The significance of Cdk5 in cell-cycle control and cancer biology has gained increased attention. Here we report the inverse correlation between the protein levels of Cdk5 and p21CIP1 from cell-based and clinical analysis. Mechanistically, we identify that Cdk5 overexpression triggers the proteasome-dependent degradation of p21CIP1 through a S130 phosphorylation in a Cdk2-independent manner. Besides, the evidence from cell-based and clinical analysis shows that Cdk5 primarily regulates nuclear p21CIP1 protein degradation. S130A-p21CIP1 mutant enables to block either its protein degradation or the increase of cancer cell growth caused by Cdk5. Notably, Cdk5-triggered p21CIP1 targeting primarily appears in S-phase, while Cdk5 overexpression increases the activation of Cdk2 and its interaction with DNA polymerase δ. The in vivo results show that Cdk2 might play an important role in the downstream signaling to Cdk5. In summary, these findings suggest that Cdk5 in a high expression status promotes cancer growth by directly and rapidly releasing p21CIP1-dependent cell-cycle inhibition and subsequent Cdk2 activation, which illustrates an oncogenic role of Cdk5 potentially applied for future diagnosis and therapy. Cancer Res; 76(23); 6888–900. ©2016 AACR.

Cdk5 activation has been reported in numerous types of tumors and high Cdk5 expression is detrimental to patients' survival (1) and poor prognosis (2). The single-nucleotide polymorphisms in the promoter and amplification of the CDK5 gene increase the risk of human cancers (3, 4), while Cdk5 activation correlates to prostate cancer metastasis and pancreatic cancer progression (5, 6). Our previous studies show that Cdk5 activates STAT3 (7) or androgen receptor (AR; ref. 8) in different cancer cells and supervises the cooperation of AR and STAT3 for prostate cancer growth (9). We also show the cell-cycle arrest or apoptosis of various cancer cells by drug-triggered Cdk5 activation (10–12). The involvements of Cdk5 in microtubule depolymerization, tumor cell mitosis (13), and Retinoblastoma-related thyroid cancer proliferation (14) were reported. These observations strongly suggest the relevant contribution of Cdk5 to cell-cycle control and cancer growth.

p21CIP1 is a member of Cdk inhibitor family containing conserved N-terminal regions for Cdk-binding, which suppresses cell-cycle progression (15). Loss of p21CIP1 causes carcinogenesis and spontaneous tumor development (16). The study indicates that the dual roles p21CIP1 plays in cell-cycle inhibition and anti-apoptosis depend on its subcellular localization (17). Nuclear p21CIP1 inhibits Cdks to restrain cell cycle; cytoplasmic p21CIP1 resists apoptosis by binding with proapoptotic proteins, such as ASK1 or caspase 3/8/10 (18). As the phosphorylation regulation of p21CIP1 determines its protein stability and potentially regulates its subcellular localization (19), it is of interest to investigate the upstream kinases to p21CIP1 in cell-cycle control.

Here, we provide evidence demonstrating the inverse correlation of Cdk5 and p21CIP1 protein levels and observed the novel and specific phosphorylation-dependent degradation of p21CIP1 by Cdk5. The Cdk5 impact on p21CIP1 was primarily restricted to the nucleus, followed by subsequent Cdk2 activation and promotion of cancer growth. Suppression of Cdk5 by RNAi in a xenograft model increased p21CIP1 protein levels and inhibited tumor growth. This study demonstrates the relevance of Cdk5 in p21CIP1-related regulation and sheds light on the potential targeting of Cdk5 in future cancer therapy.

Cell culture, transfection, and protein analysis

Breast cancer (MCF-7 and MDA-MB-231), prostate cancer (LNCaP, 22Rv1, DU145, and PC-3), lung cancer (A549), and fibroblast-like (COS-1) cell lines were all purchased from the cell bank (Bioresource Collection and Research Center, Taiwan), that performed cell line characterizations (by STR analysis) and passaged in the laboratory for fewer than 6 months after receipt or resuscitation. MCF-7 and DU145 cells were grown in 10% FBS (Gibco)-containing MEM supplemented with 0.1 mmol/L nonessential amino acids, 1 mmol/L sodium pyruvate, and penicillin/streptomycin (P/S). MDA-MB-231 and COS-1 cells were maintained in DMEM supplemented with 10% FBS and P/S. LNCaP and 22Rv1 cells were cultured in RPMI1640 supplemented with 10% FBS, 10 mmol/L HEPES, 1 mmol/L sodium pyruvate, and P/S. PC-3 and A549 cells were cultured in Ham F-12 medium containing P/S with 7% and 10% FBS, respectively. All cells were incubated under 5% CO2 at 37°C. To establish stable human cancer cell lines, the coding sequences of empty-vector (EV), wild-type (WT) Cdk5, kinase-dead Cdk5 with D144N mutation (KD; ref. 14) were cloned into mammalian expression plasmids and then transfected into cells. The transfected cells were subsequently selected in the presence of G418 (TOKU-E) for at least 4 weeks. The expression plasmid or siRNA were premixed within Lipofectamine 2000 or Lipofectamine RNAiMAX transfection reagent (Invitrogen) according to the manufacturer's instructions, and then the liposome/nucleic acids complex was transfected into cells with culture medium for 2 days. Commercial products of siRNA-Cdk5 (siCdk5) and nonspecific control siRNA (siCon) were purchased from Dharmacon (SMARTpool). Small hairpin RNAs, shCdk5 (TRCN0000021467) and shCdk2 (TRCN0000195003), for specific protein knockdown were obtained from National RNAi Core Facility of Academia Sinica (Taiwan). The 3×FLAG tagged p21CIP1 expression plasmid was kindly provided by Professor Hideo Nishitani, Graduate School of Life Science, University of Hyogo (Hyogo, Japan; ref. 20). The kinase-dead mutant pcDNA3-D144N-Cdk5 (KD) and 3×FLAG-S130A-p21CIP1 were cloned using the QuickChange Lightning Site-Directed Mutagenesis Kit (Stratagene) according to its standard instruction manual. Clinical specimens from breast cancer patients who underwent surgical resection and the procedure was approved by the Institutional Review Boards of Chang Bing Show Chwan Memorial Hospital (No. IRB-D016-B). The details of Western blot analysis and immunoprecipitation were referred to previous publication (8).

Immunostaining

Cells on coverslips were sequentially fixed, permeabilized, and blocked as described previously (8). After primary antibody binding, samples were incubated with Alexa Fluor 488- or 546-conjugated secondary antibodies (Molecular Probes). DAPI fluorescent dye was used to stain cell nuclei. The coverslips were mounted and examined by fluorescent microscopy (Olympus). The human paraffin-embedded tissue arrays (prostate cancer:PR952 and PR803a; lung cancer: LC1201; breast cancer: BR1921a) were purchased from Biomax Co. and stained by procedures of Vectastain ABC Kit and DAB Peroxidase Substrate Kit (Vector Laboratories). The slides were counterstained with hematoxylin solution (Merck) to indicate cell nuclei. The specimens were mounted and imaged by light microscopy (Olympus). Two experts blinded to the image sources evaluated the specimens in accordance with a scoring system that was based on the intensity of staining signals (Supplementary Fig. S3 for the details).

Quantitative real-time PCR

Total RNA extraction (Genemark) following by reverse transcription PCR (Applied Biosystems) was performed using SYBR Green (Roche Applied Science)-dependent real-time quantification as described previously (21). Data were analyzed as Ct values and were adjusted relative to the levels of actin housekeeping gene. Primers used in this study were listed in Supplementary Methods.

In vitro kinase assay

Recombinant full-length human p21CIP1 proteins (wild-type or S130A mutant) were expressed in the pQE30 expression system and the proteins were purified following the Qiagen manufacturer's protocol. Active Cdk5/p35 protein complex were purchased from Millipore. The kinase assays of immunoprecipitated or recombinant Cdk5 or Cdk2 were performed with histone H1 (Millipore) or recombinant p21CIP1 as the substrates like previous description (8, 22). [γ-32P]-ATP (Perkin Elmer) or cold ATP (Sigma) with phosSTOP phosphatase inhibitor cocktail (Roche) were used and the phospho-signals were detected by isotope visualization or Western blotting with specific antibody.

Identification of phosphorylation sites by mass spectrometry

The product of in vitro Cdk5-dependent phosphorylated p21CIP1 was resolved by SDS-PAGE and analyzed by LC/MS-MS (Q-Exactive mass spectrometer, Thermo Fisher Scientific), which was performed by the Mass Solutions Technology Co., Ltd.

Cell-cycle synchronization

MCF-7 cells were synchronized at different stages of the cell cycle [G1 phase: serum starvation (24 hours); S-phase: hydroxyurea treatment (3 μmol/L, 24 hours); G2–M phase: nocodazole treatment (50 ng/mL, 24 hours)] in accordance with a modified protocol (23, 24).

Tumor xenograft

Approximately 1 × 107 MDA-MB-231 or 2 × 107 22Rv1 cells were subcutaneously injected into the flanks of 5-week-old male BALB/c nude mice. The mice were randomly grouped after tumor volumes reached over 100 mm3 and intratumorally injected with a mixture of 10 μg pcDNA3-empty-vector, shCdk5, shCdk2, or pcDNA3-Cdk5 plasmids by in vivo jet-PEI transfection reagent (Polyplus) every 3 days. Tumor volumes were measured and calculated using the formula L×W×W×3.14/6, where L and W are the longest and shortest tumor axis, respectively. All animal experiments were performed according to the protocol approved by the Institutional Animal Care and Use Committee of National Chung Hsing University.

Statistical analysis

Data are presented as the mean ± SEM. The paired Student t test was used for statistical evaluation of data from the cell proliferation experiments. One-way ANOVA was utilized for animal experiments. The correlation between Cdk5 levels and p21CIP1 levels in human cancer specimens was analyzed using Fisher exact test. Statistical significance was noted with ∗, P < 0.05; ∗∗, P < 0.01. n.s. stands for no significance.

Cdk5 reduces p21CIP1 protein levels in cancer cells by increasing proteasome-mediated degradation

Figure 1A showed that p21CIP1 protein was decreased after Cdk5 overexpression in MCF-7 cells (Fig. 1A), while Cdk5 knockdown by specific siRNA resulted in increased p21CIP1 protein levels (Fig. 1B). The kinase activity of Cdk5 was in accordance with its protein levels (Supplementary Fig. S1A). The expression levels of other Cdks (including Cdk1, 2, 4, 6) were also observed with Cdk5 manipulations. Interestingly, the expression of Cdk2 after Cdk5 knockdown was increased implying the putative functional compensation between Cdk5 and Cdk2. On the basis of p21CIP1 increase with Cdk2 upregulation after Cdk5 knockdown, the authors suggested that Cdk5 might play a specific and prominent role in p21CIP1 protein regulation. Although p21CIP1 protein levels were affected by Cdk5, p21CIP1 mRNA expression was unaffected (Fig. 1C). The distribution of Cdk5 protein levels in 39 breast cancer patients' tumor specimens was shown in Fig. 1D for the quantified results and Supplementary Fig. S2A for the blotting images. The mutual growth and decline between Cdk5 and p21CIP1 protein levels in the 29 patients' tumor specimens with Cdk5 overexpression are shown in Fig. 1E (see images in Supplementary Fig. S2B), in which the Cdk5 protein levels from 29 patients were categorized into three grades. Besides, the cancer tissue arrays from 397 patients (including cancers of prostate, lung, and breast) were examined by IHC and the representative images with high Cdk5 level (low p21CIP1) and low Cdk5 level (high p21CIP1) were presented in Fig. 1F. Table 1 indicated the negative correlations between the levels of Cdk5 and p21CIP1 in the tumor sections of the prostate (P < 0.001), lung (P = 0.021), and breast (P = 0.031; analyzed by the Fisher exact test). The scoring system was characterized in Supplementary Fig. S3. Cdk5 overexpression facilitated p21CIP1 protein degradation in stable cancer cell lines identified by chase experiments with cycloheximide (CHX) treatment (Fig. 1G). The Cdk5-dependent p21CIP1 degradation was apparently rescued by treatments with proteasome inhibitors, MG132 or lactacystin (Fig. 1H and Supplementary Fig. S4A). These data suggest that the Cdk5-related decline of p21CIP1 protein is likely regulated through promoting the proteasome-mediated degradation.

Figure 1.

Cdk5 downregulates p21CIP1 protein and decreases p21CIP1 protein stability via proteasome-dependent degradation in cancer cells. Protein levels are shown after overexpression (A) or knockdown (B) of Cdk5 in MCF-7 cells. C, mRNA levels of p21CIP1 (CDKN1A) with stable Cdk5 overexpression (WT) detected by quantitative real-time PCR are shown as fold changes relative to the levels of EV group (empty vector). D, Cdk5 protein was measured in paired protein specimens from 39 breast cancer patients (T, tumor; N, normal). E, The correlation between p21CIP1 and Cdk5 protein measured in the specimens of 29 Cdk5-overexpressing patients is shown. The images of Western blotting are shown in Supplementary Fig. S2. F, The immunohistochemical images of Cdk5 and p21CIP1 proteins (brown signals) from six representative patients of three cancer types are shown. Cell nuclei were stained by hematoxylin (blue signals). Scale bar, 200 μm (for prostate cancer specimens); 100 μm (for lung and breast cancer specimens). G, Protein levels were shown in LNCaP and MCF-7 cells stably overexpressing WT-Cdk5 with cycloheximide (CHX) treatment (10 μg/mL, 0, 2, 4, or 6 hours). H, MCF-7 cells were transfected with WT-Cdk5 and treated with or without MG132 (5 μmol/L, 6 hours). Mock group was treated with DMSO as a solvent of MG132. Actin served as the internal control. The quantitative results are provided with the related images. The results are represented as the means ± SEM (n.s., no significance). *, P < 0.05; **, P < 0.01 versus the related control groups.

Figure 1.

Cdk5 downregulates p21CIP1 protein and decreases p21CIP1 protein stability via proteasome-dependent degradation in cancer cells. Protein levels are shown after overexpression (A) or knockdown (B) of Cdk5 in MCF-7 cells. C, mRNA levels of p21CIP1 (CDKN1A) with stable Cdk5 overexpression (WT) detected by quantitative real-time PCR are shown as fold changes relative to the levels of EV group (empty vector). D, Cdk5 protein was measured in paired protein specimens from 39 breast cancer patients (T, tumor; N, normal). E, The correlation between p21CIP1 and Cdk5 protein measured in the specimens of 29 Cdk5-overexpressing patients is shown. The images of Western blotting are shown in Supplementary Fig. S2. F, The immunohistochemical images of Cdk5 and p21CIP1 proteins (brown signals) from six representative patients of three cancer types are shown. Cell nuclei were stained by hematoxylin (blue signals). Scale bar, 200 μm (for prostate cancer specimens); 100 μm (for lung and breast cancer specimens). G, Protein levels were shown in LNCaP and MCF-7 cells stably overexpressing WT-Cdk5 with cycloheximide (CHX) treatment (10 μg/mL, 0, 2, 4, or 6 hours). H, MCF-7 cells were transfected with WT-Cdk5 and treated with or without MG132 (5 μmol/L, 6 hours). Mock group was treated with DMSO as a solvent of MG132. Actin served as the internal control. The quantitative results are provided with the related images. The results are represented as the means ± SEM (n.s., no significance). *, P < 0.05; **, P < 0.01 versus the related control groups.

Close modal
Table 1.

Inversed correlation between Cdk5 and p21CIP1 protein expressions in human cancers

Cdk5 Level (%)
p21CIP1 Level in human cancers (%)NegativeLowModerateHighTotal, nPPearson's R
Prostate cancer P < 0.001a R = −0.22 
 Negative 1 (0.69%) 5 (3.45%) 0 (0.00%) 3 (2.07%) 9 (6.21%)   
 Low 7 (4.83%) 21 (14.48%) 22 (15.17%) 13 (8.97%) 63 (43.45%)   
 Moderate 9 (6.21%) 26 (17.93%) 3 (2.07%) 3 (2.07%) 41 (28.27%)   
 High 4 (2.76%) 22 (15.17%) 1 (0.69%) 5 (3.45%) 32 (22.07%)   
 Total, n 21 (14.48%) 74 (51.04%) 26 (17.93%) 24 (16.55%) 145 (100%)   
Lung cancer P = 0.021a R = −0.12 
 Negative 7 (7.29%) 12 (12.50%) 8 (8.33%) 4 (4.17%) 31 (32.29%)   
 Low 3 (3.13%) 4 (4.17%) 9 (9.38%) 11 (11.46%) 27 (28.13%)   
 Moderate 6 (6.25%) 9 (9.38%) 1 (1.04%) 2 (2.08%) 18 (18.75%)   
 High 7 (7.29%) 5 (5.21%) 5 (5.21%) 3 (3.13%) 20 (20.83%)   
 Total, n 23 (23.96%) 30 (31.25%) 23 (23.96%) 20 (20.83%) 96 (100%)   
Breast cancer P = 0.031a R = −0.07 
 Negative 5 (3.21%) 19 (12.18%) 10 (6.41%) 8 (5.13%) 42 (26.92%)   
 Low 3 (1.92%) 13 (8.33%) 24 (15.38%) 15 (9.62%) 55 (35.26%)   
 Moderate 10 (6.41%) 15 (9.62%) 8 (5.13%) 7 (4.49%) 40 (25.64%)   
 High 3 (1.92%) 8 (5.13%) 3 (1.92%) 5 (3.21%) 19 (12.18%)   
 Total, n 21 (13.46%) 55 (35.26%) 45 (28.85%) 35 (22.43%) 156 (100%)   
Cdk5 Level (%)
p21CIP1 Level in human cancers (%)NegativeLowModerateHighTotal, nPPearson's R
Prostate cancer P < 0.001a R = −0.22 
 Negative 1 (0.69%) 5 (3.45%) 0 (0.00%) 3 (2.07%) 9 (6.21%)   
 Low 7 (4.83%) 21 (14.48%) 22 (15.17%) 13 (8.97%) 63 (43.45%)   
 Moderate 9 (6.21%) 26 (17.93%) 3 (2.07%) 3 (2.07%) 41 (28.27%)   
 High 4 (2.76%) 22 (15.17%) 1 (0.69%) 5 (3.45%) 32 (22.07%)   
 Total, n 21 (14.48%) 74 (51.04%) 26 (17.93%) 24 (16.55%) 145 (100%)   
Lung cancer P = 0.021a R = −0.12 
 Negative 7 (7.29%) 12 (12.50%) 8 (8.33%) 4 (4.17%) 31 (32.29%)   
 Low 3 (3.13%) 4 (4.17%) 9 (9.38%) 11 (11.46%) 27 (28.13%)   
 Moderate 6 (6.25%) 9 (9.38%) 1 (1.04%) 2 (2.08%) 18 (18.75%)   
 High 7 (7.29%) 5 (5.21%) 5 (5.21%) 3 (3.13%) 20 (20.83%)   
 Total, n 23 (23.96%) 30 (31.25%) 23 (23.96%) 20 (20.83%) 96 (100%)   
Breast cancer P = 0.031a R = −0.07 
 Negative 5 (3.21%) 19 (12.18%) 10 (6.41%) 8 (5.13%) 42 (26.92%)   
 Low 3 (1.92%) 13 (8.33%) 24 (15.38%) 15 (9.62%) 55 (35.26%)   
 Moderate 10 (6.41%) 15 (9.62%) 8 (5.13%) 7 (4.49%) 40 (25.64%)   
 High 3 (1.92%) 8 (5.13%) 3 (1.92%) 5 (3.21%) 19 (12.18%)   
 Total, n 21 (13.46%) 55 (35.26%) 45 (28.85%) 35 (22.43%) 156 (100%)   

aP < 0.05 indicates statistical significance.

p21CIP1 is a novel substrate of Cdk5

The interaction between endogenous Cdk5 and p21CIP1 was barely detectable due to degradation in normal condition; however, the interaction became prominent after MG132 treatment in three cancer cell lines (Fig. 2A). The effects of MG132 or lactacystin on p21CIP1 mRNA was shown in Supplementary Fig. S5. These results imply that Cdk5-bound p21CIP1 was unstable and rapidly degraded by proteasome. The in vitro phosphorylation of recombinant p21CIP1 by Cdk5 was detected in a dose-dependent manner (Fig. 2B). LC/MS-MS analysis revealed that Ser130 of p21CIP1 was the single site of Cdk5-dependent phosphorylation (Fig. 2C). Further confirmation of p21CIP1 phosphorylation was performed using Western blotting and the dose-dependent phosphorylation by different amounts of Cdk5/p35 complex was presented in Fig. 2D. In addition, the Ser130 phosphorylation of p21CIP1 was dose dependently reduced by in vitro treatment with a Cdk5 inhibitor, roscovitine (Fig. 2E). By using histone H1 as a substrate for in vitro Cdk5 kinase assay, WT p21CIP1 but not S130A-mutant p21CIP1 could dose dependently decrease histone H1 phosphorylation (Fig. 2F). Altogether, we determine that p21CIP1 is a novel substrate of Cdk5.

Figure 2.

p21CIP1 is a specific substrate of Cdk5 through Ser130 phosphorylation, which accelerates p21CIP1 protein degradation. A, Cdk5 immunoprecipitation (IP) followed by Western blotting was performed on cancer cells treated with or without MG132 (5 μmol/L, 6 hours) and the input proteins from the same lysate are shown in the bottom. B, Cdk5 in vitro kinase assay was performed using recombinant human p21CIP1 protein as a substrate. C, The phosphorylated p21CIP1 peptides were analyzed by mass spectrometry and revealed that Ser130 was phosphorylated (lower case p) in the indicated amino acid sequence of p21CIP1 fragment (123 to 140). D and E, Ser130 phosphorylation of p21CIP1 is shown by Cdk5 in vitro kinase assay (D) and the effects of roscovitine (Rosc) on the phosphorylation are shown in E. F, Different levels of wild-type p21CIP1 (WT) or S130A mutant p21CIP1 (SA; 0.25, 0.5, and 1 μg) and histone H1 (5 μg) were used as substrates for the Cdk5 in vitro kinase assay. The related phospho-proteins are shown. G, EV, WT-, or KD-Cdk5 were overexpressed in MCF-7 cells in the presence of MG132 treatment (5 μmol/L, 6 hours) and the protein levels are shown. H, The FLAG-tagged WT or S130A mutant (SA) p21CIP1 were overexpressed in cells with or without wild-type Cdk5 (WT) overexpression and the protein levels are shown. I, Cycloheximide (CHX, 10 μg/mL) was administered (0, 2, 4, and 6 hours) and the degradation of exogenous WT or SA FLAG-p21CIP1 proteins was monitored in the presence of Cdk5 overexpression in different cell lines. The degradation curves of p21CIP1 protein are shown beside the images and the half-lives of p21CIP1 are also provided. The protein levels are shown and actin and α-tubulin served as internal controls. The quantitative results are provided with the related images. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups.

Figure 2.

p21CIP1 is a specific substrate of Cdk5 through Ser130 phosphorylation, which accelerates p21CIP1 protein degradation. A, Cdk5 immunoprecipitation (IP) followed by Western blotting was performed on cancer cells treated with or without MG132 (5 μmol/L, 6 hours) and the input proteins from the same lysate are shown in the bottom. B, Cdk5 in vitro kinase assay was performed using recombinant human p21CIP1 protein as a substrate. C, The phosphorylated p21CIP1 peptides were analyzed by mass spectrometry and revealed that Ser130 was phosphorylated (lower case p) in the indicated amino acid sequence of p21CIP1 fragment (123 to 140). D and E, Ser130 phosphorylation of p21CIP1 is shown by Cdk5 in vitro kinase assay (D) and the effects of roscovitine (Rosc) on the phosphorylation are shown in E. F, Different levels of wild-type p21CIP1 (WT) or S130A mutant p21CIP1 (SA; 0.25, 0.5, and 1 μg) and histone H1 (5 μg) were used as substrates for the Cdk5 in vitro kinase assay. The related phospho-proteins are shown. G, EV, WT-, or KD-Cdk5 were overexpressed in MCF-7 cells in the presence of MG132 treatment (5 μmol/L, 6 hours) and the protein levels are shown. H, The FLAG-tagged WT or S130A mutant (SA) p21CIP1 were overexpressed in cells with or without wild-type Cdk5 (WT) overexpression and the protein levels are shown. I, Cycloheximide (CHX, 10 μg/mL) was administered (0, 2, 4, and 6 hours) and the degradation of exogenous WT or SA FLAG-p21CIP1 proteins was monitored in the presence of Cdk5 overexpression in different cell lines. The degradation curves of p21CIP1 protein are shown beside the images and the half-lives of p21CIP1 are also provided. The protein levels are shown and actin and α-tubulin served as internal controls. The quantitative results are provided with the related images. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups.

Close modal

Cdk5 induces p21CIP1 protein degradation through Ser130 phosphorylation in a Cdk2-independent manner

With MG132 treatment, the levels of phospho-Ser130-p21CIP1 were correspondent to Cdk5 activity, while the total p21CIP1 protein was escaped from proteasomal degradation (Fig. 2G). FLAG-SA-p21CIP1 was resistant to Cdk5-triggered degradation in different cancer cell lines (Fig. 2H). The chase experiments illustrated that SA-p21CIP1 exhibited higher protein stability than WT-p21CIP1 in different lines of Cdk5-overexpressing cells (Fig. 2I). As Cdk2 is also responsible to p21CIP1 Ser130 phosphorylation (25), Cdk5/p21CIP1 regulation was monitored in the condition of Cdk2 knockdown. The results showed that Cdk5 still significantly reduced p21CIP1 protein levels while Cdk2 was knocked down (Fig. 3A). The effects on cell growth were correspondent to the levels of p21CIP1 protein (Fig. 3B). Supplementary Fig. S6A and S6B showed the efficiency of different shRNA clones for Cdk5 and Cdk2. Altogether, by using either Cdk5 knockdown or SA-p21CIP1 to inhibit Cdk5-dependent phosphorylation, we reveal that Ser130 phosphorylation by Cdk5 is essential for p21CIP1 protein degradation in a Cdk2-independent manner.

Figure 3.

Cdk5 regulation on p21CIP1 protein is Cdk2-independent. A, Overexpression or knockdown of Cdk5 or Cdk2 was performed in MCF-7 cells and the protein levels are shown. Actin served as the internal control. The quantitative results are provided. B, The growth of MCF-7 cells with the same treatments of A are shown. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups.

Figure 3.

Cdk5 regulation on p21CIP1 protein is Cdk2-independent. A, Overexpression or knockdown of Cdk5 or Cdk2 was performed in MCF-7 cells and the protein levels are shown. Actin served as the internal control. The quantitative results are provided. B, The growth of MCF-7 cells with the same treatments of A are shown. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups.

Close modal

Cdk5 primarily targets nuclear p21CIP1

The results of cellular protein fractionation revealed that Cdk5 primarily repressed p21CIP1 protein in the nucleus of cancer cells (Fig. 4A). Oppositely, Cdk5 knockdown increased nuclear p21CIP1 protein (Supplementary Fig. S7). The nuclear distribution of endogenous p21CIP1 was significantly diminished in WT Cdk5-, but not in kinase-dead (KD) Cdk5-, overexpressing cells (Fig. 4B). Treatments of MG132 or lactacystin restored the WT-Cdk5-induced degradation of nuclear p21CIP1 protein (Fig. 4B and Supplementary Fig. S4B). The results of immunostaining showed that inhibition of Cdk5-dependent p21CIP1 phosphorylation using KD-Cdk5 or the SA-p21CIP1 mutant stabilized nuclear p21CIP1 levels (top, Fig. 4C; bottom for the quantified results). In clinical study, the tissue array data and the representative images of three cancer types showed the inverse correlation between nuclear distributions of Cdk5 and p21CIP1 (in total 397 patients, P < 0.001 for prostate cancer; P = 0.006 for lung cancer; P < 0.001 for breast cancer, Fig. 4D and E). Altogether, these results suggest that Cdk5 primarily targets the nuclear p21CIP1 protein through Ser130 phosphorylation-related degradation.

Figure 4.

Cdk5-dependent Ser130 phosphorylation contributes to nuclear p21CIP1 protein reduction. A, Empty vector (EV) or wild-type Cdk5 (WT) was overexpressed in MCF-7 cells and cellular protein fractionation and following Western blotting was performed for Cdk5 and p21CIP1 proteins. B, MCF-7 cells stably expressing empty-vector (EV), wild-type Cdk5 (WT), and kinase-dead Cdk5 mutant (KD) were treated with or without MG132 (5 μmol/L, 6 hours) before nuclear protein extraction and the levels of Cdk5 and p21CIP1 as well as input proteins are shown. PARP and α-tubulin served as markers of the nuclear and cytoplasmic fractions, respectively. Actin and α-tubulin served as internal controls. The quantitative results are provided with the related images. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups. C, COS-1 cells were transfected with WT-Cdk5, KD-Cdk5, FLAG-WT-p21CIP1, or FLAG-SA-p21CIP1 for 2 days. The subcellular localizations of Cdk5 and p21CIP1 were visualized by immunocytochemistry (Cdk5, green signal; p21CIP1, red signal). Nuclei were stained with DAPI (blue signal). Scale bar, 25 μm. The quantitative results are shown at the bottom as means ± SEM. **, P < 0.01 versus WT-Cdk5 + FLAG-WT-p21CIP1 group. D, The immunohistochemical images of nuclear Cdk5 and nuclear p21CIP1 proteins (brown signals) from six representative patients of three cancer types of tissue arrays are shown. Cell nuclei were stained by hematoxylin (blue signals). Scale bar, 50 μm. Insets show the larger magnification. E, Tissue arrays of three cancer types were evaluated for the correlation between nuclear distributions of Cdk5 and p21CIP1. The p21CIP1 nuclear distributions were categorized into the positive (blue) or negative (orange) groups corresponding to Cdk5 expression status. The statistical results were analyzed by the Fisher exact test and are indicated in the panel.

Figure 4.

Cdk5-dependent Ser130 phosphorylation contributes to nuclear p21CIP1 protein reduction. A, Empty vector (EV) or wild-type Cdk5 (WT) was overexpressed in MCF-7 cells and cellular protein fractionation and following Western blotting was performed for Cdk5 and p21CIP1 proteins. B, MCF-7 cells stably expressing empty-vector (EV), wild-type Cdk5 (WT), and kinase-dead Cdk5 mutant (KD) were treated with or without MG132 (5 μmol/L, 6 hours) before nuclear protein extraction and the levels of Cdk5 and p21CIP1 as well as input proteins are shown. PARP and α-tubulin served as markers of the nuclear and cytoplasmic fractions, respectively. Actin and α-tubulin served as internal controls. The quantitative results are provided with the related images. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups. C, COS-1 cells were transfected with WT-Cdk5, KD-Cdk5, FLAG-WT-p21CIP1, or FLAG-SA-p21CIP1 for 2 days. The subcellular localizations of Cdk5 and p21CIP1 were visualized by immunocytochemistry (Cdk5, green signal; p21CIP1, red signal). Nuclei were stained with DAPI (blue signal). Scale bar, 25 μm. The quantitative results are shown at the bottom as means ± SEM. **, P < 0.01 versus WT-Cdk5 + FLAG-WT-p21CIP1 group. D, The immunohistochemical images of nuclear Cdk5 and nuclear p21CIP1 proteins (brown signals) from six representative patients of three cancer types of tissue arrays are shown. Cell nuclei were stained by hematoxylin (blue signals). Scale bar, 50 μm. Insets show the larger magnification. E, Tissue arrays of three cancer types were evaluated for the correlation between nuclear distributions of Cdk5 and p21CIP1. The p21CIP1 nuclear distributions were categorized into the positive (blue) or negative (orange) groups corresponding to Cdk5 expression status. The statistical results were analyzed by the Fisher exact test and are indicated in the panel.

Close modal

Cdk5-related p21CIP1 reduction and Cdk2 activation contribute to cancer cell proliferation

Cdk5 overexpression significantly promoted the growth of 22Rv1 cells with the reduction of p21CIP1 protein levels (Fig. 5A). In Fig. 5B, p21CIP1 significantly inhibited cell growth (bar 3 vs. bar 1), and Cdk5 overexpression rescued p21CIP1-dependent growth inhibition (bar 4 vs. bar 3; Supplementary Fig. S8 for protein levels). However, Cdk5 overexpression was unable to rescue the growth inhibition induced by SA-p21CIP1 (bar 5 vs. bar 3). These results imply that Cdk5 promotes cancer cell growth by eliminating p21CIP1 protein. Moreover, Cdk5 overexpression in a MDA-MB-231 xenograft model also showed a significant increase of tumor growth (Fig. 5C) with the reduction of p21CIP1 proteins, whereas p27KIP1 was unaffected (top, Fig. 5D; bottom for the quantified data). In synchronized MCF-7 stable cells, WT-Cdk5 but not KD-Cdk5 decreased p21CIP1 protein levels primarily in S-phase (Fig. 5E). As Cdk2 activity is important to the progression of S-phase and regulated by p21CIP1 (26), we observed that Cdk2 was activated in WT-Cdk5-overexpressing cells (but not in KD-Cdk5-overexpressing cells) when p21CIP1 was targeted (Fig. 5F). As the activated Cdk2 interacts with DNA polymerase δ to promote the process of DNA replication, the protein interaction between Cdk2 and DNA polymerase δ was significantly enhanced by Cdk5 overexpression (Fig. 5G). Our findings suggest that Cdk5 might expedite S-phase progression and promote cell growth by degrading nuclear p21CIP1 protein and subsequently activating Cdk2.

Figure 5.

Cdk5 promotes cancer cell growth by targeting p21CIP1 and subsequent Cdk2 activation. A, Growth curves of 22Rv1 cells stably expressing empty vector (EV) or wild-type Cdk5 (WT) are shown (n = 4), with the levels of related proteins at the bottom. B, The four-day proliferation of MCF-7 cells transfected with indicated constructs is presented as the means ± SEM. The levels of related proteins are shown in Supplementary Fig. S8. C, The growth curves of MDA-MB-231 xenograft tumors expressing empty vector (EV, n = 9) or wild-type Cdk5 (WT, n = 10) are shown. The levels of related proteins in tumors are shown in D. Asterisks, the nonspecific signals. E, MCF-7 cells stably transfected with WT- or KD-Cdk5 were synchronized at the G1, S, and G2–M phases as described in Materials and Methods. Protein levels are shown. F, The in vitro kinase assays of Cdk5 and Cdk2 in different MCF-7–stable cell lines (EV, WT-Cdk5, and KD-Cdk5) were performed after immunoprecipitation (IP), while histone H1 was used as the substrate. The related input proteins for immunoprecipitation are shown at the bottom. G, Protein interaction between Cdk2 and DNA polymerase δ in MCF-7–stable cell lines was analyzed by Cdk2 immunoprecipitation, followed by Western blotting. The related input proteins are shown and α-tubulin and actin served as internal controls. The quantitative results are provided with the related images. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups.

Figure 5.

Cdk5 promotes cancer cell growth by targeting p21CIP1 and subsequent Cdk2 activation. A, Growth curves of 22Rv1 cells stably expressing empty vector (EV) or wild-type Cdk5 (WT) are shown (n = 4), with the levels of related proteins at the bottom. B, The four-day proliferation of MCF-7 cells transfected with indicated constructs is presented as the means ± SEM. The levels of related proteins are shown in Supplementary Fig. S8. C, The growth curves of MDA-MB-231 xenograft tumors expressing empty vector (EV, n = 9) or wild-type Cdk5 (WT, n = 10) are shown. The levels of related proteins in tumors are shown in D. Asterisks, the nonspecific signals. E, MCF-7 cells stably transfected with WT- or KD-Cdk5 were synchronized at the G1, S, and G2–M phases as described in Materials and Methods. Protein levels are shown. F, The in vitro kinase assays of Cdk5 and Cdk2 in different MCF-7–stable cell lines (EV, WT-Cdk5, and KD-Cdk5) were performed after immunoprecipitation (IP), while histone H1 was used as the substrate. The related input proteins for immunoprecipitation are shown at the bottom. G, Protein interaction between Cdk2 and DNA polymerase δ in MCF-7–stable cell lines was analyzed by Cdk2 immunoprecipitation, followed by Western blotting. The related input proteins are shown and α-tubulin and actin served as internal controls. The quantitative results are provided with the related images. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups.

Close modal

Cdk2 is important to Cdk5-promoted tumor growth

Cdk5 knockdown significantly retarded tumor growth in a 22Rv1 xenograft model (Fig. 6A) with p21CIP1 protein elevation (Fig. 6B). To clarify the involvement of Cdk2 in Cdk5-regulated tumor growth, Cdk2 knockdown was performed in the xenograft model while Cdk5 was overexpressed. The results showed that Cdk2 knockdown significantly blocked the Cdk5-promoted tumor growth (Fig. 6C and D for the correspondent protein levels). The significant efficiency of Cdk2 knockdown was shown in Supplementary Fig. S6C. It suggests the important role of Cdk2 in the downstream signaling to Cdk5. Besides, Cdk5 still increased tumor growth with Cdk2 knockdown (Fig. 6C), implying that Cdk5 might promote tumor growth through a Cdk2-independent pathway, which echoed to the authors' previous reports, for example, STAT3 (7, 9). Finally, we hypothesize that Cdk5 targets nuclear p21CIP1 protein and leads to downstream Cdk2 activation. Therefore the transition of G1–S phase in cell cycle is promoted and tumor growth is accelerated (Fig. 6E).

Figure 6.

Cdk2 is important to Cdk5-promoted tumor growth. A, The growth curves of 22Rv1 xenograft tumors expressing mock (n = 6) or shCdk5 (n = 8) are shown. The levels of related proteins in tumors are shown in B. C, The growth curves of 22Rv1 xenograft tumors expressing Cdk5 (n = 6), Cdk5+shCdk2 (n = 8), or shCdk2 (n = 9) are shown. The control curve is identical to the mock curve in Fig. 6A. The levels of related proteins in tumors are shown in D. Actin served as the internal control. The quantitative results are provided with the related images. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups. E, The scheme illustrates that high Cdk5 expression directly and rapidly causes nuclear p21CIP1 degradation and subsequently promotes Cdk2 activation and cancer cell growth.

Figure 6.

Cdk2 is important to Cdk5-promoted tumor growth. A, The growth curves of 22Rv1 xenograft tumors expressing mock (n = 6) or shCdk5 (n = 8) are shown. The levels of related proteins in tumors are shown in B. C, The growth curves of 22Rv1 xenograft tumors expressing Cdk5 (n = 6), Cdk5+shCdk2 (n = 8), or shCdk2 (n = 9) are shown. The control curve is identical to the mock curve in Fig. 6A. The levels of related proteins in tumors are shown in D. Actin served as the internal control. The quantitative results are provided with the related images. The data are presented as the means ± SEM. *, P < 0.05; **, P < 0.01 versus the related control groups. E, The scheme illustrates that high Cdk5 expression directly and rapidly causes nuclear p21CIP1 degradation and subsequently promotes Cdk2 activation and cancer cell growth.

Close modal

By using in vitro, in vivo, and clinical evidence, this study explored the oncogenic role of Cdk5 in a high-expressing status might cause the degradation of the cell-cycle inhibitor p21CIP1 in nucleus by Ser130 phosphorylation, subsequently activate Cdk2, and finally promote cancer growth. Therefore Cdk5 might play an oncogenic role in cancer biology and potentially exhibited as a future diagnostic marker. Cdk5 targeting could also become a future strategy for cancer treatment.

p21CIP1 expression and protein stability are tightly regulated by multiple mechanisms and contributed to cell-cycle control. The phosphorylation regulation on p21CIP1 protein controls its protein subcellular localization, partner binding affinity, and protein stability (27). This study demonstrates that p21CIP1 was a novel target of Cdk5 phosphorylation and followed by the decrease of p21CIP1 protein stability primarily in the nucleus of cancer cells (Figs. 1–4). However, the Ser130 residue of p21CIP1 can be phosphorylated by several kinases under different conditions for discrepant effects. ERK2 phosphorylates p21CIP1 at Thr57/Ser130 and induces cytoplasmic localization for protein degradation (28). Cdk2 phosphorylates p21CIP1 at Ser130 in vitro and recruits the E3 ligase complex, which contributes to p21CIP1 degradation and the release from cell-cycle inhibition (25). Pozo and colleagues found that KD-Cdk5 decreased p21CIP1 expression in medullary thyroid cancer cells, which belong to a category of neuroendocrine cancer that has distinct features compared with other types of cancer (29). Cdk5 plays unique roles in both tumor cells and neuronal cells, in which Cdk5 regulates the opposite functions of growth and differentiation, respectively. In addition to expression regulation, the issue of p21CIP1 subcellular localization under Cdk2 regulation was not commented. On the other hand, Cdk6 phosphorylation of p21CIP1 at Ser130 has minor effects on p21CIP1 protein stability or localization and fails to interact with (or inhibits) Cdk2 (30). TGFβ1 activates the stress-related JNK1 and p38α pathway and stabilizes p21CIP1 protein through Ser130 phosphorylation (31). Thus, the regulations of p21CIP1 under distinct circumstances would promote its unclear roles instead of a cell-cycle inhibitor, and the protein stability and subcellular localization of p21CIP1 regulated by kinases remain controversial. Under the inhibition of proteasomal degradation or the SA-p21CIP1 mutant (Fig. 4B and C; Supplementary Fig. S4 and S9), we showed that p21CIP1 protein was unaffected by Cdk5 and primarily localized in the nucleus of cancer cells. These results suggest that Cdk5 primarily targets nuclear p21CIP1 and contributes to cell-cycle progression, whereas cytoplasmic transport of p21CIP1 is not correlated with Cdk5 regulation.

It has been reported that p21CIP1 directly binds to the C8 subunit of the 20S proteasome and is turned over without ubiquitination-mediated targeting (32). This ubiquitination-independent degradation of p21CIP1 is accelerated by oncogenic proteins or phosphorylation on specific sites (33–35). Depletion of the 20S proteasome activator REGγ leads to the increase of nuclear p21CIP1 protein and alters the cell-cycle progression of thyroid carcinoma cells (36). Notably, the structure of the 20S proteasome contains a nuclear localization signal and can be trapped by the nuclear envelope (37). The 20S proteasome returns to the cytoplasm after nuclear envelop breakdown during mitosis and this nuclear translocation is frequently observed during the cell-cycle progression of cancerous cells and may be required for the degradation of some nuclear proteins (37). In our study, although Cdk5 degrades p21CIP1 protein via proteasomal degradation in cancer cells (Fig. 1H and Supplementary Fig. S4), no change of ubiquitinated-p21CIP1 was observed (Supplementary Fig. S10). As Bornstein and colleagues claimed that ubiquitination is not necessary after p21CIP1 phosphorylation (25), the 20S proteasome-mediated proteolysis is possibly involved in Cdk5-triggered nuclear p21CIP1 degradation in cancer cells.

Lindqvist and colleagues (38) and Zhang and colleagues (39) back-to-back raised the evidence showing the indirect regulation between Cdk5 and p21CIP1 with the connections of Akt activation and p53 downregulation under Cdk5 knockdown status, which is considerable as a cellular stress compared to high Cdk5 expression in cancer patients' tissues. Notably, our data indicates that Cdk5 still significantly reduced p21CIP1 protein under Akt inhibition (Supplementary Fig. S11). With respect to subcellular localization of p21CIP1, Hung and colleagues found that Her2 induces cytosolic localization of p21CIP1 to play anti-apoptotic role through activations of PI3K/Akt (19) and MDM2/p53 (40). Li and colleagues identified that Akt phosphorylates p21CIP1 and enhances its protein stability to promote anti-apoptosis and cell survival (41). While Cdk5 has chances to individually regulate Her2 (42), PI3K/Akt (38), and p53 (43), the network among these proteins with the issue of p21CIP1 subcellular localization becomes complicated to illustrate in one simple hypothesis. Therefore, we simply propose one direct and rapid regulation between Cdk5 (in a high expression status) and p21CIP1 in fast-growing tumors, which is independent to the findings under different pathological events (e.g., cellular stress). We believe that these findings provide a new and important perspective to the whole picture of Cdk5 regulation on p21CIP1 in cell biology.

p21CIP1 nuclear localization has also been reported to inhibit Cdk activation and PCNA function in response to DNA damage-induced cell-cycle arrest at G1–S transition (15). Harper and colleagues indicate that p21CIP1 effectively inhibits the activities of Cdk2, Cdk3, Cdk4, and Cdk6 (Ki 0.5–15 nmol/L) but poorly to Cdk5 (Ki >2 μmol/L; ref. 44). In our in vitro analysis, recombinant p21CIP1 protein significantly reduced the phosphorylation of histone H1, which is an index of Cdk5 activity (Fig. 2F). However, p21CIP1 Ser130 phosphorylation was increased opposite to the decrease of histone H1 phosphorylation, suggesting that the role of Cdk5 substrate in the reaction was from the original histone H1 shifted to the increasing p21CIP1. This finding is consistent with Harper and colleagues (44) and suggests that Cdk5 phosphorylates the substrate p21CIP1 instead of histone H1 in vitro. However, Hengst and colleagues demonstrate that a single p21CIP1 molecule is sufficient for Cdk2 inhibition and p21CIP1-saturated complex contains only one stably bound inhibitor molecule (26), suggesting that tiny change of cellular p21CIP1 protein level is critical to the regulation of cell-cycle progression. In our study, the Cdk5-reduced p21CIP1 was observed in early S-phase after hydroxyurea treatment (Fig. 5E), suggesting that Cdk5, unlike other Cdks, helps cells escape from p21CIP1 inhibition on Cdk2 by phosphorylation-dependent degradation and therefore contributes to promote S-phase progression. Interestingly, we found that the protein interaction between KD-Cdk5 and p21CIP1 were increased (Supplementary Fig. S12). The interaction between SA-p21CIP1 mutant and Cdk5 in the nucleus were also increased (Supplementary Fig. S13). These results suggest that Cdk5 might be capable of trapping p21CIP1 and relieving the inhibition of Cdk2 even when Cdk5 was inactive or p21CIP1 phosphorylation was absent. This finding leads us to conclude that Cdk5 counters the inhibitory function on p21CIP1 in cell-cycle control through activation-dependent or -independent regulations. p35 and p39 have been believed as two major physiologic regulators for Cdk5 kinase activity in recent literature. The authors previously found that Cdk5 and p35 can be two independent factors regulating the downstream Ser727 STAT3 phosphorylation and the growth of prostate cancer cells by in vitro cell culture experiments as well as clinical analysis (9).This prospect potentially highlights Cdk5 as an independent marker in future clinical analysis irrespective of the status of its regulators.

In conclusion, we had presented in vitro, in vivo, and clinical evidence demonstrating that Cdk5, as a novel and direct p21CIP1 regulator, decreases nuclear p21CIP1 protein through phosphorylation/proteasome-dependent degradation. Consistent with other recent evidence (1, 2, 5, 8, 9, 13, 14, 45, 46), we provide a new model that the aberrant elevation of Cdk5 expression in malignancy might contribute to cancer growth by promoting escape of p21CIP1 inhibition and Cdk2 activation while more work is needed to confirm the role of Cdk2 in this process (Fig. 6E). These findings shed light on the roles of Cdk5 in future cancer research for either diagnosis or therapy.

No potential conflicts of interest were disclosed.

Conception and design: P.-H. Huang, H. Lin

Development of methodology: P.-H. Huang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): P.-H. Huang, C.-H. Chang, Y.-C. Wang, C.-H. Lai, J.-H. Wang, E. Lin

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P.-H. Huang, M.-C. Chen, W.-H. Kao, C.-H. Chang, C.-H. Yue, H. Lin

Writing, review, and/or revision of the manuscript: P.-H. Huang, M.-C. Chen, W.-H. Kao, J.-T. Hsieh, H. Lin

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P.-H. Huang, Y.-C. Wang, J.-T. Hsieh, Y.-T. Lee, H.-Y. Wang, H. Lin

Study supervision: H. Lin

Other (performed the animal experiments): Y.-T. Peng

Other (conducted some preliminary results): S.-C. You

The authors thank Dr. H. C. Chen, Dr. C. M. Chen, Dr. C. M. Hsueh, and Dr. L. C. Lin (National Chung Hsing University, Taiwan) for help with animal experiments; Dr. H. C. Hung, Dr. Y. L. Liu, Dr. H. C. Cheng, Dr. J. W. Chen, Dr. H. L. Su, and Dr. C. C. Lai (National Chung Hsing University, Taiwan) for technical supports; Dr. Z. F. Chang (National Yang Ming university, Taiwan), Dr. H. Nishitani (University of Hyogo, Japan), and Dr. D. J. Mann (Imperial College London, UK) for providing p21CIP1 constructs; Dr. M. T. Lai (Chang Bing Show Chawn Memorial Hospital, Taiwan) for instructing the clinical analysis.

This work was supported by The Ministry of Science and Technology, Taiwan (101-2320-B005-004-MY3 and 103-2911-I-005-507 to H. Lin); joint grant of Taichung Veterans General Hospital and National Chung Hsing University (TCVGH-NCHU1047604 to H. Lin); and Chang Bing Show Chwan Memorial Hospital, Taiwan (CBSH9905001 to E. Lin; CBBC001 to Y.-T. Lee).

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.

1.
Levacque
Z
,
Rosales
JL
,
Lee
KY
. 
Level of cdk5 expression predicts the survival of relapsed multiple myeloma patients
.
Cell Cycle
2012
;
11
:
4093
5
.
2.
Liu
JL
,
Wang
XY
,
Huang
BX
,
Zhu
F
,
Zhang
RG
,
Wu
G
. 
Expression of CDK5/p35 in resected patients with non-small cell lung cancer: relation to prognosis
.
Med Oncol
2011
;
28
:
673
8
.
3.
Choi
HS
,
Lee
Y
,
Park
KH
,
Sung
JS
,
Lee
JE
,
Shin
ES
, et al
Single-nucleotide polymorphisms in the promoter of the CDK5 gene and lung cancer risk in a Korean population
.
J Hum Genet
2009
;
54
:
298
303
.
4.
Eggers
JP
,
Grandgenett
PM
,
Collisson
EC
,
Lewallen
ME
,
Tremayne
J
,
Singh
PK
, et al
Cyclin-dependent kinase 5 is amplified and overexpressed in pancreatic cancer and activated by mutant K-Ras
.
Clin Cancer Res
2011
;
17
:
6140
50
.
5.
Strock
CJ
,
Park
JI
,
Nakakura
EK
,
Bova
GS
,
Isaacs
JT
,
Ball
DW
, et al
Cyclin-dependent kinase 5 activity controls cell motility and metastatic potential of prostate cancer cells
.
Cancer Res
2006
;
66
:
7509
15
.
6.
Feldmann
G
,
Mishra
A
,
Hong
SM
,
Bisht
S
,
Strock
CJ
,
Ball
DW
, et al
Inhibiting the cyclin-dependent kinase CDK5 blocks pancreatic cancer formation and progression through the suppression of Ras-Ral signaling
.
Cancer Res
2010
;
70
:
4460
9
.
7.
Lin
H
,
Chen
MC
,
Chiu
CY
,
Song
YM
,
Lin
SY
. 
Cdk5 regulates STAT3 activation and cell proliferation in medullary thyroid carcinoma cells
.
J Biol Chem
2007
;
282
:
2776
84
.
8.
Hsu
FN
,
Chen
MC
,
Chiang
MC
,
Lin
E
,
Lee
YT
,
Huang
PH
, et al
Regulation of androgen receptor and prostate cancer growth by cyclin-dependent kinase 5
.
J Biol Chem
2011
;
286
:
33141
9
.
9.
Hsu
FN
,
Chen
MC
,
Lin
KC
,
Peng
YT
,
Li
PC
,
Lin
E
, et al
Cyclin-dependent kinase 5 modulates STAT3 and androgen receptor activation through phosphorylation of Ser(7)(2)(7) on STAT3 in prostate cancer cells
.
Am J Physiol Endocrinol Metab
2013
;
305
:
E975
86
.
10.
Chen
MC
,
Huang
CY
,
Hsu
SL
,
Lin
E
,
Ku
CT
,
Lin
H
, et al
Retinoic acid induces apoptosis of prostate cancer DU145 cells through Cdk5 overactivation
.
Evid Based Complement Alternat Med
2012
;
2012
:
580736
.
11.
Kuo
HS
,
Hsu
FN
,
Chiang
MC
,
You
SC
,
Chen
MC
,
Lo
MJ
, et al
The role of Cdk5 in retinoic acid-induced apoptosis of cervical cancer cell line
.
Chin J Physiol
2009
;
52
:
23
30
.
12.
Lin
H
,
Juang
JL
,
Wang
PS
. 
Involvement of Cdk5/p25 in digoxin-triggered prostate cancer cell apoptosis
.
J Biol Chem
2004
;
279
:
29302
7
.
13.
Rea
K
,
Sensi
M
,
Anichini
A
,
Canevari
S
,
Tomassetti
A
. 
EGFR/MEK/ERK/CDK5-dependent integrin-independent FAK phosphorylated on serine 732 contributes to microtubule depolymerization and mitosis in tumor cells
.
Cell Death Dis
2013
;
4
:
e815
.
14.
Pozo
K
,
Castro-Rivera
E
,
Tan
C
,
Plattner
F
,
Schwach
G
,
Siegl
V
, et al
The role of Cdk5 in neuroendocrine thyroid cancer
.
Cancer Cell
2013
;
24
:
499
511
.
15.
Sherr
CJ
,
Roberts
JM
. 
CDK inhibitors: positive and negative regulators of G1-phase progression
.
Genes Dev
1999
;
13
:
1501
12
.
16.
Martin-Caballero
J
,
Flores
JM
,
Garcia-Palencia
P
,
Serrano
M
. 
Tumor susceptibility of p21(Waf1/Cip1)-deficient mice
.
Cancer Res
2001
;
61
:
6234
8
.
17.
Cmielova
J
,
Rezacova
M
. 
p21Cip1/Waf1 protein and its function based on a subcellular localization
.
J Cell Biochem
2011
;
112
:
3502
6
.
18.
Abbas
T
,
Dutta
A
. 
p21 in cancer: intricate networks and multiple activities
.
Nat Rev Cancer
2009
;
9
:
400
14
.
19.
Zhou
BP
,
Liao
Y
,
Xia
W
,
Spohn
B
,
Lee
MH
,
Hung
MC
. 
Cytoplasmic localization of p21Cip1/WAF1 by Akt-induced phosphorylation in HER-2/neu-overexpressing cells
.
Nat Cell Biol
2001
;
3
:
245
52
.
20.
Nishitani
H
,
Shiomi
Y
,
Iida
H
,
Michishita
M
,
Takami
T
,
Tsurimoto
T
. 
CDK inhibitor p21 is degraded by a proliferating cell nuclear antigen-coupled Cul4-DDB1Cdt2 pathway during S phase and after UV irradiation
.
J Biol Chem
2008
;
283
:
29045
52
.
21.
Huang
PH
,
Huang
CY
,
Chen
MC
,
Lee
YT
,
Yue
CH
,
Wang
HY
, et al
Emodin and aloe-emodin suppress breast cancer cell proliferation through ER alpha inhibition
.
Evid Based Complement Alternat Med
2013
;
2013
:
376123
.
22.
Lin
H
,
Lin
TY
,
Juang
JL
. 
Abl deregulates Cdk5 kinase activity and subcellular localization in Drosophila neurodegeneration
.
Cell Death Differ
2007
;
14
:
607
15
.
23.
Bhatt
S
,
Xiao
Z
,
Meng
Z
,
Katzenellenbogen
BS
. 
Phosphorylation by p38 mitogen-activated protein kinase promotes estrogen receptor alpha turnover and functional activity via the SCF(Skp2) proteasomal complex
.
Mol Cell Biol
2012
;
32
:
1928
43
.
24.
Dalvai
M
,
Bystricky
K
. 
Cell cycle and anti-estrogen effects synergize to regulate cell proliferation and ER target gene expression
.
PloS One
2010
;
5
:
e11011
.
25.
Bornstein
G
,
Bloom
J
,
Sitry-Shevah
D
,
Nakayama
K
,
Pagano
M
,
Hershko
A
. 
Role of the SCFSkp2 ubiquitin ligase in the degradation of p21Cip1 in S phase
.
J Biol Chem
2003
;
278
:
25752
7
.
26.
Hengst
L
,
Gopfert
U
,
Lashuel
HA
,
Reed
SI
. 
Complete inhibition of Cdk/cyclin by one molecule of p21(Cip1)
.
Genes Dev
1998
;
12
:
3882
8
.
27.
Child
ES
,
Mann
DJ
. 
The intricacies of p21 phosphorylation: protein/protein interactions, subcellular localization and stability
.
Cell Cycle
2006
;
5
:
1313
9
.
28.
Hwang
CY
,
Lee
C
,
Kwon
KS
. 
Extracellular signal-regulated kinase 2-dependent phosphorylation induces cytoplasmic localization and degradation of p21Cip1
.
Mol Cell Biol
2009
;
29
:
3379
89
.
29.
Pozo
K
,
Hillmann
A
,
Augustyn
A
,
Plattner
F
,
Hai
T
,
Singh
T
, et al
Differential expression of cell cycle regulators in CDK5-dependent medullary thyroid carcinoma tumorigenesis
.
Oncotarget
2015
;
6
:
12080
93
.
30.
Jarviluoma
A
,
Child
ES
,
Sarek
G
,
Sirimongkolkasem
P
,
Peters
G
,
Ojala
PM
, et al
Phosphorylation of the cyclin-dependent kinase inhibitor p21Cip1 on serine 130 is essential for viral cyclin-mediated bypass of a p21Cip1-imposed G1 arrest
.
Mol Cell Biol
2006
;
26
:
2430
40
.
31.
Kim
GY
,
Mercer
SE
,
Ewton
DZ
,
Yan
Z
,
Jin
K
,
Friedman
E
. 
The stress-activated protein kinases p38 alpha and JNK1 stabilize p21(Cip1) by phosphorylation
.
J Biol Chem
2002
;
277
:
29792
802
.
32.
Touitou
R
,
Richardson
J
,
Bose
S
,
Nakanishi
M
,
Rivett
J
,
Allday
MJ
. 
A degradation signal located in the C-terminus of p21WAF1/CIP1 is a binding site for the C8 alpha-subunit of the 20S proteasome
.
EMBO J
2001
;
20
:
2367
75
.
33.
Wang
B
,
Liu
K
,
Lin
HY
,
Bellam
N
,
Ling
S
,
Lin
WC
. 
14–3-3Tau regulates ubiquitin-independent proteasomal degradation of p21, a novel mechanism of p21 downregulation in breast cancer
.
Mol Cell Biol
2010
;
30
:
1508
27
.
34.
Xu
H
,
Zhang
Z
,
Li
M
,
Zhang
R
. 
MDM2 promotes proteasomal degradation of p21Waf1 via a conformation change
.
J Biol Chem
2010
;
285
:
18407
14
.
35.
Chen
X
,
Zhang
J
,
Zhang
M
,
Liu
S
,
Yan
W
,
Jung
J
, et al
Serine 123 phosphorylation modulates p21 protein stability and activity by suppressing ubiquitin-independent proteasomal degradation
.
J Biol Chem
2012
;
287
:
34410
8
.
36.
Li
X
,
Amazit
L
,
Long
W
,
Lonard
DM
,
Monaco
JJ
,
O'Malley
BW
. 
Ubiquitin- and ATP-independent proteolytic turnover of p21 by the REGgamma-proteasome pathway
.
Mol Cell
2007
;
26
:
831
42
.
37.
Wojcik
C
,
DeMartino
GN
. 
Intracellular localization of proteasomes
.
Int J Biochem Cell Biol
2003
;
35
:
579
89
.
38.
Lindqvist
J
,
Imanishi
SY
,
Torvaldson
E
,
Malinen
M
,
Remes
M
,
Orn
F
, et al
Cyclin-dependent kinase 5 acts as a critical determinant of AKT-dependent proliferation and regulates differential gene expression by the androgen receptor in prostate cancer cells
.
Mol Biol Cell
2015
;
26
:
1971
84
.
39.
Zhang
S
,
Lu
Z
,
Mao
W
,
Ahmed
AA
,
Yang
H
,
Zhou
J
, et al
CDK5 regulates paclitaxel sensitivity in ovarian cancer cells by modulating AKT activation, p21Cip1- and p27Kip1-Mediated G1 cell cycle arrest and apoptosis
.
PLoS ONE
2015
;
10
:
e0131833
.
40.
Zhou
BP
,
Liao
Y
,
Xia
W
,
Zou
Y
,
Spohn
B
,
Hung
MC
. 
HER-2/neu induces p53 ubiquitination via Akt-mediated MDM2 phosphorylation
.
Nat Cell Biol
2001
;
3
:
973
82
.
41.
Li
Y
,
Dowbenko
D
,
Lasky
LA
. 
AKT/PKB phosphorylation of p21Cip/WAF1 enhances protein stability of p21Cip/WAF1 and promotes cell survival
.
J Biol Chem
2002
;
277
:
11352
61
.
42.
Li
BS
,
Ma
W
,
Jaffe
H
,
Zheng
Y
,
Takahashi
S
,
Zhang
L
, et al
Cyclin-dependent kinase-5 is involved in neuregulin-dependent activation of phosphatidylinositol 3-kinase and Akt activity mediating neuronal survival
.
J Biol Chem
2003
;
278
:
35702
9
.
43.
Zhang
J
,
Krishnamurthy
PK
,
Johnson
GV
. 
Cdk5 phosphorylates p53 and regulates its activity
.
J Neurochem
2002
;
81
:
307
13
.
44.
Harper
JW
,
Elledge
SJ
,
Keyomarsi
K
,
Dynlacht
B
,
Tsai
LH
,
Zhang
P
, et al
Inhibition of cyclin-dependent kinases by p21
.
Mol Biol Cell
1995
;
6
:
387
400
.
45.
Liang
Q
,
Li
L
,
Zhang
J
,
Lei
Y
,
Wang
L
,
Liu
DX
, et al
CDK5 is essential for TGF-beta1-induced epithelial-mesenchymal transition and breast cancer progression
.
Sci Rep
2013
;
3
:
2932
.
46.
Zhang
J
,
Li
H
,
Herrup
K
. 
Cdk5 nuclear localization is p27-dependent in nerve cells: implications for cell cycle suppression and caspase-3 activation
.
J Biol Chem
2010
;
285
:
14052
61
.