Inactivation of the Kinase Domain of CDK10 Prevents Tumor Growth in a Preclinical Model of Colorectal Cancer, and Is Accompanied by Downregulation of Bcl-2

Colorectal cancer, which accounts for more than 600,000 deaths worldwide annually (1), originates from benign adenomatous polyps through cumulative genetic alterations in the normal colonic mucosa and adenomatous polyps (2). The multitude of changes that occur during the development of colorectal cancer has become even more apparent in the era of genome-wide high-throughput analysis (3–6). Using publicly available genome-wide expression data, a meta-analysis was previously undertaken to identify consistently reported genes that were differentially expressed in colon cancer or adenomatous polyps in relation to normal colonic mucosa. This meta-analysis, which took into account the study sample size, and average fold change for the most significant differentially expressed genes, generated a short-list of genes that were statistically and consistently upregulated or downregulated in colorectal cancer (7). It identified cyclin-dependent kinase 10 (CDK10) gene (7) to be consistently upregulated in colorectal cancer.

CDK10, previously referred to as PISSLRE, is a member of the CDC2 CDK family of protein kinases. The evolutionary expansion of the CDK family in mammals led to the division of CDKs into cell-cycle-related (CDK1, CDK4, and CDK5) and transcriptional (CDK7, CDK8, CDK9, CDK11, and CDK20) subfamilies (8). Most of the members of this family form heterodimers with their regulatory cyclin partners for their function. For example, active complexes of CDK4 or CDK6 with cyclin D phosphorylates members of the retinoblastoma (Rb) protein family to facilitate the transition of cells from G1 to S phase. CDK1 (CDC2) or CDK2 complexed with Cyclin A are required for S phase progression and transition to the G2 phase (9). CDKs are positively regulated by phosphorylation events catalyzed by CDK activating kinases and negatively regulated by two separate families of CDK inhibitory proteins, namely the INK4 family (comprising of p15, p16, p18, and p19) and Cip/Kip family (including p21 and p27; refs. 10, 11). In addition to their role in cell cycle regulation, other members such as CDK7 and 9 are important in the initiation and elongation of RNA polymerase II (12), whereas CDK6 has been shown to induce the expression of tumor suppressor p16INK4a and the pro-angiogenic factor VEGF-A (13). CDK10 has been reported to be involved in the regulation of G2/M phase transition of the cell cycle (14), and with cyclin M as its activating cyclin partner, phosphorylates CDK substrates histone H1 and ETS2 (15). Similar to other CDKs, it contains 3 common domains that are involved in binding to cyclins and its kinase activity: (i) the adenosine triphosphate binding sequence, (ii) the PSTAIRE domain, and (iii) the kinase sequence (16).

CDK10′s role in cancer, including colorectal cancer, has not been fully elucidated. Abnormal CDK10 expression has been reported in colorectal cancer (7), lung cancer (17), gastric cancer (18), and follicular lymphoma (19). In breast cancer, CDK10 has been associated with ETS2 (20), a transcription factor known for its involvement in this disease (21).

In this study, we show that the gene expression level of CDK10 in human colorectal adenocarcinoma tissues and cell lines is significantly higher than the normal colon. High expression of CDK10 in human colorectal cancer tissues also correlates with shorter median survival. We further show that overexpression of this protein in colorectal cancer cell lines promotes their growth in vivo and in vitro, whereas its inhibition suppresses growth of patient-derived xenografts in vivo. CDK10′s inhibitory effect on apoptosis appears to be, in part, as a result of an upregulation of Bcl-2 and, to a lesser extent, Bcl-x_L.

Human colorectal cancer cells MIP101, RKO (ATCC: CRL-2577), and HCT116 (ATCC: CCL-247) were maintained in DMEM (Invitrogen) supplemented with 10% newborn calf serum (NCS) and 1% penicillin–streptomycin and 1% kanamycin (Invitrogen). 293T packaging cells (ATCC # CRL-11268) were maintained in DMEM with 10% NCS only. Human normal colon FHC cells (ATCC: CRL-1831) were maintain in DMEM:F12 (50:50) supplemented with 10% NCS, 25 mmol/L HEPES, 10 ng/mL cholera toxin, 0.005 mg/mL insulin, 0.005 mg/mL transferring and 100 ng/mL hydrocortisone. All cells were cultured at 37°C in a humidified atmosphere and 5% CO₂. Cell lines have undergone authentication by the Centre for Translational and Applied Genomics (BC Provincial Health Services Authority) using AmpFLSTR Identifiler Plus PCR Amplification Kit (MIP101 authenticated in 2014; RKO, HCT116 in 2012).

MIP101, RKO and HCT116 cells were seeded in 24-well plates (50,000 cells/well) 24 hours before transfection with CDK10 siGENOMESMARTpool siRNA (small interfering RNA, 100 nmol/L, Dharmacon), Bcl-2 siRNA (80 nmol/L; Stealth RNAi, Invitrogen), Bcl-x_L siRNA (50 nmol/L, Flexitube siRNA; Qiagen) or scramble oligonucleotide sequence (control) using DharmaFECT 1 (Dharmacon) transfection reagent according to the manufacturer's instructions. Cells were harvested for RNA analysis 72 hours after transfection. On average, there was >70% knockdown in CDK10 expression, and >80% knockdown in Bcl-2 and Bcl-x_L expression by qRT-PCR and immunoblotting.

To establish RKO and MIP101 cells with stable overexpression of either wild type CDK10 (CDK10 WT) and a dominant negative mutant of CDK10 (CDK10 DN), cells were seeded in a 6-well plate at a density of 400,000 cells per well, 24 hours before transfection. Cells in each well were transfected with 4 μg of expression plasmids, pcDNA3-HA-CDK10-WT (wild type CDK10), or pcDNA3-HA-CDK10-DN (CDK10 with kinase defective CDK10; generously provided by Dr. A. Giordano, Sbarro Institute for Cancer Research and Molecular Medicine, Temple University) or empty vector pcDNA3.0 (Invitrogen) using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Stable cells were selected against geneticin (0.5 mg/mL; Invitrogen), and resistant colonies were screened by RT-PCR and immunoblotting for HA-CDK10 expression. All stable clones were maintained in DMEM supplemented with 10% NCS, 1% penicillin–streptomycin, 1% kanamycin and 0.5 mg/mL geneticin (Invitrogen).

Frozen tissue samples of colorectal adenocarcinoma and matched normal colon from 16 patients were obtained from British Columbia Cancer Agency Tumor Tissue Repository for validating gene expression of CDK10 by quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR), and/or conventional RT-PCR. Total RNA was isolated from cells or tissues using TRizol reagent (Invitrogen) according to the manufacturer's instruction. A total of 1 μg of RNA was treated with amplification grade DNaseI (Invitrogen) followed by first strand cDNA synthesis using Superscript III (Invitrogen) reverse transcriptase. The following primers were used for RT-PCR and/or qRT-PCT: β-actin: 5′-GCCACGGCTGCTTCCAG-3′ (sense) and 5′-GGCGTAC AGGTCTTTGC-3′ (antisense); CDK10: 5′-AGCTGAAGGAGGTGGTTGTG-3′ (sense) and 5′-CCTTCAGGTCCCTGTGGATA-3′ (antisense); Bcl-2: 5′-AAGATTGATGGGATCGTTGC-3′ (sense) and 5′-GCGGAACACTTGATTCTGGT-3′ (antisense); Bcl-xL: 5′-CCTCTCCCGACCTGTGATAC-3′ (sense) and 5′-CCAAAACACCTGCTCACTCA-3′ (antisense); PCR products were separated by 2% agarose gel electrophoresis followed by ethidium bromide staining, quantified by ImageJ (National Institute of Health) and normalized to β-actin. Real-time PCR was performed using EvaGreen qPCR Mastermix- ROX (ABM) in a ABI Prism 7900HT Sequence Detection System (Applied Biosystems). The cycle time (CT) values were normalized to CT values for actin and fold changes were calculated with respect to control cells. To determine the mRNA expression of the different CDKs, the cDNA synthesis, PCR reaction conditions and quantification of transcript levels were carried out as previously (22). The primer sequences used are as follows: CDK2: 5′-ACTGGCATTCCTCTTCCCCT-3′ (sense) and 5′-GGCGAGTCACCATCTCAGCA-3′ (antisense); CDK4: 5′-GCCCAGTGCAGTCGGTGGTA-3′ (sense) and 5′-GGTTAAAAGTCAGCATTTCCAGCAG-3′ (antisense); CDK5: 5′-CAAGCTGCCAGACTATAAGCCCTA-3′ (sense) and 5′-TGCAGCAGATCCCTCCCTGT-3′ (antisense); CDK6: 5′-CAGAGCCTGGAGTGCCCACTGA-3′ (sense) and 5′CGCGATGCACTACTCGGTGTGA-3′ (antisense); CDK9: 5′-ATCCACAGAAACAAGATCCTGCAT-3′ (sense) and 5′-CTTCAGGACCCCATCACGAGT-3′ (antisense); TBP: 5′-TGCACAGGAGCCAAGAGTGAA-3′ (sense) and 5′-CACATCACAGCTCCCCACCA-3′ (antisense). The C_t values were normalized to C_t values for TBP and fold changes were calculated with respect to control cells.

Caspase-3/7 assay was carried out by mixing 20 μg of total protein extracts prepared from cells as above with Caspase-Glo 3/7 substrates (Promega) as previously described (23). In brief, caspase-3/7 assay was carried out by mixing 20 μg of total protein extracts prepared from cells as above with Caspase-Glo 3/7 substrates (Promega). The relative luminescence units (RLU) were measured using a Synergy H4 Hybrid Multi-Mode Microplate Reader (BioTek Instruments, Inc.). The percentage of apoptosis based on caspase 3/7 activity was calculated relative to that of control samples. TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling) apoptosis assay was carried out using a DeadEnd Fluorometric TUNEL System (Promega) according to the instruction manual. Observations were made in a Leica DM600B microscope with Surveyor (version 5.5.5.12) software. Images were processed by using ImageJ software (NIH). The percentage of apoptosis was calculated by counting TUNEL positive (green) cells and 4′,6-diamidino-2-phenylindole (DAPI) stained (blue) cells in >8 images for each group of samples.

For transient transfection, cells were seeded at 5,000 cells per well in a 96-well plate for 24 hours followed by transfection with CDK10 siRNA or scramble control as described above. Seventy-two hours after transfection and/or incubation with 5-Fluorouracil (5-FU), 10 μL of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; 12 mmol/L) reagent was added to each well and incubated for an additional 4 hours at 37°C with 5% CO₂. Media were aspirated and the precipitate was solubilized in 50 μL of dimethyl sulfoxide. The absorbance of each well was measured at 490 nm and 650 nm using a 96-well plate reader (Versa Max, version 4.8, Molecular Devices Co., Sunnyvale, CA) according to the manufacturer's instructions. The percentage of viable cells was calculated relative to control wells.

For the determination of IC₅₀ for 5-FU, cells with stable overexpression of CDK10 WT, or control (empty vector pcDNA3.0) were seeded in a 96-well plate for 48 hours (RKO: 1,500 cells/well, MIP101: 3,000 cells/well) followed by incubation with increasing concentrations of 5-FU (1–800 μmol/L) for 48 hours. Then, MTT assay was performed as described above. Dose–response curves were calculated for each individual experiment via sigmoidal dose–response analysis using the Hill fitting equation in the Prism 4 software (GraphPad Software Inc.) followed by determination of the IC₅₀ value.

To assess the expression of various proteins, 40 μg of total protein extracted from different colorectal cancer cell lines, following transfection with CDK10 siRNAs or scramble control, was separated on a 10% SDS-PAGE gel and then transferred to polyvinylidene difluoride membranes (Millipore). Immunodetection was performed using primary antibodies against Aurora kinase A, Bcl-2, Bcl-xL, Bax, Bad, Caspase 8 (1:1,000, Cell Signaling Technology), CDK10 (1:500, Covalab) and HA-tag (1:1,000, Applied Biological Material Inc.) followed by incubation with horseradish peroxidase (HRP)–conjugated anti-rabbit or anti-mouse secondary antibodies. The immunoblots were also probed for β-actin (1:2,000, Abcam) as loading control. Proteins were detected with ECL detection reagents (Pierce).

Stably transfected RKO cell lines were seeded in 60-mm dishes at a density of 150,000 cells. After 3 days, cells were harvested by washing with PBS then trypsinizing. Detached cells were resuspended in culture medium to quench trypsin activity and the cell suspensions were spun down. Cell pellets were resuspended and washed in ice cold PBS. Cells were then fixed in cold 70% ethanol for 30 minutes and washed twice with PBS. Cell pellets were resuspended in propidium iodide/RNase A solution and subjected to flow cytometry analysis. Labeled cells were analyzed using a FACSCalibur with CellQuest software (Becton Dickinson) and FlowJo software (Tree Star Inc.).

Tissues were fixed in 10% neutral-buffered formalin, processed, embedded in paraffin. Five micron–thick tissue sections were immunostained with anti-Ki-67 (1:200, Abcam) or anti-cleaved caspase-3 (1:200, Cell Signaling Technology) and counterstained with Hoechst 33258 (Sigma-Aldrich). Detection was by Alexa 555 nm conjugated secondary antibodies (Invitrogen). Slides were subsequently mounted in SlowFade Antifade (Life Technologies) and analyzed by fluorescence microscopy (Leica DM6000 B). All images were processed by ImageJ software.

MIP/NeoB, MIP/CDK10 DN and MIP/CDK10 WT seeded at 1,000 cells per plate in 24-well plate. After 24 hours, the cells were treated with increasing concentrations of 5-fluorouracil (5-FU; 0, 1 and 5 μmol/L) for 10 days. Cells were then washed twice with PBS and stained with 0.2% crystal violet. Colony formation was quantified using the ImageJ free software.

Tumor xenografts of colorectal cancer cell: 3 × 10⁶ cells of MIP/CDK10 WT, MIP/CDK10 DN, and MIP/NeoB cells were resuspended in 100 μL 1:1 PBS/Matrigel (BD Biosciences), and injected subcutaneously into the right flank of 3- to 4-week-old female nude athymic mice (Simonsen Laboratories) using a 26-gauge needle. Tumor dimensions were measured every 2 days using a digital caliper. Volumes were calculated using the modified formula for an ellipsoid: 1/2(length × width²).

Fresh human colorectal cancer tissues were collected at the time of surgical resection from consented patients and implanted subcutaneously in C.B-17/SCID (severe combined immunodeficient) mice (Taconic). Patient-derived xenograft (PDX) tumors that grew following this initial implantation were then subsequently used for these experiments: each tumor was harvested, but into smaller pieces (∼ 2 × 2 × 2 mm) and reimplanted into C.B-17/SCID mice, grouped to receive the following intratumoral injections: scramble siRNA (Lenti-Sc, 0.01 e⁷ TU/mm³ of tumor) and CDK10 siRNA (Lenti-SiCDK10, 0.01 e⁷ TU/mm³ of tumor).

For packaging and production of high titer lentivirus harboring CDK10 and scramble siRNA, 293T packaging cells were seeded in 10-cm culture dishes at approximately 80%–90% confluency 24 hours before transfection, with a mixture of plasmids (pLenti-GFP-siRNA plasmids: 15 μg, pCMV-dR8.74: 15 μg, pVSV-G: 4 μg and pRSV-Rev: 4 μg) using Lipofectamine 2000 (Invitrogen). After 24 hours of transfection, the medium was replaced with fresh medium without antibiotics, and incubated at 37°C for further 48 hours. Viral supernatants were collected, centrifuged at 2,000 rpm for 2 minutes to remove dead cells, and filtered using a 0.45 μm filter. Lentiviral supernatants were then concentrated by using PEG8000 and tittered by FACS analysis of GFP expression as described previously (24, 25). Tittered lentiviral preparations were then stored at −80°C in appropriate aliquots. This study was approved by the Animal Care Committee (protocol A11-0398) at the University of British Columbia, Canada.

A tissue microarray (TMA) representing 103 samples of stage II–IV colorectal cancer was used to assess CDK10 expression in colorectal cancer, and whether its level of expression correlates with clinical outcomes (disease recurrence and survival). Patient characteristics: 94 patients with colon and rectal cancers distributed across stages II, III, and IV, diagnosed at the British Columbia Cancer Agency (Vancouver, Canada) from 2000 to 2008, consented to study participation and inclusion of their tissue in the TMA. The median age at diagnosis was 59.5 years (range, 51–67; Table 1), with 85.1% of patients under the age of 70 years. Women were represented in 42.6% and males in 54.9% of the TMA, with the majority of cases representing colon cancer (63.9%) versus rectal cancer (31.9%). Eighty-four percent of patients received chemotherapy: first-line therapy for colon cancers was irinotecan alone or in combination with 5-FU for 50% of patients, oxaliplatin and 5-FU–based therapy in 37%, and capecitabine alone among 7 (8%). The remaining 5% received second line treatment. Among the 31 patients with rectal cancer, 4 received no radiation to the primary tumor, 16 received preoperative radiation, and 4 postoperative radiation (7 patients were unknown). Local recurrence, distant metastasis, and death occurring during a 60-month follow-up period were documented for each patient. Ethics approval was obtained from the institutional review boards.

For each colorectal cancer specimen, two formalin-fixed, paraffin-embedded cores were taken from representative areas of primary tumors from each patient and mounted onto the TMA block. Four micron–thick sections were made from the TMA block and subsequently deparaffinized in xylene and rehydrated. Sections were heated in citrate buffer for 15 minutes in a cooker for antigen retrieval. Endogenous peroxidase activity was blocked using 0.3% H₂O₂ and washed with PBS for 10 minutes. Immunohistochemical staining with primary antibody against CDK10 (Abgent) was carried out using the Ultravision LP Detection Kit (Thermo Fisher Scientific). Sections were treated with Ultra V Block for 5 minutes to prevent nonspecific reaction with primary antibodies, then incubated at 4°C for 24 hours with primary antibodies, followed by incubation with a primary antibody enhancer for 10 minutes at room temperature. Subsequently, sections were treated with HRP polymer for 15 minutes and the reaction product was developed using 3,3-diaminobenzidine tetrahydrochloride (Zymed). The sections were counterstained with hematoxylin and mounted with Tissue-Tek Glas 6419 (Sakura Finetek). Negative controls consisted of omission of the primary antibodies. Staining expression scores were based on the number of tumor cells with positive staining in the cytoplasm, and were categorized as follow: 0 or none (expression <10%), 1+ or weak (10%–50%), 2+ or strong (50%–80%), and 3+ or intense (>80%). Scores were provided by two independent pathologists who were blinded to clinicopathological data. The two expression scores per sample were averaged, with the average representing the patient's final expression intensity. Low CDK10 expression represented a score = 0 or 1+, and high expression as a score = 2+ and 3+. Only samples that have 2 representative scores were used in the analysis.

Kaplan–Meier method and Cox regression model were used for univariate survival analysis, based on the average score of each paired TMA core. Kaplan–Meier Method was used to estimate the survival functions, and median survival times and their 95% confidence intervals. The HRs and their 95% confidence intervals were obtained using Cox regression. These studies were approved by the Human Ethics Committee (protocol H06-03300) at the University of British Columbia.

In colorectal cancer, CDK10 was consistently reported to be upregulated in five independent, high-throughput gene expression studies with a mean fold change of 13.85 (Supplementary Table S1; ref. 7). To validate this data, CDK10 transcript levels were determined in human colorectal cancer tissues and paired normal colon tissues (Fig. 1A). On average, CDK10 expression was >2-fold higher (P < 0.05) in colorectal cancer tissues than their respective matched normal colon tissues. Similarly, mRNA levels of CDK10 in colorectal cancer cell lines MIP101, RKO, and HCT116 cells were significantly higher by 8.7-, 4.6- and 10.9-fold, respectively, in comparison with the normal FHC colon cells (Fig. 1B).

Given CDK10′s overexpression in colorectal cancer samples, we proceeded to examine the clinical relevance of CDK10 overexpression in this disease by using a TMA, which was immunostained and scored for CDK10 expression. Interestingly, patients with stage III–IV colon cancer that had low tissue expression of CDK10 had a significantly longer median survival (4.65 years; 95% CI, 2.89–5.56), when compared with individuals whose tumors expressed high levels of CDK10 (2.61 years; 95% CI, 1.36–15.24; P < 0.02; Fig. 1C and D). Given the small number of tissues representing stage II colorectal cancer, the effect of CDK10 expression could not be determined. No significant correlations were observed based on the levels of tissue CDK10 expression in colorectal cancer and patient demographics or clinicopathologic features, such as age, sex, tumor stage, tumor site, grade, and depth of invasion (Table 1). These results suggest that CDK10 may contribute to colorectal cancer pathogenesis, leading to a more aggressive disease, rapid progression and poorer overall survival.

To investigate the role of CDK10 in colorectal cancer, CDK10 expression levels were modulated to determine their effects on growth and apoptosis. Initially, normal colon cells, FHC, were transiently transfected with CDK10 to overexpress this protein, and conversely, MIP101, RKO, and HCT116 cells (which have higher levels of CDK10 expression) were transfected with CDK10 siRNA to knockdown CDK10 expression (Supplementary Fig. S1).

CDK10 expression dramatically influenced apoptosis in colorectal cancer and normal colon cells: transient overexpression of CDK10 in normal colon cell line FHC not only significantly increased cell viability by 75.6% (Fig. 2A) but also decreased caspase-3/7 activity by 50.8% (Fig. 2B) following overexpression of CDK10, in comparison with normal cells transfected with the control vector. In line with these observations, knockdown of CDK10 expression in three colon cancer cell lines (MIP101, HCT116, and RKO) increased caspase-3/7 activity by at least 2.3-fold in MIP101 and RKO cells and by 1.8-fold in HCT116 cells (Fig. 2C). Similarly, a significantly greater numbers of TUNEL-positive cells were observed, with a 3- and 4.5-fold increase of TUNEL-positive cells in CDK10-siRNA–transfected MIP101 and RKO cells, respectively, in comparison with cells transfected with control scramble siRNA (Fig. 2D and E).

To further examine the mechanisms by which CDK10 may be influencing apoptosis, the expression of key proteins involved in the apoptotic cascade was analyzed following transfection of MIP101, RKO, and HCT116 cells with CDK10 siRNA. We noted that the protein levels of anti-apoptotic Bcl-2 and Bcl-x_L were downregulated in these colorectal cancer cell lines following CDK10 silencing, with the most dramatic reduction in RKO cells (Fig. 2F; Supplementary Fig. S2). Proapoptotic members of the Bcl-2 family of proteins, such as Bax and Bad, remained unaffected. Similarly, the level of cleaved caspase-8, detectable in RKO and HCT116 cells only, did not change following CDK10 siRNA silencing, suggesting that the extrinsic pathway of apoptosis may not be involved in CDK10-mediated suppression of apoptosis in colorectal cancer.

To determine the effects of CDK10 overexpression in colorectal cancer, RKO and MIP101 cells stably overexpressing wild-type CDK10 (RKO/CDK10 WT and MIP/CDK10 WT) as well as a kinase-defective/dominant-negative form of CDK10 (RKO/CDK10 DN and MIP/CDK10 DN) and their control counterparts (RKO/NeoB and MIP/NeoB) were established (Fig. 3A). These colorectal cancer cell lines had lower CDK10 expression than HCT116 cells, and therefore, were chosen for these experiments.

In vivo, the tumor-promoting potential of CDK10 was demonstrated by the dramatic growth of tumor xenografts of MIP/CDK10 WT cells compared with tumors of MIP/NeoB cells (Fig. 3B). Interestingly, tumor xenografts of MIP/CDK10 DN cells had a significantly slower rate of growth in comparison to control MIP/NeoB cells, suggesting that the kinase activity of CDK10 may be involved in the growth-promoting properties of colorectal cancer. Interestingly, cell proliferation did not increase in tumor xenografts of MIP/CDK10 as shown by immunostaining of Ki-67 (Supplementary Fig. S3). However, increased expression of cleaved caspase-3 was observed in the tumor xenografts of MIP/CDK10 DN cells suggesting that apoptosis was contributing to tumor suppression in xenografts of MIP/CDK10 DN (Supplementary Fig. S3).

In vitro, overexpression of CDK10 dramatically increased cell proliferation in RKO/CDK10 WT, in comparison with either the kinase defective RKO/CDK10 DN (number of cells increased by 231%) or the control RKO/NeoB cells (cells increased by 356%, Fig. 3C). Interestingly, the effect of the kinase mutant also observed in MIP/CDK10 DN cells, where a 64.2% (P < 0.001) and 47.8% (P < 0.05) reduction was observed in comparison with control MIP/Neo and MIP/CDK10 WT respectively (Fig. 3D). However, there was no significant difference in cell proliferation between MIP/NeoB and MIP/CDK10 WT over a course of 145 hours, likely due to the already relatively higher basal expression of CDK10 in MIP101 cells (Fig. 1B). In line with previous observations that siRNA knockdown of CDK10 expression resulted in upregulation of the apoptotic cascade, the effect of overexpressing CDK10 in both RKO and MIP101 cells effectively reduced caspase 3/7 activity by 58.4% and 33.7% (day 4, P < 0.001 and P < 0.05) in RKO/CDK10 WT and MIP/CDK10 WT respectively, in comparison with RKO/NeoB or MIP/NeoB. Interestingly, the effect of the kinase defective mutant appears to exert an earlier effect by augmenting caspase 3/7 activity by 94.3% in RKO/CDK10 DN and 22.5% in MIP/CDK10 DN as early as 1 day of incubation, in comparison to their empty vector control cell lines (Fig. 3E and F).

Although variable CDK10 expression in colorectal cancer influenced cellular proliferation and apoptosis, the expression of CDK2, 4, 5, 6, 9 (Supplementary Fig. S4) and Aurora Kinase A (Supplementary Fig. S5) were unaffected by changes in CDK10 expression in colorectal cancer cell lines. Similarly, variable expression of CDK10 also did not influence cell-cycle progression in these cells (Supplementary Fig. S6).

Because CDK10 silencing by siRNA facilitated the induction of apoptosis in colorectal cancer cells, with a concomitant reduction in the expression of Bcl2 and Bcl-x_L, we next examined the importance of Bcl-2 and Bcl-x_L in mediating CDK10′s ability to suppress apoptosis.

In RKO/CDK10 WT cells, Bcl-x_L and Bcl-2 levels increased with CDK10 overexpression, as compared with control RKO/NeoB cells (Fig. 4A). Again, in the kinase defective RKO/CDK10 DN cells, there was a reduction in Bcl-2 and Bcl-x_L which further supports the observations that the kinase activity of CDK10 may be important in modulating apoptosis and the expression of these two anti-apoptotic proteins. To determine whether Bcl-2 and Bcl-x_L influences CDK10′s ability to suppress apoptosis, we assessed the effect of knocking-down the expression of Bcl-2 and Bcl-x_L in cells overexpressing CDK10 (RKO/CDK10 WT, Fig. 4B and C). A dramatic reversal of CDK10′s inhibitory activity on apoptosis was observed following siRNA-mediated silencing of Bcl-2 in RKO/CDK10 WT, with a significant increase in caspase 3/7 activity by 71% ± 4.8 % in comparison to scramble-control transfected RKO/CDK10 WT cells (Fig. 4B). A similar, but less dramatic increase in caspase 3/7 is observed following siRNA silencing of Bcl-x_L (Fig. 4C).

Because conventional chemotherapy induces, among others, DNA damage and subsequent activation of the mitochondrial cell death pathway, which is regulated by a balance between pro- and anti-apoptotic Bcl-2 family members, the ability of CDK10 expression to modulate a cell's response to 5-FU-mediated apoptosis was evaluated. In RKO and MIP101 cells stably overexpressing wild-type CDK10, and following exposure to incremental concentrations of 5-FU, there was a significant increase in IC₅₀ requirements in these RKO/CDK10 WT and MIP/CDK10 WT cells. Specifically, there was a 1.93 ± 0.46; and 1.86 ± 0.34-fold increase in IC₅₀ in RKO/CDK10 WT and MIP/CDK10 WT, respectively, relative to their control counterparts (RKO/NeoB and MIP/NeoB; Fig. 4D). This is supported by the results of the colony forming assay, which demonstrated 202 and 545% more colonies despite exposure to 1 and 5 μmol/L 5-FU, respectively, in MIP/CDK10 WT in comparison to MIP/NeoB. Interestingly, there were 534% and 2,610% more colonies in MIP/CDK10 WT than MIP/CDK10 DN cells following exposure to 1 and 5 μmol/L 5-FU, and similarly, 264% and 479% more colonies following exposure to 1 and 5 μmol/L 5-FU, respectively, in MIP/NeoB than MIP/CDK10 DN (Fig. 4E). These findings with MIP/CDK10 DN cells suggest that the kinase-defective CDK10 MIP101 cells were significantly more sensitive to 5-FU than MIP/CDK10 WT or MIP/NeoB, again indicating that the kinase activity of CDK10 may be important in contributing to its tumor-promoting properties in colorectal cancer.

CDK10′s ability to enhance growth and to prevent apoptosis in colorectal cancer cells in vitro indicated that targeting this gene in vivo may be of potential benefit in preventing the growth of colorectal cancer in vivo. To examine this possibility, a preclinical model using PDX was used. Constructs of GFP expressing lentivirus vectors expressing either scramble siRNA or a siRNA sequence directed against CDK10 were used for direct intratumoral injection of PDX implanted in SCID mice. A dramatic effect of CDK10 siRNA in suppressing tumor growth in vivo was observed within 10 days of treatment. By day 26 of treatment, although the average volume of tumors receiving scramble siRNA reached approximately 1,200% of the initial tumor size, the average volumes of tumors receiving CDK10 siRNA remained essentially unchanged (Fig. 5A). The GFP images of the harvested tumors indicate that tumors in both groups were expressing the GFP-tagged lentiviral siRNA (either scramble or CDK10-siRNA; Fig. 5B and C); however, CDK10 expression was effectively suppressed only in PDX tumors receiving CDK10-siRNA injections (Fig. 5D).

Also in line with our in vitro observations, tumors that were injected with CDK10-siRNA had a concomitant suppression of Bcl-2 expression (Fig. 5E), with a moderate effect on Bcl-x_L (Fig. 5F).

In this study, a potential oncogenic role for CDK10 in colorectal cancer is demonstrated. In vivo, overexpression of CDK10 in colorectal cancer cells promotes tumor growth, whereas siRNA targeting this gene in patient-derived human colorectal cancer xenografts dramatically inhibits their growth. These interesting in vivo observations correlate with CDK10′s ability to enhance cellular growth and viability, and to also inhibit apoptosis when overexpressed in colorectal cancer cells in vitro. In addition, results of a TMA of human colorectal cancers also reveal its potential as a prognostic biomarker: high tissue CDK10 expression in the primary colorectal cancer is prognostic of poor clinical outcomes, with individuals progressing in their disease and ultimately death, in a significantly shorter time interval than those with tumors with low CDK10 expression.

CDK10 was originally reported as a novel putative member of the CDK family of protein serine/threonine kinases based on structural similarities with CDKs (26, 27). As with other members of the CDK family, it is involved in regulating the cell cycle, specifically at the G₂–M phase (14). Its ability to regulate the cell cycle appears to depend on its kinase activity, as overexpression of a dominant-negative (kinase defective) mutant CDK10 in U2OS and Saos-2 osteosarcoma, and T98G glioblastoma cells, prevented cell-cycle progression and arrested cells in the G₂–M phase, leading to a suppression in cellular growth (14). Interestingly, although these initial studies demonstrated the importance of the kinase activity in CDK10 in maintaining cellular growth, overexpression of the wild-type protein in these cells did not accelerate or promote cellular proliferation, in comparison with cells expressing basal levels of this protein (14). In colorectal cancer, however, we were able to demonstrate that overexpression of CDK10 in RKO and MIP101 colorectal cancer cells contributed to a significant increase in cell growth in vitro and tumor growth in vivo (Fig. 3B, C and D). It should be noted that cell-cycle analysis performed on these cell lines in vitro did not reveal significant differences in cell-cycle regulation (Supplementary Fig. S6), suggesting that inhibition of cell death may be one of the primary modalities by which CDK10 promotes cell growth. This ability to promote cell growth appeared to be dependent on the integrity of its kinase activity, as cells overexpressing the kinase-defective mutant CDK10 failed to achieve an accelerated growth rate that was observed with cells overexpressing CDK10, but instead, had similar growth properties as those colorectal cancer cells transfected with the control vector (Fig. 3). In vivo, mutant kinase activity also contributes to a significant reduction in tumor growth, not only in comparison with tumors overexpressing CDK10, but as well, in comparison with tumors expressing basal levels of CDK10 (MIP/CDK10 DN, Fig. 3B). Similarly, we demonstrate that apoptosis is not only restored, but is in fact, augmented in cells overexpressing the kinase-defective mutant CDK10, in comparison with cells expressing basal levels of the wild type protein. Our observation that the anti-apoptotic activity of CDK10 may depend on its kinase activity is interesting because with other CDKs, such as CDK1 (28), CDK2 (28–30), CDK3 (28), CDK4 (30), CDK5 (31), CDK6 (30), and CDK11 (32), an active kinase has been shown to promote apoptosis. CDK10 therefore, appears to be a unique member of its family, in that in colorectal cancer, its kinase activity may provide a survival advantage by protecting cells from undergoing apoptosis. It is not clear how the kinase activity of CDK10 may be contributing to its anti-apoptotic function, but we are investigating the possibility that this kinase activity may be involved in the activation of a phospho-protein that in turn increases the expression of the anti-apoptotic proteins Bcl-2 and Bcl-x_L.

Our findings also indicate that CDK10′s suppressive effect on apoptosis occurs via Bcl-2 and, to a lesser extent, Bcl-x_L, as demonstrated by the reversal of CDK10′s ability to inhibit apoptosis following knock-down of Bcl-2 and Bcl-x_L (Fig. 4B and C). Knockdown of Bcl-2 eliminated the ability of CDK10 to inhibit apoptosis, despite the overexpression of the wild type CDK10 in these colorectal cancer cells. Interestingly, inhibition of CDK7 and CDK9 has also been shown to enhance apoptosis through inhibition of Bcl-2 (33). However, the observation that CDK9 expression remained unchanged in the setting of CDK10 overexpression in colorectal cancer cells suggests that, in colorectal cancer, CDK10′s inhibitory effect on apoptosis via Bcl-2 is CDK10 specific.

It is also interesting to note that the effects of lowering the expression of CDK10 by siRNA knockdown in RKO cell lines were more pronounced than colorectal cancer cells with relatively higher CDK10 expression (MIP101 and HCT116 cells). We postulate that there may be a threshold effect—as lowering CDK10 levels below a certain level (which can be achieved more easily in cells with moderate expression of CDK10) facilitates the induction of apoptosis, as seen with TUNEL assay and Bcl-2 and Bcl-x_L expression.

In line with our observations of CDK10′s growth- and survival-promoting properties in colorectal cancer cells and PDX tumors in vivo, it is not surprising that the overexpression of CDK10 in colorectal cancer also contributes to a diminished response to chemotherapy (Fig. 4D and E). These in vitro observations may also explain our findings that high levels of CDK10 in the primary human colorectal cancer tumors correlate with poor clinical outcomes: patients have a higher relapse rate and earlier progression to distant metastasis, likely because these tumors respond poorly to chemotherapy.

Taken together, the findings of this study indicate that CDK10 promotes tumorigenesis in colorectal cancer. However, the biological activity of this protein appears to be tissue-specific, as high levels of CDK10 in breast cancer and hepatocellular carcinomas appear to have an opposite effect: CDK10 inhibits cell proliferation, cell migration and anchorage-dependent growth in hepatocellular carcinoma (34). In breast cancer, CDK10 downregulation appears to be associated with tamoxifen resistance, occurring as a result of an increase in ETS2-driven transcription of c-RAF and MAPK pathway activation (35). In fact, CDK10′s effect on ETS2 has been shown to involve its activating cyclin partner, cyclin M: the CDK10/cyclin M heterodimer positively regulates ETS degradation by the proteasome thereby influencing ETS2′s proto-oncogenic activity (15). In colorectal cancer however, CDK10 appears to promote tumor growth, inhibit apoptosis, de-sensitize tumors to chemotherapy and may be a prognostic biomarker of early disease progression and poor outcome. The reason for this diverging pathological activity of CDK10 in different cancers is unclear, but it may be related to the variable expression of the different isoforms of CDK10. There is some indication that the longer isoforms of CDK10 are predominant in growing cells, whereas a shorter isoform of CDK10 that lacks exon 2 is more abundant in non-proliferating cells (36). Similarly, cyclin M, a gene product of FAM58A, has also been shown to undergo differential splicing in cancer (37) that could interfere with its interaction with CDK10 and downstream biological activities. Opposing biological activities are well known to occur as a result of alternative splicing, as observed with Bcl-x_L, which protects cells from apoptosis, as opposed to facilitating apoptosis when Bcl-x_S is expressed (38).

In summary, CDK10′s ability to promote colorectal cancer tumor growth and suppress apoptosis indicates that this gene is involved in supporting the growth of colorectal cancer, via its kinase activity. Inhibition of this protein, or its kinase activity, effectively prevents tumor growth in a PDX model, thus making it a promising novel target that can be developed for the treatment of colorectal cancer. However, a pharmacological or small-molecule inhibitor with specificity against CDK10 would be essential, given the variable results from Phase II clinical trials with other CDK inhibitors (CDK 4/6 inhibitor (refs. 39, 40); CDK9; refs. 41–43) resulting from poor target selectivity and associated toxicity profile (e.g., hematologic and gastrointestinal toxicity of CDK 4/6 inhibitors; ref. 44). As drug designers continue to develop better tools, it is possible to envision the development of a CDK10-selective inhibitor with a more tolerable toxicity profile that can be used for the treatment of colorectal cancer.

No potential conflicts of interest were disclosed.

Conception and design: L.-B. Weiswald, M.R. Hasan, I.T. Tai

Development of methodology: L.-B. Weiswald, M.R. Hasan, M. Rahman, S. Vacher, A.P. Weng, H.F. Kennecke, I. Bièche, D.T. Yapp, I.T. Tai

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L.-B. Weiswald, M.R. Hasan, J.C.T. Wong, C.C. Pasiliao, M. Rahman, J. Ren, S. Vacher, A.P. Weng, H.F. Kennecke, I. Bièche, D.F. Schaeffer, D.T. Yapp

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L.-B. Weiswald, M.R. Hasan, M. Rahman, Y. Yin, S. Vacher, H.F. Kennecke, I. Bièche, I.T. Tai

Writing, review, and/or revision of the manuscript: L.-B. Weiswald, M.R. Hasan, M. Rahman, Y. Yin, H.F. Kennecke, D.F. Schaeffer, I.T. Tai

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Rahman, J. Ren, S. Gusscott, A.P. Weng

Study supervision: I.T. Tai

The authors would like to thank Dr. Xavier Sagaert for assistance in scoring the tissue microarray for CDK10 expression and Dr. A. Giordano for providing the CDK10 plasmids.

This work was supported by operating grants from the Canadian Institutes of Health Research (CIHR, CST-85477) and from France Canada Research Funds. I.T. Tai is a CIHR new investigator and MSFHR scholar. LB. Weiswald is a recipient of a post-doctoral fellowship from the CIHR.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.