Putting the dried flowers of Marigold into the Chinese medicine crusher for 1 min, and pass the obtained powder of Marigold flowers through a standard mesh sieve of 60 mesh to obtain 60 mesh Marigold powder. Weighing 10.0 g of 60 mesh Marigold powder and 200 mL of distilled water, mix thoroughly, maintaining for 2 h at 100˚C in a water bath, extracting three times, combining the three filtrates and filter. The filtrate was spin-distilled at 45˚C under RE2000 Series Rotary Evaporator (Shanghai Yarong Biochemical Instrument Factory), until about 1 mL of spin-distilled liquid was in the spin-distillation bottle. The liquid was transferred to a Petri dish with a pasteurized pipette, wrapped in cling film and frozen in Scientz-ND Series Vacuum Freeze Dryer (Ningbo Xinzhi Biotechnology Co., Ltd.) at −20˚C for 7 h, then put into a freeze-dryer for 12 h to obtain a freeze-dried powder of water-soluble ingredients of Marigold, which was stored at 4˚C for alternate use.

Weighing 100 mg of lyophilized powder of Marigold water-soluble ingredients, dissolve in 10 mL of distilled water and prepare 10⁴ mg/L of application solution, store at 4˚C for alternate use.

A pancreatic beta-cell line established from transgenic mice, named NIT-1 cells, purchased from Guangzhou Jennio Biotech Co., Ltd. NIT-1 cells were cultured in 1640 medium (Life Technologies/Gibco, Grand Island, NY, USA) containing 10% fetal bovine serum (Biopike Biotechnology Co., Ltd., Beijing, China) and 100 unit/mL penicillin and 100 mg/L streptomycin in a cell incubator at 37˚C and saturated humidity with 5% CO₂. Cells were passaged once when they had grown to cover about 90% of the wall area of the culture flask.

The cell viability was detected with a CCK-8 kit (Nanjing Jiancheng Institute of Biological Engineering). NIT-1 cells were seeded into in 96-well plates at density of 1 × 10⁵ cells/well, and cultured for 24 h at 37˚C in 5% CO₂ incubator. Cells were treated with diferent concentrations (0, 0.4, 0.8, 1.6, 3.2, 6.4, 12.8 mg/L) of ATO (Arsenic trioxide (As₂O₃, Beijing Tanmo Quality Control Technology Co., Ltd.) and different concentrations (0, 40, 80, 160, 320, 640, 1280 mg/L) of ME for 24 h. Then, the toxic liquid was discarded and the cells were incubated in a medium mixed with 10 µL CCK8 solution and 90 µL 1640 at 37˚C, 5% CO₂, and 90% humidity for 4 h. The absorbance was subsequently measured at 450 nm with a multifunctional fuorescent plate instrument (SP-Max3500FL, Flash Biotechnology Co., Ltd., Shanghai, China). Cell survival rate was calculated according to cell survival rate = [(experimental hole − blank hole)/(negative control hole − blank hole)] × 100%.

Hoechst 33,258 staining (Beyotime Institute of Biotechnology, Jiangsu, China) was used to evaluate the morphological changes of the cells undergoing apoptosis and detect apoptosis rate. NIT-1 cells were seeded into in 6-well plates at 1 × 10⁶ cells/well, overnight cultured and after the designed treatment, cells were fixed and stained with Hoechst 33,258 solution according to the manufacturer instructions. Cells were then stimulated by ultraviolet light under a fluorescence microscope and photographed in a random 3 - 5 field of view under the microscope. Apoptotic cells were defined by the condensation of nuclear chromatin, fragmentation, or margination to the nuclear membrane. We counted 200 cells under the microscope and recorded the number of apoptotic cells. Apoptosis rate (%) was equal to apoptotic cell number/200 × 100%.

The intracellular ROS was detected using a reactive oxygen detection kit (Shanghai Beyotime Biotechnology Co., Ltd). According to the manufacturer’s instruction, NIT-1 Cells were seeded into in 6-well plates at 1 × 10⁶ cells/well, overnight cultured and after the ATO and ME treatment, adding 1000 μL of culture medium containing DCFH-DA probe with concentration of 10 μmol/L to each well and incubate for 20 min in the dark at 37˚C. Afterward, cells were washed twice with cold PBS followed by observing and photographing in a random 3 - 5 field of view under a fluorescence microscope (Olympus Corp, Tokyo, Japan). Average ROS fluorescence intensity = total optical density/cell area.

Total SOD activity was determined with a SOD assay Kit (Shanghai Beyotime Biotechnology Co., Ltd). NIT-1 cells were seeded into in 6-well plates at 1 × 10⁶ cells/well, overnight cultures, followed by the prescribed treatment, sample loading, incubation and detection were carried out according to the steps in the manufacturer. The activities of SOD were standardized by protein concentrations, which were determined by the BCA Protein Assay Kit (Shanghai Beyotime Biotechnology Co., Ltd). The total SOD activity was normalized by the protein concentrations.

MDA was tested with MDA assay Kit (Shanghai Beyotime Biotechnology Co., Ltd). NIT-1 Cells were seeded into in 6-well plates at 1 × 10⁶ cells/well, after an overnight culture and the prescribed course of therapy, cells were harvested and lysed in lysis bufer (Shanghai Beyotime Biotechnology Co., Ltd) for 10 min at 4˚C. Afterward, cell suspension was centrifuged at 13,000 rpm for 10 min and cell supernatants were collected for detecting the content of MDA. Following the manufacturer’s instructions for sample loading, incubation, and detection, the MDA content has been standardized by protein concentrations, which were determined by the BCA Protein Assay Kit (Shanghai Beyotime Biotechnology Co., Ltd). The content of MDA was normalized by the protein concentrations.

Total RNAs were isolated using a total RNA extraction kit (Tiangen Biotech, Beijing Co., Ltd, China). The quality and concentration of total RNAs were evaluated by ultra-micro ultraviolet-visible spectrophotometer (Hangzhou Suizhen Biotechnology Co., Ltd). For specific RNA sample, ratio of A₂₆₀/A₂₃₀ and A₂₆₀/A₂₈₀ both between 1.8 and 2.2 has been considered qualified and can be used for subsequent experiments. Qualified RNA sample was then reversed transcription to obtain the DNA sample, closely followed by 200 ng cDNA template for PCR amplification. The reaction conditions were as follow: 95˚C for 15 min 1 cycle (pre-denaturation), 95˚C for 10 s 40 cycles (denaturation), 60˚C for 32 s 40 cycles (annealing/extension). All the primers for mRNAs were synthesized by Shanghai Generay biological engineering Co., Ltd. (Shanghai, China), and the sequences of each primer were listed in Table 1. The glyceraldehyde-3-phosphate dehydrogenase gene (GAPDH) was used as the internal control and levels of mRNA were calculated by the equation 2^−ΔΔCt (ΔCt = Ct_mRNA − Ct_GAPDH, ΔΔCt = ΔCt_Treatment − ΔCt_Control).

The protein expression of apoptosis-related proteins was detected by Western Blot analysis, the protein concentrations were determined by the BCA Protein Assay Kit (Shanghai Beyotime Biotechnology Co., Ltd). Subsequently, equal

amounts of protein (60 μg) separated by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore, MA, USA). The membrane was then blocked with 5% non-fat milk for 2 h and incubated with primary antibodies against HO-1 and Nrf2 (1:500) (Beijing solarbio Technology Co., Ltd.,) and GAPDH (1:1000) antibody (Shanghai Jereh Bioengineering Co., Ltd.) overnight at 4˚C. The membranes were further incubated with horseradish-peroxidase-conjugated secondary antibodies (ZSGB Bio, Beijing, China, 1:6000) at room temperature for 1 h. The immuno-reactivity zone was observed with enhanced chemiluminescence reagent (Millipore, temecula, CA) and displayed with JS-M6 chemiluminescence imaging system. Image J software was utilized to analyze the average optical density of protein bands and GAPDH was used as the internal reference to correct the grayscale value of the target band protein. The relative expression level of the protein was equal to the grayscale value of the measured band/the grayscale value of the internal reference. The protein expression of the control group was used to correct the results.

All experiments were repeated three times independently, and the results were presented as mean ± standard deviation (SD). Differences among groups were performed by using one-way analysis of variance (ANOVA) analysis. Student-NewmanKeuls (SNK) test was used to compare the means of two independent groups. Non-parametric Kruskal-Wallis test was used when the original data were under variance heterogeneity. Statistical significance was set as P < 0.05. All statistical analysis were performed using the Statistical Program for Social Sciences (SPSS), version 20.0 (SPSS, Inc., Chicago, IL, USA).

As shown in Figure 1(a), after treatment with 0.4 mg/L ATO, the cell viability decreased in a dose-dependent way. In Figure 1(b) ME at 40 mg/L and 80 mg/L increased cell viability, from 100% to 105.12% and 111.64%, respectively, and the concentrations of ME above 80 mg/L decreased cell viability, as well as in a dose-dependent way. Notably, as shown in Figure 2(a) and Figure 2(b), combination of ME and ATO synergistically reduced the cell viability of NIT-1 cells to 38.60%, showing lower cell survival rate than those of single ME or ATO treated groups. These results indicate that combination of ME with ATO is capable to increase the sensitivity of ATO for killing NIT-1cells. However, when ME treatment was given 24 h after ATO treatment, the cell survival rate was higher than that of the ATO group, changing from 67.68% to 78.20%. These results indicate that ME was able to reverse the killing effect of ATO on NIT-1 cells. This treatment was used in subsequent experiments.

In the Hoechst 33,258 fluorescent staining assay, the nuclei of control cells and ME-treated cells showed an oval round shape with homogenous intensity and were very lightly stained, whereas cells treated with ATO were found in a condensed and/or fragmented shape with irregularity and bright staining (Figure 2(c)). These apoptotic morphological changes showed greater difference in cells simultaneously treated with higher doses of ATO. As shown in Figure 2(d) and Figure 2(c), there was no significant difference between the ME group and the control group in apoptotic morphological changes, and there was no significant difference in the apoptotic rate between the two groups (p > 0.05). After ME treatment, the morphological changes of apoptosis induced by ATO decreased significantly, and the apoptosis rate decreased. The rate of apoptosis in the ATO group was 2.88 times of that in the control group, whereas, after ME treatment, the rate of apoptosis in the ME + ATO group was 2.11 times of the control group. These results indicate that ME treatment can reverse ATO-induced apoptosis.

As shown in Figure 3(a), ATO exposure led to the accumulation of reactive oxygen species, but ME treatment facilitated the ATO-induced removal of ROS. Fluorescence quantitative results further showed that treatment with ATO and ATO + ME led to an 49% and 29% increase on ROS production in comparison to control cells (Figure 3(b)).

In order to comprehensively evaluate the degree of oxidative damage, we further detected the SOD activities and MDA contents in ME-treated, ATO-treated and ATO + ME-treated NIT-1 cells. The results demonstrated that SOD activity respectively were 95.98 U/mg protein, 69.66 U/mg protein, or 76.82 U/mg

protein in NIT-1 cells which was exposed to ME, ATO and ATO + ME. SOD activity was lower in the ATO and ATO + ME groups than that in the control group (85.18 U/mg protein), and it was higher than that in the ATO + ME group when compared to the ATO group (Figure 3(c)). By contrast, MDA contents were 3.78 ± 0.22 nmol/mg protein and 3.13 ± 0.19 nmol/mg protein in ATO and ME + ATO treatment cells (Figure 2(c)). MDA levels were higher in the ATO and ATO + ME groups than that in the control group (2.71 ± 0.21 nmol/mg protein), while lower in the ATO + ME group in comparison to the ATO group.

An increase in Nrf2 and HO-1 mRNA levels was clearly observed in the ATO group, while Nrf2 and HO-1 mRNA levels were lower in the ME + ATO group compared to the ATO group, as shown in Figure 4. For ATO group and ME + ATO group, the HO-1 mRNA levels were 21.65-fold and 8.17-fold of untreated group respectively (P < 0.05), the Nrf2 mRNA levels were 3.05-fold and 1.42-fold of untreated group respectively (P < 0.05). Incidentally, both Nrf2 and HO-1 mRNA levels of the ME group were slightly lower than in the untreated group.

As illustrated in Figure 5, protein expressions of Nrf2 and HO-1 have similar tendencies to the changes of mRNA levels. In detail, the protein of Nrf2 in control group, ME group, ATO group and ATO + ME group was 1-fold, 1.01-fold, 1.84-fold and 1.39-fold, respectively. Also, HO-1 was increased after treatment with ATO and ATO + ME in comparison to control group (1.84-fold, 1.39-fold).