Alpha-papillomaviruses share a tropism for the mucosal epithelium of the human genital tract, and can be classified into 15 species or clades, based upon phylogenetic similarity.1, 2 However, the extent to which this phylogenetic grouping denotes shared biological function is unclear. There is some evidence that alpha species differ in their tissue-specific tropism, such as the alpha-3 and -15 species which appear to have an affinity for the vaginal epithelium.3 Furthermore, the high-risk human papillomavirus (HPV) types that cause cervical cancer fall largely, although not completely, within 2 species, namely alpha-9, which includes HPV16 and the related high-risk types 31, 33, 35, 52 and 58, and alpha-7, which includes HPV18 and the related high-risk types, 39, 45, 59 and 68.2
HPV16 and 18 have been shown to differ in their relative potential to neoplastically transform squamous and glandular tissue of the cervix. Progression to squamous intraepithelial lesions is greater for HPV16 than HPV18,4 but HPV18 is significantly over-represented in glandular lesions5, 6 and in adenocarcinoma of the cervix7–10 in comparison to the predominant squamous cell carcinoma (SCC) type. Our aim was to assess whether such a tissue-specific pattern is also common to HPV types phylogenetically related to 16 and 18.
To address this issue, we reanalyzed data from a recently published meta-analysis of HPV type distribution in cervical cancer,7 after writing to authors for additional histology-specific HPV prevalence data when necessary. The dataset was also updated with 5 recently published studies (Table I).6, 10–13 The detailed methods used for the inclusion of studies have been reported previously.7, 14 Data on HPV-specific prevalence were extracted independently for SCC and for adeno and adenosquamous carcinoma (collectively termed ADC), when this data was available (83% of identified cases).
|HPV type/species||SCC||ADC||ADC:SCC prevalence ratio (95% CI)|
|Any HPV||11,693||90.1||2,373||80.3||0.90 (0.89–0.92)|
|Any multiple||7,253||9.1||1,408||8.5||0.98 (0.75–1.28)|
Crude type-specific prevalence is presented for the 19 most common HPV types in cervical cancer as identified in the previous meta-analysis,7, 14 namely HPV6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, 70, 73 and 82. All studies provided information on HPV16, but for other types, prevalence was estimated only among those studies testing for the HPV type in question, and thus denominators of prevalence estimates vary by type. Type-specific prevalence thus includes that in either single- or multiple-type HPV infections. ADC:SCC prevalence ratios and corresponding 95% confidence intervals (95% CI) were calculated after adjustment for continent (Africa, Asia, South/Central America, Europe, North America/Australia), PCR primers used (MY09/11, GP5+/6+, SPF10, other) and tissue source (fresh biopsies, fixed biopsies, exfoliated cells, mixture). HPV types were grouped into alpha-7, alpha-9 or other alpha species.
A total of 11,878 SCC and 2,521 ADC cases with HPV-type specific prevalence data were included in analyses (Table I). Overall HPV prevalence was lower in ADC (80.3%) compared to SCC (90.1%), and this difference was significant even after adjusting for continent, HPV testing methods and tissue source (ADC:SCC prevalence ratio 0.90; 95% CI: 0.89–0.92). The prevalence of multiple infections was similar in both ADC (8.5%) and SCC (9.1%).
Among the alpha-7 species, 4 of the 6 studied types were over-represented in ADC compared to SCC. This association was strongest and most significant for HPV18 (ADC:SCC prevalence ratio 3.16; 95% CI: 2.91–3.43), which was by far the most common alpha-7 type (37.9% in ADC vs. 12.2% in SCC), but similar patterns were also observed for HPV45, 59 and 68. HPV45 showed a significant ADC:SCC prevalence ratio of 1.65 (95% CI: 1.29–2.12), and both HPV 59 (1.64; 95% CI: 0.92–2.92) and HPV 68 (1.60; 95% CI: 0.62–4.09) were also over-represented in ADC, although, given their rarity in cervical cancer, and the smaller number of cases that had been tested, differences did not reach statistical significance. The phylogenetically related HPV39 and HPV70 showed no association with ADC. Collectively, alpha-7 types were found in 46.3% of ADC versus only 19.4% of SCC.
In contrast to the alpha-7 species, all other analyzed HPV types were under-represented in ADC compared to SCC, most of them significantly so. Among the alpha-9 species, HPV16, 31, 33, 52 and 58 were each significantly under-represented in ADC, with prevalence ratios ranging between 0.39 for HPV58 and 0.57 for HPV16. HPV35 also showed a similar pattern, although the difference did not reach statistical significance. Collectively, alpha-9 types were found in 43.3% of ADC versus 76.3% of SCC. With respect to types of other alpha species, HPV6 (ADC:SCC prevalence ratio 0.09; 95% CI: 0.01–0.74), HPV51 (0.26; 95% CI: 0.06–1.04) and HPV56 (0.23; 95% CI: 0.07–0.76) were strongly under-represented in ADC compared to SCC.
In summary, the present meta-analysis was large enough to show that the association with ADC is shared by, and unique to, members of the alpha-7 species. Given the fact that ADC accounts for only a minority of cervical cancer cases (5–25% depending on geographic region15), it can be estimated that the alpha-7 types still cause more SCC than they do ADC in terms of absolute numbers of cases. Furthermore, the oncogenic potential to cause ADC clearly varies broadly between the individual alpha-7 types. Nevertheless, findings would suggest that certain members of the alpha-7 species share some phylogenetic trait that denotes a greater tendency to cause ADC in comparison to alpha-9 and other types. This phylogenetic difference may reflect either a greater tropism for infection of cervical glandular tissue per se, and/or a better ability to neoplastically transform glandular cells once an infection is established.
Although the proportion of cervical cancer theoretically preventable by current HPV16/18 vaccines appears at least as high in ADC (70%) as in SCC (68%), one intriguing finding of this, and previous meta-analyses,14 is a significantly lower overall HPV prevalence in ADC than SCC, which persisted even in the adjusted model. This may be partly due to a proportion of misclassified ADC arising from the endometrium, in which case the HPV16/18 prevalence in cervical ADC may even be a little higher.
Gary CLIFFORD and Silvia FRANCESCHI