Ezrin and E-cadherin expression profile in cervical cytology: a prognostic marker for tumor progression in cervical cancer

One hundred and twenty-five Pap smears were obtained from women who utilized the CC screening service at the Facultad de Ciencias Químico-Biológicas of Universidad Autónoma de Guerrero (UAGro) and the Instituto Estatal de Cancerología of Guerrero state, Mexico. All participants signed an informed consent and a survey was carried out that included sociodemographic and gynecological information. Cervical smears were obtained from transformation zone and exfoliated cells were collected in: 1) DNA extraction solution to perform HPV genotyping; 2) liquid-based cytology to determine the physical status of HR-HPV and proteins expression and; 3) silanized glass slide for cytomorphological examination using the Papanicolaou technique. The diagnosis was done by a certified pathologist or cytopathologist. The Bethesda System was used to define the grade of squamous intraepithelial lesion or cervical cancer. Ethics approval to conduct this study was obtained from the Institutional Ethics Committee at the Universidad Autónoma de Guerrero and the Instituto Estatal de Cancerología.

DNA extraction was performed by SDS-Proteinase K-phenol-chloroform method. Viral genotyping was performed using the INNO-LiPA® Genotyping Extra kit (Innogenetics) according to manufacturer’s instructions. Briefly, the L1 region of HPV was PCR amplified with the SPF10 primers, the biotinylated amplicons were denatured and hybridized with specific and immobile oligonucleotides anchored to a membrane, streptavidin conjugated with alkaline phosphatase was added followed by the chromogen BCIP/NBT to reveal the reaction. The HLA-DPB1 gene was used as a control for DNA amplification, and L1 region of HPV 6 was used as a positive control.

The physical status of High-risk HPV (HR-HPV) was determined by in situ Hybridization with a tyramide signal amplification system (GenPonint Dako Cytomation, Carpinteria, CA, US). Cervical smears previously placed in monolayer on silanized slides were fixed with acetone and digested with proteinase K (1: 1000), the two probes of biotilinated viral DNA were added which recognized 13 genotypes of HR-HPV (16, 18, 31, 33, 39, 45, 51, 52, 56, 58, 59 and 68) and 2 LR-HPV genotypes (6 and 11). The slides were subjected to DNA denaturation (10 min at 95 °C) and hybridization for 20 h at 37 °C (Dako Hybridizer, Carpinteria, CA, US). The slides were then placed in an astringent solution, the primary streptavidine peroxidase was added and subsequently biotinyl-tyramide, then the secondary streptavidin and finally the DAB chromogen; cells were counterstained with Mayer’s Hematoxylin. The positive signal for In situ Hybridization was visualized as a brown deposit. To determine HPV physical status, a diffuse signal indicated episomal status, a punctiform signal integrated status and a diffuse-punctiform signal a mixed status (DakoCytomation Protocol).

Expression of Ezrin and E-cadherin proteins was determined by immunocytochemistry using the streptavidin-biotin peroxidase technique (Bio SB, Mouse/Rabbit ImmunoDetector HRP). For liquid-based citology samples, smears were prepared on silanized glass slides and fixed with 96% and 70% ethanol. For HaCaT (non-tumoral), HeLa (cervical cancer, HPV-18 positive) or SiHa (cervical cancer, HPV-16 positive) cell lines, 5 × 10⁴ cells were plated on glass coverslips in 6-well culture plates and allowed to grow for 24 h, in DMEM medium (Invitrogen, Carlsbad, CA,) supplemented with 10% FBS (Byproductos, Mexico). Cells were fixed with 4% formaldehyde; endogenous peroxidase was blocked with 3% hydrogen peroxide for 20 min and non-specific protein binding was performed with 1% BSA in PBS for 40 min. The slides were incubated with primary anti-Ezrin antibody (1:100 dilution; clone 3C12, Santa Cruz Biotechnology ScC-58,758) or anti-E-cadherin antibody (1:50 dilution; Santa Cruz Biotechnology SC-7870) in a humidity chamber during 1 h at room temperature. Biotin-bound secondary antibody was added for 30 min at room temperature followed by incubation with diaminobenzidine (DAB) for 1-2 min. Cells were counterstained with Mayer’s Hematoxylin for 10 min and the samples were dehydrated in descending degrees of ethanol and finally mounted with fast mounting medium Entellan (MERCK) and observed under a brightfield microscope (Olympus BX-43). As a negative control, we used samples processed under the same conditions as those already described, but in the absence of primary antibodies. Protein expression was visualized as brown signal and the intensity of staining was scored as low, moderate or high. To analyze the nuclear staining index, the number of cells with positive nuclei to Ezrin was divided among the number total cells present in the slide and multiplied by 100; this nuclear staining index was divided into tertiles (< 50%, 50-89% and > 90%). All the preparations were independently evaluated by three analysts, without knowing the cytological diagnosis, to establish a consensus about protein expression for each sample.

5 × 10⁴ cells (HaCaT, HeLa or SiHa) were plated on glass coverslips in 6-well culture plates and allowed to grow for 24 h in DMEM HamF/12 medium (Invitrogen, Carlsbad, CA,) supplemented with 10% FBS (Byproductos, Mexico) at 37 °C in a 5% CO₂ atmosphere. Cells were fixed with ice-cold methanol (− 20 °C) for 5 min, and permeabilized with PBS + 0.2% Triton X-100. The non-specific bindings were blocked with PBS + 1% BSA for 40 min and the samples were incubated with primary anti-Ezrin antibody (1:100 dilution; clone 3C12, Santa Cruz Biotechnology ScC-58,758) in a humidified chamber for 1 h at room temperature. Subsequently, FITC-conjugated secondary anti-mouse antibody (1:50 dilution) was added for 30 min in a humidified/dark chamber at room temperature and samples were mounted using ProLong Gold Antifade Mountant with DAPI (Invitrogen, P-36931). The images were acquired and processed on an EVOs FL Auto microscope (Thermo Fisher Scientific).

HaCaT, HeLa or SiHa cells were seeded on 100 mm plates and allowed to grow until 80% confluency. Cells were washed with PBS and lysed with RIPA buffer (50 mM Tris-HCl pH 7.6, 160 mM NaCl, 0.5 mM EDTA/EGTA, 1% Triton X-100, 10% glycerol, 1 mM PMSF and 1 μg/ml leupeptin). Protein extracts were separated by SDS-PAGE in 8% acrylamide gels, transferred to PVDF membranes and incubated overnight at 4 °C with antibodies anti-Ezrin (1:1000 dilution; clone 3C12, Santa Cruz Biotechnology) or anti-Tubulin (Merck Millipore) as loading control, overnight at 4 °C. The membrane was incubated with a secondary anti-mouse antibody (1:3000 dilution, Merck Millipore) at room temperature and revealed with chemiluminescent substrate (Luminata, Millipore).

Data ware expressed as absolute and relative frequencies. The association between variables was calculated using chi-square test (χ2). A logistic regression analysis was performed to calculate Odds Ratios (OR) and confidence intervals (CI) at 95% comparing the Ezrin and E-cadherin expression in non-SIL and LSIL group versus HSIL and CC group. To evaluate differences in Ezrin expression level between HaCaT and HeLa or SiHa cells, t-student test was used. P value < 0.05 was considered statistically significant.

Regardless of the diagnosis, Ezrin and E-cadherin expression was higher in basal and parabasal cells, compared to intermediate and superficial cells, which showed low expression of both proteins. Ezrin expression was low in non-SIL samples and high in HSIL or CC samples. In contrast, E-cadherin staining showed moderate/high intensity in non-SIL and LSIL, and low intensity in HSIL and CC (Fig. 1).

Ezrin and E-cadherin expression was significantly associated to the degree of SIL (< 0.001). 70% of cervical cancer samples show high expression of Ezrin, while 65% of non-SIL HPV (−) samples show low/negative expression of Ezrin. In the presence of HPV infection, even in absence of SIL the expression of Ezrin changes from low to moderate (58% of cases). In contrast, 95% and 75% of cervical cancer and HSIL samples, respectively, showed low or negative E-cadherin expression, whereas moderate/high expression of E-cadherin was observed in LSIL or non-SIL samples (Table 2). Moreover, we found that Ezrin and E-cadherin expression are correlated with p16ink4a and Ki-67 expression, which are validated markers at the detection of cervical lesions; in the cases that showed high expression of p16ink4a and Ki-67, the Ezrin expression was majorly high and E-cadherin expression was predominantly low (Additional file 1: Table S1). On the other hand, Ezrin and E-cadherin expression was significantly correlated to HPV infection (p < 0.001); high expression of Ezrin and low/negative expression of E-cadherin was observed in samples with multiple infections: HR-HPV and LR-HPV or several HR-HPV (Table 2). The changes in Ezrin and E-cadherin expression were related to presence of HPV-16 mainly; 95% of samples positive for HPV-16 showed high Ezrin expression and 74% showed low E-cadherin expression, and no differences were observed in expression of these proteins in relation to HPV-18 or other HPV-AR genotypes (Additional file 2: Table S2). A statistically significant correlation was observed between the HR-HPV integration and Ezrin expression; furthermore, Ezrin expression was high in 47% and 29% of samples with integrated or mixed HR-HPV genome, respectively. In contrast, E-cadherin expression was independent of physical status of HR-HPV (Table 2). We observed cases of HR-HPV integration in No SIL and LSIL group that were correlated with an increase in Ezrin expression (Additional file 3: Table S3).

Interestingly, nuclear staining for Ezrin was observed more frequently in HSIL and CC samples (Fig. 2). A statistical relation between nuclear Ezrin and cytological diagnosis was found (p < 0.001). 85% and 60% of CC and HSIL samples, respectively, showed > 90% of cells with nuclear Ezrin, whereas only 30-33% of non-SIL and 20% of LSIL samples were grouped in this category (Table 3). Moreover, Ezrin nuclear localization showed a statistically significant correlation with the HPV genotype (p = 0.001); 49% of samples positive for HR-HPV infection showed > 90% of cells with nuclear Ezrin, while 60% of the LR-HPV samples showed < 50% of cells with Ezrin-positive nuclei. No association was found between HR-HPV physical status and nuclear Ezrin (Table 3).

To evaluate the risk conferred by high Ezrin and low E-cadherin expression for development of HSIL and CC, we calculated ORs using a bivariate analysis comparing non-SIL/SIL group versus HSIL/CC group. We found that high Ezrin and low E-cadherin expression implicate 4.61 and 6.14 times more risk, respectively, for developing HSIL or CC (Table 4), suggesting that combined determination of Ezrin and E-cadherin expression could be useful for assessing the prognosis of patients. Another interesting fact that could be considered as a prognostic marker is the high percentage of nuclear staining to Ezrin, which implicates 3.68 times more risk of developing a HSIL or CC (Table 4).

In cervical cancer cells in vitro, Ezrin expression was higher in HeLa and SiHa cells compared to non-tumor cells HaCaT (Fig. 3a). Moreover, in HaCaT cells Ezrin immunoreactivity was exclusively cytoplasmic, whereas in HeLa and SiHa cells a perinuclear staining of Ezrin was observed, mainly in those cells where Ezrin expression was higher (Fig. 3b).