CDKN3 mRNA as a Biomarker for Survival and Therapeutic Target in Cervical Cancer

Abstract

The cyclin-dependent kinase inhibitor 3 (CDKN3) gene, involved in mitosis, is upregulated in cervical cancer (CC). We investigated CDKN3 mRNA as a survival biomarker and potential therapeutic target for CC. CDKN3 mRNA was measured in 134 CC and 25 controls by quantitative PCR. A 5-year survival study was conducted in 121 of these CC patients. Furthermore, CDKN3-specific siRNAs were used to investigate whether CDKN3 is involved in proliferation, migration, and invasion in CC-derived cell lines (SiHa, CaSki, HeLa). CDKN3 mRNA was on average 6.4-fold higher in tumors than in controls (p = 8 x 10-6, Mann-Whitney). A total of 68.2% of CC patients over expressing CDKN3 gene (fold change ≥ 17) died within two years of diagnosis, independent of the clinical stage and HPV type (Hazard Ratio = 5.0, 95% CI: 2.5-10, p = 3.3 x 10-6, Cox proportional-hazards regression). In contrast, only 19.2% of the patients with lower CDKN3 expression died in the same period. In vitro inactivation of CDKN3 decreased cell proliferation on average 67%, although it had no effect on cell migration and invasion. CDKN3 mRNA may be a good survival biomarker and potential therapeutic target in CC.

Fig 1. CDKN3 gene expression in normal cervical epithelium and cervical cancer samples.

The expression of cyclin-dependent kinase inhibitor 3 (CDKN3) was measured by RT-qPCR in 25 normal healthy cervix samples and 134 cervical cancer (CC) samples positive for human papillomavirus (HPV)16 (n = 90) or for other HPV types (n = 44), including 18, 31, 33, 35, 45, 51, 52, 53, 58, 59, and 68. (A) Intensity of gene expression, expressed in Log2 values, in box plots. The upper and lower boundaries of the boxes represent the 75th and 25th percentiles, respectively. The black and dotted lines within the boxes represent the median and mean values, respectively, and the whiskers represent the minimum and maximum values that lie within 1.5x the interquartile range from the end of the box. Values outside this range are represented by black circles. The fold change (FC) was calculated by dividing the median of each CC group by the median of the control group. Statistical differences between groups were calculated using the nonparametric Mann-Whitney U test. (B) Frequency (%) distribution of CDKN3 FCs, which were calculated in each tumor considering the median of the control group.

Fig 2. Survival analysis of women with cervical cancer according to CDKN3 expression.

The Kaplan-Meier curves for cyclin-dependent kinase inhibitor 3 (CDKN3) expression are shown. Patients were followed up for at least 60 months. Overall survival analyzed with Kaplan-Meier curves is shown for CDKN3 expression status, (A) all CC sample set, (B) HPV16-positive CC cases, (C) CC positive for HPVs other than HPV16. In all panels, the p-value was calculated by comparing the curves with the log-rank test. The number of patients at risk in each time intervals are noted in the tables below the curves. Censored patients are labeled with vertical bars (see material and methods).

Fig 3. Detection of CDKN3 protein by immunofluorescence and Western blot in cell lines derived from CC transfected with specific siRNAs against CDKN3 or scrambled siRNAs.

Cell lines derived from cervical cancer (CC) positive for human papilloma virus (HPV) 16 (CaSki, SiHa) and HPV18 (HeLa) were transfected with specific cyclin-dependent kinase inhibitor 3 (CDKN3) or scrambled siRNAs. Cells were harvested at 96 h after transfection and stained for CDKN3 protein. (A) Immunofluorescence staining for CDKN3 protein in SiHa, CaSki and HeLa cell lines using an anti-CDKN3 primary antibody and FITC-conjugated secondary antibody. Images were photographed at 60× magnification using a fluorescence microscope Olympus FV 1000. (B) Quantification of fluorescence intensity of anti-CDKN3 antibody-stained cells. The green fluorescence intensity was quantified using ImageJ software. Values represent the mean ± S.D. of 140 fields measured in each experiment. The statistical significance between the differences was calculated using the t test. (C) Expression of CDKN3 protein examined using western blot in SiHa, CaSki, and HeLa cell lines transfected with random siRNAs (-) and with specific CDKN3 siRNAs (+) with actin as internal control.

Fig 4. Cell proliferation curves of CC-derived cell lines transfected with specific siRNAs against CDKN3 or scrambled siRNAs.

Cell lines were transfected with specific cyclin-dependent kinase inhibitor 3 (CDKN3) (red diamonds) or scrambled siRNAs (blue diamonds). Cells were harvested at 24, 48, 72, and 96 h after transfection. Cell proliferation was measured using the MTT method, and the plates were read with a spectrophotometer at 560 nm. The experiments were repeated two times by triplicate, and the figure shows the data of one experiment. Mean ± standard deviation of triplicates, growth trend line, and growth model equations are shown for CaSki (A), HeLa (B) and SiHa (C). Cell growth follows an exponential trend predicted by the equation y_t = y_{0 *} e^rx, where y_t = absorbance at specific time, x = incubation time in hours, y₀ = a constant absorbance at time = 0, and r = rate growth. In panel D, we calculated the percentage of cell-proliferation inhibition of CaSki, HeLa and SiHa cell lines at 24, 48, 72, and 96 h post-transfection. At each time point, the mean absorbance of triplicates of cells transfected with the scrambled siRNAs was set as 100%.

Fig 5. Cell migration and invasion of cervical cancer (CC)-derived cell lines transfected with specific siRNAs against cyclin-dependent kinase inhibitor 3 (CDKN3) or scrambled siRNAs.

Cell lines were transfected with specific CDKN3 (black bars) or scrambled siRNAs (white bars). Cells were harvested 48 h after transfection. Cell migration (A) and invasion (B) was assayed in a 24-well transwell polycarbonate membrane inserts and in a 24-well transwell matrigel invasion chamber, respectively. After incubation time, cells that migrated through the filters or invaded through the matrigel membranes were collected and stained with MTT colorimetric assay. The experiments were repeated two times by triplicate, and the figure shows the mean ± S.D. of one experiment.

Fig 6. Analysis of CDKN3 transcripts.

Figure shows gel electrophoresis of mRNA variants of CDKN3 gene identified in CC, controls, and cell lines using reverse transcription-polymerase chain reaction (RT-PCR). Panel A and B show wild-type (wt) CDKN3 mRNA transcript. (A) RT-PCR was performed with external primers F1/R1 (756 bp, black arrow). (B) A 2-μL sample of 1:5 dilution of reactions of (A) were re-amplified using the nested primers F2/R5. In some samples, other weak transcripts were observed below wtCDKN3 transcript (721 bp) including variants cx1, cx2, cx3, cx4, and cx5 (200, < 200, 400, 450, and 500 bp, respectively). (C) RT-PCR identified the “f” variant (453 bp) using primers F1/R9f. Most samples showed an additional lower band (cx6 variant, 370 bp). (D) RT-PCR identified “i” variant (633 bp) using primers F6i/R5. (E) RT-PCR identified “k” variant (392 bp) using primers F4/R5. Lower panel shows RT-PCR amplification of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene as internal control. Cell lines were transfected with random siRNAs (-) or specific CDKN3 siRNAs (+), treated, and then incubated as described in Material and methods. RT-PCR was performed on RNAs extracted 48 h after transfections.

Fig 7. Sequencing analysis of CDKN3 transcripts.

All CDKN3 transcripts found in analyzed samples were sequenced. (A) Graphical representation of sequencing findings in transcripts detected in samples including exons, introns, alternative splicings, and stop codons. Variant “a” represents wild-type (wt) CDKN3 mRNA and variants f, i, and k are transcripts previously reported that were also detected in our samples. Graphics of newly detected transcripts not previously reported are also shown (cx1, cx2, cx3, cx4, cx5, and cx6) based on sequences, which are partially shown in (B). Gray light boxes represent the 3ʹUTR, 5ʹUTR, and nontranslated sequences. Coding exons are represented by light blue boxes and introns are depicted by continuous black lines. Alternative splicings are represented by black dashed lines. Start codons (AUG) are shown at the beginning of every transcript. Normal stop codons (UAA) and stop codons generated by alternative splicing (*) are shown. Pink shaded rectangle in the 5ʹʹ region depicts the coding region of first 34 amino acids necessarily for CDK2 interaction. Shaded green rectangle in the 3ʹ region represents carboxyl terminal region of CDKN3, which participates in stabilization of CDK2 interaction. Green, blue, and red bars above transcripts in (A) indicate recognition sites of TaqMan probes, anti-CDKN3 antibody, and siRNA-CDKN3, respectively.

Fig 8. Histological analysis of CDKN3 protein expression.

CDKN3 protein expression was determined by immunohistochemistry using formalin-fixed paraffin-embedded tissue sections. Representative experiments in HPV16-positive squamous cell carcinomas (A–D) and adenocarcinomas (E and F), and squamous cell carcinomas positive for other HPVs (G–I) are shown. Specific signals are shown as brown staining (counterstained with hematoxylin, original magnification ×800; bars, 20 μm). High-grade squamous intraepithelial lesion (J), normal cervical epithelium (K and L), and experimental controls without primary antibody (M–O) are also shown.