Sodium butyrate inhibits aerobic glycolysis of hepatocellular carcinoma cells via the c‐myc/hexokinase 2 pathway

Abstract Aerobic glycolysis is a well‐known hallmark of hepatocellular carcinoma (HCC). Hence, targeting the key enzymes of this pathway is considered a novel approach to HCC treatment. The effects of sodium butyrate (NaBu), a sodium salt of the short‐chain fatty acid butyrate, on aerobic glycolysis in HCC cells and the underlying mechanism are unknown. In the present study, data obtained from cell lines with mouse xenograft model revealed that NaBu inhibited aerobic glycolysis in the HCC cells in vivo and in vitro. NaBu induced apoptosis while inhibiting the proliferation of the HCC cells in vivo and in vitro. Furthermore, the compound inhibited the release of lactate and glucose consumption in the HCC cells in vitro and inhibited the production of lactate in vivo. The modulatory effects of NaBu on glycolysis, proliferation and apoptosis were related to its modulation of hexokinase 2 (HK2). NaBu downregulated HK2 expression via c‐myc signalling. The upregulation of glycolysis in the HCC cells induced by sorafenib was impeded by NaBu, thereby enhancing the anti‐HCC effect of sorafenib in vitro and in vivo. Thus, NaBu inhibits the expression of HK2 to downregulate aerobic glycolysis and the proliferation of HCC cells and induces their apoptosis via the c‐myc pathway.


| INTRODUC TI ON
Primary liver cancer is responsible for over 8 million deaths worldwide annually and is the sixth most frequently diagnosed cancer and the third leading cause of cancer-related death, which makes it a serious global health burden. 1 Hepatocellular carcinoma (HCC) accounts for 75%-85% of cases of primary liver cancer. Thus, finding a safe and effective strategy to treat patients with HCC is urgently needed. 2 In 2011, energy metabolism reprogramming was identified to be a novel hallmark of cancer. 3 In 1930, Otto Warburg found that in the presence of sufficient oxygen, cancer cells tended to convert glucose to lactate via glycolysis instead of catabolizing it via oxidative phosphorylation (OXPHOS), a phenomenon termed 'aerobic glycolysis' or 'Warburg effect'. 3 Enhanced aerobic glycolysis was observed in HCC, 4 lung cancer, 5 breast cancer 6 and other types of cancer.
High-speed energy production, synthesis of metabolic substrates and lactate production via aerobic glycolysis promote hepatocellular carcinogenesis by regulating the proliferation, growth, immune evasion, invasion, metastasis, angiogenesis and sorafenib resistance of HCC. 7 Moreover, a recent review indicated that key enzymes in the glycolytic pathway modulate several important phenotypes of cancer, including DNA repair, mitosis, regulation of exosomes and ectosomes, autophagy, and mitochondria-dependent apoptosis, via noncanonical functions. 8 Previous experimental studies found that glycolysis inhibitors, such as aspirin, 9 simvastatin 10 and shikonin, 11 could exert anti-HCC effects and enhance the sensitivity of the HCC cells to sorafenib, an oral multikinase inhibitor, in the treatment of advanced HCC. Hence, targeting the glycolytic pathway and its key enzymes could exert antitumour effects by regulating diverse key phenotypes of cancer and has become a novel strategy in HCC treatment. Moreover, active hepatic stellate cells (HSCs) were reported to use aerobic glycolysis as their energy source. 12 Inhibiting the key enzymes in this pathway, including HK2, 13 6-phosphofructo-2-kinase/ fructose-2,6-bisphosphatase-3 (PFKFB3) 14 and pyruvate kinase M2 (PKM2), 15 could suppress the activation of hepatic stellate cells and the subsequent liver fibrosis, which is the main contributor to liver cirrhosis and HCC.
Sodium butyrate (NaBu) is a sodium salt metabolite of shortchain fatty acids (SCFAs), which is produced by the gut microbiota NaBu is generally produced by Firmicutes and exerts antioxidant, immunomodulatory and anticancer effects. 16,17 Furthermore, NaBu modulates glycolysis in breast cancer cells, 18 dendritic epidermal T cells 19 and lung cancer cells. 20 According to a previous study, the anti-HCC effects of NaBu are dependent on the induction of apoptosis; cell cycle arrest and autophagy; and the inhibition of proliferation, migration, invasion and epithelial-to-mesenchymal transition. 17 Activation of the tumour suppressor protein p53, 21 modulation of β-catenin, 22 c-myc, 23 mitogen-activated protein kinase (MAPK), Akt-mTOR pathway 24

| Cell culture
The human HCC cell lines, HCC-LM3, Bel-7402 and SMMC-7721; the LO2 normal liver cell line; and the HepG2 human hepatoblastoma cell line 27 were purchased from the Chinese Academy of Sciences Committee Type Culture Collection cell bank (Shanghai, China). All cell lines were cultured in DMEM supplemented with 10% FBS, 100 U/ml penicillin and 100 g/ml streptomycin at 37°C under a 5% CO 2 humidified atmosphere.

| CCK8 assay
The cells in the logarithmic growth phase were added to the wells of a 96-well plate at the density of 2 × 10 4 /ml and incubated for 24 h, after which the culture medium was removed and replaced with a fresh medium containing 0, 1, 2, 3, 4 or 5 mmol/L NaBu. The cells were re-incubated for 24, 48 or 72 h. Next, 10 µl of the CCK8 solution was added to each well, and the cells were incubated for 2 h at 37℃. A microplate reader was used to detect the absorbance at 450 nm. The effects of NaBu on cell viability and the half-maximum inhibition concentration (IC 50 ) of NaBu on different cell lines were calculated using the CalsuSyn software. When exploring the effects of NaBu on the sensitivity of HCC-LM3 cells to sorafenib in vitro, the combination index (CI) and the dose reduction index (DRI) were calculated using the CalsuSyn software.

| Colony formation
Cells were seeded at the rate of 800 cells/well in a 6-well plate and incubated for 5 days. Then, 2, 3 or 4 mmol/L NaBu was added to the cells, and the plates were incubated for another 10 days and then fixed with ethanol and stained with 0.1% crystal violet.

| Flow cytometric analysis of apoptosis
Cells at the density of 4 × 10 5 /ml were seeded into 6-well plates and incubated for 24 h. After which, 3 mmol/L of NaBu was added to the cells and the plate was incubated for another 48 h. Next, the cells were collected and centrifuged at 800×g for 5 min. The cells were washed twice with PBS and then treated with 1 × binding buffer containing Annexin-V/PI for 15 min at room temperature. The cells were assessed by flow cytometer, and the results obtained were analysed with the FlowJo software (FlowJo LLC, Ashland, OR, USA). To investigate the effect of the HK2 overexpression on the apoptosis rate of HCC cells treated with NaBu, the HK2-overexpressing HCC cells were stained with phycoerythrin (PE) according to the manufacturer's protocol and then detected by flow cytometry. The results obtained were analysed by using the FlowJo software.

| Western blotting
Proteins were extracted from the nucleus and cytoplasm using nuclear and cytoplasmic protein extraction kits. The total protein and the nuclear protein were extracted by radioimmunoprecipitation assay lysis in a buffer containing protease inhibitors and phenylmethanesulfonyl fluoride (Epizyme Biotech). The protein concentration was measured using the BCA Kit (Beyotime). Then, 30 μg of the protein was loaded onto sodium dodecyl sulphate-polyacrylamide gels and transferred onto the polyvinylidene difluoride membranes.
The membranes were blocked with 5% bovine serum albumin (BSA) for 1 h at room temperature and incubated with primary antibodies overnight at 4℃ and with secondary antibodies for 1 h at room temperature. Finally, the blots were scanned using an Odyssey twocolour infrared laser imaging system (LI-COR Biosciences, Lincoln, NB, USA). The details of the primary antibodies used in the present study are presented in Table S1.

| Real-time (RT) PCR
The primers used in the present study are listed in Table S2. Total RNA was obtained with the TRIzol reagent. cDNA was reversetranscribed from total RNA using a reverse transcription kit. Then, RT-PCR was performed using the 7900 HT Fast PCR System in accordance with the manufacturer's protocol (Yeasen).

| Biochemical assays
The lactate levels of the culture medium and the tumour tissues were determined using the Lactate Testing Kit, while the serum levels of ALT, AST and creatinine were detected using appropriate kits, all produced by the Jiancheng Bioengineering Institute (Nanjing, China).
The levels of cell glucose consumption were detected as previously described, 28 and the results were normalized to the protein level.   After 20 days, the mice were anaesthetized with 1.25% pentobarbital and sacrificed. Their tumours were removed and photographed. Their blood, liver, kidney, lung and heart were collected for toxicity analyses.

| Haematoxylin and eosin staining
The tumour tissues were fixed with paraformaldehyde and then embedded in paraffin. The paraffin blocks were cut into 5μm-thick sections and subjected to haematoxylin and eosin staining.
These sections were then incubated with appropriate secondary antibodies. The positively stained areas were observed by light microscopy.

| TUNEL assay
Tumour sections (5μm thick) were dewaxed, dehydrated and digested with proteinase K. Then, the sections were incubated with the TUNEL reaction mixture. TUNEL-positive cells were observed and imaged by light microscopy.

| Statistical analysis
All experiments were performed at least thrice. Data are shown as mean ± standard deviation (SD). Differences between the groups were analysed by Student's t-tests or one-way analysis of variance.
p < 0.05 was considered to indicate statistical significance.

| NaBu inhibits cell proliferation and induces apoptosis both in vitro and in vivo
The toxicity of NaBu on the HCC and normal liver cell lines was Flow cytometric analysis indicated that 3 mM NaBu increased the apoptosis rate in the HCC-LM3 and Bel-7402 cell lines ( Figure 1C).
Bcl-2 is an anti-apoptotic protein whereas Bax, cleaved caspase 3 and cleaved caspase 9 are pro-apoptotic proteins. Western blotting analysis revealed that NaBu decreased the expression of Bcl-2 but increased the expression of cleaved caspase 9 and 3 ( Figure 1D).
NaBu did not affect the expression of Bax ( Figure 1D). These results suggested that NaBu induced the apoptosis of the HCC cells in vitro.
Furthermore, the results of Western blotting showed that NaBu did not affect the expression of Bcl-2, Bax or cleaved caspase 3 in the LO2 cell line ( Figure S1), which indicated that NaBu had a lower capacity to induce apoptosis in the LO2 cell line.
Subsequently, a xenograft model was generated by injecting the HCC-LM3 cell line to investigate the effects of NaBu on proliferation, apoptosis and glycolysis of the HCC cells in vivo. The gavage of 200 mg/kg NaBu significantly decreased the tumour volume ( Figure 1E), whereas it did not affect the weight of the mice ( Figure 1F). Haematoxylin and eosin staining showed that there were more necrotic lesions in the NaBu group than in the NC group. TUNEL assay exposed that there were more apoptotic cells in the NaBu group than in the NC group. Ki67 has been viewed as a proliferation marker in cancer cells. In the present study, Ki67 staning showed that there were more positive cells in the NC group than in the NaBu group ( Figure 1G). The results of Western blotting showed that 200 mg/kg NaBu inhibited the expressions of Bcl-2 and PCNA in the mouse tumour tissues ( Figure 1H). These results indicated that NaBu suppressed the proliferation and induced the apoptosis of the HCC cells in vivo.
Moreover, NaBu did not affect the mouse serum levels of ALT, AST or creatinine ( Figure S2A). Furthermore, no obvious pathological changes were observed in the liver, lung, heart or kidney sections of the NaBu-treated mice ( Figure S2B), which implied that the gavage of 200 mg/kg of NaBu daily for 16 days did not harm these organs.
Collectively, the above results show that NaBu induces apoptosis and suppresses the proliferation of the HCC cells both in vivo and in vitro.

| NaBu inhibits aerobic glycolysis in the HCC cells both in vitro and in vivo
Next, the effects of NaBu treatment on aerobic glycolysis in the NaBu-sensitive cell lines HCC-LM3 and Bel-7402 were determined.
The results showed that NaBu treatment for 48 h inhibited lactate

| The modulation of aerobic glycolysis, proliferation, and apoptosis of the HCC cells by NaBu is associated with the inhibition of HK2
In the present study, among the eight glycolysis-related genes, the mRNA level of HK2 witnessed the highest decrease in the HCC-LM3  Figure 3A and B). The results of Western blotting indicated that the protein expression of HK2 in the HK2-EV and HK2-OE HCC cell lines could be downregulated by NaBu ( Figure 3F). Figure 3D and E, compared with the HK2-EV group, the lactate production and glucose consumption in both the HCC cell lines were enhanced by HK2 expression. Moreover, the inhibitory effects of NaBu on lactate production and glucose consumption were impaired in the HK2-OE group compared with the HK2-EV group ( Figure 3C and D). These results suggest that HK2 is essential for the inhibition of aerobic glycolysis upon NaBu treatment in the HCC cells.

As shown in
As shown in Figure 3E, the viability of the HCC cells in the HK2-OE group was higher than that in the HK2-EV group. Furthermore, HK2 overexpression reversed the decrease in proliferation induced by  Figure 3G).
These findings imply that HK2 is essential for the inhibition of aerobic glycolysis and cell proliferation and for the induction of apoptosis upon NaBu treatment in the HCC cells.

| NaBu inhibits HK2 via the c-myc pathway
Next, the signalling pathways involved in NaBu-induced HK2 inhibition were explored. The HK2 expression was modulated by several signalling pathways and transcription factors, including    Figure 4L). The above results imply that NaBu downregulates HK2 expression via the c-myc pathway, thereby suppressing aerobic glycolysis in the HCC cells.

| NaBu enhances the chemosensitivity of sorafenib in the HCC cells both in vitro and in vivo
Sorafenib is a first-generation targeted therapeutic drug for advanced HCC. Glucose consumption and lactate export are increased and mitochondrial OXPHOS is impaired in the HCC cells after sorafenib treatment. 30 Moreover, lactate can function as a direct signal instructing cancer-associated fibroblast that secretes hepatocyte growth factor, thus reducing the sensitivity of the tumours to sorafenib treatment. 31 Inhibition of key enzymes in the glycolytic pathway via lentivirus transfection or drugs enhances the sensitivity of the HCC cells to sorafenib. 10,32,33 In the present study, NaBu was found to inhibit the glycolysis influx of the HCC cells. However, the effect of NaBu treatment on the sensitivity of the HCC cells to sorafenib remains to be investigated.  Figure 1A). Thus, the ratio of the IC 50 of sorafenib to that of NaBu on the cell line was approximately 1:500. Next, different concentrations of sorafenib were combined with NaBu (at a fixed ratio of 1:500) to determine whether NaBu can enhance the sensitivity of the HCC-LM3 cell line to sorafenib. As shown in Figure 5A, compared with sorafenib or NaBu alone, the combination of sorafenib and NaBu (1:500) exerted a greater cytotoxic effect on the HCC-LM3 cell line. Moreover, the results of the median dose-effect analysis indicated that the combination index (CI) in the HCC-LM3 cells was <1, which implied the synergistic effect of sorafenib and NaBu ( Figure 5B). The DRI of NaBu was >1, which suggested that NaBu can lower the dose of sorafenib ( Figure 5B).
Moreover, the inhibitory effects of sorafenib on the colony-forming ability and PCNA expression of the HCC-LM3 cell line were further enhanced by NaBu treatment (Figure 5C and D). These results denote that NaBu enhances the inhibitory effects of sorafenib on the proliferation of the HCC cells.
Flow cytometric analysis showed that 2 mM NaBu and 4 μM sorafenib exerted comparable effects on the apoptosis rate of the HCC-LM3 cell line, and NaBu treatment was found to enhance the apoptosis induced by sorafenib ( Figure 5E). Moreover, the reduction in Bcl2 expression following sorafenib treatment in the HCC-LM3 cell line was further suppressed by NaBu ( Figure 5D).
The lactate production, glucose uptake and protein expression of the rate-limiting enzymes involved in the glycolytic pathway were improved in the HCC-LM3 cell line after sorafenib treatment, whereas mitochondrial OXPHOS was impaired ( Figure 5F and G).
These results were consistent with those obtained from our previous study. 28 However, these trends were reversed upon NaBu treatment ( Figure 5F and G). The above findings signify that NaBu Collectively, these results establish that NaBu treatment enhances the chemosensitivity of the HCC cells to sorafenib both in vitro and in vivo.

| DISCUSS ION
In the 1920s, Warburg et al. observed that rat liver carcinoma tissues consumed higher amounts of glucose than the surrounding normal liver tissues. Moreover, the liver carcinoma tissues tended to catabolize glucose to lactate via glycolysis instead of converting it to CO 2 and H 2 O via the citric acid cycle even in the presence of oxygen. 3 This phenomenon termed the 'Warburg effect' or aerobic glycolysis has been observed in HCC, 4 lung cancer 34 and other kinds of cancer. In aerobic glycolysis, 2 moles of ATP are produced from 1 mole of glucose. In contrast, 36 moles of ATP are produced from 1 mole of glucose in mitochondrial OXPHOS. 35 Although aerobic glycolysis seems inefficient in terms of ATP production, the speed of ATP production via aerobic glycolysis is much higher than that via mitochondrial OXPHOS per unit time. 7 High-speed glucose fermentation via aerobic glycolysis helps cancer cells consume more glucose than normal cells. 36 This rapid ATP generation via aerobic glycolysis aids in the proliferation of the cancer cells. 37 Glycolytic intermediates, such as dihydroxyacetone phosphate and 3-phosphoglycerate, support the biosynthetic programs of the cancer cells and promote their proliferation and growth. 38 Moreover, lactate production via aerobic glycolysis acidifies the tumour microenvironment, which enhances immune evasion, invasion, metastasis, angiogenesis, drug resistance and radioresistance of the cancers. 7,38,39 A recent review has systematically summarized the noncanonical functions of key enzymes involved in the glycolytic pathway. 8 For instance, HK2 regulates autophagy and mitochondria-dependent apoptosis. PKM2 regulates mitosis, cell cytokinesis, homologous recombination repair, apoptosis, tumour cell exosomes and ectosomes, and PFKFB3 regulates DNA repair. 8 Hence, targeting the enzymes that play a key role in aerobic glycolysis could regulate various phenotypes of the cancer cells and has been regarded as a novel HCC treatment strategy. 40 Preclinical studies have signified that glycolytic inhibitors, such as chrysin (targeting HK2), 41  NaBu is a sodium salt of the SCFA butyrate. Experimental studies have shown that NaBu exerts a protective effect against various liver diseases, including autoimmune hepatitis, 49 liver fibrosis 50 and non-alcoholic fatty liver disease. 51 NaBu exerts dual effects on the HCC cells, that is a low-dose promotes their proliferation, 52 whereas a high-dose inhibits their proliferation, invasion and metastasis. Moreover, a high dose induces G0/G1 arrest, apoptosis and autophagy of the HCC cells, thus exerting anti-HCC effects. 17,23,24,52 NaBu has been reported to inhibit aerobic glycolysis in the lung 20 and breast cancer cells. 18 However, the effects of NaBu on aerobic glycolysis in the HCC cells and whether these effects are related to the modulation of key enzymes in the glycolytic pathway are yet to be unravelled. In the present study, NaBu was found to apoptosis and inhibit the proliferation of the NaBu-sensitive HCC cell HK2 is the first rate-limiting enzyme in the glycolytic pathway and catalysed the conversion of glucose to glucose-6-phosphate. 53 Five isoforms of HK(HK1, HK2, HK3, HK4 and HKDC1) have been identified in humans. 54 In normal hepatocytes, HK4 or glucokinase is the major isoform. 55 However, in the HCC cells, HK2 is usually the only expressed HK. This isoform is not expressed in most normal adult tissues, including hepatocytes. 56 Moreover, no adverse physiological effects have been found in mice with systematic HK2 deletion. 57 DeWaal et al. reported that the upregulation of HK2 begins during liver cirrhosis, increases in dysplasia and reaches peaks in HCC. 56 Furthermore, the upregulation of HK2 and the enhancement of glycolysis have been reported to play pertinent roles in the activation of HSCs and liver fibrosis, 13,58 which is the main contributor to liver cirrhosis and HCC. 59 These findings show that targeting HK2 could be an excellent and safe strategy for treating HCC, even in the early phase of hepatocellular carcinogenesis. HK2 binds to VDAC1 on the outer mitochondrial membrane via its hydrophobic N terminal, which augments ATP production and promotes the proliferation of the HCC cells. 60,61 Moreover, the binding of HK2 and VDAC1 suppresses the binding of Bax to VDAC1 and the release of cytochrome C, which prevents the HCC cells from mitochondriaassociated apoptosis. 41 Hepatic HK2 deletion has been reported to suppress diethylnitrosamine-induced hepatocarcinogenesis in mice.
Knockdown of HK2 in the HCC cells inhibited the proliferation and aerobic glycolysis while enhancing mitochondrial OXPHOS and apoptosis. Moreover, inhibition of HK2 increased the sensitivity of the HCC cells to sorafenib and metformin. 56 2-Deoxy-d-glucose (2-DG), a synthetic glucose analog that can inhibit HK2 activity, has been shown to inhibit the proliferation, metastasis and invasion of the HCC cells and induce their apoptosis. 42,43 Moreover, 2-DG enhances the sensitivity of the HCC cells to sorafenib. 42 32 In the present study, the results of the CCK8 assay indicated that HCC-LM3 and Bel-7402 cell lines were more sensitive to NaBu treatment than SMMC-7721, HepG2 and LO2 cell lines. This finding indirectly demonstrates that the modulation of HK2 and glycolysis plays a key role in the anti-HCC effects of NaBu.
C-myc is a common oncogene that enhances aerobic glycolysis in the cancer cells by transcriptionally activating GLUT1, HK2, PKM2 and LDH-A. 62 NaBu suppresses the expression of c-myc in HCC cells. 23 In the present study, the mRNA expression of c-myc was significantly inhibited in both HCC-LM3 and Bel-7402 cell lines by NaBu. The results of Western blotting showed that NaBu inhibited the expression of c-myc in the total and nuclear lysate in a dose-depedent manner. NaBu suppressed the expression of c-myc in the tumour tissues. After transfection with c-myc-overexpressing lentivirus, the inhibitory effects of NaBu on the expressions of total c-myc, nuclear c-myc, and HK2 and lactate production and glucose uptake were attenuated. Furthermore, these effects were reversed by HK2 knockdown. In addition, when the c-myc expression was ablated by siRNA transfection, 3 mM NaBu failed to modulate the protein expression of HK2 in the HCC cells, which implies that cmyc was actually the target of NaBu when inhibiting HK2 expression ( Figure 4). The above findings indicate that the modulatory effect of NaBu on the c-myc pathway is essential for the inhibition of HK2 and aerobic glycolysis by NaBu in the HCC cells.
Sorafenib is the first-line target drug to treat late-stage HCC and prolong the survival of patients. However, few patients derive long-term benefits from sorafenib owing to the early occurrence of sorafenib resistance. 63 Many mechanisms underly sorafenib resistance in the HCC cells. Enhanced lactate production and glucose consumption and impaired mitochondrial OXPHOS have been observed in the HCC cells after sorafenib administration. 28,30 Key enzymes in the glycolytic pathway, including GLUT1, HK2, PFK1, PKM2 and LDH-A, are overexpressed in sorafenib-resistant HCC cell lines, which upregulates aerobic glycolysis. 9,10,64 Inhibiting these enzymes has been reported to exhibit synergistic effects with sorafenib in treating HCC. 28 66 In the present study, aerobic glycolysis was enhanced and mitochondrial OXPHOS was impaired in the HCC-LM3 cell line after sorafenib treatment, which was reversed by NaBu treatment. Accordingly, NaBu augmented the ability of sorafenib to inhibit proliferation and induce apoptosis in the HCC cells both in vitro and in vivo ( Figure 5).
There are some limitations to our research. For instance, NaBu treatment suppressed the protein expressions of PFK1 and LDH-A in HCC-LM3 and Bel-7402 cell lines in vitro. However, the results of IHC staining and Western blotting implied that 200 mg/kg NaBu did not affect the expression of PFK1 or LDH-A in mouse tumour tissues. This discrepancy warrants further investigation.

| CON CLUS ION
Our findings suggest that NaBu, a gut microbiota metabolite, inhibits the expression of HK2 and subsequently downregulates aerobic glycolysis and cell proliferation, and induces apoptosis by suppressing the c-myc pathway. NaBu impairs the enhancement of aerobic glycolysis in the HCC cells by sorafenib and improves the effect of the drug on HCC cells both in vitro and in vivo. The findings of the present study may aid us in better understanding the role of butyrate in hepatocellular carcinogenesis and the treatment of HCC.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.