TuAK was purchased from Zhejiang Lin-An Medicinal Herbs, Co., Ltd., (Zhejiang, China; Batch no. 061208) and was authenticated by Zhejiang Chinese Medical University. A voucher specimen is maintained in the university. 5-Fluorouracil (5-FU) was from Nantong Jinghua Pharmaceutical Group, Co., Ltd., (Nantong, China). Chloroquine (CQ) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Other chemical reagents (grade AR) were from Sigma-Aldrich, Amresco (Solon, OH, USA) or Aladdin Reagents Co., Ltd. (Shanghai, China). The following antibodies were used: survivin (ZSGB-Bio, Beijing, China), Bax (ZSGB-Bio), Bcl-2 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), caspase-9 (Abcam, Cambridge, MA, USA), LC3 (Sigma-Aldrich), Atg7 (Cell Signaling Technology, Danvers, MA, USA) β-actin (Lianke Bio, Hangzhou, China) and GAPDH (Lianke Bio). The secondary antibodies were conjugated to HRP (Jackson Immunolab, West Grove, PA, USA).

Dried TuAK was ground to fine particles, soaked in 8x volume of 95% ethanol overnight, and extracted twice by reverse flow, 2 h each time. The ethanol collected was further extracted by 2x volume of ligarine twice. The ligarine extract was recovered and concentrated by distillation in reduced pressure. The concentrated extract was filtered to obtain the organic solvent-based extract. The ethanol residuum was extracted by dH₂O twice, 10x and 8x volume respectively, and concentrated to obtain the aqueous extract of TuAK. One mega-extraction using 15 kg of TuAK was performed and the extracts obtained were stored at 4°C with desiccants for the following experiments. The organic solvent-based extract of TuAK is hereafter referred to as TuAKe.

SGC-7901, AGS and HEK293 cells were purchased from the Institute of Biochemistry and Cell Biology, Chinese Academy of Science (Shanghai, China). Cells were cultured in Dulbeccos modified Eagles medium (DMEM; Gibco, Carlsbad, CA, USA) containing 10% fetal bovine serum (FBS; Gibco) at 37°C with 5% CO₂. For the viability assays, cells were seeded in 96-well plates, 1×10⁴ cells/well. Desired drugs, compounds and extracts were added to the wells immediately afterwards and incubated with the cells for 24 h. The cell viability was measured by spectrometry (Multiscan MK3 spectrophotometer; Thermo Fisher Scientific, Inc., Waltham, MA, USA) using CCK-8 kit (Lianke Bio) at OD 450 nm.

Cells were seeded in 6-well culture plates and treated by desired drugs or extracts, or by phosphate-buffered saline (PBS) as control for 24 h, then trypsinized and washed by PBS. Aliquots of resuspended cells (5×10⁵/ml) were stained by cell cycle and apoptosis analysis kit (Beckman Coulter, Brea, CA, USA) based on the manufacturers instructions. FCM analysis was performed immediately afterwards (FACSAria™; BD Biosciences, San Jose, CA, USA).

Cells were seeded in 24-well culture plates on glass coverslides, and treated at the desired conditions, then fixed and observed under a fluorescence microscope (IX71-22FL/PH; Olympus, Tokyo, Japan, or Axio A1; Carl Zeiss AG, Oberkochen, Germany).

Cells after desired treatment were scraped, spun down, washed by PBS and resuspended in RIPA buffer with standard protease and phosphatase inhibitors. After quantification of protein concentration by BCA method (Lianke Bio), cell lysates were mixed with loading buffer and subjected to SDS-PAGE with 20 µg protein per lane. The proteins were later transferred to PVDF membranes, which were incubated by primary and secondary antibodies. The membranes were developed with an enhanced chemiluminescence kit (Lianke Bio). Each experimental condition was repeated for at least three times. Quantitative analysis of protein levels determined by band intensities was carried out by ImageJ (v1.48u).

Total RNA was extracted by TRIzol^® reagent (Invitrogen, Carlsbad, CA, USA) and reverse transcripted to cDNA by a commercial kit (Takara Bio, Shiga, Japan). Primers for β-actin, BCL2 and BAX were synthesized by Invitrogen (Shanghai, China) with the following sequences: β-actin forward, (5-ACAACTTTGGTATCGTGGAAGGAC-3) and β-actin reverse, (5-AGGTGGAGGAGTGGGTGTCG-3), BCL2 forward, (5-ATTGGGAAGTTTCAAATCAGC-3), BCL2 reverse, (5-TGCATTCTTGGACGAGGG-3), BAX forward, (5-GACACCTGAGCTGACCTTGG-3) and BAX reverse, (5-GAGGAAGTCCAGTGTCCAGC-3). PCR was performed in three independent experiments with multiple repeats (Applied Biosystems StepOnePlus; Applied Biosystems, Foster City, CA, USA) and the products were analyzed by agarose electrophoresis to compare mRNA levels.

Cells after treatment were scraped in PBS and spun down. The pellets were fixed by 2.5% glutaraldehyde for one week, then washed by 0.1 M PBS and fixed in 1% osmic acid for 1 h, washed by ddH₂O, dehydrated by serial concentrations of ethanol from 50 to 100%. Then the pellets were dehydrated by 100% acetone, permeated by acetone/resin (1:1, v/v) for 2 h and by pure resin overnight. The resin was polymerized by serial heating at 37, 45 and 60°C, sectioned and stained by uranium acetate-lead citrate, and then observed by EM equipped by the companion software (Tecnai; Phillips, Amsterdam, the Netherlands).

Cultured cells were infected with lentivirus expressing shRNA targeting ATG7 gene and control scrambled shRNA (Novobio, Shanghai, China) in the presence of polybrene. Twenty-four hours after infection, knocking-down of Atg7 was confirmed by western blotting, and cells were treated by TuAK for another 24 h, with or without CQ for the last 6 h, followed by cell viability assay.

The institutional ethics committee of Zhejiang Chinese Medical University approved this study and waived the necessity to obtain informed consent based on its nature as a retrospective and observational study. A total of 30 inward gastric cancer patients admitted to the First Affiliated Hospital of Zhejiang Chinese Medical University were grouped into two. Ten patients in the control group received conventional chemotherapy using FAM regimen, namely combination of 5-FU, adriamycin and mitomycin C. Twenty patients in the TuAK group received TuAK-based decoction in addition to the FAM regimen. The evaluated Karnofsky performance status (KPS) scores of all patients were over 60 upon admission to the hospital, and showed no statistically signifcant difference between the two groups. KPS scores of each patient were re-evaluated after 8 weeks of treatment and compared.

Each experiment was performed at least three times with multiple repeats, and the data were analyzed by SPSS (v13.0). The statistical significance was calculated by one-way ANOVA to compare between multiple groups, or by Students t-test when only two groups were compared. P<0.05 was considered statistically significant.

In cultured human gastric cancer cell lines SGC-7901 and AGS, within 24 h of treatment, the inhibitory effect of the aqueous extract of TuAK was trivial compared with the organic extract, TuAKe (Fig. 1A and B). Based on the inhibition curves, in the two cell lines, the IC₅₀ of TuAKe was 35–45 µg/ml, while the aqueous extract had little inhibitory effect. TuAKe, the more potent extract, was used in the following experiments. In long-term experiments, consecutive treatment of 50 µg/ml TuAKe daily killed ~100% SGC-7901 cells and ~80% of AGS cells within 7 days, while lower dose of TuAKe (30 µg/ml) inhibited ~70% of SGC-7901 cells and 50% of AGS cells (Fig. 1C and D).

We analyzed the apoptotic rate in TuAKe-treated SGC-7901 cells. In Fig. 2A, FCM was carried out to detect the apoptotic cells induced by TuAKe after co-staining of Annexin V-FITC and PI. After treatment of 24 h, 5-FU induced significant increase of apoptosis; 25 µg/ml TuAKe increased the apoptotic rate but without statistical significance. After cells were treated by 50 or 100 µg/ml TuAKe, cell apoptosis rates were significantly increased (13.3 and 11.2%, respectively) to similar level as the 5-FU-treated group (12.7%).

Cell cycle distribution was analyzed by FCM in SGC-7901 cells treated by TuAKe for 24 h (Fig. 2B). The percentage of cells in the G0/G1 phase increased from 38.5% in the control (PBS) group to 46.6–63.9% in TuAKe-treated groups (25–100 µg/ml); on the other hand, the percentage of cells in the G2/M phase decreased from 31.5% in the control group to 12.5–17.6% in TuAKe-treated groups. Thus, TuAKe treatment resulted in cell cycle arrest in the G0/G1 phase.

To further analyze the molecular evidence of apoptosis, western blottings for several apoptosis-related markers were carried out. Different concentrations of TuAKe could significantly reduce the level of survivin; whereas, there was significant decrease in Bcl-2 as well as increase in Bax and cleaved caspase-9 levels (Fig. 2C and D). By qRT-PCR, it was further demonstrated that TuAKe resulted in decreased BCL2 and increased BAX gene expression (Fig. 2E).

Autophagy flux analysis was carried out using HEK293 and AGS cells (Fig. 3A). HEK293, a cell line commonly used in studies on autophagy as a tool, was less sensitive to TuAKe compared with the two gastric cancer cell lines used in this study. The IC₅₀ of TuAKe on HEK293 was higher than the maximum dose tested, 75 µg/ml (Fig. 4B), compared with 35–45 µg/ml in tumor cells (Fig. 1A). As seen in Fig. 3A, in both HEK293 and AGS cells, TuAKe increased the level of LC3-II, the conjugate of cytosolic autophagy-related protein LC3-I with phosphatidylethanolamine (PE), an indicator of either autophagy activation or blockage of autophagic degradation. In the autophagy flux analysis, CQ, the lysosome inhibitor, increased the level of LC3-II; combination of CQ and TuAKe further increased the level of LC3-II, indicating the effect of TuAKe was to promote the formation of LC3-II instead of blocking its degradation, thus, demonstrating induction of autophagy by TuAKe.

The RFP-GFP-LC3 tag is a well-defined tool to monitor the dynamic formation and degradation of LC3-II-positive autophagosomes. Presence of yellow puncta (RFP-GFP-LC3-II) implies autophagosomes. Moreover, because GFP is degraded faster than RFP in the lysosomes, presence of red puncta (RFP only) implies autolysosomes under degradation at the late stage. In both HEK293 and AGS cells stably expressing RFP-GFP-LC3-II, the control cells were free of puncta (Fig. 3B and C), while TuAKe induced significant increase in the number of yellow puncta. Increase in the number of observable autophagosomes further supported induction of autophagy by TuAKe demonstrated by flux analysis (Fig. 3A).

Observed by EM (Fig. 3D), TuAKe induced significant increase in the number of intracellular membranous organelles with high electron density, typical of double-membraned autophagosomes (Fig. 3Dc-d). The size of autophagosome-like structures observed under EM was comparable with the LC3-II positive puncta from fluorescent microscopy.

In both HEK293 and AGS cells, after suppression of autophagy by silencing of ATG7 gene, decrease of Atg7 and inhibition of autophagy was confirmed by western blotting (Fig. 4A). After targeted suppression of autophagy, the inhibition rate of TuAKe (10–75 µg/ml) on HEK293 cells was significantly decreased from 5–32% in the control group to 2–10% in the siRNA ATG7 group (Fig. 4B). In AGS cells, decrease of inhibition by TuAKe after ATG7 knockdown was only significant when TuAKe was used in higher doses (40–50 µg/ml); however, the inhibition rates between the two groups was highly significant with the net difference of 30–40%.

We evaluated the KPS scores in 30 gastric cancer patients, 10 received conventional FAM chemotherapeutic regimen and 20 received FAM regimen plus TuAK-based decoction. As shown in Table I, among the 20 patients in the TuAK group, the KPS scores in 10 had improved, 7 were stable and 3 were worsened; among the 10 patients from the FAM group, 1 had improved KPS score, 3 were stable and 6 were worsened. More patients in the TuAK group obtained improved or stable KPS scores (85%) compared with those in the FAM only group (40%), suggesting TuAK-based medication could improve the life quality of gastric cancer patients undergoing chemotherapy.