Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide with a high incidence and mortality [1,2]. Besides surgical management, chemotherapy is another important therapeutic option for HCC, but it would cause severe toxicities which are commonly difficult to tolerate by HCC patients [3,4]; for instance, fluorouracil and doxorubicin, two drugs currently used for treating HCC, are not satisfactory since they can produce severe side effects and are not too effective [5]. Thus, searching for new candidate agents with better effectiveness and lower toxicity for treating HCC have received more and more attention.

Traditional Chinese Medicines (TCMs) have a long history in the treatment of various cancers, and they were reported to have low cost and less side effects [6,7]. Senecio scandens Buch,-Ham. (Asteraceae), known as “Qianliguang” in Chinese, is one of the most popular species used as a Chinese folk medicine, and it is mainly distributed in southwest China [8]. It is used for the treatment for various ailments, such as enteritis, bacterial diarrhea, respiratory tract infections, and conjunctivitis in China [9]. However, there is no report on the anti-tumor effect of S. scandens and its related active components. In the present study, the antitumor effects of polysaccharides extracted from the root of S. scandens (PRS) against HCC were investigated, which could lead to further development of PRS as a potential drug for treating HCC.

Chemicals and reagents

Fetal bovine serum (FBS) and Roswell park memorial institute (RPMI)-1640 culture medium were purchased from Thermo Fisher Scientific (Waltham, MA USA); 5-Fluorouracil (5-FU) and MTT were obtained from Sigma-Aldrich Inc. (St. Louis, MO, USA); primary antibodies of caspase-3, caspase-9, Bcl-2 and Bax were from R&D Systems, Ltd. (Minneapolis, USA). All the other chemicals and reagents were of analytical grade.

Extraction of polysaccharides from S. scandens

The dried whole plant of S. scandens was powdered, and defatted with 95 % ethanol for 24 h. The pretreated powder of S. scandens (50 g) were extracted by refluxing with distilled water at selected extraction times (2, 3 and 4 h), number of extraction (1, 2 and 3), and ratio of water to raw material (10, 15 and 20 mL/g). After filtration, the solution were concentrated and precipitated with ethanol to a final concentration of 80 % (v/v) overnight at 4 °C. After centrifugation (8000 rpm for 10 min), the precipitates were collected, and then washed respectively with 95 % ethanol, anhydrous ethanol and acetone. The crude polysaccharides were dried, and the percentage yield (%) was calculated as the following formula: PRS extraction yield (%) = Y₀/Y, where Y₀ (g) is the weight of dried crude polysaccharides (PRS); Y (g) is the weight of dried powder of S. scandens.

Experimental design of response surface methodology (RSM)

RSM with a Box-Behnken design (BBD, three levels and three factors) was applied to determine the effect of extraction time, number of extractions, and ratio of water to raw material on the extraction yield of PRS. Based on the results of single factor experiment, the experimental range of each factor was determined. The experimental arrangement is shown in .

Cell culture

A549, HL60, S180 and H22 cell lines were obtained from the Type Culture Collection of Chinese Academy of Sciences (Shanghai, China). All cells were maintained in RPMI medium supplemented with 10 % FBS, 100U/mL penicillin and 100 μg/mL streptomycin. The cell lines were cultured in a cell incubator (Thermo Scientific, Waltham, MA, USA) at 37 °C in 5 % CO₂/95 % air.

Animals

BALB/c mice (18-22 g) were obtained from the Experimental Animal Center of the First Affiliated Hospital of Chinese PLA General Hospital (Beijing, China) and acclimatized for 7 days before they were used for the experiment. The mice were housed under controlled environment (23 ± 1 °C and a 12-h light/12-h dark cycle) with free access to food and water. Research and animal experimental procedures were strictly in accordance with “Principles of Laboratory Animal Care” (NIH publication no. 85-23, revised 1985) and approved by the Animal Care and Use Committee of the First Affiliated Hospital of Chinese PLA General Hospital (approval no. A 90-12-2015t09).

Determination of anti-tumor effect of PRS in vitro

A549, HL60, S180 and H22 cell proliferations were determined by using MTT assay as previously described [10]. Briefly, cells (2.0 × 10⁴/0.2 mL) were seeded into 96-well plates and allowed to attach for 24 h. A series of concentrations of PRS (10, 20, 40, 80, 160, 320 and 640 μg/mL) were added and incubated for 48 h. Subsequently, 5 mg/mL MTT (10 μL) solutions were added. After incubating at 37 ^oC for 4 h, the supernatant was discarded and dimethyl sulfoxide (100 μL) was added to each well. The plates were read at 570 nm with a Mode 680 96-well microplate reader (Bio-Rad, Tokyo, Japan). Inhibition rate (H) was calculated as in Eq 1.

H (%) = {(A₀ – A_x)/A₀}100 …………. (1)

where A₀ is the absorbance of controls and A_x the absorbance of cells treated with PRS.

Determination of anti-tumor effect of PRS in vivo

The diluted tumor cells were subcutaneously inoculated (0.2 mL, 1 × 10⁶ cells/mouse) into the right armpits of the mice to establish the animal model. After 24 h, mice were divided randomly into five groups. PRS was administered orally (ig) to each group at different dosages (20, 40, 80 mg/kg). The positive control group was treated with 5-FU at a dosage of 5 mg/kg intraperitoneally (ip), and the control group was treated with normal saline (ig). The weights of the mice were recorded before drug administration which lasted for 14 days. On the 15th day, all the mice were sacrificed and the segregated tumor weights were recorded.

Western blot assay

Western blot assay was used to investigate the apoptotic mechanisms of PRS on H22 cells [11]. Different concentrations of PRS (20, 40 and 80 μg/mL) were added to the H22 cells and incubated for 48 h. Then, the cells were harvested and total proteins were extracted. Total protein sample (40 μg) was separated using SDS/PAGE. The separated proteins were transferred onto polyvinylidene difluoride (PVDF) membrane, and subsequently subjected to immunoblotting analysis with corresponding primary antibodies and horseradish peroxidase-conjugated secondary antibodies (Beyotime Biotech, Haimen, China). After that, chemiluminescence reagents (Beyotime Biotech, Haimen, China) were added for visualization of the protein bands. Antibodies against β-actin was used to normalize for protein loading.

Statistical analysis

The data are expressed as mean ± standard deviation (SD) and were analyzed by ANOVA. Significance of differences between groups was evaluated using Student’s t-test. RSM was analyzed with Design Expert Version 8.0.5 software (State-Ease, USA). P < 0.05 was regarded as statistically significant.

Model fitting and statistical analysis

As shown in , independent variables (extraction time, number of extraction, and ratio of water to raw material) on the extraction yield of PRS were examined by BBD design (). Multiple regression analysis was performed, and the predicted the yield (Y) of PRS could be obtained by the second-order polynomial equation as in Eq 2.

Y = 6.71 + 0.091 A + 0.61 B + 0.27 C – 0.005 AB – 0.033 AC – 0.16 BC – 0.72 A² – 1.05 B² – 0.48 C² …………………….. (2)

The analysis results of ANOVA are shown in . The “lack of fit F-value” (3.01) was not significant due to the relative pure error. Based on the validity analysis of the three variables, it showed that the independent variables of B and C, the interaction term of B × C, and the two quadratic terms of A², B² and C² significantly affected the yield of PRS (p < 0.001). The R² and R²_Adj values (0.9890 and 0.9750, respectively) showed a high degree correlation between the observed values and the predicted values. In addition, the statistical analysis also showed a very low p-value (< 0.0001) and a very high F value (585.70), indicating the established model was highly significant.

Optimized PRS extraction conditions and validated models

By using the Design-Expert software, 3D graphic surface for the effects of three independent variables are shown in . From the model, the optimal conditions for obtaining the highest PRS yield were: extraction time 3.06 h, number of extractions 2.27, and ratio of water to raw material 16.17 mL/g. The maximum yield of PRS predicted by the model was 6.83 %. To validate the accuracy of the model, the optimal condition was modified as follows: extraction time 2.11 h, number of extractions 2, and ratio of water to raw material 16.17 mL/g. A verification experiment was carried out under the modified conditions. As a result, the mean yield of PRS obtained from actual experiments was 6.85 ± 0.12 %. The results indicated that the RSM was effective for optimizing the extraction conditions for PRS.

Effect of PRS on cells growth in vitro

The cytotoxic effect of PRS on A549, S180, HL60 and H22 cell lines were evaluated by MTT assay. As a result, PRS significantly inhibited the cell proliferations of H22, A549, S180 and HL60 cell lines, and the IC₅₀ values were shown in . PRS showed higher inhibitory effect against the proliferation of H22 cells (IC₅₀ value of 42.4 μg/mL ) than the other cell lines with the highest inhibition ratio of 73.3% at 320 μg/mL.

In vivo anti-tumor effect of PRS

The in vivo antitumor effect of PRS on transplanted H22 tumors in mice was showed in . The results indicated that oral administration of PRS (20, 40 and 80 mg/kg; p < 0.05, p < 0.05 and p < 0.01, respectively) could significantly reduce the tumor weight in a dose-dependent manner. Also, the body weights of the mice treated with PRS were higher than that of control, especially at the doses of 40 and 80 mg/kg (p < 0.05, p < 0.05). The 5-FU could also significantly (p < 0.01) reduce the tumor weight, but the body weight of the mice was sharply decreased in this group (p < 0.01).

Effect of PRS on expressions of Bcl-2, Bax, caspase-3 and caspase-9

To study the pharmacological mechanisms for the anti-tumor effects of PRS, the effects of PRS on apoptosis related proteins including Bcl-2, Bax, caspase-3 and caspase-9 were investigated in H22 cells. As shown in , PRS treatment (20, 40 and 80 μg/mL) significantly up-regulated the expression of Bax, and down-regulated Bcl-2. Furthermore, similar to Bax, caspase-3 and caspase-9 were up-regulated following treatment with PRS (20, 40 and 80 μg/mL), compared with the control ().

It is well known that RSM is an effective method for optimizing the experimental conditions of complex processes [14].  The main advantage of RSM technique over other statistical techniques is that it can evaluate multiple parameters and their interactions, but the number of experiments needed is reduced [15,16]. Many researches have showed that RSM is an appropriate method used to optimize polysaccharide extraction processes [17,18]. RSM applied with a three-level three-factor BBD was employed in the present study, and the optimal conditions for PRS were as follows, extraction time, 2.11 h, number of extractions 2, and ratio of water to raw material, 16.17 mL/g.

There is increasing evidence showed that natural products and many antitumor chemotherapeutic agents can induce apoptosis of cancer cells [19]. It has also been indicated that the ratio of Bax/Bcl-2 is a key factor in the regulation of the apoptotic process, and a high ratio of Bax/Bcl-2 might promote apoptosis activity [19,20]. Furthermore, caspase-9 and caspase-3 are important activators in mitochondria-mediated apoptosis pathway, and caspase-3 can be activated by caspase-9 [21]. Thus, after treating with PRS, the expression of caspase-9 and caspase-3 proteins were up-regulated and Bax/Bcl-2 ratio increased in the H22 cells, indicating that PRS could induce mitochondrial-mediated apoptosis.

The findings of the present study indicate that PRS possesses significant anti-tumor effect on H22 tumor in vitro and in vivo, and that the mechanism of the anti-tumor effect is closely related to the induction of mitochondria-mediated apoptosis.