FZKA, containing 12 components, was obtained from Guangdong Kangmei Pharmaceutical Company Ltd. (Guangdong, China), as previously reported (Xiao-Bing Yang et al., 2014). The components of FZKA include Radix Pseudostellariae 30 g, Rhizoma Atractylodis Macrocephalae 15 g, Milkvetch Root 30 g, Hedyotis Difusa 30 g, Solanum Nigrum 30 g, Chinese Sage Herb 30 g, Indian Iphigenia Bulb 30 g, Coix Seed 30 g, Akebia Trifoliata Koidz 30 g, Snake Bubble Ilicifolius 30 g, Curcuma Zedoaria 15 g, Licorice 10 g. For in vitro experiments, the granules were dissolved in RPMI-1640 medium to a final concentration of 20 mg/ml and centrifuged at 14,000 rpm for 10 min; the supernatant was then filtered using 0.22 μm filter before use and the pH value of the cultured cells with the media was adjusted to 7.2–7.4 after FZKA addition. For in vivo experiments, animals were treated with FZKA by intragastric administration.

The initial batch to batch consistency study was performed using HPLC, as previously reported (Zheng et al., 2016). Briefly, the sample solutions were put into the HPLC system (250 mm × 4.6 mm, 5 μm, ACE, Scotland). The mobile phase consisted of deionized water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B). The gradient elution program was as follows: 5% B at 0–5 min, 5–20% B at 5–10 min, 20–40% B at 10–15 min, 40–95% B at 15–40 min, and 95–100% B at 40–45 min. The flow rate was 1.0 ml/min, and the detection wavelength was set at 280 nm. The injection volume was 10 μl and the column temperature was maintained at 30°C. The efficacy of different batch of FZKA is dependable (Wang et al., 2020).

NSCLC cell lines including A549, H1299, PC9 and H1650 were obtained from Guangzhou Cellcook Biotech Co. (Guangzhou, China). All cells were grown at 37°C in a humidified 5% CO₂ and 95% air and cultured in RPMI-1640 medium (Life Technologies, Carlsbad, CA, United States) containing 10% FBS (Gibco, United States) and 0.5% penicillin-streptomycin sulfate (Invitrogen Life Technologies, Carlsbad, CA, United States). Annexin V-FITC Apoptosis Detection Kit and Cell Counting Kit (CCK-8) were purchased from Shanghai Yisheng Biotechnology Co. (Shanghai, China). BODIPY™ 581/591 C11 was obtained from Thermo Fisher Scientific (Waltham, MA). Lentiviral vectors for overexpression constructs were purchased from GeneCopoeia (Rockville, United States). The antibodies were obtained from the following sources: GPX4 (ab125066) and GAPDH (ab9485) were purchased from Abcam (Cambridge, United Kingdom); SLC3A2 (4F2hc/CD98) (47213S), SLC7A11 (12691S), horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody (7074S), were from Cell Signalling Technology (Danvers, MA); SLC7A11 (bs-6883R) for immunohistochemistry was obtained from Bioss Biological Technology Co. Ltd. (Beijing, China). FerroOrange and GSSG/GSH Quantification Kit were purchased from Dojindo Molecular Technologies Company (Kumamoto, Japan).

Cell proliferation was measured by CCK-8 assay according to the manufacturer’s protocol. Briefly, the NSCLC cells were administered with different treatments, and incubated with the CCK-8 reaction solution for 1.5 h. After that, the optical density (OD) values were measured at the wavelength of 450 nm to evaluate cell viability.

5-ethynyl-2′-deoxyuridine (EdU) proliferation assay was performed to measure cell proliferation. Cells were plated in 96-well plates at a density of 8 × 10³ cells/well. After adding FZKA for 24 h, cells were treated with 50 µM EdU (RiboBio, Guangzhou, China) and fixed with 4% paraformaldehyde in PBS for 30 min. After permeabilization with 0.5% TritonX-100 for 10 min, the cells were stained with 1× Apollo reaction reagent. Then the DNA contents were stained with Hoechst 33,342 for 30 min. The photographs were obtained using fluorescence microscope.

Cells were treated with FZKA and ferroptosis inhibitors. Annexin V-FITC Apoptosis Detection Kit was used to observe the quadrant distribution of cell death. Both floating and adherent cells were collected and washed 3 times with PBS. Finally, 10 μl Annexin V-FITC and 5 μl PI were added into the cells at room temperature for 15 min. The quadrant distribution of cell death was measured using flow cytometry with the acquisition criteria of 10,000 events for each sample.

The mitochondrial ultrastructure was observed by transmission electron microscopy (TEM). 2 × 10⁶ cells were seeded into 100 mm cell culture dishes and exposed to FZKA decoction and erastin for 10 h, respectively. After that, cells were collected, washed three times with PBS, and fixed with 2.5% glutaraldehyde. Samples were then pretreated according to standard procedures, including staining, dehydration, embedding, and slicing to obtain ultra-thin sections. During the analysis, images were acquired using a HITACHIH-7650 transmission electron microscope (Hitachi, Tokyo, Japan).

C11-BODIPY (10 μM) was added to FZKA treated or untreated cells for 0.5 h, then cells were collected by trypsin. Oxidation of the poly-unsaturated butadienyl portion of C11-BODIPY resulted in a shift of the fluorescence emission peak from ∼590 to ∼510 nm. Samples were analyzed using flow cytometry (Exc: 488 nm, Em: 510 nm) after washing twice with PBS, and the results were analyzed by NovoExpress software.

To clarify Fe²⁺ ions generation via the nanoparticles in cells, FerroOrange (1 μM, an intracellular Fe²⁺ ions probe, Ex: 543 nm, Em: 580 nm) dispersed in serum-free medium was added to the cells, and cells were incubated for 30 min in a 37°C incubator. Cells were then collected by trypsin. Finally, the fluorescence of cells were captured using flow cytometry after washing twice with PBS.

Lentiviral vectors including GPX4 (CS-M0369-Lv105) and Negative Control (EX-NEG-Lv105) were purchased from GeneCopoeia. For lentivirus production, HEK293T packaging cells were transfected with 10 μg lentiviral vectors using the calcium phosphate method. After 48 h of incubation, the viral supernatant was collected and filtered. NSCLC cells were incubated overnight with the viral supernatant and supplemented with 10 μg/ml polybrene. Puromycin at a dose of 2 μg/ml was used to select the cell line overexpressing GPX4 stably.

Western blot was conducted as previously reported (Wang et al., 2018b). Briefly, the cells were harvested, washed and lysed with 1 × RIPA buffer, and their protein concentrations were measured using Bradford method. SDS-PAGE was used to separate the protein in each sample. Proteins were transferred from gel to membrane. Then, the membrane was blocked and incubated with indicated primary antibodies. The blots were rinsed before probed with secondary antibodies. The reactive bands were visualized by ECL and scanned using the Bio-Rad ChemiDoc XRS + Chemiluminescence imaging system (Bio-Rad, Hercules, CA, United States). All the results were analyzed by ImageJ software.

Total RNA was isolated using Trizol (Invitrogen, CA, United States). Transcriptor first strand cDNA synthesis kit (Roche, Basel, Switzerland) was used to convert RNAs to cDNAs. And FS essential DNA green master (Roche, Basel, Switzerland) was used to perform qRT-PCR. Complementary DNA from various cell samples was amplified with specific primers. GPX4: 5′-AGT​GAG​GCA​AGA​CCG​AAG​T-3′ and 5′- AAC​TGG​TTA​CAC​GGG​AAG​G-3′; GAPDH: 5′- GAACGGGAAGCTCACTGG -3′ and 5′- GCC​TGC​TTC​ACC​ACC​TTC​T -3′. Data were analyzed with 2^−ΔΔCt for relative changes in gene expression.

The intracellular level of GSH and GSSG (glutathione disulfide) were performed by GSSG/GSH Quantification Kit (Dojindo, Kumamoto, Japan), following the manufacturer’s instruction. The concentration of total glutathione or GSSG was calculated via standard curve. GSH level was calculated as: GSH = (total glutathione-GSSG) × 2. The ratio of GSH/GSSG was calculated as (GSH)/(GSSG).

All animal experiments were approved by the Ethics Committee of Guangdong Provincial Hospital of Chinese Medicine (2020079). A total of 1.0 × 10⁶ A549 cells were subcutaneously injected into the right flank of the athymic BALB/c nude mice (aged 4–6 weeks, weight 18–20 g, female; Vital River, Beijing, China). When the tumor mass became palpable (at day 5 after injection), the mice were randomly divided into three groups: control, FZKA (31 g/kg) and combination with FZKA and liproxstatin-1(30 mg/kg). Tumors were measured every 5 days with digital calipers. The tumor volume (in mm³) was calculated using the formula: Volume=(L×W²)/2. Mice were sacrificed around day 25 after injection, when some of the tumors reached the size limit set by the institutional animal care and use committee. Tumors were weighed after careful resection.

The protein levels of GPX4, SLC7A11 and SLC3A2 expression were detected immunohistochemically on paraffin-embedded xenograft tumor tissue sections. Briefly, sections were treated with 10 mM sodium citrate buffer (pH 6.0) for heat-induced retrieval of the antigen and immersed in 3% hydrogen peroxide solution to inhibit endogenous peroxidase activity, followed by incubation of the sections in 5% bovine serum albumin to block nonspecific binding. The sections were incubated with primary antibodies against GPX4 (1:250), SLC7A11 (1:100) and SLC3A2 (1:100) at 4°C overnight and then incubated with biotinylated secondary antibody followed by the Liquid DAB Substrate Chromogen System according to the manufacturer’s instructions. Protein expression level was evaluated by counting at least 500 tumor cells in at least five representative high-power fields. The percentage of positive tumor cells and the staining intensity were multiplied to produce a weighted score for each case (Lu et al., 1998).

Statistical analysis was performed using the SPSS statistical software. Statistical evaluation for data analysis used Student’s t-test when there were only two groups (two sided) and differences between groups were assessed by one-way ANOVA. All data are reported as Mean ± SD. Differences between groups were considered significant statistically when p ≤0.05.

Our previous studies have shown that FZKA inhibited the growth of NSCLC cell lines including A549, PC9, and H1975 cells (Wang et al., 2020; Zheng et al., 2019). In the present study, we further observed the effect of FZKA on NSCLC cell growth inhibition in another NSCLC cell types including H1650 and H1299 using CCK-8 assay. We reconfirmed that FZKA decreased H1650 and H1299 cell viability in a dose-and time-dependent manner (Figure 1A). Similar findings were also demonstrated by EdU proliferation assay, which detects EdU incorporated into cellular DNA during cell proliferation (Figure 1B). Intriguingly, we found that blocking ferroptosis by UAMC 3203 and liproxstatin-1 could significantly reversed changes in the quadrants of Annexin V-/PI+ (Q1) and Annexin V+/PI+ (Figure 1C). The cells in Q1 quadrant in the top left had been supposed as non-apoptotic cells, and ferroptotic cell death was included in Q1 quadrant (Gai et al., 2020). The data showed that inhibiting ferroptosis could decrease the percent of Q1 induced by FZKA, which suggests that ferroptosis plays an important role in the inhibition effect of FZKA in NSCLC cells and FZKA may promote ferroptosis in NSCLC cells.

To identify whether NSCLC cell ferroptosis was induced by FZKA treatment, lipid peroxidation and intracellular-free iron, as two key characteristics of cell ferroptosis, were then detected in NSCLC cells after treatment with FZKA (Lei et al., 2019). C11-BODIPY was used as a lipid peroxidation probe in mammalian cells (Ortega Ferrusola et al., 2009). The intracellular labile Fe (II) levels in the living cells were measured by FerroOrange (Hirayama et al., 2020). The results showed that FZKA increased the levels of lipid peroxidation (Figure 2A) and intracellular-free iron (Figure 2B) in A549 and PC9 cells. The same results were also observed in H1299 and H1650 cells (Supplementary Figure S1). Moreover, the characteristic changes of ferroptosis on mitochondria, including swollen, decreased cristae, mitochondria vacuolization and increased membrane density, were further observed under TEM in NSCLC cells (Figure 2C, Supplementary Figure S2).

To further observe the role of ferroptosis in FZKA-treated NSCLC cells, two ferroptosis inhibitors including UAMC 3203 and liproxstatin-1 were applied to block ferroptosis. As shown in Figure 3A and Supplementary Figure S3, the FZKA-induced elevation of lipid peroxidation was almost reversed by treating with UAMC 3203 and liproxstatin-1. Our further data showed that blocking ferroptosis remarkedly reversed the inhibition effect of FZKA on NSCLC cell lines (A549, PC9, H1650 and H1299), as shown in the Figure 3B. These results indicated that ferroptosis plays a vital role in FZKA treated NSCLC cells.

GPX4 is a key inhibitor of lipid peroxidation and ferroptosis. The down-regulation of GPX4 could directly or indirectly trigger ferroptosis as a result of lipid peroxidation inhibition. We detected the expressions of GPX4 at protein and mRNA level after treatment with FZKA. Our data found that the protein level of GPX4 was significantly decreased following the application of FZKA in NSCLC cells (Figure 4A). Meanwhile, the mRNA level of GPX4 were also decreased by treating with FZKA (Figure 4B). The above data indicated that GPX4 might be a main molecular in the FZKA-induced NSCLC cell ferroptosis process.

The cystine-glutamate antiporter system xc⁻, which is composed of the subunits SLC7A11 and SLC3A2, plays a protective role against cell ferroptosis (Cui et al., 2018). We revealed that the protein levels of SLC7A11 and SLC3A2 were obviously decreased following the application of FZKA in a dose-dependent manner in H1650 and H1299 cells (Figure 5A). Glutathione is a tripeptide, that is, derived from cysteine, glutamate, and glycine, among which cysteine is the rate-limiting precursor. As expected, the amount of GSH was significantly decreased by FZKA. GSH is highly reactive with lipid ROS, and their reaction generates glutathione disulfide (GSSG). A reduced ratio of GSH/GSSG is considered to be a marker of oxidative stress. In our study, we found that FZKA reduced the ratio of GSH/GSSG significantly (Figure 5B). System xc⁻ and GSH are at the upstream of GPX4, therefore, our data showed that system xc⁻/GSH/GPX4 axis plays an important role in FZKA-induced NSCLC cell ferroptosis.

Since GPX4 plays a crucial role in the process of cell ferroptosis, we further confirmed that GPX4′s role in FZKA-induced NSCLC ferroptosis. We then over-expressed GPX4 in H1299 and PC9 cells by transfecting lentivirus (Figure 6A). As expected, the induced effect of NSCLC ferroptosis by FZKA was substantially decreased following GPX4 overexpression as shown by lipid peroxidation assay (Figure 6B and Supplementary Figure S4). Interestingly, NSCLC cell viability inhibition by FZKA was also partially reversed when over-expressed GPX4 (Figure 6C). This data further confirmed that GPX4 contributes to the effect of FZKA-induced NSCLC cell ferroptosis and FZKA-suppressed NSCLC cell growth, suggesting that GPX4 plays a crucial role in FZKA-treated NSCLC cells.

To validate the effect of FZKA-induced NSCLC cell ferroptosis in vivo, we constructed NSCLC cell xenograft model. As shown in Figures 7A,B, mice tumor growth was obviously inhibited by FZKA treatment undoubtedly. Notably, when cell ferroptosis was blocked by liproxstatin-1, the inhibition effect of tumor growth by FZKA was significantly rescued. Then system xc⁻ and GPX4 were detected by Western blot and immunohistochemistry. As expected, the data was consistent with in vitro results showing downregulated expression of system xc⁻ and GPX4 in the FZKA-treated group (Figures 7C,D). Totally, our xenograft model data reconfirmed the effect of FZKA-induced NSCLC cell ferroptosis and system xc⁻/GPX4 axis plays a crucial in the process.