Berberine, anti-rabbit IgG, anti-mouse IgG and anti-β-actin IgG were purchased from Sigma-Aldrich (Shanghai, China). Antibodies for HIF-1α (cat. no. 14179), mTOR (cat. no. 2983) and phospho-mTOR (Ser2448) (cat. no. 2971) were purchased from Cell Signaling Technology Inc. (Beverly, MA, USA). 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (2-NBDG) was purchased from Thermo Fisher Scientific (Eugene, OR, USA). All other reagents including DFX and MG132 were obtained from Sigma-Aldrich unless stated otherwise.

Human colon cancer cell lines HCT116 and KM12C were obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). HCT116 cells were cultured in McCoy'5A medium (Thermo Fisher Scientific, Hudson, NH, USA). KM12C cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Thermo Fisher Scientific). All the medium was supplemented with 10% fetal bovine serum (FBS; GE Healthcare Life Sciences, HyClone Laboratories, Logan, UT, USA), 100 µg/ml of streptomycin (Life Technologies; Thermo Fisher Scientific) and 100 units of penicillin.

Cell viability was detected using the 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Cells were seeded at a density of 1×10⁴ cells/well in 96-well plates and then treated with 0–100 µM berberine as indicated. At the time-points (24 h for HCT116 cells and 15 h for KM12C cells), 50 µl of 5 mg/ml MTT solution was added to each well and incubated at 37°C for 4 h. The formazan crystals formed were then dissolved in 150 µl of dimethyl sulfoxide (DMSO), and the absorbance of the solution was then obtained on a microplate reader at λ570 nm. Results are presented as percentage loss of cell viability compared with the control.

Cells were seeded in a 6-well plate one day before the experiment. Cells were treated with berberine for 24 h (HCT116) or 15 h (KM12C), and then washed with phosphate-buffered saline (PBS), harvested by trypsinization, counted, and were seeded into 6-well dishes at 600 cells/well. The cells were incubated for another 10 days, fixed and stained with 1% crystal violet in ethanol. Then, the colonies in 6-well plates were photographed using a scanner (Epson Perfection V330; Epson Corp., Nagano, Japan).

Glucose uptake of cells was measured by 2-NBDG uptake as previously described (11). Briefly, the cells were seeded in a 12-well plate at a density of 70–80%, and treated with 0–100 µM of berberine for 24 h (HCT116) or 15 h (KM12C). After treatment, the cells were harvested and resuspended in Krebs-Ringer's HEPES Buffer (KRB) solution, and incubated with 100 nmol/l 2-NBDG at 37°C in 5% CO₂ for 30 min. The 2-NBDG uptake reaction was terminated by removing the incubation medium and washing the cells twice with pre-cold PBS. Cells were resuspended in 1 µg/ml propidium iodide (PI) solution to exclude dead cells. Glucose uptake was determined by measuring the fluorescence intensity of 2-NBDG in PI-negative cells.

Cells were seeded in a 6-well plate at a density of 70–80% and treated with 0–100 µM berberine for 24 h (HCT116) or 15 h (KM12C). Total RNA was isolated using TRIzol reagent (Takara Biotechnology Co., Ltd., Dalian, China). RNA was reverse-transcribed into cDNA using PrimeScript^™ RT reagent kit (Takara Biotechnology Co., Ltd.) according to the manufacturer's protocol. Real-time quantitative polymerase chain reaction (PCR) was carried out with the SYBR-Green I fluorescent dye method (SYBR^® Premix Ex Taq™ II; Takara Biotechnology Co., Ltd.) and the StepOnePlus Real-Time PCR apparatus (Applied Biosystems; Thermo Fisher Scientific). The sequences of primers used were as follows: Forward, 5′-TATTGCACTGCACAGGCCACATTC-3′ and reverse, 5′-TGATGGGTGAGGAATGGGTTCACA-3′ for HIF-1α; forward, 5′-GGCATTGATGACTCCAGTGTT-3′ and reverse, 5′-ATGGAGCCCAGCAGCAA-3′ for GLUT1; forward, 5′-TCACGGAGCTCAACCATGAC-3′ and reverse, 5′-CTGCAGTAGGGTGAGTGGTG-3′ for HK2; forward, 5′-GCCCGACGTGCATTCCCGATTCCTT-3′ and reverse, 5′-GACGGCTTTCTCCCTCTTGCTGACG-3′ for LDHA; forward, 5′-CGTGTACTACAATGAGGCTGC-3′ and reverse, 5′-CTGGTCTGAAGATCTGGCCG-3′ for β-tubulin. The amplification specificity was checked by melting curve analysis. The relative expression of miRNA was calculated by the 2^−ΔΔCt method as described by Livak and Schmittgen (12).

Cells at 60–80% confluence were washed with PBS and lysed directly into SDS-PAGE loading buffer. A total of 20 µg of protein was analyzed by SDS-PAGE and transferred to PVDF membranes. All primary antibodies were used at 1:1,000 in 5% milk in Tris-buffered saline with 0.05% Tween. Immunopositive bands were visualized by Amersham ECL™ Plus Western Blotting Detection kit (GE Healthcare, Chicago, IL, USA).

The SigmaPlot version 11.0 software package (Systat Software, Inc., San Jose, CA, USA) was used for statistical analysis. The results are presented as mean ± standard error (SEM). Data were analyzed by one way analysis of variance (ANOVA) or the Student's t-test. P<0.05 was considered to indicate a statistically significant result.

Berberine has been shown to enhance glucose uptake in cell lines including 3T3 adipose cells (5,13) and L6 myotubes (7). To test its effect on colon cancer cells, two colon cancer cell lines, HCT116 and KM12C, were treated with 0, 6.25, 12.5, 25, 50 and 100 µM of berberine for 24 or 15 h respectively, and the glucose uptake of these cell lines was assessed. In contrast to berberine's reported effect of enhancing glucose update, we unexpectedly revealed that berberine inhibited glucose uptake in the colon cancer cell lines. As shown in Fig. 1A and B, treatment with the different concentrations of berberine significantly reduced glucose uptake in these two cell lines. At concentrations between 6.25 and 100 µM, berberine decreased glucose uptake of HCT116 by 40–90%, and by 10–75% in KM12C cells, which was slightly less sensitive. We then validated the anticancer effect of berberine by detecting cell proliferation after berberine treatment. MTT assay indicated that berberine decreased the viability of the HCT116 and KM12C cells in a concentration-dependent manner (Fig. 1C and D). To further investigate the long-term effect of berberine on cell growth, colony forming assay was performed. Similarly, berberine inhibited the growth of colon cancer cells, manifested as the reduction in the number of colonies of the HCT116 and KM12C cells in a concentration-dependent manner (Fig. 1E). These results indicated that berberine inhibits glucose uptake in colon cancer cells, and these effects may contribute to its antitumor effect.

We then further investigated the mechanism by which berberine inhibits glucose uptake. Glucose transporter 1 (GLUT1) is the dominate glucose transport gene in colon cancer cells (14,15), which facilitates the transport of glucose across the plasma membranes. We examined the transcription of GLUT1 after berberine treatment. As shown in Fig. 2A and B, treatment with berberine inhibited the mRNA level of GLUT1 in a concentration-dependent manner. At concentrations between 6.25 and 25 µM, berberine decreased the mRNA level of GLUT1 in the HCT116 cells by 50–70%, and similar to the data of glucose uptake, KM12C cells were less sensitive. Berberine (6.25–25 µM) reduced the mRNA level of GLUT1 by 10–35% in the KM12C cells. These data indicated that the inhibitory effect of berberine on glucose uptake may be mediated by the transcription inhibition of GLUT1.

The altered energy metabolism of cancer cells is not only attained by enhancing glucose uptake, but also a higher rate of glycolysis. Thus, we examined the transcription of two glycolytic enzymes following berberine treatment: Lactate dehydrogenase A (LDHA) and hexokinases 2 (HK2). LDHA catalyzes the inter-conversion of pyruvate and L-lactate with concomitant inter-conversion of NADH and NAD⁺. HK2 phosphorylates glucose to produce glucose-6-phosphate (G6P), the first step in most glucose metabolism pathways. As shown in Fig. 2C-F, treatment with 0–100 µM berberine inhibited the mRNA levels of LDHA and HK2 in a concentration-dependent manner as assessed by qPCR. Thus, our data indicated that berberine may downregulate glucose metabolism in colon cancer cells by inhibition of the transcription of glucose metabolism-related genes.

We next explored how berberine regulates the transcription of the above glucose metabolism-related genes. It is well known that the enhanced glucose uptake and glycolysis of cancer cells is partly due to highly expressed HIF-1α. HIF-1α is upregulated in low O₂ concentrations but also by oncogene activation or loss of tumor suppressors (16). Upregulated HIF-1α would further facilitate the transcription of genes involved in glucose uptake and glycolysis-related genes, including GLUT1, LDHA and HK2 (17,18). Therefore, we next examined the effect of berberine on the mRNA and protein level of HIF-1α. As determined by qPCR, we found that treatment with berberine did not alter the mRNA level of HIF1A in the HCT116 and KM12C cells (Fig. 3A and B). Western blotting demonstrated that KM12C cells expressed HIF-1α protein even in normoxic condition. Treatment with 0–100 µM berberine decreased the protein level in a concentration-dependent manner (Fig. 3C). These data indicated that berberine negatively modulates glucose metabolism through regulation of HIF-1α at the post-transcriptional level.

We then aimed to resolve whether berberine affects HIF-1α degradation. It is known that HIF-1α protein degradation is regulated by prolyl hydroxylation and then the ubiquitin-dependent proteasome pathway (19). Briefly, HIF-1α is hydroxylated by prolyl hydroxylases (PHD1-3) at proline (Pro)-402 and −564. Hydroxylated HIF-1α then binds to the von Hippel-Lindau tumor-suppressor protein (VHL). VHL is the recognition component of an E3 ubiquitin-protein and ubiquitylated HIF-1α is rapidly degraded by the proteasome (20). We first treated KM12C cells with desferrioxamine (DFX), an iron chelator, which is known to inhibit hydroxylation of HIF-1α by chelating the iron required for activity of the HIF-1α-specific proline hydroxylases, and thus blocks HIF-1α degradation (21). As shown in Fig. 4A, when HIF-1α degradation was blocked by DFX, berberine was still able to decrease the expression of HIF-1α. Similarly, treatment with proteasome inhibitor, MG132, was unable to inhibit berberine's effect on the reduction of HIF-1α expression (Fig. 4B). These results demonstrated that the negative post-transcriptional regulation of berberine on HIF-1α expression was not by proteasomal degradation.

We then aimed to ascertain whether berberine interrupts HIF-1α protein translation. Mammalian target of rapamycin (mTOR) activity is a major determinant of the rate of HIF-1α protein synthesis (22,23). We then detected the activity of the mTOR pathway in KM12C cells after the treatment of berberine. As shown in Fig. 4C, berberine significantly inhibited the phosphorylation of mTOR (p-mTOR) while not altering the total expression level of mTOR protein. Taken together, our data indicated that berberine inhibits HIF-1α protein expression by inhibition of the mTOR pathway, and thereby interruption of HIF-1α protein synthesis.