The anti-tumor activities of Neferine on cell invasion and oxaliplatin sensitivity regulated by EMT via Snail signaling in hepatocellular carcinoma

Tumor invasion and chemotherapy resistance, which are associated with epithelial-mesenchymal transition (EMT), remain as major challenges in hepatocellular carcinoma (HCC) treatment. Neferine, a natural component of Nelumbo nucifera, have been proven the antitumor efficiency in cancer, but the effects of Neferine on HCC invasion and chemosensitivity need to be elucidated. Applying multiple assays of cell proliferation, flow cytometry, immunofluorescence staining, qRT-PCR, Western blot, fluorescence molecular tomography imaging, the influences of Neferine on EMT-regulated viability, apoptosis, invasion, and oxaliplatin (OXA) sensitivity were assessed in HCC cells of HepG2 and Bel-7402, as well as in xenograft animal models in vivo. Here, we reported that Neferine had no obvious effects on HCC cells proliferation, but significantly enhanced cytotoxicity and apoptosis caused by OXA in vitro and in vivo. Through an upregulation of E-cadherin and downregulation of Vimentin, Snail and N-cadherin, Neferine suppressed EMT-induced migration and invasion abilities of HCC cells. TGF-β1 cancelled the effects of Neferine on the migration and invasion of HCC cells. Snail overexpression or TGF-β1-induced EMT attenuated Neferine-mediated OXA sensitization in HCC. Together, our data suggest that Neferine enhances oxaliplatin sensitivity through an inhibition of EMT in HCC cells. Neferine may be used as an OXA sensitizer in HCC chemotherapy.

Our previous study found that HCC cells with EMT phenotype acquired OXA-resistance 8 . Since that, EMT suppression may be synergistic with conventional chemotherapy to improve HCC prognosis.
Neferine (CAS No.: 2292-16-2) is a major bisbenzylisoquinoline alkaloid derived from the seed embryo of lotus (Nelumbo nucifera, a traditional medicinal plant), demonstrating the nontoxic nature of it 9 . Previous studies had shown that Neferine exerted extensive cardio-protective effects, such as anti-hypertensive, anti-arrhythmic 10 , anti-agglutinating, anti-thrombotic 11 , cholinesterase inhibition 12 , as well as its anti-diabetes functions 13 . Importantly, Neferine was suggested to have antitumor effects and sensitized the cancer cells to chemotherapeutic agents. The multidrug resistance was documented to be reversed in human gastric carcinoma cells by Neferine 14 . Neferine inhibited lung cancer cells growth 15,16 and potentiated anti-cancer effect on lung cancer when combined with doxorubicin 17 . Meanwhile, Neferine could inhibit the proliferation of HCC cells 18,19 and osteosarcoma cells 20 . Neferine was also reported to induce autophagy of ovarian cancer cells 21 . However, the effects of Neferine on HCC invasion and chemosensitivity need to be elucidated.
Based on the existed evidence and understanding, the present study aims to prove our hypothesis that Neferine may have some anti-tumor effects on the invasion and chemosensitivity through EMT regulation in HCC cells.

Results
The effects of Neferine on cells viability in HCC and normal hepatic cells. L02, HepG2 and Bel-7402 cells were treated with Neferine at different concentrations for 24 hrs and 48 hrs. As shown in Fig. 1a, Neferine at the dose up to 10 μ M exhibited no significant inhibition of HepG2 cells viability (p = 0.7311). HepG2 cells was a little less sensitive to Neferine than Bel-7402, whose viability slightly decreased without statistical difference at 10 μ M (p = 0.0810). However, Neferine treated at high doses (> 10 μ M) significantly inhibited cells viability in HepG2 and Bel-7402 cells (p < 0.05) in a dose-dependent manner. L02 cells sustained cell viability of over 85% at 80 μ M of Neferine, showing no cytotoxic effects of Neferine on normal liver cells. Therefore, a dose of 10 μ M Neferine nontoxic to HCC cells was chosen for combined treatment with OXA in further experiments.
OXA-inhibited HCC growth was potentiated by Neferine. The dose-dependent viability of HCC cells was firstly measured in the group of OXA treatment alone for 48 hrs to determine its IC 50 , which was 5.99 ± 0.35 μ M for HepG2 and 4.81 ± 0.39 μ M for Bel-7402. OXA at 5 μ M, which was the nearest concentration to the IC 50 values of OXA treatment alone, was then chosen for further co-treatment combined with Neferine.
The cells were subsequently treated with OXA at different concentrations combined with Neferine at 10 μ M for 48 hrs. Neferine significantly improved OXA-induced cytotoxicity in HepG2 and Bel-7402 cells. As compared with the control group without any treatment, 5 μ M OXA induced 45.91 ± 1.86% growth inhibitions in HepG2, which was considerably increased to 62.35 ± 1.35% in co-treatment group of OXA + Neferine (p < 0.01). In Bel-7402 cells, the growth inhibition rate in OXA treatment group was significantly improved in the combined treatment group of 5 μ M OXA and 10 μ M Neferine (47.23 ± 3.27% vs. 64.22 ± 2.86%, p < 0.05, Fig. 1b). Meanwhile, Neferine obviously reduced the IC 50 values of OXA from 5.99 ± 0.35 μ M to 3.15 ± 0.32 μ M for HepG2, from 4.81 ± 0.39 μ M to 2.29 ± 0.20 μ M for Bel-7402 cells (Fig. 1c, p < 0.01, respectively). In colony formation assays, the group of OXA + Neferine had a significantly decreased colony amounts than the OXA treatment group (p < 0.01, Fig. 1d). These results indicated that Neferine significantly increased the chemosensitivity of HCC cells to OXA, implying a synergistic effect between OXA and Neferine. To quantitatively analyze the interactions between Neferine and OXA, an arithmetic method of isobologram indicated a synergistic effect between Neferine and OXA (Fig. 1e).

Establishment of TGF-β1-induced EMT model in HCC cells.
Since EMT played an important role in invasion and chemoresistance of cancer cells, we explored whether EMT and invasion capacity were regulated by Neferine in HCC cells. We found that Neferine alone inhibited cells migration/invasion abilities ( Supplementary Fig. S1, p < 0.01, respectively), and promoted HCC epithelial cells HepG2 and Bel-7402 to more epithelial phenotype by increasing E-cadherine expression and decreasing Vimentin/N-cadherin/Snail expression ( Supplementary Fig. S2). We next established a model of TGF-β 1-induced EMT in HCC cells in time-and dose-dependent manners to further investigate the association between EMT progress and Neferine-promoted chemosensitivity of HCC cell. HepG2 cells and Bel-7402 cells were treated with different concentrations (5 ng/ml, 10 ng/ml, 20 ng/ml) of TGF-β 1 for 24 hrs and 48 hrs. qRT-PCR and Western blot revealed that TGF-β 1 at the dose of 10 ng/ml for 48 hrs induced obvious EMT of HCC cells with decreased expression level of E-cadherin but increased expression level of N-cadherin and Vimentin as compared with the untreated cells (Fig. 3a,b). The expression of Snail, a transcription factor of EMT, rapidly increased at 24 hrs with the administration of TGF-β 1 (Fig. 3a,b). Being stimulated with 10 ng/ml TGF-β 1 for 48 hrs, cells gradually elongated and dispersed. The administration of TGF-β 1 changed cells morphology from pebble-like epithelial to spindle-like mesenchymal Scatter points "S" at the horizontal axis represent the IC 50 of HCC cells when treated with OXA alone and scatter points "R" at vertical axis represent the IC 50 of HCC cells when treated with Neferine alone. Scatter points "Q" represent the IC 50 of OXA when co-treated with Neferine. "Q" points appeared under the straight line the connecting "S" and "R" points, suggesting that synergism effect existed between Neferine and OXA. Nef: Neferine, ** p < 0.01. and pseudopodium stretching after 72 hrs (Fig. 3c). Taken together, 10 ng/ml TGF-β 1 for 48 hrs was chosen to induce HCC EMT in further experiments.
Neferine reversed TGF-β1-induced EMT phenotype in HCC cells. Due to the important roles of EMT in invasion and chemoresistance of cancer cells, we further explored whether EMT was regulated by Neferine in HCC cells. qRT-PCR, Western blot and immunofluorescence staining were respectively applied to detect the expression levels and co-localization of EMT biomarkers in Neferine and/or OXA-treated HCC cells with the administration of 10 ng/ml TGF-β 1. TGF-β 1 induced EMT phenotype as compared to the control groups, as well as the same phenomenon occurred in TGF-β 1 co-treatment with OXA groups. Comparing to the TGF-β 1 groups, Neferine treatment alone and combined with OXA groups resulted in a down-regulated mRNA and protein expression of Vimentin and N-cadherin, concomitant with an up-regulated expression of E-cadherin in TGF-β 1-treated HCC cells. Moreover, mRNA and protein expression of transcription factor Snail, which is critical for EMT induction, was significantly decreased in Neferine treatment and co-treatment with OXA groups. Expression of Slug, another EMT transcription factor, was decreased in Neferine-treated HepG2 but not in Bel-7402 cells. The expression of the other EMT-related transcription factors of Twist and Zeb1 was not significantly regulated by Neferine treatment (Fig. 4a,b). Cellular immunofluorescence labeled by E-cadherin and Vimentin antibodies confirmed the suppression effects of Neferine on TGF-β 1-induced EMT in HepG2 and Bel-7402 cells (Fig. 4c).

Knockdown of Snail inhibited EMT phenotype and migration, but enhanced OXA sensitivity.
To determine the roles of Snail in HCC cells' migration and chemosensitivity, we knocked down the expression of Snail with siRNA in HCC cells and performed further study. The expression changes of EMT markers in HepG2 and Bel-7402 cells were assayed by qRT-PCR and Western blot. Comparing to the si-control groups, Snail knockdown significantly decreased the expression levels of mesenchymal markers of N-cadherin and Vimentin, and increased the expression levels of epithelial marker of E-cadherin (Fig. 5a). Wound healing and transwell assays indicated that knockdown of Snail significantly inhibited HCC cells' migration and invasion abilities as compared to the si-control groups (Fig. 5b,c. p < 0.01, respectively). Si-Snail groups enhanced the sensitivity of HCC cells against OXA as compared with the si-control groups by decreasing the IC 50 values of OXA from 6.844 ± 1.108 μ M to 3.411 ± 0.375 μ M for HepG2 and from 5.339 ± 0.338 μ M to 2.340 ± 0.304 μ M for Bel-7402 (Fig. 5d, p < 0.05, respectively), implying the correlation of Snail and chemosensitivity.

Snail overexpression induced EMT and eliminated migration suppression and OXA sensitivity regulated by Neferine in HCC cells. The above results exhibited the suppression of Snail expression by
Neferine and the association of Snail with migration and chemosensitivity in HCC cells. We further investigated the functions of Snail overexpression on EMT, Neferine-suppressed migration and Neferine-promoted OXA sensitivity in HCC cells. Snail overexpression by Snail plasmid (pcDNA3.1-Snail) transfection induced an obvious EMT progress, which was recognized by decreased E-cadherin expression and increased Vimentin/N-cadherin expression (Supplementary Figure S3a-c). Neferine treatment caused obvious migration and invasion suppression in the blank vector transfection groups (Fig. 6a,b, pcDNA3.1-NC vs. pcDNA3.1-NC + Neferine, p < 0.05, respectively). Consequently, EMT induced by Snail overexpression significantly enhanced the migratory and invasive capabilities of Snail-transfected HCC cells as compared to pcDNA3.1-NC groups (Fig. 6a,b, p < 0.001, respectively). Migratory and invasive capabilities were significantly enhanced in pcDNA3.1-Snail groups which were pre-treated with Neferine for 48 hrs than the control groups transfected with the blank vector (pcD-NA3.1-NC) which were pre-treated with Neferine ( Fig. 6a,b, p < 0.001, respectively), implying that overexpression of Snail eliminated migration suppression by Neferine. Based on the chemoresistance induced by EMT, Snail transfection significantly attenuated OXA sensitization by Neferine in HepG2 and Bel-7402 cells. CCK-8 assays   Supplementary Fig. S5. (c) Representative double indicated that pcDNA3.1-Snail groups had significantly decreased OXA sensitivity than pcDNA3.1-NC groups (Fig. 6c, p < 0.01, respectively).
Neferine increased OXA sensitivity via EMT inhibition in xenograft nude mice. We established subcutaneous xenograft tumor models to verify the effects of Neferine on EMT-regulated OXA sensitivity of HCC in vivo. As shown in Fig. 7a, HepG2 cell-derived tumors at the implantation sites treated with OXA alone were statistically larger than that treated with OXA and Neferine (0.73 ± 0.01 cm 3 vs. 0.12 ± 0.03 cm 3 , p < 0.01). Consistently, Bel-7402 cell-derived tumors in OXA treatment group grew more rapidly than that in OXA/ Neferine co-treatment group (0.64 ± 0.12 cm 3 vs. 0.15 ± 0.05 cm 3 , p < 0.05).

Discussion
Over years, a great deal of efforts has been taken to increase chemosensitivity in HCC patients. However, owing to little and limited understanding of chemoresistant mechanisms, the prognosis of HCC remains poor. New therapeutic strategies which potentiate chemotherapeutic sensitivity safely provide a promising approach for effective HCC treatment. The present study indicated that co-administration of Neferine, a natural component of Nelumbo nucifera, could enhance OXA chemosensitivity in HCC.
Regarding to the anti-tumor effects of Neferine against HCC, Paramasivan et al. 18 reported that Neferine at the dose of 10 μ M exhibited obviously cytotoxicity by inducing reactive oxygen species mediated intrinsic pathway of HepG2 apoptosis. However, we did not find any significant effect of Neferine at the concentration of 10 μ M on the proliferation of both HepG2 and Bel-7402 cells. Consistent to our finding, Yoon 19 provided evidences that Neferine at the concentration of 20 μ M exhibited significant cytotoxicity against Hep3B cells via ER stress and autophagy induction, while 10 μ M Neferine showed no distinct effect. Even at the dose of 20 μ M, Neferine, which was extracted by the researchers themselves, did not affect the growth of Sk-Hep1 cells and primary normal hepatocytes of THLE-3. Different liver (cancer) cells may have diverse tolerance to Neferine. Meanwhile, the purity decided by extraction methods may have significant influences on the stability and antitumor activity of Neferine.
Scientific RepoRts | 7:41616 | DOI: 10.1038/srep41616 sensitivity to epidermal growth factor receptor inhibitors in lung cancer cells, while mesenchymal-like cells were resistance to drug treatment 33 . Meanwhile, Snail suppressed TGF-β -induced apoptosis and was sufficient to trigger EMT in hepatocytes 34 . EMT inhibition could be a useful strategy to cause a loss of anti-apoptotic signal and/or trigger apoptotic responses to sensitize cancer cells to chemotherapy. Accompanying with caspase-3 activation,  Moreover, EMT-generated properties of cancer stem cells (CSCs) are important reasons contributing to chemoresistance in human cancers 35,36 . EMT facilitates the generation of CSCs with the mesenchymal traits which are required for dissemination and chemoresistance 37 . The cells with CSC phenotype (CD44high, CD24low) in breast cancer were found to be resistant to neoadjuvant chemotherapy 38 . Furthermore, Snail was associated with CSCs-like traits acquisition and mediated cell survival in ovarian cancer effectively 39 . PI3K pathway, which was confirmed to be activated in Snail-expressing cells 31 , was directly linked to CSCs expansion and maintenance via promoting the proliferation of CSCs in breast cancer 40 and prostate cancer 41 . Targeting CSCs therapy therapeutically seems to overcome drug resistance. Antibiotic salinomycin could kill breast CSCs preferentially and induced the differentiation of mesenchymal-like cancers in vivo, as assessed by increased E-cadherin expression and decreased vimentin expression 42 . We demonstrated that Snail overexpression induced EMT and eliminated OXA sensitization effects of Neferine in HCC. However, the roles of EMT-elicited HCC CSCs in OXA sensitivity promoted by Neferine needs to be further clarified.
In conclusion, Neferine significantly suppressed EMT so as to inhibit cell mobility but increased OXA sensitivity via Snail signaling in HCC. Our findings suggest that Neferine may be a potent OXA sensitizer in HCC to improve the patients' chemotherapy response. Cell proliferation and clonogenic assay. Cells were seeded on a 96-well plate at a density of 5.0 × 10 3 /well overnight and then were subjected to various concentrations of OXA with or without Neferine for 48 hrs. Cell viability was assayed by Cell Counting Kit-8 (CCK-8, Dojindo Molecular Technologies, Inc., Tokyo, Japan). Briefly, added 10 μ L CCK-8 to each well and the absorbance at 450 nm was measured. The wells absent of drugs were used as the controls. The 50% inhibitory concentration (IC 50 ) was calculated from the survival curves. Each assay was performed in triplicate. An arithmetic method of isobologram was employed to quantitatively analyze the effect of drug synergism 43 .

Methods
For colony formation assays, cells were seeded in 60-mm dishes at a density of 1 × 10 3 /dish. Add OXA and/or Neferine after 24 hrs and then cultured for 2 weeks to form colonies. Fixed the cells with methanol and stained with 0.1% crystal violet for 20 mins.
Apoptosis assay. A FACS Canto II flow cytometer (BD Biosciences, San Jose, CA) was used to quantitate the apoptosis rate by Annexin V-FITC Apoptosis Detection Kit (BD Biosciences). HCC cells in different groups were harvested and suspended in 100 μ L binding buffer. Added 5 μ L of Annexin V-FITC and 5 μ L of propidium iodide (PI) for 15 mins in darkness. 100 μ L binding buffer was added and the cells' apoptosis were detected in 1 hr.

Cell migration and invasion assays.
Wound healing assay. Cells were seeded in 6-well plate for 24 hrs.
The confluent cell monolayers were scratched by a 200 μ L pipette tip straightly. The cells were washed with PBS for 2-3 times and subsequently cultured in fresh medium with 2% FBS for 48 hrs. The wound healing of the scratched cells were photographed under a DMIL LED AE2000 inverted microscope (Leica, Wetzlar, Germany).
Transwell invasion assay. The invasion abilities of HCC cells were assessed using Matrigel-coated upper inserts contained polycarbonate filters in 8-μ m pore size (BD Biosciences). Culture medium containing 10% FBS was placed in the lower chambers to act as a chemoattractant. 4 × 10 4 cells suspended in 200 μ L serum-free DMEM were seeded in the upper chambers and incubated at 37 °C for 48 hrs. The cells that penetrated the Matrigel-coated filter were stained with 0.1% crystal violet hydrate solution.
Xenograft model. To investigate the effects of Neferine on OXA sensitivity in vivo, a xenograft animal model was constructed in male BALB/c mice (4-week old). Briefly, 5 × 10 6 HepG2 and Bel-7402 cells were suspended in 150 μ L serum-free DMEM and injected subcutaneously into the left flank regions. One week later, the mice received an intraperitoneal injection of OXA (0.8 mg/kg/w) with or without Neferine (20 mg/kg/d) for 3 weeks.

Transient transfections of Snail cDNA.
To explore the roles of Snail in Neferine-regulated EMT in HCC cells, 1 × 10 5 cells seeded into 6-well plates were transiently transfected with a pcDNA3.1 expression vector containing a full-length of human Snail sequence (pcDNA3.1-Snail) using Lipofectamine-2000 (Invitrogen) in accordance with the manufacture's protocol. HCC cells were transfected in parallel with the corresponding blank vector as the control (pcDNA3.1-NC). 48 hrs after the transfection, cells with different treatment were collected for qRT-PCR and Western blot analyses.