Introduction

Cervical cancer is a major contributor to cancer-related morbidity and mortality in world and is the fourth most common cancer in women [1]. In India, 453 million women (Age > 15) were at risk for developing cervical cancer. A total of 142,844 cases of carcinoma cervix were reported causing 67,477 deaths with incidence of 20.2 per 100,000 women [1]. Diagnosis of precancer and cancer cases are mostly depends on Pap smear preparation of cervical epithelial cells.

Liquid-based cytology (LBC), including SurePath and ThinPrep, has emerged as better source to the conventional Pap smear and has reduced the incidence of cervical cancer worldwide with improved detection of high-grade squamous intraepithelial lesion (HSIL) and in prevention of invasive cancers. In a study of Rozemeijer et al. 2016, they found that SurePath was associated with increased CIN2+ detection. The implementation of cytological testing and screening for detection of premalignant cervical lesions has proven to be invaluable in reducing the incidence of this malignancy. Availability of residual material in liquid-based cytology provides subsequent molecular testing for Human Papilloma Virus (HPV), methylation, immunocytochemistry, and DNA ploidy [2, 3]. Ancillary studies are helpful in providing information beyond morphology utilize in understanding of disease progression from precancer to cancer.

Screening of cervical precancer and cancer is based on the fact that infection with oncogenic HPV plays a crucial role in development of cervical cancer. However, other molecular changes such as alteration of gene product regulating oncogenesis, tumor suppression, DNA repair, apoptosis, metastasis, and invasion are necessary for developing malignant lesions. DNA mutations including deletions or epigenetic alterations may result in such alteration. DNA methylation also plays role in maintaining genomic stability and gene expression regulation [4].

Studies report HPV as a main etiologic agent for most of premalignant and malignant lesions [5, 6]. HPV DNA testing is of importance, but HR-HPV (Higher Risk-HPV) testing often fails in accurate differentiation between patients whose lesions will persist or progress to invasive lesion and those whose lesions will relapse spontaneously. Integration of viral oncogene in the effected cell clone results in deregulated viral oncogene expression and this appears to results in chromosomal instability and progression from precancerous to cancerous lesion and aneuploidy of the transformed cells [7]. Gene promoter hypermethylation, chromosomal aberrations, and disruption of normal cell cycle triggered by the oncogenic virus may also lead to variation in nuclear DNA content [5, 8]. There is evidence that the persistence of cervical carcinoma is related to aneuploidy [9]. Aneuploidy reflects a situation of uncontrolled increase of DNA and loss of essential information and plays an important role in neoplastic transformation.

P16 protein tightly regulates the cell cycle and its expression in normal dysplastic cells is very low. Due to transforming activity of E7 oncogene of HR-HPV, p16 is strongly expressed in dysplastic cervical cells. P16 measures active HPV gene expression rather than presence of viral gene only and has been detected in studies as an important biomarker for detection of cervical intraepithelial neoplasia (CIN) [10, 11]. Tumor suppressor gene silencing and other cancer-associated gene methylation in CpG islands of promoter region are a frequent occurrence in human cancer. Methylation in CpG regions is associated with transcriptional block and loss of essential protein. This aberrant methylation detection may be applied as a diagnostic approach based on methylation-specific PCR techniques [12]. However, no single technique can diagnose precancerous lesions with high sensitivity and specificity. Hence, the current study was designed to assess the diagnostics of P16INK4a immunoexpression, p16 promoter hypermethylation, Human Papilloma Virus (HPV) detection, and DNA ploidy studies individually and in combination in LBC samples of cervical precancer and cancer.

Materials and methods

Patient samples and procedures

The study sample comprised of 95 high-grade (41 HSIL and 54 SCC) and 69 low-grade cytology (26 Normal, 22 ASC-US, and 21 LSIL). Hospital-based collection was done from the department of Obestrics and Gynaecology, Queen Mary’s Hospital, King George’s Medical University and Dr. Ram Manohar Lohia Combined Hospital, Lucknow, India. Samples consecutively collected during cervical screening of women. We subjected PAP screening and samples showing cytological evidence of epithelial abnormalities as per the Bethesda guidelines were included in the study. The age-matched controls were also included.

Cytology

Samples were collected in ThinPrep (TP) vials (Hologic, Inc. USA) containing 20 ml PreservCyt (ThinPrep, Hologic, Boxborough, MA) as per the instructions provided. Smears were prepared on ThinPrep 2000 processor (CYTYC Corporation and cytologic screening was done by the conventional Pap staining. All slides were evaluated by two pathologists (NH, NA) and diagnosis was made using the Bethesda System (2014). Age-matched cases with normal cytology were used as a control for Methylation, HPV, p16 staining, and DNA ploidy analysis by flow cytometry.

DNA extraction

DNA extraction was done by DNA extraction kit (Invitrogen, USA). The quality and quantity of DNA was measured by taking optical density on A260/280 parameter by Nanodrop (DeNovix, USA). Isolated DNA was stored at – 20 °C until further processing.

Expression of p16 protein by immunocytochemistry (ICC)

The ICC procedure was performed on fresh smear made from material remaining in LBC vials. Briefly, ICC process has done by standard protocol including rehydration of the smears in decreasing concentration of alcohol. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide (Merck, India) in methanol for 30 min. Antigen retrieval was done by placing the slides in citrate buffer (pH 6.0) in water bath at 94 °C for 40 min. Slides were cooled at room temperature (RT) and incubated with p16 primary antibody (G175-405, Biogenex, Fremont, CA USA) at RT for 1 h, followed by treatment with polymer-based secondary antibody kit with DAB (DAKO, Denmark). Positive reactions were visualized using diaminobenzidine, DAB (1:50). Sections were finally counter-stained with 0.1% hematoxylin. The positive cells expressing the p16 positivity were assessed by the cytoplasmic as well as nuclear staining at high magnification (40×) (Fig. 1a–h).

Fig. 1
figure 1

PAP and p16 expression in LBC smears

Full size image

Methylation-specific PCR for p16 promoter hypermethylation

Bisulphite modification

With the bisulfite modification, unmethylated cytosines are converted to uracil, while methylated cytosines remain unmodified. Bisulfite conversion of genomic DNA (500 ng-1 µg) was performed using a commercially available EpiTect plus DNA Bisulfite Kit (Qiagen, USA). Briefly, 500 ng–1 µg of DNA was denaturated by incubation in bisulfite mix and DNA protect buffer for 5 h at conditions provided in the kit. After desalting, DNA was desulfonated in desulphonation buffer at room temperature for 15 min, followed by washing with wash buffer. Treated DNA was eluted in elution buffer and immediately stored at − 80 °C till further processing. The modified DNA was used as template for methylation-specific PCR (MSP).

Conventional MSP

Methylation status of the p16 gene was determined by the method of MSP. Conventional MSP PCR was carried out in a 20 µl reaction mixture containing 2 µl of bisulfite converted DNA, 2 µl of 10X PCR buffer, 0.5 µl of 25 mM dNTPs, 0.5 μM of each forward and reverse primer, and 0.75U Taq DNA polymerase (Invitrogen, USA), and the volume was adjusted with nuclease-free water. PCR was performed on PCR system (Bio-Rad Laboratories, Hercules, CA, USA) with thermal cycling conditions of first denaturation at 95 °C for 5 min, followed by 35 cycles of 95 °C for 30 s, 67.3 °C for 30 s, and 72 °C for 30 s.

Control without DNA and positive control for Unmethylated (U) and Methylated (M) reactions was run in each batch. The PCR product was visualized in agarose gel stained with ethidium bromide under ultraviolet illumination. DNA methylation was determined by the presence of a 150 bp and 151 bp fragments in those samples amplified with the p16-M and p16-U primers, respectively. Positive PCR products in p16-M and/or p16-U were interpreted as methylated and PCR products in only p16-U were interpreted as unmethylated. Heterogeneous methylation status was determined by presence of both M and U bands in same sample. Methylated band intensity of standards and cases was measured by Image J software (NIH, USA). Quantification of methylation status was done by comparing the intensity with standards of known concentration. Universally methylated, unmethylated DNA standards (Qiagen, USA), and no template control (NTC) were included in each PCR run. Standards of 100, 75, 50, 25, and 5% from universally methylated and unmethylated DNA were prepared and run in each MSP assay to quantify percent of methylation in unknown target sample (Fig. 2a, b).

Fig. 2
figure 2

Gel with a Methylated and b Unmethylated p16 promoter. M is 100 bp DNA ladder, L1-L6: Methylated and unmethylated DNA controls of 100, 75, 50, 25 and 5% (r2 = 0.988); L7: 2a: methylated case (71.17%), b unmethylated case

Full size image

The sequences of the primers for methylation-specific PCR were as follows:

Gene

Primer sequence

Annealing T (°C)

Size (bp)

p16-Ink4a—UF

5′ TTATTAGAGGGTGGGGTGGATTGT 3′

67.3

151

p16-Ink4a—UR

5′ CAACCCCAAACCACAACCATAA 3′

p16-Ink4a—MF

5′ TTATTAGAGGGTGGGGCGGATCGC 3′

67.3

150

Pp16-Ink4a—MF

5′ GACCCCGAACCGCGACCGTAA 3′

DNA ploidy by flow cytometry in cervical precancer and cancer

LBC sample preparation

5 ml of LBC sample was centrifuged (2000 rpm for 5 min) to get the cell pellet followed by 3 wash of phosphate buffer saline (PBS, pH 7.4). Epithelial cell count was done using Neubauer chamber and sample with count of 1 × 105 cells or more were considered appropriate for flow cytometry analysis.

Internal standard

Internal DNA standards (DNA QC Particle kit B.D. Biosciences, Singapore) were used for verification of instrument performance in term of coefficient of variation (CV = 1–3%) and linearity (1.95–2.05) of the fluorescence pulse detector.

Staining and acquisition of samples

Cells were stained with Telford reagent as described by Mishra et al. [13] acquired on flowcytometer (FACSCalibur: Becton, Dickinson & Co.; San Jose, CA), equipped with 488-nm argon laser and four-color filter using Cell Quest Pro software. Events were acquired in R1 gate drawn around the epithelial cells based on light scatter properties of cells. All measurements were performed at low flow rate (Figs. 3a–d, 4a–d). The cell cycle profiling of all samples was done in an unchanged background using the same instrument and by the same observer.

Fig. 3
figure 3

Acquisition of NILM case. a–c Shows acquisition of stained cells on FSC versus SSC, FL2-A versus FL2-W and FL2-A versus Count on Cell Quest Pro software (B.D Biosciences, Singapore). d Shows analysis of acquired FCS file on ModFit LT 3.2 (Verity software house) and cases was found to be Diploid

Full size image
Fig. 4
figure 4

Acquisition of HSIL case. a–c Shows acquisition of stained cells on FSC versus SSC, FL2-A versus FL2-W and FL2-A versus Count on Cell Quest Pro software (B.D Biosciences, Singapore). d On ModFit LT analysis case was found to be aneuploid

Full size image

Cell cycle analysis and ploidy

Data obtained were analyzed using the ModFit LT software (DNA Modelling System) version 3.2 (Verity Software House, Inc Topsham, ME).

HPV DNA testing

HPV was detected using PGMY09/11 primers as designed to amplify a 450 bp HPV L1 gene fragment [14]. PCR was carried out in a volume of 20 μl containing 50 ng/μl of genomic DNA, 4 mM of MgCl2 (Invitrogen, USA), 200 μM of each dNTP, 5 pmol of PGMY09/11 primers, 5U of AmpliTaq DNA polymerase (Invitrogen, USA), and final volume make up to 20 μl by adding nuclease-free water. PCR was performed on thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA) with thermal cycling conditions of first denaturation at 95 °C for 9 min, followed by 35 cycles of 95 °C for 60 s, 55 °C for 60 s, and 72 °C for 60 s. Positive and negative controls (without DNA) were included in each run. Agarose gel electrophoresis was performed to check the amplification of 450 bp for HPV positive cases (Fig. 5).

Fig. 5
figure 5

HPV PCR Conventional PGMY09/11: Agarose gel showing cases L1-negative, L2-L6-positive (450 bp product) and M-100 bp DNA ladder

Full size image

Statistical analysis

Statistical analysis was performed using the SPSS (Statistical Package for the Social Sciences) software package, version 16.0. The results are presented in Mean ± SD and percentages. The Chi-square test was used to compare the categorical variables and unpaired t test was used to compare the continuous variables. Diagnostic (sensitivity and specificity) of the cell cycle parameters was done using the cut-off values obtain from receiver-operating characteristics (ROC) curve analysis. A p value of less than 0.05 was considered significant. The sensitivity for CIN2+ of the p16, p16 promoter hypermethylation, and HPV is defined as the proportion of high-grade cytology (HSIL and SCC) detected by the different tests. In the calculations of specificity and negative predictive values (NPV), it is assumed that women with low-grade cytology (Normal, ASC-US, and LSIL) do not have CIN2+.

Ethical approval

The study was conducted at the molecular pathology lab of the Department of Pathology, Dr. Ram Manohar Lohia Institute of Medical Sciences, Lucknow, India. Ethical approval was obtained from Institutional Ethics Committee of Ram Manohar Lohia Institute of Medical Sciences before recruiting patients. All participants signed an informed consent and a full explanation was given about techniques. Participants were excluded from the study if they were not willing to participate or cases with malignancy under follow-up/prior therapy.

Results

P16 expression in cervical precancer and cancer cases

The mean (± SD) of p16 expression in ASC-US, LSIL, HSIL, and SCC was 16.3 (± 3.0), 18.4 (± 5.7), 34.5 (± 9.9), and 41.3 (± 15.0), respectively. The test positivity rate of p16 expression in women with ASC-US, LSIL, HSIL, and SCC was 22.7, 38.9, 68.0, and 85.3% (Table 1).

Table 1 Positivity rate of p16 expression, p16 methylation, and HPV in cervical precancer and cancer cases
Full size table

P16 gene promoter methylation and HPV in cervical precancer and cancer cases

The mean (± SD) of the methylation index in ASC-US, LSIL, HSIL, and SCC was 38.2 (± 19.4), 39.1 (± 24.6), 85.8 (± 34.8), and 98.0 (± 42.8), respectively. The test positivity rate of p16 promoter methylation in women with ASC-US, LSIL, HSIL, and SCC cytology was 36.3, 76.1, 92.6, and 92.5%, respectively (Table 1). The test positivity rate of HPV in women with ASC-US, LSIL, HSIL, and SCC was 45.4, 76.2, 87.8 and 92.5% respectively (Table 1).

Cell cycle parameter value in cervical precancer and cancer cases

The test positivity rate of diploid G1 value in women with LSIL, HSIL, and SCC was 50.0, 65.0, and 96.8% respectively. The test positivity rate of diploid S value in women with LSIL, HSIL, and SCC was 50.0, 70.0, and 71.4% respectively (Table 2). Diploid G1 and diploid S values significantly (p < 0.05 or p <  0.01) discriminate LSIL versus HSIL and LSIL versus SCC.

Table 2 Positivity rate of cell cycle parameter values in cervical precancer and cancer
Full size table

Combined positivity of p16 expression and p16 promoter methylation in ASC-US, LSIL, HSIL, and SCC was 31.2, 63.6, 77.1, and 97.3 respectively (Table 1). However, the positivity of HPV with p16 methylation in ASC-US, LSIL, HSIL, and SCC was 31.8, 57.1, 82.9, and 96.3%, respectively (Table 3). The combined positivity of p16 expression, p16 methylation, and HPV in prediction of ASC-US, LSIL, HSIL, and SCC were 62.9, 50.0, 71.4, and 88.1%, respectively. However, combination of p16 Expression + P16 methylation + HPV + DipG1 in detection of LSIL, HSIL, and SCC were 50.0, 60.8, and 72.2% respectively (Table 3).

Table 3 Combined positivity of p16 Expression, p16 methylation, HPV, and Diploid G1 in cervical precancer and cancer
Full size table

Discussion

Introduction of liquid-based cytology reduced the number of unsuitable sample for cytological diagnosis; however, there are number of number of disadvantages associated with this cytology-based screening. Study of Sørbye et al. reported an increased number of inadequate cervical cytology for laboratories switching from conventional cytology to ThinPrep. However, there was reduction of inadequate cervical cytology samples in laboratories switching from conventional cytology to SurePath [15]. Replacement of ThinPrep with SurePath also showed higher reduction in inadequate cervical cytology. In recent times, a study on interobserver reproducibility in the histologic diagnosis of CIN reported that significant discrepancies occur in the diagnostic interpretation of low-grade lesions [16]. Therefore, both cytologic and histologic diagnoses are limited with regard to an exact diagnosis in low-grade CIN lesions, and a new biomarker for dysplastic or abnormal cervical cells is required to aid in the exact diagnosis and to evaluate a patient’s prognosis.

The conventional Pap test has played a significant role in screening of cervical carcinoma; however, its sensitivity and specificity are limited. For high-grade lesions, conventional cytology relatively has low sensitivity (50–70%) and performance of assay varies across population [17]. In the conventional cytology, due to detection and sampling failures in the conventional cytology, high-grade cervical lesions are often missed [18]. However, frequent testing is necessary to compensate for the low sensitivity of a single cytology test, as specificity is higher (95%), a considerable number of women with insignificant abnormalities, and who do not harbour any underlying high-grade lesion, have needless follow-up procedures [19]. In a pooled analysis showing the performance of conventional Pap smear for HSIL+ had a sensitivity of 55.2% (95% CI 45.5–64.7) for histologically conformed CIN2+ [20]. In the event of any abnormality, Pap smears detect 88.2% (95% CI 80.2–93.2) of all CIN2+ [20]. Specificity of Pap smears is between 71 and 97% depending on cutoff (ASCUS, LSIL, or HSIL). In a study of Sulik et al., they demonstrated higher sensitivity (90%; 95% CI 77–96%) of LBC than the conventional cytology (79%; 95% CI 59–91%) for CIN 2 or more severe lesions [21]. Study by Bernstein et al. reported the LBC as good as or superior to the conventional cytology for diagnosing CIN [22]. While various studies have reported LBC as good as or superior to the conventional cytology for diagnosing CIN, the Pap test false positive rate for premalignant and malignant lesions lies at an approximate 30% and false-negative rate lies between 6 and 55% [23, 24]. However, the impact of replacing conventional cytology by LBC as primary test method depends on the type of LBC test used [25]. In Netherlands, the 72 month cumulative cancer frequency for ThinPrep and SurePath was 66.8 and 44.6, respectively, as compared to 58.5 per 100,000 normal conventional cytology samples. For SurePath, hazard of invasive cancer was lowered to 19% and 15% for ThinPrep when compared with the conventional cytology [26]. In Denmark, introduction of SurePath significantly increased the detection of CIN3+ by 85% compared with conventional cytology for women between age group of 23–29 years; however, the increase of 11% with ThinPrep cytology was not statistically significant [27]. For the women between age group of 30–44 years, the increase with SurePath was 58%, while increase of 16% with ThinPrep was not significant. It is also very well known that HPV undeniably plays a role in the development of most cervical cancers [28, 29]. The major advantage of using LBC is the possibility to carry out ancillary techniques including HPV detection, molecular tests, methylation studies, and DNA ploidy as done in our study.

Aberrant promoter hypermethylation of tumor suppressor genes has been shown to be involved in human neoplasia [30]. The genomic organization at 9p21 houses two members of the INK4 family of cyclin-dependent kinase inhibitors (CDKIs), p15 and p16, and an unrelated gene p14ARF. Notably, p14ARF utilizes two of the same exons as p16, but is translated in an alternative reading frame. P16 is the most commonly altered gene in human malignancies [31]. The p16 promoter has recently been shown to be methylated in several human malignancies.

Promoter hypermethylation analysis of p16 found to be of importance in detection of premalignant lesion with positivity rate of 36.4, 76.2, 92.7, and 92.6% in women with ASC-US, LSIL, HSIL, and SCC cytology, respectively. p16 (p16INK4A) gene plays an important role in cell cycle regulation. Promoter hypermethylation of p16 gene has been reported earlier at a mean of 31% of cervical malignant lesions and also in lung carcinoma cases [32, 33]. However, study of Virmani et al. reported the methylation of p16 in 24% of high-grade lesions, 42% of invasive lesions, and was present in only 3% of nondysplasia/low-grade lesions [30].

Our findings showed that the expression of p16 increased from low-grade to high-grade cytology and showed a test positivity rate of 22.7% in women with ASC-US to 85.3% for women with SCC cytology. However, HPV was present in 45.4% cases of ASC-US to 92.5% cases of SCC cytology. P16 expression with HPV infection suggests that the detected low-grade lesion is at increased risk of progression to cancer due to possible genomic incorporation of oncogenic HPV. Study of Klaes et al. demonstrated that p16 immunostaining allows for a precise identification of small CIN or cervical cancer lesions [16]. Comparison of the positive rate of HPV test versus positive rate of p16 staining in our study in cases with ASC-US (45.4% vs. 22.7%) and LSIL (76.1% vs. 38.9%) with that reported in most studies was lower in our study. HPV and p16 expressions in ASC-US and LSIL cases were concordant with the finding of Yuan-ying et al. [34].

However, expression of p16 is believed to be superior to HPV DNA detection in diagnosis of cervical neoplasia; however, reported results lack consistency in triage of ASC-US/LSIL. Few studies have reported the p16 expression more sensitive and specific as compared to HPV detection in malignant lesions [35, 36]. Most studies show that p16 have lower sensitivity, but higher specificity for CIN2+ than HPV DNA testing [37, 38].

We found that detection of HPV increases the sensitivity as compared to p16 immunostaining and p16 methylation; however, the specificity was lowered (60.8 vs. 65.2 vs. 81.1%) (Table 4). Combination of p16 expression + p16 methylation + HPV increases the sensitivity to 100.0%; however, the specificity was lowered. However, combination of p16 expression + p16 methylation showed improved sensitivity as compared to p16 methylation + HPV and p16 expression + p16 methylation + HPV (55.0 vs. 52.1 and 47.8% (Table 5). The analysis of our results suggests that the use of HPV DNA detection and p16 hypermethylation studies serves as a better adjunct marker to cytology in detection of precancerous lesions.

Table 4 Comparison of methods in high-grade cytology (HSIL and SCC) and low-grade cytology (Normal, ASC-US, and LSIL)
Full size table
Table 5 Comparison of combination of methods in high-grade cytology (HSIL and SCC) and low-grade cytology (Normal, ASC-US, and LSIL)
Full size table

DNA ploidy has been effectively performed in diagnosis and prognosis of cervical, ovarian, and endometrial cancer worldwide. Image cytometry is the most widely used technique; others being laser scanning cytometry and flow cytometry. The analysis of cell cycle profile on the same liquid-based cytology sample, with abnormal result (aneuploidy or high S phase fraction) provides powerful complement to select women with developing lesions. We have found significantly higher S phase value in HSIL and SCC cases as compared to controls. Furthermore, we have found that Dip G1 and DipS phase values significantly differentiate LSIL, HSIL, and SCC cases from control with positivity rate of 50.0, 70.0, and 71.4, respectively. Our findings reinforce the hypothesis that higher diploid S value and lower diploid G1 may be associated with progression of cervical carcinoma.

In liquid cytology, study of Saxena et al. reported sensitivity and specificity of 96.77 and 100%, respectively, for diploid G0/G1 values in discrimination of cases from controls; however, sensitivity was 100% for total S phase and aneuploidy [39]. In contrast to this finding, Singh et al. reported aneuploidy in 49.3% (39/79) of mild, 77.7% (28/36) of moderate, and 91.6 (11/12) of severe dysplasia patients. In ASCUS and control groups, aneuploidy was present in 14.0% (8/57) and 8.6% (6/69) patients respectively [9]. As compared to patients with higher S phase value, lower S phase value favours a better survival [40]. A study of Lai et al. suggested that diploidy is a feature of better survival as compared to aneuploidy [41].

Conclusion

Beyond p16 expression and HPV testing, our findings of p16 methylation DNA ploidy analysis need to be extended to larger study sample. Findings of our study suggest that p16 promoter hypermethylation and DNA ploidy analysis could be of importance along with p16 expression and HPV detection in cervical precancer and cancer. These ancillary studies can be performed routinely in samples collected for LBC in conjunction with cyto-pathological examination where the gold-standard is histologically confirmed high-grade cervical lesions (CIN2+).