Article Text
Abstract
Objectives We determined the human papillomavirus (HPV) types present in invasive cervical cancer (ICC) of women in Cameroon in order to estimate the potential efficacies of HPV prophylactic vaccines.
Methods This is a retrospective study using 181 formalin-fixed paraffin-embedded cervical tissue samples of ICC collected from the Institute of Pathology, Gyneco-Obstetric and Pediatric Hospital, Yaoundé, Cameroon. HPV was detected by PCR using modified GP5+/GP6+ (MGP) primers. Genotyping was performed by reverse-blot hybridisation, which allowed the detection of 9 of the 14 high-risk HPV types.
Results Of the 181 samples, 91.7% were squamous cell carcinomas and 6.6% were adenocarcinomas. Counting all the single and multiple infections, the three most common high-risk types in descending order were HPV16 (88%), HPV45 (32%) and HPV18 (14.8%). 54.9% of cases were infected with a single HPV type and 45.1% had two or more HPV infections.
Conclusions The frequencies of HPV16, HPV45 and multiple infections are all higher than previously reported. These observations have significant implications on the consideration of vaccination strategies because each vaccine has different duration and efficacies in cross-protection of different HPV types. The method used proved to be sensitive and cost-efficient for retrospective studies where fresh materials are not available.
- HPV
- INFECTION
- CERVICAL NEOPLASIA
- DNA AMPLIFICATION
- AFRICA
Statistics from Altmetric.com
Introduction
Invasive cervical cancer (ICC) is the second most common cancer among women worldwide accounting for 15% of cancers in women, after breast cancer. It is the leading cause of cancer mortality among women in developing countries with >80% of the cases arising in those regions. It is also replacing infectious diseases as a major cause of death. Every year, about 500 000 new cases of cervical cancers are diagnosed and 250 000 women die from the disease (585 278 incident cases and 327 899 cases of attributable deaths were predicted for the year 2010). Approximately one million women are living today with cervical cancer, and most of them do not have access to screening and treatment facilities.1–6
Infection with one of the few oncogenic human papillomavirus (HPV) types is a necessary cause of ICC. More than 118 different HPV types have been isolated and sequenced, and about 40 of them are known to infect the genital tract and 14 have oncogenic potentials.2
Currently, the best way to prevent cervical cancer is through organised gynaecological screening programmes with associated treatment of the detected pre-cancerous lesions. In developed countries, introduction of those programmes has resulted in a significant reduction in death rates, burden of disease and cost of treatment. However, they have not been effectively implemented in some countries where they are most needed due to barriers of cost, trained personnel and infrastructure.1 ,6
While the type distribution in cervical cancer does vary somewhat worldwide, HPV16 and HPV18 are the most prevalent types, being found in >70% of samples from cervical cancer around the world and being the two types targeted by current vaccine formulations.1 ,6 Data on HPV-type distribution in women with ICC and its precursor lesions are essential to predict the potential impact of new prophylactic vaccine against HPV16 and HPV18, as well as to determine priorities for inclusion of HPV types in future HPV vaccines and HPV-based screening tests. It is known that a majority of worldwide ICC cases are associated with HPV16 and/or HPV18, but no reliable information exists about the type-specific contribution of HPV in ICC in many, mostly developing, countries, and data are limited or missing from many regions in Africa, such as the region of West–Central Africa where Cameroon is situated.
In Africa, about 21% of women in the general population are estimated to harbour cervical HPV infections, and 72.9% of ICC is attributed to HPV16 and/or HPV18.4 ,7 Cameroon has a population of 5.24 million women aged 15 years and older who are at risk of developing cervical cancer. Current estimates indicate that every year 1474 women are diagnosed with cervical cancer and 995 die from the disease.1 ,6 There is no national cervical cancer screening programme, but sporadic screening limited to some main cities are conducted, in spite of 42% of the population of Cameroon being rural. This condition probably contributes to the high incidence and mortality rates of cervical cancer in the country. Based on hospital registry data, women most at risk for cervical cancer are those >35 years with a median age at diagnosis of 49 years; most of them having an advanced and incurable disease at presentation.7
The present study was conducted in the framework of the Swiss–Cameroon cooperation. This study aims to provide information about HPV genotype distribution in samples of ICC from Cameroon, where the incidence of ICC is one of the highest in the world. It also offers a comparison with data of other countries worldwide in order to predict and evaluate how HPV vaccination and HPV-based screening might influence cervical cancer prevention.
Methods
Sample collection
Cases were recruited between January 2006 and December 2010 from Yaoundé and Douala areas. All cases were referred for diagnostic purposes to the Pathology Department (ZS) of the Gyneco-Obstetric and Pediatric Hospital, Yaoundé, Cameroon. The biopsies were retrieved from the archives of the anatomopathology services of the two capitals (Douala and Yaoundé), which hold all specimens from Cameroon for research and complementary diagnostics purposes. Some samples were from preoperative biopsies and others from tissue fragments of surgical specimens.
In total, 181 formalin-fixed paraffin-embedded samples were selected, following initial diagnosis of ICC. An independent histological examination of all the samples was carried out by a pathologist of the Clinical Pathology Department at the University Hospitals of Geneva, Switzerland. DNA extraction and HPV genotyping were carried out in the molecular biology laboratory of the same department.
DNA extraction and reverse-blot hybridisation
Five to ten sections of 10 μm from each sample were used to extract DNA. The sections were incubated in xylene to remove paraffin. Xylene was subsequently removed by rinsing the tissue pellet in 100% ethanol, followed by rehydration in 70% ethanol. The pellet was air-dried and lysed in 600 μL lysis solution (50 mM Tris, pH 8.3, 1 mM NaEDTA, 0.5% Tween-20) and digested overnight at 56°C in 350 μg/mL proteinase K. Phenol chloroform extraction was carried out twice on the lysed material and DNA was precipitated, resuspended in water and quantified on NanoDrop ND-1000 (NanoDrop Technologies, USA). DNA was successfully obtained from 181 samples. A total of 100 ng DNA was used by PCR using modified general primers GP5+/GP6+ (MGP) and genotyping was performed by reverse-blot hybridisation using the probes from Dr R. Sahli (personal communication), which allows the detection of 14 high-risk HPV types (16, 18, 31, 33, 35, 39, 45, 51, 56, 58, 59, 66, 73 and 82) and 12 low-risk types (6, 11, 34, 40, 42, 43, 44, 53, 54, 57, 70 and 84).8–11 Note that the oligonucleotides used for 26, 52, and 68 lie outside the PCR fragment using MGP primer pairs, and therefore would not be detected in this study.
In order to determine whether HPV negativity was due to poor quality materials from which amplifiable DNA may not be extracted, beta-globin of 200 bp was amplified by PCR to test the integrity of DNA extracted. For histopathology for cervical intraepithelial neoplasia grade 3 (CIN3), the sensitivity and specificity of GP5+/GP6+ primers are 100% and 23.5%, respectively.12 MGP has the advantage of being more sensitive in detecting multiple infections.9
Results
Patient characteristics
The average age of the patients at the time of diagnostics was 51.8 years, bearing in mind that the ages of 59 patients (32.6%) were not known. The patients for whom we do not have the age were those who did not know their age and did not have identification cards. Further, initial clinical information from remote health services in rural areas is incomplete. Such were common situations in Cameroon.
Between the initial diagnostics and the independent histological examination of sections before DNA extraction, there was a concordance of 98.3% (178 of 181 cases). HPV DNA positivity was detected in 175 (96.7%) cases. The result was that all six HPV-negative cases were negative for beta-globin (results not shown), confirming that HPV negativity was due to the poor qualities of these DNAs. The majority of samples (n=166; 91.7%) were squamous cell carcinomas (SCC) and 12 (6.6%) were adenocarcinomas (ADC). In three cases, there was a discrepancy between the initial diagnosis and the review diagnosis of tissue sections. In the 175 cases that were positive for HPV DNA, 9 of the 14 high-risk types were detected, of which 96 cases (54.9%) were infected with a single HPV type and 79 cases (45.1%) had two or more HPV infections (table 1).
Prevalence of HPV types in ICC samples
The most common HPV types, taking into account single and multiple infections, in descending order, were HPV16, HPV45, HPV18, HPV82, HPV35, HPV56, HPV39, HPV51 and HPV33. HPV16 was identified in 154 HPV-positive cases (88%), which consisted of almost equal proportions of single and co-infections with other types (44.6% and 43.4%, respectively; table 2). HPV45 was the second most prevalent high-risk HPV type in Cameroonian women with ICC, and the majority of HPV45 was found in co-infections with HPV16 or HPV18 (figure 1). HPV82 was detected in 6.9% of the cases but only in co-infections with HPV16, HPV45, or both. Fifteen cases (8.6%) of ICC were found to have one of the other established high-risk HPV types as single or multiple infections. In both SCC and ADC, the prevalence of HPV16 was similar (86% and 83%). There were also seven cases with low-risk HPV types: HPV6, HPV11 and HPV83, all of which are present in co-infections with high-risk types (table 2). A graphical representation of the data including only HPV16 and HPV18 as single and multiple infections is shown in figure 2.
Discussion
This is one of the few published studies on the HPV genotype distribution among women with ICC in the West–Central African region and the first one in Cameroon, a country where no data on this specific subject are currently available.1 ,6 Among the 175 HPV-positive cases analysed, ADC constituted 6.6% of all ICCs. This is in agreement with the study of Guan and colleagues, which showed that ADCs represent <10% of ICC worldwide and 3% in Africa.4 The fact that HPV18 was detected in 14.8% of the cases is also in accordance with the results from previous studies supporting 19.8%±4.1% prevalence rate in Africa, and 18% in Zimbabwe.4 ,13 In Africa, some studies showed that the prevalence of HPV18 in ICC could be up to 25%, which is higher than in Europe and the world average (5% and 10%, respectively).2
Three outcomes are quite different from previously reported studies: first, the high prevalence of HPV16 (88%), which is significantly higher than the 48% shown in the meta-analyses, or the 53.1%±4.4% HPV16 prevalence in sub-Saharan Africa reported by Guan and colleagues2 ,4; second, the high prevalence rate of HPV45 (32%) instead of around 10% (11±2.2%) previously reported in Africa4; and third, the high multiple-infection prevalence in ICC, 45.1% compared with 19% reported in Africa.2
In a meta-analysis of HPV-type distribution in different continents, Li and coworkers have compared different PCR protocols in HPV genotyping and demonstrated that the primers that were used (GP primers) in this study gave higher prevalence of HPV16 than other primers such as PGMY or MYO primers.5 This could explain the relative high HPV16 prevalence rate identified in this study. It has been shown that HPV45 ranks at the third place among the most frequent oncogenic types in ICCs, following HPV16 and HPV18, with a prevalence of 6% worldwide and 4% in Europe.2 ,4 ,5 HPV45 is largely represented in Africa, reaching 11% in the study of Guan et al,4 supporting that more severe histopathological diagnostic is associated with HPV45, like HPV16 and HPV18. In Guinea, it represented 18.6%, while HPV18 represented only 14%.14 In a meta-analysis of HPV distribution in West Africa, HPV45 was found in 13.3% of ICC compared with 10.5% of HPV18.3 ,14 In light of these observations, the high prevalence of HPV45 found in this study in Cameroon is less surprising. What is worth noting is that the majority of HPV45 was present in co-infections with HPV16. There are substantial data suggesting that in HIV-positive women there is higher prevalence of HPV and more frequent multiple type co-infections, and one small study in Cameroon comparing HPV infections and HIV status showed HPV45 to be the dominant co-infection type.15 ,16 More studies are required to address the association between HIV infection and the prevalence of HPV types in women with normal cytology or with ICC. Africa has one of the highest prevalence of multiple HPV infections worldwide, reaching 11.9–19%.2 ,4 The increasing prevalence of multiple infections worldwide has been observed since the 1990s.5 This phenomenon could be explained, on the one hand, by the increasing sensitivities of the genotyping tests, and on the other hand, by the increasing number of HPV types included in the genotyping protocols.5 Cameroon has an HIV prevalence of 5.1% in the general population and 2.8% among women.6 In women who are HIV-positive, the prevalence of multiple HPV co-infections has been reported to be much higher. One study on HPV genotypes in HIV-positive women in South Africa found that 72.5% had multiple HPV infections.17 However, the data sets containing multiple infections remain limited; as most PCR-based commercial tests used in the studies published so far do not give information on HPV types other than HPV16 and HPV18. It is well known that different HPV types, taken separately, differ a lot in their relative carcinogenic potential and some of them could interact or act in synergy to induce the development or the progress of a lesion.4 ,18 The attribution of an ICC to a type HPV is therefore more and more complicated through the growing prevalence of co-infections. It is challenging but necessary to evaluate the expected impact of prophylactic vaccination as well as to carry out screening tests, which should contribute to improve risk stratification among HPV-positive women.
We also found a significant presence of HPV82 among women with ICC (6.9%), but only in co-infection status, mainly with HPV16. HPV82 has not been described in previous studies in Africa as an emerging HPV type. In view of these major differences, it is worth repeating the study in Cameroon with a larger sample set. These observations, if confirmed, would have considerable implications on the screening and vaccination programmes in Cameroon. The sensitive and comprehensive HPV genotyping technique will generate data that should help designing future HPV vaccines.
Currently, two vaccines are available on the market: the bivalent Cervarix, which protects against HPV16 and HPV18; and the quadrivalent Gardasil, which provides an additional protection against two low-risk types, HPV6 and HPV11. The immunogenicity induced by these vaccines is excellent, providing a protection against 70–75% of ICCs, which corresponds to the percentage of ICCs assigned to HPV16 and HPV18.19 The data reported here support that 55% of the ICC was caused by HPV16, HPV18 or HPV16+18 (figure 2), which is preventable by the vaccines. The rest (36%) are ICC cases co-infected with other HPV types as well as HPV16, HPV18 or HPV16+18 (figure 2). A large proportion of co-infecting type was HPV45. The funding for Gardasil, the quadrivalent vaccine, has been approved and pilot programmes are being conducted in Cameroon by GAVI alliance. The advantage of this vaccine is that it protects against HPV6 and HPV11, which reduces the burden of common warts. The disadvantage is that it offers only 8% (95% CI −67% to 49%) cross-protection against HPV45 compared with 79% (95% CI 61% to 89%) offered by the bivalent vaccine.20–22 However, the clinical data so far suggest that the efficiency against persistent infections with the types HPV45 would tend to decrease with time and that the effect of cross-protection could disappear within years.20 It would be interesting, in light of our data, to wait for the clinical data from the next-generation nonavalent vaccine that provides protection against nine HPV types (6, 11, 16, 18, 31, 33, 45, 52 and 58).23 Studies on the frequency of integration of viral oncogenes E6 and E7 into host genome show that HPV45 has the highest frequency among patients with CIN3, followed by HPV16 and HPV18, while HPV31 and HPV33 show much lower integration frequencies.24 Although integration is not a prerequisite for transformation into tumour (only 62% ICC show integration of HPV), the high frequency of integration of HPV45 suggests a high oncogenic potential.24 Further studies are necessary to determine the virulence of other HPV types in women protected by vaccination.
The median age of women at the time of diagnosis was 52 years, which lies within the age group with the highest incidence rate of ICC.25 However, due to the missing information on the age of a significant proportion of patients, the information on the correlation between age at diagnosis and the HPV type could not be obtained. Furthermore, the results are probably not totally representative of the whole female population of Cameroon suffering from ICC as all the biopsies were collected in the hospital centres of the two largest cities of the country. There might be clusters of HPV types in women in numerous isolated villages of Cameroon.
Nevertheless, the strength of this study is the opportunity to collaborate within the framework of the cooperation between Switzerland and Cameroon, a solid collaboration already existing over almost 20 years and bringing together, in this precise case, gynaecologists and pathologists of both countries concerned. All the histopathological diagnostics have been revised and confirmed in the Service of Clinical Pathology of the ‘Geneva University Hospitals’. The study protocol allows the typisation of the 14 high-risk and 12 low-risk HPV types, which is one of the most comprehensive lists currently used by laboratories worldwide. Furthermore, the protocol has the advantage over commercially available HPV tests because it is suitable for large retrospective studies where only fixed tissue samples are available.
Key messages
Predominant high-risk human papillomavirus (HPV) types are HPV16 and HPV45, and multiple infections are more frequent than previously reported in the sub-Saharan African region.
Significant presence of HPV82 among women with invasive cervical cancer (ICC) (6.9%), but only in co-infection status, mainly with HPV16.
Available vaccines could prevent at least 55% of ICC.
References
Footnotes
Handling editor Jackie A Cassell
Acknowledgements The authors are grateful for the help received from the staff of the immunohistochemistry laboratory and the molecular biology laboratory of the Department of Clinical Pathology of Geneva University Hospitals. The study was made possible with the facilities of the service of clinical pathology under the direction of Professor Laura Rubbia-Brandt and the support of the head of the laboratory (TAM).
Contributors PP, PV and DP had the idea and initiated the project. ZS collected the samples. ZS, PV, J-CP and JG performed the histopathological analyses. LH contributed to the establishment of the HPV genotyping protocol and DP did most of the analyses in the laboratory. LR-B, and TM provided the facilities of the laboratory. DP and LH drafted the article, and the manuscript approved by all the authors.
Funding ‘Solidarité Internationale Genève’, Geneva University Hospitals.
Competing interests None.
Patient consent Due to the retrospective aspect of this study, and the fact that it was very difficult or impossible to trace patients, who often came from remote rural areas, consent from patients was not obtained.
Ethics approval National Ethics Committee of Cameroon/Geneva Ethics commission (CER 11-203R; 11-048R).
Provenance and peer review Not commissioned; externally peer reviewed.