DMEM/F12, CelLytic MT mammalian tissue lysis/extraction reagents, phloretin, chlorogenic acid, caffeine, caffeic acid and ferulic acid were purchased from Sigma-Aldrich; Merck KGaA. FBS was purchased from Gibco; Thermo Fisher Scientific, Inc. Polyclonal rabbit anti-SGLT1 (1:250; cat. no. 07-1417), GLUT2 (1:250; cat. no. 07-1402-I), 5' AMP-activated protein kinase (AMPK; 1:250; cat. no. 07-1417) and phospho-AMPK (1:250; cat. no. 07-363) were purchased from Merck KGaA. Monoclonal mouse alkaline phosphatase (ALP; 1:500; cat. no. NB110-3638) antibodies were obtained from Novus Biologicals, LLC. Monoclonal mouse anti-β-actin (1:500; cat. no. ab8224) was purchased from Abcam. 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (2-NBDG) was obtained from Thermo Fisher Scientific, Inc. Phlorizin was purchased from Tocris Bioscience. All other chemicals with high purity were obtained from commercial sources.

Coffea arabica beans were kindly provided by Hillkoff, and a voucher specimen (no. NU003806) was deposited at the herbarium of the Faculty of Biology, Naresuan University, Phitsanulok, Thailand. Dried Coffea arabica beans were weighed and blended thoroughly, followed by 100˚C water infusion for 30 min. Subsequently, the liquid extract was filtered through filter paper (Whatman plc; GE Healthcare Life Sciences) three times. The filtrate was evaporated using a lyophilizer (Labconco). The chemical constituents of CBE, caffeine, chlorogenic acid, caffeic acid and ferulic acid (Sigma-Aldrich; Merck KGaA), were used as references for validation and quantitation by high-performance liquid chromatography (HPLC) with diode array detection on an Agilent 1200 series chromatograph (LabX) at the Science and Technology Service Centre, Faculty of Science, Chiang Mai University (Chiang Mai, Thailand) according to the ISO/IEC 17025 method (International Organization for Standardization, 2005) (15). The flow rate was 0.5 ml/min. The mobile phase was a binary solvent system consisting of 2% acetic acid dissolved in HPLC water (solvent A) and absolute methanol (solvent B). The column used was an Eclipse XDB-C18 (4.6x150 mm; 5 µm) and the absorption wavelength selected was 280 nm.

The Caco-2 human colorectal adenocarcinoma cell line was purchased from American Type Culture Collection. Cells in the 2nd-22nd passage were grown in DMEM/F12 containing 20 and 1% penicillin-streptomycin in a humidified atmosphere with 5% CO2. Caco-2 cells were seeded in 24- and 96-well plates at a density of 5x10⁴ cells/well and were cultured at 37˚C for 18-21 days to create a differentiated Caco-2 monolayer. Differentiated Caco-2 monolayer forms a continuous barrier between the apical and basolateral compartments and express several morphological and functional characteristics of the absorptive enterocytes as found in the small intestine (16). The medium was replaced every 3 days during culture.

Differentiated Caco-2 cells were washed three times with PBS (pH 7.4) and incubated with 2 mM 2-NBDG in the presence or absence of 0.1-1,000 µg/ml CBE, a molar ratio of four major active compounds of chlorogenic acid (29.62 µg/ml), caffeine (5.11 µg/ml), caffeic acid (3.16 µg/ml), ferulic acid (1.54 µg/ml), phlorizin (0.5 mM) and phloretin (1 mM) for 4 h. The medium was removed and the cells were washed three times with ice-cold PBS. The fluorescence intensity was measured using a Synergy™ HT microplate reader (BioTek Instruments, Inc.) at excitation and emission wavelengths of 485 and 535 nm, respectively.

The viability of Caco-2 cells after exposure to CBE and other test compounds was determined using an MTT assay. Caco-2 cells were seeded at a density of 5x10⁴ cells/well in 96-well plates and cultured for 18-21 days at 37˚C. On the day of the experiment, culture medium with or without test compounds was added and the cells were incubated for 4 h at 37˚C. After exposure, the cells were washed with PBS, culture medium containing MTT reagent was added and the cells were incubated for the next 4 h at 37˚C. At the end of the experiment, the MTT solution was aspirated and the cells were washed once with ice-cold PBS. DMSO was added to each well, and cells were incubated for a further 30 min at 37˚C. The absorbance of dissolved formazan was measured at a wavelength of 570 nm using an M965 AccuReader microplate reader (Metertech, Inc.). The sample detected at a wavelength of 680 nm was also used as a reference.

The dose of CBE used in this experiment was inferred from the glucose transport experiment, which revealed that CBE at a dose of 100 µg/ml reduced glucose transport to a similar extent to that observed with CBE at a dose of 1,000 µg/ml. Therefore, CBE at a dose of 100 µg/ml was selected for further experiments. Inhibition of sucrase enzyme activity was determined as previously described, with some modifications (17). Briefly, differentiated Caco-2 cells were rinsed three times with Hank's balanced salt solution (HBSS) (Sigma-Aldrich; Merck KGaA). The cells were subsequently incubated with HBSS containing 28 mM sucrose, which is an α-glucosidase substrate, in the presence or absence of 100 µg/ml CBE, a molar ratio of four major active compounds of CBE at 100 µg/ml of chlorogenic acid (29.62 µg/ml), caffeine (5.11 µg/ml), caffeic acid (3.16 µg/ml), ferulic acid (1.54 µg/ml) and acarbose, disaccharidase inhibitor (50 µM) at 37˚C for 40 min. After incubation, the incubated solution was collected. The glucose concentration was assayed using commercial enzymatic assay kits (cat. no. BLT0002) obtained from Biotechnical, Co., Ltd. Cells were lysed by three freeze-thaw cycles and protein concentration was determined using a Bradford assay (Bio-Rad Laboratories, Inc.). The data obtained are expressed as units per mg protein.

Previous studies have demonstrated that sterol regulatory element-binding protein-1 (SREBP-1c) and hepatocyte nuclear factor 1α (HNF1α) are involved in SGLT1 and GLUT2 expression and function. Therefore, the present study measured SREBP-1c and HNF1α mRNA expression in Caco-2 cells after 4 h of exposure to CBE and other test compounds. Total RNA was extracted and purified from Caco-2 cells using TRIzol^® reagent (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. First strand cDNA was synthesized using a SensiFAST™ cDNA synthesis kit (Bioline; Meridian Bioscience) and qPCR was performed using SYBR Real-Time PCR MasterMix (Bioline; Meridian Bioscience) on a CFX Touch real-time PCR system (Bio-Rad Laboratories, Inc.). PCR amplifications were performed in a 20-µl volume with the following thermocycling conditions: A polymerase enzyme activation step at 95˚C for 2 min; followed by 40 cycles consisting of 5 sec of denaturation at 95˚C, 10 sec of annealing at 60˚C depending on primers, and 10 sec of elongation at 72˚C. Forward and reverse primers were purchased from Macrogen, Inc. and used at a final concentration of 0.4 µM. The specific primer sets for human SREBP-1c, HNF1α and GAPDH were as follows: Human (h)SREBP-1c forward, 5'-ATACCACCAGCGTCTACC-3' and reverse, 5'-CACCAACAGCCCATTGAG-3'; hHNF1α forward, 5'-TACACCTGGTACGTCCGCAA-3' and reverse, 5'-CACTTGAAACGGTTCCTCCG-3'; and hGAPDH forward, 5'-AGCCTTCTCCATGGTGGTGAAGAC-3' and reverse, 5'-CGGAGTCAACGGATTTGGTCG-3'. The results were calculated using the 2^-ΔΔCq method (18), normalized to GAPDH mRNA levels and reported as the relative fold change. qPCR amplification was performed in duplicate for each synthesized cDNA set.

To measure protein expression of the glucose transporters SGLT1, GLUT2, AMPK and phosphorylated AMPK (p-AMPK), which is the SGLT1 and GLUT2 regulator (19,20), subcellular fractions were prepared using differential centrifugation. Caco-2 cells were lysed using lysis buffer containing 1% complete protease inhibitor mixture according to the manufacturer's protocol. Briefly, samples were disrupted using a homogenizer and centrifuged at 5,000 x g for 10 min at 4˚C. The supernatant was designated as whole-cell lysate. Half of the supernatant was recentrifuged at 100,000 x g for 2 h at 4˚C. The supernatant fraction from this step was designated as the cytosolic fraction, while the pellet was re-suspended with the same buffer and used as the membrane fraction. The protein concentration in each sample was also determined using a Bradford assay (Bio-Rad Laboratories, Inc.), and the samples were stored at -80˚C prior to use. For western blot analysis, samples were resolved in 4X Laemmli solution, electrophoresed by 10% SDS-PAGE and transferred onto PVDF membranes (Merck KGaA). Non-specific binding was eliminated by blocking with 5% (w/v) non-fat dry milk in 0.05% Tween-20 in Tris-buffered saline (TBS-T) for 1 h at room temperature and incubated overnight at 4˚C with polyclonal anti-rabbit SGLT1 (1:250; cat. no. 07-1417), polyclonal anti-rabbit GLUT2 (1:250; cat. no. 07-1402-I), polyclonal anti-rabbit AMPK (1:250; cat. no. 07-1417), polyclonal anti-rabbit p-AMPK (1:250; cat. no. 07-363), monoclonal anti-mouse ALP (BBM marker) (1:500; cat. no. NB110-3638), monoclonal anti-mouse Na+K+ATPase (BLM marker) (1:500; cat. no. NB300-146) or anti-mouse β-actin (1:500; cat. no. ab8224) antibodies. The PVDF membrane was washed with TBS-T buffer and incubated with horseradish peroxidase-conjugated ImmunoPure secondary goat anti-rabbit or anti-mouse IgG (1:1,000 dilution; cat. no. AP132P; Merck KGaA) for 1 h at room temperature. Proteins were detected using Clarity Western ECL Substrate (Bio-Rad Laboratories, Inc.) and semi-quantitatively analyzed using ImageJ version 1.44p (National Institutes of Health).

Data are presented as the mean ± the standard error of the mean. Statistical analysis was performed using SPSS version 23 (IBM Corp.). A one-way ANOVA followed by a post-hoc Dunnett's test was used to compare differences between multiple groups. P<0.05 was considered to indicate a statistically significant difference.

The HPLC chromatogram of CBE reliably detected chlorogenic acid, caffeine, caffeic acid and ferulic acid with retention times of 17.163, 19.761, 21.867 and 29.626 min, respectively, compared with their respective references (Fig. 1). A relatively high content of chlorogenic acid was obtained (296.2 mg/g) compared with other phenolic compounds. In addition, caffeine, a major component in Coffea arabica beans, was found at 51.1 mg/g of CBE, whereas caffeic acid and ferulic acid were present at 31.6 and 15.4 mg/g of CBE, respectively.

To further demonstrate the inhibitory effect of CBE and its major constituents on glucose absorption, a glucose transport experiment was performed. As shown in Fig. 2A, CBE at 10-1,000 µg/ml inhibited glucose transport compared with the control similar to phlorizin (an SGLT1 inhibitor) and phloretin (a GLUT2 inhibitor). These data indicated that CBE may interfere with the transport function of SGLT1 and GLUT2 resulting in decreased glucose uptake into Caco-2 cells. In addition, the major constituent, caffeic acid, also reduced glucose uptake to a similar degree to 100 and 1,000 µg/ml CBE, suggesting that inhibition of glucose uptake by CBE may be partly affected by caffeic acid.

Cell viability under exposure to test compounds was also examined. CBE (0.1-1,000 µg/ml), caffeine, chlorogenic acid, caffeic acid, ferulic acid, phlorizin and phloretin did not affect Caco-2 cell viability compared with that of control cells (Fig. 2B). These data suggested that CBE and its major components exerted an inhibitory effect on glucose absorption without any cytotoxic effect.

To determine the effect of CBE on sucrase activity, the glucose concentration in the substrate solution was evaluated. As shown in Fig. 3, co-incubation with CBE at 100 µg/ml decreased the glucose concentration compared with the control group. Furthermore, the constituents of CBE (caffeine, chlorogenic acid and caffeic acid) were also reduced compared with the control. The inhibitory effect was similar to that of the positive control, acarbose, which is a sucrase enzyme inhibitor. Therefore, CBE may not only reduce glucose transporter expression but also decrease disaccharidase enzyme activity, leading to decreased glucose transport across the human intestinal cell membrane.

To further determine whether the reduced rate of glucose transport caused by CBE was due to decreased membrane expression of SGLT1, cellular protein expression of SGLT1 was determined using western blotting. As shown in Fig. 4A and B, SGLT1 membrane protein expression was decreased by CBE and its constituent, caffeic acid, compared with the control group. Similarly, phlorizin (an SGLT1 inhibitor) also decreased SGLT1 protein expression in Caco-2 cells, whereas cells treated with caffeine, chlorogenic acid, ferulic acid and phloretin did not exhibit any notable differences. Accordingly, CBE-induced downregulation of SGLT1 membrane protein expression resulted in a reduction of glucose uptake in human intestinal epithelia.

As mentioned above, GLUT2 is also associated with intestinal glucose absorption in both the BBM and BLM. Therefore, the present study further examined the effect of CBE on GLUT2 expression using western blotting. The results showed that GLUT2 membrane protein expression was reduced by CBE, caffeic acid and phloretin (a GLUT2 inhibitor) (Fig. 4C and D). However, caffeine-, chlorogenic acid-, ferulic acid- and phlorizin-treated cells did not exhibit any notable differences (Fig. 4B). Therefore, the results suggested that CBE downregulated glucose transport across the jejunal epithelial cell membrane via GLUT2 inhibition.

To further elucidate the mechanism by which CBE reduced glucose uptake, the effect of CBE on intestinal AMPK activation and expression was further investigated. As shown in Fig. 5, phosphorylation of AMPK, a rate-limiting enzyme decelerating gluconeogenesis, was increased in the CBE-treated group compared with the control group, similar to the caffeic acid-treated group. Neither CBE constituents nor glucose transporter inhibitors activated p-AMPK. Therefore, CBE and caffeic acid exerted a suppressive effect on glucose absorption, primarily via inhibition of SGLT1 expression at the translational level, resulting in modulation of AMPK activity.

The present study further investigated the mechanisms by which CBE inhibited glucose uptake. SREBP-1c and HNF1α serve a critical role in glucose transporter regulation. Therefore, the expression levels of these transcription factors in Caco-2 cells were determined using qPCR. CBE at 100 µg/ml suppressed HNF1α mRNA expression, similar to caffeic acid and phloretin, a GLUT2 inhibitor (Fig. 6). However, there was no effect on SREBP-1c expression. Therefore, these data suggested that CBE and its major component, caffeic acid, had an inhibitory effect on GLUT2 transport function and expression via HNF1α expression.