Human PC9 and HCC827 NSCLC cell lines were generously provided by Dr Jin Kyung Rho (Asan Medical Center, Ulsan University). The cells were cultured in RPMI-1640 medium (HyClone; Cytiva) supplemented with 10% fetal bovine serum (FBS) and antibiotics (100 U/ml penicillin and 100 µg/ml streptomycin) in a humidified incubator with 5% CO₂. All cell lines used in the present study were authenticated by short tandem repeat profiling (KOMA Biotech, Inc.) and were routinely screened for microbial contamination as part of standard quality control procedures. The TRPML3-GCaMP6 construct, generated by inserting the full-length GCaMP6 sequence into the p3XFLAG-CMV-7.1-TRPML3 vector (17), was kindly provided by Dr Hyun Jin Kim (Sungkyunkwan University). The pCMV-TRPML3 expression vector containing wild-type TRPML3 (NM_018298) was obtained from OriGene Technologies, Inc. Reagents and their sources were as follows: osimertinib (Selleck Chemicals), Cell Fractionation Kit (Cell Signaling Technology, Inc.), anti-TFEB (cat. no. 4240s), anti-Nanog (cat. no. 3580s; both from Cell Signalling Technology, Inc.), TRPML agonist ML-SA1 (MilliporeSigma; cat. no. SML0627), TRPML inhibitor ML-SI1 (GlpBio; cat. no. GC19764), anti-Myc (cat. no. sc-40; Santa Cruz Biotechnology, Inc.), anti-CD44 (cat. no. MA5-13890; Thermo Fisher Scientific, Inc.), anti-CD133 (cat. no. ab16048), anti-lamin-B1 (cat. no. ab19898; both from Abcam), anti-TRPML1 (cat. no. MBS9205163; MyBioSource, Inc.), anti-TRPML3 (cat. no. 13879-1-AP; Proteintech Group, Inc.) and anti-GAPDH (cat. no. bs2188R; BIOSS). For transient gene silencing, synthetic small interfering RNAs (siRNAs) targeting TRPML3 and scrambled control siRNAs were purchased from Integrated DNA Technologies, Inc. Their sequences were as follows: TRPML3 siRNA sense, 5′-GUAGAAUUUACCUCUACUCAUUCAT-3′ and antisense, 3′-AUCAUCUUAAAUGGAGAUGAGUAAGUA-5′; scrambled siRNA sense, 5′-CGUUAAUCGCGUAUAAUACGCUAT-3′ and antisense, 3′-AUACGCGUAUUAUACGCGAUUAACGAC-5′. The siRNAs and expression vectors were introduced using Lipofectamine RNAiMAX and Lipofectamine 3000, respectively, according to the manufacturer's protocols (Thermo Fisher Scientific, Inc.). In brief, each mixture of siRNA (30 nM per 35-mm dish) with RNAiMAX and expression vectors (2.5 µg per 35-mm dish) with Lipofectamine 3000 was incubated for 5 min and 15 min, respectively, at 24°C. Each sample was then subjected to subsequent experiment following an additional incubation of 72 and 24 h, respectively. To generate a stable TRPML3 knockout cell line, an sgRNA targeting TRPML3 (RefSeq NM_018298.11) was designed and synthesized by Macrogen, Inc. (5′-CAGCTACTCACAACTACCTCAGG-3′). It was inserted into the pSp-U6-Cas9-2A-Puro plasmid; cells were transfected with either the TRPML3-specific construct or the empty vector control. After puromycin selection (2–10 µg/ml) for two weeks, the TRPML3 knockout was confirmed using western blotting.

Resistant sub-lines of PC9 and HCC827, designated as PC9/OR and HCC827/OR, respectively, were generated by gradually increasing osimertinib concentrations (1×10⁻⁴, 1×10⁻³, 1×10⁻², and 1×10⁻¹ µM) over five months. Surviving cells were cultured in 1×10⁻¹ µM osimertinib for an additional month to confirm the stability of resistance. Viability at each subculture step was assessed using 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide (MTT) assays.

To measure gene expression changes in lung adenocarcinoma (LUAD) tissues during the development of osimertinib resistance, a publicly available RNA-seq dataset (GSE253742) from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) was analyzed, representing paired LUAD tumor samples collected before and after osimertinib treatment (22). Data were processed using Expression Differential Expression Gene Analysis (ExDEGA v5.1.1.1) software (Ebiogen Inc.). Briefly, raw data were subjected to quality control with FastQC, followed by adapter trimming and low-quality read removal using Fastp (23). Processed reads were aligned to the reference genome with STAR (24), and quantified using Salmon (25). Read count normalization was performed using the TMM + CPM method in EdgeR (26). Data mining and visualization were performed using ExDEGA and Multi Experiment Viewer (MeV 4.9.0; http://www.tm4.org) (27).

Total RNA was isolated from cell lines using TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.), following the manufacturer's protocol. Subsequently, 1 µg of RNA was converted into cDNA using the High-Capacity RNA-to-cDNA kit (Applied Biosystems; Thermo Fisher Scientific, Inc.), following the manufacturer's protocol. Quantitative PCR (qPCR) was conducted using VeriQuest SYBR Green qPCR Master Mix (Affymetrix; Thermo Fisher Scientific, Inc.) on an ABI PRISM 7900 Sequence Detection System (Applied Biosystems; Thermo Fisher Scientific, Inc.). The thermal cycling conditions were as follows: initial denaturation at 95°C for 20 sec, followed by 40 cycles of denaturation at 95°C for 3 sec and annealing/extension at 60°C for 30 sec. Primers were as follows: TRPML3 forward, 5′-TCCGTTGGGAATCATGCTTAT-3′ and reverse, 5′-AGTGCTGACAGATTGCCATAGC-3′; GAPDH forward, 5′-ATGGAAATCCCATCACCATCTT-3′ and reverse, 5′-CGCCCCACTTGATTTTGG-3′. GAPDH expression was used to normalize data and controls for variations in mRNA concentration. Expression levels were calculated using the 2^−ΔΔCq method (28).

To generate 3D spheroid cultures, cells were resuspended in culture medium supplemented with threefold the standard FBS concentration at a density of 6×10⁴ cells/ml for PC9 and its subline and 4.2×10⁴ cells/ml for HCC827 and its subline. Cell suspensions were mixed with VitroGel Hydrogel Matrix (Well Bioscience Inc.; http://www.thewellbio.com) at a 2:1 (v/v) ratio; the 50 µl of the mixture was added per well in U-bottom 96-well plates (S-bio Sumitomo Bakelite Co., Ltd.). Plates were centrifuged at 1,000 × g for 5 min at 24°C, and 200 µl of regular culture medium was added per well. Aggregates were incubated in the hydrogel matrix at 37°C in a 5% CO₂ atmosphere for three days to form multilayered spheroids. These spheroids were treated with osimertinib or left untreated (DMSO vehicle only) for an additional 9 days, with medium refreshed every 3 days. Spheroids were imaged using a NIKON Eclipse TS100 microscope equipped with an FL-20BW CMOS camera (Tucsen Photonics Co., Ltd.). Images were captured with Mosaic 2.4 software (Tucsen Photonics Co., Ltd.) and analyzed using ImageJ software (version 1.51k; National Institutes of Health) to measure spheroid diameters and other parameters. Viability was measured using the Cell Counting Kit-8 (CCK-8) assay kit (Dojindo Laboratories, Inc.), according to the manufacturer's protocol. In brief, 20 µl of CCK-8 solution was added to each well and incubated at 37°C for 1 h, after which the absorbance was measured at 450 nm using an iMark microplate reader (Bio-Rad Laboratories, Inc.).

Western blotting was performed using established methods (29). Cell lysates prepared in RIPA buffer (Invitrogen; Thermo Fisher Scientific, Inc.) were sonicated for 5 sec at 50% amplitude and incubated on ice for 10 min. Debris was pelleted at 14,440 × g for 10 min at 4°C. Supernatant protein concentrations were quantified using the BCA assay. Equal amounts of protein (20 µg per sample) were separated using 8% SDS-PAGE, followed by transfer to 0.2-µm polyvinylidene difluoride membranes (Cytiva). The membrane was blocked with 5% bovine serum albumin (GenDEPOT, LLC) in Tris-buffered saline with 0.1% Tween 20 at 4°C for 2 h. Primary antibodies against TRPML1, TRPML3, TFEB, lamin-B1, Myc, CD44, CD133, Nanog and GAPDH were used at a 1:1,000 dilution. The membranes were incubated overnight at 4°C with the primary antibodies, followed by incubation with horseradish peroxidase (HRP)-linked secondary IgG antibody (1:2,500; Santa Cruz Biotechnology, Inc.; cat. no. sc-516102 for anti-mouse, and cat. no. sc-2357 for anti-rabbit) for 2 h at 4°C. Protein bands were visualized and quantified using the AzureSpot 2.0 software (Azure Biosystems, Inc.). Whole-cell lysate from spheroids cultured in VitroGel Hydrogel Matrix (Well Bioscience, Inc.) were extracted using VitroGel Organoid Recovery Solution (Well Bioscience, Inc.) according to the manufacturer's protocol. Briefly, spheroids were transferred to 1.5-ml tubes, washed with PBS, and resuspended in 1 ml of pre-warmed (37°C) VitroGel Organoid Recovery Solution. The solution was gently pipetted to isolate the spheroids, followed by a brief incubation at 37°C. After centrifugation at 100 × g for 5 min, pellets were collected and lysed in RIPA buffer. To measure TFEB translocation from the cytoplasm to the nucleus, cytosolic and nuclear fractions were isolated using a Cell Fractionation Kit (Cell Signaling Technology, Inc.), following the manufacturer's instructions. Briefly, cell pellets were resuspended in Cytoplasm Isolation Buffer and incubated at 4°C for 5 min. Suspensions were centrifuged at 500 × g for 5 min at 4°C, and cytosolic supernatants were collected. Nuclear pellets were resuspended in Membrane Isolation Buffer, re-pelleted at 8,000 × g for 5 min at 4°C, and resuspended in Cytoskeleton/Nuclear Isolation Buffer. Both fractions were analyzed using western blotting.

Cells were seeded at a density of 7×10⁵ per 60 mm culture dish. Following incubation under the designated conditions, they were detached using 0.05% Trypsin-EDTA (Gibco; Thermo Fisher Scientific, Inc.) and fixed in 70% ethanol at −20°C for 2 h, then stained with 0.5 ml of propidium iodide/RNase staining buffer (BD Biosciences) for 15 min. Flow cytometry was performed on at least 20,000 cells using a FACScan analyzer (BD CellQuest^TM Pro version 5.2; BD Biosciences).

Cell viability was assessed using the Viability Assay Kit (Cellrix; MediFab) following the manufacturer's protocol. Cells were seeded at a density of 1×10⁴ per well in 96-well plates, treated with 10 µl of assay reagent containing water-soluble tetrazolium salt, and incubated at 37°C for 30 min. The production of water-soluble formazan was evaluated by measuring the absorbance at 450 nm using an iMark microplate reader (Bio-Rad Laboratories, Inc.).

Intracellular Ca²⁺ levels were measured by incubation with Fura-2/AM (MilliporeSigma) at 37°C for 50 min, followed by perfusion with a HEPES-buffered solution containing 140 mmol/l NaCl, 5 mmol/l KCl, 1 mmol/l MgCl₂, 1 mmol/l CaCl₂, 10 mmol/l HEPES, and 10 mmol/l glucose, adjusted to pH 7.4 and 310 mOsm. For Ca²⁺-free conditions, the CaCl₂ was replaced with 1 mmol/l EGTA. Osimertinib and other compounds, including ML-SI1 and MS-SA1, were diluted in a Ca²⁺-free HEPES buffer. Fluorescence excitation was performed at 340 and 380 nm, with emission at 510 nm-captured using an sCMOS camera (Andor Technology Ltd.). Ratio-metric data (F₃₄₀/F₃₈₀) were analyzed using Meta-Fluor Fluorescence Ratio Imaging software (Molecular Devices, LLC). GCaMP6 fluorescence intensities were measured at an excitation wavelength of 488 nm, with emission at 510 nm measured using an sCMOS camera (Andor Technology Ltd.) and analyzed using Meta-Fluor Fluorescence Ratio Imaging software (Molecular Devices, LLC).

Lysosomal pH was measured using a LysoSensor Yellow/Blue DND-160 (Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. Cells cultured in 35 mm confocal dishes (SPL Life Sciences) were incubated with the dye for 10 min at 37°C. After a brief wash with HEPES buffer, cells were continuously exposed to osimertinib diluted in HEPES buffer. Fluorescent images were captured using an sCMOS camera (Andor Technology Ltd.) with excitation at 329 and 384 nm and emission at 440 and 540 nm, respectively. Data were analyzed using MetaMorph software (Molecular Devices, LLC), and lysosomal pH was determined by calculating the fluorescence intensity ratio of 384 nm (yellow, acidic) to 329 nm (blue, less acidic), with fold changes expressed relative to the control (scr-cells pre-osimertinib treatment).

The lysosomal area in individual cells was measured using LysoTracker Red DND-99 (Thermo Fisher Scientific, Inc.), according to the manufacturer's instructions. Cells were plated in 35 mm confocal dishes and transfected with either siML3 or scrambled RNA as a control. After 72 h, the cells were treated with 0.01 µM osimertinib, incubated for 24 h, and stained with LysoTracker Red DND-99 for 30 min. Fluorescence was excited at 577 nm and emitted at 590 nm, and images were captured using a confocal microscope (FV1000; Olympus Corporation). Image analysis was performed using ImageJ software (version 1.52v; National Institutes of Health).

Data were analyzed using Origin 2020 software (OriginLab Corporation). Results are presented as the mean ± standard deviation (SD) based on a minimum of three independent experiments. Statistical significance was evaluated using one-way analysis of variance (ANOVA), followed by Tukey's post-hoc test for multiple comparisons. P<0.05 was considered to indicate a statistically significant difference.

To identify genes involved in osimertinib-induced lysosomal pH dysregulation, the GEO dataset GSE253742 was analyzed, which includes RNA-seq data from eight clinical LUAD specimens (four naive tumors and four osimertinib-residual tumors). Of the 40,929 genes analyzed, TRPML3 was the only gene whose expression was significantly increased (5.121±0.045; |fold change|≥2, log2 ≤3, P<0.05) in osimertinib-residual tumors within a subset of 26 genes related to lysosomal acidification (GO:0007042) (Fig. 1A). Based on these findings, the expression was next measured in NSCLC cell lines, specifically PC9 and HCC827, and in their osimertinib-resistant sublines. TRPML3 mRNA levels were significantly higher in osimertinib-resistant cells than in their parental lines (Fig. 1B). Additionally, osimertinib exposure and resistance increased TRPML3 protein expression, whereas TRPML1 expression remained unchanged regardless of treatment or resistance (Fig. 1C and D). These findings suggest that TRPML3 plays a critical role in the lysosomal response associated with osimertinib resistance in NSCLC.

siRNA-mediated TRPML3 knockdown (siML3) was next performed in PC9 cells and their osimertinib-resistant subline PC9/OR. TRPML3 expression was efficiently reduced without affecting TRPML1 expression (Fig. 2A). Next, the effects of TRPML3 depletion on osimertinib-induced G0/G1 cell-cycle arrest was examined. Cells transfected with siML3 or scrambled RNA (control) were incubated for 48 h with or without 0.01 µM osimertinib. In PC9 cells, TRPML3 knockdown did not alter osimertinib-induced G0/G1 arrest but increased the sub-G0 population (Fig. 2B), suggesting increased cell death. By contrast, TRPML3 knockdown in PC9/OR cells restored their sensitivity to osimertinib-induced G0/G1 arrest (Fig. 2C). All tested TKIs, including osimertinib (Fig. 2D), lapatinib (Fig. 2E), gefitinib (Fig. 2F) and erlotinib (Fig. 2G), reduced PC9 viability in a dose-dependent manner. By contrast, PC9/OR cells displayed cross-resistance to all TKIs except for lapatinib. Finally, TRPML3 knockdown restored the sensitivity to osimertinib and other TKIs in these resistant cells. These findings indicate that TRPML3 plays a critical role in mediating TKI resistance in NSCLC cells.

To better mimic tumors in vivo, 3D spheroids derived from TRPML3-overexpressing and -silenced cells treated with osimertinib treatment for 9 days were employed. was initiated on day 0 and maintained for 9 days. Spheroid viability and size were assessed using CCK-8 assays and bright-field imaging, respectively (Fig. 3A). Western blotting confirmed changes in TRPML3 expression without affecting TRPML1 expression (Fig. 3B). Blotting with antibodies against the myc tag and TRPML3 verified pCMV-TRPML3 expression in PC9 and HCC827 cells, whereas targeting TRPML3 with sgRNA reduced endogenous TRPML3 expression in PC9/OR and HCC827/OR cells. On day 0, increased expression of the spheroid markers CD44, CD133 and Nanog was observed in spheroids, but not in parental 2D cultures (Fig. 3C and D). TRPML3 modulation did not affect spheroid formation. Next, the effects of TRPML3 overexpression (Fig. 3E and F) and silencing (Fig. 3G and H) on osimertinib-induced toxicity were assessed in spheroids derived from osimertinib-sensitive (PC9 and HCC827) and -resistant (PC9/OR and HCC827/OR) cells. Osimertinib produced a dose-dependent decrease in spheroid viability in sensitive cell-derived spheroids (Fig. 3E and F), whereas those derived from resistant cells exhibited greater tolerance (Fig. 3G and H). TRPML3 overexpression in PC9 (ML3 OE-PC9) and HCC827 (ML3 OE-HCC827) spheroids attenuated osimertinib-induced viability reductions at 0.1 µM (PC9) and 1 µM (HCC827). Conversely, TRPML3 silencing significantly reduced osimertinib tolerance in spheroids derived from PC9/OR and HCC827/OR cells (Fig. 3G and H). Consistent with these findings, intracellular TRPML3 levels influenced osimertinib-induced reductions in spheroid size. Spheroids were treated with osimertinib at the induced concentrations; bright-field images were captured every 3 days until day 9. Those derived from PC9, HCC827 and their sublines continued to increase in size in the absence of osimertinib (Fig. 3I-P). By contrast, osimertinib treatment reduced spheroid size in both PC9 (Fig. 3I) and HCC827 (Fig. 3J) cells in a dose-dependent manner. Spheroids from TRPML3-overexpressing PC9 (Fig. 3K) and HCC827 (Fig. 3L) cells exhibited increased resistance to osimertinib, suppressing size reduction, whereas those derived from PC9/OR (Fig. 3M) and HCC827/OR (Fig. 3N) cells maintained their size even at higher osimertinib concentrations (0.01 and 0.1 µM for PC9/OR; 0.1 µM for HCC827/OR). TRPML3 silencing in PC9/OR (Fig. 3O) and HCC827/OR (Fig. 3P) cells restored osimertinib sensitivity, reducing spheroid size at 0.1 µM osimertinib.

To improve understanding of the TRPML3-mediated TKI resistance, it was determined whether TRPML3 channel activity contributed to osimertinib resistance by measuring [Ca²⁺]_i. Osimertinib treatment induced intracellular Ca²⁺ oscillations in PC9 cells, independently of extracellular Ca²⁺ (Fig. 4A). These oscillations were more pronounced in resistant PC9/OR cells (Fig. 4B) and were inhibited by ML-SI1, a mucolipin channel blocker (Fig. 4C). Ca²⁺_i oscillation frequency was significantly reduced in TRPML3-deficient cells but not in TRPML1-deficient cells (Fig. 4D-F). In TRPML3-deficient PC9 cells, Ca²⁺ spikes were significantly reduced, whereas PC9/OR cells with TRPML3 overexpression exhibited enhanced oscillations (Fig. 4G). To determine whether these oscillations depend on lysosomal Ca²⁺ release via TRPML3, the TRPML3-GCaMP6 construct was used. GFP fluorescence intensity closely mirrored the Fura-2 fluorescence pattern in osimertinib-treated cells, consistent with TRPML3-mediated lysosomal Ca²⁺ release (Fig. 4H and I).

Lysosomal pH changes were next monitored using LysoSensor Yellow/Blue DND-160. The baseline fluorescence intensities were recorded before treatment, followed by continuous perfusion with HEPES buffer containing osimertinib. Osimertinib rapidly increased lysosomal pH in both PC9 (Fig. 5A) and HCC827 cells (Fig. 5B), reaching saturation at 60–65% of baseline. By contrast, osimertinib-resistant cells (PC9/OR and HCC827/OR) exhibited both a slower initial increase in lysosomal pH and a lower saturation level. TRPML3 deletion in osimertinib-resistant cells accelerated initial lysosomal pH elevation and increased its saturation level compared with that in scrambled control cells. These findings indicated that TRPML3 plays a critical role in the regulation of lysosomal pH changes induced by osimertinib in PC9 cells. Moreover, osimertinib-resistant NSCLC cells increased sequestration of osimertinib within lysosomes, a process that depends on TRPML3 expression. This suggests a link between TRPML3-mediated lysosomal Ca²⁺ release and regulation of lysosomal acidity.

Building on the role of TRPML3 in lysosomal acidity regulation, it was determined whether it also plays a role in osimertinib-induced lysosomal biogenesis, which has been primarily associated with TPRML1. TFEB translocation from the cytosol to the nucleus was analyzed by fractionating total cellular proteins. Osimertinib increased the time-dependent nuclear translocation of TFEB in both PC9 (Fig. 6A) and HCC827 (Fig. 6B) cells. However, TRPML3 deletion completely blocked TFEB translocation. LysoTracker Red DND-99 staining revealed that the osimertinib-induced increase in lysosomal area was significantly reduced in TRPML3-deficient PC9 (Fig. 6C) and HCC827 (Fig. 6D) cells. In TRPML3-deficient cells, the lysosomal area was comparable to that in untreated control cells (scr-PC9 and scr-HCC827). These findings indicate that TRPML3 is essential for the regulation of TFEB nuclear translocation and lysosomal biogenesis in response to osimertinib-induced lysosomal pH alteration.