Introduction

Bafilomycin A1 (BA1) is a class of macrolide antibiotics produced from Streptomycetes. Several biological functions such as antiparasitic, antifungal, immunosuppressive, and ***antitumor effects have been recognized for BA11. However, information on its mechanism of action remains limited. The best-known function of BA1 is to inhibit autophagy and specifically target the Vo subunit of the vacuolar-type ATPase (V-ATPase)2,3,4. V-ATPase is present in various cellular organelles including endosomes, lysosomes, Golgi complexes, and secretory vesicles5,6. The most significant effects of V-ATPase inhibition by BA1 relate to its role in the endolysosomal system, particularly in affecting lysosomal acidification and inhibiting the fusion of autophagosomes and lysosomes, which may contribute to cancer development and growth7,8,9,10,11. Studies of the anticancer effect of BA1 have demonstrated that BA1 inhibits both autophagy and apoptosis, interferes with cancer cell viability, and promotes cell death through both apoptotic and non-apoptotic pathways12,13,14,15,16. These results suggest that BA1 is a candidate drug for cancer treatment. However, to effectively inhibit cancer cell growth, autophagic flux, and endolysosomal system/vesicle acidification, BA1 is typically used at concentrations exceeding 0.1 µM. Such high concentrations can induce severe acidosis and harmful effects in normal cells, thereby obstructing its clinical application17.

In the present study, we found that BA1 at nanomolar concentrations significantly inhibited growth in both in vitro cell cultures and in vivo zebrafish models. Furthermore, BA1 induces caspase-independent cell death associated with endolysosomal system impairment, endoplasmic reticulum (ER) stress, and intracellular iron deficiency, which are reminiscent of lysosomal storage diseases18,19. However, the detailed mechanism remains unclear, necessitating further investigation of BA1-induced intracellular responses and signaling pathways.

Results

CRC cells were sensitive to low concentration of BA1

To investigate the effect of BA1, MTT assays were performed in seven CRC cell lines (HCT116, HT-29, SW480, LoVo, DLD-1, SNU-70, and SNU-796), five lung cancer cell lines (HCC827, H820, H1975, H460 and A549) as well as a normal colon fibroblast cell line (CCD-18Co). The results showed that CRC cells were much more sensitive to low concentrations of BA1 than lung cancer cells and normal colon fibroblast cells, in which there was almost no effect from the indicated concentrations of BA1 (Fig. 1A). Further colony formation and FACS analyses of LoVo and SW480 cells confirmed that BA1 inhibited cell proliferation and induced cell death (Fig. 1B,C). Cell migration was also reduced by BA1, as detected by a migration assay (Fig. 1D). These results indicated that BA1, as an anticancer agent, is specifically effective at low concentrations against CRC cells.

Fig. 1
figure 1

Low concentration of BA1 reduces cell viability and proliferation in CRC. (A) Cell viability was determined via MTT assay. (B) Cell proliferation was examined by the colony formation assay (top). The graph shows quantification of colony formation (bottom). (C) Cell death was analyzed by flow cytometry using annexin V/PI staining. (D) Horizontal migration was examined using wound healing assay at the indicated time points. Representative cell images were captured by a light microscope. Scale bar: 1 mm. #p > 0.05 (no significance), *p < 0.05, **p < 0.01, ***p < 0.001, compared with control.

Full size image

BA1-induced cell death was rescued by ferric citrate

To determine the type of cell death induced by BA1, cells were co-treated with BA1 and with an apoptosis inhibitor (Z-VAD-FMK), ferroptosis inhibitor (ferrostatin), necroptosis inhibitor (necrostatin-1), paraptosis inhibitor/protein synthesis inhibitor (CHX), or iron (III)-citrate as an iron supplement, and cell viability was measured using the MTT assay. Cell death was rescued by iron (III)-citrate, but not by other inhibitors (Fig. 2A), as confirmed by cell count and FACS analyses (Fig. 2B,C). The cleaved forms of caspase-3, and caspase-9 as apoptosis markers, were examined by western blotting and were not observed in BA1-treated cells (Fig. 2D). Etoposide was used as the positive control for cleaved caspase-3 and caspase-9. These results suggested that BA1-induced cell death may be related to iron deficiency. Therefore, intracellular iron balance was observed using FerroOrange, that specifically stains labile iron (II) ions (Fe2+). The intracellular fluorescence intensity of FerroOrange decreased when BA1 was added and was restored when iron (III)-citrate was added with BA1, as observed using a confocal microscope (Fig. 2E) and a microplate reader (Fig. 2F). Consistent with these results, ferritin expression was decreased by BA1 and was restored by iron (III)-citrate (Fig. 2G). In addition, Sudan-Black-B (SBB) staining, which specifically reacts with lipofuscin, an aggregate of oxidized proteins, lipids, and metals, showed increased accumulation of lipofuscin granules after BA1 incubation, which also was reversed in the presence of iron (III)-citrate (Fig. 2H).

Fig. 2
figure 2

Cellular iron is associated with cell death triggered by BA1. Cells were treated with BA1 (LoVo: 2 nM, SW480: 1.5 nM) with or without iron (III)-citrate. (A) Cell viability was determined via the MTT assay. (B) Cell proliferation was analyzed by cell counting. Representative cell images were captured by a light microscope. Scale bar: 1 mm. (C) Cell death was analyzed by flow cytometry using annexin V/PI staining. (D) The expression of caspase-3 and caspase-9 was detected by western blotting. Etoposide (100 µM) treatment was used as a positive control for inducing cleaved caspase-3 and -9. (E,F) Intracellular iron was examined using by FerroOrange. Representative cell images were captured by a laser confocal scanning microscope, scale bar: 10 μm (E), and the relative fluorescence intensity was measured using a microplate reader (F) after cell staining with FerroOrange. (G) The expression of ferritin heavy chain was detected by western blotting. (H) Representative image of lipofuscin stained by Sudan Black B, captured by laser confocal microscope. Scale bar: 10 μm. #p > 0.05 (no significance), *p < 0.05, **p < 0.01, ***p < 0.001, compared with control.

Full size image

BA1 impaired the endolysosomal system and autophagic flux

Since BA1 is known to increase endolysosomal pH, pH changes in endolysosomal vesicles were measured using acridine orange, BHEt-1, LysoTracker, and LysoSensor. Contrary to expectations, the pH of endolysosomal vesicles was not increased by low concentrations of BA1; cell death was induced, but the number and size of acidic perinuclear vesicles increased (Fig. 3A–C, and Supplementary Fig. S1A). Moreover, treatment with iron (III)-citrate as a supplement with BA1 recovered the number and size of acidic vesicles (Fig. 3B, and Supplementary Fig. S1B). To verify the activity of BA1, a high concentration (100 nM) of BA1 was added to LoVo and SW480 cells for 1 h, resulting in increased pH, as expected (Supplementary Fig. S1B).

Fig. 3
figure 3

BA1 increased size of perinucleic acidic vesicles independent of endolysosomal pH. Cells were treated with BA1 (LoVo: 2 nM, SW480: 1.5 nM) with or without iron (III)-citrate. (A) Representative confocal image of cells stained with acridine orange. Orange-red fluorescence indicates acidic compartments. Scale bar: 10 μm. (B) Cells were stained with LysoTracker Red DND-99, and pH-sensitive probe (BHEt-1). (C) Quantitative intracellular pH was measured using LysoSensor Yellow/Blue DND-160 by a microplate reader. Scale bar: 10 μm. #p > 0.05 (no significance), *p < 0.05, **p < 0.01, ***p < 0.001, compared with control.

Full size image

Since acridine orange, LysoTracker, and BHEt-1 staining showed increased volume/size and number of perinuclear vesicles, early and late endolysosomal systems, as well as autophagy flux, were further examined by transfection of plasmid DNAs, GFP-EEA1wt, mCh-Rab7A, GFP-Rab7A, pLAMP1-mCherry LC3-GFP, or mRFP-GFP-LC3 and/or staining with LysoTracker. EEA1 is an early endosome marker, and EEA1-positive vesicles were not increased by BA1 treatment, unlike LysoTracker (Fig. 4A). However, the number and size of vesicles positive for Rab7a, a late endosome marker, and LAMP1, a lysosomal marker, were increased along with LysoTracker, indicating that late endolysosomal vesicles were increased (Fig. 4B, C). In addition, autophagy flux was examined using the mRFP-GFP-LC3 plasmid, where GFP is converted to RFP in an acidic environment. With this method, the autophagosome is yellow while the autolysosome is red. The results showed an increase in yellow puncta (autophagosomes), but not red puncta (autolysosomes), after BA1 treatment (Fig. 5A). This result indicated the blocking/inhibition of autophagy flux, which is consistent with the western blotting results for the autophagy markers p62 and LC3-II (Fig. 5C). Further examination of LC3-GFP, an autophagosome marker, with pLAMP1-mCherry, mCh-Rab7a, or LysoTracker showed increased numbers and sizes of vesicles positive for LC3 co-localized with LAMP1, LysoTracker, or Rab7 (Fig. 5B, and Supplementary Fig. S2). These results confirmed that autophagosomes merged with lysosomes, but that the endolysosomal systems were possibly dysfunctional. Thus, lysosomal impairment was assessed. The ptf-Galectin-3 plasmid, with mRFP- and GFP-tandem-tagged Gal3, was used to observe lysosomal damage and the pH environment. GFP-mRFP double-positive galectin-3 puncta (yellow) were induced after treatment with BA1, indicating that the lysosomal membrane was damaged, but not sufficiently acidic to quench GFP (Fig. 6A). The GFP signal can be easily quenched, whereas mRFP is more stable under acidic lysosomal conditions20. Cathepsin B (CTSB), a lysosomal cysteine endopeptidase, was also examined by western blotting, which showed increased inactivation of pro-CTSB and decreased activation of CTSB following BA1 treatment, indicating lysosomal membrane permeabilization (Fig. 6B). Leupeptin, a lysosomal protease inhibitor, restored the ferritin levels decreased by BA1, but not by MG132 (Fig. 6C). Moreover, pEGFP-N1-TFEB used to observe the localization of TFEB showed translocation of TFBE from the cytoplasm to the nucleus after treatment with BA1 (Fig. 6D). Furthermore, disruption/impairment of endocytosis or recycling of the receptor protein EGFR was examined by transfection of EGFR-GFP, which showed decreased expression of membrane and aggregated cytoplasmic EGFR after treatment with BA1 (Fig. 6E). Lipid droplet accumulation, along with increased expression of ADRP by BA1, was examined using live tomography microscopy and RT-qPCR (Fig. 6F,G). Taken together, these results indicate that BA1 induces the dysfunction or impairment of the endolysosomal system.

Fig. 4
figure 4

BA1 impaired late endolysosomes. Endolysosomal vesicles were examined after treatment of cells with BA1 (1.5 or 2 nM) for 24 h. The number of puncta/cell for each probe was calculated. (A,B) Representative live images were captured by laser confocal microscope after transfection with the GFP-EEA1 wt (A) or GFP-Rab7A (B) plasmids along with LysoTracker Red DND-99 staining. (C) Representative images were captured by laser confocal microscope after immunofluorescence staining with lysotracker (red), LAMP1 (green), and DAPI (blue). Scale bar: 10 μm. #p > 0.05 (no significance), *p < 0.05, **p < 0.01, ***p < 0.001, compared with control.

Full size image
Fig. 5
figure 5

BA1 disrupted the endolysosomal system including autophagy. Cells were treated with BA1 (1.5 or 2 nM) with or without iron (III)-citrate for 24 h. (A) Representative live images captured by laser confocal microscope after transection of the mRFP-GFP-LC3 tandem plasmid. Scale bar: 10 μm. The number of LC3 puncta/cell was calculated. Relative LC3 puncta represent autolysosomes (red puncta) and autophagosomes (yellow puncta). (B) Representative live images captured by laser confocal microscope after transfection of LC3-GFP plasmid along with LysoTracker Red DND-99 staining. The number of LC3-GFP, and lysotracker puncta/cell was calculated. (C) Expression of p62 and LC3 were detected by western blotting. #p > 0.05 (no significance), *p < 0.05, **p < 0.01, ***p < 0.001, compared with control.

Full size image
Fig. 6
figure 6

BA1 induced lysosomal damage. Cells were treated with BA1 (1.5 or 2 nM) with or without iron (III)-citrate for 24 h. (A) Representative live images captured by laser confocal microscope after transfection of the ptf-Galectin-3 plasmid (top). Relative number of RFP puncta (red), or RFP and GFP double positive puncta (yellow) were counted (bottom). Scale bar: 10 μm. (B) Expression of CTSB was detected by western blotting at the indicated time points. (C) Expression of ferritin was detected by western blotting after co-treatment with iron (III)-citrate for 24 h, or leupeptin, or MG132 for 6 h. (D,E) Representative live images captured by laser confocal microscope after transection of pEGFP-N1-TFEB (D) or EGFR-GFP (E) plasmids. Scale bar: 10 μm. (F) Representative live cell images captured by holotomography microscope. Arrow indicates lipid droplets. Scale bar: 7 μm. The ratio of lipid droplet volume to cell volume was calculated. (G) Expression of ADRP mRNA level detected by RT-qPCR. #p > 0.05 (no significance), *p < 0.05, **p < 0.01, ***p < 0.001, compared with control.

Full size image

BA1-induced endolysosomal impairment is associated with ER stress response

Since it has been reported that endolysosome dysfunction induces ER stress, RT-qPCR and/or western blotting of gene products related to the ER stress response, EIF2AK3 (PERK), DDIT3 (CHOP), and HSPA5, were performed, and an increase in these molecules was observed after treatment with BA1 (Fig. 7A,B). A splice variant of XBP1 that plays a critical role in activating UPR target genes was also detected (Fig. 7C). Staining with ER-Tracker also identified morphological changes, such as formation of ER whorls by BA1 treatment, confirming the induction of ER stress by BA1 (Fig. 7D). BA1-induced ER stress response was rescued by iron (III)-citrate supplementation. These results indicated that BA1 induced ER stress is likely a consequence of disrupted iron homeostasis due to lysosomal dysfunction. Iron (III)-citrate supplementation effectively resolves this stress by providing an alternative source of iron that doesn’t rely on lysosomal processing, thus restoring cellular iron balance and mitigating the downstream effects on mitochondrial function and ER stress21,22.

Fig. 7
figure 7

BA1 induced ER stress. Cells were treated with BA1 (1.5 or 2 nM) with or without iron (III)-citrate. (A,B) Expression of the ER stress markers detected by western blotting (A), and RT-qPCR (B). (C) Detection of XBP1 splicing by PCR analysis. Tunicamycin treatment was used as a positive control for inducing ER stress. (D) Representative images were captured by laser confocal microscope after staining with ER-Tracker Blue-White DPX. Scale bar: 10 μm. #p > 0.05 (no significance), *p < 0.05, **p < 0.01, ***p < 0.001, compared with control.

Full size image

BA1 inhibited tumor growth in a zebrafish model

Suppression of tumorigenesis in vivo by BA1 treatment in LoVo and SW480 cells was investigated in a xenograft zebrafish model. CMDil-labeled cells were grafted onto Tg(flk1:EGFP) zebrafish embryos, cultured in E3 medium with DMSO or BA1, and refreshed every 24 h for 5 days (Fig. 8). Both showed a reduced tumor area. These results indicate that BA1 has an anticancer effect, which is consistent with the in vitro results.

Fig. 8
figure 8

BA1 treatment had an anticancer effect in vivo. (A) Representative laser confocal images of CM-Dil-labeled cancer cells (red) in the vasculature (green) of zebrafish larvae. (B) Relative area penetrated by CM-Dil-labeled cancer cells after treatment with DMSO or BA1. Scale bar: 20 μm. *p < 0.05, **p < 0.01, ***p < 0.001, compared with control.

Full size image

Discussion

CRC remains a significant global health concern, with high incidence and prevalence rates worldwide23. However, available therapies are limited, underscoring the urgent need for novel drug discovery approaches. Recent research has highlighted the potential of targeting the lysosomal stress response as a therapeutic strategy for neurodegenerative disorders and cancers24,25,26. Lysosomes are integral components of the broader lysosomal system and play crucial roles in intracellular transport, signaling, and metabolism. This system encompasses various intracellular vesicles including endosomes, multivesicular bodies, autophagosomes, and autolysosomes. The subcellular positioning and size of lysosomes within cells are dynamic, and respond to various stimuli, stress factors, and pH levels. Cancer cells are characterized by numerous relatively large acidic lysosomes that are thought to be more fragile than those in normal cells27,28. BA1, a widely used autophagy inhibitor targeting the endolysosomal system, has shown promising anticancer effects in various cancer cell types, including colon cancer11,12,13,14,15,29,30,31,32,−33. Its primary mechanism of action involves growth inhibition and cell cycle arrest, wherein the impaired fusion of autophagosomes with dysfunctional lysosomes inhibits autophagy and triggers apoptosis. However, most studies reporting cancer cell death used BA1 concentrations higher than 0.1 µM. Notably, some exceptions have been reported, where lower concentrations of BA1 induced apoptosis and growth inhibition in specific cancer types, such as hepatocellular carcinoma29, the pancreatic cancer cell line Capan-1 (5 nM)30, gastric cancer cell line MKN-131, pediatric B-cell acute lymphoblastic leukemia cells32, and diffuse large B-cell lymphoma (5 nM)33. Therefore, BA1 may play various roles in cell death under different cellular conditions.

Our results demonstrated that CRC cells are highly sensitive to low concentrations (~ 2 nM) of BA1, compared to CCD-18Co noncancerous colon cells and lung cancer cells. Unlike previous studies showing that high concentrations of BA1 induced apoptosis in CRC cells34,35, our results indicated that caspase-independent cell death was rescued by Fe3+ supplementation but not by the other inhibitors we used. This suggests that iron homeostasis is associated with BA1-induced cell death. A recent report showed that inhibition of lysosomal acidification triggers cellular iron deficiency and HIF-1α activation, resulting in inhibition of cell proliferation and non-apoptotic cell death21. The iron-deficiency response caused by V-ATPase inhibition is likely due to the retention of iron in lysosomes. Supplementation with Fe3+ may allow iron to be imported through transporters, thus bypassing the endolysosomal pathway and promoting cell survival21.

Additionally, BA1 treatment disrupted autophagic flux, as evidenced by the accumulation of autophagosomes and lysosomes detected by mRFP-GFP-LC3 and LAMP1 markers, whereas autolysosomes were reduced by BA1, as previously reported. Interestingly, further examination revealed that autophagosomes and late endosomes or lysosomes were likely fused, as indicated by the increased colocalization of LAMP1 or Rab7a with mRFP-LC3. Contrary to previous studies6,7,8,9,10, BA1 did not increase endolysosomal pH at the concentration used in our study. This observation was confirmed using various methods, including LysoTracker, BHEt-136, and acridine orange staining, and LysoSensor. While these results complicate interpretation, two possibilities can be suggested. First, the lysosomal pH may have increased slightly, but the techniques used were not sensitive enough to measure it. Increased autolysosome formation, despite the lack of mRFP-GFP-LC3 cleavage to mRFP-LC3, suggests that the lysosomal environment was not acidic enough for digestive function. Another possibility is that the observed lysosomal dysfunction may not be due to a loss of acidic pH but rather due to an alternative mechanism of lysosomal impairment at low BA1 concentrations. This could directly affect the functionality of lysosomal enzymes or membrane integrity, warranting further investigation. Supporting the hypothesis of lysosomal dysfunction, we observed leakage of cathepsin B, increased galectin-3 levels, translocation of TFEB from the cytoplasm to the nucleus, accumulation of lipid droplets37, induced ER stress38, and accumulation of lipofuscin39. These findings are consistent with those of previous studies on lysosomal dysfunction in cancers and other diseases24,25,26,27.

Recent advances in endolysosomal-targeting therapeutics have shown promise across various cancer types, including colorectal cancer (CRC)40,41,42,43. BaP1, a benzo[a]phenoxazine derivative, demonstrates selective anticancer activity against CRC cells by inducing LMP, leading to apoptotic cell death44. DQ661, a new lysosomal inhibitor targeting palmitoyl-protein thioesterase 1 (PPT1), has demonstrated efficacy against multiple tumor models by inhibiting autophagy, macropinocytosis, and mTORC1 signaling43. Disrupting iron homeostasis through lysosomal targeting has emerged as a potential strategy, particularly in cancers dependent on iron for survival and growth45,46,47. Additionally, LMP inducing agents have shown efficacy in cancer when combined with conventional chemotherapeutics47,48. Therefore, combination therapies involving BA1 at lower concentrations and other agents such as iron chelators could potentially enhance treatment efficacy while mitigating potential side effects, such as acidosis in normal cell49. This approach might allow for more targeted treatment of cancer cells that are more dependent on iron and autophagy for survival, while minimizing damage to healthy tissues45.

In conclusion, our study provides new insights into the sensitivity of CRC cells to low BA1 concentrations and the complex interplay between lysosomal function, iron homeostasis, and cell death mechanisms in CRC cells. These findings contribute to the growing body of evidence supporting the targeting of lysosomal function as a promising approach to cancer therapy, particularly for CRC. Further investigation into the molecular pathways involved in BA1-induced cell death and their relationship with iron homeostasis may lead to the development of more effective and targeted therapies for patients with CRC. It is important to note that the specific effects can vary depending on the experimental system, cell type, and the precise nature of the V-ATPase inhibition by BA1 underlying its anticancer effects.

Materials and methods

Cell culture and reagents

Human colorectal cancer (CRC) cell lines (HCT116, HT-29, SW480, LoVo, and DLD-1), a normal colon fibroblast cell line (CCD-18Co), and human non-small cell lung carcinoma (NSCLC) cell lines (HCC827, H820, H1975, H460, and A549) were obtained from the American Type Culture Collection (Manassas, VA, USA). The CRC cell lines SNU-70 and SNU-796 were purchased from the Korea Cell Line Bank (Seoul, South Korea). CCD18Co cells were cultured in DMEM (WELGENE, Daegu, South Korea), while CRC and NSCLC cells were cultured in RPMI 1640 (WELGENE) supplemented with 10% fetal bovine serum (WELGENE) and 1% penicillin/streptomycin (WELGENE) at 37 °C in a humidified atmosphere containing 5% CO2. Bafilomycin A1 (BA1) (B1793, Sigma-Aldrich, St. Louis, MI, USA), Z-VAD-FMK (S7023, Selleckchem, USA), ferrostatin-1 (S7243, Selleckchem), necrostatin-1 (S8037, Selleckchem), etoposide (S1225, Selleckchem), cycloheximide (CHX) (S7418, Selleckchem), MG132 (474787, Sigma-Aldrich), leupeptin hemisulfate (S7380, Selleckchem), tunicamycin (T7765, Sigma-Aldrich) were dissolved in dimethyl sulfoxide (DMSO) (D2650, Sigma-Aldrich). Iron (III)-citrate (F3388; Sigma-Aldrich) was dissolved in distilled water. Cells were seeded in an appropriate culture dish. After overnight incubation, cells were treated with DMSO, or BA1 (1.5–2 nM) with or without iron (III)-citrate (10 µM), or other indicated chemicals.

Cell viability and proliferation assay

Cell viability was measured using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Cells were treated with the indicated concentrations of BA1, ferrostatin-1, Z-VAD-FMK, iron (III)-citrate, necrostatin-1, or CHX for 48 h, followed by the addition of 50 µL of MTT solution (0.5 mg/mL, Sigma-Aldrich) to each well. After 4 h of additional incubation, the MTT solution was discarded and DMSO was added. The absorbance was measured at 570 nm using a microplate reader (SpectraMax Plus 384, Molecular Devices, USA). Cell proliferation was measured by counting. The number of cells was counted using a hemocytometer after trypan blue staining (Sigma-Aldrich). Cell images were captured using an inverted light microscope (Olympus IX71, Olympus Corp.).

Cell migration assay

Horizontal cell migration was assessed using a wound healing assay. Cells were grown to 90% confluence, and then the cell monolayer was scraped with a sterile micropipette tip, and the medium was changed to remove detached cells. The degree of migration was quantified by calculating the area of the migrated cells using the image processing program ImageJ (NIH, Bethesda, MD, USA) at the indicated time points after treatment with DMSO or BA1.

Colony formation assays

The cells were treated with BA1 for 24 h and then grown for 7 days. Cells were then fixed and stained using the Hemacolor® Rapid staining (Sigma-Aldrich). Stained cells were washed with distilled water, and air dried at ~ 20 °C. The number of colonies were quantified using the Image J program (NIH, Bethesda, MD, USA).

RNA isolation and reverse transcription-quantitative polymerase chain reaction (RT-qPCR)

Total RNA was extracted from the cells using TRIZOL (Invitrogen, CA, USA). cDNA was synthesized from total RNA using a reverse transcription kit (Labopass, Cosmo Genetech, Seoul, South Korea) following the manufacturer’s instructions. RT-qPCR was performed with gene-specific primers and SYBR Green Q Master (Labopass) using a QuantStudio 3 Real-Time PCR Instrument (Applied Biosystems, MA, USA). The oligonucleotide primers used for RT-qPCR analysis were: EIF2AK3 Forward: 5′-GCTGTCGGACCTCGCAGTGG-3′, Reverse: 5′-TCCGGCTCTCGTTTCCATGTCTG-3′; DDIT3 Forward: 5′-GGAGCATCAGTCCCCCACTT-3′, Reverse: 5′-TGTGGGATTGAGGGTCACATC-3′; HSPA5 Forward: 5′-ACCGCTGAGGCTTATTTGGG-3′, Reverse: 5′-GTCTTTGGTTGCTTGGCGTT-3′; ADRP Forward: 5′-TAACAACACGCCCCTCAACT-3′, Reverse: 5′-GCACCTTGGTCCTGAGCATT-3′; GAPDH Forward: 5′-ACCCACTCCTCCACCTTTGA-3′, Reverse: 5′-CTGTTGCTGTAGCCAAATTCGT-3′. Ct values of the target genes were normalized to those of an endogenous reference gene (GAPDH). Each gene was assessed in triplicates, repeated in three independent experiments.

Flow cytometry assay

Cell death was measured using the FITC Annexin V Apoptosis Detection Kit I (BD Pharmingen™, NJ, USA) following the manufacturer’s instructions. Cells death was analyzed by Flow Cytometer (BD Bioscience, BD FACSCanto™ II Flow Cytometry System, NJ, USA). The percentage of cell death was quantified by summing the proportion of cells in quadrants Q2 and Q4 of the flow cytometry dot plot.

Measurement of intracellular Fe2+

Intracellular Fe2+ was measured using FerroOrange (F374, Dojindo, Kumamoto, Japan), following the manufacturer’s instructions. Fluorescence images were obtained using a confocal laser scanning microscope (LSM 800; Carl Zeiss, Oberkochen, Germany) and visualized using an appropriate objective lens. Quantitative measurement of the FerroOrange intensity of cells was performed at a wavelength of 543 nm excitation and 580 nm emission using an EnSpire MultiPlate Reader (PerkinElmer, MA, USA). Each experiment was performed in triplicate and repeated in three independent experiments.

Western blotting analysis

Cells were lysed in RIPA lysis buffer with protease/phosphatase inhibitor cocktail (Sigma-Aldrich) at 4 °C for 1 h. Proteins were separated by SDS-PAGE and transferred to PVDF membranes (IPVH00010, Millipore, MA, USA). The membranes were subsequently incubated with the indicated primary antibodies and either goat anti-mouse IgG (#7076, Cell Signaling, MA, USA) or goat anti-rabbit IgG (#7074, Cell Signaling) secondary antibodies conjugated with horseradish peroxidase (HRP). Chemiluminescence was detected using an enhanced chemiluminescence (ECL) system (TransLab, South Korea). The following primary antibodies were used: p62 (ab56416, Abcam, Cambridge, UK), LC3 (PM036, MBL, Nagoya, Japan), caspase-3 (#9662, Cell Signaling), caspase-9 (human specific, #9502, Cell Signaling), BiP (#3177, Cell Signaling), p-PERK (Thr980) (#3719, Cell Signaling), p-eIFα (Ser51) (#9721, Cell Signaling), CHOP (#2895, Cell Signaling), cathepsin B (D1C7Y) (#31718, Cell Signaling), ferritin heavy chain (sc-376594, Santa Cruz Biotechnology, CA, USA), GAPDH (60004-1-lg, Proteintech, Rosemont, IL, USA). The densitometry quantification of the western blot was determined using the Image J program (NIH, Bethesda, MD, USA). The result of gels images was cropped (Supplementary Fig. S3).

Autophagy flux and endo-lysosomal system monitoring

Plasmids GFP-EEA1wt (#42307), GFP-Rab7A (#61803), mCh-Rab7A (#61804), pLAMP1-mCherry (#45147), EGFR-GFP (#32751), ptf-Galectin-3 (#64149), and pEGFP-N1-TFEB (#38119) were purchased from Addgene (Cambridge, MA). The GFP-LC3 vector was kindly provided by Professor Daniel J. Klionsky (Department of Molecular Sciences, University of Michigan, USA). The mRFP-GFP-LC3 tandem vector was provided by Professor Yoshimori (Department of Genetics, Graduate School of Medicine, Osaka University, Osaka, Japan). Transfection of plasmid DNAs was performed using GeneFect (TLC-001, TransLab) following the manufacturer’s instructions to examine endolysosomal systems such as early endosomes, late endosomes, lysosomes, and autophagy flux. Fluorescence images of live cells were obtained using a confocal laser scanning microscope (LSM 780, Carl Zeiss) with an appropriate objective lens.

Immunofluorescence staining

Cultured cells were washed twice with ice-cold PBS and fixed with 4% formaldehyde for 10 min at ~ 20 ℃. The fixed cells were washed thrice with ice-cold PBS, after which, the cells were permeabilized with 0.1% TritonX-100 in PBS for 10 min, blocked with 3% BSA, anti-LAMP1 (sc-20011, Santa Cruz Biotechnology) antibodies overnight at 4 ℃ in the dark, and then incubated with Alexa Fluor 488 secondary antibodies. Each coverslip was mounted with medium containing the nuclear stain DAPI (H1200, Vector Laboratories, USA). Fluorescence images were obtained using a confocal laser scanning microscope (LSM 800) and visualized using an appropriate objective lens.

Measurement of endo-lysosomal pH

Endolysosomal pH was examined using Acridine Orange Staining Solution (ab270791, Abcam), LysoTracker Red DND-99 (L7528, Invitrogen, CA, USA), pH-sensitive probes (BHEt-1)36, or LysoSensor Yellow/Blue DND-160 (L7545, Invitrogen) following the manufacturer’s instructions. Fluorescent images of live cells were obtained using a confocal laser scanning microscope (LSM 900, Carl Zeiss) with an appropriate objective lens. Quantitative measurements of lysosomal pH using the LysoSensor™ Yellow/Blue DND-160 (5 µM) was followed the method described by Li Ma et al.50. Fluorescence was measured using a microplate reader (SpectraMax Plus 384).

ER monitoring using ER tracker

Cultured cells were stained with 1 µM of ER-Tracker™ Red (E34250, Invitrogen, CA, USA) or ER-Tracker™ Blue-White DPX (E12353, Invitrogen) for 1 h. Fluorescence images were examined, and captured using a confocal laser scanning microscope (LSM 800) with an appropriate objective lens.

Sudan Black B staining

Cultured cells were fixed in 4% paraformaldehyde for 10 min at ~ 20 °C, then stained with lipofuscin using the Sudan Black B Stain Kit (SBK-1; ScyTek Laboratories, USA) according to the manufacturer’s instructions. Confocal microscopy (LSM 900) was performed, and lipofuscin was visualized as dark blue or black vesicles.

Live holotomography microscopy

The cells were seeded into a TomoDish (Tomocube Inc., Daejeon, South Korea) and incubated overnight. The 3D label-free live cell images were obtained with a 3D holotomography microscope (HT-2 H, Tomocube Inc.) at 37 ℃ in a 5% CO2 atmosphere at a wavelength of 532 nm. The ratio of lipid droplet volume to cell volume was quantified using TomoAnalysis software (Tomocube Inc., Daejeon, Republic of Korea).

In vivo zebrafish tumor model

Zebrafish (Danio rerio) experimental protocols were approved by the local ethics board (Sookmyung Women’s University Animal Care and Use Committee, SMWU-IACUC-1712–036, SMWU-IACUC-2405-015) and performed in accordance with relevant guidelines and regulations as previously described51. The authors confirm that all animal experiments in the present study have been performed in accordance with ARRIVE guidelines. LoVo or SW480 cells were xenografted into embryos at 2 dpf, and 5 or 7 nM of BA1 was added to the embryonic media for 5 days, one day post injection. Then, the 2 dpf zebrafish embryos were anesthetized with 0.4 mg/mL tricaine solution in E3 medium. Images were acquired using a confocal microscope (LSM 700; Carl Zeiss) at the Chronic and Metabolic Diseases Research Center of Sookmyung Women’s University.

Statistical analysis

Experimental results are presented as mean ± SD from at least three independent experiments. Comparisons between two independent groups were performed using a two-tailed Student’s t-test and compared with the control. The ANOVA test was used to compare multiple groups. Statistical significance was reached at p < 0.05.