Prediction of target proteins of FBW7 was completed by Ubibrowser (http://ubibrowser.ncpsb.org.cn/ubibrowser/). After logging in the website, the query protein was selected as E3, and the “FBW7” was input. Then the “Explore” button was clicked, “Continue” button was chosen, and results were obtained.

Analysis of chromodomain helicase DNA binding protein 3 (CHD3) expression in 374 liver hepatocellular carcinoma (LIHC) and 50 normal samples was conducted on starBase (http://starbase.sysu.edu.cn/), where we logged in, selected “Pan-Cancer”, clicked “Gene Differential Expression”, input “CHD3”, and chose “LIHC” to obtain results.

Culture (humidity, 95% air, 5% CO₂, 37 °C) of HCC cell lines HCCLM3 (SCSP-5093, Shanghai Institute of Biochemistry and Cell Biology, Shanghai, China) and HepG2 (SCSP-510, Shanghai Institute of Biochemistry and Cell Biology, Shanghai, China) as well as normal liver cell line HL-7702 (BFN608006124, Shanghai Cell Bank, Shanghai, China) was carried out with Dulbecco’s Modified Eagle’s Medium (DMEM; A4192101, Thermo Fisher Scientific, Waltham, MA, USA) incorporating 10% fetal bovine serum (FBS, 16140063, Gibco, Thermo Fisher Scientific, Waltham, MA, USA). Mycoplasma tests were performed periodically by polymerase chain reaction (PCR) to confirm that cells were not contaminated by mycoplasma, followed by authentication with short tandem repeat (STR) profiling.

Short hairpin RNA CHD3 (shCHD3) (5^′-CCCAGAAGAGGAUUUGUCA-3^′) was constructed on shRNA plasmid DNA (SHC004, Sigma-Aldrich, Darmstadt, Germany), while mammalian promoter plasmid vector (PP2356, Sigma-Aldrich, Darmstadt, Germany) was applied in the construction of CHD3 overexpression plasmid and FBW7 overexpression plasmid (their sequences are listed in Supplementary File 1), with the empty shRNA plasmid and the empty mammalian promoter plasmid as the negative controls, respectively. Besides, pCMV-N-Myc (D2756, Beyotime Biotechnology, Shanghai, China) and pUC-CMV-N-Flag (D2722, Beyotime Biotechnology, Shanghai, China) served as the vectors of Myc-FBW7 as well as Flag-CHD3. pCMV-HA-ubiquitin (Ub) plasmid was bought from Ke Lei Biological Technology (kl-zl-0513, Shanghai, China). Transfection of above plasmids into HCCLM3 and HepG2 cells was accomplished through Lipo 2000 (11668500, Thermo Fisher Scientific, Waltham, MA, USA). Firstly, cells were trypsinized (15090046, Thermo Fisher Scientific, Waltham, MA, USA), seeded in 24-well plates (1 $\times$ 10⁵/well) and cultured until 80% confluence, subsequent to which 2 µL Lipo 2000 (5 min, room temperature (RT)) and 0.8 µg plasmid were diluted in 50 µL Opti-modified eagles medium (MEM) (31985062, Thermo Fisher Scientific, Waltham, MA, USA), mixed and cultivated (20 min, RT). Finally, 24/48 h cell incubation was conducted in 100 µL/well complexes.

Following total RNA obtainment from cells with TRIzol kit (12183555, Thermo Fisher Scientific, Waltham, MA, USA) and cDNA synthesis using SuperScript™ First-Strand Synthesis System for RT-PCR (11904018, Thermo Fisher Scientific, Waltham, MA, USA), cDNA amplification was completed with TB Green® Fast qPCR Mix (RR430S, Takara Biomedical Technology, Beijing, China) in a StepOnePlus PCR system (Thermo Fisher Scientific, Waltham, MA, USA) under the conditions of 1 cycle of predenaturation (30 s, 95 °C) as well as 40 cycles (5 s, 95 °C; 15 s, 60 °C). Synthesis of primer sequences by RiboBio (Guangzhou, China) is displayed in Table 1. $\beta$-actin were endogenous controls. Data analysis was accomplished using 2^{-Δ⁢Δ⁢CT} method [19].

Based on CCK-8 kit (C0040, Beyotime Biotechnology, Shanghai, China), we examined cell viability and toxicity. For cell viability, transfected cells experienced digestion, seeding in 96-well plates (2000/well), and 24/48/72/96 h cultivation, followed by culture (1 h, 37 °C) in CCK-8 solution (10 µL/well).

For cell toxicity, 5000 cells in the exponential growth phase underwent 24-h incubation in a 96-well plate (100 µL/well). As mentioned before, oxaliplatin (OXA; MZ3501-500MG, Mocan Biotechnology Co., Ltd., Shanghai, China) was prepared and diluted in sterile saline (0.9%) [20]. Subsequently, different concentrations of OXA (0.25, 0.5, 1, 2, and 4 µM) were sequentially cultivated with cells (37 °C, humidity, 5% CO₂, 24 h). Then, cells were further immersed in 10 µL/well CCK-8 solution (1–2 h, 37 °C).

Measurement of optical density (OD) value (450 nm, an enzyme-linked immunosorbent assay reader) was followed by cell viability data obtained using the equation: Cell viability (%) = (average OD value of experimental group – average OD value of blank control group)/(average OD value of negative control group – average OD value of blank control group) $\times$ 100%. Finally, with x-axis and y-axis separately representing the concentration and the corresponding cell viability using GraphPad Prism software (version 7.0, GraphPad Software, San Diego, CA, USA), a curve was drawn. The IC₅₀ value of OXA refers to the concentration at which cell viability (%) is reduced by OXA to 50%.

Cells (5 $\times$ 10⁵) in a 6-well plate were cultured until 80% fusion and scratched exploiting a pipette tip. Post washing with phosphate-buffered saline (PBS; 20012050, Thermo Fisher Scientific, Waltham, MA, USA) to discard cell debris, cells were cultivated in serum-deprived medium (37 °C, 48 h). The cells were photographed through a microscope (0 h, 48 h; magnification: $\times$100; IXplore Standard, Olympus, Tokyo, Japan). The migrated distance was obtained using Image J, version 1.48 (National Institutes of Health, Bethesda, MD, USA).

Suspension of the digested and PBS-washed cells was performed in the serum-free medium. Later, 2 $\times$ 10⁵/mL cells were transferred to the Matrigel (354234, Corning, Corning, NY, USA)-coated upper filter of a 24-well transwell chamber (140629, Thermo Fisher Scientific, Waltham, MA, USA), with the lower chamber containing 10% FBS-supplemented medium. 48 h (37 °C) later, fixation (4% paraformaldehyde, 4 °C, 0.5 h; BL-G002, SenBeiJia, Nanjing, China) and staining (0.1% crystal violet, RT, 0.5 h; BP-DL134, SenBeiJia, Nanjing, China) of cells invading the membrane were conducted. The results were viewed under the microscope (magnification: $\times$200, IXplore Standard, Olympus, Tokyo, Japan).

Seeded at 2 $\times$ 10³ cells/well in ultra-low attachment 24-well culture plates (3473, Corning, Corning, NY, USA), cells were rest in serum-deprived RPMI-1640 medium (11875085, Gibco, Thermo Fisher Scientific, Waltham, MA, USA) supplemented with B27 (20 µL/mL; 17504044, Thermo Fisher Scientific, Waltham, MA, USA), epidermal growth factor (EGF, 20 ng/mL; PHG0313, Thermo Fisher Scientific, Waltham, MA, USA), basic fibroblast growth factor (bFGF, 20 ng/mL; 13256-029, Thermo Fisher Scientific, Waltham, MA, USA) and 1% penicillin-streptomycin (P/S; V900929, Sigma-Aldrich, Darmstadt, Germany). After 10-day cultivation in an incubator (37 °C, 5% CO₂), the microscope (IXplore Standard, Olympus, Tokyo, Japan) was used for cell observation with images taken (magnification: $\times$200).

Following cell centrifugation (4 °C, 12,830 $\times$g, 5 min) in radioimmunoprecipitation assay (RIPA) lysis buffer (C500007, Sangon Biotech, Shanghai, China), the supernatant was collected for protein concentration detection by bicinchoninic acid (BCA) assay (C503021, Sangon Biotech, Shanghai, China), and separation of proteins (45 µg) was performed with Spectra™ Multicolor Broad Range Protein Ladder (5 µL; 26634, Thermo Fisher Scientific, Waltham, MA, USA) and SDS-PAGE Preparation kit (C631100, Sangon Biotech, Shanghai, China). After transference of protein and marker onto polyvinylidene difluoride (PVDF) membranes (F019533, Sangon Biotech, Shanghai, China) and blockage (RT, 1 h) in 5% bovine serum albumin (E661003, Sangon Biotech, Shanghai, China), membranes underwent cultivation (overnight, 4 °C) with primary antibodies FBW7 (rabbit, 1:2000; ab109617, Abcam, Cambridge, UK), CHD3 (rabbit, 1:1000; #4241, Cell Signaling, Danvers, MA, USA) and $\beta$-actin (mouse, 1:1000; ab8226, Abcam, Cambridge, UK), washing with Tris-buffered saline containing Tween 20 (C520009, Sangon Biotech, Shanghai, China) and culture (1 h, RT) with secondary antibodies (Abcam, Cambridge, UK) against rabbit (ab205718, 1:2000) and mouse (ab205719, 1:2000). Reaction of membranes and enhanced chemiluminescence (ECL) luminescence reagent (C510043, Sangon Biotech, Shanghai, China) was followed by visualization employing GelView 6000Plus Smart Gel Imaging System (Biolight Biotechnology, Guangzhou, China). Image J (1.48v, NIH, Bethesda, MD, USA) was employed to analyze the data acquired.

HEK-293 cells (Procell Life Science&Technology, Wuhan, China) were cultivated in Minimum Essential Medium (PM150467, Procell Life Science&Technology, Wuhan, China) incorporating 1% P/S and 10% FBS.

To explore the interaction between exogenous FBW7 and exogenous CHD3, HEK-293 cells transfected with Myc-FBW7 and Flag-CHD3 expression vectors underwent lysis in RIPA lysis buffer. The lysates experienced culture (4 °C, overnight) with BeyoMag™ Anti-Myc or Anti-Flag Magnetic Beads (P2118, P2115; Beyotime Biotechnology, Shanghai, China). With the application of an immunoprecipitation kit (ab206996, Abcam, Cambridge, UK), HEK-293 cells were lysed, and precleared by Protein A/G Sepharose® beads. The lysates were cultured with antibodies against FBW7 (1:400) or CHD3 (1:50) at 4 °C with gentle rocking overnight, for which rabbit Immunoglobulin G (IgG) (A7016, Beyotime Biotechnology, Shanghai, China) was the negative control. Next, Protein A/G Sepharose® beads were mixed with the antigen-antibody complex (4 °C, 1 h) to acquire antigen-antibody-bead complex. As for ubiquitination detection, the lysates of HEK-293 cells with transfection of Myc-FBW7 expression vector, Flag-CHD3 expression vectors as well as HA-Ub and treatment of MG-132 (S1748, Beyotime Biotechnology, Shanghai, China) were cultivated (4 °C, overnight) with the antibody against ubiquitinylated proteins (1:2000; Millipore, Merck KGaA, Darmstadt, Germany) and BeyoMag™ Anti-Flag Magnetic Beads (P2115, Beyotime Biotechnology, Shanghai, China).

Subsequently, all bound proteins above were eluted with the beads, and rinsed off in the sodium dodecyl sulfate (SDS) sample buffer and finally heated at 90 °C for 5 min. The protein level was obtained by western blot. Image J (1.48v, NIH, Bethesda, MD, USA) was employed to analyze the band intensities.

Statistical data (form: means $\pm$ standard deviation) were dissected using GraphPad 8.0 (GraphPad Software, San Diego, CA, USA). The paired-samples t test was adopted for contrast of cancer tissues and normal tissues adjacent to cancer, whereas multiple groups were compared through one-way ANOVA followed by Tukey’s or Dunnett’s post hoc test, with two-way ANOVA used to analyze double factors among multiple groups. p 0.05 was deemed statistically noteworthy.

The ubibrowser predicted CHD3 as a target protein of FBW7 (Fig. 1A). The analysis of starBase exhibited a higher expression level of CHD3 in LIHC samples relative to normal samples (Fig. 1B, p = 3.1 $\times$ 10^-9). Similarly, qRT-PCR and western blot data revealed that CHD3 level in HCC cell lines (HCCLM3 and HepG2) was evidently raised when contrasted with normal cell line HL-7702 (Fig. 1C,D, p 0.001). To explore the functions of CHD3 on HCC, shCHD3 was transfected into HCCLM3 and HepG2 cells, the success of which was confirmed by downregulation of CHD3 in these cells (Fig. 2A, p 0.001). Consequently, both of HCCLM3 and HepG2 cells in shCHD3 group presented reduced OD value, migration and invasion rates, and diameter of spheres (Fig. 2B–I, p 0.05). To investigate how CHD3 regulates OXA resistance in HCC cells, shCHD3 was transfected into HCCLM3 and HepG2 cells. Following treatment with gradient concentrations of OXA, cell viability assessment was performed exploiting CCK-8 assay. The results showed that in HCCLM3 cells treated with 0.5 µM and 1 µM OXA, the viability was weakened in shCHD3 group relative to shNC group (p 0.05, Fig. 2J). Similarly, in HepG2 cells exposed to 2 µM OXA, the viability was attenuated in shCHD3 group relative to shNC group (p 0.05, Fig. 2K). HCCLM3 and HepG2 cells had declined IC₅₀ values after CHD3 knockdown relative to shNC-treated cells (p 0.01, Fig. 2L).

Fig. 1.

CHD3 was upregulated in HCC. (A) The predicted target proteins of FBW7 (Ubibrowser). (B) The differentially expressed CHD3 in LIHC (n = 374) and normal samples (n = 50) (starBase). (C,D) CHD3 mRNA (C) and protein (D) expressions between HCC cell lines (HCCLM3 and HepG2) and normal cell line (HL-7702) (quantitative reverse transcription-polymerase chain reaction; western blot, $\beta$-actin as the loading control). ^{^⁢^⁢^}p 0.001 vs. HL-7702 cell line. All experiments were repeated independently at least three times, n = 3. Data were presented as the means $\pm$ standard deviation. Abbreviation: CHD3, chromodomain helicase DNA binding protein 3; HCC, hepatocellular carcinoma; FBW7, F-box and WD repeat domain containing 7; LIHC, liver hepatocellular carcinoma.

Fig. 2.

CHD3 knockdown attenuated HCC cell proliferation, migration, invasion and stemness. (A) CHD3 mRNA expression of HCCLM3 and HepG2 cells after transfection of shCHD3 (qRT-PCR, $\beta$-actin as the loading control). (B,C) OD value of HCCLM3 (B) and HepG2 cells (C) after transfection of shCHD3 (CCK-8 assay). (D) Representative images of wound healing assay in HCCLM3 and HepG2 cells transfected with shCHD3 (magnification: $\times$100, scale: 100 µm). (E) Migration rate of HCCLM3 and HepG2 cells after transfection of shCHD3 (wound healing assay). (F) Representative images of transwell assay in HCCLM3 and HepG2 cells transfected with shCHD3 (magnification: $\times$250, scale: 50 µm). (G) Invasion rate of HCCLM3 and HepG2 cells after transfection of shCHD3 (transwell assay). (H) Representative images of sphere formation assay in HCCLM3 and HepG2 cells transfected with shCHD3 (magnification: $\times$200, scale: 200 µm, arrows: sphere formation). (I) Diameter of spheres of HCCLM3 and HepG2 cells after transfection of shCHD3 (sphere formation). (J–L) Viability of HCCLM3 and HepG2 cells transfected with shCHD3 and treated with gradient concentrations of OXA (CCK-8 assay), with the IC₅₀ values calculated for OXA. * p 0.05, ** p 0.01, *** p 0.001 vs. shNC group. All experiments were repeated independently at least thrice, n = 3. Data were presented as the means $\pm$ standard deviation. Abbreviation: CHD3, chromodomain helicase DNA binding protein 3; HCC, hepatocellular carcinoma; qRT-PCR, quantitative reverse transcription-polymerase chain reaction; shCHD3, short hairpin RNA CHD3; OXA, oxaliplatin; OD, optical density; shNC, short hairpin RNA negative control.

As per western blot data, the co-transfection of FBW7 and CHD3 overexpression plasmids decreased CHD3 protein level in HCCLM3 and HepG2 cells, while transfection of CHD3 overexpression plasmid alone barely impacted FBW7 protein expression (Fig. 3A), implicating that FBW7 overexpression declined the stability of CHD3. Then, we dug into the interaction of FBW7 with CHD3. Myc-FBW7 and Flag-CHD3 were co-transfected into HEK-293 cells. It was discovered that IP of Myc-FBW7 by anti-Myc antibody brought about Co-IP of Flag-CHD3, and Flag-CHD3 was co-immunoprecipitated with Myc-FBW7 by anti-Flag antibody (Fig. 3B,C). Additionally, IP analysis showed that endogenous FBW7 and CHD3 were immunoprecipitated from HCCLM3 and HepG2 cells, with endogenous FBW7 and CHD3 detected, respectively (Fig. 3D). These results mirrored FBW7 can interact with CHD3. Hence, to validate whether FBW7 modulates CHD3 stability via this activity, Flag-CHD3 and HA-Ub were co-expressed in HEK-293 cells with/without Myc-FBW7. Overexpression of FBW7 reduced CHD3 expression and may promote ubiquitination degradation of CHD3. Further, the effect of FBW7 overexpression were antagonized by MG-132 (Fig. 3E). This suggests that FBW7 promotes ubiquitination degradation of CHD3 through the MG-132 (Fig. 3E). These findings implied that FBW7 mediated ubiquitination and proteasomal degradation to regulate CHD3 protein expression.

Fig. 3.

FBW7 targeted CHD3 for ubiquitination. (A) Western blot of exogenous FBW7 and CHD3 in HCCLM3 and HepG2 cells co-expressing FBW7 and CHD3 ($\beta$-actin as the loading control). (B,C) FBW7 interacted with CHD3 at the exogenous level. Immunoblotting analysis of lysates after immunoprecipitation from HEK-293 cells transfected with Myc-FBW7 and Flag-CHD3. (D) FBW7 interacted with CHD3 at the endogenous level. Lysates from HCCLM3 and HepG2 cells were immunoprecipitated with anti-FBW7, followed by immunoblotting for FBW7 and CHD3 (IgG as a control). (E) Flag-CHD3 and HA-Ub were co-expressed with Myc-FBW7 in HEK-293 cells. In cells treated with/without 10 mM MG-132 (6 h), CHD3 was immunoprecipitated with anti-Flag antibody, and the poly-Ub status of CHD3 was detected by immunoblotting ($\beta$-actin as loading control). All experiments were repeated independently at least thrice, n = 3. Data were described as the means $\pm$ standard deviation. Abbreviation: FBW7, F-box and WD repeat domain containing 7; CHD3, chromodomain helicase DNA binding protein 3; HCC, hepatocellular carcinoma; IgG, Immunoglobulin G; HA-Ub, HA-ubiquitin; poly-Ub, polyubiquitination.

In HCCLM3 and HepG2 cells, FBW7 overexpression apparently raised FBW7 level and decreased CHD3 expression, whereas CHD3 overexpression notably increased CHD3 level with FBW7 expression barely changed (Fig. 4, p 0.05). Moreover, CHD3 overexpression counteracted the inhibitory effect of FBW7 overexpression on the expression of CHD3 (Fig. 4, p 0.05). We also proved the migration rate, invasion rate and diameter of spheres in HCCLM3 and HepG2 cells were dramatically elevated after CHD3 overexpression, but were prominently declined after FBW7 overexpression (Fig. 5, Fig. 6A–C, p 0.05). However, CHD3 overexpression reversed the effect of FBW7 overexpression on the migration rate, invasion rate and diameter of spheres in HCCLM3 and HepG2 cells (Fig. 5, Fig. 6A–C, p 0.05). In the cytotoxicity assay, we found the IC₅₀ values of HCCLM3 and HepG2 cells were increased after CHD3 overexpression, but were decreased after FBW7 overexpression, (Fig. 6D–G, p 0.05). CHD3 overexpression reversed the effect of FBW7 overexpression on the IC₅₀ values (Fig. 6D–G, p 0.05).

Fig. 4.

FBW7 overexpression targeted to repress CHD3 level in HCC cells. (A,B) FBW7 and CHD3 mRNA expression of HCCLM3 (A) and HepG2 (B) after transfection of FBW7 and CHD3 overexpression plasmids (qRT-PCR, $\beta$-actin as the loading control). (C,D) FBW7 and CHD3 protein expression of HCCLM3 (C) and HepG2 (D) after transfection of FBW7 and CHD3 overexpression plasmids (western blot). * p 0.05, ** p 0.01, *** p 0.001 vs. NC group; ^{^⁢^}p 0.01, ^{^⁢^⁢^}p 0.001 vs. CHD3 group; ^#p 0.05, ^#⁢#⁢#p 0.001 vs. FBW7 group. All experiments were repeated independently at least thrice, n = 3. Data were delineated as the means $\pm$ standard deviation. Abbreviation: FBW7, F-box and WD repeat domain containing 7; CHD3, chromodomain helicase DNA binding protein 3; HCC, hepatocellular carcinoma; qRT-PCR, quantitative reverse transcription-polymerase chain reaction; NC, negative control.

Fig. 5.

FBW7 overexpression mitigated the influences of CHD3 on migration and invasion of HCC cells. (A) Representative images of wound healing assay in HCCLM3 and HepG2 cells transfected with FBW7 and CHD3 overexpression plasmids (magnification: $\times$100, scale: 100 µm). (B,C) Migration rate of HCCLM3 (B) and HepG2 cells (C) after transfection of FBW7 and CHD3 overexpression plasmids (wound healing assay). (D) Representative images of transwell assay in HCCLM3 and HepG2 cells transfected with FBW7 and CHD3 overexpression plasmids (magnification: $\times$250, scale: 50 µm). (E,F) Invasion rate of HCCLM3 (E) and HepG2 cells (F) after transfection of FBW7 and CHD3 overexpression plasmids (transwell assay). * p 0.05, ** p 0.01 *** p 0.001 vs. NC group; ^{^}p 0.05, ^{^⁢^⁢^}p 0.001 vs. CHD3 group; ^#p 0.05, ^#⁢#p 0.01 vs. FBW7 group. All experiments were repeated independently at least thrice, n = 3. Data were described as the means $\pm$ standard deviation. Abbreviation: FBW7, F-box and WD repeat domain containing 7; CHD3, chromodomain helicase DNA binding protein 3; HCC, hepatocellular carcinoma; NC, negative control.

Fig. 6.

FBW7 overexpression mitigated the influences of CHD3 on stemness of HCC cells. (A) Representative images of sphere formation assay in HCCLM3 and HepG2 cells transfected with FBW7 and CHD3 overexpression plasmids (magnification: $\times$200, scale: 200 µm, arrows: sphere formation). (B,C) Diameter of spheres of HCCLM3 (B) and HepG2 (C) cells after transfection of shCHD3 (sphere formation). (D–G) Viability of HCCLM3 and HepG2 cells transfected with FBW7 and CHD3 overexpression plasmids and treated with gradient concentrations of OXA (CCK-8 assay), and the IC₅₀ values for OXA. * p 0.05, ** p 0.01 *** p 0.001 vs. NC group; ^{^}p 0.05, ^{^⁢^⁢^}p 0.001 vs. CHD3 group; ^#p 0.05, ^#⁢#p 0.01, ^#⁢#⁢#p 0.001 vs. FBW7 group. All experiments were repeated independently at least thrice, n = 3. Data were described as the means $\pm$ standard deviation. Abbreviation: FBW7, F-box and WD repeat domain containing 7; CHD3, chromodomain helicase DNA binding protein 3; HCC, hepatocellular carcinoma; NC, negative control.