1 Introduction
Esophageal cancer is a global health issue, ranking sixth as a cause of cancer-related deaths worldwide [
1]. Esophageal squamous cell carcinoma (ESCC) is one of the main subtypes of esophageal cancer. In 2020, approximately 544 000 people died worldwide from ESCC, with 90% of the deaths occurring in developing countries [
1,
2]. Despite established combination therapy regimens, the patient’s prognosis remains poor [
3,
4]. In addition, half of the patients still have inoperable or metastatic disease, making ESCC incurable [
5]. Therefore, new strategies are urgently needed to improve the prognosis of patients with ESCC.
Developing new drugs for therapy or chemoprevention is time-consuming and expensive [
6–
8], taking approximately 10–15 years from compound discovery to drug approval [
9]. Therefore, screening FDA-approved drugs is an effective and cost-reduction strategy. This approach has two main advantages: it shortens the drug development timeline and reduces risk because preclinical trials have already been performed [
10]. Here, we found that dronedarone, a non-iodinated benzofuran derivative of amiodarone, potently inhibited the proliferation of ESCC by screening the FDA-approved drug library. Dronedarone has been previously reported to control ventricular responses and maintain sinus rhythm during atrial fibrillation episodes [
11]. However, the mechanism behind its inhibitory effect on ESCC has not yet been investigated.
Cyclin-dependent kinases 4 and 6 (CDK4/6) control cell cycle progression from the G1 to S phase. D-type cell cycle proteins (D1, D2, and D3) bind to CDK4/6 before DNA synthesis [
12]. The CDK4/6-cytosolic protein D complex phosphorylates RB1 to regulate the progression from G1 to the S phase after entering the nucleus. Evidence has shown that inhibitors of CDK4/6 have therapeutic effects on breast cancer [
13,
14]. Nevertheless, the clinical applications of these treatments in other types of tumors have not been thoroughly investigated.
In our study, we illustrated that dronedarone, as an inhibitor of CDK4/6, effectively inhibited the CDK4/6-RB1 axis, blocking the cell cycle at the G1/S phase and preventing ESCC cell proliferation. Notably, dronedarone inhibited the growth of ESCC tumors via the CDK4/6-RB1 axis in vivo. Thus, our work confirmed that dronedarone is a promising candidate for suppressing the development of ESCC and may prove to be an effective treatment for this type of cancer.
2 Materials and methods
2.1 Reagents and antibodies
Dronedarone was purchased from Shanghai Cheng Shao Biotechnology Co. It was then dissolved in saline. CDK4 antibody was purchased from Cell Signaling Technology (Boston, USA); CDK6, RB1, p-RB1 S807, p-RB1 S811, and p-RB1 T826 antibodies were purchased from Abcam (Cambridge, the UK). RB1 kinase was purchased from the Proteintech Group (Wuhan, China; Table S1).
2.2 Cell culture
The human ESCC cell lines KYSE150 and KYSE450 were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The immortalized human fetal esophageal epithelium (SHEE) cell by the expression of HPV-18 oncoproteins E6 and E7 was gifted by Prof. Enmin Li from Shantou University. SHEE cells were used for the 25th–31st passage without the malignant phenotype [
15,
16]. KYSE150 cells were cultured in RPMI-1640 medium, whereas KYSE450 and SHEE cells were cultured in DMEM. All cells were cultured in medium containing 10% FBS, 1% penicillin, and 1% streptomycin and incubated at 37 °C with 5% CO
2. Before culturing, all cells were verified to be mycoplasma-free and analyzed by STR analysis (Supplementary Material 1).
2.3 Cytotoxicity assay
SHEE cells (8000 cells/100 μL/well), KYSE150 cells (8000 cells/100 μL/well), and KYSE450 (12 000 cells/100 μL/well) were evenly spread in 96-well plates separately with three replicate wells for each group. The cells were treated with different concentrations of dronedarone (0, 6.25, 12.5, 25, and 50 μmol/L) at 37 °C for culture. The cells were washed with PBS and fixed with 4% paraformaldehyde for 30 min after 24 or 48 h. They were stained with DAPI solution for 1 h before being analyzed for cell number using the IN Cell Analyzer 6000. Cell viability was calculated by comparing the treatment group with the control group and multiplied by 100%. All cytotoxicity assays were performed by three biological replicates.
2.4 Cell proliferation assay
SHEE cells (3000 cells/100 μL/well), KYSE150 cells (3000 cells/100 μL/well), and KYSE450 (3000 cells/100 μL/well) were evenly spread in 96-well plates separately with three replication wells in each group. The cells were treated with different concentrations of dronedarone (0, 1, 1.5, 2, and 2.5 μmol/L) at 37 °C for 0, 24, 48, 72, and 96 h. The cells were either counted in IN Cell Analyzer 6000 or detected through MTT (0.5 mg/mL) after being stained with DAPI solution for 1 h. The EdU incorporation assay was used to evaluate cell proliferation. All cell proliferation assays were performed with three biological replicates.
2.5 Anchorage-independent cell growth assay
KYSE150 and KYSE450 cells at the logarithmic growth stage were selected for anchorage-independent cell growth assay. The cells were resuspended with FBS–BME, and the concentration was reset to 2.4 × 104 cells/mL. KYSE150 and KYSE450 cells were seeded in the top gum with dronedarone (0, 1, 1.5, 2, and 2.5 μmol/L) and incubated at 37 °C for a week. The colonies were measured in IN Cell Analyzer 6000. All these assays were performed with three biological replicates.
2.6 Colony formation assay
KYSE150 cells (200 cells/2 mL/well) and KYSE450 cells (400 cells/2 mL/well) were evenly spread in 6-well plates. The cells were treated with dronedarone (0, 1, 1.5, 2, and 2.5 μmol/L) at 37 °C for 10 days. The cells were fixed with 4% paraformaldehyde for 30 min, followed by staining with 0.1% crystal violet staining solution. The colonies were photographed using a Nikon D3500. All colony formation assays were performed with three biological replicates.
2.7 Cell sample preparation and phosphorylated proteomics analysis
Cell lysates were prepared from KYSE150 cells treated with DMSO or 2.5 μmol/L dronedarone. Proteins were collected after 4 h of lysis with RIPA lysis solution (with a PMSF ratio of 100:1). The collected material was fractionated by high-pH Agilent 300 to enlarge the C18 column for reverse-phase HPLC (5 µm particles, 4.6 mm, 250 mm length). The peptide was fractionated into 60 parts by using an acetonitrile gradient of 8% to 32% (pH 9.0). The peptides were then separated using an ultra-high-performance liquid system and ionized by injecting them into the ion source. Finally, the peptides were examined using Orbitrap-Fusion-Lumos mass spectrometry.
2.8 Western blot
The protein concentration was measured using the BCA kit. SDS-PAGE gel electrophoresis was conducted at 80 V for 100 min to separate each protein sample. The protein samples were transferred onto 0.45 μm PVDF membranes for 90 V for 120 min. The membranes were sealed with 5% skim milk for 2 h before incubating with the primary antibodies for 12–16 h at 4 °C. Thereafter, the secondary antibody was incubated for 2 h at room temperature. Exposure was performed using a chemiluminescence instrument. All Western blot assays were carried out three times. The raw data for all Western blot assays are provided in Supplementary Material 2.
2.9 Computational docking model
The molecular structure files for CDK4 and CDK6 were downloaded from the PDB website. The 3D structures of dronedarone were prepared beforehand using Schrödinger software. Molecular docking simulations were performed between dronedarone and CDK4 (PDB: 2W99) or CDK6 (PDB: 3NUP). The molecular structure with the highest simulated docking score was selected to identify the binding sites.
2.10 Protein purification
The CDK4/CDK6 plasmids were purchased from You Bio (Changsha, China). These plasmids were designed using the mature peptide coding gene of CDK4/CDK6 as a template with the restriction sites marked as underlined. The primers used were ordered from Sangon Biotech (Shanghai, China), and their details are shown in Table S2. The target genes were obtained through PCR amplification. The prokaryotic expression vector pet28a was cut by restriction enzymes separately. The constructed fusion expression plasmids (CDK4-pet28a and CDK6-pet28a) were then transformed into Escherichia coli Rosetta cells. The single-cell clone was picked and incubated in liquid LB medium for 10–12 h. Induction was performed by adding 0.5 mM IPTG and incubating at 4 °C with 180 rpm/min for 16 h. The bacteria were then collected and centrifuged at 12 000 rpm for 20 min at 4 °C. The protein was purified in the ÄKTA Purifier 100 protein purification system. The purified CDK4 or CDK6 protein was then separated using Superdex 75 10/300 GL, and its purity was analyzed by SDS-PAGE.
2.11 Pull-down assay
The protein (500 μg) was mixed with dronedarone-Sepharose 4B (or DMSO-Sepharose 4B) beads and placed in the reaction buffer. The treatment buffer contained 50 mmol/L Tris-HCl, pH 7.5, 5 mmol/L EDTA, 150 mmol/L NaCl, 1 mmol/L DTT, 0.01% NP 40, 0.2 mmol/L PMSF, and 1× protease inhibitor. The mixture was shaken overnight at 4 °C before the beads were washed three times with washing buffer (50 mL). Finally, the binding efficiency was visualized by Western blot.
2.12 In vitro kinase assay
Inactive RB1 proteins (500 ng) were incubated with active CDK4 (100 ng) or CDK6 (100 ng) for in vitro kinase assays. The kinase reaction system was divided into two parts: (1) 1 μL of CDK4 (100 ng) or CDK6 (100 ng), 1 μL of dronedarone (0, 1, 1.5, 2, and 2.5 μmol/L), and 13 μL of kinase buffer; and (2) 1 μL of RB1 proteins (500 ng), 1 μL of ATP (200 mmol/L), and 8 μL of kinase buffer. Subsequently, the samples were boiled for 5 min at 95 °C after adding 5 μL of 6× loading buffer. Finally, 20 μL of the samples was obtained, and the kinase efficiency was visualized by Western blot.
2.13 Cell cycle assay
KYSE150 (3 × 105 cells/dish) and KYSE450 (3 × 105 cells/dish) were sown in the dishes. The cells were first deprived of nutrients for 24 h and then cultured in serum-containing media for another 24 h. The cells were treated with propidium iodide for 30 min and RNAase for 1 h. The outcome was evaluated using flow cytometry.
2.14 Construction of ESCC knockout cell lines
The most efficient sgRNA sequences of CDK4 and CDK6 (Table S2) were selected. Their accuracy was verified by comparing them with the NCBI database. The sgCDK4 and sgCDK6 sequences were then cloned into the lentiCRISPR-V2 vector. For lentiviral infection, logarithmic growth phase cells from KYSE150 and KYSE450 were selected. After 24 h of treatment, the cells were screened using the puromycin-containing medium. The cells with stable knockout sequences were obtained by continuing the culture.
2.15 Construction of an overexpressed RB1 cell line
The Homo sapiens RB1 overexpression plasmid was purchased from HonorGene (Changsha, China). For lentiviral infection, logarithmic growth phase cells from the knock-out CDK4 and CDK6 cells were selected. After 24 h of treatment, the cells were screened using the puromycin-containing medium. The cell lines with stable overexpression of RB1 were obtained by continuing the culture.
2.16 In vivo experiments
The tumor tissue was cut into 0.1 g pieces and transplanted into SCID/CB17 immunodeficient mice (aged 5 to 6 weeks; provided by Beijing Viton Lever Laboratory Animal Technology Co., Ltd.) to establish the ESCC patient-derived xenograft (PDX) model [
17]. The mice had unrestricted access to food and water while housed in a 12/12 h light/dark cycle. The successful establishment of the esophageal cancer xenograft model was indicated by the appearance of tumors on the back of the mice after 1 week. When the tumor volume of mice reached 100–200 mm
3, they were randomly grouped. The PDX mice, including case EG20 (
n = 8), case LEG34 (
n = 11), and case LEG110 (
n = 10), were classified into three groups: (1) control, (2) 30 mg/kg, and (3) 120 mg/kg. Each group was given a daily gavage of saline, 30 mg/kg dronedarone hydrochloride solution, or 120 mg/kg dronedarone hydrochloride solution.
The mice’s body weight was measured every 2 days, and the tumor volumes were assessed every 3 days. The control group’s average tumor volume was monitored until it reached 1000 mm3; thereafter, tumor tissues from all the groups were collected after the mice were euthanized. The PDX model used tumors from patients without radiotherapy, chemotherapy, and no other diseases. All tumors were collected from surgical resection, and patients were informed of this situation and signed consent. All animal experiments were conducted with the approval of the Ethics Committee of Zhengzhou University (ZZUIRB2022-72).
2.17 Immunohistochemistry
The tumor tissues were collected and then dipped in 4% formalin, where they were left to fix for 72 h. Subsequently, the tumor tissues were dehydrated and embedded. The tumor tissues were cut to 4 μm and baked at 65 °C for 2 h. The tumor tissues were dewaxed for antigen repair and cultured with the primary antibody for 12–16 h at 4 °C. The tissue was stained with hematoxylin for 1–2 min, and DAB staining solution was applied while using the second antibody. The tissues were dehydrated, blighted, and covered with coverslips after adding neutral gum.
2.18 Correlation analysis
The Western blot strips were analyzed for their grayscale using ImageJ software. To obtain the data, the destination bands were compared with the internal reference. The correlation coefficient and P values for the study group were calculated using Spearman’s test in GraphPad Prism 7.
2.19 Statistical analysis
The results of the study were presented as mean and standard deviation values. Statistical analysis was performed using the Statistical Package for the Social Sciences (IBM, Inc., Armonk, NY, USA), and significant differences were determined through one-way or two-way ANOVOA or non-parametric tests. A P value of less than 0.05 was considered to be statistically significant.
3 Results
3.1 Dronedarone inhibited the proliferation of ESCC in vitro
To find drugs for ESCC chemoprevention, we screened a library of FDA-approved drugs for cytotoxicity against KYSE450. We discovered that dronedarone exhibited a toxic effect on KYSE450 (Fig.1 and S1A, Table S3). We observed that the suppressive effect of dronedarone on KYSE150 and KYSE450 was much higher than that on SHEE (Fig.1 and 1C, Tab.1). The results from cell proliferation assays showed that at 96 h, KYSE150 and KYSE450 cells decreased by 47.8% and 47.6%, respectively, with 2.5 μmol/L dronedarone, while SHEE cells were not inhibited (Fig.1–1F). The EdU incorporation assay also confirmed that the blockage of DNA synthesis led to the inhibition of KYSE150 and KYSE450 cell proliferation (Fig. S1B). The anchorage-independent cell growth assay showed that the growth of KYSE150 and KYSE450 cells decreased by 80.3% and 74.9%, respectively, with 2.5 μmol/L dronedarone (Fig.1). This result was also proved by the colony formation assays (Fig.1). Therefore, the findings demonstrated that dronedarone decreased the growth of ESCC cells but had a weak effect on SHEE cells.
3.2 Phosphoproteomics revealed that dronedarone downregulated the levels of p-RB1 S807, p-RB1 S811, and p-RB1 T826
To explore the molecular mechanism by which dronedarone inhibits the progression of ESCC, we treated KYSE150 cells with DMSO or 2.5 μmol/L dronedarone for 24 h. We examined the differential expression of phosphorylation sites by phosphoproteomics. The results revealed that dronedarone treatment had 350 upregulated phosphorylation sites on 245 proteins and 910 downregulated phosphorylation sites on 532 proteins in KYSE150 cells (Fig.2 and 2B, Table S4). We then analyzed the top 20 downregulated protein sites and found that three of them were associated with RB1, p-RB1 S807, p-RB1 S811, and p-RB1 T826 (Fig.2 and 2D, Table S4). Therefore, the downregulation of RB1 phosphorylation sites may be a critical step for dronedarone to inhibit ESCC cell proliferation. Subsequently, after performing kinase prediction using GPS5.0 software on the downregulated protein sites, we found a strong link between CDK4 and CDK6 with RB1 (Fig.2). TCGA database analysis indicated a strong positive correlation between expression levels of CDK4/CDK6 and RB1 (Fig.2 and 2G). Moreover, we selected 15 pairs of human normal and tumor tissues from patients with ESCC and verified the correlation between CDK4/CDK6 and RB1 or p-Rb S807/811 through Western blot assay. The results showed a positive correlation between CDK4/6 and RB1 or p-RB1 S807/S811 at the protein level (Fig. S2A and S2B). These findings suggested that dronedarone might exert an inhibitory effect on ESCC cells through the CDK4/6-RB1 axis. Finally, the results of Western blot demonstrated that the levels of p-RB1 S807, p-RB1 S811, and p-RB1 T826 were reduced in KYSE150 cells treated with dronedarone (Fig.2). Therefore, we demonstrated that dronedarone decreased the phosphorylation of p-RB1 S807, p-RB1 S811, and p-RB1 T826 sites by inhibiting CDK4/CDK6 kinase activity, resulting in the inhibition of the proliferation of ESCC cells.
3.3 Dronedarone bound to CDK4 and CDK6 kinase
To investigate whether CDK4/CDK6 are targets of dronedarone, we then performed computational docking models between dronedarone and CDK4 kinase and CDK6 kinase by Schrödinger software, respectively. Dronedarone directly bound to ASP97, ASP158, and THR177 sites of CDK4 kinase and GLU21, ASN150, and ASP163 sites of CDK6, respectively (Fig.3 and 3B). Furthermore, dronedarone was found to bind CDK4/CDK6 directly by ex vivo and in vitro pull-down assays (Fig.3 and 3D). In addition, after mutating ASP97, ASP158, and THR177 of CDK4, dronedarone’s capacity to bind to the CDK4 kinase protein decreased significantly. Similar results were observed in the case of binding with CDK6 (GLU21, ASN150, and ASP163; Fig.3 and 3F). These results suggested that dronedarone bound directly to CDK4 kinase at ASP97, ASP158, and THR177 sites and CDK6 kinase at GLU21, ASN150, and ASP163 sites.
3.4 Dronedarone blocked ESCC at G0/G1 phase by suppressing the CDK4/CDK6-RB1 axis
Although dronedarone can directly target CDK4 and CDK6, the molecular mechanism of its inhibition of cancer cell proliferation remains unclear. After dronedarone treatment, we examined the expression levels of CDK4, CDK6, p-RB1 S807, p-RB1 S811, p-RB1 T826, and RB1 in KYSE150 and KYSE450 cells. Interestingly, we found that the expression of CDK4, CDK6, and RB1 remained unchanged, whereas p-RB1 S807, p-RB1 S811, and p-RB1 T826 showed a dose-dependent decrease. These results revealed that dronedarone binding to CDK4/CDK6 decreased its kinase activity, reducing RB1 phosphorylation levels at S807, S811, and T826 (Fig.4 and 4B). To confirm our hypothesis, we performed
in vitro kinase assays. The findings demonstrated that dronedarone dose-dependently suppressed CDK4/CDK6 kinase, resulting in a reduction in the levels of p-RB1 S807 and p-RB1 S811 (Fig.4 and 4D). However, CDK4/CDK6 had no significant effect on the level of p-RB1 T826, which was the opposite of the results mentioned above (Fig. S3A and S3B). During the G1 to S phase, DNA synthesis requires phosphorylation of RB1 [
18]. Cell cycle experiments demonstrated a 35.6% and 42.9% increase in the G0/G1 phase with 2.5 μmol/L dronedarone in KYSE150 and KYSE450 cells, respectively. Furthermore, palbociclib, a CDK4/6 inhibitor, was used as a positive control to restrict the G0/G1 phase of the ESCC cells (Fig.4) [
19]. The activity of the E2F transcription factor family was controlled by the phosphorylated RB1 protein. The mechanism inhibiting occupancy in G1 by cyclin A2 and cyclin B1 promoters is regulated by E2F [
20]. We observed a dose-dependent reduction of cyclin A2 and B1 in KYSE150/KYSE450 cells (Fig.4 and 4B). These results showed that inhibition of p-RB1 could decrease levels of E2F-regulated proteins, including cyclin A2 and cyclin B1, after dronedarone treatment. Moreover, we found that dronedarone induced cell death through early and late stages of apoptosis (Fig. S3C and S3D). We also examined the protein levels of the apoptosis indicators (caspase-3 and BAX) and the dormancy indicator nuclear receptor subfamily 2 group F member 1 (NR2F1) [
21,
22]. The results showed that caspase-3 and BAX increased in KYSE150 and KYSE450 cells in a dose-dependent manner, but the dormancy index NR2F1 remained unchanged in KYSE150 and KYSE450 cells (Fig. S3E). A previous study reported that dronedarone inhibits the progression of ovarian cancer by downregulating c-MYC [
23]. To verify whether c-MYC is a key target for dronedarone in ESCC, we examined the protein level of c-MYC in KYSE150 and KYSE450 cells treated with different concentrations of dronedarone (Fig. S3E). The results showed that c-MYC remained unchanged in KYSE150 and KYSE450 cells after dronedarone treatment. Moreover, we performed a pull-down assay and found that dronedarone could not directly bind to c-MYC protein in KYSE150 and KYSE450 cells (Fig. S3F). These data suggested that c-MYC was not a direct target of dronedarone in ESCC. These outcomes led us to the conclusion that dronedarone inhibited the proliferation of ESCC by directly targeting CDK4/CDK6 and downregulating kinase activity, causing p-RB1 S807 and p-RB1 S811 to become less phosphorylated, which caused cell cycle arrest in the G0/G1 phase. Our findings indicated that dronedarone prevented the growth of ESCC by specifically targeting CDK4/CDK6 and reducing kinase activity. These changes led to decreased phosphorylation of p-RB1 S807 and p-RB1 S811, ultimately resulting in cell cycle arrest in the G0/G1 phase.
3.5 Depletion of CDK4/CDK6 decreased cell sensitivity to dronedarone
We examined the significance of CDK4/CDK6 as a potential target for ESCC by analyzing the expression levels of CDK4 and CDK6 in esophageal cancer through TCGA. The analysis revealed that CDK4 and CDK6 were strongly expressed in esophageal cancer, with ESCC showing higher levels of expression than esophageal adenocarcinoma (EAC; Fig.5 and S4A). Subsequently, we knocked out CDK4 or CDK6 in KYSE150 and KYSE450 cells (Fig.5). After CDK4 or CDK6 was knocked out, the proliferation of KYSE150 and KYSE450 cells was reduced (Fig.5 and S4B). The results of plate cloning assay revealed that the clone-forming ability of KYSE150 and KYSE450 cells significantly decreased (Fig.5). Subsequently, we evaluated the dronedarone susceptibility of KYSE150 and KYSE450 when CDK4 and CDK6 were knocked out. Deleting CDK4 and CDK6 significantly reduced KYSE150 and KYSE450s reactivity to dronedarone treatment (Fig.5). In addition, we found that KYSE150 and KYSE450 cells could prevent the cells from entering the G0/G1 phase after CDK4 and CDK6 were knocked out (Fig.5 and S4C). To further explore whether dronedarone inhibits ESCC at the G0/G1 phase through the CDK4/6-RB1 axis. We treated sgCDK4 or sgCDK6 KYSE150 and KYSE450 cells with dronedarone or overexpression of RB1 (Fig. S5A). The results of the cell cycle assay showed that knock-out CDK4 or CDK6 could block ESCC at the G0/G1 phase. The sgCDK4/sgCDK6 cells treated with dronedarone had no significant change compared with the sgCDK4/sgCDK6 cells (Fig. S5B and S5C). Knock-out CDK4 or CDK6 cells showed that cell cycle arrest was rescued after overexpression of RB1 (Fig. S6A and S6B). These results proved that dronedarone blocked ESCC at the G0/G1 phase by suppressing the CDK4/CDK6-RB1 axis. Thus, dronedarone inhibited ESCC proliferation mainly through CDK4 and CDK6.
3.6 Dronedarone inhibited ESCC growth in vivo via the CDK4/CDK6-RB1 axis
To investigate the inhibitory effects of dronedarone on ESCC in vivo, we developed ESCC-derived xenograft models in SCID/CB17 immunodeficient mice. The results showed that the sizes (Fig.6), volumes (Fig.6), and weights (Fig.6) of tumors were significantly reduced after dronedarone treatment. Moreover, the inhibition rate of the 30 mg/kg group was 25.9%–38.2%, and the 120 mg/kg group inhibition rate was 22.6%–28.5% compared with the control group (P < 0.01; Fig. S7A). Furthermore, the 30 mg/kg group and 120 mg/kg group had no significant reduction in body weight (Fig. S7B). HE staining of the heart, liver, spleen, lungs, and kidneys of mice in the EG20, LEG34, and LEG110 groups indicated that dronedarone treatment reduced the tumor load of mice without apparent side effects (Fig. S7C). In addition, we examined a biomarker of tumor proliferation, Ki67, and found a significant decrease in Ki67 levels in the tumor treated with dronedarone (Fig.6). The levels of p-RB1 S807 and p-RB1 S811 were significantly reduced in the tumors of the dronedarone treatment group (Fig.6). These results indicated that dronedarone inhibited the growth and proliferation of ESCC tumors in vivo through the CDK4/CDK6-RB1 axis.
4 Discussion
Cancer chemoprevention is defined as using natural, synthetic, or biological substances to slow down or reverse the development of cancer cells [
24]. Cancer chemoprevention using FDA-approved drugs is a cost-effective strategy to improve patient survival [
25,
26]. Accumulating evidence suggests that cancer chemoprevention can effectively reduce the occurrence of multiple cancers. For example, tamoxifen reduces breast cancer effectively [
27], and aspirin has benefits for colorectal cancer. However, chemopreventive drugs for ESCC have not been used in clinics [
6–
8,
26]. Here, by screening the FDA-approved drug library, we discovered that dronedarone inhibited the growth of ESCC. Dronedarone is a non-iodinated benzofuran derivative of amiodarone and has decreased the growth of ovarian and breast cancer [
23,
28]. In this study, dronedarone effectively inhibited the growth of ESCC
in vitro and
in vivo (Fig.1 and Fig.6).
One of the characteristics of carcinogenesis is a failure of cell cycle regulation [
18,
29,
30]. The process of the cell cycle is facilitated by the complex of the cell cycle protein-dependent kinase and cell cycle protein. This complex initiates phosphorylation of RB1, which in turn relaxes E2F target gene repression [
31,
32]. Inhibitors of CDK4/6 bind to the ATP pocket of the kinase, halting the cell cycle [
33–
35]. Thus, CDK4/6 inhibitor treatments may be applied to treat multiple types of breast cancer [
34,
35]. Evidence reveals that limited targeted therapy for esophageal cancer may benefit from targeting CDK4 and CDK6 [
36,
37]. The TCGA database also showed that CDK4 and CDK6 were highly expressed in esophageal cancer (Fig.5). Our study discovered that dronedarone directly targeted CDK4 and CDK6 to inhibit the progress of ESCC by blocking p-RB1 S807 and p-RB1 S811
in vitro and
in vivo (Fig.4–Fig.6), which has not been identified in previous studies.
CDK4/6 is the effective target for clinically treating tumors, although CDK4/6 inhibitors have side effects in clinical practice. Here, our study demonstrated that dronedarone had a significant inhibitory effect on ESCC by binding to CDK4/6. Interestingly, previous reports have shown that dronedarone inhibits ovarian cancer by downregulating c-MYC [
23]. However, whether c-MYC is a direct target for dronedarone has not been demonstrated. Our study indicated that CDK4 and CDK6, not c-MYC, were direct targets for dronedarone. Moreover, in Fig.6–6C, following treatment with dronedarone at 30 and 120 mg/kg, the sizes, volumes, and weights of tumors were considerably reduced compared with the control group. However, we found no significant difference between the 120 and the 30 mg/kg groups. For the
in vivo experiments, we selected two doses of dronedarone in mice: 30 (equivalent to the recommended concentration in humans of 800 mg/d for antiarrhythmic effects) and 120 mg/kg (the maximum tolerated dose in mice). Previous studies suggested that the use of high doses of dronedarone (1200 and 1600 mg/d) do not show any significant difference in preventing atrial fibrillation recurrence compared with the recommended dose (800 mg/d). Additionally, many patients experience adverse events while taking higher doses than the recommended dose [
38]. Therefore, this evidence demonstrated that the absorption of dronedarone was already saturated at 30 mg/kg, and increasing the drug concentration did not raise the blood concentration.
Notably, studies have revealed that cancer treatment and radiation therapy can cause various cardiovascular diseases, which include cardiotoxicity, vascular toxicity, and arrhythmias [
39,
40]. Ribociclib, a CDK4/6 inhibitor, may induce QTc interval prolongation in around 10% of treated patients [
41,
42]. However, dronedarone, an antiarrhythmic drug, is an excellent option for patients who experience cardiotoxicity or arrhythmias following cancer treatment or radiation therapy. Moreover, the current CDK4/6 inhibitors have different rates of discontinuation due to adverse reactions. Palbociclib and ribociclib have a 7.5% discontinuation rate, whereas abemaciclib has a higher rate of 20%. The most common reasons for discontinuation are myelosuppression for palbociclib and ribociclib and diarrhea for abemaciclib. However, dronedarone did not have myelosuppression, and only 3.2% of people discontinued due to gastrointestinal adverse events while taking dronedarone [
43]. Therefore, dronedarone has fewer side effects than existing CDK4/6 inhibitors. Besides, dronedarone protects against cardiac toxicity or arrhythmias due to cancer treatment or radiotherapy. Furthermore, our study observed no weight changes in mice in the dronedarone group through only three PDX models. Therefore, the clinical effectiveness of dronedarone needs to be verified in the real world when it is used as a chemoprevention drug in the clinic for ESCC.
In conclusion, dronedarone could effectively inhibit the proliferation of ESCC in vivo and in vitro. Dronedarone decreased the expression of p-RB1S807 and p-RB1S811 by directly binding to CDK4 and CDK6 kinases, causing cell cycle arrest in the G1 phase. Therefore, dronedarone could serve as a practical and rapid clinical application for the chemoprevention of ESCC.
PDF
(27086KB)
