Human papillomavirus detection using the Abbott RealTime high-risk HPV tests compared with conventional nested PCR coupled to high-throughput sequencing of amplification products in cervical smear specimens from a Gabonese female population

Infections with human papillomavirus (HPV) are the most common sexually transmitted infections worldwide [1]. Persistent cervical infection is a high risk factor for cervical cancer, which is the fourth most common malignancy in women worldwide. Moreover, there is a clear etiological link between the persistence of HPV infections in epithelial cells of the cervical mucosa, precancerous lesions (i.e., cervical intraepithelial neoplasia 3; CIN3) and cervical cancer [2]. Today, more than 200 HPV genotypes have been defined, among which 50 can infect the cervical epithelia. However, only 16 HPV types (16, 18, 31, 33, 34, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, 70) are classified as high risk (hrHPVs) and 14 (excluding 34 and 59) show strong evidence of a causal link to cervical cancer [3].

Based on the clear causal link between hrHPVs and cervical cancer, the guidelines of the American Society for Colposcopy and Cervical Pathology recommend HPV DNA testing instead of cytology to carry out triage in a screened population. Molecular tests for HPV detection may be a good alternative to cytology for early screening because an HPV-negative result is sufficient to declare an extremely low risk of developing CIN3 or worse (CIN3+) for 5 years in contrast to a negative cytology result [4]. Other recent studies have confirmed that HPV type identification has significantly higher sensitivity than cytology [5]. Consequently, this higher sensitivity may be the reason behind an increase in the number of false positives, i.e. samples lacking high-grade lesions (CIN2+) [6]. Therefore, a single positive HPV test may be insufficient to identify women who run the risk of developing cervical neoplasia [7]. In response to this caveat, Meijer et al. (2009) published novel guidelines for assessing the performance of a new HPV DNA assay. Molecular assays must detect hrHPV infections that are preferentially associated with CIN2+, and thus they should have at least 90% sensitivity with GP5+/6+ primers or the Hybrid Capture 2 (hc2) test [8]. The hc2 test was approved by the United States Food and Drug Administration (FDA) as the gold standard for the detection of 13 hrHPV genotypes in 1999. However, this test has some disadvantages, including the non-discrimination of HPV16 or HPV18 from other hrHPVs and the absence of an internal control [7]. Furthermore, the hc2 test is associated with cross-reactivity of the probe mixture with untargeted HPV types [7]. To overcome this limitation, new HPV DNA testing assays (Abbott, Wiesbaden, Germany; LG Life Science, Seoul, Korea, etc.) based on real-time PCR, have been developed and can distinguish HPV16 or HPV18 from other major hrHPVs, and some are currently approved by the FDA for clinical use [7]. These molecular tests are recommended by the Guideline Development Group for the screen-and-treat strategy because their contribution may be greater than visual inspection with acetic acid and/or Lugol’s iodine (VIA/VILI) to reduce cervical cancer and its related mortality. Moreover, their implementation in low- and middle-income countries may be easier than cytology-based screening programs [9]. Furthermore, the quality of the molecular tests is not affected by most bacterial infections concomitant to the HPV infection. Finally, the level of technical training of the healthcare staff has only a minor impact on the performance of the test.

This study describes the first use of the Abbott RealTime High-Risk HPV assay for HPV detection in cervical cell samples of women with or without cervical abnormalities from a Gabonese female population. Moreover, this assay was compared with the use of conventional nested PCR coupled to HTS, which identifies the exact hrHPV genotypes revealed with the AbRT assay. The real-time PCR method (AbRT assay) can detect a total of 14 different high-risk HPV genotypes whereas the conventional nested PCR that uses the primers MY09/11 and GP5+/6+ can amplify around 30 different low- and high-risk HPV genotypes [13, 14]. The genotypes are assayed from DNA extracted from cervical smears whose cervical anomalies have been revealed by a VIA/VILI test and confirmed by cytology. Although the manufacturer recommends using the Abbott automaton (m200sp, Abbott Molecular) with this assay for HPV detection, extractions were performed manually as described in literature as a practicable alternative. In a recent study, Kocjan et al. (2012) showed that similar results are obtained for HPV detection from biological tissues of head-neck squamous cell cancers regardless of the extraction method used with the AbRT assay [15, 16]. Moreover, we did not observe any differences in amplification intensity of the human β-globin (a housekeeping gene) between the AbRT kit and a classic qPCR protocol (data not shown).

In this study, the AbRT generally gave consistent genotyping results for the hrHPV screened in HSIL+ cases in comparison with the conventional nested PCR. These data are similar to other studies that compared the AbRT kit with this nested PCR method or with other molecular methods of HPV genotyping [15, 16]. However, nested PCR appears to be more sensitive than the AbRT kit for the detection of hrHPV in HSIL- cases (i.e. LSIL, ASCUS and normal) even if in most cases, at least one hrHPV genotype was detected by both methods (Table 3). Given that the detection threshold of the AbRT assay is between 500 and 5000 genome copies, genotypes detected using both methods in this study may have a higher number of copies than this threshold [17], suggesting that cases for which only one genotype was detected by the nested PCR method contain fewer than 500 copies. However, this hypothesis remains to be tested because we have no data on the actual viral loads of our samples. Furthermore, although the correlation between the number of reads and viral load is not perfect, cases with genotypes detected by both methods were generally those with highest numbers of sequenced reads [10]. Owing to the depth of sequencing with a minimum of several hundred thousand reads per sample, a minor HPV genotype can be detected if it is represented in at least 2% of the total reads [10]. Therefore, in a given sample, HTS can unambiguously detect several genotypes whose relative proportions potentially vary [18, 19]. Moreover, although the sequenced amplicon size was only 150 bp, this region of the L1 gene is highly amenable to molecular analysis, which can discriminate between the different HPV genotypes. Finally, the HTS approach for sequencing HPV amplicons may overcome the limit of direct amplicon sequencing using the Sanger method, which is simple to implement from a biopsy, but cannot be performed on DNA extracted from a (LBC) cervical smear, especially when the cytological grade is low or normal, due to the potential simultaneous presence of several HPV genotypes [20]. Although HPV genotyping assays using consensus or broad-spectrum primers are less sensitive than type-specific or targeted primers [21], consensus PCR is less affected by competitive primer binding in mixed HPV infections. For instance, the presence of high viral loads of HPV16 or, in some cases, the presence of some hrHPVs can mask the HPV52 genotype in anogenital samples [22] or inhibit the amplification and detection of HPV31 and 33 [23]. Moreover, unfavorable amplification kinetics may occur when an HPV genotype belonging to the alpha-9 subgenus is in the presence of other genotypes (Iftner et al. (2016).

Several studies have shown that the sensitivity of the AbRT assay is similar to that of the hc2 test and that the specificity of detection of HPVs associated with the development and/or spread of CIN2+ can reach 92% [16, 24, 25]. GP5+/6+ PCR has the same clinical performance as the hc2 test [8]. However, the addition of the first nested PCR step with MY09/11 primers improves the sensitivity of the conventional PCR test, but leads to a decrease in its specificity for the detection of HPV genotypes shown to be associated with the development of HSIL+ into cancer [26]. Furthermore, the nature of the samples used in this study clearly influences the differences observed between the two methods. In contrast to samples from biopsies, LBC samples recover numerous superficial cells in addition to the cells located in a putative precancerous lesion. However, these superficial cells may be infected by the same hrHPV genotype or other hrHPV genotypes. The use of a sensitive HPV detection method may reveal these supplementary infections and not just the HPV implicated in the (pre)cancerous lesion. In effect, due to this high sensitivity, most studies on the clinical validation of HPV detection tests are carried out on FFPE tissue blocks, not LBC samples, and cytological results are generally confirmed by a histological analysis to confirm the grade of the observed lesion. Therefore, the differences between the two molecular methods for HPV detection are most likely due to the nature of the samples used in this study and may have led to variation in the performance of the test and the specificity of the AbRT compared with the data found in literature [15, 24, 27‐29]. However, the use of LBC led to a better preparation of the sample associated with an enhanced quality of the cytological results. Recent studies have shown improved slide reading when LBC systems are used instead of conventional Pap smears in a routine clinical setting. Moreover, LBC using PreservCyt (Thinprep) has been approved by the FDA for molecular HPV tests [30]. This FDA approval provides the opportunity to carry out two biological analyses from a single sample, a certain advantage in low-income countries or when the population has limited access to healthcare.

This study on the performance of the AbRT assay shows that it is an easy and reliable molecular tool for HPV detection in HSIL+ LBC samples. However, as expected, the AbRT assay is less sensitive than conventional nested PCR for the detection of HPV in HSIL-. Moreover, the sequencing of amplicons using a HTS approach effectively identified the genotype of hrHPV detected with the AbRT assay. Finally, our data showed that the use of AbRT in co-testing with VIA/VILI in low-income countries holds promise as an interesting alternative to other combinations, such as VIA and cytology.