All cell culture plasticware were obtained from Falcon Labware (Franklin Lakes, NJ, USA) and Techno Plastic Products (Trasadingen, Switzerland). Trypsin, McCoy's 5A, Leibovitz's L-15, RPMI-1640 and DMEM media, and phosphate-buffered saline were obtained from Mediatech, Inc. (Herndon, VA, USA). Penicillin and streptomycin were obtained from Sigma-Aldrich (St. Louis, MO, USA). The MTS assay kit, CellTiter 96 Aqueous Solution Cell Proliferation Assay, was obtained from Promega (Madison, WI, USA). The Annexin V-FITC apoptosis detection kit was obtained from BD Biosciences (Rockville, MD, USA). PI/RNase staining buffer was obtained from BD Biosciences Pharmingen (San Diego, CA, USA). Caspase 3 and 9 ELISA kits were obtained from BioVison (Mountain View, CA, USA). Baicalin and wogonoside were obtained from Indofine Chemical Co. Inc. (Hillsborough Township, NJ, USA). Baicalein and wogonin were obtained from Sigma-Aldrich.

The roots of Scutellaria baicalensis were obtained from Chengde (Hebei, China). The voucher samples were deposited at the Tang Center for Herbal Medicine Research at The University of Chicago. Dried S. baicalensis roots were ground to powder, and the powdered roots were extracted with 70% ethanol for 2 h. The extraction method was boiling under reflux. The filtrate was collected and the extraction procedure was repeated one more time on the residue. The combined filtrate was condensed under vacuum and lyophilized to yield dried S. baicalensis extract (SbE).

Fecal samples were obtained from five adult volunteers, who were non-smokers and had not consumed antibiotics for ≥3 months before the study. The samples were collected by the donors in plastic cups, and were processed within 30 min of passage. All five fecal samples were mixed and an aliquot of 5 g of the mixed feces was homogenized with 20 ml of phosphate buffer (pH 7.0) to obtain a fecal slurry. The slurry was filtered through muslin to remove particulate material. One microliter of the fecal slurry was mixed with 4 ml anaerobic medium containing 2.5 mg of SbE. They were anaerobically incubated at 37°C for 0, 2 or 8 h. Then, 1 ml of reaction mixture was extracted three times with 400 μl n-butanol/each time. The combined n-butanol solution was dried under nitrogen steam spray in a water bath (60°C). Then the residue was dissolved in methanol. The methanol solution was centrifuged at 17,000 × g for 10 min before HPLC analysis.

The HPLC system was a Waters 2960 instrument (Milford, MA, USA) with a quaternary pump, an automatic injector, a photodiode array detector (Model 996), and Waters Empower software for peak identification and integration. The separations were carried out on a Phenomenex Prodigy ODS(2) column (150×2.0 mm, 5 μm). A binary gradient solvent system of acetonitrile (eluent A) −0.03% (v/v) phosphoric acid in water (eluent B) was used as follows: 13% A and 87% B (0 min), 28% A and 72% B (17 min), 35% A and 65% B (27 min), 90% A and 10% B (30–31 min), 13% A and 87% B (34–39 min). The flow-rate of 0.8 ml/min was used and absorbance was detected at 280 nm. All tested solutions were filtered through Millex 0.2-μm nylon membrane syringe filters before use. The contents of the constituents were calculated using standard curves of flavonoids.

The human colorectal cancer cell lines HCT-116 (McCoy's 5A), SW-480 (Leibovitz's L-15), HT-29 (McCoy's 5A), NSCLC non-small cell lung cancer cells (DMEM), and human breast cancer cell lines MCF-7 (RPMI-1640), MDA-MB-231 (RPMI-1640) were obtained from American Type Culture Collection (Manassas, VA, USA). The cells were grown in the indicated medium supplemented with 10% FBS and 50 IU penicillin/streptomycin in a humidified atmosphere with 5% CO₂ at 37°C.

Baicalin and baicalein were dissolved in DMSO and were stored at −20°C before use. Cells were seeded in 96-well plates (1×10⁴ cells/well). After 24 h, indicated concentrations of drugs were added to the wells. The final concentration of DMSO was 1%. Controls were exposed to culture medium containing 1% DMSO without drugs. All experiments were performed in triplicate and repeated 3 times. Following the indicated incubation period, cell proliferation was evaluated using an MTS assay according to the manufacturer's instructions. Briefly, the medium was replaced with 100 μl of fresh medium and 20 μl of MTS reagent (CellTiter 96 Aqueous Solution) in each well, and the plate was returned to the incubator for 1–2 h. A 60-μl aliquot of medium from each well was transferred to an ELISA 96-well plate and its absorbance at 490 nm was recorded. Since 1% DMSO did not influence the proliferation of the two cell lines, results were expressed as percent of control (DMSO vehicle set at 100%).

HCT-116 cells were seeded in 24-well tissue culture plates. On the second day, the medium was changed and cells were treated with test compounds. Cells were incubated for 48 h before they were harvested. These cells were fixed gently with 80% ethanol in a freezer for 2 h and were then treated with 0.25% Triton X-100 for 5 min on an ice bath. Cells were resuspended in 300 μl of PBS containing 40 μg/ml propidium iodide (PI) and 0.1 mg/ml RNase (28). Then the cells were incubated in the dark for 20 min at room temperature, and cell cycle analysis was performed using a FACScan flow cytometer (Becton-Dickinson, Mountain View, CA, USA) and FlowJo 7.1.0 software (Tree Star, Ashland, OR, USA). For each measurement, ≥10,000 cells were counted.

The apoptosis assay was performed by flow cytometry following a previously described procedure (29). Briefly, HCT-116 cells were seeded in 24-well tissue culture plates. After culturing for 1 day, the medium was changed and test compounds were added. After treatment for 48 h, cells floating in the medium were collected. The adherent cells were detached with trypsin. Then, culture medium containing 10% FBS (and floating cells) was added to inactivate trypsin. After gentle pipetting, the cells were centrifuged for 5 min at 1,500 g. The supernatant was removed and cells were stained with Annexin V-fluorescein isothiocyanate (FITC) and propidium iodide (PI) according to the manufacturer's instructions. Untreated cells served as controls. The cells were analyzed immediately after staining using a flow cytometer. For each measurement, ≥20,000 cells were counted.

HCT-116 cells were seeded in 6-well tissue culture plates. After 24 h, the medium was changed and baicalein was added. After treatment for 24 h, cell lysates were collected. Expression levels of caspase 3 and 9 were determined by the colorimetric method according to the manufacturer's instructions. Briefly, cell lysates were diluted with 50 μl of 2X reaction buffer (containing 10 mM DTT) to a protein concentration of 0.5 mg/ml in an ELISA 96-well plate. Then, 5 μl of colorimetric tetrapeptide substrate (DEVD-pNA for caspase 3 and LEHDpNA for caspase 9) and cell lysate were added, and plate was incubated at 37°C for 24 h. Then, the absorbance was recorded at 405 nm. The change in caspase activity was calculated as absorbance of baicalein treated cells/absorbance of untreated controls.

The possible binding modes of baicalein at the catalytic domains of human caspase 3 and caspase 9 were predicted using the docking program Surflex-Dock (Tripos, St. Louis, MO, USA). The structure of baicalein was generated (through Ligand model in Sybyl), and protein crystal structures were obtained (PDB code 3H0E for caspase 3 and 2AR9 for caspase 9). To prepare for docking analysis, the protein structures were prepared by adding hydrogen atoms and missing sidechain atoms and removing water molecules. Intermolecular interaction between baicalein and caspases were analyzed, and the key pharmacophore in the ligand was identified (30,31).

Data are presented as mean ± standard error (SE) (n=3). A one-way ANOVA was employed to determine the statistical significance of the results. When necessary, a Student's t-test was used to compare the two groups. The level of statistical significance was set at P<0.05.

HPLC analysis was used to monitor SbE flavonoid changes during the biotransformation of human enteric microbiome. As shown in Fig. 1, four flavonoids were detected in SbE, i.e., baicalin, wogonoside, baicalein and wogonin. Baicalin is the major constituent in SbE. To determine whether fecal compounds influence SbE flavonoid analysis, we assayed the vehicle fecal sample. A major peak at retention time (Rt) of 9.061 min was observed in the chromatogram of the fecal sample. For the SbE, the closest flavonoid peak to this fecal peak is baicalin, with an Rt of 11.366 min. This fecal peak was separated from the baicalin peak at baseline. Thus, compounds from fecal sample did not influence SbE flavonoid determination.

When SbE was cultured with human enteric microbiome for 2 h, compared to un-transformed SbE, the baicalin peak was significantly reduced. Data showed that 75.3% of baicalin in SbE was converted to baicalein after being cultured for 2 h. Furthermore, after being cultured for 8 h, baicalin was not detected, while baicalein was the major constituent in the reaction mixture, indicating that all baicalin in SbE was converted to baicalein. Similar results were also observed in wogonoside. Our data showed that the gut microbiome can quickly transform baicalin to baicalein.

In this study, six human cancer cell lines from three of the most common human cancers were used, including colorectal cancer (HCT-116, SW-480 and HT-29), non-small cell lung cancer (NSCLC), and breast cancer (MCF-7 and MDA-MB-231).

As shown in Fig. 2A–C, while 48-h treatment with baicalin inhibited cancer cell growth in relatively high concentrations, baicalein caused much stronger growth suppression in all three colorectal cancer cell lines. At 60 μM, only baicalin showed 7.6±1.7% of an antiproliferative effect on HCT-116 cells (P<0.05 vs. control), while no significant effects were observed on the other two cancer cell lines. In the same concentration (60 μM), baicalein inhibited cancer cell growth by 73.7±3.4% in HCT-116, 40.3±2.9% in SW-480, and 22.3±1.8% in HT-29 cells, respectively (all P<0.01 vs. control). Among the three cell lines, baicalein showed the most potent antiproliferative effects in HCT-116 cells with an IC₅₀ value of 40.1 μM.

Baicalein also showed significant antiproliferative effects on the other three cancer cell lines, but NSCLC cells were not sensitive to baicalein treatment. Compared to baicalein, baicalin showed weaker antiproliferative effects on NSCLC and MDA-MB-231 cells (Fig. 2D and F). MCF-7 cells were the only exception. Although baicalein showed significant effects on MCF-7 cells at concentrations even as low as 20 μM, when treatment concentration was increased to 100 μM, baicalin still did not inhibit cancer cell growth (Fig. 2E).

Our data suggested that baicalin showed less antiproliferative effects in the tested concentrations. Its enteric microbiome metabolite, baicalein, showed very significant antiproliferative effects in different human cancer cell lines, especially HCT-116 cells (Fig. 2).

To confirm the in vitro antiproliferative effect of baicalein on HCT-116 colorectal cancer cells, the in vivo antitumor activities of baicalin and baicalein were evaluated. Firefly luciferase-tagged HCT-116 cells were inoculated into the flanks of athymic nude mice. Beginning on day 1, animals were also administered with baicalin or baicalein at 30 mg/kg or vehicle intraperitoneally every other day. Tumor growth was measured by xenogeny bioluminescence imaging on a weekly basis. Representative xenogen imaging results at weeks 0–4 are shown in Fig. 3A. Tumor size at indicated time-points as assessed by imaging signal intensities is summarized in Fig. 3B. The data showed that at weeks 2 and 3, baicalin suppressed tumor growth, but there is no significant differences compared to control. At week 4, baicalin significantly inhibited tumor growth (P<0.05). For the baicalein, at week 2, baicalein exhibited significantly decreased xenogeny imaging signal intensities compared with those of the control (P<0.05). Weeks 3 and 4 exhibited more significant antitumor effects than week 2 (both P<0.01). Our data suggested the metabolite baicalein showed more significant antitumor effects than those of its parent compound baicalin. The enteric microbiome metabolism plays an important role in enhancing the anti-cancer activity of SbE.

Since HCT-116 colorectal cancer cells were sensitive to baicalein treatment both in vitro and in vivo, we selected this cell line for further mechanistic evaluations. As shown in Fig. 4, compared to the control, the effects of baicalein on the cell cycle profile were observed at concentrations as low as 20 μM. Treatment of HCT-116 cells with 20, 40 and 60 μM baicalein for 48 h decreased G1 phase to 25.0, 19.0 and 26.7%, respectively, compared to 41.8% in vehicle treated cells, while increasing S phase to 49.7, 60.7 and 73.6%, respectively, compared to 43.3% in vehicle treated cells (all P<0.01). Thus, baicalein significantly decreased the number of cancer cells in G1 phase and increased the number of cancer cells in S phase. On the other hand, baicalin treatment did not influence the cell cycle at 20 and 40 μM. Treatment with 60 μM of baicalin for 48 h changed cell proportions by increasing G1 phase and decreasing G2/M phase (P<0.05 vs control). These results suggested that the metabolite baicalein, not parent compound baicalin, significantly induced S phase cell cycle arrest in HCT-116 cells.

The apoptotic effects of baicalin and baicalein were evaluated by flow cytometry after staining with Annexin V and PI. Annexin V can be detected in both early and late stages of apoptosis, whereas PI stained cells only in late apoptosis or necrosis. Early apoptotic cells were positive for Annexin V and negative for PI (lower right quadrant); late apoptotic cells stained for both Annexin V and PI (upper right quadrant). As shown in Fig. 5, following treatment with 20, 40, 60 and 80 μM of baicalein for 48 h, compared to the control (6.7%), the percentage of early apoptotic SW-480 cells was increased to 11.2, 12.5, 28.8 and 37.5%, respectively (P<0.05, P<0.01, P<0.01 and P<0.01). However, baicalin did not induce apoptosis at the concentrations of 20–60 μM. Only 80 μM of baicalin showed an antiproliferative effect, which induced 10.5% of early apoptotic cells (P<0.05). These data demonstrate that baicalein significantly induces cell apoptosis.

Caspase 3 and caspase 9 are two key proteins of the caspase family of proteases, which are highly conserved in multicellular organisms and function as central regulators of apoptosis. They have been identified as playing a key role in the progression of apoptosis (32). To characterize the potential mechanism of baicalein's anticancer activity, we assayed the activities of two caspases since baicalein increased cancer cell apoptosis. As shown in Fig. 6A, treatment of HCT-116 cells with 40 μM baicalein for 24 h upregulated caspase 3 and 9 activities significantly. These activities were further enhanced with 60 μM of baicalein, increasing caspase 3 and 9 activities to 103.7±18.5 and 95.4±16.4% above those of vehicle treated cells, respectively (both P<0.01). Our results suggested that baicalein significantly induced the expression of caspase 3 and 9.

To further explore the potential effects of baicalein on caspase 3 and 9, we performed docking analysis to characterize the physical interactions of baicalein with these caspases. We examined baicalein docking for human caspase 3 (PDB code: 3H0E) and human caspase 9 (PDB code: 2AR9). The Surflex-Dock program was used to predict the binding sites of baicalein to caspase 3 and 9. The energetically most favorable positions for baicalein interaction with these caspases are shown in Fig. 6B. The in silico modeling suggested that baicalein forms hydrogen bonds with residues Ser251 and Asp253 at the active site of caspase 3, while baicalein forms hydrogen bonds with residues Leu227 and Asp228 in caspase 9 through its hydroxyl groups. In addition, baicalein is predicted to show significant binding affinity for caspase 3 (CScore 3.76) and caspase 9 (CScore 3.18), suggesting that baicalein may directly interact with these caspases.