Eight cervical cancer cell lines (HeLa229, MS751, ME180, SiHa, HeLa, CasKi, C33A and HCC94) were purchased from the American Type Culture Collection (ATCC; Rockville, MD, USA) and maintained in the Department of Experimental Research, Sun Yat-sen University Cancer Center. They were grown in RPMI-1640 medium (Gibco-BRL, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS; HyClone Laboratories, Logan, UT, USA) and 1% antibiotics (100 U/ml penicillin and 100 µg/ml streptomycin).

Written informed consent was obtained from each of the cervical cancer patients and the study was approved by the research ethics committee of the Cancer Center of Sun Yat-sen University for the use of these clinical materials for study purposes. The Institutional Review Board (IRB) at the Sun Yat-sen University approved all experimental methods. Included in the present study were tumor specimens from 189 consecutive patients undergoing surgical treatment for cervical cancer at the Department of Sun Yat-sen University Cancer Center from 2006 to 2010. The tumor staging and clinicopathological classification of the tumors were accessed in accordance with the International Federation of Gynecologists and Obstetricians (FIGO, 2009) criteria. The classification of the histological types was based on the World Health Organization (WHO) recommendations.

The 189 patients were in Ib1-IIa2 stage cervical cancer and all received a radical hysterectomy and lymphadenectomy. Patients had not undergone anticancer treatment, such as immunotherapy, chemotherapy, hormone therapy or radiotherapy, before specimen collection. Patients who received postoperative chemotherapy and/or radiotherapy were those with high-risk factors, including: pelvic lymph node metastasis (PLNM), high differentiation grade, positive parametrial involvement, positive lymphovascular space involvement, positive surgical margin, deep stromal invasion and large tumor size (>4 cm). Patients with only deep stromal invasion or positive surgical margins received postoperative radiotherapy, while patients with only lymphovascular space involvement, high differentiation grade, or large tumor size (>4 cm) received postoperative chemotherapy. No patients received immunotherapy after surgery.

In total, 189 early-stage cervical cancer patients were included (aged 23–68 years, with a median age of 46 years). The media age is used to determine the age gap of patients. Thus, we separated the groups by a 46-year age gap in the present study. Demographic and clinicopathological information were collected from impatient medical records (Table I). In addition, six pairs of cervical cancer tissues and the matched tumor-adjacent morphologically normal tissues, which were collected from early-stage cervical cancer patients who underwent surgery in 2015, were frozen in liquid nitrogen at −80°C until further use.

To explore the role of TIMELESS in the development of cervical cancer, we evaluated its mRNA expression pattern in cervical cancer cell lines and tissues. RNA extraction from the cultured cell lines and surgically dissected tissue samples was carried out using the TRIzol reagent (Invitrogen) following the manufacturer's guidelines. Agilent Bioanalyzer 2100 was used to evaluate the RNA quality (RIN, 2.4–8.8; median 5.9). For PCR-mediated synthesis and amplification of TIMELESS cDNA, an initial amplification reaction using TIMELESS-specific primers was performed with a denaturation step at 95°C for 10 min, followed by 30 denaturation cycles at 95°C for 60 sec, primer annealing at 55°C for 30 sec, and primer extension at 72°C for 30 sec. A final extension at 72°C for 5 min was carried out before the reaction was stopped and stored at 4°C on completion of the cycling steps. TIMELESS-specific primers were designed using the Primer Express v. 2.0 software (Applied Biosystems) as follows: forward 5′-CCATACATCAGTGG ACCAACC-3′ and reverse 5′-CTCCTCCGGGCTTCTGA-3′. To control the variability in expression levels, expression data of TIMELESS were normalized to the expression of the housekeeping gene, GADPH. Primers for GADPH were: forward 5′-TTGAGGTCAATGAAGGGGTC-3′ and reverse 5′-GAAGGTGAAGGTCGGAGTCA-3′. Quantitative RT-PCR (qRT-PCR) was performed in a total volume of 10 µl using the LightCycler 480 instrument (Roche Diagnostics, Penzberg, Germany) with the following conditions: 95°C for 5 min, followed by 45 cycles of 95°C for 30 sec, 55°C for 30 sec and 72°C for 15 sec, and a final extension step of 72°C for 5 min. The relative quantitative value was expressed by the 2^−ΔΔCt method, where Ct represents the threshold cycle for each transcript. All experiments were repeated three times.

We also evaluated TIMELESS protein expression in cervical cancer cell lines and tissues. The cervical tissues and cell lines were prepared and lysed using the cell total protein extraction kits, on ice, according to the manufacturer's instructions (Millipore, Billerica, MA, USA). Western blots were performed according to standard methods in our previous publication (32). Briefly, we separated equal amounts of protein samples (30 µg) on 9% SDS polyacrylamide gels and transferred them to PVDF membranes (Immobilon P; Millipore, Bedford, MA, USA). Anti-TIMELESS rabbit monoclonal (1:1,000; Abcam, Cambridge, MA, USA) and anti-rabbit (1:2,000; Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibodies were used to detect TIMELESS protein. After detection, the blotted membranes were stripped, and anti-GADPH was detected using a mouse monoclonal antibody (1:2,000; Sigma-Aldrich, St. Louis, MO, USA). The secondary antibody (anti-mouse antibody; Santa Cruz Biotechnology) was diluted 1:2,000 in both samples. TIMELESS signals were examined on X-ray film using an ECL prime Western blotting detection reagent (Amersham Biosciences).

Paraffin-embedded specimens were obtained from 189 early-stage cervical cancer patients for immunohistochemical analysis. The procedures of immunostaining were carried out with standard protocols. In detail, paraffin tissue slides (4 µm) were dried at 60°C for 1 h, deparaffinized for 10 min in xylene (twice), and rehydrated through a series of incubations in graded ethanol solutions (100, 100, 95, 90 and 80%) for 5 min, respectively. Endogenous peroxidase activity was blocked with 3% hydrogen peroxide in methanol for 15 min at room temperature to quench endogenous peroxidase activity. Following this, all deparaffinized sections were immersed and boiled in buffered ethylenediaminetetraacetic acid (pH 8.0) for 2 min in a pressure cooker to retrieve antigen, and then cooled to room temperature. Then, sections were immersed with 1% bovine serum albumin (BSA) to avoid non-specific binding. Subsequently, the slides were immunostained with a primary antibody against TIMELESS (Abcam) diluted at 1:250 at 4°C overnight in a moist chamber. Phosphate-buffered saline (PBS) was used as a negative control. After washing with PBS buffer, the slides were incubated with prediluted anti-rabbit secondary antibody (Abcam) and then incubated with streptavidin-horseradish-peroxidase complex (Abcam). The tissue sections were immersed in 3-amino-9-ethylcarbazole, counterstained with 10% Mayer's haematoxylin, dehydrated and mounted in Crystal Mount. For visualization, tissue slides were stained with DAB (3,3-diaminobenzidine) for 1 min at room temperature. Finally, the tissue specimens were counterstained with hematoxylin, dehydrated and mounted using neutral balsam (Sigma-Aldrich).

All immunostained tissue slides were assessed independently by two observers without prior knowledge of clinical data. To evaluate IHC expression of TIMELESS, we applied a scoring system based on multiplying the proportion of positively stained tumor cells and the intensity of staining. The staining intensity (SI) was graded according to the following criteria: 0, no staining; 1, weak staining; 2, modest staining; and 3, strong staining. The proportion of positive tumor cells was scored as follows: 0 (no positive tumor cells); 1 (<10% positive tumor cells); 2 (10–50% positive tumor cells); 3 (51–80% positive tumor cells), and 4 (>80% positive tumor cells). The slides were rescored if the difference between the two pathologists was >3. The optimal cut-off values were distinguished and categorized by the median of the immunoreactive score (IRS) results: an SI score of 6 was used to identify tumors with high TIMELESS expression and an SI score of 4 was used for low TIMELESS expression.

All calculations were carried out using the SPSS version 18.0 software package (SPSS, Inc., Chicago, IL, USA). The relationship of TIMELESS protein expression levels with clinical and histomorphological characteristics were determined using a Chi-squared test. The overall survival (OS) and disease-free survival (DFS) curves in association with TIMELESS expression were plotted using the Kaplan-Meier method and the difference between the groups was tested by a log-rank test. Multivariate analysis was conducted independently for each biomarker, including only significant clinical/histomorphological factors from the univariate analysis. Univariate and multivariate survival analyses were carried out using the Cox regression model. A statistically significant difference was determined by a two-sided P-value of <0.05.

Significantly higher TIMELESS protein expression and mRNA levels were observed in all eight cervical cancer cell lines compared to the normal cervical tissue sample (P<0.05; Fig. 1). TIMELESS protein expression was also significantly upregulated in the six cervical cancer cases compared with matched adjacent non-cancerous cervical tissues (P<0.05; Fig. 2A). Similarly, qRT-PCR data also indicated that TIMELESS mRNA was significantly upregulated in cervical cancer specimens compared to their matched adjacent non-cancerous regions (P<0.05; Fig. 2B). By real-time RT-PCR evaluation, the tumor/adjacent non-cancerous (T/N) ratio of TIMELESS mRNA expression was >5-fold in all cases, with the highest ratio being ~18-fold (Fig. 2B). In agreement with the results of the western blot analysis and real-time RT-PCR assay, IHC showed high expression of TIMELESS protein in cervical cancer lesions (Fig. 2C). Together, our results showed that TIMELESS expression was elevated at both the transcriptional level and translational levels in the cervical cancer cell lines and tissues.

We collected 189 paraffin-embedded early-stage cervical cancer tissue samples for IHC, including 83 cases of stage IB1, 28 cases of stage IB2, 55 cases of stage IIA1 and 23 cases of stage IIA2 tumors. Following IHC staining, TIMELESS expression was assessed as positive in 124/189 (65.6%) cases and weakly positive or negative in 65 cases (34.4%). TIMELESS expression was mainly observed in the nucleus and rarely in the cytoplasm of epithelial cells (Fig. 3A). Notably, TIMELESS expression was found in most cancerous lesions in the primary cervical tumors, but was reduced in the adjacent normal cervical regions. TIMELESS protein expression was generally negative in normal cervical tissues, weak in stage IB1 and IB2 cervical cancer tissues, and strong in stage IIA1 and IIA2 cervical cancer tissues; TIMELESS expression obviously increased with the progression of tumor grades (P<0.05; Fig. 3A). Mean optical density (MOD) represents the average response strength of all positive cells in the field of vision. Quantitative IHC data revealed that the MOD values of TIMELESS were significantly associated with the cervical cancer tumor grades (Fig. 3B), and were higher in cervical cancer tissues with PLNM than in those without PLNM (Fig. 3C).

Next, we examined the relationship between TIMELESS expression and established clinicopathological and investigative parameters using the Chi-squared test and Fisher's exact test (Table II). In patients with early-stage cervical cancer, high TIMELESS expression was markedly associated with age (P=0.011), clinical stage (P<0.001), PLNM (P<0.001), squamous cell carcinoma (SCC) antigen (P=0.003), tumor recurrence (P=0.015), vital status (P<0.001), differentiation grade (P<0.001), property of surgical margin (P=0.036) and lymphovascular space involvement (P=0.001; Table III). Moreover, patients with increased TIMELESS protein expression showed strong tendencies to receive postoperative radiotherapy (P=0.002). However, no significant correlations between TIMELESS expression and the histological type, HPV infection, tumor size, myometrium invasion, or infiltration of the parauterine organ were observed (P>0.05; Table III). Patients with aberrant TIMELESS protein expression also showed no obvious tendency to receive chemotherapy, concurrent chemotherapy and radiotherapy, chemotherapy with TP regimen (paclitaxel plus cisplatin), or radical hysterectomy (P>0.05; Table III). TP chemotherapy is the main chemotherapy of cervical cancer patients which combines paclitaxel with cisplatin.

Association coefficient analyses (Spearman's correlation) further confirmed the above-mentioned relationships between the TIMELESS protein expression levels and the clinicopathological features (Table III).

We examined whether patients with TIMELESS-positive samples had a poorer survival rate than those with TIMELESS-negative samples. Survival time was markedly different between patients with TIMELESS-positive and TIMELESS-negative samples based on the log-rank test. Patients with higher TIMELESS expression had shorter OS (log-rank, P<0.001; Fig. 4A) and DFS (log-rank, P=0.001; Fig. 4B) than those with lower TIMELESS expression. When examining the whole study cohort, patients with high TIMELESS expression had a significantly lower cumulative 5-year survival rate than those with low TIMELESS protein expression (73.1 vs. 96.8%, respectively; P<0.05).

We also explored the prognostic value of TIMELESS protein expression in different subgroups of patients with early-stage cervical cancer. Subgroups of patients were stratified according to age, HPV infection, FIGO stage, PLNM, SCC antigen, tumor size, differentiation grade, chemotherapy, radiotherapy, concurrent chemotherapy and radiotherapy, TP chemotherapy, radical hysterectomy, myometrium invasion, properties of the surgical margin, infiltration of parauterine organ and lymphovascular space involvement. TIMELESS protein expression was strongly associated with the OS duration of patients aged >46 years (log-rank test; P=0.008), aged ≤46 years (log-rank test; P<0.001), with HPV infection (log-rank test; P<0.001), without HPV infection (log-rank test; P=0.030), without radiotherapy (log-rank test; P<0.001), with SCC antigen ≤1.5 ng/ml (log-rank test; P=0.004), with clinical stage IB1-IB2 (log-rank test; P<0.001), with TP chemotherapy (log-rank test; P=0.005), with histological differentiation grade (grades 1 and 2, log-rank test, P<0.001; grades 2 and 3, log-rank test; P=0.011), with tumor size (tumor size <4 cm, log-rank test; P<0.001; tumor size ≥4 cm, log-rank test; P=0.033), with radical hysterectomy (log-rank test; P<0.001), without infiltration of parauterine organ (log-rank test; P<0.001), without PLNM (log-rank test; P<0.001), with SCC (log-rank test; P<0.001), without lymphovascular space involvement (log-rank test; P<0.001), without properties of the surgical margin (log-rank test; P<0.001), without myometrium invasion (log-rank test; P=0.001), with myometrium invasion (log-rank test; P=0.001) and with chemotherapy (log-rank test, P<0.001; Figs. 4 and 5).

Univariate analysis based on the Cox regression model demonstrated that upregulation of TIMELESS protein expression was a significant prognostic biomarker of poor OS (P<0.001; Table IV). In addition, early-stage cervical carcinoma patients with PLNM, with lymphovascular space involvement, and those who had undergone radiation therapy, showed poorer prognosis (Table IV). In our multivariate analysis, only TIMELESS protein expression and tumor recurrence remained significant (Table IV), and served as independent prognostic factors for poor OS in cervical cancer patients (P=0.002, 95% CI, 2.285–42.639; and P<0.001, 95% CI, 2.430–10.673, respectively).