Detection of cervical abnormalities in a developing country using measurement of Brn-3a in cervical smears

Gynecologic Oncology 100 (2006) 89 – 94 www.elsevier.com/locate/ygyno

Detection of cervical abnormalities in a developing country using measurement of Brn-3a in cervical smears Daniel Ndisang a,*, Felipe Lorenzato b, Michael Sindos c, Albert Singer c, David S. Latchman a

a,d

Medical Molecular Biology Unit, Institute of Child Health, University College London, 30 Guilford Street, London WC1N 1EH, UK b Instituto Materno Infantil de Pernambuco (IMIP), Rua dos coelhos, 300 Boa Vista Recife-Pernambuco, Brazil c Department of Women’s and Children’s Health, The Whittington Hospital, Highgate Hill, London N19 5NF, UK d Birkbeck, University of London, Malet Street, London WC1E 7HX, UK Received 6 June 2005 Available online 29 August 2005

Abstract Objective. We have previously shown that Brn-3a is elevated in biopsies from women in the United Kingdom with high grade cervical neoplasia, and that it specifically trans-activates the HPV URR in vitro and in vivo. The aim of this study is to establish the relationship of Brn-3a, HPV E6 and pathological diagnosis in cervical smear from women in a developing country with a high prevalence of cervical disease. This is a follow-up of our previous work in which for the first time Brn-3a and E6 levels in cervical smears from women in United Kingdom were shown to correlate with the histological diagnosis of cervical neoplasia and were even better in predicting underlying premalignant disease than conventional procedures. Method. Cervical smears from 295 women with cervical abnormalities attending gynecological clinics in Brazil were used to make RNA. The mRNA levels of Brn-3a and HPV E6 were measured and the data obtained were used to establish the relationship between Brn-3a and the histological diagnosis. Results. The cellular transcription factor Brn-3a was readily measured in cervical smears from the Brazilian population. Its presence was shown to be frequently associated with the expression of HPV E6. The measured level of Brn-3a parallels the severity of the cervical ailment and predicts the pathological categories. Conclusions. The ability of Brn-3a to predict for cervical ailments is independent to the geographical characteristics of the population, and hence it could be used routinely as an adjunct to colposcopy and pathological diagnosis in developing and developed countries. D 2005 Elsevier Inc. All rights reserved. Keywords: Brn-3a transcription factor; Pap smear; Trans-activation; HPV-16; cervical abnormality; Cervicitis; URR; E6

Introduction Cervical cancer is the second most common cancer in women world wide, about 0.5 million new cases are reported each year with nearly 80% in developing countries: 68,000 in Africa, 77,000 in Latin America and 245,000 in Asia [1,2]. In many developed countries, cervical cancer programs have not only reduced significantly the incidence Abbreviations: Upstream regulatory region, (URR); Early open reading frame, (EORF). * Corresponding author. Fax: +44 20 7905 2301. E-mail address: [email protected] (D. Ndisang). 0090-8258/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2005.07.109

of this ailment, but also helped in reducing associated mortality [3]. However, accumulated costs of screening, diagnosis, treatment and management of cervical cancer in developing countries have prevented many women from participating fully in these programs [4– 6]. Thus, in Brazil, mortality due to cervical cancer is very high as compared to the UK [5 –7], even though a high proportion of women in Brazil over 45 years have had at least one cervical cancer test [5 –7]. It is therefore imperative that novel cost-effective cervical programs or adjuncts that would increase the efficacy of conventional procedures such as colposcopy and pathology be made available in the developing world [7 –9]. Unfortunately, the current methods used commonly

90

D. Ndisang et al. / Gynecologic Oncology 100 (2006) 89 – 94

in developing countries are similar to those employed in developed countries. Most of these methods are not only invasive but depend on expensive referrals or follow-up programs to reduce cervical mortality [7,8,10]. Interestingly, it is well established that infection with HPV is directly associated with the natural history of cervical cancer, and recently WHO mandated HPV as the primordial cause of cervical cancer [10,11]. HPV is transmitted sexually and there are many different types of HPV associated with human genital infections [12,13,14]. Of these, the oncogenic HPV-16 and HPV-18 are the most common types directly associated with cervical cancer [12]. These viruses play a key role in cervical tumorigenesis by encoding specific transforming EORF (early open reading frame) oncoproteins E6 and E7 [15 –18]. Cellular factors are able to bind to specific motifs of the URR (upstream regulatory region) which regulates the production of E6 and E7 oncoproteins. Brn-3a is member of the POU family of cellular transcription factors, expressed in cervical cells [19,20]. We have previously shown that Brn-3a is highly elevated in biopsies from women with CIN3 [21,22] and that Brn-3a trans-activates the URR driving expression of the EORF E6/ E7 gene in cervical cells [17 –20]. Moreover, artificial reduction of Brn-3a levels significantly reduces HPV EORF E6/E7 gene expressions, growth rate and anchorage independence in vitro and tumor growth in vivo [23,24]. This suggests a critical role for the Brn-3a transcription factor in the natural history of cervical cancer. Moreover, in our previous studies in the United Kingdom, we showed that measurement of Brn-3a in cervical smears not only improved the detection of high grade disease better than most conventional procedures, but also improved the detection of underlying cervical neoplasia in women referred with persisting borderline or mildly dyskaryotic smears [25,26]. It was also shown that elevated Brn-3a levels correlate with the presence of HPV activity as measured by E6 mRNA and cervical abnormality, which was critical for its ability as a diagnostic and prognostic marker [21,25,26]. Our previous studies were done in the United Kingdom representing a developed country with good screening procedures. It was important therefore to ascertain whether Brn-3a and E6 could be measured in cervical smears in Brazil, representing a developing country with a different population from that of the UK.

Materials and methods Patients The study group in Recife, Brazil, which has the highest rate of cervical disease in Brazil and one of the highest in the world [5,7,8], consisted of 295 women with cervical abnormalities attending one of the following gynecological units: (1) Instituto Materno Infantil De Pernambuco (IMIP)

Recife-Pernambuco Brazil, (2) The Gynaecological Oncology Research Unit Bar de Lucena Hospital-Recife and (3) The Cancer Hospital Recife. The mean age for the study group was 40.2 years (the median 38 years and the range 17– 76 years). Ethical permission was obtained from the hospital’s ethics committee after review of the study protocol. All the women provided informed consent to obtain and use their cervical material for the purpose of the study. Most of the women acknowledged the desire to participate for a long-term follow-up program. Furthermore, Recife has a very high prevalence of cervicitis. Cervicitis is an inflammation of the uterine cervix usually attributed to infection caused by sexually transmitted diseases [28 – 30]. This cervical abnormality is very common, affecting more than half of all women at some point during their adult lives [29,30]. Increased risk is associated with HPV infection as a result of high-risk sexual behavior [28 –30]. However, cervicitis could progress to cervical neoplasia if left untreated [28 – 30]. Unfortunately the knowledge of this abnormal cervical ailment in the developing world is very scanty due to poor diagnosis [5,28]. For this reason, the cost of managing such an obscure cervical disease in developing countries using conventional procedures is very high [6– 8]. Therefore, to investigate whether Brn-3a could be measured in cervicitis and thus select for women with this ailment, one hundred and forty cervicitis samples were collected from the three different catchments in Brazil. Study protocol Cervical cells were obtained and processed using our previously described routine methodology but with some slight modifications to reduce cost [25,26]. In brief, two cervical smears were obtained with a spatula and/or cytobrush: the first for cytology and the second for Brn-3a and HPV-16 E6 analysis. The second specimen was placed in a 10-ml 1 PBS and stored on ice for less than 30 min prior to a 2-min 1500  g centrifugation (in developing countries, PBS is a cheaper but effective alternative to RPMI media). The concentrated cervical cells recovered were used for RNA extractions. The Brn-3a mRNA and HPV-16 E6 mRNA levels were determined subsequently using our previously described RT-PCR assay [25]. To express the levels of Brn-3a and E6 mRNA, we have taken an arbitrary scale of values in the range 0.0 – 1.0, with 0.0 representing a lack of detection of mRNA and 1.0 the highest achievable level. Thus, Brn-3a positivity was defined as presence of the lowest detectable level or above Brn-3a mRNA (0.1 or above).

Results In this study, 295 cervical smears from women attending the gynecological units in Recife were analyzed for the

D. Ndisang et al. / Gynecologic Oncology 100 (2006) 89 – 94

presence of Brn-3a mRNA and HPV E6 mRNA. Brn-3a was positive in 214 (73%) of the cervical samples while HPV E6 was positive in 198 (67%). A linearity with an absolute R = 0.7593 was observed between Brn-3a and E6 using Microsoft Excel ordinal regression statistical analysis (Fig. 1). To determine whether Brn-3a as an adjunct to pathological diagnosis could predict cervical disease in Recife, Brazil, the levels of Brn-3a and E6 were correlated with histological analysis, and their respective mean levels in each histological category tabulated as shown in Table 1. The results showed that, as in our previous work in London, UK, the mean levels of Brn-3a and E6 increase with the severity of the cervical disease. By using the Student’s t test, it was observed that there was a significant difference between the mean of Brn-3a in the different histologic categories ( P < 0.05). This observation was also true for E6 ( P = 0.004), though not as strongly as for Brn-3a ( P = 0.001). Brn-3a and E6 were also significantly higher in women with lesions as compared to women with no cervical disease (LGSIL P = 0.001 for Brn-3a and P = 0.003 for E6; HGSIL P = 0.00001 for Brn-3a and P = 0.0003 for E6; cancer P = 0.00005 for Brn-3a and P = 0.00005). This correlates with the work of Sindos et al. [25] in the UK who showed that the presence of Brn-3a in cervical smears is associated with 300% greater risk of HGSIL as compared to a negative Brn-3a smear. Moreover, using Student’s t test assuming unequal variance, we showed with the Brazilian population that Brn-3a was able to discriminate strongly between the histologic categories of LGSIL and HGSIL ( P = 0.01, CI 95%) as compared to E6 ( P = 0.04, CI 95%). Importantly, the segregation of LGSIL and HGSIL using conventional procedure has always been difficult and as a result costly. Nevertheless, as Brn-3a could discriminate between the different HGSIL and LGSIL categories, it could be possible to tailor a cost-effective diagnosis for pre-cancer lesions using Brn-3a as an adjunct to other classical procedures in referring women for further treatment. Furthermore, to establish whether Brn-3a could predict cervical diseases as an adjunct to colposcopy, the levels of

91

Table 1 Brn-3a and E6 Levels in Pap smears from patients categorized on the basis of the histological diagnosis Category

Count

Percentage

Mean Brn-3a

Mean E6

Negative LGSIL(HPV-CIN1) HGSIL(CIN2-CIN3) Cancer Total

34 28 48 43 153

22.2 18.3 31.4 28.1 100

0.17* 0.38* 0.57*,b 0.5*

0.16a 0.36a 0.38a,b 0.47a

a Student’s t test for differences between the mean of non-disease and disease for E6, in all categories P < 0.05. b Student’s t test for differences between the mean of Brn-3a mRNA and E6 mRNA, only where P < 0.05. * Student’s t test for differences between the mean of non-disease and disease for Brn-3a, in all categories P < 0.05.

Brn-3a and E6 were correlated according to colposcopic impression [26,27] (Table 2). The results indicated that Brn3a and E6 levels progressively increased along the classification gradient, samples with normal colposcopy having the lowest Brn-3a and E6 values, via intermediate values for minor lesion and finally, higher values for moderate to major lesions. Again, the Student’s t test showed significant differences between the mean values for Brn-3a in the colposcopic categories ( P < 0.05 between categories). This was also true for the E6 values ( P < 0.05 between categories). To investigate if Brn-3a and E6 could discriminate between the minor and moderate lesions, where colposcopic classifications and management could be difficult, we used similar Student’s t test with unequal variance at 95% confidence interval which showed P = 0.02 for Brn-3a and E6 P = 0.03. This showed that Brn-3a was not only significantly higher in moderate lesions than in minor lesions, but it also discriminated better than E6 between these categories. Thus, Brn-3a was able to discriminate between these two problematic colposcopic categories in Brazil, as it did in our previous studies in the UK. Furthermore, Brn-3a was positive in 96 (68.6%) of the cervicitis samples while E6 was positive in 80 (57.1%) of them. Interestingly, the mean Brn-3a value for cervicitis was higher than the mean Brn-3a value for the normal or negative samples but lower than the mean values for the disease categories (Table 3). A similar trend was also observed with the E6 values. This shows that Brn-3a and E6 means progressively increased from non-disease via cervicitis to Table 2 Brn-3a and E6 levels in samples categorized on the basis of colposcopy Category

Fig. 1. Correlation of Brn-3a and HPV E6 in cervical smears from women in Recife, Brazil. The levels of HPV E6 mRNA and Brn-3a mRNA were determined by reverse transcriptase-polymerase chain reaction assay in disease and non-disease cervical smears from 295 women. The invariant cyclophilin mRNA levels were recorded to allow normalization of the data. The graph and statistical analysis were carried out using Microsoft Excel.

Normal Minor lesion Moderate lesion Major lesion/cancer Total

Count 33 35 45 40 153

Percentage 21.6 22.9 29.4 26.1 100

Mean 0.14* 0.37* 0.53* 0.50*

Brn-3a

Mean E6

a

0.12 0.26a 0.50a 0.46a

a Student’s t test for differences between the mean of non-disease and disease for E6, in all categories P < 0.05. * Student’s t test for differences between the mean of non-disease and disease for Brn-3a, in all categories P < 0.05.

92

D. Ndisang et al. / Gynecologic Oncology 100 (2006) 89 – 94

Table 3 Diagnosis on the basis of Brn-3a and E6 levels in associated with both histology and colposcopy Category

Count

Normal/negative Cervicitis Low grade High grade Cancer Total

(34 + Z33) 15.0 140 31.4 (28 + 35) 14.1 (48 + 45) 20.9 (43 + 40) 18.6 140+ (153 + 153) 100

Percentage Mean Brn-3a Mean E6 0.15* 0.3* 0.37* 0.55*,b 0.5*

0.14* 0.26a 0.31a 0.44a,b 0.46a

a

Student’s t test for differences between the mean of non-disease and disease for E6, in all categories P < 0.05. b Student’s t test for differences between the mean of Brn-3a mRNA and E6 mRNA, only where P < 0.05. * Student’s t test for differences between the mean of non-disease and disease for Brn-3a, in all categories P < 0.05.

low grade disease. Hence, to investigate whether Brn-3a levels in cervicitis correlates with E6 levels, an ordinal regression analysis was performed. A strong linear correlation was observed between the levels of Brn-3a mRNA and E6 mRNA (R = 0.7365) (Fig. 2), suggesting that Brn-3a is important in the etiology of cervicitis by regulating the transcription of HPV EORF E6 transforming protein. Thus, to investigate the diagnostic potential of Brn-3a in association with colposcopy and histology for non-disease (negative and normal) and low grade cervical diseases (cervicitis, minor legions), further analysis using ANOVA t test was performed. The result showed that there was a significant difference between Brn-3a mean value for nondisease and cervicitis ( P = 0.001, for E6 P = 0.004) as well as between cervicitis and minor lesions ( P = 0.00001, for E6 P = 0.0002). Hence, Brn-3a either alone or in conjunction with E6 could strongly select for cervicitis among low grade diseases such as LGSIL and from women with no cervical abnormality. Taking the results together this indicates that Brn-3a is a potential diagnostic maker that cannot only be employed routinely in predicting persistent underlying high grade and moderate lesions but also low grade cervical abnormality such as cervicitis.

UK, where Brn-3a was shown to discriminate between the histological HGSIL category better than HPV DNA detection [21,22,25,26]. Thus, taking the results together, Brn-3a’s ability to select for high grade and underlying cervical abnormality is independent to the geographical characteristic of the population and hence could be used routinely as an adjunct to colposcopy and pathological diagnosis in developed and developing countries. In our study in Brazil, 85% of women had a form of cervical abnormality. Of these, 70% presented abnormal smears that were positive for Brn-3a and 65% positive for E6. Moreover, Recife also has a high prevalence of cervicitis accounting for 47.4% of the Pap smears analyzed. Brn-3a was positive in 68.6% of cervicitis samples while E6 was positive in 57.1%. The mean Brn-3a value for cervicitis was lower than that for the low grade lesions but higher than the mean Brn-3a value for non-disease. This observation was similar for E6. Taking the results together, this suggests a transient HPV activity in cervicitis, regulated via the URR by the Brn-3a cellular transcription factor. Importantly, studies have shown that some women with minor abnormalities associated with HPV could progress to high-grade disease [27 –29]. It would be expensive, however, to monitor and manage minor abnormalities including cervicitis [30,31]. In the developed world such as the UK, the current guide lines call for women to be kept under cytologic surveillance and only referred for colposcopy if a repeat smear is abnormal on at least two or three occasions depending on the severity of diagnosis. This policy, however, is expensive and very difficult to administer. Moreover, in a developing country like Brazil, such a policy could not be realized because of high cost. Nevertheless, as we have shown here and in accordance with our previous work, Brn-3a could contribute significantly to the solution of this problem, since it indicates the real-time activity of oncogenic HPV viruses as opposed to HPV DNA testing. This is because HPV DNA is necessary but not sufficient to cause cervical cancer, being detected in

Discussion In the work described here, we have demonstrated that Brn-3a can be measured in cervical smears from a population in a developing country and its presence is closely associated with HPV mRNA positivity. This is in accordance with our previous findings where Brn-3a levels in cervical smears from a developed country were shown to correlate with the natural history of cervical cancer and was even better at predicting the underlying pre-malignant disease than conventional procedures [25,26]. Moreover, we have shown here that Brn-3a levels could assist in the management of different pathological categories. Interestingly, these findings are in agreement with our previous work in London,

Fig. 2. Correlation of Brn-3a and HPV E6 in cervical smears from cervicitis patients. The levels of HPV E6 mRNA and Brn-3a mRNA were determined by reverse transcriptase-polymerase chain reaction assay in cervical smears from 140 cervicitis patients in Recife, Brazil. Cyclophilin mRNA levels were recorded to allow normalization of the data. The graph and statistical analysis were carried out using Microsoft Excel.

D. Ndisang et al. / Gynecologic Oncology 100 (2006) 89 – 94

95 –100% of cases [1,2,10,11]. Transcription of the HPV EORF E6/E7 transforming proteins is a consistent feature of cervical carcinomas that has been shown in several studies to be regulated by the Brn-3a POU domain transcription factor [17 – 24]. Thus, the levels of E6 activated by Brn-3a in cervical material are an objective indication of transformation and hence disease. Moreover, artificial reduction of Brn-3a levels significantly reduces HPV EORF E6/E7 gene expressions, growth rate and anchorage independence in vitro and tumor growth in vivo [23,24]. Furthermore, Brn-3a is highly elevated in biopsies from women with CIN3 [21,22] as compared to normal control. Hence, measurement of Brn-3a alone could be a costeffective procedure in selecting for women at risk of rapid progression of disease should they be infected with HPV, thus enabling specific but appropriate treatment and management. This could dramatically reduce the costs for cervical programs not only in developed countries such as the UK, where costs and over treatment are hampering the NHS cervical policy, but also in developing countries such as Brazil where follow-up, treatment and management are very expensive. Further study is needed to ascertain whether certain women are predisposed to accelerated progression of CIN by presenting high Brn-3a levels. In addition, it would be important to determine the level in cervical smears of the antagonistic Brn-3b transcription factor, since Brn-3b has been shown not only to inhibit the activity of Brn-3a in cervical tumorigenesis in vitro and in vivo [23,24] but also to play a significant role in other gynecological oncogenesis [32]. It is also important to investigate the association of an environmental factor such as nicotine and Brn-3a in cervical tumorigenesis. Also, further studies are needed using our current Brn-3a and E6 procedures as well as with other techniques to evaluate the high prevalence of cervicitis among the Brazilian population. Finally, using ordinal regression analysis, a weak linear correlation was observed between the levels of Brn-3a mRNA and E6 mRNA in our UK study (R = 0.607), as compared to our Brazil study where a strong linear correlation was observed between Brn-3a mRNA and E6 mRNA (R = 0.7593) ( P < 0.05). This is in agreement with our studies which show that differences between different HPV-16 variants in their disease-causing abilities [14] are associated with differences in their response to Brn-3a in co-transfection assays and in human cervical smears [33]. This suggests that Brn-3a has a greater efficiency in trans-activating the HPV-16 URR in variants found in Brazil than the URR of the HPV-16 variants found in London, suggesting common and different HPV-16 variants between UK and Brazil. Importantly, Brn-3a was readily measured in cervical smears from women in Brazil, representing a developing country with a different population, compared to women in United Kingdom, and correlates with the severity of the cervical abnormalities. Thus, Brn-3a’s ability to predict for high grade and underlying cervical disease is independent to

93

the geographical characteristics of the population and hence could be used routinely in developed and developing countries as an adjunct to colposcopy and pathological diagnosis.

Acknowledgment This work was supported by the Association for International Cancer Research (AICR).

References [1] Parkin DM, Pisani P, Ferlay J. Global cancer statistics. CA Cancer J Clin 1999;49:33 – 64. [2] WHO Report on Human Papillomavirus and Cervical Cancer (WHO) databases (http://www.who.int/vaccine_research/diseases/hpv/en/). [3] Bosch FX, Manos MM, Munoz N, Sherman M, Jansen AM, Peto J, et al. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. J Natl Cancer Inst 1995;85:796 – 802. [4] Durst M, Gissman L, Ikenberg H, Zur Hausen H. A papilloma virus DNA from a cervical carcinoma and its prevalence in cancer biopsy sample from different geographic regions. Proc Natl Acad Sci U S A 1983;38:12 – 5. [5] Gregg J. Mixed blessings: cervical cancer screening in Recife. Brazil Med Anthropol 2000;19:41 – 63. [6] Azevedo G, Mendonca S. Cancer in the female population in Brazil. Rev Saude Publica 1993;27:68 – 75. [7] Lorenzato FR, Singer A, Ho L, Santos LC, Batista Rde L., Lubambo TM, et al. Human papillomavirus detection for cervical cancer prevention with polymerase chain reaction in self-collected samples. Am J Obstet Gynecol 2002;186:962 – 8. [8] Lorenzato F, Ho L, Terry G, Singer A, Santos LC, De Lucena Batista R, et al. The use of human papillomavirus typing in detection of cervical neoplasia in Recife (Brazil). Int J Gynecol Cancer 2000; 10:143 – 50. [9] Matos YF, Costa HL, Faundes AE. Prevalence of cervical squamous intraepithelial lesions in women with HIV. Int J Gynaecol Obstet 2003;83:63 – 4. [10] Woodman BJ, Collins S, Winter H, Bailey A, Ellis J, Prior P, et al. Natural history of cervical human papillomavirus infection in young women: a longitudinal cohort study. Lancet 2001; 357:1831 – 6. [11] Human papillomaviruses. IARC Monographs on the Evaluation of Carcinogenic Risks to Human 1995;64. Lyon: International Agency for Research in Cancer. [12] Bosch FX, Manos MM, Munoz N, Sherman M, Jansen AM, Peto J, et al. Prevalence of human papillomavirus in cervical cancer a world perspective. J Natl Cancer Intl 1995;85:796 – 802. [13] Durst M, Gissmann L, Ikenberg H, Zur Hausen H. A papilloma virus DNA from a cervical carcinoma and its prevalence in cancer biopsy sample from different geographic regions. Proc Natl Acad Sci U S A 1983;80:3812 – 5. [14] Bible JM, Mant C, Best JM, Kell B, Starkey WG, Shanti G, et al. Cervical lesions are associated with human papillomavirus type 16 intratypic variants that have high transcriptional activity and increased usage of common mammalian codons. J Gen Virol 2000;81: 1517 – 27. [15] Cripe TP, Haugen JP, Turk F, Tabatabai PG, Schmid M, Durst T, et al. Transcriptional regulation of human papilloma virus-16 E6 – E7 promoter by keratinocyte-dependent enhancer and by viral E2 transactivator and repressor gene products: implications for cervical carcinoma. EMBO J 1987;6:3745 – 53. [16] Gloss B, Bernard HU, Seedorf K, Klocl G. The upstream regulatory

94

[17]

[18]

[19]

[20]

[21]

[22]

[23]

D. Ndisang et al. / Gynecologic Oncology 100 (2006) 89 – 94 region of human papilloma virus contains an E2 protein independent enhancer which is specific for cervical carcinoma cells and regulated by glucocorticoid hormones. EMBO J 1987;6:3735 – 43. Morris PJ, Dent CL, Ring CJ, Latchman DS. The octamer-binding site in the HPV 16 regulatory region produces opposite effects on gene expression in cervical and non-cervical cells. Nucleic Acids Res 1993;21:1019 – 23. Dent CL, McIndoe GA, Latchman DS. The constitutively expressed octamer-binding protein OTF-1and a novel octamer binding protein expressed specifically in cervical cells bind to an octamer-related sequence in human papilloma 16 enhancer. Nucleic Acids Res 1991;19:4531 – 5. Lillycrop KA, Budhram-Mahadeo VS, Lakin N, Terrenghi G, Wood JN, Polak JM, et al. A novel POU family transcription factor I closely related to Brn-3 but has a distinct expression pattern in neuronal cells. Nucleic Acids Res 1992;20:5093 – 6. Morris PJ, Theil T, Ring CJ, Lillycrop KA, Mo¨ro¨y T, Latchman DS. The opposite and antagonistic effects of the closely related POU family transcription factors on the activity of a target promoter are dependent upon differences in the POU domain. Mol Cell Biol 1994;14:6907 – 14. Ndisang D, Morris PJ, Chapman C, Ho L, Singer A, Latchman DS. The HPV activating transcription factor Brn-3a is overexpressed in CIN3 lesions. J Clin Invest 1998;101:1687 – 92. Ndisang D, Budhram-Mahadeo VS, Singer A, Latchman DS. Widespread elevated expression of the HPV-activating cellular transcription factor Brn-3a in the cervix of women with CIN3. Clin Sci 2000;98: 601 – 2. Ndisang D, Budhram-Mahadeo VS, Latchman DS. Brn-3a transcription factor plays a critical role in regulating human papilloma virus gene expression and determining the growth characteristics of cervical cancer cells. J Biol Chem 1999;274:28521 – 7.

[24] Ndisang D, Budhram-Mahadeo VS, Pedley B, Latchman DS. The Brn-3a transcription factor plays a key role in regulating the growth of cervical cancer cells in vivo. Oncogene 2001;20:4899 – 903. [25] Sindos MD, Ndisang D, Pisal N, Chow C, Singer A, Latchman DS. Measurement of Brn-3a levels in Pap smears provides a novel diagnostic marker for the detection of cervical neoplasia. Gynaecol Oncol 2003;90:366 – 71. [26] Sindos MD, Ndisang D, Pisal N, Chow C, Deery A, Singer A, et al. Detection of cervical neoplasia using measurement of Brn-3a in cervical smears with persistent minor abnormality. Int J Gynecol Cancer 2003;13:515 – 7. [27] Rieper JP. Value of colposcopy in the management of cervicitis. An Bras Ginecol 1957;44:137 – 40. [28] Anderson JR. Genital tract infections in women. Med Clin North Am 1995;79:1271 – 98. [29] Soper DE. Sexually transmitted disease and pelvic inflammation disease. Prim Care Women 1995339 – 47. [30] Souter WP, Wisdom S, Brough AK, Monaghan JM. Should patients with mild atypia in a cervical smear be referred for colposcopy. Br J Obstet Gynaecol 1986;93:70 – 4. [31] Bolger BS, Lewis BV. A prospective study of colposcopy in women with mild dyskaryosis or koilocytosis. Br Obstet Gynaecol 1988;95: 1117 – 9. [32] Ndisang D, Lee SA, Patel C, Danis J, Faulkes D, D’Arrigo C, et al. Expression of Brb-3b transcription factor correlates with expression of HSP-27 in breast cancer biopsies and is required for maximal activation of HSP-27. Promoter Cancer Res 2005;65: 3072 – 3080. [33] D. Ndisang, D. Faulkes, D. Gascoyne, M. Sindos, A. Singer, C. Cason, and D.S. Latchman, Differential regulation of different human papilloma virus variants by the cellular Brn-3a POU family transcription factor. Oncogene [in press].