Lung cancer organoids were derived from surgery samples or transbronchial biopsies of lung cancer patients at Beijing Chest Hospital, Capital Medical University, Beijing, China. The study has got the approval of the Ethical Committee of Beijing Chest Hospital, Capital Medical University (Trial No. 11, 2020). Patients participating in this study signed informed consent forms. On arrival, tumor tissues were washed with cold PBS, cut into small pieces, washed with Advanced DMEM/F12 (Thermo Fisher Scientific, Waltham, MA, United States; containing 1× Glutamax, 10 mM HEPES and antibiotics), and digested with collagenase (Sigma-Aldrich, Cat #C9407, 2 mg/ml) for 1–2 h at 37°C. After washing twice with fresh medium (2% fetal calf serum, FCS) and centrifugation (400 rcf, 4 min), dissociated cells were seeded into growth factor reduced Matrigel (Corning, Cat # 356252) with the presence of Advanced DMEM/F12 at 37°C for 30 min. Next, the surface of the solidified mixture of cell suspension/Matrigel was sealed with complete human organoid medium (HOM, 500 μL), which comprised Advanced DMEM/F12 supplementing with series additives as described by Lampis et al. (2017) and Puca et al. (2018), replacing every 3 days. When the organoids ranged up to 200–500 μm in diameter (about 1 week), organoids were dissociated and passaged weekly using TrypLE Express (Gibco, Grand Island, NY, United States). The PDOs (2 × 10⁶ cells/tube, P3) were frozen using the Recovery Cell Culture Freezing Medium (Gibco) and stored at −80°C before drug screening.

A collection of almost 1,100 natural products were obtained from MedChemExpress (Shanghai, China). The natural product library was reformatted into 96-well source plates with concentration of 3.3 mM for automated robotic screening. At parallel, the cells were also treated with an equal volume (0.1%) of DMSO as a negative control and 1 μM final of staurosporine (MCE, shanghai, China) as a positive control. Plate-to-plate normalization and assay quality control were calculated according to them. A 3D cell viability assay was implemented, which determines the number of cell viability according to the ATP level using commercially available luminescence detection reagent (CellTiter-Glo 3D, #G9683, Promega, Madison, WI). In brief, organoids were processed as described earlier and plated in a 96-well low binding assay plate at a density of 6,000 cells per well in 50 μL comprising 10% growth factor reduced Matrigel. Additional 40 μL culture medium without Matrigel was added. Organoids were maintained in medium described earlier and drugged 2 days later by adding 10 μL culture medium comprising 33 μM natural product to get a final concentration of 3.3 μM (Supplementary Figure S1A). The assay was terminated at day 5 by adding 50 μL CellTiter-Glo. Assay quality and robustness were evaluated with the signal window (SW) and Z factor. Triplicate wells treated with staurosporine and vehicle solution (DMSO) were employed as bottom wells and top wells, respectively (Supplementary Figure S1B). The assay showed the signal windows (SW) were much larger than 10, and the Z factor values were between 0.5 and 1, which indicated that the assay was qualified for high-throughput screening (Supplementary Figures S1C,D).

Human lung cancer cell lines NCI-H1299 (ATCC Cat# CRL-5803, RRID: CVCL_0060) and NCI-H460 (ATCC Cat# HTB-177, RRID: CVCL_0459) were purchased from the American Type Culture Collection (ATCC; Manassas, VA, United States). Cells were maintained in RPMI-1640/1641/1642 medium (Gibco) supplemented with 10% FCS (Gibco) and 1% penicillin–streptavidin (Gibco) at 37°C in 5% CO₂.

NCI-H460 (0.75 × 10³ cells/well) and NCI-H1299 (0.75 × 10³ cells/well) cell lines were seeded into 96-well plates and treated with vehicle or HF for 1, 3, and 5 days. After incubation, cell viability was detected using luminescence detection reagent (CellTiter-Glo, #G9243, Promega, Madison, WI). The CellTiter-Glo assay determines the number of viable cells in culture by quantifying ATP, which indicates the presence of metabolically active cells. Luminescence readout is directly proportional to the number of viable cells in culture. As for the foci assay, NCI-H460 (2 × 10³ cells/well) and NCI-H1299 (2 × 10³ cells/well) were seeded into 6-well plates, treated with HF as the way of cell viability assay, then fixed and stained the cells with crystal violet solution, and took photos with a camera and bright-field microscope. Each vial of frozen cells was thawed and maintained for a maximum of 10 passages.

The cell cycle and apoptosis were detected as previously described (Shi et al., 2018). Cells were cultured and treated with DMSO and HF (0.05 and 0.2 μM) in both NCI-H460 and NCI-H1299 for 24 h, followed by single staining with PI (Beyotime) for cell cycle analysis and double staining with PI and Annexin V-FITC (Beyotime) for apoptosis analysis. Data analysis was performed using NovoExpress v1.3.4.

A standard Western blot analysis of whole-cell protein lysates was performed using primary antibodies against cleaved PARP, PARP (Cell Signaling Technology, #9542, 1:1,000, RRID:AB_2160739), cleaved caspase-3 (Cell Signaling Technology Cat# 9661, 1:1,000, RRID:AB_2341188), caspase-3 (Cell Signaling Technology Cat# 9662, 1:1,000, RRID:AB_331439), cylclinD1 (Cell Signaling Technology Cat# 2978, 1:1,000, RRID:AB_2259616), p27 (Cell Signaling Technology Cat# 3686, 1:1,000, RRID:AB_2077850), p21 (Cell Signaling Technology Cat# 2947, 1:1,000, RRID:AB_823586), pRb (Cell Signaling Technology Cat# 8516, 1:1,000, RRID:AB_11178658), and Rb (Cell Signaling Technology Cat# 9309, 1:1,000, RRID:AB_823629) to check the changes of apoptosis and G0/G1 phase markers. As well, for the alteration of effectors, the signaling pathways were examined with the primary antibodies including p-AKT1/2 (Cell Signaling Technology Cat# 4060, 1:1,000, RRID:AB_2315049), AKT1/2 (Proteintech Cat# 10176-2-AP, 1:2000, RRID:AB_2224574), p-ERK (Cell Signaling Technology Cat# 4370, 1:1,000, RRID:AB_2315112), and ERK (Proteintech Cat# 16443-1-AP, 1:2000, RRID:AB_10603369). Equal amounts of protein, which were blotted with an anti-β-actin antibody (Proteintech Cat# 60008-1-Ig, 1:2000, RRID: AB _2289225), was used as loading control.

NCI-H460 and NCI-H1299 cells were incubated with DMSO or HF (0.05 and 0.2 μM) for 48 h; after harvest, total RNA was isolated using the TriZolTM UP Plus RNA Kit. RNA was sent to BGI (Beijing, China) for sequencing and analysis. In brief, after total RNA was fragmented into short fragments, mRNA was enriched using oligo (dT) magnetic beads, followed by cDNA synthesis. Double-stranded cDNA was purified and enriched by PCR amplification, after which the library products were sequenced using BGIseq-500. The heatmap of DEGs ( log 2 ⁡ F C ≥ 1, p ≤ 0.001) and KEGG analysis ( log 2 ⁡ F C ≥ 1, p ≤ 0.05) in NSCLC cell lines were performed by the BGI using the Dr. TOM approach, a customized data mining system from BGI. Altered (upregulated or downregulated) expression of genes was expressed as log 2 ⁡ F C , which represents log-transformed fold change ( log 2 ⁡ F C = log 2 [ B ] − log 2 [ A ] , while A and B represent values of gene expression for different treatment conditions).

NCI-H1299 and NCI-H460 cells were plated in 96-well plates and treated with various concentrations of cisplatin or/and HF, either alone or in combination, for 72 h. Cell viability was determined as described before. The synergy effect was evaluated by measuring the IC₅₀ and highest single agent (HSA) reference model (Yadav et al., 2015). In addition, Western blotting analysis and flow cytometry assay were performed to detect the effect of drug combination on signaling pathways and cell cycle arrest.

Cisplatin-resistant PDOs (PDO-R1 and PDO-R2) were derived from two cisplatin-resistant patients and were plated in 96-well plates and treated with various concentrations of HF and cisplatin. Synergistic effects were observed under a microscope, and Western blotting assay was performed to detect the effect of drug combination on signaling pathways.

Data statistical analysis was performed using GraphPad Prism 8.0. The HF IC₅₀ values were analyzed using non-linear regression (curve fit). The cell cycle and apoptosis data were analyzed using Excel using Student’s t test. A p value < 0.05 was considered as statistically significant. All data subjected to statistical evaluations were gathered with at least three independent repeats of experiments.

PDOs have been reported for applications in preclinical drug discovery (Raja et al., 2016; Huang et al., 2017; Choudhury et al., 2020; Maloney et al., 2020; Skardal et al., 2020). A Two-stage screening strategy was employed whereby almost 1,100 natural products were primarily screened in 1 PDO and then validated in six PDO models which completed by K2 Oncology Co., Ltd. (Beijing 100061, P.R. China). The screening process is illustrated in Supplementary Figures S1A–D, and the results showed that HF is exactly one of the top hits with strong anticancer potential (Supplementary Figure S1E). Then we performed further anticancer effect verification of HF in two cisplatin-resistant lung cancer PDOs. The clinical information and data of these two PDOs are summarized in Figure 1A. Both patients were treated with the platinum-based regimen, and the RECIST evaluation was performed after two cycles (Figure 1B). The cisplatin resistance were maintained after PDOs were successfully established (Figures 1C,D).

Next, we verified the effect of HF in lung cancer cell lines, NCI-H1299 and NCI-H460. NCI-H460 showed partial cisplatin sensitivity, but the maximal inhibition rate was very limited (Figure 4B). NCI-H1299 is a cisplatin-resistant lung cancer cell line (Fan et al., 2020) (Chen M. et al., 2021), and its cisplatin resistance is also validated in this study (Figure 4D). The dose–response curve of HF in both cell lines exhibited identical IC₅₀ (0.07 μM) (Figures 2A,B). The significant dose-dependent and time-dependent inhibition of HF on lung cancer cell proliferation were observed in both cell lines (Figures 2C,D). In addition, HF remarkably suppressed colony formation of the two cell lines in a dose-dependent manner (Figures 2E–H). Thus, these results indicated a remarkable growth inhibition of HF in lung cancer cell lines.

To explore the reasons for growth inhibition of lung cancer cells by HF, cell cycle distribution, and apoptosis induction were evaluated on HF-treated lung cancer cell lines. Through flow-cytometry assay, we noticed a dose-dependent G0/G1 phase arrest (Figures 3A–D) in both NCI-H1299 and NCI-H460 following 24-h treatment of HF. Additionally, HF induced dose-dependent apoptosis in both cell lines (Figures 3E–H). In detail, both the early and late apoptosis were significantly induced by HF (Supplementary Figure S2). Cyclin D1 (Sherr, 1996; Lukas et al., 1996) and p21 (Cheng et al., 1999; Pestell et al., 1999) are two canonical cell cycle markers for the G1 phase. Western blotting confirmed G0/G1 phase arrest with the decreased cyclin D1 expression and increased p21 expression in accordance with the flow cytometry tendency (Figure 3I). PARP (Lu et al., 2019) and caspase (Julien and Wells, 2017) activation are the key events of apoptosis, and the expression levels of their full-length and cleaved forms were examined with Western blotting. The decreased PARP and caspase 3 and the increased cleaved PARP and cleaved caspase3 suggested that HF induced apoptosis in a dose-dependent manner as well (Figure 3J). Taken together, HF induced G0/G1 phase arrest and apoptosis in a dose-dependent manner in NCI-H460 and NCI-H1299, which well explained the significant lung cancer cells growth inhibition.

To further investigate the impact of HF on cisplatin resistance in lung cancer, lung cancer cell lines, NCI-H1299 and NCI-H460, were employed for the dural drug synergy test. NCI-H1299 shows resistance to cisplatin, with IC₅₀ larger than 100 mM and maximal inhibition rate smaller than 10%. NCI-H460 showed limited response to cisplatin, with maximal inhibition rate of about 20%, which induced the absolute IC₅₀ larger than 10 mM and indicated a very low efficacy (Figures 4B,D). The synergy effect was evaluated by measuring the IC₅₀ and the highest single agent (HSA) reference model (Yadav et al., 2015). With the existence of HF, significant left shift of dose–efficacy curve and decrease in IC₅₀ were observed (Figures 4B,D). By adding 5, 14, and 41 nM of HF, the IC₅₀ of cisplatin decreased from larger than 10 mM to 8, 2.4, and 0.4 μM for NIC-H460 and 9, 2.5, and 0.5 μM for NCI-H1299, respectively. Consistently, by adding serial diluted cisplatin to HF, the IC₅₀ of HF also decreased, respectively (Figures 4A,C). To further analyze the synergy effect of cisplatin with HF, synergy score matrixes were calculated with the highest single agent (HSA) reference model. The average synergy score of cisplatin with HF is 21.585 and 24.851 for NCI-H1299 and NCI-H460, respectively. The HSA synergy scores were visualized with heatmap (red areas in the model graph) and 3D hillmap (Figures 4E–H). These data suggest the synergic effect of HF and cisplatin in lung cancer cell lines.

To further explore the potential value and the underlying synergy mechanism of HF with cisplatin, 1 and 10 μM of cisplatin and 50 and 200 nM HF were selected for cell cycle, apoptosis, and pathway markers detection using Western blot in NCI-H460 and NCI-H1299. When HF was combined with 1 μM of cisplatin, significant G0/G1 phase arrest was observed with the increase in p21 and p27 and the decrease in p-Rb and cyclin D1 in a dose-dependent manner (Figure 5 upper). HF combined with 10 μM of cisplatin showed similar trend but a stronger dose-dependent synergistic effect than low-dose cisplatin. In addition, the apoptosis markers also showed similar synergy effect with the indicator of caspase 3 and cleaved caspase 3 (Figure 5 middle). Similar tendency of G0/G1 phase arrest was further approved using flow cytometry with cell models treated with cisplatin alone, HF alone, and cisplatin combined with HF (Figures 6A–C). Obviously, cisplatin alone barely changed the cell cycle progression, but HF could induce prominent G0/G1 phase arrest at 200 nM. Altogether, these data demonstrated that HF would be a potential cisplatin sensitizer via G0/G1 phase arrest and apoptosis.

RNA sequencing was performed to profile gene expression in both NCI-H460 and NCI-H1299 with the treatment of HF at 0.05 and 0.2 μM, respectively. (The datasets presented in this study can be found in online repositories. The names of the repository and accession number can be found below: http://www.ncbi.nlm.nih.gov/sra, accession number PRJNA769938). In both cell lines, >4000 genes were differentially expressed under the two concentrations. We observed substantial overlap between the two cell lines with regard to the genes that were differentially expressed following incubation with HF-1798 genes (DEGs, log₂FC ≥ 1, p ≤ 0.001) were shared in NCI-H460 and NCI-H1299 with the treatment of different concentration of HF (Figure 7A). In addition, HF led to alterations of 5478 DEGs in NCI-H460 cells and 3523 DEGs in NCI-H1299 cells versus each corresponding control group (Figures 7B,C). The KEGG analysis of the DEGs revealed enrichments of cancer-associated pathways, including PI3K/AKT and MAPK signaling pathways (Figures 7D,E). Moreover, the genes related with PI3K/AKT and MAPK signaling pathways were sensitive to HF exposure (Figures 7F,G,I,J).

Upregulation of p-AKT might enhance tumor progression and mediate resistance to drugs (Liu R. et al., 2020), and it has been well known that the binding of ligand to a transmembrane receptor, receptor tyrosine kinase (RTK), activates both PI3K/AKT and MAPK pathways to promote cell survival (Cao et al., 2019). Herein, we hypothesized that dual inhibition on the two pathways would exert a more stable and stronger growth and survival suppression than targeting individual pathway. The ratio of p-AKT (S473) to total Akt examined by Western blot was decreased by HF in a dose-dependent manner (Figure 5 lower). Here, the observed downregulation of p-Akt in our cisplatin-resistant lung cancer cell lines by HF, indicating possible tumor suppression. Together, HF inhibited the cisplatin-resistant cell models growth via blocking the PI3K/AKT signaling pathway.

The identical analysis and assays as before were performed to examine the effect of HF on the MAPK signaling pathway, which includes a small G protein (Ras) and three protein kinases (Raf, Mek, and Erk) and is activated with translocation of Erk (MAPK) to the nucleus (McCain, 2013). In protein level, p-Erk/Erk examined by Western blot exhibited dose-dependent decrease with the exposure of HF (Figure 5 lower). Activation of the MAPK signaling pathway would strengthen tumor progression and mediate drug resistance as well (Cocco et al., 2019; Li et al., 2019). The decline of p-Erk1/2 indicated the potential activity of HF in lung cancer suppression. Thus, HF inhibited the cell models growth simultaneously via blocking the MAPK signaling pathway.

To study the response to the combination treatment of cisplatin and HF, the efficiency of tumor destruction was observed under a microscope, and exposure of PDOs to HF and combination of cisplatin and HF led to substantially reduced survival (Figure 8A). In addition, significant p-AKT (S473) and p-ERK decrease via Western blot appeared in cisplatin and HF combination groups compared to single drug treatment groups or control group (Figures 8B–E). Therefore, the results suggested that HF had the capacity to expand its sensitizer effect for cisplatin to preclinical cisplatin-resistant PDO models.