Cancer Letters 239 (2006) 98–102 www.elsevier.com/locate/canlet
Papillomavirus infection in the conjunctiva of individuals with and without AIDS: An autopsy series from Uganda Charles Ateenyi-Agabaa, Elisabete Weiderpassb,c,*, Massimo Tommasinod, Anouk Smetd, Annie Arsland, Min Daid, Edward Katongole-Mbiddee, Pierre Hainautd, Peter J.F. Snijdersf, Silvia Franceschid a
Makerere University Medical School, P.O. Box 7072, Kampala, Uganda b Karolinska Institutet, P.O. Box 281, SE-171 77, Stockholm, Sweden c Cancer Registry of Norway, Montebello, N-0310 Oslo, Norway d International Agency for Research on Cancer, 150 cours Albert Thomas, F-69372 Lyon cedex 08, France e Uganda Cancer Institute, P.O. Box 3935, Kampala, Uganda f Vrije University Medical Center, Postbus 7057, NL-1007 MB Amsterdam, The Netherlands Received 24 May 2005; received in revised form 20 July 2005; accepted 26 July 2005
Abstract HIV and genus beta human papillomavirus (HPV) types have been associated with squamous cell carcinoma of the conjunctiva (SCC). To determine whether conjunctival HPV infection is associated with AIDS, we analysed 136 lesion-free eye biopsies and tested for genera alpha, beta and gamma HPV types. Only infections with genera beta and gamma HPV types was found. After adjustment for age and gender, no excess of genera beta or gamma HPV infection was found in individuals who had died of or with AIDS compared to those who had died of other infectious diseases [relative risk (RR)Z1.3; 95% confidence interval (CI): 0.4–4.8], or chronic diseases or trauma (RRZ0.9; 95% CI: 0.3–2.9). Our findings suggest that infection with genera beta or gamma HPV types in lesion-free conjunctivas is common, but not greatly enhanced by the presence of AIDS. q 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Conjunctiva; Papillomavirus; AIDS; Africa; Immunity
1. Introduction Squamous cell carcinoma of the conjunctiva (SCC) is relatively common in some parts of sub-Saharan * Corresponding author. Address: Elisabete Weiderpass, PhD, The Cancer Registry of Norway, Montebello, N-0310 Oslo, Norway. Tel.: C47 2333 3982; fax: C47 2245 1370. E-mail address:
[email protected] (E. Weiderpass).
Africa, including Uganda [1]. In addition to heavy solar ultraviolet radiation, infection with HIV is an established risk factor for SCC [2–4]. HIV infection is likeliest to act via immunosuppression and activation of potentially oncogenic viruses [5], notably human papillomaviruses (HPVs). Hundreds of HPV types have been detected and recently classified into genera [6]. Genus alpha, often referred to as mucosal types, includes HPV types such as HPV 16 and 18, which
0304-3835/$ - see front matter q 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.canlet.2005.07.024
C. Ateenyi-Agaba et al. / Cancer Letters 239 (2006) 98–102
can cause squamous cell cancer of the cervix, of the anogenital tract and of the head and neck [7]. They were the first HPV types suspected to play a role in SCC, but a few studies from Africa [8] and elsewhere [9], as well as [9–11] a seroprevalence study from Uganda [4] failed to support this possibility. Genus beta papillomaviruses, conversely, include HPV types that are consistently found in skin carcinomas of patients with Epidermodysplasia Verruciformis (EV) (e.g. HPV 5, 8, 19–25 and 36–38) [12] and are often referred to as EV types or skin types. Genus gamma HPV types are associated with benign skin lesions and have been little studied so far [6]. We have reported for the first time [8] a 12-fold increased risk of SCC in individuals from Uganda whose conjunctival biopsies harboured genus beta HPV types [7,12]. We also showed, in the same patients, a high proportion of sunlight-related mutations of the tumour suppressor gene TP53 [13]. In our case-control study, however, we did not know whether study participants suffered from HIV infection or AIDS [8]. In order to elucidate whether the excess of HPV infection we found in SCC cases was attributable to concurrent HIV infection rather than to the presence of SCC, we compared the HPV prevalence in conjunctival biopsies of individuals who had died of or with AIDS and individuals who had died of other causes.
2. Materials and methods Biopsy specimens from the conjunctiva were collected between the first and 31 August 2002 from 136 individuals who had died and undergone an autopsy at the New Mulago Hospital, Kampala, Uganda, within 24 h of death (medianZ12 h). Individuals were classified into three groups. (1) AIDS (nZ33): Whenever AIDS or AIDS-defining illnesses were mentioned as the cause of death or in the medical history records. (2) Infectious diseases (nZ 45): When an infection was reported as the cause of death. These mainly included tuberculosis (nZ15), pneumonia (nZ10), meningitis (nZ9) and malaria (nZ 5). (3) Chronic diseases or trauma (nZ58): These mainly included cardiovascular diseases (nZ13), accidents (nZ8), diabetes and respiratory problems (nZ6 each). Four cancer deaths (oesophagus, stomach, lung and cervix) were also included in group 3.
99
Individuals with conjunctival lesions were excluded. A minimal quantity of saline solution was injected close to the limbus and a biopsy was taken from the nasal part of the conjunctiva using a sterile surgical blade. Biopsies were immediately frozen at K80 8C and then shipped to the International Agency for Research on Cancer (IARC). The present study was approved by the Ethical Committees of Mulago Hospital and the IARC. DNA was extracted using BioRobot (Qiagen). To analyse the quality of target DNA, beta (b)-globin gene-specific primers were used and all conjunctival biopsies were found to be b-globin-positive. To reduce the risk of contamination during DNA extraction, preparation of the polymerase chain reaction (PCR) mix and PCR were performed in three different rooms. To monitor the possible occurrence of cross contamination between different samples during the DNA extraction and/or PCR, one negative control (tubes containing only distilled water) every 10 samples was included. No contamination of the negative controls was detected during the study. The prevalence of genus alpha and genera beta/gamma HPV types was evaluated using two independent assays that are highly specific for each HPV subgroup. Testing for genera beta/gamma HPV types was performed at the laboratory of Infections and Cancer Biology at IARC, Lyon, France, by means of a newly developed consensus PCR system using five overlapping forward and four overlapping reverse primers that encompass the region from nucleotide position 6539– 6610 of the HPV 4 genome. Typing was performed by reverse line blotting [14,15] for genus beta HPV types 5, 8, 9, 12, 14, 15, 17, 19, 20–25, 36–38, 47 and 49, and genus gamma types 4, 48, 50, 60 and 65. Testing for 36 genus alpha HPV types, including HPV 6, 11, 16, 18, 26, 31, 33–35, 39, 40, 42–45, 51– 59, 61, 66, 68, 70, 71 (equivalent to CP8061), 72, 73, 81 (equivalent to CP8304), 82 (IS39 and MM4 subtypes), 83 (equivalent to MM7), 84 (equivalent to MM8) and CP6108 was performed in the laboratory of Molecular Pathology of the Vrije University Medical Center, Amsterdam, The Netherlands, using general primer-mediated GP5C/6CPCR and by hybridisation of PCR products in an enzyme immunoassay using two oligoprobe cocktails [14]. Finally, to evaluate the presence of TP53 mutations, DNA was extracted and exons 5–9 of
100
C. Ateenyi-Agaba et al. / Cancer Letters 239 (2006) 98–102
TP53, including splice junctions, were screened for the presence of a somatic mutation by Denaturing High Performance Liquid Chromatography using primers and conditions described elsewhere [13]. 2.1. Statistical analysis The association between HPV infection and various characteristics of study individuals was evaluated using univariate and multivariate odds ratios (ORs) and the corresponding 95% confidence intervals (CIs). Multivariate ORs were computed by means of unconditional multiple logistic regression including terms for age (!25, 25–34, 35–44, R45), sex and cause of death (i.e. AIDS, infectious diseases other than AIDS, and chronic diseases or trauma).
3. Results Individuals who had died of AIDS were significantly younger (median age Z30; range 15–58 years) than either those who had died of infectious diseases
other than AIDS (median age Z40; range 17–85 years; c21 for trend Z14.6; p!0.001) or of chronic diseases or trauma (median age Z44; range 15–94 years; c21 Z8.0; pZ0.005). Women represented 51.5% of individuals who had died of AIDS, but 35.6% (c21 Z2.0; pZ0.16) and 27.6% (c21 Z5.2; pZ 0.02), respectively, of those who had died of other infectious diseases and chronic diseases or trauma. Overall, conjunctival infection with genera beta/gamma HPV types was found in 14.7% of individuals (Table 1). In the univariate analysis, age greater than 25 years, female gender, indoor occupation and AIDS-related death (OR Z2.0; 95% CI: 0.6–6.2) were each associated with slightly increased ORs for HPV infection, but all 95% CIs were broad and included unity. In the multivariate analysis, that included age, gender and cause of death, there was no significant difference HPV prevalence by cause of death (OR among individuals who had died of AIDS and from other infectious diseases versus individuals who had died of chronic diseases or trauma Z1.3; 95% CI: 0.4–4.8 and 0.9; 95% CI: 0.3–2.9, respectively, Table 1).
Table 1 Odds ratios (ORs) for infection with genera beta/gamma human papillomavirus (HPV) types in biopsies from the conjunctiva and corresponding 95% confidence intervals (CIs) according to various characteristics: 136 autopsies, Kampala, Uganda, 2002 Total no of subjects
HPV positive No
Age (years) !25 25–34 35–44 R45 Sex Male Female Tribe Other than Ganda Ganda Occupation Outdoors Mainly indoors Cause of death Chronic disease or trauma Infectious diseases AIDS Total
%
Univariate OR (95% CI)
Multivariate OR (95% CI)a
16 38 30 45
1 7 7 5
6.3 18.4 23.3 11.1
1 3.4 (0.4–30.1) 4.6 (0.5–41.0) 1.9 (0.2–17.4)
1 3.3 (0.4–29.4) 5.0 (0.5–46.2) 2.3 (0.2–21.8)
87 49
11 9
12.6 18.4
1 1.6 (0.6–4.1)
1 1.3 (0.5–3.7)
52 79
8 12
15.4 15.2
1 0.98 (0.4–2.6)
– –
89 36
11 7
12.4 19.4
1 1.7 (0.6–4.8)
– –
58 45 33 136
7 6 7 20
12.1 13.3 21.2 14.7
1 1.1 (0.3–3.6) 2.0 (0.6–6.2)
1 0.9 (0.3–2.9) 1.3 (0.4–4.8)
Some figures do not add up to the total because of some missing values. a Adjusted for age, sex and cause of death as appropriate.
C. Ateenyi-Agaba et al. / Cancer Letters 239 (2006) 98–102 Table 2 Detection of genera beta/gamma human papillomavirus (HPV) types in biopsies from the conjunctiva by type and multiplicity of infection: 136 autopsies, Kampala, Uganda, 2002 HPV types Absent Present HPV 8 HPV 12 HPV 14 HPV 15 HPV 17 HPV 19 HPV 20 HPV 21 HPV 22 HPV 23 HPV 36 HPV 60 HPV 65 a b
Single
Multiple
No
No
Total No
%
– 16 4a 1 – – 1a 1 1a – 3 3b – 1 1
– 4 2 – 1 1 1 1a – 1 4a – 1 – –
116 20 6 1 1 1 2 2 1 1 7 3 1 1 1
85.3 14.7 4.4 0.7 0.7 0.7 1.5 1.5 0.7 0.7 5.2 2.2 0.7 0.7 0.7
Including one individual with AIDS. Including three individuals with AIDS.
Eleven genus beta and two genus gamma types were detected in either single- (nZ16) or multipletype (nZ4) infections (Table 2). The most frequently found types were HPV 22 (nZ7), HPV 8 (nZ6), HPV 23 (nZ3), and HPV 17 and 19 (nZ2 each). Of the individuals who died of AIDS, six had single-type infections and one had a multiple-type infection. HPV 23 was found in three of these individuals, and HPV 8, 17, 19, 20, and 22 were all found in one individual. All 136 conjunctival biopsies tested negative for genus alpha HPV types. The analysis of TP53 mutations was started with 26 individuals who had died of or with AIDS and had a sufficient amount of DNA left after HPV testing. Since, all these individuals had wild-type TP53, the search for mutations was discontinued.
4. Discussion Use of post-mortem specimens allowed us to obtain, for the first time, information on the HPV status of conjunctival biopsies from individuals who did not have any clinical eye lesions. The main findings of our present study are that: (1) genus beta HPV types, but not genus alpha types, were detectable
101
in a substantial proportion (13%) of conjunctival biopsies from adults of both sexes; and (2) the prevalences of genera beta/gamma HPV types did not vary substantially between individuals with and without AIDS. The frequent detection of genus beta HPV types in SCC patients [8] points, therefore, to a possible role for these types in cancer aetiology, rather than to a confounding effect of HIV infection that is often associated with SCC [3,4]. Some misclassification of the cause of death is possible between individuals who had died of or with AIDS and individuals who had died of infectious diseases other than AIDS, as some of the latter-mentioned might have had undiagnosed HIV infections, or other immune impairment. No clear difference in the prevalence of genera beta/ gamma HPV types was found, however, between individuals who had died of or with AIDS and those whose cause of death was chronic disease or trauma, for whom misclassification is much less likely. Taken together, therefore, our findings suggest that infection with genera beta/gamma HPV types in the conjunctiva is not greatly enhanced by immune impairment. Our study also expanded to previous negative findings in normal conjunctiva in respect to the lack of genus alpha HPV types in dyplastic and cancerous lesions of the conjunctiva [8,9]. HPV 6 or 11 were, however, detected in benign papillomas of the conjunctiva [9] and, possibly, the pterygum [8]. HPV types that belong to the genus beta represent an extremely broad and heterogeneous group [6] and have been evaluated previously using complicated sets of different PCR assays [8]. The major strengths of the present study include, therefore, the use of a new PCR system that can detect a broad range of these HPV types. HPV 8, 22 and 23 seemed to predominate. Interestingly, HPV 22 and 23 are so closely related to HPV 38 that we detected them, using different sets of PCR primers, in SCC [8] and non-melanomatous skin cancer [15]. Infections with multiple genus beta HPV types were relatively rare, even in individuals who died of AIDS. Although there seems to be a lack of HPV infection among individuals below 25 years of age, no significant trend of increase in HPV prevalence emerged with increasing age, possibly on account of the small study size. Gender, tribe and heavy sun
102
C. Ateenyi-Agaba et al. / Cancer Letters 239 (2006) 98–102
exposure, as indicated by outdoor occupation, did not seem associated with HPV prevalence. Conversely, heavy sun exposure was strongly associated with SCC risk in Uganda [4,8]. The causal role of solar ultraviolet light in SCC is confirmed, at a molecular level, by the elevated frequency of TP53 mutations (56% of SCC biopsies) and, particularly, by the high prevalence of tandem CC-to-TT transitions [13]. We did not find, however, evidence that TP53 mutations might be present early in biopsies of healthy conjunctiva, even in individuals with AIDS who were heavily exposed to the sun.
Acknowledgements This study was supported by grants from the Swedish International Development Cooperation Agency (SIDA) for the collaboration between Makerere University (Kampala, Uganda) and the Karolinska Institutet (Stockholm, Sweden).
References [1] D.M. Parkin, J. Ferlay, M. Hamdi-Cherif, F. Sitas, J.O. Thomas, H. Wabinga, et al., Cancers of the eye and orbit in: D.M. Parkin, J. Ferlay, M. Hamdi-Cherif, F. Sitas, J.O. Thomas, H. Wabinga, S.L. Whelan (Eds.), Cancer in Africa: Epidemiology and Prevention, IARC Press, Lyon, 2003, pp. 282–285. [2] C. Ateenyi-Agaba, Conjunctival squamous-cell carcinoma associated with HIV infection in Kampala, Uganda, Lancet 345 (1995) 695–696. [3] R. Newton, A review of the aetiology of squamous cell carcinoma of the conjunctiva, Br. J. Cancer 74 (1996) 1511–1513. [4] R. Newton, J. Ziegler, C. Ateenyi-Agaba, L. Bousarghin, D. Casabonne, V. Beral, et al., The epidemiology of conjunctival squamous cell carcinoma in Uganda, Br. J. Cancer 87 (2002) 301–308.
[5] IARC, Monographs on the Evaluation of the Carcinogenic Risks to Humans Volume 67: Human Immunodeficiency Viruses and Human T-cell Lymphotropic Viruses, IARC Press, Lyon, 1996. [6] E.M. de Villiers, C. Fauquet, T.R. Broker, H.U. Bernard, H. zur Hausen, Classification of papillomaviruses, Virology 324 (2004) 17–27. [7] IARC, Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 64: Human Papillomavirus, IARC Press, Lyon, 1995. [8] C. Ateenyi-Agaba, E. Weiderpass, A. Smet, W. Dong, M. Dai, B. Kahwa, et al., Epidermodysplasia verruciformis human papillomavirus types and carcinoma of the conjunctiva: a pilot study, Br. J. Cancer 90 (2004) 1772–1779. [9] H.L. Eng, T.M. Lin, S.Y. Chen, S.M. Wu, W.J. Chen, Failure to detect human papillomavirus DNA in malignant epithelial neoplasms of conjunctiva by polymerase chain reaction, Am. J. Clin. Pathol. 117 (2002) 429–436. [10] M.A. Palazzi, C.M. Erwenne, L.L. Villa, Detection of human papillomavirus in epithelial lesions of the conjunctiva, Sao Paulo Med. J. 118 (2000) 125–130. [11] W. Tulvatana, P. Bhattarakosol, L. Sansopha, W. Sipiyarak, E. Kowitdamrong, T. Paisuntornsug, et al., Risk factors for conjunctival squamous cell neoplasia: a matched case-control study, Br. J. Ophthalmol. 87 (2003) 396–398. [12] A. Antonsson, O. Forslund, H. Ekberg, G. Sterner, B.G. Hansson, The ubiquity and impressive genomic diversity of human skin papillomaviruses suggest a commensalic nature of these viruses, J. Virol. 74 (2000) 11636–11641. [13] C. Ateenyi-Agaba, M. Dai, F. Le Calvez, E. KatongoleMbidde, A. Smet, M. Tommasino, et al., TP53 mutations in squamous-cell carcinomas of the conjunctiva: evidence for UV-induced mutagenesis, Mutagenesis 19 (2004) 399–401. [14] A. A. T. P. Brink, B. Lloveras, I. Nindi, D. A. M. Heideman, D. Kramer, R. Pol, M. J. Fuente, C. J. L. M. Meijer, P. J. F. Snijders. Development of a general primer PCR and reverse line blotting system for the detection of beta- and gamma cutaneous human papillomaviruses (BGC-PCR). J. Clin Microbiol. In press. [15] S. Caldeira, R. Filotico, R. Accardi, I. Zehbe, S. Franceschi, M. Tommasino, p53 mutations are common in human papillomavirus type 38-positive non-melanoma skin cancers, Cancer Lett. 209 (2004) 119–124.