Vaccine-related HPV genotypes in women with and without cervical cancer in Mozambique: Burden and potential for prevention

Abstract

Knowledge about the burden of Human Papillomavirus (HPV) infections in Sub-Saharan Africa is very limited. We collected cervical samples from 262 women from the general population and 241 tumor samples from women with invasive cervical cancer in Mozambique and tested them for HPV genotyping by the SPF₁₀-LiPA₂₅ PCR system. Among the 195 women without cervical abnormalities by cytology HPV prevalence was 75.9%. In this group of women, the most frequently identified HPV types among HPV-positive women were in descending order of frequency: HPV51 (23.6%), HPV35 (19.6%), HPV18 (14.2%), HPV31 (13.5%) and HPV52 (12.8%). In women with cervical cancer HPV DNA detection was 100%. The type-specific distribution of the most frequent types in descending order of frequency was: HPV16 (47.0%), HPV18 (31.3%), HPV51 (14.8%), HPV52 (14.3%), HPV45 (12.6%), HPV35 (10.4%), HPV33 (4.8%) and HPV31 (2.6%). HPVs 16/18 and HPVs 16/18/31/45 were detected in 71.7% and 80.9% of cervical cancer tissue, respectively. While HPVs 51 and 35 were the two most common types in cytologically normal women in Mozambique, HPVs 16 and 18 remained the two most frequently identified types in cervical cancer. The introduction of an efficacious HPV 16/18 vaccine could potentially prevent the occurrence of 72% of cervical cancer cases and up to 81% of the cases if full cross-protection against HPVs 31 and 45 is assumed. © 2007 Wiley-Liss, Inc.

Introduction

Type-specific vaccines against Human Papillomavirus (HPV) may be an efficacious strategy to combat invasive cervical cancer in African high-risk countries such as Mozambique where there is a high incidence of invasive cervical cancer and where there are no effective screening programs in place.1 However, there is limited information as to the distribution of HPV genotypes in women with and without cervical cancer in Sub-Saharian Africa. Given the recent preliminary evidence showing that HPV 16/18 vaccines could also partially cross-protect against incident infections by HPVs 31 and 45,2 burden estimates of potentially preventable disease related to HPVs 16/18/31/45 need to be updated. This is important in order to be able to project the full potential efficacy of implementing HPV vaccines in this population.

To estimate the burden of type-specific HPV infections, we analyzed cervical samples from women with and without cervical cancer using the SPF₁₀-LiPA₂₅ system version 1.3 A secondary objective of the study was to confirm, using these newer assays, the finding reported previously by our group that HPV 35 is the most common HPV type in both cytologically normal women and women with squamous intraepithelial lesions (SIL).4

Material and methods

Between August and October, 1999, 262 women aged 14–61 years were randomly selected from Manhiça, a rural area in Southern Mozambique, by means of a pre-established age-stratified scheme. The study design is described in more detailed elsewhere.4 In brief, study participants underwent a complete gynaecological examination with collection of cervical samples for screening of cytological abnormalities and HPV genotyping. In addition, 241 paraffin embedded tissue blocks of invasive cervical cancer samples were obtained from the Maputo Central Hospital, Mozambique. Ethics approval was provided by the Ministerio da Saude of Mozambique and the Ethics Review Committees of the Hospital Universitari de Bellvitge (Hospitalet de Llobregat, Spain) and the Hospital Clinic of Barcelona (Barcelona, Spain).

From the cervical swabs DNA extraction and purification was performed using a commercial kit from Qiagen (QIAamp® DNA Mini Kit) and eluted to a final volume of 200 μl. The purified DNA extracts were stored at −20°C.

In the invasive cervical cancer group, consisting of 241 women, the so-called sandwich technique was applied to obtain the paraffin sections. The first and the last sections were stained with Hematoxylin-Eosin (H&E) on microscope slides which were reviewed by pathologist for diagnosis and determination of specimen adequacy. From the paraffin inner 5-μm sections, DNA was extracted. DNA extraction was performed by Proteinase K digestions for 16 hr at 56°C temperature.

In the general population samples, the human β–globin target gene determination using PCO3/PCO5 primers that amplify a 209 base-pair fragment was used as a guide for specimen adequacy. β-globin amplification was visualized by 1.5% agarose gel electrophoresis.

SPF₁₀-LiPA₂₅ was performed using 10μl of the DNA extract in a final reaction volume of 50 μl.3 β-globin PCR was performed also using 10 μl of the DNA extract in a final reaction volume of 50 μl. The amplified PCR products were tested using a probe hybridization with a cocktail of conservative probes recognizing at least 54 mucosal HPV genotypes in a microtiter plate format for the detection of HPV DNA or β-globin.5 Optical densities (OD₄₅₀) were read on a microtiter plate reader. HPV DNA positive samples were subsequently analyzed by HPV SPF₁₀-LiPA₂₅ (version 1: produced at Labo Biomedical Products, Rijswijk, The Netherlands), a reverse hybridization technique that detects 25 high-risk and low-risk HPV types (6, 11, 16, 18, 31, 33, 34, 35, 39, 40, 42, 43, 44, 45, 51, 52, 53, 54, 56, 58, 59, 66, 68, 70, 74).3 The sequence variation within the SPF₁₀-LiPA₂₅ primers allows the recognition of these different HPV genotypes, except for the Types 68 and 73 as their interprimer regions are identical and cannot be distinguished on this test. After PCR, 10 μl of the amplimers was used to perform reverse hybridization for HPV genotype identification. The positive hybridization on the strips is visualized as a purple band by means of a precipitating color substrate on the probe site. SPF₁₀-LiPA₂₅ detection and typing was performed at the facilities of DDL Diagnostic Laboratories (DDL, Voorburg, Netherlands). In the invasive cervical cancer group samples that turned out negative were further analyzed by repeating the PCR using a 1:10 dilution of the DNA extract in order to reduce the possible interference by inhibiting substances.

Samples were classified as valid if the results were HPV positive or both HPV negative and β-globin positive. HPV type-specific distributions were calculated among HPV positive women. The percentages of these distributions were calculated by including in the numerator the number of women testing positive for a particular genotype regardless of coinfection with other types, and in the denominator, the total number of HPV positive women. Thus, the sum of the percentages may exceed 100 since women with multiple infections were counted several times in the numerator but only once in the denominator of the percentage distribution.

Results

Table I shows in detail the type-specific HPV distribution among HPV-positive women with and without cervical cancer. Of the 262 women from the general population included in the study cytological reading was adequate in 244 women: 195 (77.4%) were cytologically normal, 19 (7.5%) had ASCUS, 13 (5.2%) had LSIL, 16 (6.3%) had HSIL, and 1 (0.4%) had squamous-cell carcinoma. In total, 195 women with normal cytology and 49 women with ASCUS or higher lesions (ASCUS+) yielded valid PCR results. Of the 241 cervical cancer specimens initially collected, 230 contributed valid results by SPF₁₀-LiPA₂₅.

Table I. Type-Specific HPV Distribution among HPV-Positive Women with and Without Cervical Cancer

Women from general population (cytology results)				Women with invasive cervical cancer
Normal (n = 195)		≥ASCUS (n = 49)		Cervical cancer (n = 230)
HPV + women	148 (75.9)1		44 (89.8)1		230 (100.0)1
HR, high risk; LR, low risk. Type-specific distributions are based among HPV positive women. Sum of distribution percentages exceeds 100 because, in the numerator, women with multiple infections are counted for each of the HPV types found. Classification of HR and LR HPV types is based according to the publication by Muñoz et al.6 Five HPV positive samples of cervical cancer tested negative by LIPA and were excluded from the type-specific statistical analysis. This explains why the sum of women with multiple and single infections does not add up to 230. 1 The values indicate N (%).
HR types		HR types		HR types
51	35 (23.6)	35	13 (29.5)	16	108 (47.0)
35	29 (19.6)	52	12 (27.3)	18	72 (31.3)
18	21 (14.2)	51	8 (18.2)	51	34 (14.8)
31	20 (13.5)	58	8 (18.2)	52	33 (14.3)
52	19 (12.8)	18	5 (11.4)	45	29 (12.6)
39	14 (9.5)	33	5 (11.4)	35	24 (10.4)
68/73	14 (9.5)	16	4 (9.1)	33	11 (4.8)
16	12 (8.1)	39/68/73	3 (6.8)	31	6 (2.6)
66	12 (8.1)	56	3 (6.8)	68/73	6 (2.6)
33	11 (7.4)	31	2 (4.5)	53	4 (1.7)
53	9 (6.1)	45	2 (4.5)	58	4 (1.7)
56	7 (4.7)	53	2 (4.5)	39/68/73	2 (0.9)
45	6 (4.1)	39	1 (2.3)	39	1 (0.4)
39/68/73	4 (2.7)	68/73	1 (2.3)	59	1 (0.4)
58	3 (2.0)			66	1 (0.4)
LR types		LR types		LR types
70	11 (7.4)	44	3 (6.8)	11	4 (1.7)
44	10 (6.8)	74	3 (6.8)	6	3 (1.3)
6	10 (6.8)	11	2 (4.5)	70	1 (0.4)
54	8 (5.4)	40	2 (4.5)	74	1 (0.4)
11	7 (4.7)	70	2 (4.5)
40	6 (4.1)	43	1 (2.3)
74	6 (4.1)	6	1 (2.3)
43	5 (3.4)
34	1 (0.7)
42	1 (0.7)
Single infections	74 (50.0)		18 (40.9)		144 (64.0)
Multiple infections	74 (50.0)		26 (59.1)		81 (36.0)
2 types	45 (30.4)		16 (36.4)		53 (23.6)
3 types	13 (8.8)		7 (15.9)		21 (9.3)
4 types	9 (6.1)		3 (6.8)		4 (1.8)
≥5 types	7 (4.7)		–		3 (1.3)
Total no. of infections	281		83		350

HPV prevalence estimates were 75.9% in women with cervical normalcy, 89.8% in women with ASCUS+ lesions and 100% in women with cervical cancer. HPVs 51, 35, 18, 31, 52, 39, 68/73 and 16 were the 8 most frequently identified genotypes among HPV positive women with normal cytology. The corresponding ranking among women with ASCUS+ was HPVs 35, 52, 51, 58, 18, 33, 16 and 39/68/73. In women with cervical cancer HPVs 16 and HPV 18 were the two most frequently identified genotypes (47.0% and 31.3%, respectively), followed by HPV Types 51, 52, 45, 35, 33 and 31. The percentage of women with multiple infections was 50% among women with normal cytology, 59.1% among women with ASCUS+ lesions and 36.0% among women with cervical cancer.

Table II summarizes the full potential of prevention of current HPV vaccines showing the percentage of women with HPV infections attributed to vaccine-related HPV genotypes, including putative HPV 31/45 cross-protection, in the 3 groups of women. The percentage of cervical cancer cases potentially prevented by a vaccine that included HPV 16/18 VLPs ranges from 71.7% assuming no cross-protection to 80.9% assuming full cross-protection against HPVs 31/45. The corresponding percentages for SIL were 20.7% and 31.0%, respectively. Finally, a HPV 16/18 efficacious vaccine could potentially prevent HPV type-specific vaccine-related infections in up to 26.7% of women with normal cytology assuming cross-protection. As shown in Table II, the highest carcinoma to normal HPV type-specific prevalence ratios were observed for HPV 16 (7.6) followed by HPV 45 (4.1) and HPV 18 (2.9). In contrast, HPV 31 is more likely to be detected in cervical cells from cytologically normal women than in cervical cancer specimens.

Table II. Full Preventative Potential of Current HPV Vaccines in Mozambique: Number and Percentage of Women with HPV Infections Attributed to Vaccine-Related HPV Genotypes using the SPF10-LIPA₂₅ PCR Assay by Diagnostic Group

HPV types	Diagnostic group1			Cervical carcinoma to “normal” HPV prevalence ratios
	Normal	SIL	Cervical carcinoma
	N (%) [n = 195]	N (%) [n = 29]	N (%) [n = 230]
Percentages are computed among all women in that diagnostic group regardless of HPV status. 1 Includes women with a valid diagnostic result by both cytology and SPF10-LIPA25 PCR. 2 Includes single and multiple infections involving that specific type(s). 3 Concomitant infections involving HPV types 16, 18, 31 or 45 were: HPVs 16/18/31 (n = 1), HPVs 18/31/45 (n = 1), HPVs 16/31 (n = 1) and HPVs 16/18 (n = 2). 4 Concomitant infection involving HPV types 16, 18, 31 or 45 were: HPVs 16/18 (n = 13), HPVS 16/18/45 (n = 2), HPVS 16/31 (n = 3), HPVS 18/31 (n = 1), HPVS 31/45 (n = 1), HPVs 16/45 (n = 3) and HPVs 18/45 (n = 2).
HPV + (any type)	148 (75.9)	27 (93.1)	230 (100.0)	1.3
HPV 162	12 (6.2)	3 (10.3)	108 (47.0)	7.6
HPV 182	21 (10.8)	3 (10.3)	72 (31.3)	2.9
HPV 16/18 combined2	30 (15.4)	6 (20.7)	165 (71.7)	4.7
HPV 312	20 (10.3)	2 (6.9)	6 (2.6)	0.3
HPV 452	6 (3.1)	1 (3.4)	29 (12.6)	4.1
HPV 31/45 combined2	25 (12.8)	3 (10.3)	34 (14.8)	1.2
HPV 16/18/31/45 combined2	523 (26.7)	9 (31.0)	1864 (80.9)	3.0

Discussion

We have previously reported that HPV 35, as detected using a modified MY09/11 PCR system, was the most commonly identified HPV genotype both in cytologically normal women from the general population and in women with cervical neoplasia in rural Mozambique, a country with a high incidence of cervical cancer.4 Because of findings discrepant from the reported worldwide genotype distribution,6, 7 doubts arose to the possibility that certain HPV types might be under-detected depending on the sensitivity levels of the technologies used or, alternatively, that this was the true underlying HPV distribution for this particular population. Our study, using the SPF₁₀-LiPA₂₅ PCR system, identifies HPVs 51 and 35 as the first and second most common types among women with normal cytology and HPVs 35 and 52 as the first and second most common among women with ASCUS+ lesions. These results confirm the relative importance of HPV 35 in women without cervical cancer in Mozambique. The data from our study also confirm the high prevalence of HPV infections in this high-risk population: 75.9% among women with normal cytology.

Despite the high prevalence of HPV 35 in women from the general population of Mozambique, the impact of HPV 35 in cervical carcinogenesis is relatively marginal, and only 10.4% of women with cervical cancer are infected with this genotype in this population. Indeed, HPV16 and HPV18 are the two most frequently detected types among cervical cancer cases (Table I). These results are in agreement with other studies showing discrepant type-specific distributions between women with and without cervical cancer.8

Concerning the relevance of other genotypes our study highlights the relevance of HPV Types 51 and 52 in women from the general population and in women with cervical cancer when using the SPF₁₀-LiPA₂₅ PCR assay. Thus, HPVs 51 and 52 are after HPVs 16 and 18, the 3rd and 4th most common types in cervical cancer specimens.

We found a high percentage of multiple types in the 3 groups of women: 50%, 59.1% and 36.0% in women with normal cytology, ASCUS+ and cervical cancer, respectively. The high rate of multiple infections in cervical cancer could be caused by the presence of underlying accompanying CIN lesions or subclinical infections in normal mucosa present in the specimen. The analyzed tissue samples were not microdissected to explore this possibility. The HIV status was not available for the cervical cancer group, however, HIV testing was done unlinked and anonymously in the community group of our study, showing an overall prevalence of 12%,9 a figure that may be even higher among women with cervical cancer. Thus possible immune deficiencies resulting in simultaneous multiple infections and concomitant CIN lesions within the samples cannot be ruled out. Multiple HPV infections with an average of more than 3 genotypes is most common in HIV positive women.10

Our findings among cervical cancer cases are consistent with other smaller studies in Mozambique that also found that HPVs 16 and 18 were the most common types in tumor samples from these patients.11–14 The results of these studies are consistent with the world-wide distribution of HPV types in cervical cancer.7, 15

Despite of the fact that HPVs 51 and 35 are the most frequently detected types in women without cervical abnormalities, HPV Types 16 and 18 are still the most frequently identified genotypes in cervical cancer specimens. This finding is consistent with the recent evidence from a large prospective study that showed that HPV 16 and 18 infections are more likely to escape immunity and persist over time gaining advantage over the other high-risk types with regard to disease progression and development of neoplastic lesions.16 Our cross-sectional data are consistent with these observations, as the HPV 16 and the HPV 18 prevalence among cervical cancer cases were respectively about 7.6-fold and 3-fold higher that the corresponding prevalence among cytologically normal women (Table II), suggesting indeed that these HPV types are more likely to persist and cause cancerous lesions than the other types. Interestingly, HPV 45 also shows a 4-fold higher prevalence among cervical cancer cases relative to that among women without cervical lesions. In contrast, infection by HPV 31 is about 4 times more likely to be detected in the general population group than in the cervical cancer group, suggesting that it is more likely to be cleared or less likely to persist/progress on to cervical cancer (Table II).

Despite the predominant detection of HPVs 16 and 18 in cervical cancer, it is worth noting the relative importance of HPVs 51, 52, 45 and 35. This information is relevant in deciding the composition of future polyvalent vaccines.

As summarized in Table II the implementation of an efficacious HPV vaccine against HPVs 16/18 is likely to prevent about 72% of the cervical cancer cases, 21% of SIL cases and 15% of infections in normal women in Mozambique. If current HPV 16/18 vaccines induced full cross-protection against HPV Types 45 and 31, the preventive potential could increase up to 81% for cervical cancer, 31% of SIL cases and 27% of HPV infections in normal women.

In summary, our study confirms that HPV vaccines have a strong potential to greatly reduce invasive cervical cancer in rural Mozambique. However, future polyvalent vaccines should take into account the relative importance of HPVs 51, 52, 45 and 35 in cervical cancer in Mozambique.

Acknowledgements

We are grateful to the women participating in the study. We want to thank Dr. M. Araño and Dr. C. Romagosa for helping in the collection of the cervical samples. We are grateful to the technical support provided by the staff of the Manhiça Health Center and the Manhiça Health Research Center (CISM), as well as to the staff of the Pathology Department of the Maputo Central Hospital.