MLN4924 (cat. no. S7109) and sorafenib (cat. no. S7397) were purchased from Selleck Industries LLC (Shanghai, China). Fetal bovine serum (FBS) medium (cat. no. 11995500) was obtained from Gibco (Thermo Fisher Scientific, Inc., Waltham, MA, USA). Cell Counting Kit-8 (CCK-8) (cat. no. CK04) was purchased from Dojindo Laboratories (Kumamoto, Japan). Annexin V/PI kit (cat. no. 556547) was from BD Biosciences (Franklin Lakes, NJ, USA). Anti-mouse lgG-HRP (cat. no. sc-2005), anti-rabbit lgG-HRP (cat. no. sc-2004) and IκBα (cat. no. sc-371) antibodies were obtained from Santa Cruz Biotechnology Inc. (Dallas, TX, USA). Chemiluminescent HRP substrate (cat. no. WBKLS0500) was obtained from EMD Millipore (Billerica, MA, USA). Cleaved caspase-3 (cat. no. 9661S), cleaved PARP (cat. no. 6987S), p-ERK (cat. no. 4370S), ERK (cat. no. 4695S), p21 (cat. no. 2947S), p27 (cat. no. 3686S), E-cadherin (cat. no. 14472S), N-cadherin (cat. no. 14215S), vimentin (cat. no. 3932S), Noxa (cat. no. 14766S), Deptor (cat. no. 11816S), p-4EBP1 (cat. no. 2855S), p-S6K (cat. no. 9202S), NF-κB p65 (cat. no. 8242S), NEDD8 (cat. no. 2745S) and p-mTOR (cat. no. 5536S) antibodies were obtained from Cell Signaling Technology, Inc. (Danvers, MA, USA). MMP9 antibody (cat. no. 10375-2-AP) was obtained form Proteintech (Wuhan, China). NAE1 (NHA2310) was provided by Novogene Co., Ltd. (Beijing, China).

The human HCC LM3 and 97H cell lines were purchased from the Cell Bank of the Type Culture Collection of Fudan University (Shanghai, China) and they are neither misidentified nor contaminated. Cells were cultured in Dulbecco's modified Eagle's medium (DMEM). All experiments were carried out in DMEM containing 10% FBS, 100 U/ml penicillin and 100 mg/ml streptomycin (Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C in 5% CO₂.

For the cell viability assay, the cells were seeded into 96-well microtiter plates at a density of 3,000 cells/well and allowed to adhere for 12 h. Cells were then treated with different chemicals for the indicated durations and then exposed to CCK-8 (10 µl/well) for 2 h at 37°C. Absorbance was measured at 450 nm on a Tecan Sunrise microplate reader (Tecan Group AG, Männedorf, Switzerland).

For the clonogenic assay, 800 cells were plated in triplicate in 6-well plates. After overnight incubation at 37°C, different chemicals were added into the medium to reach various concentrations (0, 25, 50, 100, 300 and 1,000 nM). Medium was changed every 3–4 days while maintaining the previous concentrations. After 9–10 days of culture, cell colonies were fixed with ice-cold methanol, followed by 0.05% crystal violet staining for 15 min. Colonies containing >50 cells in each well were counted manually under an Olympus BX41 light microscope (Olympus Corp., Tokyo, Japan). A gridded plastic sheet was attached to the bottom of each well to keep track of colonies counted (26). All fields were counted.

Cells were harvested and lysed with lysis buffer [20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM DTT, 1 mM sodium orthovanadate, 1 µg/ml leupeptin and 1 mM phenylmethylsulfonyl fluoride] on ice for 1 h. Afterwards, cell lysates were centrifuged for 15 min at 11,000 × g at 4°C. Protein concentrations of the supernatants were determined using the bicinchoninic acid assay (BCA) assay. A total of 50 mg of protein loaded per lane and separated using 8–12% SDS-PAGE and then transferred to polyvinylidene difluoride (PVDF) membranes. Membranes were then blocked with 5% non-fat milk and incubated 4°C for 12 h with the specified primary and secondary antibodies. All antibodies were diluted at 1:1,000 for western blot analysis. Protein bands were visualized using an ECL detection kit (EMD Millipore).

Cells were treated with indicated chemicals and then washed twice with phosphate-buffered saline (PBS). After incubation with Annexin V-FITC in binding buffer [10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl₂, 0.1% bovine serum albumin (BSA) and pH 7.4] for 15 min, the cells were immediately exposed to 2 µg/ml PI for 5 min before analysis by flow cytometry. Annexin V-positive cells represented apoptotic cells and were quantified as previously described (27).

The siRNA oligos were purchased from Shanghai GenePharma Co., Ltd., (Shanghai, China) as follows: siNFκB p65, 5′-GAUUGAGGAGAAACGUAAAdTdT-3′; a non-target siRNA (siControl), 5′-UCUACGAGGCACGAGACUU-3′. Briefly, cells were transfected with various siRNAs in MEM medium with 90 nM of each siRNA duplex, using Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific, Inc.) following the manufacturer's protocol. After transfection for 48 h, cells were harvested for western blot analysis.

Twenty-four 5 week-old male BALB/c nude mice with body weight of ~18.5 g were purchased from Beijing Bioscience Co., Ltd. (Beijing, China) and maintained with 12-h light/12-h dark cycles at a temperature of 25°C with a humidity level of 40–60% with food and water provided ad libitum in the Laboratory Animal Center of Army Medical University (Chongqing, China). Animal studies were approved by the Institutional Animal Care and Use Committee of Southwest Hospital (Chongqing, China). 97H cells (1×10⁷ cells in 100 µl serum-free DMEM medium) were inoculated subcutaneously into the right flank of the nude mice (n=6/group). After the third day, the mice were randomized and treated with 60 mg/kg MLN4924 in 10% cyclodextrin (13), 30 mg/kg sorafenib in ethanol/castor oil (v/v 1:1) or the combination of the two chemicals via daily oral gavage.

Tumor growth (6 for each group) was measured in all three dimensions once a week for three weeks. Tumor volume was calculated using the formula V = 4/3(π)XYZ, where X, Y and Z represent the radius of the tumor in each dimension. After three weeks, the mice were placed in the euthanasia chamber before turning on gas from a CO₂ tank. The flow rate of CO₂ (~30% of the euthanasia chamber volume per min) was added to the existing air in the chamber. Mice were usually expected to reach unconsciousness within 2–3 min, followed by waiting at least 2 min without seeing a breath and a heartbeat. To ensure death, cervical dislocation was used following CO₂ death. Tumors were harvested, photographed, weighed and the results were plotted.

The NEDD8 and NAE expression at mRNA levels for 337 patients with HCC were obtained from the TCGA database and analyzed (Tables I and II) [The results shown here are part based upon data generated by TCGA Research Network (http://cancergenome.nih.gor)]. In order to validate the TCGA data, the HCC specimens used in this study for IHC staining were obtained from 26 patients with HCC who underwent curative resection at the Department of Hepatobiliary Surgery, Southwest Hospital. This study was approved by the Ethics Committee of Southwest Hospital. Patient information as shown in Table SI. For each patient, the diagnosis of HCC was confirmed by pathologic examination. The HCC specimens were harvested and then fixed in 10% formalin for 48 h. Then the tumor tissues were embedded and sliced at 5-µm thickness. Immunohistochemistry was performed using the ABC Vectastain kit (Vector Laboratories Inc., Burlingame, CA, USA) with rabbit anti-NEDD8 and anti-NAE1 antibodies All antibodies were diluted at 1:200 for immunohistochemistry staining. Sections were developed with DAB and counterstained with hematoxylin. Finally, the sections were photographed under a light microscope (Olympus BX41; Olympus Corp.) with magnification at ×200. The staining was evaluated by different specialized pathologists and was performed without any knowledge of the patient characteristics (28,29). For the staining intensity if no significant difference was noted for paired HCC tissues and adjacent normal tissues, the staining intensity was determined by selecting same object area using Spot Denso function of an AlphaEaseFC software (Protein Simple, San Jose, CA, USA), and the integrated density value (IDV) was compared between cancer tissues and adjacent normal tissues.

The Pearson's Chi-squared test (χ²) test was used to analyze the relationship between NEDD8/NAE1 expression and the clinicopathological features of HCC cases using the SPSS 15.0 software (SPSS, Inc., Chicago, IL, USA). For cell proliferation, migration, apoptotic cells, patient survival, tumor volume as well as tumor weight, statistical analysis was performed using the Student's t-test for comparison of two groups or one-way analysis of variance (ANOVA) for comparison of more than two groups followed by Tukey's multiple comparison test. For multiple testing, a Bonferroni post hoc test of P-values was made using GraphPad Prism 6 (GraphPad, Inc., San Diego, CA, USA). Data are expressed as mean ± SEM of at least three independent experiments. A P-value <0.05 was considered to be statistically significant.

As activation of CRLs requires the covalent binding of NEDD8, we aimed to ascertain whether NEDD8 is related to HCC patients. We analyzed the relationship between NEDD8 expression and clinicopathologic characteristics of the HCC patients with complete information using TCGA data. The result indicated that NEDD8 expression is associated with histologic grade (Table I). Similar to this finding, NEDD8-activating enzyme (NAE) expression was found to be related to histologic grade, tumor size as well as clinical stage of HCC patients (Table II). Kaplan-Meier analysis showed that the overall survival rate was significantly low in HCC patients with high expression of NEDD8 (Fig. 1A, P=0.0472, n=337) or NAE1 (Fig. 2A, P=0.0146, n=337) compared to patients with low expression of these proteins, respectively. This suggests that high expression levels of NEDD8 and NAE1 in HCC are significantly associated with worse patient prognosis. In order to further confirm this finding, we used our department patient samples to detect NEDD8 and NAE expression levels by IHC assay. The results demonstrated that protein levels of NEDD8 and NAE1 were highly expressed in tumor tissues compared to matched adjacent non-tumor tissues (Figs. 1B and 2B).

As shown in Fig. 2C, MLN4924, a small molecular inhibitor of NAE, inhibited cullin-1 neddylation. We next determined whether inactivation of neddylation modification by MLN4924 would inhibit cell growth and clonogenic survival in HCC LM3 and 97H cell lines. The results showed that MLN4924 could significantly inhibit cancer cell growth and clonogenic survival in a dose-dependent manner (Fig. 2D and E).

In order to determine whether MLN4924 further inhibits cell proliferation by sorafenib, LM3 and 97H cells were treated with MLN4924 and sorafenib alone or in combination. CCK-8 and clonogenic assays were then performed to evaluate cell proliferation and survival. MLN4924 significantly inhibited HCC cell proliferation with IC₁₀ and IC₁₅ concentrations of 50 and 100 nM, respectively (Fig. 3A). Sorafenib weakly inhibited cell proliferation even at 2 µM, but when combined with low concentrations of MLN4924 the cytotoxicity of sorafenib was enhanced in a dose-dependent manner (P<0.05) (Fig. 3B). In addition, MLN4924 at 50 nM similar to sorafenib at 2 µM weakly inhibited the colony formation in both LM3 and 97H cells. However, following the combination treatment, enhanced inhibition of colony formation of sorafenib was noted (Fig. 3C and D). Taken together, these findings suggest that the combination of sorafenib and MLN4924 appears to cause an additive effect with a maximal antitumor activity in the treatment of HCC.

Annexin V/PI staining assay was used to determine apoptosis levels in the LM3 and 97H cell lines. Cells were grown as monolayers and exposed to sorafenib alone or in combination with MLN4924 for up to 48 h. Apoptosis levels were determined by assessing the cells that were positive for Annexin V staining. As shown in Fig. 4A and B, MLN4924 and sorafenib alone at low concentrations did not significantly induce apoptosis, whereas the combination of both drugs significantly induced apoptotic cells up to 25%. In order to further validate this finding, we analyzed apoptosis-related proteins cleaved caspase-3 and its downstream target PARP using western blot analysis. As shown in Fig. 4C, sorafenib alone induced cleaved caspase-3 and cleaved PARP accumulation in a dose-dependent manner, and this was enhanced by MLN4924 in a dose-dependent manner. Cell growth inhibition induced by these two drugs alone or in combination appears to be mediated by the induction of apoptosis.

We determined the effect of sorafenib alone or in combination with MLN4924 on HCC cell migration capacity. In LM3 cells, sorafenib at 2 µM induced an ~15% inhibition, whereas MLN4924 at 50 nM induced a 10% inhibition of cell migration. The combination of both drugs induced a 60% inhibition in cell migration, which was statistically significant (Fig. 5A). Likewise, in 97H cells, single drug treatment induced a 15–30% inhibition of cell migration for sorafenib at 5 µM and MLN4924 at 100 nM. Whereas the combination of the two drugs induced up to a 65% inhibition of cell migration, which was again statistically significant (Fig. 5A). Subsequently, migration markers were analyzed by western blot analysis. Sorafenib upregulated E-cadherin, and downregulated N-cadherin and vimentin in a dose-dependent manner. Compared to sorafenib alone treatment, the combination treatment further enhanced this action (Fig. 5B). Taken together, MLN4924 further suppresses cell migration induced by sorafenib by upregulating E-cadherin and downregulating N-cadherin and vimentin.

We assessed the efficacy of sorafenib, MLN4924, and the combination of both in an in vivo 97H xenograft mouse model. As shown in Fig. 6A and B, sorafenib and MLN4924 alone moderately inhibited tumor growth in nude mice, while the combination of the two drugs had a nearly complete inhibition of tumor growth. At the end of the experiment, the average tumor weight in the Control group, the sorafenib group, the MLN4924 group and the combination therapy group was 1.52, 0.80, 0.89 and 0.22 g, respectively, with a statistical difference between the sorafenib group and the combination therapy group (P<0.05) (Fig. 6B). The tumor growth index in the sorafenib group, the MLN4924 group and the combination therapy group was 47.0, 65.4 and 21.2%, respectively, with statistical difference between each group (P<0.05) (Fig. 6C). Finally, the drug dosages used were not toxic to the animals as indicated by the minimal loss of body weight (Fig. 6D). Taken together, the results demonstrate that MLN4924 indeed sensitizes HCC tumors to sorafenib in an in vivo xenograft mouse tumor model.

To determine the potential mechanism of MLN4924 as a sorafenib sensitizer, cell proliferation- and migration-associated substrates of CRL/SCF E3 ligase were analyzed by western blot analysis. We found that MLN4924 inhibited cullin neddylation, and induced the accumulation of Noxa (a pro-apoptotic protein, shown to be a CRL5 substrate). However, we did not observe a further increase in Noxa levels after combination treatment, suggesting it was not critical for sorafenib sensitization. However, the levels of proliferation- and migration-associated proteins Deptor (an mTOR inhibitor), IκBα (an inhibitor of NF-κB), p21 and p27 (two inhibitors of cyclin-dependent kinases) were higher in the combination treatment group compared to the MLN4924 or sorafenib treatment only group (Fig. 7A). These findings were further validated by decreases in p-mTOR, p-4EBP1 and p-S6K (two downstream targets of mTOR) and NF-κB p65 (a downstream target of IκBα). Given that p-ERK is the main target of sorafenib and that MLN4924 did not have any effect on p-ERK levels, this suggests that ERK was not involved in MLN4924-mediated sorafinib sensitization. A previous study reported that MMP9, a downstream target of NF-κB, was important for regulating cell migration (30). However, we found that MMP9 was not associated with sorafineb sensitization triggered by MLN4924 in HCC cells, as indicated by no significant changes to MMP9 levels in the presence or absence of MLN4924. Furthermore, we found that silencing of NF-κB p65 and mTOR inhibitor rapamycin could significantly inhibit cell proliferation and migration (Fig. 7B-D), which are consistent with the effect of the combination treatment group. Taken together, these results indicate that the accumulation of SCF E3 ligase substrates p21, p27, Deptor and IκBα induced by MLN4924 was associated with the chemo-sensitization effects of sorafenib on HCC cell lines (Fig. 7E).