CDK1 serves as a therapeutic target of adrenocortical carcinoma via regulating epithelial–mesenchymal transition, G2/M phase transition, and PANoptosis

CDK1 expression in normal and tumor group was investigated in the following databases: http://​gepia.​cancer-pku.​cn/​ [17], http://​gdac.​broadinstitute.​org/​ [18] and https://​www.​oncomine.​org/​ [19]. The mRNA level of CDK1 in ACC was analyzed on the website, http://​gepia.​cancer-pku.​cn/​, http://​ualcan.​path.​uab.​edu/​index.​html [20]. The prognostic value of CDK1 for ACC was analyzed on the website, http://​sangerbox.​com/​Index.

The human ACC cell lines, NCI-H295R and SW-13, were acquired from Procell Life Science &Technology (Wuhan, China). SW-13 cells were maintained in Leibovitz's L-15, with 10% FBS and 1% P/S (Procell, CM-0451). NCI-H295R cells were maintained in DMEM/F12, with 6.25 μg/mL insulin, 6.25 ng/mL selenium, 6.25 μg/mL transferrin, 5.35 μg/mL linoleic acid, 10% FBS and 1% P/S (Procell, CM-0399).

For stable CDK1 overexpression in SW-13 cells (SW13_CDK1), the pSLenti-SFH-EGFP-P2A-Puro-CMV-CDK1-3xFLAG-WPRE and corresponding control lentivirus were obtained from Hanbio (Shanghai, China). After transfection into SW-13 for 72 h, 2 μg/mL puromycin was added to the culture medium screen for stable cloning. Two shRNAs were used to construct a stable CDK1-knockdown NCI-H295R cell line (NCI-H295R_shCDK1). The sequences of the two shRNAs were as follows: 5′-GTGGAATCTTTACAGGACTAT-3′ and 5′-GATTCAGAAATTGATCAACTC-3′. Puromycin (2 μg/mL) was added to cultures to obtain stable CDK1 knockdown in NCI-H295R cells.

The inhibitors used in the pharmacological activity test for ACC were obtained from MedChemExpress (New Jersey, USA). The CDK1 inhibitors were prazosin hydrochloride, HY-B0193A, 19237-84-4; cucurbitacin E (CurE), HY-N0417, 18444-66-1; Ro-3306, HY-12529, 872573-93-8; and AZD-5438, HY-10012, 602306-29-6; the GSK3 inhibitors were IX, HY-10580, 667463-62-9 and mitotane, HY-13690, 53-19-0; the pan-caspase inhibitor was Z-VAD-FMK (VAD), HY-16658B, 161401-82-7; the inhibitor of autophagy was 3-methyladenine (3-MA), HY-19312, 5142-23-4; and the necroptosis inhibitor was necrosulfonamide (Nec), HY-100573, 1360614-48-7. Additional reagents were cisplatin (Pt), HY-17394, 15663–27-1; paclitaxel (Ptx), HY-B0015, 33069-62-4; and cycloheximide (CHX), HY-12320, 66-81-9.

Cell proliferation was measured by CCK8 kit (Dojindo, Kumanmoto, Japan). SW-13_CDK1, NCI-H295R_shCDK1 and corresponding control cells were seeded in 96-well plates at 3 × 10³ cells per well and cultured for 24, 48 and 72 h. Subsequently, CCK8 reagent was added (10 μL/well) and after incubation for 1 h, the absorbance at 450 nm was measured. To calculate the IC₅₀ of CDK1 inhibitors, cells were cultured in 96-well plates with 3 × 10³ cells per well at 37 °C with different concentration of the CDK1 inhibitor, CurE (0–10 μmol/L). After culturing for 24, 48 and 72 h, 10 μL of CCK8 reagent was added per well. After incubation for 1 h, the absorbance at 450 nm was measured.

NCI-H295R and SW-13 cells were pretreated with different inhibitors including Z-VAD-FMK (50 μmol/L), 3-MA (2 mmol/L), and necrosulfonamide (2 μmol/L) for 2 h. Afterwards, the SW-13 cells were incubated with 0.1 μmol/L CurE and NCI-H295R cells were treated with 1 μmol/L CurE for 24 h. The cell viability was measured according to the above method.

The EdU kit was used according to the manufacture’s protocol. SW-13_CDK1, NCI-H295R_shCDK1, and the corresponding control cells were cultured for 24 h. Afterwards, the EdU reagent (50 μmol/L final) was added to each and incubation at 37 °C was continued for another 24 h, after which the cells were fixed, permeabilized and stained with Hoechst 33342 and Apollo 567.

A lactate dehydrogenase (LDH) detection kit (Dojindo, Japan) was used to measure the level of LDH released from cells. Cells were seeded at 3 × 10³/well in 96-well plates and treated with CurE for 24 h and 48 h. After incubation with CurE, the medium was discarded and 100 μL of working reagent was added to each well. Thirty minutes later, the reaction was ended with the stop solution and the activity was measured at 490 nm.

The roles of CDK1 and CurE in invasion and migration of NCI-H295R and SW-13 cells were evaluated using a 6.5 mm transwell (#3422, Corning Costar, NY, USA). For the migration assay, 2 × 10⁴ cells were put into the upper chamber. For the invasion assay, the transwell chamber was coated with Matrigel dissolved in Biocoat (Corning USA). The Matrigel was diluted with FBS-free medium at a ratio of 1:6 and then 1 × 10⁵ cells were added into the transwell chamber. After incubation, the cells that crossed the insert were fixed, stained with 1% crystal violet, photographed, and counted under a visible light microscope (Nikon Eclipse Ti-U, Tokyo, Japan).

Cell apoptosis was determined using the annexin V-FITC/PI detection kit (Solarbio, Beijing, China). SW-13 cells were incubated with 0, 0.03, 0.1 and 0.3 μmol/L CurE and NCI-H295R cells with 0, 0.3, 1 and 3 μmol/L CurE for 24 and 48 h. After incubation, the cells were harvested and stained with annexin V-FITC and PI for 10 min. Flow cytometry (FACSVERSE, BD Biosciences, New Jersey, USA) was used to determine the percentage of early apoptotic cells, late apoptotic cells, pyroptotic cells and necrotic cells. The data on apoptosis rate was quantified by FlowJo software (v10). ACC cells were pretreated with necrosulfonamide (2 μmol/L) for 2 h and then treated with CurE for 24 h. The percentages of early apoptotic, late apoptotic, pyroptotic and necrotic cells were calculated using the above method.

SW-13 cells were incubated with 0, 0.03, 0.1 and 0.3 μmol/L CurE and NCI-H295R cells with 0, 0.3, 1 and 3 μmol/L CurE for 24 and 48 h. Then cells were harvested by centrifugation and fixed in 75% ethanol for 24 h at 4 °C. After centrifugation, cells were stained with PI for 20 min, fluorescence was measured, and data was processed with FlowJo software (v10).

SW-13_CDK1 cells, SW-13 cells administrated with 0.1 μmol/L CurE for 24 h and corresponding control cells were collected in TRIzol reagent. The next generation sequencing of the prepared isolates was conducted on an Illumina sequencer in Novogene Corporation (Beijing, China). The integrity and concentration of total RNA were qualified by the Agilent 2100 bioanalyzer and NanoDrop spectrophotometer. Illumina sequencer was used to build and qualify, then pool and sequence for the establishment of the New England Biolabs (NEB) libraries. The Gene ontology (GO) analysis were conducted on the DAVID database (v6.8) (https://​david.​ncifcrf.​gov/​).

To knock down Slug, Twist and ZBP1 expression, 100 nmol/L relative target siRNA and relative control siRNA (Genechem, Suzhou, China) were transfected into NCI-H295R and SW-13 cells using Lipofectamine 3000 reagent and cultured for 48 h at 37 °C. After transfection, SW-13 cells were incubated with 0.1 μmol/L CurE and NCI-H295R cells were incubated with 1 μmol/L CurE for 24 h.

The protein lysates of control and CurE treated cells were obtained as described. Equal quantities of proteins were incubated with CDK1 antibodies or IgG control and rotated overnight at 4 ℃. Protein complexes were collected with protein A/G agarose beads (Beyotime Technology) followed by washing 5 times and prepared for Western blotting.

RIPA reagent with protease and phosphatase inhibitor cocktail was used to extract total protein (Applygen, Beijing, China) for 30 min on ice. The lysates were centrifuged at 12,000 g for 15 min at 4 °C and supernatants were collected. Equivalent amounts of protein were separated by SDS-PAGE, transferred to PVDF membranes, and incubated with primary antibodies (Additional file 8: Table S1) at 4 °C for 12 h. The blots were then washed and incubated with HRP-conjugated secondary antibodies. After washing, the blots were incubated with SuperECL reagent, protein bands were visualized, and image densities were quantitated using a Tanon imaging system.

All animal experiments were carried out following the NIH Guide for the Care and Use of Laboratory Animals. Female BALB/c-nu nude mice (16–18 g) were obtained from the Charles River Co. (Beijing, China) and housed at the Institute of Materia Medica animal barrier facility.

In the first part, mice were randomly divided into six per group, and aliquots of 1 × 10⁶ SW-13_NC, SW-13_CDK1, NCI-H295R_shNC, or NCI-H295R_shCDK1 cells were injected subcutaneously into the right flank (n = 6). The tumor volume (mm³) and body weight (g) were measured every 2 days. At the end of experiment, the mice were euthanized, and the tumors were excised, weighed and photographed.

To determine the effect of CDK1 inhibitor treatment, SW-13 cells (1 × 10⁸ cells/mL) were subcutaneously injected into the right sides of the mice. After 14 days, when the tumor volume had reached about 100 mm³, the mice were randomly assigned to five groups (n = 6) for the following treatments: (1) no treatment, (2) 40 mg/kg mitotane i.p., (3) 10 mg/kg CurE i.p., (4) 20 mg/kg CurE, i.p., and (5) 10 mg/kg CurE plus 40 mg/kg mitotane, i.p., once a day for 3 weeks. Body weight (g) and tumor volume (V, mm³) were measured every 2 days. All the mice were euthanized after 3 weeks, and the organs and tumors were collected, weighed and photographed.

Immunohistochemistry (IHC) and immunofluorescence (IF) were carried out using kits according to manufacturer’s instructions. The expression of Slug and Twist was detected via IHC staining in the tumor tissues formed from SW-13_NC, SW-13_CDK1, NCI-H295R_shNC, NCI-H295R_shCDK1 cells in the xenograft assays. The expression of ZBP1 was detected in tumor tissues from mice treated with 10 mg/kg CurE (low) and 20 mg/kg CurE (high) via IF staining. A Nikon Eclipse Ti microscope was used to image the stained sections.

To investigate the expression pattern of CDK1 in ACC, bioinformatic analysis was conducted on the GTEx, TCGA, and Oncomine databases. Results suggested that CDK1 was abnormally elevated in the majority of tumors relative to healthy controls (Additional file 1: Fig. S1 a–b). CDK1 was abnormally elevated in ACC (Additional file 1: Fig. S1c) and the elevated expression of CDK1 closely tracked with the pathological stage (Additional file 1: Fig. S1d) and nodal metastatic status of ACC patients (Additional file 1: Fig. S1e). Correlation analysis results also indicated that CDK1 expression was related to TNM stage (P < 0.001), tumor status (P < 0.001) and pathological stage (P < 0.001). In addition, the expression level of CDK1 was associated with the outcome of primary therapy (P < 0.001), overall survival (OS) (P < 0.001), disease-specific survival (DSS) (P < 0.001) and progression-free interval (PFI) (P < 0.001) (Table 1). The Kaplan-Meier analysis also indicated a relationship between CDK1 expression and OS (P < 0.0001), DSS (P < 0.0001) and PFI (P < 0.0001) survival probability of ACC patients (Additional file 1: Fig. S1f–h). ROC curve of CDK1 expression and OS, DSS and PFI survival suggested that CDK1 could be used as a prognostic indicator of ACC (Additional file 1: Fig. S1i–k). The above results indicated that CDK1 was abnormally elevated in ACC and that its expression level could serve as a prognostic biomarker as well as a potential therapeutic target for ACC.

To assess the influence of CDK1 on the proliferation of ACC cells, the growth in liquid medium, 3D Matrigel, soft agar, and nude mice was evaluated. Upregulation of CDK1 increased the growth of SW-13 cells in culture medium (Fig. 1a–b), promoted colony formation in soft agar (Fig. 1c), invasiveness in 3D Matrigel (Fig. 1d), and the growth of tumors in nude mice (Fig. 1e). In contrast, downregulation of CDK1 in NCI-H295R cells remarkably suppressed their growth (Fig. 1f–g), colony formation in soft agar (Fig. 1h), and invasiveness in 3D Matrigel (Fig. 1i). Downregulating CDK1 in NCI-H295R cells also suppressed the growth of these tumor cells in nude mice (Fig. 1j). Thus, CDK1 is significantly associated with the proliferation of ACC cells and could be utilized as a potential target for ACC therapy.

The expression of CDK1 in ACC cell lines was determined by Western blotting, and the results showed that CDK1 expression was lower in SW-13 cells than in NCI-H295R cells (Additional file 2: Fig. S2a). To investigate the underlying mechanism of CDK1’s effects in ACC, the sequences of RNA isolated from cells in the SW-13_CDK1 group were compared to those from the SW-13_NC group. There were 66 significantly upregulated genes and 104 downregulated genes in the SW-13_CDK1 group (Additional file 2: Fig. S2b). The GO enrichment results of differentially-expressed genes (DEGs) suggested that CDK1 participated in cell-cell adhesion, muscle structure development, angiogenesis and cell proliferation. (Additional file 2: Fig. S2c). KEGG pathway enrichment of DEGs also indicated that CDK1 could regulate cell adhesion molecules, epithelial cell signaling, and cell cycle pathways (Additional file 2: Fig. S2d). Reactome enrichment indicated that CDK1 might play an essential role in GPCR signaling (Additional file 2: Fig. S2e). Based on these RNA-seq results, we speculated that CDK1 played a vital part in the proliferation and metastasis of ACC cells.

To further evaluate the potential of CDK1 as a therapeutic target for ACC, we screened several CDK1 inhibitors for antitumor activity on ACC cell lines. Results revealed that CurE had the best inhibitory activity with good dose-dependency on SW-13 and NCI-H295R cells (Fig. 3a). Subsequent experiments verified that CurE (Fig. 3b) could time- and dose-dependently suppress the growth of NCI-H295R and SW-13 cells at nanomolar concentrations (Fig. 3c). CurE also inhibited colony formation of SW-13 and NCI-H295R cells (Fig. 3d) in the soft agar assay and suppressed their invasion and migration (Fig. 3e). CurE was tested in a nude mouse xenograft model of ACC (Fig. 3f) and it significantly and dose-dependently suppressed tumor growth in comparison with the vehicle group. The tumor weight and volume were reduced after treatment of CurE, and the therapeutic effect of 20 mg/kg of CurE (ip) was significantly better than that of the positive control drug, mitotane, at 40 mg/kg (ip). More importantly, the tumors were almost gone after 21 days of co-administration of 10 mg/kg CurE and 40 mg/kg mitotane (Fig. 3g–i). The body weight (Fig. 3j), morphological characteristics of organs including liver, heart, kidney, spleen, and lung remained stable (Fig. 3k) under the administration of CurE. No noticeable toxicity was identified through physiological blood tests (Additional file 3: Fig. S3). From the above results, we conclude that CurE can suppress the proliferation of ACC cells without observable toxic effects.

To investigate the molecular mechanism of CurE’s antitumor effect against ACC, RNA-seq was carried out. We found 115 downregulated genes and 274 upregulated genes in SW-13 cells treated with CurE (Additional file 4: Fig. S4a). GO enrichment of the DEGs suggested that CurE principally regulated the apoptotic signaling pathway, the G2/M phase transition in the cell cycle, epithelial cell–cell adhesion, cell proliferation, and angiogenesis pathways (Additional file 4: Fig. S4b). The KEGG pathway enrichment indicated that CurE could regulate the cell cycle, the synthesis and metabolism of amino acids, cell adhesion molecules, P53 and PPAR signaling pathway, DNA replication, necroptosis and cellular senescence (Additional file 4: Fig. S4c). Reactome enrichment suggested that CurE participated in the S phase, G2/M checkpoints, MyD88-related pathway, CDK-mediated phosphorylation, antigen processing, proteasome degradation, apoptosis and programmed cell death (PCD) (Additional file 4: Fig. S4d). Overall, these results suggested that CurE could inhibit ACC cells by regulating the cell cycle, metastasis, and PCD.

The effect of CurE on morphology was observed in NCI-H295R and SW-13 cells at 3, 6 and 12 h. Spherical membrane protrusions were observed in NCI-H295R and SW-13 cells (Fig. 5a). CurE time- and dose-dependently promoted the release of LDH in SW-13 and NCI-H295R cells (Fig. 5b), corresponding to increased damage to the plasma membrane allowing leakage. The cytotoxicity of CurE could be partially rescued by co-administration of the apoptosis/pyroptosis inhibitor, VAD, and the necroptosis inhibitor, Nec, but it could not be reversed by using them alone or using the autophagy inhibitor, 3-MA (Fig. 5c). PI-annexin V staining was used to determine the type of cell death. The early phase of apoptosis was represented by single-positive staining for annexin V, while the late phase of apoptosis or pyroptosis showed double-positive staining [21]. Annexin V-PI double staining suggested that CurE time- and dose-dependently increased the percentage of necrotic and pyroptotic/apoptotic cells of SW-13 (Fig. 5d) and NCI-H295R (Fig. 5e).

Western blotting results indicated that the anti-apoptotic protein, Bcl-2, was remarkably reduced after treatment with CurE, while the pro-apoptotic proteins, Bax and Bim, were elevated in comparison with the control group. CurE increased the expression level of cleaved caspase 3 (CASP3), cleaved PARP and cleaved caspase 7 (CASP7) in SW-13 (Fig. 5f) and NCI-H295R (Fig. 5i) cells in a time-dependent manner. The characteristics of pyroptosis include gasdermin family-regulated pore generation in the cytomembrane, cell turgescence, and cytomembrane disruption, together with leakage of the proinflammatory cytokines [22]. Caspase 1 could cleave gasdermin D (GSDMD-FL) to produce an N-terminal cleavage product (GSDMD-N) under activation, which could induce pyroptosis by generating cytomembrane pores and leaking inflammatory cytokines [21]. Nuclear factor-κB (NF-κB) chiefly facilitates the production of proinflammatory effectors, like NLRP3, Myd88 and caspase 1 [23]. The results also indicated that CurE time-dependently elevated the expression of NF-κB and caspase 1 in SW-13 (Fig. 5g) and NCI-H295R (Fig. 5j) cells and significantly upregulated the expression of GSDMD-N. After treatment with CurE, the expression of Myd88 was found to be significantly increased, whereas no significant changes were seen in NLRP3.

Necroptosis is a caspase-independent necrotic cell death, which is mediated by strictly regulated cellular signaling pathways. It is chiefly regulated via receptor-interacting protein 3 (RIP3), RIP1, and mixed lineage kinase domain-like (MLKL) [24]. ZBP1 can induce PANoptosis of cells when it senses the Z-RNA produced during infection with influenza virus [25]. Results of Western blotting suggested that the expression levels of RIP1, RIP3 and MLKL were remarkably elevated after administration of CurE, and that the ZBP1 expression was significantly elevated in comparison with the control group (Fig. 5h, k). Based on the above results, CurE could induce PANoptosis of ACC cells in a caspase-independent way.

To further explore the mechanism of CDK1 in regulating PANoptosis, we evaluate the relationship between CDK1 and the PANoptosome. Co-IP experiments showed that CurE could induce the formation of PANoptosomes and that CDK1 could bind to the PANoptosome in SW-13 cells (Fig. 6a) and NCI-H295R cells (Fig. 6b). Immunofluorescence experiments showed that CDK1 co-localized with ZBP1 in SW-13 (Fig. 6c) and NCI-H295R cells (Fig. 6d). Silencing ZBP1 could rescue the cell death triggered by CurE in SW-13 and NCI-H295R cells (Fig. 6e). And when ZBP1 was knockdown, it could inhibit the marker proteins of apoptosis, pyroptosis, and necroptosis which were triggered by CurE (Fig. 6f). In addition, when caspase 3 was knocked down, the percentage of necrotic cells was raised after treatment with CurE, whereas the overall cell survival rate didn’t significantly change in NCI-H295R and SW-13 cells. Similarly, when GSDMD was knocked down, the proportion of necrotic cells was increased, but the overall cell survival didn’t change significantly (Additional file 6: Fig. S6a–h). The above results indicated that when the pathways of apoptosis and pyroptosis were inhibited, CurE was able to suppress the proliferation of ACC cells via triggering cell necroptosis instead. These results also suggested that CurE inhibited the proliferation of ACC cells through triggering PANoptosis. Furthermore, immunofluorescence experiments also showed that ZBP1 was upregulated with administration of CurE in vivo (Fig. 6g). In addition, the experimental results showed that knockdown of CDK1 could significantly reverse the inhibitory effect of CurE on ACC cell lines (Additional file 7: Fig. S7a–f), which indicated that the above effects of CurE were due to the CDK1 inhibition. In conclusion, CDK1 regulated the PANoptosis of ACC cells through binding with the PANoptosome in a ZBP1-dependent way.