Oncogenic KIT or PDGFRA receptor tyrosine kinase mutations are compelling therapeutic targets in gastrointestinal stromal tumor (GIST), and treatment with the KIT/PDGFRA inhibitor imatinib is the standard of care for patients with advanced GIST. Polyclonal emergence of KIT/PDGFRA secondary mutations is the main mechanism of imatinib progression, making it challenging to overcome KIT/PDGFRA-inhibitor resistance. It is unclear whether there are other therapeutic targets in advanced GIST. Using genome-wide transcriptomic profiling of advanced versus early-stage GIST and CRISPR knockout functional screens, we demonstrate that CDK1 is frequently highly expressed in advanced GIST but not in early-stage GIST across three patient cohorts. High expression of CDK1 was associated with malignancy in GIST. CDK1 was critically required for advanced GIST, including imatinib-resistant GIST. CDK1 ablation led to robust proliferation inhibition. A mass spectrometry-based proteomics screen further revealed that AKT is a novel substrate of CDK1 kinase in GIST. CDK1 bound AKT and regulated its phosphorylation, thereby promoting GIST proliferation and progression. Importantly, a pharmacologic inhibitor of CDK1, RO-3306, disrupted GIST cell proliferation in CDK1 highly expressed GIST but not in CDK1-negative GIST cells and nontransformed fibroblast cells. Treatment with RO-3306 reduced tumor growth in both imatinib-resistant and imatinib-sensitive GIST xenograft mouse models. Our findings suggest that CDK1 represents a druggable therapeutic target in GIST and warrants further testing in clinical trials.

Significance:

These findings propose CDK1 as a novel cell-cycle–independent vulnerability in gastrointestinal stromal tumors, representing a new therapeutic opportunity for patients with advanced disease.

Gastrointestinal stromal tumor (GIST) is the most common type of sarcoma that rises from the gastrointestinal tract (1, 2). Activating mutations of receptor tyrosine kinases KIT or PDGFRA are key to the pathogenesis of nearly 85% of GIST (1, 3, 4). Most patients with GIST with a KIT or PDGFRA mutation can be effectively treated with tyrosine kinase inhibitors (TKI), such as imatinib or other agents (1, 5–10), but secondary polyclonal mutations conferring drug resistance frequently emerge, leading to GIST progression (1, 11, 12). Sunitinib and regorafenib, the approved second-line and third-line agents for advanced GIST, are effective. Unfortunately, the median progression-free survival (PFS) with second-line sunitinib is only 6 to 7 months (7). Similarly, regorafenib treatment offers modest clinical benefit after patients have progressed on imatinib and sunitinib with a median PFS not exceeding 6 months (5). Therefore, there is a need for novel therapeutic targets and effective agents to treat patients with advanced GIST.

GIST spans a biological behavior spectrum that ranges from benign to a malignant/advanced state, with some low-risk lesions remaining stable for years whereas others progress rapidly to widespread metastatic disease. Most early-stage/benign GISTs already contain KIT or PDGFRA mutations and yet have exceedingly low potential for malignant progression (13, 14). There is substantial interest in defining the genomic and transcriptomic features of advanced GIST to prevent cancer or identify disease when it is most curable. The evolution of advanced GIST is a complex multistep process (1, 15). Deciphering the molecular changes contribute to the development of advanced GIST may provide insights into GIST biology and therapeutic target identification. We previously reported a genomic characterization of advanced versus early-stage GIST and demonstrated that DMD deletion/inactivation promotes GIST invasion and metastasis (16).

Recent advances in genetic perturbation tools have allowed assessment of wild-type gene function at scale, yielding important insights into tumor biology (17), drug resistance (18), and signal transduction (19). In this study, through genome-wide transcriptomic profiling of advanced versus early-stage GIST and CRISPR functional screening, we found that CDK1 is frequently highly expressed in advanced GIST but not in early-stage GIST. Highly expressed CDK1 is oncogenic, promoting GIST growth and proliferation. We proved evidence that CDK1 bind to AKT and regulates its phosphorylation, thereby promoting GIST proliferation and progression. The CDK1 chemical inhibitor, RO-3306 (20, 21), suppresses both imatinib-resistant and imatinib-sensitive GIST growth in various cell line models and patient-derived mouse xenograft models. Our findings provide a preclinical rationale for targeting CDK1 as a therapeutic strategy for patients with GIST.

Tumor and tissue samples

De-identified snap-frozen tumor biopsies and matched normal samples were from patients with GIST at Ren Ji Hospital Shanghai Jiao Tong University School of Medicine. All samples were collected with institutional review board approval. Informed written consent was obtained from all human participants.

Lentiviral gRNA library essentiality screens

GIST430/654 was transduced with the GeCKOv2 library at a low multiplicity of infection (MOI < 1). Seven days postinfection (and 6 days postpuromycin selection), cells were split into two independent replicate populations of minimum 200-fold library coverage (∼25 million cells). A population-doubling 0 (PD0) sample was collected for genomic DNA extraction. The other replicate population was passaged without selection while maintaining a library fold-coverage of 200× for an additional 15 doublings. Genomic DNA was purified from PD0 and PD15 cell pellets, and guide sequence PCR was amplified with sufficient gDNA to maintain representation and quantified using deep sequencing.

Data processing and analysis

Sequencing reads were processed by counting the number of unique reads for each sgRNA in each experimental condition. sgRNAs with less than 50 counts in the PD0 control sample were removed from downstream analyses. The log2-fold change in abundance of each sgRNA was calculated for final population samples for each of the cell lines after adding a count of one as a pseudocount, and the median of nontargeting controls in the GeCKOv2 library was subtracted from each sgRNA to generate a sgRNA score. Gene-based CRISPR scores (CS) were defined as the median score of all sgRNAs targeting a given gene.

Cell lines and reagents

HEK293T was obtained from the ATCC. GIST-T1 (KIT exon 11: V560_Y578del primary mutation) was generously provided by Dr. Takahiro Taguchi. GIST882, GIST48, GIST430/654, and GIST48B cell lines (RRID:CVCL_M441) were generously provided by professor Jonathan Fletcher at Harvard Medical School. GIST882 (KIT exon 13: K642E primary mutation) was derived from a TKI-naïve metastatic GIST. GIST48 (KIT exon 11: V560D primary mutation + exon 17: D820V activation loop secondary imatinib-resistance mutation) and GIST430/654 (KIT exon 11: V560_L576del primary mutation + exon 13: V654A ATP-binding pocket secondary imatinib-resistance mutation) were established from patients progressing clinically on imatinib. GIST48B was derived from GIST48, cultured with 500 nmol/L 17-AAG (HSP90 inhibitor). The GIST cells were cultured in an IMDM medium (Hyclone, #SH30228.01) with 10% FBS (Thermo Fisher Scientific, #10099141) and penicillin/streptomycin (GIBCO, #15140122). HEK293T was maintained in RPMI1640 medium (Hyclone, #SH30027.01) with 10% FBS and penicillin/streptomycin. All the KIT genotype was confirmed with Sanger sequencing. Cell lines were routinely tested for microbial contamination (including Mycoplasma) and authenticated with KIT sanger sequencing, whole exome sequencing and whole transcriptome sequencing assays. All cells were cultured in a 5% CO2 humidified atmosphere at 37°C. For in vitro cell culture studies, imatinib (Selleck, #S1026), RO-3306 (Selleck, #S7747), BEZ235 (Selleck, #S1009), and MK-2206 (Sigma, #A6730) were dissolved with DMSO. Imatinib was dissolved in saline for in vivo animal studies.

Transcriptome sequencing

The transcriptome sequencing data have been reported previously from our group (22).

Primary tumor cell culture

Tumor tissues were obtained from surgical resection of the patient with GIST tumor samples. Tumor tissues were collected in serum-free IMDM medium and cut into small fragments (5 mm3) with a sterile scalpel or scissors. Add collagenase type I (Gibco, #17100-017) to 50 to 200 U/mL with 3 mmol/L CaCl2 and incubate at 37°C for 6 hours. Disperse cells by passing through a cell strainer. Wash dispersed cells several times by centrifugation in PBS. Seed cells into culture dish containing IMDM media. After then, the cells were cultured in IMDM containing 10% FBS and penicillin–streptomycin mixed solution.

Plasmid constructs and lentivirus production

Two short hairpin RNAs (shRNA) that target human CDK1 or PDK1 were designed using the GPP Web Portal (https://portals.broadinstitute.org/gpp/public/). CDK1 shRNA-1 and shRNA-2 vector `were generated by cloning these two shRNAs into the pLKO.1 vector (Sigma plasmid, #SHC001) individually. ShRandom control was constructed by cloning a shRNA with a random sequence into the pLKO.1 vector. Lentivirus particles were generated by cotransfecting these shRNA constructs with helper virus packaging plasmids pCMVΔ8.9 and pHCMV-VSV-G into HEK293T cells using Lipofectamine 3000 (Invitrogen, #L3000015). Lentivirus were harvested after 24, 36, 48, and 60 hours, and frozen at −80°C in aliquots at appropriate amounts for infection. GIST-T1 and GIST430/654 cells were infected with lentivirus for 16 hours in the supernatant containing 8 μg/mL polybrene (Sigma, #107689), and then treated with puromycin (Millipore, #540411) 1 day after infection.

Cell viability assays

Cells were plated at 1 to 2×104 cells per well in a 96-well flat-bottomed plate. Imatinib and RO-3306 with seven concentrations, 0, 10, 25, 50, 100, 250, 500, and 1,000 nmol/L were respectively added to 96-well plates after 48 hours. Viability studies were performed using the CellTiter-Glo luminescent assay (Promega) after 72 hours (for GIST-T1, GIST48, and GIST48B) or 144h (for GIST430/654, GIST882, GIST-1 primary, and BJ) drug treatment. GIST cell lines transduced with shRNA lentivirus were plated at 2,500 to 10,000 cells per well in 96-well plate. Viability studies were performed once every 3 days. Luminescence was analyzed using a BioTek Gen5 Microplate Readers (BioTek, #H1210-018).

Soft agar assay

Six-well plates were first layered with 0.6% bottom Noble agar (BD Difco, #214220) containing RPMI1640 medium with 10% FBS and penicillin/streptomycin. GIST-T1 cells (5,000 cells per well) and GIST430/654 cells (10,000 cells per well) were transduced with shCtrl and shCDk1 lentivirus and seeded in 0.35% top agar containing 10% FBS and penicillin/streptomycin. Cells were allowed to grow for 4 weeks and then stained with 1 mL of 1 mg/mL methyl thiazol tetrazolium (MTT; Sigma, #M5655) for 3 hours. Colonies were counted by ImageJ software (NIH, RRID:SCR_003070). All assays were performed in triplicate wells, with the entire study replicated three times.

Colony formation assay

Colony formation assay was conducted by seeding GIST-T1 and GIST430/654 cells (500 cells per well) transduced with the shCtrl, shCDK1-1, shCDK1-2 lentivirus into 6-well plates and allowed to grow for 4 weeks. Then, the cells were fixed with 4% paraformaldehyde for 15 minutes and stained with crystal violet solution (Shanghai Sangon Biotechnology Co., #E607309-0100) for 15 minutes. After rinse with distilled water, the colony images were obtained using a scanner (Microtek, Scannermarker 1000XL) and counted by ImageJ software. All assays were carried out in in triplicate wells, with the entire study replicated three times.

Western blotting analysis

Whole cell lysates from cell lines were prepared using lysis buffer (1% NP-40, 50 mmol/L Tris-HCl pH 8.0, 100 mmol/L sodium fluoride, 30 mmol/L sodium pyrophosphate, 2 mmol/L sodium molybdate, 5 mmol/L EDTA, and 2 mmol/L sodium orthovanadate) containing protease inhibitors (10 μg/mL aprotinin, 10 μg/mL leupeptin, and 1 mmol/L phenylmethylsulfonyl fluoride). The lysates were then rocked overnight at 4°C and cleared by centrifugation at 14,000 rpm for 30 minutes at 4°C. The lysate protein concentrations were determined using a Quick Start Bradford 1× Dye Reagent (Bio-Rad, #5000205). Electrophoresis and Western blotting were performed using standard techniques. The hybridization signals were detected by chemiluminescence (Immobilon Western; Millipore Corporation) and captured using an Amersham Imager 600 imagers (GE Healthcare; #29083461, RRID:SCR_000004). The primary antibodies were list as follows: CDK1 for WB (Santa Cruz Technology, #sc-54, RRID:AB_627224), CDK1 for IHC (Santa Cruz Technology, #sc-54), pAKTS473 [Cell Signaling Technology (CST), #9271, RRID:AB_329825], pAKTT308 (CST, #9275, RRID:AB_329828), AKT (CST, #9272), AKT2 (CST, #2962, RRID:AB_329872), PCNA (Santa Cruz Technology, #sc-56, RRID:AB_628110), PARP (CST, #9542, RRID:AB_2160739), PDK1 (CST, #3062, RRID:AB_2236832), and β-actin (Sigma, #A4700, RRID:AB_476730). Relative protein quantification was performed with ImageJ software.

qRT-PCR

Genomic DNA was isolated using QIAamp DNA Mini Kit (Qiagen, #51306). qPCR was performed for target-gene-expression analysis or exon detection of CDK1 using the iQ SYBR Green Supermix (Bio-Rad, #170-8882). Samples were run in triplicate with nontemplate control. Amplification accuracy was verified by melting curve analysis.

SA-β-gal staining

GIST-T1 and GIST430/654 cells were transduced with the Ctrl, CDK1-KD-1, CDK1-KD-2 lentivirus in six-well plates and allowed to grow for 6 days. Cells were fixed in 4% paraformaldehyde and subjected to SA-β-gal staining with Cell Senescence SA-β-Gal Staining Kit (Beyotime, C0602). Random fields were chosen to show SA-β-gal positivity.

Immunoprecipitation and mass spectrometry

Cells were lysed by IP buffer containing protease inhibitors (10 μg/mL leupeptin, 10 μg/mL aprotinin, and 1 mmol/L phenylmethylsulfonyl fluoride), mixed with 1 μg anti-FLAG antibody (Sigma, #F1804, RRID:AB_262044) and 20 μL protein G-sepharose (Thermo Fisher Scientific, #101242), incubated overnight and eluted by boiling with SDS loading buffer. The eluted samples were detected by SDS-PAGE followed by Coomassie staining (Colloidal Blue Staining Kit; Invitrogen, #LC6025). For mass spectrometry, IP samples were eluted by shaking with 8 mol/L urea and 100 mmol/L Tris-Cl (pH 8.0) and analyzed by mass spectrometry.

Coimmunoprecipitation

Cells were lysed by IP buffer containing protease inhibitors (10 μg/mL leupeptin, 10 μg/mL aprotinin, and 1 mmol/L phenylmethylsulfonyl fluoride), mixed with 2 μg anti-Flag antibody and 20 μL protein G-sepharose and incubated overnight. Immunoprecipitates were eluted by boiling with SDS loading buffer. IP samples and whole cell lysates were analyzed by Western blotting.

In vitro binding assay

We first immobilized 0.5 mol/L recombinant AKT1 onto glutathione-Sepharose 4B beads for 60 minutes at 4°C. The immobilized AKT1 with beads was incubated with coimmunoprecipitation (co-IP) lysate for 120 minutes at 4°C. After removal of the supernatant, the beads were washed three times with PEM buffer (100 mmol/L PIPES at pH 6.8, 1 mmol/L EDTA, 0.5 mmol/L MgSO4). Then the samples were boiled in SDS sample buffer and subjected to SDS-PAGE and immunoblot analysis with the indicated antibodies.

Xenograft tumor models

Female BALB/c nude mice (5–6-weeks old; weight 18–25 g) were obtained from Shanghai Lingchang BioTech Co. Ltd. All animals were maintained in the specific pathogen free (SPF) facility of Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences according to the international, national and institutional guidelines for humane animal treatment and complied with relevant legislation. The mouse studies were approved by the Institutional Animal Care and Use Committee (IACUC) of the Shanghai Institutes for Biological Sciences. Prior to implantation, GIST cells were harvested during exponential growth. GIST cells (3 × 106) were formulated as a 1:1 mixture with Matrigel (BD Biosciences) and were subcutaneously injected into BALB/c nude mice. The resulting tumors were measured once per 3 days. Tumor volume was calculated using the formula: tumor volume (mm3) = [(W2 × L)/2] in which width (W) is defined as the smaller of the two measurements and length (L) is defined as the larger of the two measurements. For drug treated mouse, when the average tumor volume reached approximately 75 mm3, the mice were randomized into treatment groups with eight mice per group. For GIST-T1 xenograft tumor mice, the dose of imatinib was 25 mg/kg twice daily and RO-3306 was 4 mg/kg every 2 days by oral gavage for 30 days. And for GIST PDX xenograft tumor mice, the dose of imatinib was 50 mg/kg twice daily and RO-3306 was 4 mg/kg every two days by oral gavage for 30 days. After 30 days treatment, all mice were killed and tumors were collected. Tumor volume and tumor weight was used to evaluate antitumor activity.

IHC

IHC was performed on tissue and tumor sections using CDK1 antibody (Santa Cruz Technology, #sc-54) or Ki67 antibody (Leica Biosystems, #NCL-L-KI67-MIB1). 4 μmol/L slides were deparaffinized in xylene and hydrated in a graded series of alcohol. Slides were then boiled by microwave for 12 minutes in citrate buffer (pH 6). IHC reactions were visualized by diaminobenzidine staining, using an EnVision+ system (Dako).

Statistical analysis

Statistical analysis was performed using GraphPad Prism 8 (GraphPad Software, RRID:SCR_002798). The difference between two groups was analyzed by one-tailed or two-tailed unpaired sample t test. IC50 of imatinib or RO-3306 in each group was analyzed using nonlinear regression model.

Genes highly expressed in advanced GIST

According to the well-established clinicopathological criteria (23), primary GIST is classified as low-risk, intermediate-risk, and high-risk. Low-risk and intermediate-risk are stages of GIST development that precede transition to clinically aggressive “high-risk” GIST. Herein, GIST with low risk or intermediate risk is collectively called early-stage GIST. In contrast, GIST with high risk or metastasis is called advanced GIST. We first performed and analyzed whole transcriptome sequencing (RNA sequencing) data on 43 GIST samples from 43 patients, including 11 early-stage GIST and 32 advanced GIST samples, as we reported recently (Supplementary Table S1; ref. 22). This analysis identified 568 genes with a higher expression level in advanced GIST than in early-stage GIST (Fig. 1A; Supplementary Tables S2 and S3, P < 0.05, FDR <0.25, unpaired t test, two-sided, log2-fold change >1).

Figure 1.

Genome-wide CRISPR screening identifies CDK1 as a vulnerability of advanced GIST cells. A, Volcano plot of RNA-seq data displaying the pattern of differential gene expression between early-stage and advanced GIST. Significantly higher expressed genes are highlighted in black, with the dashed lines showing thresholds of fold change > 2 and P < 0.05. The CDK1 gene is indicated by the red point. B, Design of the genome-wide CRISPR screen to identify essential genes in advanced GIST cells. C, Box plot showing the distribution of sgRNA frequencies at different population doublings. Median, upper, and lower quartiles are shown. D, Ranked CS for each gene from genome-wide CRISPR screen in GIST430/654 cells. All genes analyzed in the CRISPR screen are plotted in gray, and the 568 genes identified in the whole transcriptome sequencing are plotted in black. CDK1 gene is indicated by the red point.

Figure 1.

Genome-wide CRISPR screening identifies CDK1 as a vulnerability of advanced GIST cells. A, Volcano plot of RNA-seq data displaying the pattern of differential gene expression between early-stage and advanced GIST. Significantly higher expressed genes are highlighted in black, with the dashed lines showing thresholds of fold change > 2 and P < 0.05. The CDK1 gene is indicated by the red point. B, Design of the genome-wide CRISPR screen to identify essential genes in advanced GIST cells. C, Box plot showing the distribution of sgRNA frequencies at different population doublings. Median, upper, and lower quartiles are shown. D, Ranked CS for each gene from genome-wide CRISPR screen in GIST430/654 cells. All genes analyzed in the CRISPR screen are plotted in gray, and the 568 genes identified in the whole transcriptome sequencing are plotted in black. CDK1 gene is indicated by the red point.

Close modal

Genome-wide CRISPR screen identifies CDK1 as a vulnerability of advanced GIST cells

The CRISPR-Cas9 knockout screen is a powerful approach in identifying genes that are essential for tumor growth and proliferation (18). To functionally screen for genes that are essential in imatinib-resistant GIST cells, a genome-scale CRISPR knockout screen was performed in imatinib-resistant/advanced GIST430/654 cells (Fig. 1B), established from an advanced/metastatic GIST patient refractory to imatinib, harboring the primary KIT exon 11 mutation (V560_L576del) and the secondary exon 13 V654A missense mutation (24). We introduced a lentiviral genome-wide CRISPR-Cas9 knockout (GeCKO v2) library (25) into GIST430/654 cells, which contained 123,411 unique sgRNAs targeting 19,050 genes (6 sgRNAs per gene) and 1,000 nontargeting negative control sgRNAs. Transduced cells were purified by selection with puromycin and then passaged with an average representation of 200 cells per sgRNA. We monitored evolving cell populations over ∼15 doublings by deep sequencing of sgRNAs. After 15 population doublings (PD), the sgRNA distribution was significantly shifted compared with PD0 cells, indicating that the screen functioned as designed (Fig. 1C). For each sgRNA, we first calculated the log2-fold change in abundance from the screen endpoint compared with the initial point, followed by subtraction of the median score of the negative control sgRNAs. We defined a CS as the median score of all sgRNAs targeting the gene after 15 PDs. To identify genes essential for viability in GIST430/654 cells, we rank-ordered genes by CS from most negative to positive (Supplementary Table S4). Among the 568 genes identified in the whole transcriptome sequencing that were expressed at higher levels in advanced GIST, CDK1 had the second lowest CS, ranking as the top two essential hits in imatinib-resistant GIST cells, showing that CDK1 perturbation causes a decrease in the fitness of advanced GIST (Fig. 1D). We decided to investigate CDK1 rather than the other top candidates because CDK1 is more druggable, and hence might have more translational potential.

Frequent high expression of CDK1 in advanced GIST

Frequent high expression of CDK1 in advanced GIST was identified in the discovery GIST cohort (Supplementary Table S1; Fig. 2A). Interestingly, CDK1 high expression found in one metastasis was also detected in all other metastases from a particular GIST individual (Fig. 2B), consistent with a biological advantage of CDK1 overexpression. High CDK1 expression was not found in benign/early-stage GIST but rather was remarkably frequent in advanced GIST (Fig. 2A; Supplementary Table S1). In contrast, the other CDK family members, such as CDK4, CDK6, and CDK9 were uniformly expressed in early-stage and advanced GIST (Fig. 2A), suggesting the unique expression pattern of CDK1 was biologically selected. Next, the expression of CDK1 was validated in another GIST cohort containing 40 early-stage GISTs and 52 advanced GISTs by immunoblotting. CDK1 expression was demonstrated in 38 of 52 (73%) advanced GISTs, whereas no-to-low expression was demonstrated in 40 of 40 (100%) early-stage GISTs (Fig. 2C). High CDK1 expression was detected in advanced GISTs irrespective of whether they had KIT or PDGFRA mutations or wide type genotype (Supplementary Table S5; Supplementary Table S1). Then, we investigated in a tissue microarray (TMA) consisting of 502 GIST samples using CDK1 IHC. In line with the RNA-seq analysis, CDK1 was highly expressed in advanced GIST (Fig. 2D and E). Furthermore, the high expression level of CDK1 was associated with malignancy in GIST (P < 0.0001, two-tailed Fisher test; Fig. 2E). CDK1 expression level is positively associated with CCNB1, CDC25A, or CDC25B expression level, but is not positively associated with c-Myc expression level in GIST (Supplementary Figs. S1A–S1D). CDK1 expression level is also positively associated with Ki-67, a hallmark of proliferation (26), expression in GIST (Fig. 2F; Supplementary Figs. S2A and S2B). In conclusion, CDK1 was highly expressed in advanced GIST, whereas no-to-low expression was observed in early-stage GIST.

Figure 2.

Frequent high expression of CDK1 in advanced GIST. A, Results from the RNA-seq on the discovery GIST cohort to detect CDK1, CDK4, CDK6 and CDK9 expression (early-stage GIST n = 11, advanced n = 32). B, High expression of CDK1 found in one metastasis was also detected in all other metastases from a particular GIST individual. C, Protein blotting of the validation GIST cohort demonstrated no-to-low expression in early-stage/benign GIST but frequent high expression in advanced GIST. D, CDK1 immunohistochemistry in a larger GIST cohort (early-stage GIST n = 325, advanced n = 177). Immunohistochemical detection in formalin-fixed paraffin-embedded surgical clinical samples shows robust CDK1 expression in advanced GIST, whereas CDK1 expression is inhibited in early-stage GIST. E, Contingency table summarizing CDK1 expression assessed by immunohistochemistry in 325 early-stage and 177 advanced GISTs in a tissue microarray. High expression of CDK1 correlates with malignancy in GIST (P < 0.0001, two-tailed Fisher test). F, Correlation of RNA expression between MKI67 and CDK1 on the discovery GIST cohort. GISTs with TKI treatment have been removed from the analyses. **, P < 0.01; ns, not significant.

Figure 2.

Frequent high expression of CDK1 in advanced GIST. A, Results from the RNA-seq on the discovery GIST cohort to detect CDK1, CDK4, CDK6 and CDK9 expression (early-stage GIST n = 11, advanced n = 32). B, High expression of CDK1 found in one metastasis was also detected in all other metastases from a particular GIST individual. C, Protein blotting of the validation GIST cohort demonstrated no-to-low expression in early-stage/benign GIST but frequent high expression in advanced GIST. D, CDK1 immunohistochemistry in a larger GIST cohort (early-stage GIST n = 325, advanced n = 177). Immunohistochemical detection in formalin-fixed paraffin-embedded surgical clinical samples shows robust CDK1 expression in advanced GIST, whereas CDK1 expression is inhibited in early-stage GIST. E, Contingency table summarizing CDK1 expression assessed by immunohistochemistry in 325 early-stage and 177 advanced GISTs in a tissue microarray. High expression of CDK1 correlates with malignancy in GIST (P < 0.0001, two-tailed Fisher test). F, Correlation of RNA expression between MKI67 and CDK1 on the discovery GIST cohort. GISTs with TKI treatment have been removed from the analyses. **, P < 0.01; ns, not significant.

Close modal

To understand the basis for the increase in CDK1 expression, we examined the copy number of the CDK1 locus using quantitative PCR. The analysis revealed increased CDK1 copy number (∼8 copies) in two GISTs as compared with nonneoplastic cells from the same patients (∼2 copies), demonstrating the tumor restricted nature of CDK1 amplification (Supplementary Fig. S3; Supplementary Table S5). Altogether, the data are consistent with a genomic amplification in the CDK1 locus in the advanced GIST with high expression of CDK1.

CDK1 promotes GIST tumor proliferation and progression

The biological function of CDK1 was investigated using various human GIST models. CDK1 biological function was evaluated by CDK1 ablation in GIST cells with high CDK1 expression. Lentivirus-mediated shRNA transduction into CDK1 highly expressed imatinib-resistant GIST430/654 cells decreased endogenous CDK1 expression (Fig. 3A and B). CDK1 knockdown with two CDK1 shRNAs induced strong proliferation inhibition in both short-term (Fig. 3C) and long-term (Fig. 3D) proliferation assays and increased cell apoptosis (Fig. 3B). CDK1 knockdown also inhibited three-dimensional (3D) anchorage-independent growth in GIST cells (Fig. 3E). To determine whether the inhibition of cell proliferation manifested in vivo, we generated both control and CDK1-ablated GIST430/654 xenografts in nude mice. CDK1 inhibition dramatically attenuated tumor growth (Fig. 3F). To further test the oncogenic role of CDK1 in GIST tumorigenesis, we extended these findings to a second human GIST cell line (GIST-T1, established from an advanced/metastatic GIST harboring the KIT exon 11 deletion, V560_L576del) with high CDK1 expression. Consistent with the above data, CDK1 knockdown reduced the cell growth and proliferation in vitro and in vivo (Fig. 3C–F). Flow cytometric analysis showed that CDK1 knockdown inhibited cell-cycle progression at the G1–S checkpoint, reducing the proportion of cells in S phase and increasing the proportion of cells in G0–G1 phase (Fig. 3G), which cannot be explained by CDK1’s role as a regulator of cell-cycle driving G2–M cell-cycle progression (27, 28). CDK1 knockdown caused evident cellular senescence, as indicated by positive SA-β-Gal staining (Fig. 3H), which is consistent with the increased proportion of cells in G0–G1 phase after CDK1 ablation. Collectively, these results demonstrate that highly expressed CDK1 is oncogenic, promoting GIST tumorigenesis and progression.

Figure 3.

CDK1 ablation inhibits tumor growth and proliferation in advanced GIST. A, The level of CDK1 knockdown in imatinib-resistant GIST430/654 and imatinib-sensitive GIST-T1 cell lines was measured by qRT-PCR. B, GIST430/654 and GIST-T1 cells were treated with scrambled shRNA (Ctrl) or shRNA (CDK1-KD-1 and CDK1-KD-2) against CDK1. Western blot analysis detected the indicated protein levels. Arrows, full length and cleaved PARP. C, Lentivirus-mediated CDK1 knockdown reduces the viability of GIST430/654 and GIST-T1 cells, as assessed by the CellTiter-Glo viability assay. D, Crystal violet staining assays show that CDK1 knockdown suppresses GIST cell proliferation. Mean percentage areas are shown. E, CDK1 knockdown suppresses anchorage-independent growth of GIST430/654 and GIST-T1 cells. Mean colony numbers are shown. F, CDK1 knockdown inhibits the growth of GIST430/654 and GIST-T1 xenografts in nude mice. Growth curves of transplanted tumors are shown. Error bars, mean ± SEM of six replicates. G, Cell cycle analyses demonstrating that CDK1 knockdown decreases the cell cycle in both GIST430/654 and GIST-T1 cells, reducing the proportion of cells in S phase and G2–M phase and increasing the proportion of cells in G0–G1 phase. Experiments were performed in triplicate. H, Cellular senescence analyses demonstrating that CDK1 knockdown increases cellular senescence in both GIST430/654 and GIST-T1 cells. Top, representative images of GIST cell senescence by SA-β-Gal staining. Bottom, SA-β-Gal staining statistics of cells in the individual groups. Data obtained from counting of at least 200 cells per independent experiment and analyzed using Student t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (P < 0.05 by unpaired t test, two-sided). The P values refer to the comparisons with the controls. The data represent the mean ± SE.

Figure 3.

CDK1 ablation inhibits tumor growth and proliferation in advanced GIST. A, The level of CDK1 knockdown in imatinib-resistant GIST430/654 and imatinib-sensitive GIST-T1 cell lines was measured by qRT-PCR. B, GIST430/654 and GIST-T1 cells were treated with scrambled shRNA (Ctrl) or shRNA (CDK1-KD-1 and CDK1-KD-2) against CDK1. Western blot analysis detected the indicated protein levels. Arrows, full length and cleaved PARP. C, Lentivirus-mediated CDK1 knockdown reduces the viability of GIST430/654 and GIST-T1 cells, as assessed by the CellTiter-Glo viability assay. D, Crystal violet staining assays show that CDK1 knockdown suppresses GIST cell proliferation. Mean percentage areas are shown. E, CDK1 knockdown suppresses anchorage-independent growth of GIST430/654 and GIST-T1 cells. Mean colony numbers are shown. F, CDK1 knockdown inhibits the growth of GIST430/654 and GIST-T1 xenografts in nude mice. Growth curves of transplanted tumors are shown. Error bars, mean ± SEM of six replicates. G, Cell cycle analyses demonstrating that CDK1 knockdown decreases the cell cycle in both GIST430/654 and GIST-T1 cells, reducing the proportion of cells in S phase and G2–M phase and increasing the proportion of cells in G0–G1 phase. Experiments were performed in triplicate. H, Cellular senescence analyses demonstrating that CDK1 knockdown increases cellular senescence in both GIST430/654 and GIST-T1 cells. Top, representative images of GIST cell senescence by SA-β-Gal staining. Bottom, SA-β-Gal staining statistics of cells in the individual groups. Data obtained from counting of at least 200 cells per independent experiment and analyzed using Student t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001 (P < 0.05 by unpaired t test, two-sided). The P values refer to the comparisons with the controls. The data represent the mean ± SE.

Close modal

CDK1 interacts with and phosphorylates AKT

High CDK1 expression was demonstrated in some but not all advanced GISTs, whereas no-to-low expression was demonstrated in all early-stage GISTs (Fig. 2D), which strongly suggests that CDK1 is not absolutely essential for cell proliferation in GIST. What are the molecular mechanisms underlying CDK1 oncogenic role in advanced GIST? CDK1 is the founding member of the cyclin-dependent protein kinase (CDK) family (29). To identify the potential substrates for CDK1 in the GIST context. FLAG-tagged CDK1 was transduced into imatinib-resistant GIST430/654 cells as bait. CDK1 was recovered along with proteins that potentially formed complexes with CDK1 through co-IP with anti-FLAG beads, and the samples were subjected to mass spectrometry-based proteomics screening. Top protein hits included known CDK1 interaction proteins, including HNRNPUL1 (30) and CCNB1 (Fig. 4A; ref. 31). Strikingly, AKT1 protein emerged as one of the top hits with the highest confidence. This result was validated in a co-IP assay (Fig. 4B). AKT2 does not interact with CDK1 in GIST (Supplementary Fig. S4). CDK1 can activate PDK1, which could then phosphorylate AKT in embryonic stem cells (32). Interestingly, AKT1-CDK1 interaction was confirmed in both PDK1-wild-type GIST430/654 cells and PDK1-deficient GIST430/654 cells (Fig. 4B; Supplementary Fig. S5). These data indicate that CDK1 can interact with AKT independently of the PDK1 collaboration in the GIST context. AKT1–CDK1 interaction decreases in the face of pharmacologic inactivation of the AKT (Fig. 4C). The interaction between endogenous CDK1 and AKT1 was confirmed with co-IP with anti-CDK1 antibody and anti-AKT antibody, respectively (Fig. 4D). When GIST cells were treated with CDK1 inhibitor RO-3306, co-IP of CDK1 with AKT1 (detection of AKT with CDK1 IP) was decreased in GIST430/654 and GIST-T1 cell lines (Fig. 4D, left). Co-IP of AKT with CDK1 (CDK1 immunoblot bands with AKT IP) was also decreased in GIST430/654 cells (Fig. 4D, right). In line with these findings, in vitro binding assays demonstrated a direct binding of CDK1 to AKT, which was dramatically inhibited by RO-3306 treatment (Fig. 4E). With HEK293T cell model and the PDK1-deficient GIST430/654 model, we further demonstrate that CDK1 induces AKT activation loop (T308) and hydrophobic motif (S473) phosphorylation in the absence of PDK1 (Fig. 4F and G). These observations indicate that CDK1 is sufficient to induce AKT activation/phosphorylation independently of the canonical PDK1 collaboration in GIST.

Figure 4.

CDK1 interacts with AKT and phosphorylates AKT. A, Identification of CDK1-interacting proteins through co-IP followed by mass spectrometry–based proteomics. GIST430/654 cells were transduced with FLAG-CDK1 lentivirus and a pull-down assay was carried out using anti-FLAG antibody. The table shows high-confidence hits in the mass spectrum. B, Confirmation of AKT-FLAG-CDK1 interaction in PDK1-WT GIST430/654 cells or PDK1-deficient GIST430/654 cells. The GIST430/654 cells were transduced with FLAG-CDK1, and cell lysates were subjected to a co-IP assay using anti-Flag antibody, followed by immunoblotting with anti-AKT antibody.C, The GIST430/654 cells were transduced with FLAG-CDK1, followed by AKT inhibitor, MK-2206 (3 μmol/L), treatment for 24 hours and cell lysates were subjected to a co-IP assay using anti-Flag antibody, followed by immunoblotting with anti-AKT antibody. D, Endogenous CDK1 interacts strongly with AKT in GIST430/654 and GIST-T1 cell lines. CDK1 and AKT IP immunoblot of cells treated with vehicle or CDK1 inhibitor (RO-3306, 1 μmol/L) for 24 hours. Left, cell lysates were immunoprecipitated with anti-CDK1 antibody and immunoblotted with anti-AKT antibody. Right, cell lysates were immunoprecipitated with anti-AKT antibody and immunoblotted with anti-CDK1 antibody. E, Immunoblots depicting the binding of recombinant AKT to CDK1 after CDK1 IP. RO-3306 was used to show CDK1-dependent interaction. GIST cells were treated with RO-3306 (1 μmol/L) for 24 hours. F, HEK293T cells were transfected with FLAG-CDK1 or HA-AKT plasmids as indicated. Two days posttransfection, whole-cell lysates were assessed for CDK1 and AKT expression and accumulation of phosphorylated AKT as indicated. G, PDK1-deficient GIST430/654 cells were transduced with FLAG-CDK1 or HA-AKT lentivirus as indicated. Two days after infection, whole-cell lysates were assessed for CDK1 and AKT expression and accumulation of phosphorylated AKT as indicated. H, GIST430/654 and GIST-T1 cells were treated with RO-3306 (1 μmol/L) for 24 hours. Western blot analysis detected the indicated protein levels. Arrows, full length and cleaved PARP.

Figure 4.

CDK1 interacts with AKT and phosphorylates AKT. A, Identification of CDK1-interacting proteins through co-IP followed by mass spectrometry–based proteomics. GIST430/654 cells were transduced with FLAG-CDK1 lentivirus and a pull-down assay was carried out using anti-FLAG antibody. The table shows high-confidence hits in the mass spectrum. B, Confirmation of AKT-FLAG-CDK1 interaction in PDK1-WT GIST430/654 cells or PDK1-deficient GIST430/654 cells. The GIST430/654 cells were transduced with FLAG-CDK1, and cell lysates were subjected to a co-IP assay using anti-Flag antibody, followed by immunoblotting with anti-AKT antibody.C, The GIST430/654 cells were transduced with FLAG-CDK1, followed by AKT inhibitor, MK-2206 (3 μmol/L), treatment for 24 hours and cell lysates were subjected to a co-IP assay using anti-Flag antibody, followed by immunoblotting with anti-AKT antibody. D, Endogenous CDK1 interacts strongly with AKT in GIST430/654 and GIST-T1 cell lines. CDK1 and AKT IP immunoblot of cells treated with vehicle or CDK1 inhibitor (RO-3306, 1 μmol/L) for 24 hours. Left, cell lysates were immunoprecipitated with anti-CDK1 antibody and immunoblotted with anti-AKT antibody. Right, cell lysates were immunoprecipitated with anti-AKT antibody and immunoblotted with anti-CDK1 antibody. E, Immunoblots depicting the binding of recombinant AKT to CDK1 after CDK1 IP. RO-3306 was used to show CDK1-dependent interaction. GIST cells were treated with RO-3306 (1 μmol/L) for 24 hours. F, HEK293T cells were transfected with FLAG-CDK1 or HA-AKT plasmids as indicated. Two days posttransfection, whole-cell lysates were assessed for CDK1 and AKT expression and accumulation of phosphorylated AKT as indicated. G, PDK1-deficient GIST430/654 cells were transduced with FLAG-CDK1 or HA-AKT lentivirus as indicated. Two days after infection, whole-cell lysates were assessed for CDK1 and AKT expression and accumulation of phosphorylated AKT as indicated. H, GIST430/654 and GIST-T1 cells were treated with RO-3306 (1 μmol/L) for 24 hours. Western blot analysis detected the indicated protein levels. Arrows, full length and cleaved PARP.

Close modal

Because CDK1 is a kinase that catalyzes the transfer of phosphate groups from ATP to target substrate proteins, we hypothesized that CDK1 may phosphorylate AKT while interacting with this protein. Consistent with our hypothesis, knocking down CDK1 using two shRNAs showed a reduction in AKT phosphorylation in GIST430/654 and GIST-T1 cells (Fig. 3B). Although we observed differences in the phosphorylation of AKT, we did not see the effects of CDK1 knockdown on AKT protein expression. Similar to these findings, RO-3306 treatments led to a decrease in AKT phosphorylation in GIST430/654 and GIST-T1 cells without affecting total AKT levels (Fig. 4H). Together, these results demonstrate a CDK1–AKT interaction and subsequent phosphorylation of AKT in GIST cells.

AKT phosphorylation has been shown to play an important role in cancer progression by mediating a variety of biological responses including cell growth, proliferation, and survival (33). To determine whether CDK1 promotes GIST tumor growth and proliferation by phosphorylation of AKT, a series of rescue experiments were performed in various GIST models. Endogenous CDK1 was knocked down in GIST430/654 cells. CDK1 knockdown reduced the GIST growth and proliferation, whereas enforced active myristoyl-AKT expression (34) in CDK1-ablated GIST cells significantly attenuated growth and proliferation inhibition properties in vitro and in vivo (Fig. 5A–D). These data show that CDK1 oncogenic roles in GIST largely through AKT phosphorylation.

Figure 5.

CDK1 promotes GIST tumor growth and proliferation largely through AKT. A, Viability of engineered GIST430/654 cells was assessed by the CellTiter-Glo viability assay. B, Engineered GIST430/654 cell proliferation was assessed by crystal violet staining assays. Mean percentage areas are shown. C, Anchorage-independent growth of engineered GIST430/654 cells was assessed by soft agar assays. Mean colony numbers are shown. D, Growth of xenografts of engineered GIST430/654 cells in nude mice. Growth curves of transplanted tumors are shown. Error bars, mean ± SEM of six replicates. *, P < 0.05; **, P < 0.01.

Figure 5.

CDK1 promotes GIST tumor growth and proliferation largely through AKT. A, Viability of engineered GIST430/654 cells was assessed by the CellTiter-Glo viability assay. B, Engineered GIST430/654 cell proliferation was assessed by crystal violet staining assays. Mean percentage areas are shown. C, Anchorage-independent growth of engineered GIST430/654 cells was assessed by soft agar assays. Mean colony numbers are shown. D, Growth of xenografts of engineered GIST430/654 cells in nude mice. Growth curves of transplanted tumors are shown. Error bars, mean ± SEM of six replicates. *, P < 0.05; **, P < 0.01.

Close modal

Effect of CDK1 on GIST cell signaling

We next compared the signal transduction pathways regulated by oncogenic CDK1. CDK1 ablation suppressed phosphorylation of AKT (see above, Fig. 3B), but not KIT kinase (Supplementary Fig. S6). Apoptotic responses to CDK1 ablation were assessed by measuring cleaved PARP, which has been well correlated with a commitment to programmed cell death (35). CDK1 ablation resulted in the induction of PARP cleavage (Fig. 3B). CDK1 ablation with shRNAs decreased proliferating cell nuclear antigen (PCNA) expression (Fig. 3B), which is consistent with CDK1 promoting GIST proliferation (Fig. 3C–F). In summary, the growth inhibitory, cellular senescence and apoptotic effects seen upon CDK1 ablation reflect a specific requirement for CDK1 in advanced GIST and highly expressed CDK1 confers “oncogene addiction” in advanced GIST.

Chemical inhibition of CDK1 disrupts GIST cell proliferation

The above functional data suggest that inhibition of oncogenic CDK1 may have therapeutic potential. To assess the therapeutic potential of the above observations, we assessed the sensitivity of GIST models to small molecule inhibitors of CDK1. Pharmacologic inhibition of CDK1 by RO-3306 (21) and ablation of CDK1 by shRNAs showed similar consequent decreases in AKT phosphorylation and an increase in apoptosis (Fig. 3B). Next, the CDK1 inhibitor RO-3306 was tested against a range of GIST cell lines with different activating and secondary resistance mutations (Fig. 6A). RO-3306 was shown to inhibit the viability of both imatinib-resistant (GIST430/654, GIST48, and GIST48B) and imatinib-sensitive (GIST-T1 and GIST882) lines with CDK1 high expression with sub-200 nmol/L potencies (Fig. 6A–D; Supplementary Fig. S7). Interestingly, GIST-1 (established from a metastatic GIST with no CDK1 expression) primary cultured cells displayed no sensitivity to RO-3306 (Fig. 6A–D; Supplementary Fig. S7). Similarly, there is no evidence for viability inhibition in RO-3306-treated nontransformed fibroblast cell line (Fig. 6A–D; Supplementary Fig. S7). BEZ235 (PI3K inhibitor; ref. 36) or MK-2206 (AKT inhibitor; ref. 37) was shown to inhibit the viability of both imatinib-resistant (GIST430/654 and GIST48) and imatinib-sensitive (GIST-T1 and GIST882) lines (Supplementary Figs. S8A–S8C). CDK1-AKT1 interaction decreases in the face of pharmacologic inactivation of the AKT (Fig. 4C).

Figure 6.

Antitumor activities of CDK1 inhibition against imatinib-resistant and imatinib-sensitive GIST models. A, CellTiter-Glo growth inhibition assay of CDK1 inhibitor RO-3306 in the panel of GIST cell lines and GIST primary cells. GIST-1, established from a metastatic GIST with no CDK1 expression, is a primary GIST cell culture. BJ is a nontransformed fibroblast cell line. B, IC50 values (nmol/L) for inhibition of cell proliferation in imatinib-resistant and imatinib-sensitive GIST cells. C, Protein blotting to show the CDK1 expression levels in the panel of GIST cell lines and GIST primary cells. D, GIST cells were treated with RO-3306 (1 μmol/L) for 24 hours. Western blot analysis detected the indicated protein levels. E–G, Antitumor activities of CDK1 inhibition against imatinib-resistant GIST xenografts. Mice bearing patient-derived xenograft GIST tumors (ex 11+ ex 17) were treated with RO-3306 at 4 mg/kg every 2 days or imatinib at 50 mg/kg twice a day orally. Control tumor-bearing animals received vehicle. Relative tumor volume is shown. All data points include 8 animals per group. E, Growth curves of tumor volume changes in the GIST PDX models in various treatment groups. The data represent the means ± SEM. F, The tumor weights of the indicated groups corresponding to E. The data represent the means ± SD. G, Body weight changes in the GIST xenograft study. The data represent the means ± SEM. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 (P < 0.05 by unpaired t test, two-sided). The P values refer to the comparisons with the controls. H, Cell viability assay reveals the synergistic effect of RO-3306 and imatinib in GIST-T1. Light gray bars, control values. “Multiplication” indicates expected effect of combined treatment if single-treatment effects are multiplied; gray arrow, actual effect of combination. I–K, Antitumor activities of CDK1 inhibition against imatinib-sensitive GIST xenografts. Mice bearing GIST-T1 tumors were treated with RO-3306 at 4 mg/kg every 2 days or imatinib at low dose (25 mg/kg) twice a day orally as single agents and in combination. Control tumor-bearing animals received vehicle. Relative tumor volume is shown. All data points include 8 animals per group. I, Growth curves of tumor volume changes in the GIST-T1 xenograft models in various treatment groups. The data represent the means ± SEM. J, The tumor weights of the indicated groups corresponding to I. The data represent means ± SD. K, Body weight changes in the GIST xenograft study. The data represent the means ± SEM. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 (P < 0.05 by unpaired t test, two-sided). The P values refer to the comparisons with the controls.

Figure 6.

Antitumor activities of CDK1 inhibition against imatinib-resistant and imatinib-sensitive GIST models. A, CellTiter-Glo growth inhibition assay of CDK1 inhibitor RO-3306 in the panel of GIST cell lines and GIST primary cells. GIST-1, established from a metastatic GIST with no CDK1 expression, is a primary GIST cell culture. BJ is a nontransformed fibroblast cell line. B, IC50 values (nmol/L) for inhibition of cell proliferation in imatinib-resistant and imatinib-sensitive GIST cells. C, Protein blotting to show the CDK1 expression levels in the panel of GIST cell lines and GIST primary cells. D, GIST cells were treated with RO-3306 (1 μmol/L) for 24 hours. Western blot analysis detected the indicated protein levels. E–G, Antitumor activities of CDK1 inhibition against imatinib-resistant GIST xenografts. Mice bearing patient-derived xenograft GIST tumors (ex 11+ ex 17) were treated with RO-3306 at 4 mg/kg every 2 days or imatinib at 50 mg/kg twice a day orally. Control tumor-bearing animals received vehicle. Relative tumor volume is shown. All data points include 8 animals per group. E, Growth curves of tumor volume changes in the GIST PDX models in various treatment groups. The data represent the means ± SEM. F, The tumor weights of the indicated groups corresponding to E. The data represent the means ± SD. G, Body weight changes in the GIST xenograft study. The data represent the means ± SEM. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 (P < 0.05 by unpaired t test, two-sided). The P values refer to the comparisons with the controls. H, Cell viability assay reveals the synergistic effect of RO-3306 and imatinib in GIST-T1. Light gray bars, control values. “Multiplication” indicates expected effect of combined treatment if single-treatment effects are multiplied; gray arrow, actual effect of combination. I–K, Antitumor activities of CDK1 inhibition against imatinib-sensitive GIST xenografts. Mice bearing GIST-T1 tumors were treated with RO-3306 at 4 mg/kg every 2 days or imatinib at low dose (25 mg/kg) twice a day orally as single agents and in combination. Control tumor-bearing animals received vehicle. Relative tumor volume is shown. All data points include 8 animals per group. I, Growth curves of tumor volume changes in the GIST-T1 xenograft models in various treatment groups. The data represent the means ± SEM. J, The tumor weights of the indicated groups corresponding to I. The data represent means ± SD. K, Body weight changes in the GIST xenograft study. The data represent the means ± SEM. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 (P < 0.05 by unpaired t test, two-sided). The P values refer to the comparisons with the controls.

Close modal

Antitumor activities of a CDK1 inhibitor against imatinib-resistant and imatinib-sensitive GIST xenografts

The antitumor effects of the CDK1 inhibitor RO-3306 were assessed in a GIST patient-derived xenograft (PDX) mouse model (bearing KIT exon 11/17, deletion/N822K, alterations). As expected, this xenograft model was refractory to imatinib even at a high dose of 50 mg/kg (Fig. 6E and F). The results correlated well with the clinical observations because patients with this KIT exon 11/17 aberration often show resistance to imatinib therapies. In contrast, RO-3306 at a dose of 4 mg/kg led to GIST PDX regression (Fig. 6E and F) and was well tolerated (Fig. 6G). In this model, RO-3306 reduced the tumor volume by 72% after 28 days of therapy and was more effective than imatinib (Fig. 6E and F).

Finally, the synergistic effects of RO-3306 and imatinib were observed in GIST-T1 cells harboring the exon 11 (V560_Y578del) mutation (Fig. 6H). Another GIST xenograft model using GIST-T1 cells was investigated. In the GIST-T1 xenograft model, imatinib (at a low dose of 25 mg/kg) and RO-3306 (4 mg/kg) stabilized the tumor growth when each was given as a single agent, but their combination reduced the tumor volume by 81% after 28 days of therapy (Fig. 6I–K), showing the synergistic effects of RO-3306 and imatinib in vivo.

GISTs are life-threatening when advanced/metastatic or not amenable to surgical removal. In a few patients with advanced GISTs refractory to imatinib, treatment with sunitinib followed by regorafenib has provided limited tumor control (1, 38). In this year, avapritinib and ripretinib were approved by the FDA as novel TKIs. However, avapritinib was only approved for GIST with PDGFRA mutations in exon 18, which benefits a small fraction of patients with GIST (∼5%; refs. 8, 10). Some patients with PDGFRA mutations treated within the recent phase I trial have progressed following an initial response to avapritinib (10). Ripretinib is a switch-control TKI active against a broad spectrum of KIT and PDGFRA mutations (39). According to a recent phase III trial data, the median PFS in the ripretinib group was only 6.3 months and the objective response rate was 9% (40). In addition, ripretinib probably does not benefit KIT/PDGFRA wide type GIST. Hence, additional active treatments are needed for advanced GIST patients.

CDK1 is the founding member of the cyclin-dependent protein kinase (CDK) family (29). Its precise functions in the context of the dividing cell are poorly understood (41). Unrestricted cell proliferation, one of the hallmarks of malignant tumors, is often driven by alterations in CDK activity. Altered CDK expression and/or activity is observed in many human cancers (42, 43). Given that CDKs are often overactive in many human cancers, inhibitors targeting CDKs are used to treat cancers in clinic by preventing unregulated proliferation of cancer cells. The US FDA have approved several CDK4/6 inhibitors (palbociclib, ribociclib, and abemaciclib) for the treatment of certain type of breast cancer (44–46). A priori, the essential nature of CDK1 might be predicted to preclude its selection as a potential target for cancer treatment. However, in GIST context, CDK1 inhibitor RO-3306, dramatically disrupts GIST cell proliferation in CDK1 highly expressed advanced GIST, but not in CDK1-negative GIST cells or nontransformed fibroblast cells. Therefore, with appropriate selection of the cellular context, its unique expression pattern may offer opportunities for therapeutic exploitation. In addition, the validation of CDK1 as a clinical target has been accelerated by the CDK1 selective tool compound RO-3306 as one example. RO-3306 remains one of only a few inhibitors that have shown selectivity for CDK1 over other members of the CDK family.

CDK1 inhibition might be useful in the treatment of human tumors overexpressing MYC (47). CDK1 inhibition has also been proposed as a route to causing chemosensitivity, enhancing the effects of PARP inhibitors (48). Oncogenic CDK1 expression found in one metastasis was also detected in all other metastases from a particular GIST individual. High expression of CDK1 was not found in benign/early-stage GIST but rather was remarkably frequent in advanced GIST, suggesting a biological advantage for CDK1 overexpression.

CDKN2A is frequently inactivated in advanced GIST, and this can result in a dysregulated regulation of RB, which is known to interact with CDK1. All the CDK1 highly expressed GIST cell lines were sensitive to RO-3306 or CDK1 ablation no matter CDKN2A or RB is inactivated or not. We explored the CDK1 inhibitor RO-3306 to determine the efficacy of treatment in GIST PDX tumor models. PDX models have unique advantages for conducting preclinical trials, as they mostly mimic the molecular, histologic, and genomic characteristics of the tumors. This tendency reveals an increased acceptance for its predictive value and a high degree of representation of the clinical outcome, which could increase the overall success of transitioning to a new clinically approved treatment and delivery to patients. They are a precious approach to integrating preclinical and clinical drug development. A CDK1 inhibitor (RO-3306) acts on advanced GIST with an antitumor growth effect on PDX tumor models, which may be due to its effects. The in vivo tumorigenicity analysis showed the antitumor effect of the CDK1 inhibitor on GIST PDX models. We also explored the CDK1 inhibitor RO-3306 alone or in combination with imatinib to determine the efficacy of treatment in GIST-T1 xenograft mouse model. Notably, we demonstrated that the adverse effects of RO-3306, even plus imatinib were well tolerated and that RO-3306 is a bearable drug with a similar body weight curve to the vehicle-treated group (Fig. 6). On the basis of the anticancer evidence of the PDX models, the high expression activation level of CDK1 has a more dramatic response to CDK1 inhibition treatment compared with the corresponding controls. With the GIST xenograft data, our study has pointed to CDK1 as a potential new therapeutic target for clinical translation to treat patients with advanced GISTs, especially imatinib-resistant GISTs, which is urgently needed.

We provided the evidence that CDK1 plays an oncogenic role in GIST progression. Here, we confirmed the hypothesis of the essential role of CDK1 in advanced GIST and that pharmacological inhibition could suppress GIST progression. We explored the molecular mechanism of the oncogenic role of CDK1 and identified AKT as a novel substrate of CDK1. CDK1 inhibition decreases AKT phosphorylation. AKT phosphorylation has been shown to play an important role in cancer progression by mediating a variety of biological responses including cell growth, proliferation, and survival (33). Constitutive activation of PI3K/AKT pathway is indispensable for GIST development and maintenance. Accordingly, several studies based on targeted inhibition of PI3K or AKT with GIST xenograft models (36, 37) highlighted PI3K/AKT pathway essentiality in GIST (49). Our studies further established a regulation between CDK1 and PI3K/AKT pathway, providing the preclinical rationale that triggered the clinical development of CDK1–AKT pathway inhibitors.

In summary, we report, for the first time, that CDK1 is frequently overexpressed in advanced GIST and promotes GIST growth and proliferation. CDK1 kinase binds to AKT and regulates its phosphorylation, thereby promoting GIST proliferation and progression. The CDK1 inhibitor RO-3306 shows antitumor activities against imatinib-sensitive and imatinib-resistant GIST xenografts. Together, this evidence supports a personalized therapeutic option for a novel molecular subgroup of patients with CDK1-aberrant GIST receiving CDK1 inhibitor treatment.

X. Lu reports a patent no. 202110172517.9 pending. Y. Wang reports a patent no. 202110172517.9 pending. No disclosures were reported by the other authors.

X. Lu: Conceptualization, data curation, software, formal analysis, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. Y. Pang: Data curation, validation, investigation. H. Cao: Resources, investigation. X. Liu: Data curation, software, investigation. L. Tu: Resources. Y. Shen: Resources, validation, investigation. X. Jia: Investigation, project administration. J.-C. Lee: Resources, methodology. Y. Wang: Conceptualization, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing.

This work was supported by grants from the National Natural Science Foundation of China (82072974, 81572642), the National Key R&D Program of China (2016YFC1302100), the Basic Research Project of Shanghai Science and Technology Commission (20JC1419200, 16JC1405600), the fund of the Key Laboratory of Tissue Microenvironment and Tumor of Chinese Academy of Sciences (No. 202002), the Technology Foundation for Selected Overseas Chinese Scholar, the Ministry of Human Resources and Social Security, China, and the funds by the Chinese Academy of Sciences. We also thank Dr. Jonathan Fletcher at Brigham and Women’s Hospital/Harvard Medical School for critical comments and for GIST430/654, GIST48, GIST882, and GIST48B cell lines, Dr. Takahiro Taguchi for the GIST-T1 cell line. We also thank all members of the Wang laboratory for thought-provoking discussion, technical assistance, and help with manuscript preparation.

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.
Corless
CL
,
Barnett
CM
,
Heinrich
MC
. 
Gastrointestinal stromal tumours: origin and molecular oncology
.
Nat Rev Cancer
2011
;
11
:
865
78
.
2.
Burningham
Z
,
Hashibe
M
,
Spector
L
,
Schiffman
JD
. 
The epidemiology of sarcoma
.
Clin Sarcoma Res
2012
;
2
:
14
.
3.
Heinrich
MC
,
Corless
CL
,
Duensing
A
,
McGreevey
L
,
Chen
CJ
,
Joseph
N
, et al
PDGFRA activating mutations in gastrointestinal stromal tumors
.
Science
2003
;
299
:
708
10
.
4.
Hirota
S
,
Isozaki
K
,
Moriyama
Y
,
Hashimoto
K
,
Nishida
T
,
Ishiguro
S
, et al
Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors
.
Science
1998
;
279
:
577
80
.
5.
Demetri
GD
,
Reichardt
P
,
Kang
YK
,
Blay
JY
,
Rutkowski
P
,
Gelderblom
H
, et al
Efficacy and safety of regorafenib for advanced gastrointestinal stromal tumours after failure of imatinib and sunitinib (GRID): an international, multicentre, randomised, placebo-controlled, phase 3 trial
.
Lancet
2013
;
381
:
295
302
.
6.
Demetri
GD
,
von Mehren
M
,
Blanke
CD
,
Van den Abbeele
AD
,
Eisenberg
B
,
Roberts
PJ
, et al
Efficacy and safety of imatinib mesylate in advanced gastrointestinal stromal tumors
.
N Engl J Med
2002
;
347
:
472
80
.
7.
Heinrich
MC
,
Maki
RG
,
Corless
CL
,
Antonescu
CR
,
Harlow
A
,
Griffith
D
, et al
Primary and secondary kinase genotypes correlate with the biological and clinical activity of sunitinib in imatinib-resistant gastrointestinal stromal tumor
.
J Clin Oncol
2008
;
26
:
5352
9
.
8.
Dolgin
E
. 
Avapritinib approved for GIST subgroup
.
Cancer Discov
2020
;
10
:
334
.
9.
Falkenhorst
J
,
Hamacher
R
,
Bauer
S
. 
New therapeutic agents in gastrointestinal stromal tumours
.
Curr Opin Oncol
2019
;
31
:
322
8
.
10.
Grunewald
S
,
Klug
LR
,
Muhlenberg
T
,
Lategahn
J
,
Falkenhorst
J
,
Town
A
, et al
Resistance to avapritinib in PDGFRA-driven GIST is caused by secondary mutations in the PDGFRA kinase domain
.
Cancer Discov
2020
;
10
:
334
.
11.
Gramza
AW
,
Corless
CL
,
Heinrich
MC
. 
Resistance to tyrosine kinase inhibitors in gastrointestinal stromal tumors
.
Clin Cancer Res
2009
;
15
:
7510
8
.
12.
Liegl
B
,
Kepten
I
,
Le
C
,
Zhu
M
,
Demetri
GD
,
Heinrich
MC
, et al
Heterogeneity of kinase inhibitor resistance mechanisms in GIST
.
J Pathol
2008
;
216
:
64
74
.
13.
Agaimy
A
,
Wunsch
PH
,
Hofstaedter
F
,
Blaszyk
H
,
Rummele
P
,
Gaumann
A
, et al
Minute gastric sclerosing stromal tumors (GIST tumorlets) are common in adults and frequently show c-kit mutations
.
Am J Surg Pathol
2007
;
31
:
113
20
.
14.
Mikami
T
,
Nemoto
Y
,
Numata
Y
,
Hana
K
,
Nakada
N
,
Ichinoe
M
, et al
Small gastrointestinal stromal tumor in the stomach: identification of precursor for clinical gastrointestinal stromal tumor using c-kit and alpha-smooth muscle actin expression
.
Hum Pathol
2013
;
44
:
2628
35
.
15.
El-Rifai
W
,
Sarlomo-Rikala
M
,
Andersson
LC
,
Knuutila
S
,
Miettinen
M
. 
DNA sequence copy number changes in gastrointestinal stromal tumors: tumor progression and prognostic significance
.
Cancer Res
2000
;
60
:
3899
903
.
16.
Wang
YX
,
Marino-Enriquez
A
,
Bennett
RR
,
Zhu
MJ
,
Shen
YP
,
Eilers
G
, et al
Dystrophin is a tumor suppressor in human cancers with myogenic programs
.
Nat Genet
2014
;
46
:
601
6
.
17.
Boehm
JS
,
Zhao
JJ
,
Yao
J
,
Kim
SY
,
Firestein
R
,
Dunn
IF
, et al
Integrative genomic approaches identify IKBKE as a breast cancer oncogene
.
Cell
2007
;
129
:
1065
79
.
18.
Shalem
O
,
Sanjana
NE
,
Hartenian
E
,
Shi
X
,
Scott
DA
,
Mikkelsen
TS
, et al
Genome-scale CRISPR-Cas9 knockout screening in human cells
.
Science
2014
;
343
:
84
7
.
19.
Parnas
O
,
Jovanovic
M
,
Eisenhaure
TM
,
Herbst
RH
,
Dixit
A
,
Ye
CJ
, et al
A genome-wide CRISPR screen in primary immune cells to dissect regulatory networks
.
Cell
2015
;
162
:
675
86
.
20.
Costa-Cabral
S
,
Brough
R
,
Konde
A
,
Aarts
M
,
Campbell
J
,
Marinari
E
, et al
CDK1 is a synthetic lethal target for KRAS mutant tumours
.
PLoS One
2016
;
11
:
e0149099
.
21.
Vassilev
LT
,
Tovar
C
,
Chen
SQ
,
Knezevic
D
,
Zhao
XL
,
Sun
HM
, et al
Selective small-molecule inhibitor reveals critical mitotic functions of human CDK1
.
Proc Natl Acad Sci U S A
2006
;
103
:
10660
5
.
22.
Pang
YZ
,
Xie
FF
,
Cao
H
,
Wang
CM
,
Zhu
MJ
,
Liu
XX
, et al
Mutational inactivation of mTORC1 repressor gene DEPDC5 in human gastrointestinal stromal tumors
.
Proc Natl Acad Sci U S A
2019
;
116
:
22746
53
.
23.
von Mehren
M
,
Randall
RL
,
Benjamin
RS
,
Boles
S
,
Bui
MM
,
Conrad
EU
 3rd
, et al
Soft tissue sarcoma, version 2.2016, NCCN clinical practice guidelines in oncology
.
J Natl Compr Canc Netw
2016
;
14
:
758
86
.
24.
Bauer
S
,
Yu
LK
,
Demetri
GD
,
Fletcher
JA
. 
Heat shock protein 90 inhibition in imatinib-resistant gastrointestinal stromal tumor
.
Cancer Res
2006
;
66
:
9153
61
.
25.
Joung
J
,
Konermann
S
,
Gootenberg
JS
,
Abudayyeh
OO
,
Platt
RJ
,
Brigham
MD
, et al
Genome-scale CRISPR-Cas9 knockout and transcriptional activation screening
.
Nat Protoc
2017
;
12
:
828
63
.
26.
Yerushalmi
R
,
Woods
R
,
Ravdin
PM
,
Hayes
MM
,
Gelmon
KA
. 
Ki67 in breast cancer: prognostic and predictive potential
.
Lancet Oncol
2010
;
11
:
174
83
.
27.
Malumbres
M
,
Barbacid
M
. 
Cell cycle, CDKs and cancer: a changing paradigm
.
Nat Rev Cancer
2009
;
9
:
153
66
.
28.
Ubersax
JA
,
Woodbury
EL
,
Quang
PN
,
Paraz
M
,
Blethrow
JD
,
Shah
K
, et al
Targets of the cyclin-dependent kinase Cdk1
.
Nature
2003
;
425
:
859
64
.
29.
Otto
T
,
Sicinski
P
. 
Cell cycle proteins as promising targets in cancer therapy
.
Nat Rev Cancer
2017
;
17
:
93
115
.
30.
Petrone
A
,
Adamo
ME
,
Cheng
C
,
Kettenbach
AN
. 
Identification of candidate cyclin-dependent kinase 1 (cdk1) substrates in mitosis by quantitative phosphoproteomics
.
Mol Cell Proteomics
2016
;
15
:
2448
61
.
31.
Deota
S
,
Rathnachalam
S
,
Namrata
K
,
Boob
M
,
Fulzele
A
,
Radhika
S
, et al
Allosteric regulation of cyclin-B binding by the charge state of catalytic lysine in CDK1 is essential for cell-cycle progression
.
J Mol Biol
2019
;
431
:
2127
42
.
32.
Wang
XQ
,
Lo
CM
,
Chen
L
,
Ngan
ES
,
Xu
A
,
Poon
RY
. 
CDK1-PDK1-PI3K/Akt signaling pathway regulates embryonic and induced pluripotency
.
Cell Death Differ
2017
;
24
:
38
48
.
33.
Hoxhaj
G
,
Manning
BD
. 
The PI3K-AKT network at the interface of oncogenic signalling and cancer metabolism
.
Nat Rev Cancer
2020
;
20
:
74
88
.
34.
Sergina
NV
,
Rausch
M
,
Wang
D
,
Blair
J
,
Hann
B
,
Shokat
KM
, et al
Escape from HER-family tyrosine kinase inhibitor therapy by the kinase-inactive HER3
.
Nature
2007
;
445
:
437
41
.
35.
Herceg
Z
,
Wang
ZQ
. 
Functions of poly(ADP-ribose) polymerase (PARP) in DNA repair, genomic integrity and cell death
.
Mutat Res
2001
;
477
:
97
110
.
36.
Van Looy
T
,
Wozniak
A
,
Floris
G
,
Sciot
R
,
Li
H
,
Wellens
J
, et al
Phosphoinositide 3-kinase inhibitors combined with imatinib in patient-derived xenograft models of gastrointestinal stromal tumors: rationale and efficacy
.
Clin Cancer Res
2014
;
20
:
6071
82
.
37.
Zook
P
,
Pathak
HB
,
Belinsky
MG
,
Gersz
L
,
Devarajan
K
,
Zhou
Y
, et al
Combination of imatinib mesylate and AKT inhibitor provides synergistic effects in preclinical study of gastrointestinal stromal tumor
.
Clin Cancer Res
2017
;
23
:
171
80
.
38.
Nishida
T
,
Doi
T
,
Naito
Y
. 
Tyrosine kinase inhibitors in the treatment of unresectable or metastatic gastrointestinal stromal tumors
.
Expert Opin Pharmacother
2014
;
15
:
1979
89
.
39.
Smith
BD
,
Kaufman
MD
,
Lu
WP
,
Gupta
A
,
Leary
CB
,
Wise
SC
, et al
Ripretinib (DCC-2618) is a switch control kinase inhibitor of a broad spectrum of oncogenic and drug-resistant KIT and PDGFRA variants
.
Cancer Cell
2019
;
35
:
738
51
.
40.
Blay
JY
,
Serrano
C
,
Heinrich
MC
,
Zalcberg
J
,
Bauer
S
,
Gelderblom
H
, et al
Ripretinib in patients with advanced gastrointestinal stromal tumours (INVICTUS): a double-blind, randomised, placebo-controlled, phase 3 trial
.
Lancet Oncol
2020
;
21
:
923
34
.
41.
Michowski
W
,
Chick
JM
,
Chu
C
,
Kolodziejczyk
A
,
Wang
Y
,
Suski
JM
, et al
CDK1 controls global epigenetic landscape in embryonic stem cells
.
Mol Cell
2020
;
78
:
459
76
.
42.
Dickson
MA
,
Tap
WD
,
Keohan
ML
,
D’Angelo
SP
,
Gounder
MM
,
Antonescu
CR
, et al
Phase II trial of the CDK4 inhibitor PD0332991 in patients with advanced CDK4-amplified well-differentiated or dedifferentiated liposarcoma
.
J Clin Oncol
2013
;
31
:
2024
8
.
43.
Assi
T
,
Kattan
J
,
Rassy
E
,
Nassereddine
H
,
Farhat
F
,
Honore
C
, et al
Targeting CDK4 (cyclin-dependent kinase) amplification in liposarcoma: a comprehensive review
.
Crit Rev Oncol Hematol
2020
;
153
:
103029
.
44.
Hortobagyi
GN
,
Stemmer
SM
,
Burris
HA
,
Yap
YS
,
Sonke
GS
,
Paluch-Shimon
S
, et al
Ribociclib as first-line therapy for HR-positive, advanced breast cancer
.
N Engl J Med
2016
;
375
:
1738
48
.
45.
Patnaik
A
,
Rosen
LS
,
Tolaney
SM
,
Tolcher
AW
,
Goldman
JW
,
Gandhi
L
, et al
Efficacy and safety of abemaciclib, an inhibitor of CDK4 and CDK6, for patients with breast cancer, non-small cell lung cancer, and other solid tumors
.
Cancer Discov
2016
;6:
740
53
.
46.
Kwapisz
D
. 
Cyclin-dependent kinase 4/6 inhibitors in breast cancer: palbociclib, ribociclib, and abemaciclib
.
Breast Cancer Res Treat
2017
;
166
:
41
54
.
47.
Goga
A
,
Yang
D
,
Tward
AD
,
Morgan
DO
,
Bishop
JM
. 
Inhibition of CDK1 as a potential therapy for tumors over-expressing MYC
.
Nat Med
2007
;
13
:
820
7
.
48.
Johnson
N
,
Li
YC
,
Walton
ZE
,
Cheng
KA
,
Li
DA
,
Rodig
SJ
, et al
Compromised CDK1 activity sensitizes BRCA-proficient cancers to PARP inhibition
.
Nat Med
2011
;
17
:
875
82
.
49.
Garcia-Valverde
A
,
Rosell
J
,
Serna
G
,
Valverde
C
,
Carles
J
,
Nuciforo
P
, et al
Preclinical activity of PI3K inhibitor copanlisib in gastrointestinal stromal tumor
.
Mol Cancer Ther
2020
;
19
:
1289
97
.