogens is necessary to achieve malignant transformations in HPV-infected cells in vitro [6 –8]. Carcinogen-DNA adducts are present in the cervix [9 –13]. The cervix contains enzymes that can metabolise carcinogens [14,15] and expression of these enzymes may be modified by HPV infection [15] and may help to determine cervical cancer risk [16]. Furthermore, increased cancer risk following exposure to both a virus (HBV) and a chemical carcinogen (aflatoxin) has been described previously [17]. As the tobacco-specific nitrosamine (4-(methylnitrosoamine)-1-(3-pyridyl)-1-butanone (NNK) was more
606
N.N. Acladious et al. / Gynecologic Oncology 93 (2004) 605–609
prevalent in cervical mucus samples from smokers than non-smokers [18], we have investigated whether exposure to alkylating agents, as measured by N7-methyldeoxyguanosine (N7-MedG) levels in cervical DNA, is higher in those women who experience treatment failure than those who do not.
Materials and methods A prospective multi-centre study was conducted on a cohort of women living in the North West of England who were treated for CIN after abnormal cervical smear test results. This study has been described in full elsewhere [5]. Briefly, women who attended one of three colposcopy clinics at three hospitals (Royal Bolton Hospital, Saint Mary’s Hospital, Rochdale Royal Infirmary) for treatment of CIN between August 1, 1995 and July 31, 1999 were recruited. Pregnant patients, cervical cancer patients and those with conditions causing diminished immune competence (such as corticosteriod treatment, human immune deficiency virus infection, organ transplantation, chemotherapy, radiotherapy) were excluded from the study. At recruitment, information on demographic factors, smoking and other risk factors was obtained and a pretreatment biopsy was taken and stored at 70jC until analysed. Women were treated for CIN with all the centres having the facilities for treatment by loop diathermy, cone biopsy, CO2 laser and cold coagulation. Upon follow-up, women who had treatment failure of CIN were then identified. For the purposes of this nested case-control study, cases were defined as those patients who had treatment failure of CIN within 2 years following treatment, the diagnosis of which had to be confirmed histologically. Controls were selected from patients with no treatment failure of CIN in this time period, which was confirmed by two negative smears, the second of which was taken at least 24 months after treatment. Controls were matched for lesional and treatment factors (the size of the lesion, the method of treatment, the degree of CIN and the state of the excision margins) known to affect the risk of treatment failure. Those women with other smear results during follow-up (e.g., borderline, inadequate, abnormal results without histological confirmation or reverted to normal without treatment) were excluded as cases or controls. Nine hundred and fifty-eight patients were recruited into the study cohort; 77 cases were identified and 154 controls were selected based on the matching criteria. DNA was extracted from the pre-treatment cervical biopsy using the Qiagen genomic DNA extraction kit following the standard protocol (Qiagen, Crawley, UK) except that proteinase K and ribonuclease A digestion was carried out overnight at 4jC, followed by 1 h at 37jC, to reduce loss of N7-MedG [19]. N7-MedG was quantified as the ring-opened form of the adduct by an immunoslotblot assay [20]. The detection limit
of the assay is 0.1 Amol N7-MedG/mol dG; those samples with values below this limit were entered in the analysis as 0.05 Amol/mol dG. Transformation of N7-MedG levels into log N7-MeG resulted in a more normal distribution and analysis was then performed using parametric procedures: results for N7-MedG levels are summarised as the geometric mean (95% CI). HPV testing was performed by PCR as previously described [5]. Characteristics of matched cases and controls were summarised in terms of medians and inter-quartile ranges (IQRs) for quantitative variables and in terms of frequency and percentage for the categorical variable. Treatment failure of CIN was modelled using multiple conditional logistic regression. The significance of variables in the model was assessed using the Wald test and using a 5% significance level. Categorical predictor variables considered were coded using the standard indicator (dummy) variables approach and odds ratio estimates with 95% confidence intervals (CIs) obtained. As there were relatively small numbers of samples in this matched comparison (41 out of a total of 61 analysed), associations between N7MedG levels and smoking and HPV status were also explored in all the analysed samples. No significant difference was found in the results when data from all the 61 subjects were analysed.
Results and discussion DNA extractions from single cervical biopsies yielded 2.9 F 2.14 Ag DNA (mean F SD; range 0.1 –26.7 Ag). Sufficient DNA for analysis of N7-MedG (>2.5 Ag DNA) was extracted from the biopsies of 61 patients (20 cases and 41 controls). Of these 41 controls, 21 were matched to the Table 1 Characteristics of sub-set of cases and matched controls from nested casecontrol study Characteristic
Cases (n = 20)
Controls (n = 21)
Median age (IQR) (years) Number HPV-positive at treatment/total sample (%) Number HPV-positive at follow-up/total sample (%) Smoking status: current/ex/never (%) Median time smoked (IQR) (years)a Median age started smoking (IQR) (years)a Median number of cigarettes smoked per day (IQR)a Median number of cigarettes smoked per day now (IQR)b Median time since last smoked (IQR) (years)c
27 (5.5) 10/20 (50.0)
26 (8) 17/21 (81.0)
6/15 (40.0)
1/10 (10.0)
17/1/2 (85/5/10)
11/3/7 (52/14/33)
10 (3)
10 (10.5)
15 (3.5)
15 (6)
10 (12.5)
5 (5)
10 (12.5)
8 (5)
3 (–)
8 (–)
a
Current and ex-smokers only. Current smokers only. c Ex-smokers only. b
N.N. Acladious et al. / Gynecologic Oncology 93 (2004) 605–609
607
Fig. 1. Determination of N7-MedG in cervical DNA. Panel A, immunoslot blot analysis standards and cervical biopsy DNA. All slots contain 1 Ag of DNA. Columns A and B rows 1 – 8 contain methylated standards ranging from 0 to 5.4 fmol N7-MedG/Ag DNA. The remaining slots are all human DNA samples. Panel B, standard curve for N7-MedG quantitation.
20 cases resulting in 20 matched case-control groups; the remaining 20 controls could not be matched to any cases. Of these 61 DNA samples, 33 were analysed more than twice (with a mean CV 9.0 F 7.6). The population characteristics of this matched sub-set can be seen in Table 1. No differences between cases and controls were statistically significant, although there was weak evidence that controls were more likely than cases to be HPV-positive at the time of diagnosis ( P = 0.081, OR 0.31; 95% CI 0.08 –1.16). The lack of HPV follow-up data led to a considerable lack of power to detect any difference between cases and controls ( P = 0.21, OR 4.00; 95% CI 0.45 –35.8). N7-MedG was detected in cervical DNA with levels ranging from non-detected (<0.1 Amol/mol dG) to 4.83 Amol/mol dG. A typical example of an immunoslotblot with a standard curve is shown in Fig. 1. The N7-MedG levels for the matched samples are summarised in Table 2. N7MedG levels in cases (geometric mean 0.99 Amol/mol dG) were significantly higher than those found in controls (0.33 Amol/mol dG; P = 0.01). The difference in N7-MedG levels between cases and the matched controls was 1.24 Amol/mol
dG, (95% CI 0.40 –2.08; P = 0.006). Cases were also more likely than controls to have N7-MedG levels greater than the median level (14 out of 20 vs. 6 out of 20, respectively; OR 5.44, 95% CI 1.18 – 27.12). There was no detectable association between N7-MedG levels and HPV infection with the geometric mean N7MedG level (95% CI) in HPV-positive women being 0.44 (0.28 –0.69) Amol/mol dG and 0.91 (0.34 –2.44) Amol/mol dG in women without HPV infection ( P = 0.12). The effect of log N7-MedG on treatment failure was assessed both independently and jointly with HPV status at the time of
Table 2 N7-MedG levels (Amol/mol dG) by treatment failure status with respect to smoking status Cases
All Smokers Non/ex smokers
Controls
n
Geometric mean (95% CI)
n
Geometric mean (95% CI)
20 17 3
0.99 (0.52 – 1.90) 1.04 (0.54 – 2.00) 0.74 (0.002 – 261.8)
21 11 10
0.32 (0.19 – 0.54) 0.27 (0.10 – 0.72) 0.42 (0.23 – 0.64)
608
N.N. Acladious et al. / Gynecologic Oncology 93 (2004) 605–609
diagnosis of CIN using multiple conditional logistic regression modelling. Log N7-MedG content, after adjustment for HPV status at time of treatment, was significantly associated with increased risk of treatment failure (OR 5.74, 95% CI 1.05 – 31.23). For a 10-fold increase in N7-MedG content of cervical DNA, there was an estimated 5- to 6-fold increase in the odds of CIN treatment failure. N7-MedG is a specific marker of exposure to methylating agents, including those present in cigarette smoke [21]. Current smokers had similar N7-MedG levels to those of current non-smokers with a geometric mean (95% CI) of 0.61 (0.34 – 1.09) and 0.48 (0.24 – 0.97) Amol/mol dG, respectively ( P = 0.62). There was no smoking-related difference in adduct levels in case or control populations analysed individually (Table 2). In current smokers, there was no association between N7-MedG levels and numbers of cigarettes smoked per day (r = 0.24; P = 0.22). In addition, there was no significant difference in the distribution of adduct levels as 16/28 (57.1%) smokers and 4/12 (33.3%) non-smokers had adduct levels above the median ( P = 0.17). The lack of an association between smoking and N7-MedG levels may reflect the small sample size, or alternatively, it could suggest that women in this study were misreporting their own smoking habit or that other sources of exposure, for example, dietary may be important [22,23]. A more specific exposure marker, such as cotinine in urine or serum, would help to identify the contribution of tobacco-specific nitrosamines [24] to cervical DNA alkylation. Alkylating agents are known mutagens and carcinogens that can cause cancer in humans at high doses [25]. Exposures, at lower doses, will lead to DNA damage but evidence linking such exposures to human cancers is more limited [26]. Pre-treatment levels of N7-MedG in cervical DNA were associated with an increased risk of treatment failure, which given the prospective nature of the study, provides strong evidence of a causal relationship between exposure and outcome. That such an association was detected within 2 years of treatment may highlight the synergistic nature of the interaction between a chemical carcinogen and a virus. However, the main limitation of this study is the small numbers of biological samples that were analysed which, for example, precluded a detailed investigation of the joint effect of smoking and the level of N7MedG on the risk of treatment failure. Whilst it is possible that the results may have occurred by chance or by some other (unaccounted for) confounding factor, the data are consistent with the known carcinogenic nature of methylating agents. Further work is thus warranted to confirm this association.
Acknowledgment Work in the laboratory was supported by Cancer Research UK.
References [1] Soutter WP, de Barros Lopes A, Fletcher A, Monaghan JM, Duncan ID, Paraskevaidis E, et al. Invasive cervical cancer after conservative therapy for cervical intraepithelial neoplasia. Lancet 1997; 349:978 – 80. [2] Andersen ES, Pedersen B, Nielsen K. Laser conization: the results of treatment of cervical intraepithelial neoplasia. Gynecol Oncol 1994; 54:201 – 4. [3] Shafi MI, Dunn JA, Buxton EJ, Finn CB, Jordan JA, Luesley DM. Abnormal cervical cytology following large loop excision of the transformation zone: a case controlled study. Br J Obstet Gynaecol 1993;100:145 – 8. [4] Cox T. Management of cervical intraepithelial neoplasia. Lancet 1999;53:857 – 9. [5] Acladious NN, Sutton C, Mandal D, Hopkins R, Zaklama M, Kitchener H. Persistent human papillomavirus infection and smoking increase risk of failure of treatment of cervical intraepithelial neoplasia (CIN). Int J Cancer 2002;98:435 – 9. [6] Nakao Y, Yang X, Yokoyama M, Pater MM, Pater A. Malignant transformation of human ectocervical cells immortalized by HPV 18: in vitro model of carcinogenesis by cigarette smoke. Carcinogenesis 1996;17:577 – 83. [7] Li SL, Kim MS, Cherrick HM, Doniger J, Park NH. Sequential combined tumorigenic effect of HPV-16 and chemical carcinogens. Carcinogenesis 1992;13:1981 – 7. [8] Garrett LR, Perez-Reyes N, Smith PP, McDougall JK. Interaction of HPV-18 and nitrosomethylurea in the induction of squamous cell carcinoma. Carcinogenesis 1993;14:329 – 32. [9] Ali S, Astley SB, Sheldon TA, Peel KR, Wells M. Detection and measurement of DNA adducts in the cervix of smokers and nonsmokers. Int J Gynecol Cancer 1994;4:188 – 93. [10] Simons AM, Mugica van Herckenrode C, Rodriguez JA, Maitland N, Anderson M, Phillips DH, et al. Demonstration of smoking-related DNA damage in cervical epithelium and correlation with human papillomavirus type 16, using exfoliated cervical cells. Br J Cancer 1995;71:246 – 9. [11] King MM, Hollingsworth A, Cuzick J, Garner RC. The detection of adducts in human cervix tissue DNA using 32P-postlabelling: a study of the relationship with smoking history and oral contraceptive use. Carcinogenesis 1994;15:1097 – 100. [12] Phillips DH, She MN. DNA adducts in cervical tissue of smokers and non-smokers. Mutat Res 1994;313:277 – 84. [13] Melikian AA, Sun P, Prokopczyk B, El-Bayoumy K, Hoffmann D, Wang X, et al. Identification of benzo[a]pyrene metabolites in cervical mucus and DNA adducts in cervical tissues in humans by gas chromatography-mass spectrometry. Cancer Lett 1999; 146:127 – 34. [14] Patel KR, Astley S, Adams DJ, Lacey CJ, Ali SW, Wells M. Expression of cytochrome P450 enzymes in the cervix. An immunohistochemical study. Int J Gynecol Cancer 1993;3:159 – 63. [15] Farin FM, Bigler LG, Oda D, McDougall JK, Omiecinski CJ. Expression of cytochrome P450 and microsomal expoxide hydrolase in cervical and oral epithelial cells immortalized by human papillomavirus type 16 E6/E7 genes. Carcinogenesis 1995;16: 1670 – 4. [16] Warwick A, Sarhanis P, Redman C, Pemble S, Taylor JB, Ketterer B, et al. Theta class glutathione S-transferase GSTT1 genotypes and susceptibility to cervical neoplasia: interactions with GSTM1, CYP2D6 and smoking. Carcinogenesis 1994;15:2841 – 5. [17] Ross RK, Yuan J-M, Yu MC, Wogan GN, Qian G-S, Tu J-T, et al. Urinary aflatoxin biomarkers and risk of hepatocellular carcinoma. Lancet 1992;339:943 – 6. [18] Prokopczyk B, Cox JE, Hoffmann D, Waggoner SE. Identification of tobacco-specific carcinogen in the cervical mucus of smokers and nonsmokers. J Natl Cancer Inst 1997;89:868 – 73.
N.N. Acladious et al. / Gynecologic Oncology 93 (2004) 605–609 [19] Haque K, Cooper DP, van Delft JHM, Lee SM, Povey AC. Accurate and sensitive quantitation of N7-methyldeoxyguanosine-3V-monophosphate by 32P-postlabelling and storage-phosphor imaging. Chem Res Toxicol 1997;10:660 – 6. [20] Harrison KL, Wood M, Lees NP, Hall CN, Margison GP, Povey AC. Development and application of a sensitive and rapid immunoassay for the quantitation of N7-methyldeoxyguanosine in DNA samples. Chem Res Toxicol 2001;14:295 – 301. [21] Bianchini F, Wild CP. 7-methyldeoxyguanosine as a marker of exposure to environmental methylating agents. Toxicol Lett 1994; 72:175 – 84. [22] Simons AM, Phillips DH, Coleman DV. Damage to DNA in cer-
[23] [24] [25]
[26]
609
vical epithelium related to smoking tobacco. Br Med J 1993; 306:1444 – 8. Sillett RW, Wilson MB, Malcolm RE, Ball KP. Deception among smokers. Br Med J 1978;2:1185 – 6. Hecht SS. DNA adduct formation from tobacco-specific N-nitrosamines. Mutat Res 1999;424:127 – 42. Leone G, Voso MT, Sica S, Morosetti R, Pagano L. Therapy related leukemias: susceptibility, prevention and treatment. Leuk Lymphoma 2001;41:255 – 76. Hecht SS. Approaches to cancer prevention based on an understanding of N-nitrosamine carcinogenesis. Proc Soc Exp Biol Med 1997;216:181 – 91.