FYLM consists of ten Chinese herbal medicines, including 18 g of Huangqi [Astragalus membranaceus (Fisch.) Bge.], 24 g of Banzhilian (Scutellaria barbata D. Don), 12 g of Baizhu (Atractylodes macrocephala Koidz.), 20 g of Baihuasheshecao [Oldenlandia diffusa (Willd.) Roxb.], 9 g of Renshen (Panax ginseng C. A. Mey.), 15 g of Fuling [Poria cocos (Schw.) Wolf], 15 g of Zhebeimu (Fritillaria thunbergii Miq.), 20 g of Shancigu [Cremastra appendiculata (D. Don) Makino], 20 g of Yiyiren [Coix lacryma—jobi L. var. ma—yuen (Roman.) Stapf] and 6 g of Gancao (Glycyrrhiza uralensis Fisch.). All the Chinese medicine pieces were provided by Bozhou Huqiao Pharmaceutical Co., Ltd (Bozhou, China) and authenticated by Prof. Feng Li of Shandong University of Traditional Chinese Medicine. For in vivo animal studies, all herbs (159 g in total) were soaked in ten times cold distilled water for 1 h and then decocted for 40 min. After filtering the solution, the dregs of decoction were decocted again in eight times distilled water for 30 min. After combining two filtrates, the FYLM decoction was concentrated to 66 ml under a rotary evaporator (N-1300, Shanghai Ailang Instrument Co., Ltd, Shanghai, China). At last, 2.4 g (crude drug)/ml FYLM decoction was prepared and stored at −20°C.

The composition analysis of FYLM decoction was detected using a chromatograph (instrument model: UltiMate 3000 RS) and a Q Exactive high-resolution mass spectrometer [Thermo Fisher Scientific (China) Co. Ltd.]. 1 ml of 80% methanol was mixed with 200 µl of FYLM solution and vortexed for 10 min. After centrifugation at 20,000 g for 10 min at 4°C, the supernatant was filtered for subsequent studies. Chromatographic analysis was performed on an AQ-C18 column (150 mm × 2.1 mm, 1.8 µm, Welch) at 35°C with an injection volume of 5 µl. The chromatographic gradient elution procedure: 98% aqueous phase, 2% organic phase at 1 min; 80% aqueous phase, 20% organic phase at 5 min; 50% aqueous phase, 50% organic phase at 10 min; 20% aqueous phase, 80% organic phase at 15 min; 5% aqueous phase, 95% organic phase at 20 min; 5% aqueous phase, 95% organic phase at 27 min; 98% aqueous phase, 2% organic phase at 28 min; 98% aqueous phase, 2% organic phase at 30 min. The mass spectrometry analysis was performed using a full mass/dd-MS2 detection method with a positive and negative ion switching scan. Data collected from the high-resolution liquid mass were initially analyzed by CD2.1 (Thermo Fisher) and then compared to a database (mzCloud).

Twenty male Sprague-Dawley rats (weighting 180 ± 10 g) were purchased from Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China) [animal License number: SCXK (Jing) 2016-0006]. Animals were housed in a 12/12 h light/dark cycle at 23°C ± 2°C with 50%–60% relative humidity and free of the specific pathogens [Laboratory use of animal license number: SYXK(Lu)2017-0022], with free access to food and water (Geng et al., 2021).

After 1 week of adaptive feeding, the animals were randomly divided into two groups, the control group and the FYLM administration group. The clinical dose of FYLM in adults is 159 g/60 kg and the gavage dose for rats is 6.3 times the clinical dose based on body surface area. The rats of FYLM administration group were given FYLM at the dose of 16.7 g/kg by gavage. While rats in the control group were given saline gavage (1.5 ml/100 g body weight). After 6 days of continuous gavage, rats were anesthetized with 3% pentobarbital sodium (45 mg/kg, intraperitoneal injection) (Sun et al., 2004), and blood was collected from the abdominal aorta. After resting for 1 h at room temperature, the serum was obtained by centrifugation at 3,000 rpm for 15 min at 4°C. Subsequently, it was inactivated in a water bath at 57°C for 30 min. After that, the bacteria were removed by filtration through a 0.22 µm microporous membrane and stored in a refrigerator at −80°C for in vitro experiments.

Mouse Lewis lung carcinoma (LLC) cell line and human NSCLC cell line (H1975) were purchased from the Cell bank, Type Culture Collection of Chinese Academy of Sciences (Shanghai, China). LLC cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM, Gibco, Beijing, China) containing 10% fetal bovine serum (FBS, ExCell Bio, Shanghai, China) and 1% Penicillin Streptomycin solution (Yu et al., 2020a). H1975 cells were cultured in RPMI 1640 Medium (Invitrogen, 11875-093) with 10% FBS, 1% Glutamax (Invitrogen, 35050061), and 1% Sodium Pyruvate 100 mM Solution (Invitrogen, 11360070) (Jiang et al., 2021). All cells were cultured in a humidified incubator at 37°C and 5% CO₂.

EGFR triple-mutant cells were generated by transfection with lentiviruses harboring the gene sequence encoding EGFR^{Del19/T790M/C797S} mutations in LLC cells. Construction of the overexpression lentiviral vector and concentration of viruses was performed by Shanghai Genechem Co., Ltd. (Shanghai, China). Based on the results of the pre-experiment, the optimal multiplicity of infection (MOI) for EGFR^{Del19/T790M/C797S} overexpression and empty vector lentivirus infection of LLC cells was 100. Before infection, LLC cells were plated in 6-well plates at 3 × 10⁵ cells per well overnight, and then the supernatant was replaced with 1 ml complete medium supplemented with lentivirus. After 14 h of infection, the lentivirus-containing medium was discarded and replaced with 2 ml fresh complete medium. After 72 h of infection, the stably transfected cell lines were selected by 2 μg/ml puromycin. The transfection efficiency was assessed by observing fluorescence intensity, and the EGFR^{Del19/T790M/C797S} overexpression levels were identified by real-time qPCR and western blotting.

LLC triple-mutant cells and H1975 cells were seeded at 1 × 10⁴ cells per well in 96-well plates for 24 h. Then, cells were treated with different concentrations of FYLM-containing serum. After 48 h of intervention, 100 µl medium containing 20% MTT (5 mg/ml, Solarbio, Beijing, China) was added to each well. After incubation at 37°C for 4 h, 150 µl dimethylsulfoxide (DMSO) was added to each well and shaken on a shaker at low speed for 10 min to dissolve the formazan crystals. In the end, the absorbance values of each well were measured at 490 nm (Lin et al., 2015) using a full-wavelength microplate reader (BioTek, United States).

Total RNA was extracted from cells or tumor tissues by Trizol reagent (Invitrogen, Thermo Fisher Scientific, United States). A reverse transcription reaction was performed using a 5x All-In-One RT MasterMix kit (Abm, Canada) to synthesize cDNA from total RNA. qPCR assays were detected by the SYBR Green PCR Master Mix kit (DBI Bioscience, Germany) in accordance with the manufacturer’s instructions on a real-time fluorescence quantitative PCR instrument (Roche LightCycler 480II, Mannheim, Germany). The relative expression of genes was measured using the 2^−△△CT method (Michaelidou et al., 2013), and GAPDH was used as a normalization control. Each sample was tested three times in duplicate. The primer sequences are shown in Table 1.

After treatments, cells or tumor tissues were lysed for 20 min on ice in RIPA lysis buffer (Solarbio, Beijing, China) supplemented with protease inhibitor (PMSF) and phosphatase inhibitor (APExBIO, MA, United States). And then the supernatant-containing proteins were collected by centrifugation at 12,000 rpm for 20 min at 4°C. Protein concentration measurements were performed using a BCA protein assay kit (Beyotime, Shanghai, China). Protein samples of 40 µg were isolated on 8% or 12% SDS-polyacrylamide gels and transferred onto the polyvinylidene fluoride (PVDF) membranes (0.45 µm, Millipore, Billerica, MA, United States). After blocking with 5% skim milk in TBST at room temperature for 2 h, membranes were incubated separately with the specific primary antibodies at 4°C overnight. The primary antibodies were as follows: EGFR (#4267, Cell Signaling Technology, 1:1000), p-EGFR (#3777s, Cell Signaling Technology, 1:1000), Cyclin B1 (#4138, Cell Signaling Technology, 1:1000), Bcl-2 (#3498, Cell Signaling Technology, 1:1000), Cleaved Caspase-3 (#9661, Cell Signaling Technology, 1:1000), PRC1 (BM3910, BOSTER, Wuhan, China, 1:1000), β-catenin (BA0426, BOSTER, Wuhan, China, 1:1000), c-Myc (PB9092, BOSTER, Wuhan, China, 1:1000), c-Jun (#9165, Cell Signaling Technology, 1:1000), Akt (ab179463, Abcam, 1:10000), p-Akt (ab192623, Abcam, 1:1000), β-actin (BA2305, BOSTER, Wuhan, China, 1:5000), GAPDH (A00227-1, BOSTER, Wuhan, China, 1:1000). After washing 3 times with TBST for 5 min each time, the membranes were incubated with secondary antibodies conjugated to horseradish peroxidase (BOSTER, Wuhan, China, 1:5000) for 1 h at room temperature. Finally, the bands were detected by an automated chemiluminescence gel imaging system (GE Amersham Imager600, United States), and the grayscale value was measured by ImageJ software (NIH, Bethesda, MD).

6-week-old male C57BL/6J mice, weighing 18–20 g, were purchased from Jinan Pengyue experimental Animal Breeding Co., Ltd. (Jinan, China) [animal License number: SCXK(Lu)2019-0003]. LLC cells stably transfected with EGFR^{Del19/T790M/C797S}-mutant or empty vector lentivirus were harvested and washed 2 times with PBS. Then, 1 × 10⁶ cells were suspended in 100 µl PBS and injected subcutaneously into the right side of the flank region of C57BL/6J mice. The equivalent dose was converted from human and mouse body surface area (9.1 times), and the dose administered to mice was 24 g/kg. Body weight and tumor volume of mice were recorded every 3 days, and tumor volume was represented by 0.5 × length × width² (mm³) (Chen et al., 2021). When the tumor volume reached about 200 mm³ (7 days after tumor inoculation) (Uchibori et al., 2017), the mice were randomly divided into seven groups (n = 5), empty vector control group (0.5% CMC-Na, Selleckchem, Houston, TX, United States, catalog number: S6703), overexpression control group (0.5% CMC-Na, Selleckchem, Houston, TX, United States, catalog number: S6703), 24 g/kg FYLM group, 48 g/kg FYLM group, osimertinib group (50 mg/kg, AZD9291, Selleckchem, Houston, TX, United States, catalog number: S7279), osimertinib +24 g/kg FYLM group (50 mg/kg osimertinib plus 24 g/kg FYLM) and osimertinib +48 g/kg FYLM group (50 mg/kg osimertinib plus 48 g/kg FYLM). After 2 weeks of continuous gavage, the tumor tissue was obtained and stored at −80°C or fixed with paraformaldehyde for further experiments.

The tumor tissues were dehydrated and embedded in paraffin, and then cut into 5 µm sections. Sections were stained with hematoxylin and eosin and then observed using an optical microscope.

Blood samples were collected from the mouse retro-orbital venous plexus. After resting at room temperature, the blood samples were centrifuged at 1,500 g for 30 min to obtain the supernatant. According to the manufacturer’s instructions, alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), creatinine (CRE), blood urea nitrogen (BUN) levels in serum samples were determined using the kit from Nanjing Jiancheng (Nanjing, China).

The tumor sections were dewaxed, antigen repaired in a microwave oven, and incubated with 3% H₂O₂ for 20 min at room temperature. The sections were incubated overnight at 4°C with an anti-Ki67 antibody (GB111499, Servicebio, 1:500) and an anti-β-catenin antibody (GB11015, Servicebio, 1:1000). Finally, the protein expression of Ki67 and β-catenin was observed under the optical microscope.

TUNEL staining of tumor sections was performed to detect apoptosis using the TUNEL assay kit (G1507, Servicebio, Wuhan, China) according to the manufacturer’s instructions. The paraffin sections were dewaxed, rehydrated, and incubated with proteinase K working solution for 20 min at 37°C. After incubation with 3% H₂O₂ at room temperature for 20 min, the sections were blocked with TUNEL reaction at 37°C for 1 h in a flat wet box. Positive expression was observed under an optical microscope.

The data were analyzed using SPSS 23.0 software (SPSS, Inc., Chicago, IL, United States). Mean ± standard deviation (SD) was used to reveal data. One-way ANOVA analysis followed by Fisher’s least-significant difference (LSD) was used to compare multiple-group statistical differences. A value of p < 0.05 was considered statistically significant.

To analyze the main components of FYLM, we used combination of Ultra-Performance Liquid Chromatography and Quadrupole-Orbitrap mass spectrometry (UPLC-Q-Orbitrap-MS) method. Chinese medicine samples matched 710 compounds in mzCloud. As shown in Table 2, 36 kinds of compounds were identified by screening with the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP, https://old.tcmsp-e.com/tcmsp.php). We found that the main components of huangqi include Choline, Betaine, Nicotinic acid, Coumarin, Chlorogenic acid, Caffeic acid, Ononin, Daidzein, Soyasaponin I, Linoleic acid. The main components of banzhilian include 4-Hydroxybenzaldehyde, Phenylacetaldehyde, Quercetin, Apigenin, Scutellarin, Eriodictyol, Baicalin, Hispidulin. The main components of baizhu include L-Histidine, DL-Arginine, L-Glutamic acid, L-Isoleucine, Uridine, L-Tyrosine. And the main components of baihuasheshecao include Geniposidic acid, 4-Methoxycinnamic acid, Rutin. The total ion chromatograms of FYLM were shown in Figure 1A (negative ion mode) and Figure 1B (positive ion mode). Figure 1C manifested the chemical structure formula corresponding to the typical peaks.

To determine whether EGFR^{Del19/T790M/C797S} is stably expressed in LLC cells, we analyzed fluorescence intensity, mRNA and protein levels of EGFR. By fluorescence microscopy, both overexpression and empty vector lentiviruses-transfected LLC cells exhibited enhanced fluorescence signals, indicating normal expression of fluorescent marker genes and successful transfection (Figure 2A). Based on the real-time qPCR results (Figure 2C), the EGFR gene expression in the overexpression group was 37-fold higher than that in the empty vector group. Similarly, the protein levels of EGFR were significantly higher in the overexpression group (Figures 2B, D). The above results suggested that LLC cells were successfully transfected with EGFR^{Del19/T790M/C797S} expressing lentivirus. Next, cell viability was measured by treating with various concentrations of FYLM-containing serum for 48 h in EGFR^{Del19/T790M/C797S}-mutated and EGFR^L858R/T790M-mutated cells. As shown in Figure 2E, FYLM-containing serum dose-dependently inhibited the proliferation of LLC-triple mutant cells (IC₅₀ = 14.27%). In addition, FYLM-containing serum can inhibit the proliferation of H1975-L858R/T790M cells (Figure 2F, IC₅₀ = 17.72%). In order to further verify the effect of FYLM-containing serum on LLC-triple mutant cells, we detected the protein expression of p-EGFR by western blotting. The results suggested that low and high doses of FYLM-containing serum can reduce EGFR phosphorylation compared with the control group. In addition, the combination of FYLM-containing serum and osimertinib had a more significant inhibitory effect compared with the osimertinib alone group. However, there was no significant differences between high and low doses (Figures 2G–I).

To explore the inhibitory effect of FYLM on tumor progression in vivo, we first constructed a xenograft mouse model harboring EGFR-Del19/T790M/C797S-mutant via subcutaneously inoculating LLC-triple mutant cells into C57BL/6J mice. Tumor-bearing mice were continuously orally gavage for 2 weeks with 0.5% CMC-Na, FYLM (24 g/kg/d), FYLM (48 g/kg/d), osimertinib (50 g/kg/d), a combination treatment of osimertinib (50 g/kg/d) and FYLM (24 g/kg/d), or a combination treatment of osimertinib (50 g/kg/d) and FYLM (48 g/kg/d). As a result, the body weight of mice in each group steadily increased without significant differences, and no obvious toxicity was observed compared to the control treatment group (Figures 3B, G). The tumor weight and tumor volume of the empty vector group were slightly smaller than those of the overexpression group, but the differences were not statistically significant (Figures 3C–E). FYLM or osimertinib inhibited tumor progression compared with the control group. Among them, the combination of FYLM (48 g/kg/d) and osimertinib had the most significant inhibitory effect (Figures 3H–J). Furthermore, to evaluate the safety of combination treatment with FYLM and osimertinib, we measured spleen weight and biochemical analysis of liver function [alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP)] and kidney function [creatinine, blood urea nitrogen (BUN)]. The results showed no significant differences among the groups. Therefore, the combination of FYLM and osimertinib has no significant toxicity (Figures 4A–F).

HE staining revealed a decrease in nuclear density, lighter staining and lower nucleoplasm ratio in the combined treatment group (Figure 5A). To investigate the effect of FYLM on tumor growth, we detected the levels of cell proliferation, cell cycle, and apoptosis-related protein Ki-67, cyclin B1, Bcl-2 and cleaved Caspase-3 in xenograft mouse tumor tissues by western blot and immunohistochemistry. Consistent with the results of phenotypic experiments, the combination of FYLM and osimertinib reduced the positive expression of Ki-67 protein in a dose-dependent manner in the immunohistochemistry assay (Figure 5B). The western blot results showed that the combination of FYLM and osimertinib downregulated the protein expression level of cyclin B1 (Figures 5C, D). Furthermore, TUNEL staining results illustrated that combination of osimertinib and FYLM (24 and 48 g/kg) increased apoptosis (Figure 6A). Moreover, western blot assays revealed that combination treatment of osimertinib and FYLM (24 and 48 g/kg) inhibited Bcl-2 expression compared with osimertinib or FYLM alone. However, cleaved Caspase-3 protein expression were upregulated in combination treatment of osimertinib and FYLM (24 g/kg) group and the combination treatment of osimertinib and FYLM (48 g/kg) group compared with the osimertinib alone group (Figures 6B–D).

To clarify the mechanism synergistic anti-cancer effect of FYLM, we investigated the protein expression of PRC1, Wnt pathway-related protein (β-catenin, c-Myc, c-Jun) and EGFR pathway-related protein (p-EGFR, EGFR, p-Akt and Akt) in xenograft tumor tissues by western blot and immunohistochemistry. Immunohistochemistry assays showed a decrease of β-catenin protein in both the combination treatment of osimertinib and FYLM (24 g/kg) group and the combination treatment of osimertinib and FYLM (48 g/kg) group (Figure 7A). Western blot assays revealed that combination treatment of osimertinib and FYLM (24 and 48 g/kg) suppress PRC1, β-catenin, c-Myc and c-Jun expression relative to the treatment with osimertinib or FYLM alone (Figures 7B–G). The Wnt pathway and the EGFR pathway are key pathways related to drug resistance, and they crosstalk each other. Inhibition of Wnt pathway and EGFR pathway can delay drug resistance. No significant changes in EGFR and Akt expression in the combination treatment group compared to other individual treatment groups in terms of real-time qPCR and immunoblotting of mouse tumor tissues (Figures 8A, B, E, G). However, the combination of osimertinib and FYLM (24 and 48 g/kg) reduced the expression of p-EGFR/EGFR and p-Akt/Akt ratios (Figures 8C, D, F). These results suggest that the combination of FYLM and osimertinib inhibits the PRC1/Wnt/EGFR pathway, which may be the mechanism of FYLM alleviating TKI resistance.