Background

Two hundred different HPV types, classified into 49 species, are currently recognized by the International HPV Reference Center [1]. From these, about 40 are found in the female genital tract [2] and can be divided, according to their ability to generate malignancies, into low-risk types (LR-HPV) (e.g. HPV 6 and 11) which are rarely detected in high grade cervical lesions (HSIL) but produces the majority of genital warts [3, 4] and high-risk types (HR-HPV), which cause lesions with potential for oncogenic progression (e.g. HPV 16 and 18) [5]. Longitudinal studies have shown that individual HPV types grouped under the HR-HPV class differ enormously in their oncogenic potential and the pace in which they may drive the evolution of high-grade premalignant lesions to cervical cancer. Knowledge on the distribution of individual HPV types in different geographical areas is fundamental for the optimization of cervical cancer preventive strategies such as vaccination and HPV-DNA primary screening, which should take into consideration the local contribution of HPV types [6].

The majority of HPV infected women will never develop cancer since most HPV infections are transient and asymptomatic. As expected, HPV is more prevalent in young sexually active women (18 to 30 years) [7], decaying with aging. However, a second peak of HPV has been described after menopause in some studies [8], but was not verified in larger cohorts worldwide, except for Africa and Central America [7]. HPV persistence and progression from infection to disease vary according to the virus genotype, even among HR-HPV types [9]. Another important factor for disease progression seems to be the age; women with an HPV genital infection and older than 40 years have much higher risk of progression to cervical cancer than younger women [10].

Invasive Cervical Cancer (ICC) is worldwide the third most frequent cancer among women [11]. In Brazil, ICC is preceded only by non-melanoma skin cancer and breast cancer. According to the Brazilian National Cancer Institute, 16,340 new cases are expected to occur in 2016 with an estimated incidence of 15.85 cases per 100,000 women. In São Paulo State, where the present study was conducted, the estimated incidence for 2016 was of 9.50 per 100,000 women, in contrast to, for instance, another Brazilian city, Manaus located in the North region, where an alarming projection of 53.73 cases per 100,000 women, not adjusted by age, was made [12].

The distribution and prevalence of the HPV genotypes vary with the grade of cervical disease, age and the geographical location of the patients [8, 13–15]. Prevalence of HPV 16 and 18 increases in parallel to the severity of the lesion [16–18] being associated to approximately 70 % of the ICC cases worldwide but are also the most frequent types observed in women with normal cytology [7]. In a meta-analysis of 1 million women with normal cytological findings, HPV18 was found in many regions as the second most frequent [19], also verified in some Brazilian studies [20–22]. Notwithstanding HPV 31 occupies the second position in frequency in Europe and HPV 52 in Africa, pointing to the importance of other HR-types non 16/18 in certain regions [19]. Illustrating the geographical variability, in Brazil, HPVs 31 and 33 are the second most prevalent types [23, 24] among selected populations, respectively in the Northeast and Central regions. Surprisingly, HPV 66 was detected in 22 % of the HPV positive samples from Campo Grande, Mato Grosso do Sul, [25] and HPV 58 was the most frequent type (19.8 %) in HIV-infected women followed by HPV 53 (15.5 %) in the Southeast region [26].

As a result of accumulated knowledge over the past three decades on the causal relationship between HPV infection and ICC, detection of the viral DNA became an attractive approach to identify women at risk for developing ICC [27, 28]. Evidence from randomized trials now supports the incorporation of screening methods that clearly focus on the detection of the HPV genome [29] since there is overwhelming evidence that this is more sensitive than cytology for identifying CIN3, the true premalignant precursor of ICC [30–32]. Investigating HPV genotypic distribution and disease-association became an important tool in epidemiological studies, further supporting vaccine and screening tests development and improvement, by ensuring their coverage and effectiveness [33, 34].

Since June 2013, the Brazilian Ministry of Health has announced the decision to introduce the quadrivalent HPV vaccine in the national immunization program, the vaccination schedule (0 and 6 months) was updated in January 2016. From 2014, this vaccine was offered to girls and adolescents 11–13 years old and from 2015 for adolescents between 9 and 11 years [35]. In this context, HPV genotyping may be an essential tool to assess the program effectiveness by identification of the genotypic profile pre and post HPV vaccination.

PapilloCheck assay is a robust method for HPV genotyping, which is based on PCR amplification of a fragment of approximately 350 bp from the E1 region of HPV genomes, using broad-spectrum consensus primers, allowing the simultaneous detection and genotyping of 24 different HPV types by DNA chip technology including low-risk HPVs [36].

We aimed to uncover the HPV type-distribution hiding under an HPV-HR result in the HC2 test or the 12 HR-HPVs cocktail in the Cobas HPV test to analyze the HPVs genotype distribution in women undergoing routine cervical cancer screening and women with a previous abnormal Pap test in São Paulo and Barretos cities and their correlation to cytological results and age groups.

Methods

Study population

Specimens were collected in between October, 2009 and April, 2011, in two reference centers of São Paulo State, Brazil: Hospital das Clínicas from the São Paulo Medical School, Universidade de São Paulo and Hospital de Câncer de Barretos. Participants, with age ranging from 14 to 95 years, average 43 yo, median 43yo (SD = ±14 yo), consisted of 403 women undergoing routine screening (n = 156 from São Paulo and 247 from Barretos), in addition to 262 referral women with a previous abnormal Pap test in the year preceding study enrollment (163 from São Paulo and 99 from Barretos). A cervical sample was obtained and processed as described elsewhere [14].

Briefly, all samples were submitted to Cobas 4800 HPV (Roche Molecular Systems, Pleasanton, CA, USA) and the Hybrid Capture 2 (HC2, Qiagen, Hilden, Germany) assays. Only the HR-HPV cocktail probe was employed on the HC2 testing. Samples from women found positive for high-risk HPVs DNA by either the HC2 or Cobas HPV (n = 431) were submitted to HPV genotyping by the PapilloCheck assay (Greiner Bio-One, Frickenhausen, Germany) [14] plus a random sample of 223 negative by both assays and 11 samples with indeterminate or invalid results, totalizing 665 samples as shown in Table 1.

Table 1 Distribution of samples according to prior molecular testing results
Full size table

Molecular tests

Cervical scrapes obtained by the SurePath collection kit (BD SurePath ™ - TriPath, Burlington, NC, EUA) and preserved in liquid cytology medium were transported to the lab where DNA was extracted using the QIAamp DNA Mini Kit (Qiagen, Gaithersburg, USA), according to manufacturer’s instructions. For each reaction, 5 μL of DNA eluate was used in the PapilloCheck assay. Specimens containing the target DNA are hybridized to specific oligonucleotide probes immobilized on a DNA chip and detected by the binding of a Cy5-dUTP labeled oligonucleotide probe to the tag sequence. The DNA chip is scanned by the CheckScanner apparatus at wavelengths of 532 and 635 nm. This test detects HPV genotypes 6, 11, 16, 18, 31, 33, 35, 39, 40, 42, 43, 44, 45, 51, 52, 53.56, 58, 59, 66, 68, 70, 73 and 82. In addition, human ADAT1 gene (adenosine deaminase, tRNA specific 1) is used as an internal control to assess the quality of the DNA [36].

Cytology

Six hundred sixty five smears were examined by staff cytopathologists at both institutions and classified according to the Bethesda system [37]: NILM (Negative for Intraepithelial Lesion or Malignancy), ASC-US (Atypical Squamous Cells of Undetermined Significance), ASC-H (Atypical squamous cells – cannot exclude HSIL), LSIL (Low-Grade Squamous Intraepithelial Lesion); HSIL (High-Grade Squamous Intraepithelial Lesion) and ICC (Invasive Cervical Cancer). For analysis purposes in this study ASCUS and ASC-H were grouped under ASCUS class.

Statistical analyses

HPV genotype distribution analyses were performed using GraphPad Prism version 5.0. The x 2 and p trend test were used to evaluate statistical significance between genotype, cytological diagnosis and age. For all analyses, a p < 0.05 was considered statistically significant.

Results

Liquid based cytology

Cytological diagnosis of these samples included 499 (75 %) women with negative cytology (NILM) and 166 (25.0 %) with some cytological abnormality as follows: 54 (8.1 %) ASCUS; 66 (9.9 %) LSIL; 43 (6.5 %) HSIL and 3 (0.5 %) ICC.

When stratified by age group, LSIL was less frequent in women older than 45 years old (p trend = 0.01), whereas HSIL was more frequent (p trend = 0.03). All 3 ICC cases were found in women older than 45 years: 48, 52 and 84 years old.

HPV genotype distribution

Among the 665 samples analyzed 323 (48.6 %) were HPV positive by the PapilloCheck assay. Single type infection was observed in 65 % (210/323) of them, while 59 (18 %) had co-infection by 2 HPV types, 34 (11 %) by three, 14 (4 %) by four and 6 (2 %) by five or more types. The frequency of HPV infection decreased with aging (p trend <0.001); in women under 31 years old the prevalence of HPV DNA was 58.2 % (92/158) decreasing to 52.3 % (116/222) in women between 31 and 45 years and 40.4 % (115/285) in women older than 45 years old.

The most frequent HR-HPV type among HPV positive women was 16 (23.2 %), followed by 56 (21.0 %), 52 (8.7 %), 31 (7.7 %), 53 (7.7 %), 51 (7.4 %), 39, 59 and 66 (6.5 % each), 33 (5.3 %), 58 (5 %), 18 (5 %), 82 (4 %), 45 and 70 (4 %) each, 68 (3.4 %), 73 (2.5 %) and 35 (2.5 %). The frequency of LR- HPV types was: 42 (12.1 %), 6 (6.2 %), 44 (4.3 %), 43 (4 %), 40 (2.8 %) and 11 (1.5 %). No association between LR-HPV and HR-HPV types with age group was observed.

When stratified by women undergoing routine screening and those with a previous abnormal Pap test (referral), HPV 16 was detected more often in referral than in the screening group; 14.9 % vs 8.9 %. This was also observed for HPV 56, the second most common genotype in both groups, found in 13.7 and 7.9 % respectively. Table 2 shows the distribution of the more frequent HR-HPV genotypes in both groups.

Table 2 Frequency of HR-HPVs in screening (n = 403) and referral (n = 262) groups as determined by the Papillocheck assay
Full size table

Comparing the HPV result obtained by the PapilloCheck to the cytology diagnosis, we found 41.8 % (209/499) of HPV positive results among women with negative cytology; 55.6 % (30/54) ASCUS; 80.3 % (53/66) LSIL; 67.4 % (29/43) HSIL and 100 % (3/3) among women with ICC. Eight HSIL cases, 18.6 % (8/43) were positive either by Hybrid Capture 2 (HC2), Cobas HPV or both, while negative for the Papillocheck assay, whereas in the three ICC cases HPV was detected by all three methods.

HPV 56 and HPV 16 were the most frequent in negative cytology being found in 19.1 % (40/209) and 17.7 % (37/209) of the patients, respectively. In LSIL cases, HPV 56 was the most frequent type present in 28.8 % (19/66), while in HSIL and ICC cases, HPV 16 was the predominant type, detected in 37.2 % (16/43) and 66.7 % (2/3) of these samples, respectively whilst, the third ICC case harbored HPV 33. Distribution of HPV types according to the cytological diagnosis is shown in Fig. 1.

Fig. 1
figure 1

Frequency of individual HPV types according to cytological diagnosis

Full size image

Concerning samples that showed negative results in molecular screening by HC2 and Cobas HPV on the original study, 13.9 % (31/223) were HPV positive by the Papillocheck. Of these, 54.8 % (17/31) were positive for LR-HPV, whereas 45.2 % (14/31) were associated with HR- HPVs types 16, 33, 39, 51, 53, 56, 58, 59, 66, 68, 70 or 73. In contrast, 49 samples positive by both Cobas HPV and HC 2 methods tested negative by the PapilloCheck assay (19.8 %; 49/247).

The two HR-HPV types found more frequently in all ages from the screening group were HPV 16 and 56. HPV 16 was observed in 15.5 % (15/97) of samples from women under 31 years and in 7.7 % (14/183) of samples from women older than 45 years, followed by HPV-56 that was detected in 9.3 % (9/97) of the patients under 31 years and in 6.0 % (11/183) of the older than 45 years. In samples from women aged 31–45 years old, the most frequent HR-HPV type was HPV 56 (9.8 %; 12/123), followed by HPV 16 (5.7 %; 12/123). The most frequent HPV types according to age groups are depicted in Fig. 2.

Fig. 2
figure 2

Distribution of HPV genotypes by age (screening group, n = 403)

Full size image

When HR-HPV genotypes are clustered in HPV species according to the classification of de Villiers et al. [38], it may be observed that the most common species were alpha-9 and alpha-6, both composed exclusively of high-risk HPVs; (16, 31, 33, 35, 52, 58) and (53, 56, 66) respectively, followed by species alpha-7, alpha-1, alpha-10, alpha-5, alpha-8 and alpha-11 (Fig. 3).

Fig. 3
figure 3

Frequency of HPV genotypes among HPV+ samples as grouped by phylogenetic species

Full size image

Discussion

In the present study, the specimens analyzed were enriched from a group of HPV positive samples and a random subsample of specimens displaying HPV negative molecular testing results, as described above. Thus, it has to be considered that this rate of HPV infection does not represent the true prevalence in women submitted to cervical cancer screening, in whom HR-HPV DNA was detected in about 9 % of subjects under routine screening, nor the prevalence in the referral group, of 24 % in women with a previous abnormal Pap test [14].

Differences in the prevalence and distribution of HPV types worldwide have been observed in several studies [1, 39], and can be related to the complex interaction between distinct HPV types with host immunogenetic factors, age, geographic location of the patients or use of different sets of primers in HPV PCR and genotyping assays used in these studies [40, 41].

HPV 16 was the most frequent type in the present series as found in almost all screening population studies, irrespective of the study design [42–44]. An interesting finding of this study was that HPV 56 was the second most common type, as previously reported by Levi and colleagues [14], confirming the results of Bruno et al. [15] in a study conducted in women from the state of Bahia, Brazil, where it was detected in 14 % of the cervical samples. This finding was not verified in other studies performed in Brazil; therefore, presence of HPV 56 in such a high frequency demands further investigation.

In the current study, HPV 56 was the most frequent type among HPV+ women bearing NILM and LSIL cytology, found respectively in 19.1 and 28.8 % of these samples. In contrast Oliveira-Silva et al. [22] studied 84 NILM samples from Brazilian women reactive by Hybrid Capture II, observing only one HPV 56 infection (0.8 %) while Martin et al. [45], in Spain, reported the presence of HPV 56 in 2.4 and 7.9 % of their NILM and LSIL cytology specimens.

In fact, other studies using PapilloCheck assay reported rates of HPV 56 above the expectation [15, 46], which may be attributed to an increased sensitivity of the test for this particular type or, alternatively, reduced specificity [36]. Undertaking other molecular techniques such as type-specific PCR or DNA direct sequencing of these samples could dismiss the potential lack of specificity of this method.

In this study, HPV 16 was the predominant type found in 38.1 and 66.7 % of the HPV+ HSIL and ICC cases respectively. These findings corroborate the results of a systematic review and meta-analysis in Latin America and the Caribbean including 2446 cases of HSIL and 5540 of ICC that reported HPV 16 as the most frequent type, verified in 46.5 % of HSILs and 53.2 % of ICC cases [47]. In another study conducted in São Paulo, samples from histologically confirmed invasive cervical cancer (n = 175) also harbored HPV 16 as the most frequent type (77.6 %), followed by HPV 18 (12.3 %), HPV 31 (8.8 %), HPV 33 (7.1 %) and HPV 35 (5.9 %) [48].

Detection of of LR-HPV types in the current study is explained by the inclusion of probes for these genotypes in the PapilloCheck assay, depicting them when present in co-infections with the HR-HPV genotypes, or in single type infections in the HPV screening (Cobas-/HC2-) negative group. Regarding the LR-HPV types, HPV 42 was the most common as observed in other studies using the same method of genotyping [15, 49, 50]. This higher frequency of HPV 42 compared to other low risk types could potentially be attributed to specific features of the Papillocheck assay, as observed for HPV 56.

Currently, HPV DNA testing has been incorporated into cervical cancer screening algorithms to triage cases with indeterminate cytologic results (ASC-US) [51]; clinical follow-up after treatment, monitoring of patients without cytological lesions or only minor lesions that were identified by colposcopy; resolution of discordant results of cytology, colposcopy, or histology [52, 53]; or as an stand-alone primary screening tool [54, 55]. As seen in this study, HPV DNA is detected in the majority of the precursor lesions or ICC [56]. Several HPV tests are available in the market, and we evaluated methodologies with different targets and principles, nevertheless no test is infallible and false-negative results may possibly occur [57].

In the current study, HPV DNA was not detected in 32.6 % of HSIL samples. This rate is higher than found by Guo et al. [58], whom reported that 13.6 % of their HSIL cases displayed a negative HPV-DNA molecular test. False-negative results may be due to a low burden of HPV DNA, presence of a new type of HPV not targeted by the method used, inadequate sampling, cytological and histopathological missclassification or might have been the result of methodological or reproducibility errors [59, 60]. The fact that the E1 region, targeted by the Papillocheck assay, is located next to the E2 region in the HPV genome, could render it more vulnerable to partial deletion, that often occurs during the HPV DNA integration into the host cell DNA [36] verified on more advanced squamous intraepithelial lesions.

When stratified by age group, women under 31 years old had the highest HPV frequency which mostly represents transient infections [61]. However, for women aged 31 years old or more, test positivity may indicate persistent infection, which would require more intense monitoring, when molecular testing could be more effective.

Multiple infections were detected in 35 % of samples, and were more frequent in women under 31 years old. This observation is consistent with others studies [45, 62], and suggests that it may be associated to more intense sexual activity in younger women. It is well known that the prevalence of HPV infection and the occurrence of multiple infections decrease with an increase in age [63, 64].

Current US guidelines recommend genotyping only for HPV 16 and 18, and detection of either of these two HPV types is sufficient to warrant direct referral to colposcopy, avoiding additional triage methods. However, some studies suggested that it may be also clinically useful to identify other HPV types such as HPV 31, 33, 33/58 (combined), which showed comparable or increased risk for the development of CIN 2+ [65–67].

Conclusions

The nonavalent HPV vaccine appears to be safe and effective in preventing persistent infection and precancerous lesions associated with HPV types 16/18/31/33/45/52/58, as well as genital warts related to HPV types 6 and 11 [68]. Moreover, the currently available HPV vaccines have shown partial cross-protection against HR-HPV types [69] by alpha7/alpha9 high-risk HPV species but it should not provide cross-protection against HPV 56, which is an alpha 6 group species. Although HPV 56 was a common finding in our study, especially in cases of NILM and LSIL, it was not found in ICC and in HSIL cases suggesting lower oncogenic potential, corroborating the data from de Oliveira et al. [48], who did not find HPV 56 in any of the 175 ICC cases from São Paulo.

The prevalence and genotype distribution of human papillomavirus (HPV) provides the basis for designing HPV prevention programs. HPV types 53, 66, 68, 70, 73 and 82 included in the Papillocheck assay and other types such as HPVs 26 and 67 have been rarely but consistently identified as single HPV infections in about 3 % of cervical cancer tissues [70]. To describe the genotypic profile pre and post HPV immunization is an important step in prevention strategies for cervical cancer and further studies are needed to elucidate the potential role of these HPV types in the development of malignancies, eventually justifing their inclusion in new HPV vaccine formulations. Such decisions have to include careful estimation of effectiveness and cost-benefit.