Intracellular Human Papillomavirus E6, E7 mRNA Quantification Predicts CIN 2+ in Cervical Biopsies Better Than Papanicolaou Screening for Women Regardless of Age

We prospectively followed 3133 women between 19 and 75 years old in a higher risk urban screening population. Included in this study were 1088 women younger than 30 years and 2045 30 years and older. Routine cervical cytology alone or with HPV DNA cotesting was performed, and women were referred for colposcopy and biopsy based on American College of Obstetricians and Gynecologists (ACOG) recommendations.15 These recommendations include referral based on low-grade squamous intraepithelial lesion (LSIL) or high-grade squamous intraepithelial lesion (HSIL) cytology or an atypical squamous cell of undetermined significance (ASCUS) cytology and an HPV DNA positive reflex test. Of the women sent for colposcopy and biopsy who also had HPV E6, E7 mRNA testing, 114 were 30 years and older and 132 were younger than 30 years. Cervical cytology specimens were collected using a cytobrush and preserved using either PreservCyt (Hologic, Inc, Marlborough, Massachusetts) or SurePath (Becton-Dickinson, Burlington, North Carolina) liquid-based cytology fixatives. Sample age was restricted to less than 2 weeks of age to maintain integrity of the RNA.7 The cytology smears were classified using the Bethesda System 2001 criteria.31 

HeLa cells were used as a positive cell control for the E6, E7 mRNA assay. These cells were obtained from the American Type Culture Collection (ATCC) and grown according to the manufacturer's instructions. Normal human ectocervical cells (Lonza, Inc, Walkersville, Maryland) were used as negative control cells for the E6, E7 mRNA assay and were grown in medium supplied by the company. In addition, positive and negative HPV control cells were purchased in liquid-based cytology fixative (IncellTrol; IncellDx, Menlo Park, California).

A 1-mL aliquot was removed from the residual liquid-based cervical cytology specimen. The cells were pelleted by centrifugation at 400g and washed once in phosphate-buffered saline (PBS), pH 7.4. The cells were fixed and permeabilized (IncellDx) for 1 hour at ambient temperature. Following fixation and permeabilization, the cells were washed twice and pelleted by centrifugation. In cell E6, E7 mRNA quantification was performed by resuspending the cells in hybridization mixture and a cocktail of oligonucleotide probes (HPV OncoTect; IncellDx).6,7,16,17 Hybridization was performed at 43°C for 30 minutes and was followed by a 5-minute high stringency wash and a 15-minute high stringency wash. The cells were resuspended in PBS, pH 7.4, with 2% fetal calf serum for flow cytometric analysis.

Cell analysis was performed with a compact tabletop flow cytometer (GoldFish analyzer; IncellDx) with an argon laser capable of analyzing forward scatter, side scatter, and three fluorescence channels (IncellDx). Ectocervical cells were identified using forward scatter and side scatter as previously described18–20 and gated using software for HPV E6, E7 mRNA overexpression.

To determine the performance of cervical cytology in predicting high-grade (CIN 2+) lesions by biopsy in our cohort, we followed the ACOG recommendations for referring colposcopy and biopsy in women based on cervical cytology. As demonstrated in Table 1, only 9% of women 30 years and older and 12.5% of women younger than 30 years with an ASCUS cytology had CIN 2+ on biopsy. Similarly, only 18% of women 30 years and older and 12.5% of women younger than 30 years with an LSIL cytology had CIN 2+ on biopsy. Forty-six percent of women with HSIL cytology had CIN 2+ on biopsy. These rates for cytology prediction of biopsy-proven lesions are consistent with previous publications.21 

Quantification of cells overexpressing E6, E7 mRNA (HPV OncoTect) correlates with the severity of cervical disease as determined by biopsy. To determine the performance of HPV E6, E7 mRNA quantification for the detection of CIN2+ on biopsy, we assayed cervical scrapings samples from 3133 women, 246 of whom had follow-up colposcopy and biopsy based on ACOG recommendations. Of the women sent for colposcopy and biopsy who also had HPV E6, E7 mRNA testing, 114 were 30 years and older and 132 were younger than 30 years. To test the hypothesis that an assay detecting overexpression of HPV oncogenes (which typically follows integration), would demonstrate superior predictive value for CIN 2+ compared to an assay such as HPV DNA, which detects only the presence or absence of high risk HPV, we compared the sensitivity and specificity of HPV E6, E7 mRNA for CIN 2+ and CIN 3+ in both sets of women enrolled in the study. In women 30 years and older, HPV E6, E7 mRNA had a sensitivity of 89% for CIN 2+ and 92% for CIN 3+ (Table 2). Interestingly, in women younger than 30 years, HPV E6, E7 mRNA had a sensitivity of 82% for CIN 2+ and 93% for CIN 3+ (Table 2). The positive predictive value for CIN 2+ was 71% in all women and 64% in women younger than 30 years. For women younger than 30 years with an LSIL cytology, HPV E6, E7 mRNA had a 100% sensitivity and 61% specificity for CIN 2+ on biopsy (data not shown). The sensitivity and specificity of HPV E6, E7 mRNA in women younger than 30 years demonstrate the utility of this cervical cancer screening method in an age group in which high risk HPV DNA assays lack clinical utility.

To compare the performance of E6, E7 mRNA quantification with abnormal cytology for the detection of high-grade cervical lesions, we determined the sensitivity, specificity, and positive predictive value (PPV) of E6, E7 mRNA quantification and abnormal cytology for CIN 2+ and CIN 3+ lesions (Table 3). In women 30 years and older, the sensitivity for the detection of CIN 2+ or CIN 3+ lesions is identical for E6, E7 mRNA and abnormal cytology, yet E6, E7 mRNA is much more specific and exhibits a much greater PPV in both histologic categories (Table 3). In women younger than 30 years, E6, E7 mRNA quantification demonstrates a slightly lower sensitivity than abnormal cytology for CIN 2+ (88% compared to 96%, respectively) and similar sensitivity for CIN 3+ (92% compared to 93%, respectively). The most profound finding is the specificity of E6, E7 mRNA compared to abnormal cytology for CIN 2+ (88% compared to 2%, respectively) and for CIN 3+ (81% compared to 2%, respectively). Similarly the PPV for E6, E7 mRNA was greater than that of abnormal cytology for the detection of CIN 2+, 64% versus 19%, and for the detection of CIN 3+, 36% versus 10%. Of the nine LSIL cytology samples that were CIN 2+, E6, E7 mRNA was positive in all nine; however, 63 LSIL samples had a biopsy diagnosis of CIN 1 or benign and E6, E7 mRNA was positive in only 13 of 63 LSIL-positive samples with CIN 1 or benign histology.

The use of HPV testing for cervical cancer screening is growing worldwide. HPV DNA assays determining the presence or absence of high-risk HPV genotypes have been extensively utilized over the past decade; however, newer HPV RNA-based assays are becoming popular because of the hope of increased specificity compared to HPV DNA assays including genotyping. One of the issues with other HPV RNA assays such as GenProbe Aptima and BioMerieux NucliSENS EasyQ HPV is that they genotype HPV from the E6, E7 transcript22 rather than provide a quantitative assessment that can distinguish overexpression of E6, E7 mRNA in cells, the hallmark of cervical carcinogenesis. Here, we report the use of a high-throughput assay that quantifies the effects of HPV on two levels. First, this assay quantifies the amount of E6, E7 expression in cells on a cell-by-cell basis, using high-throughput flow cytometry. Second, this assay quantifies the number of cells overexpressing E6, E7 in the squamous cell compartment of the cervical cytology specimen (Figure).

The role of HPV-encoded E6 and E7 genes in cervical cancer is well defined.23,24 In fact, E6 and E7 mRNA expression is ubiquitous (31 of 31 cervical cancers) in cervical cancer.13 Of significance for cervical cancer screening, E6, E7 mRNA expression in cancer occurs irrespective of HPV genotype. In other words, once a high-risk genotype overexpresses E6, E7 mRNA, the biological behavior no longer depends on the genotype.12,13 These data in combination with studies showing E6, E7 expression correlating with the severity of cervical lesions7 are consistent with the data reported in the present study. Sotlar et al12 found sliced E6, E7 oncogene transcripts by RT-PCR in 18% of CIN 0 (no dysplastic epithelium), 58% of CIN 1, 77% of CIN 2, and 84% of CIN 3.12 Similar results are presented here, using HPV OncoTect rather than RT-PCR for the detection of E6, E7 mRNA. In our screening cohort, the sensitivity of HPV OncoTect was 89% for CIN 2+ and 100% for CIN 3+ lesions in women 30 years and older. In women younger than 30 years, the sensitivity of HPV OncoTect for CIN 2+ lesions was 88%, and 92% for CIN 3+ lesions. Abnormal cytology exhibited a sensitivity of 89% for CIN 2+ and 100% for CIN 3+ in women 30 years and older and 96% sensitivity for CIN 2+ and 93% sensitivity for CIN 3+ in women younger than 30 years. The specificity of HPV OncoTect was greater than 80% for CIN 2+ and CIN 3+ in both groups of women compared to a specificity of abnormal cytology of less than 10% for CIN 2+ and CIN 3+ in both groups.

The specificity of our E6, E7 mRNA assay was 94.5% based on normal cytology and greater than 80% based on biopsy samples that were CIN 2, which includes benign biopsies and CIN 1. If CIN 1 is excluded from the analyses, as it may not be a true negative endpoint, the specificity of our E6, E7 mRNA assay exceeds 95% (data not shown) for CIN 2+ in all age categories. Our data from this screening cohort including women younger than 30 years demonstrate equivalent sensitivity and superior specificity compared to that of HPV RNA and HPV DNA assays, comparatively studied in an abnormal cytology referral population.25 

Natural history studies of HPV infection in women under 30 clearly demonstrate the transient nature of HPV infections in this age group, including infections with high-risk genotypes.4,26 Abnormal cervical cytology and histology were associated with HPV persistence as defined by consistently positive tests for high-risk HPV DNA.4,26 At best, sequential high-risk HPV DNA-positive tests are a surrogate for HPV infections that result in cellular transformation, the process leading to cervical cancer that is driven by overexpression of E6 and E7. A diagnostic test that is sensitive and specific enough to detect precervical cancer lesions in women under 30 is needed, as 15% of cervical cancer occurs in women ages 20–34 and the potential productive years lost in young women who succumb to cervical cancer is great. The performance of our assay in women under 30 years was comparable if not better than the performance of other HPV DNA or HPV RNA assays in women more than 30 years old. In particular, we demonstrated a 100% sensitivity and 61% specificity for CIN 2+ lesions in women younger than 30 with LSIL cytology. This finding, if confirmed by larger studies, would be particularly useful in this age group as therapeutic procedures can have consequences for child-bearing in young women.27,28 

Our data suggest that HPV E6, E7 overexpression is a more specific marker of high-grade cervical lesions that require intervention and treatment. Although one might argue that more prospective studies are necessary to define the natural history of high-grade CIN 2+ lesions overexpressing E6, E7, ethical issues make these studies difficult because of the risks of not treating affected women. Of note, however, 4 untreated women in our study with CIN 1 and overexpression of E6, E7 progressed to CIN 2+ on follow-up biopsy (data not shown). In addition, a prospective study from Spain that used liquid-based cervical cytology samples collected over 24 months demonstrated that of LSIL cytology samples, 91% that were positive for E6, E7 overexpression by the same HPV OncoTect assay used in the present study progressed, while 85% that were E6, E7 negative regressed over the period of the study.29 The balance of the study samples carried a stable diagnosis over time.

In summary, the increased specificity for more severe, treatable lesions as demonstrated here likely would reduce the number of unnecessary colposcopies that would result from overcalling CIN 2+ lesions using HPV DNA. The incidence of overcalling (false positives) CIN 2+ lesions has been shown to be as high as 28% in some studies.30 The assay used in the current study uses a small benchtop flow cytometer that can accommodate tubes or 96-well microtiter plates. This instrumentation already exists in many developing countries to perform CD4 counts in human immunodeficiency virus-infected individuals. The combination of an assay that can use an installed instrumentation base worldwide or with dedicated instrumentation without sacrifice in performance has global implications for improving screening for cervical cancer, which is still the second leading cause of cancer death worldwide.