Four ethnic groups—the Yi, Bai, Dai, and Naxi people—have used P. polyphylla var. yunnanensis to treat cancer (Yan et al., 2021). The Yi ethnic group mix about 10 g (fresh weight) the rhizomes of P. polyphylla var. yunnanensis and Rohdea chinensis (Baker) N. Tanaka and make it into a decoction or use 3 g (dry weight) of the same mixture for the treatment of gynecological tumors. The Bai ethnic group employs 15 g each of a powder from P. polyphylla var. yunnanensis, Viola philippica Cav., Hedyotis diffusa Willd., and Scutellaria barbata D. Don and make it into a decoction for use once a day for 1–2 weeks to treat stomach and esophageal cancers. The Naxi ethnic group uses powder from P. polyphylla var. yunnanensis (20 g), H. diffusa Willd. (50 g), S. barbata D. Don (50 g), and Engleromyces goetzi P. Henn. (20 g) to prepare a decoction, which is consumed every day over 1–2 months for cancer treatment; the types of cancer were not specified. For the Dai ethnic group, there was a record treatment for cancer, but the prescription and type of cancer are unknown (Yan et al., 2021).

There are five types of cancer treatments using P. polyphylla var. yunnanensis including brain neoplasms and thyroid, laryngeal, gastric, and lung cancer, which have been recorded in the book of Yunnan Kang Ai Zhong Cao Yao (Hu and Xuan, 1982). There are 10 kinds of traditional Chinese medicine recipes for lung cancer, each using 30 g of powder as the basis for a decoction, comprising P. polyphylla var. yunnanensis, C. lacryma-jobi, S. scandens, H. cordata, Wolfiporia cocos (F. A. Wolf) Ryvarden & Gilb., Polyporus umbellatus, M. tenacissima, Asparagus densiflorus (Kunth) Jessop, Platycodon grandiflorus (Jacq.) A. DC., and Sageretia thea (Osbeck) Johnst. in various amounts. These anticancer prescriptions that include P. polyphylla var. yunnanensis, but the pharmacological and chemical properties of this species that could be effective for treating anticancer are unclear.

Deionized water was used throughout the study. Ethanol was analytical grade. ¹H and ¹³C NMR spectra were collected on an Avance III 600 spectrometer (Bruker, Switzerland) with tetramethylsilane (TMS) as an internal standard in pyridine-d ₅ (Cambridge Isotope Laboratories, Inc.). An ultra-high-performance liquid chromatograph connected to a triple quadrupole tandem mass spectrometer (UPLC-Xevo TQ MS, Waters, United States) was used to obtain electrospray ionization–mass spectrometry (ESI-MS) data. Silica gel (200–300 mesh, Qingdao Haiyang Chemical Co., Ltd., Qingdao, China), a macroporous resin column (D101, Jiangsu Donghong Chemical Co., Ltd., China), and thin-layer chromatography (TLC) plates (Qingdao Haiyang Chemical Co., Ltd., China) were used to monitor the compound. The TLC spots were visualized under ultraviolet (UV) light at 254 nm and stained by spraying the TLC plates with 5% (v/v) H₂SO₄ in alcohol followed by heating. Semi-preparative high-performance liquid chromatography (HPLC) analysis was performed on an Agilent 1260 series system (Agilent Technologies, United States) equipped with a diode array detector, and preparative HPLC was conducted with a Hanbon series system (Jiangsu Hanbon Science & Technology Co., Ltd., China) equipped with a UV detector, using two types of C18 columns produced by Agilent (Zorbax SB-C18 column, 5 μm; ϕ 9.4 × 250 mm and ϕ 4.6 × 250 mm).

As an identification of Good Agriculture Practice (GAP) and geographical indication products, the rhizomes of P. polyphylla var. yunnanensis which is native and local distribution (called “Daodi” medicine in Chinese) were obtained from Lijiang Yunxin Green Biological Development Co., Ltd., in NW Yunnan Province, China, where located in Three Parallel Rivers and one of hotpot on biodiversity conservation in the world, in May 2020 and identified based on morphological features by Professor Heng Li (CAS Key Laboratory for Plant Diversity and Biogeography of East Asia, Kunming Institute of Botany, Chinese Academy of Sciences, China).

The dried rhizomes of P. polyphylla var. yunnanensis (2.1 kg) were crushed into powder with a pulverizer and extracted in 80% (v/v) ethanol (3 × 6 L) with an ultrasonicator (40 KHZ) at 60°C for 2 h. The ethanol extract was filtered by filter paper, and the resulting filtrate was concentrated as crude extract (A, 400 g). A 300-g aliquot of the crude extract was separated on a macroporous resin column with four concentrations of ethanol to obtain an aqueous fraction (A ₁ , 170 g), a 50% ethanol fraction (A ₂ , 23 g), a 80% ethanol fraction (A ₃ , 5 g), and a 95% ethanol fraction (A ₄ , 4 g) (Supplementary Figure S7, in the Supplementary Material S2).

Fraction A3 (2.1 g) was isolated by column chromatography (silica gel, 200–300 mesh, CH₂Cl₂/CH₃OH, 19:1→0:1, v/v) as fractions A3–1 (186.3 mg), A3–2 (498.6 mg), A3–3 (300.1 mg), and A3–4 (103.2 mg). An aliquot from fraction A3–2 (109 mg) was separated by semi-preparative HPLC (Zorbax SB-C18 column, 9.4 × 250 mm; CH₃CN/0.1% [v/v] H₃PO₄ solution, 1 mL/min, 30°C, CH₃CN, 0–15 min, 0%–20%; 15–20 min, 20%–35%; 20–25 min, 35%–55%; 25–35 min, 55%–60%; 35–40 min, 60%–75%; 40–45 min, 70%–100%; 45–55 min, 100%–100%) to obtain four saponins: Paris saponin Pb (1, 12.9 mg, pennogenin-3-O-α-L-rhamnopyranosyl- (1→2)-[α-L-rhamnopyranosyl-(1→4)]-β-D-glucopyranoside, t_R = 35.05 min), Paris saponin H (2, 11.1 mg, pennogenin-3-O-α-L-arabinofuranosyl-(1→4)-[α-L-rhamnopyranosyl-(1→2)]-β-D-glycopyranoside, t_R = 35.76 min), Paris saponin III (3, 23.3 mg, diosgenin-3-O-α-L-rhamnopyranosyl-(1→2)-[α-L-rhamnopyranosyl-(1→ 4)]-β-D-glucopyranoside, t_R = 42.68 min), and Paris saponin I (4, 26.0 mg, diosgenin-3-O-α-L-arabinofuranosyl-(1→4)-[α-L-rhamnopyranosyl-(1→2)]-β-D-glucopyranoside, t_R = 44.24 min). An aliquot of fraction A3–3 (40.1 mg) was similarly separated by preparative HPLC (Zorbax SB-C18 column, 9.4 × 250 mm; CH₃CN/H₂O, 3 mL/min, 30°C, CH₃CN, 0–20 min, 30%–60%; 20–25 min, 60%–30%) to obtain two saponins: Paris saponin VII (5, 15.2 mg, pennogenin 3-O-α-L-rhamnopyranosyl-(1→2)-[α-L-rhamnopyranosyl-(1→4)-α-L-rhamnopyranosyl-(1→4)]-β-D-glucopyranoside, t_R = 18.26 min) and Paris saponin II (6, 7.6 mg, diosgenin-3-O-α-L-rhamnopyranosyl-(1→2)-[α-L-rhamnopyranosyl-(1→4)-α-L-rhamnopyranosyl-(1→4)]-β-D-glucopyranoside, t_R = 22.01 min) (Supplementary Figure S8, in the Supplementary Material S2).

The contents of total saponins in dry rhizomes were determined by the method described in the Chinese Pharmacopoeia (National Pharmacopoeia Committee, 2020). The dry rhizomes were crushed into powder and passed through a 50 mesh sieve (0.28 mm). The powder (498.8 mg) was placed in a conical flask with a cover, to which 25 mL ethanol was added before the weight of the system was measured. The system was refluxed for 30 min and cooled to room temperature. The weight of the entire system was determined again and brought to its original weight with ethanol. Five extracts (A, A₁, A₂, A₃, and A₄) in the analysis and determination of total saponin were prepared at concentrations of 9.45, 4.45, 11.50, 10.90, and 12.75 mg/mL in methanol, respectively. The liquid was filtered through a 0.45-μm membrane before use for HPLC analysis.

Samples were analyzed by HPLC (Agilent 1260 series system, Zorbax SB-C18 column, 5 μm, ϕ 4.6 × 250 mm), with a mobile phase consisting of CH₃CN/H₂O, with a flow rate of 1 mL/min, at 30°C, starting with CH₃CN, 0–20 min, followed by 30%–60% CH₃CN (in water, v/v); 20–25 min, 60%–30%. The injection volumes were 10 μL for the dry rhizome powder and 5 μL for each of the six extracts. Eight standard saponins (Paris saponins I, II, VII, III, V, VI, Pb, and H) were purchased from Sichuan Weikeqi Biological Technology Co., Ltd. (purity ≥98%).

Three human liver cancer cell lines (SMMC-7721, HepG2, and SK-HEP-1) and the non-small-cell lung cancer (NSCLC) cell line A549 were obtained from ATCC (Manassas, VA, United States). Normal human cell lines (liver cell line LO2 and lung cell line BEAS-2B) were obtained from ATCC (Manassas, VA, United States). Cells were cultured in RMPI-1640 or DMEM medium (Biological Industries, Kibbutz Beit-Haemek, Israel) supplemented with 10% (v/v) fetal bovine serum (Biological Industries) at 37°C in a humidified atmosphere with 5% (v/v) CO₂.

The effects of the extracts and isolated saponins on cell growth were evaluated by the MTS assay (3-(4,5-dimethylthiazol-2-yl)-5(3-carboxymethoxyphenyl)-2-(4-sulfopheny)-2H-tetrazolium; Promega, Madison, WI, United States) on human tumor cell lines (Mosmann, 1983; Cory et al., 1991; Malich et al., 1997). In a test experiment of cell inhibition, the cell inhibition effects of six crude extracts (100 μg/mL) and six isolated saponins (40 μM) were roughly measured to ascertain whether to measure the IC₅₀ value by MTS assay.

Each well of a 96-well plate (Multiskan Fc, Therm) was seeded with 100 μL cells at a density of 3,000–5,000 cells per well, with the cell suspension containing 10% (v/v) fetal bovine serum (DMEM or RMPI1640, Biological Industries, Kibbutz Beit-Haemek, Israel). After 12–24 h of inoculation and culture at 37°C, the extracts or saponins were added at the indicated concentrations after being dissolved in dimethyl sulfoxide (DMSO, Biological Industries, Kibbutz Beit-Haemek, Israel) in a final volume of 200 μL per well, each in triplicate. The six crude extracts were added at concentrations of 100, 20, 4, 0.8, and 0.16 μg/mL; the six isolated saponins were added at concentrations of 10, 5, 2.5, 1.25, 0.625, and 0.3125 μM. After culture for 48 h at 37°C, 20 μL of MTS solution and 100 μL of DMEM were added to each well; three blank wells were set aside with 20 μL MTS solution and 100 μL DMEM. After an incubation for another 2–4 h, the absorbance at 492 nm was measured. Cisplatin and Paclitaxel (MeilunBio) were used as a positive control in each experiment. The cell growth curve was drawn with the concentration as the abscissa and the cell survival rate as the ordinate. The IC₅₀ value of each saponin was calculated by the two-point method (Reed and Muench, 1938).

Cell cycle distribution was determined by propidium iodide (PI, BD, Pharmingen) staining (Kong et al., 2016; Wei et al., 2021). Briefly, three cancer cell lines were cultured in six-well plates at a density of 3 × 10⁵ cells per well for 24 h before treatment with each of the six isolated saponins at the indicated concentrations for 24 h. The cells were collected, centrifuged, and washed twice with ice-cold PBS. Cells were fixed in 70% (v/v) ice-cold ethanol at 4°C overnight and washed again in PBS twice. The cells were incubated with 100 μL RNase A at 37°C for 30 min in the dark, after which 400 μL PI was added to the cells with RNase A in the dark at 4°C for 15 min. The cell cycle stage was analyzed by flow cytometry (BD, FACSCelesta, United States) with FLOWJO software (FlowJo, Ashland, OR, United States).

The probe 5,6-dichloro-1,1′,3,3′- tetraethyl-imidacarbocyanine iodide (JC-10) was used as a measure of mitochondrial membrane potential (Hirose et al., 1974; Chaudhari et al., 2008). The effects of each of the six isolated saponins on mitochondrial membrane potential were determined using a JC-10 assay kit (×200, 10 μL/tube, five tubes), which was purchased from Beijing Solarbio Science & Technology Co., Ltd. (Beijing, China). Cells were cultured in a confocal dish with this method. After 12–24 h of inoculation and culture at 37°C, the purified saponins were added at the indicated concentrations. The cells were incubated at 37°C in a humidified atmosphere with 5% (v/v) CO₂ for 24 h. The culture medium was removed, and the confocal dishes were washed twice with 1 mL PBS, followed by 0.5 mL incubation buffer containing 2.5 μL JC-10 (×200) at 37°C for 20 min in the dark. After incubation, the supernatant was removed, and the cells were washed twice with 1 mL JC-10 buffer and PBS. PBS (1 mL) was added to each well of the confocal dish, and the stained cells were observed with a confocal laser scanning microscope (TCS SP8X, Leica, Germany) using a ×40 objective. The wavelengths for excitation were 490 nm (green fluorescence) and 525 nm (red fluorescence). The wavelengths for emission were in the range of 495–565 nm (monomers for JC-10) and 550–620 nm (polymer for JC-10).

Double staining with annexin V-fluorescein isothiocyanate (FITC) and PI was used to determine the effects of the six saponins on cell apoptosis by flow cytometry (Homburg et al., 1995; Vermes et al., 1995; Casciola-Rosen et al., 1996). The effects of the six isolated saponins on apoptosis were determined using an annexin V-FITC apoptosis kit (BD, Pharmingen); the chemotherapy drug doxorubicin (DOX, Beyotime Institute of Biotechnology, Shanghai, China) was used as the positive control. Cells were cultured in six-well plates at a density of 3 × 10⁵ cells per well for 24 h before being treated with each of the six isolated saponins at the indicated concentrations for 24 h. The cells were collected, centrifuged, and washed twice with ice-cold phosphate-buffered saline (PBS). The cells were stained with annexin V-FITC and PI for 15 min in the dark at room temperature before analysis by flow cytometry (BD, FACSCelesta, America). Annexin V-FITC⁺ and PI⁺ (Q2) were considered late-stage apoptotic cells, while annexin V-FITC⁺ and PI⁻ (Q3) cells were considered early-stage apoptotic cells. Annexin V-FITC^– and PI⁺ (Q1) were necrotic or mechanically damaged cells, and annexin V-FITC^– and PI⁻ (Q4) cells were healthy living cells.

The saponins 1–6 (Figure 1) were isolated from the dried rhizomes of P. polyphylla var. yunnanensis and were identified to be Paris saponin Pb (1, pennogenin-3-O-α-L-rhamnopyranosyl-(1→2)-[α-L-rhamnopyranosyl-(1→4)]-β-D-glucopyranoside) (Nakano et al., 1989; Hua et al., 2015), Paris saponin H (2, pennogenin-3-O-α-L-arabinofuranosyl-(1→4)-[α-L-rhamnopyranosyl-(1→2)]-β-D-glycopyranoside) (Chen and Zhou, 1987; Wang et al., 2007; Pang et al., 2015), Paris saponin III (3, diosgenin-3-O-α-L-rhamnopyranosyl-(1→2)-[α-L-rhamnopyranosyl- (1→4)]-β-D-glucopyranoside) (Han et al., 1999; Wang et al., 2001; Wang et al., 2007), Paris saponin I (4, diosgenin-3-O-α-L-arabinofuranosyl-(1→4)-[α-L-rhamnopyranosyl-(1→2)]-β-D-glucopyranoside) (Chen and Zhou, 1987; Huang et al., 2009; Jing et al., 2017), Paris saponin VII (5, pennogenin 3-O-α-L-rhamnopyranosyl-(1→2)-[α-L-rhamnopyranosyl-(1→4)-α-L-rhamnopyranosyl-(1→4)]-β-D-glucopyranoside) (Chen et al., 1983; Zhang et al., 2014), and Paris saponin II (6, diosgenin-3-O-α-L-rhamnopyranosyl-(1→2)-[α-L-rhamnopyranosyl-(1→4)-α-L-rhamnopyranosyl-(1→4)]-β-D-glucopyranoside) (Munday et al., 1993; Jing et al., 2017) by their nuclear magnetic resonance (NMR) and mass spectrometry (MS) data, which were provided in the Supplementary Material. These values were all consistent with the known saponins in the reported literature.

The Chinese Pharmacopoeia details a method for the determination of total saponin contents in P. polyphylla extracts (National Pharmacopoeia Committee, 2020), whose levels in dry TCM powder should reach at least 0.6% (National Pharmacopoeia Committee, 2020) to be effective and should include Paris saponin I (C₄₄H₇₀O₁₆), Paris saponin II (C₅₁H₈₂O₂₀), and Paris saponin VII (C₅₁H₈₂O₂₁). Accordingly, total saponin contents in the powder of dry rhizomes and six ethanol or aqueous extracts were measured using eight saponins as standards (Table 1). Six saponins (Paris saponins I, II, VII, III, H, and Pb) were detected in dry rhizome powder; importantly, the contents for Paris saponins I, II, and VII were much higher (6.96%) than the contents set by the Chinese Pharmacopoeia. Paris saponins I, II, and VII and total saponins reached their highest contents in fraction A₃, which was obtained by elution from the macroporous resin column with 80% ethanol. These two sets of saponins comprised 37.06% (Paris saponins I, II, and VII) and 59.38% (total saponins) of this fraction. Therefore, fraction A₃ was subjected to further separation for the purification of the six detected saponins.

Steroidal saponins accumulate to different levels in Paris plants each year and as a function of their natural habitat (Wang et al., 2015). In addition, distinct Paris species present different steroidal saponin profiles (Wei et al., 2014); for example, 112 steroidal saponins were detected in P. polyphylla var. yunnanensis (Zhou et al., 2022). The saponin contents of the P. polyphylla var. yunnanensis rhizomes tested here were much higher than those set by the standards of the Chinese Pharmacopoeia in 2020, which used plants grown in Lijiang, Yunnan province, China. The samples analyzed here mainly contained the eight Paris saponins I, II, VII, III, V, VI, H, and Pb (Table 1).

The growth inhibitory activity of all extracts and saponins were obtained from the dried P. polyphylla var. yunnanensis rhizomes. A preliminary experiment indicated that five crude extracts (A, A₁, A₂, A₃, and A₄) inhibit cell division when applied at 100 μg/mL in the pre-experiment. To facilitate comparisons, the IC₅₀ values of five extracts (A, A₁, A₂, A₃, and A₄) were determined in the three human liver cancer cell lines SMMC-7721, HepG2, and SK-HEP-1 as well as in the NSCLC cell line A549 by the MTS method (Table 2). Among all crude extracts, A₃ was the most active against all four human cancer cell lines tested, with IC₅₀ values of 1.226 μg/mL (SMMC-7721), 1.878 μg/mL (HepG2), 0.878 μg/mL (SK-HEP-1), and 1.287 μg/mL (A549).

Six main saponins (1–6) were obtained from the further isolated and identified in the A₃ fraction (Figure 1), and their IC₅₀ values for growth inhibition were determined using the same four human cancer cell lines. Importantly, all IC₅₀ values were lower than those obtained with the positive control cisplatin, a commonly used chemotherapy drug (Table 2). In addition, the normal liver and lung cells (LO2 and BEAS-2B) had also been evaluated for six saponins (Table 3). However, none of the purified saponins exhibited the same level of activity as the other chemotherapy drug paclitaxel. Moreover, each saponin showed a distinct activity when tested against the four cancer cell lines, with no single saponin being most effective against all cell lines. More specifically, Paris saponin II (6) was most active against NSCLC, as evidenced by the low IC₅₀ values. Likewise, Paris saponins I (4) and Pb (1) were most active against the three liver cancer cell lines. A previous review mentioned that toxicity evaluation suggested that Rhizome Paridis had slight liver toxicity (Ding et al., 2021). However, the present data for normal cells supported this view (Table 3).

Paris saponins can be divided into diosgenin-type (saponins 3, 4, and 6) and pennogenin-type (saponins 1, 2, and 5) as a function of the presence of a hydroxyl group at site 17 (chain R1 in Figure 1) (Man et al., 2013; Wang et al., 2018). However, the results showed that the activities of these two kinds of saponins provided limitation, related different results compared with pre-evidence (Table 2). Although, the species of Paris has been extensively used for cancer therapy in indigenous medicine knowledge (Yan et al., 2021), research on the anticancer properties and potential mechanisms is still a preliminary exploration mainly due to the limitation of their low in plants. With the development of artificial cultivation techniques of Paris species for the past few years, more and more studies of metabolites and pharmacology are being performed to further identify effective steroidal saponins as potential anticancer drugs (Long et al., 2015; Qin et al., 2018). The results of MTS assay further confirmed the correlation between total saponin content in crude extracts and anticarcinogenic activity. What’s more, the difference in the anticancer activity of six saponins should be highly associated with the type, binding mode and number of glycosyl (Figure 1; Tables 1, 2).

To test whether inhibition of the growth of three human cancer cells also affected cell cycle progression (G0/G1, S, and G2/M), induction of A549, SMMC-7721, and HepG2 cell cycles was detected by propidium iodide (PI) staining and flow cytometry after six isolated saponin treatments (Figures 2Bn–Gn). The same cell lines with DMSO were only treated as a negative control (Figure 2An). Substantial changes were detected in the cell cycle progression of the three human cancer cell lines when treated with Paris saponins (Figures 2, 3).

Compared to the DMSO control, saponin 2 increased the proportion of A549 cells in the S and G2/M phases, while saponins 1, 4, and 6 increased the proportion of cells in the G2/M phase (Figures 2A1–G1). Saponin 5 increased the proportion of cells in S phase, and saponin 3 increased the proportion of cells in the G0/G1 phase.

The six Paris saponins also induced cell cycle arrest in the SMMC-7721 cell line (Figures 2A2–G2). Compared to the DMSO control, saponin 6 increased the proportion of cells in the G2/M phase by 27.10%, while the other five saponins increased the proportion of cells in the G0/G1 phase. The HepG2 cell line was also sensitive to Paris saponins (Figures 2A3–G3). Compared to the DMSO control, saponins 4 and 6 increased the proportion of cells in the G2/M phase by 17.2% and 25.8%, respectively; saponins 1 and 3 increased the proportion of cells in S phase by 23.7% and 27.5%, respectively; and saponins 2 and 5 increased the proportion of cells in the S and G2/M phases.

Many anticancer medications inhibit the growth of tumor cells by targeting the S and G2/M phases of the cell cycle in cancer cells. Indeed, both S and G2/M phases are very important for the proliferation of malignant cancer cells (Sui et al., 2016). A closer look at the flow cytometry data revealed that the proportion of HepG2 cells at the S or G2/M phase treated with any one of the six saponins is higher relative to the DMSO control, suggesting that all six saponins block the cell cycle at the S or G2/M phases (Figure 3C). For the other liver cancer cell line, SMMC-7721, only saponin 6 caused cell cycle arrest at the G2/M phase (Figure 3B). For the A549 lung cancer cell line, only saponin 4 showed a significant effect on the relative proportions of cells in each cell cycle stage, with an increase in the proportion of cells at the G2/M phase, suggestive of cell cycle arrest (Figure 3A).

These results indicate that the six Paris saponins inhibit cancer cell growth by inducing cell cycle arrest at the S or G2/M phase.

There were two main apoptotic pathways in tumor cells, including the cell surface death receptor-mediated pathway and the mitochondrial apoptosis pathway, the latter being at the core of apoptosis induction (Sui et al., 2016). The six isolated saponins were therefore assessed for their ability to induce early apoptotic cell death in three human cancer cell lines (A549, SMMC-7721, and HepG2) using a mitochondrial membrane potential assay kit based on the potential-dependent accumulation of the JC-10 dye inside mitochondria (Figures 4, 5; Supplementary Figures S9–S11). Red fluorescence from the JC-10 probe indicates a normal mitochondrial membrane potential, while a shift to green fluorescence reflects a drop in membrane potential, which is suggestive of a cell in the early stage of apoptosis.

Cancer cells showed red or yellow-green fluorescence in the absence of any treatment, indicating that their mitochondrial membrane potential is normal (Figures 4A1–A3), which used as negative control (NC). As a positive control for cells with depolarized mitochondrial membranes, cells were treated with the proton ion mitochondrial uncoupling agent carbonyl cyanide 3-chlorophenylhydrazone (CCCP), which resulted in strong green fluorescence (Figures 4B1–B3). Each cancer cell line was also treated with one of the six isolated saponins (Figures 4Cn–Hn). However, the merge by red and green fluorescence for three human cancer cells of six saponins were showed in Figure 4. Also, the merge, green channel and red channel for each human cancer cells of six saponins were showed in Supplementary Figures S9–S11 in the Supplementary Material S2, respectively. Compared to the NC and CCCP control, the green fluorescence was generally higher than red fluorescence in SMMC-7721 cells, next is A549 cells. And for HepG2 cells, the red fluorescence was higher than green fluorescence (Figure 4). Importantly, all purified saponins displayed an increase in green fluorescence over the untreated controls. This finding indicated that the six saponins purified from dried P. polyphylla var. yunnanensis rhizomes increase the permeability of the mitochondrion membrane, leading to a decrease in mitochondrial membrane potential.

JC-10 fluorescent probe can enter mitochondria in normal cells when the membrane potential of mitochondria is high, and it exists in the form of polymer, showing red fluorescence. During the period of cell apoptosis, the mitochondrial membrane potential decreases, and JC-10 is dissociated in the cytoplasm in the form of monomer, showing green fluorescence (Hirose et al., 1974; Castedo et al., 2002). In this study, the ratio of relative fluorescence intensity was used to evaluate the changes of mitochondrial membrane potential, which was indicated by the ratio of red/green fluorescence (Figure 5). The lower the ratio of relative fluorescence intensity was, the more red fluorescence and the less green fluorescence were, which was indicated more severe the damage to the cell membrane, resulting in lower mitochondrial membrane potential. As shown in Figure 5, the ratio of relative fluorescence intensity was decreased in different levels after treating with six saponins, which indicated that the six saponins caused the decrease of mitochondrial membrane potential by damaging mitochondrial membrane, and suggested that cancer cells were in the early stage of cell apoptosis.

To sum up, six Paris saponins may increase the permeability of the mitochondrion membrane in different levels, leading to a decrease in mitochondrial membrane potential.

Anticancer drugs inhibit the proliferation of tumor cells by increasing the apoptosis rate (Foo et al., 2015). Encouraged by the results from the mitochondrial membrane potential assay above, the effects of the six isolated saponins on apoptosis were determined by double staining with annexin V-FITC and PI in three human cancer cell lines, including A549, SMMC-7721, and HepG2 (Figure 6). Cancer cells were treated with 0.1% DMSO as a negative control (Figure 6An) and the chemotherapy drug doxorubicin (DOX) as a positive control (Figure 6Bn). These six isolated saponins significantly increased early-stage and late-stage apoptotic cells, as reflected by the proportion of annexin V-FITC⁺/PI⁻ and annexin V-FITC⁺/PI⁺ cells, respectively (Figures 6Cn–Hn). A concentration of 5 μM for each saponin was sufficient to promote apoptosis in the three cancel cell lines tested here. The effects of six saponins on promoting apoptosis were equivalent to that of DOX, or even more than that of DOX.

Apoptosis rate of each cell line was calculated in the absence or presence of each purified Paris saponin (Figure 7). Saponin 1 was the most effective in inducing apoptosis in A549 and SMMC-7721 cells and was also effective in HepG2 cells. Saponins 2 and 5 also induced strong apoptosis in A549 cells, while saponins 5 and 6 exhibited a substantial induction of apoptosis in SMMC-7721 cells. For HepG2 cells, the highest apoptosis rate was saponin 5, and next were saponins 2 and 1. Notably, saponins 5 and 1 consistently induced stronger apoptosis than the positive control DOX in all three cancer cell lines. In conclusion, six saponins induced significant apoptosis in the three human cancer cell lines, supporting their potential as inhibitory saponins against the proliferation of cancer cells.

We found six saponins inhibited cancer cell growth in three human cancer cells was further demonstrated (Table 2). Moreover, the anticancer activity of six saponins were associated with induction of early- and late-stage apoptosis and cell cycle G2/M or S phase arrest in there human cancer cells (Figures 2, 6). Apoptosis, or programmed cell death, plays a crucial role in controlling cell number in many developmental and physiological settings and in chemotherapy-induced tumour-cell killing, through regulating the normal balance between cell death and survival (Hu and Kavanagh, 2003). Therefore, apoptosis inducers of cancer cells have therapeutic potential in chemoprevention and chemotherapy of cancer (Long et al., 2015). The study showed that the breakdown of mitochondrial membrane is able to enhance the permeability of outer mitochondrial membrane. The release of induced cytochrome and other apoptotic factors from mitochondria into the cytosol is critical to activate the intrinsic apoptotic pathway (Long et al., 2015). Herein, the main effects of six saponins in mitochondrial apoptosis pathway were detected by JC-10 staining and Annexin V- FITC & PI staining in human cancer cells, respectively (Figures 4, 6). Our findings suggest that anti-cancer activity of six saponins is associated with apoptotic induction through a decrease in mitochondrial membrane potential in A549, SMMC-7721, and HepG2 cells.