CYP51A1 drives resistance to pH-dependent cell death in pancreatic cancer

Abstract

Disrupted pH homeostasis can precipitate cell death and represents a viable therapeutic target in oncological interventions. Here, we utilize mass spectrometry-based drug analysis, transcriptomic screens, and lipid metabolomics to explore the metabolic mechanisms underlying pH-dependent cell death. We reveal CYP51A1, a gene involved in cholesterol synthesis, as a key suppressor of alkalization-induced cell death in pancreatic cancer cells. Inducing intracellular alkalization by the small molecule JTC801 leads to a decrease in endoplasmic reticulum cholesterol levels, subsequently activating SREBF2, a transcription factor responsible for controlling the expression of genes involved in cholesterol biosynthesis. Specifically, SREBF2-driven upregulation of CYP51A1 prevents cholesterol accumulation within lysosomes, leading to TMEM175-dependent lysosomal proton efflux, ultimately resulting in the inhibition of cell death. In animal models, including xenografts, syngeneic orthotopic, and patient-derived models, the genetic or pharmacological inhibition of CYP51A1 enhances the effectiveness of JTC801 in suppressing pancreatic tumors. These findings demonstrate a role of the CYP51A1-dependent lysosomal pathway in inhibiting alkalization-induced cell death and highlight its potential as a targetable vulnerability in pancreatic cancer.

Previously, the opioid analgesic drug JCT801 was reported to induce cell death via disruption of pH homeostasis in pancreatic cancer cells. Here, the authors investigate the metabolic mechanisms underlying JCT801-induced cell death, identifying cholesterol synthesis gene, CYP51A1, as a suppressor of alkalization-induced cell death.

Introduction

Therapy resistance presents a challenge in cancer treatment, primarily due to alterations in regulated cell death (RCD) pathways, including both apoptotic and non-apoptotic mechanisms¹. To overcome the resistance of cancer cells to apoptosis induced by conventional chemotherapy, alternative approaches are needed to induce non-apoptotic forms of RCD effectively. In addition to established forms of RCD, such as necroptosis and ferroptosis^2–4, several types of cell death show promise in eliminating drug-resistant cancer cells^5–7. A previously reported model known as ‘alkaliptosis’ describes a non-apoptotic form of RCD triggered by lethal intracellular alkalinization, which is induced by the compound JTC801⁸. Given that decreased extracellular pH is associated with tumor development and chemoresistance⁹, the use of alkalinizing agents may represent a promising therapeutic strategy in oncology^10,11.

JTC801, an antagonist of opioid-related nociceptin receptor 1 (OPRL1), exhibits potent anti-nociceptive effects in acute pain animal models¹². However, JTC801-induced pH-dependent cell death is not dependent on OPRL1 as well as other regulators associated with known forms of RCD, such as caspases for apoptosis, mixed-lineage kinase domain-like pseudokinase (MLKL) for necroptosis, or glutathione peroxidase 4 (GPX4) for ferroptosis⁸. In most cases, JTC801-induced cell death can be triggered by NF-κB activation, leading to the downregulation of carbonic anhydrase 9 (CA9) or the upregulation of ATPase H⁺ transporting V0 subunit D1 (ATP6V0D1), thereby disrupting lysosomal pH^8,10. CA9 is an enzyme that helps maintain pH balance by catalyzing the reversible hydration of carbon dioxide, which facilitates the transport of bicarbonate ions and protons across cell membranes. When CA9 is excessively downregulated, the impaired H⁺ efflux can activate other transporters or regulators to increase intracellular pH levels as an adaptive mechanism. This alkalinization disrupts cellular homeostasis and can activate signaling pathways that induce cell death. Despite this, neither CA9 overexpression nor ATP6V0D1 inhibition fully prevents anticancer activity of JTC801, suggesting potential involvement of negative feedback or complex compensatory mechanisms.

Cholesterol is a component of cell membranes and plays a critical role in maintaining membrane integrity and fluidity, which in turn influences the sensitivity of cell death pathways¹³. It also plays a role in forming lipid rafts, specialized membrane microdomains for signaling and protein trafficking¹⁴. Cancer cells often exhibit heightened cholesterol synthesis compared to normal cells, driven by their rapid proliferation and increased demand for cell membrane formation¹³. This augmented cholesterol synthesis is regulated by various enzymes and transcription factors, including β-hydroxy β-methylglutaryl-CoA (HMG-CoA) reductase and sterol regulatory element binding transcription factor 2 (SREBF2). Modulating cholesterol metabolism in tumor microenvironment may offer an opportunity to target tumor cell vulnerabilities^15,16.

In this study, we unveil a role for the cholesterol biosynthesis enzyme, cytochrome P450 family 51 subfamily A member 1 (CYP51A1), as a suppressor of JTC801-induced cell death in human pancreatic ductal adenocarcinoma (PDAC) cells. CYP51A1, which is transcriptionally regulated by SREBF2, prevents cell death by increasing lysosomal proton release through transmembrane protein 175 (TMEM175). Genetic and pharmacological CYP51A1 inhibition enhances JTC801-induced tumor suppression, establishing the cholesterol metabolism pathway as a potential target to modulate cell death sensitivity for PDAC therapy.

Results

CYP51A1 is upregulated during JTC801 treatment

To explore the key regulators influencing pH-dependent cell death, we performed mRNA transcriptomics assays on the human PDAC cell line MIAPaCa2 with and without JTC801 treatment. This analysis revealed gene alterations induced by JTC801 (Supplementary Fig. 1a), with an emphasis on genes associated with the cholesterol metabolic pathway (Supplementary Fig. 1b, c), including eight genes: sterol C5 desaturase (SC5D), 24-dehydrocholesterol reductase (DHCR24), DHCR7, squalene epoxidase (SQLE), CYP51A1, methylsterol monooxygenase 1 (MSMO1), farnesyl-diphosphate farnesyltransferase 1 (FDFT1), and hydroxysteroid 17-beta dehydrogenase 7 (HSD17B7; Fig. 1a).

Fig. 1. JTC801 induces CYP51A1 upregulation.

a Screening MIAPaCa2 cells to identify the 8 most differentially expressed genes after 24 h treatment with JTC801 (3.5 μM) using mRNA transcriptomics technology. b Comparative analysis of transcriptomics data in MIAPaCa2 cells after 24 h treatment with JTC801 (3.5 μM) and proteomics data from Wayne analysis in PANC1 cells treated with JTC801 (3.5 μM) for 24 h. c Selecting 6 representative molecules from Wayne’s analysis in (b). d, e Lipidomics-based analysis of metabolites differentially produced and their influence on Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways in SW1990 and PANC1 cells following a 12 h treatment with JTC801 (3.5 μM). f Western blot analysis of CYP51A1 protein expression in PDAC cells treated with JTC801 (3, 3.5, and 4 μM) for 24 h. g Western blot analysis of CYP51A1 protein expression in PDAC cells treated with JTC801 (3.5 μM) for 3, 6, 12, and 24 h. h Western blot analysis of indicated protein expression in MIAPaCa2 and SW1990 cells following 24 h treatment with JTC801 (3.5 μM) or RSL3 (1 μM). i Western blot analysis of indicated protein expression in MIAPaCa2 and SW1990 cells following 24 h treatment with JTC801 (3.5 μM) or H₂O₂ (500 μM). j Western blot analysis of indicated protein expression in MIAPaCa2 and SW1990 cells following 24 h treatment with JTC801 (3.5 μM) or staurosporine (500 nM). Data shown in (f–j) represent three independent experiments and are presented as mean ± SD. Statistical significance was determined using one-way ANOVA with Tukey’s multiple comparisons test. Source data are available in the accompanying Source Data file.

In a prior study, we identified potential JTC801-binding proteins in the human PDAC cell line PANC1¹⁰. Among these proteins, CYP51A1 was found to interact with JTC801 in PANC1 cells (Fig. 1b, c). Additionally, lipid metabolomics analysis of PDAC cells (PANC1 and SW1990) treated with JTC801 for 12 h revealed several alterations in cholesterol products (Supplementary Fig. 1d, e), underscoring changes in cholesterol metabolic pathways (Fig. 1d, e).

CYP51A1, a cholesterol synthase located in the endoplasmic reticulum (ER)¹⁷, was further validated through qPCR and western blot analysis. This validation confirmed that JTC801 treatment led to a dose- and time-dependent upregulation of CYP51A1 expression in human PDAC cell lines (Fig. 1f, g and Supplementary Fig. 1f). We also investigated whether JTC801 influenced the expression of other cell death markers associated with ferroptosis, necroptosis, and apoptosis, such as GPX4, the phosphorylation of MLKL (p-MLKL), and cleaved PARP1 (c-PARP1). JTC801 did not affect the expression of these markers (Fig. 1h–j). Furthermore, unlike these markers, CYP51A1 expression remained unaltered upon the stimulation of ferroptosis, necroptosis, or apoptosis by RSL3, H₂O₂, and staurosporine, respectively (Fig. 1h–j).

Collectively, these results demonstrate that JTC801 not only binds to CYP51A1 but also promotes its upregulation, thereby distinguishing this compound from other cell death inducers.

SREBF2 mediates CYP51A1 upregulation

Given that CYP51A1 is a protein involved in the cholesterol synthesis pathway localized in the ER and that ER stress can induce de novo cholesterol synthesis¹⁸, we investigated whether JTC801 induces ER stress. However, western blot analysis showed no significant changes in ER stress-related markers (e.g., heat shock protein family A [Hsp70] member 5 [HSPA5], endoplasmic reticulum to nucleus signaling 1 [ERN1], eukaryotic translation initiation factor 2 alpha kinase 3 [EIF2AK3], and activating transcription factor 6 [ATF6]) upon treatment with JTC801 (Supplementary Fig. 2a, b). Furthermore, a thioflavin T assay, which detects intracellular unfolded proteins¹⁹, indicated that JTC801 did not lead to an accumulation of unfolded proteins (Supplementary Fig. 2c). We also assessed the impact of inhibitors (4μ8C, ISRIB, and melatonin) that target various components of the ER stress signaling pathway, including ERN1, EIF2AK3, and ATF6, and found that they had no effect on the JTC801-induced protein expression of CYP51A1 (Supplementary Fig. 2d). Additionally, combining these inhibitors with JTC801 did not enhance JTC801-induced cell viability inhibition (Supplementary Fig. 2e).

Previous research has established CYP51A1 as a target gene for SREBF2 in PDAC cells, particularly under conditions of low extracellular pH²⁰. Our analysis of the Cancer Genome Atlas (TCGA) database for PDAC patients further corroborated this, revealing a positive correlation between CYP51A1 and SREBF2 gene expression in tumors (Fig. 2a). As ER cholesterol levels decrease, SREBF2 translocates from the ER to the Golgi apparatus, where it undergoes hydrolytic cleavage and subsequently enters the nucleus to regulate gene transcription^21,22. Consistently, western blot analysis of nuclear extracts revealed that JTC801 promotes the nuclear accumulation of SREBF2 in comparison to SREBF1 (Fig. 2b). Confocal imaging assays also confirmed the augmented nuclear translocation of SREBF2 in PDAC cells after JTC801 treatment (Fig. 2c).

Fig. 2. SREBF2 drives CYP51A1 upregulation.

a Examining the correlation between CYP51A1 and SREBP2 using datasets from the Cancer Genome Atlas (TCGA). b Western blot analysis of protein expression in nuclear (N) or cytosolic (C) extracts in control and CYP51A1-knockdown PDAC cells. (p) = precursor; (m) = mature. c Immunofluorescence analysis of the nuclear localization of SREBF2 in the presence or absence of JTC801 treatment (3.5 µM, 24 h). n = 3 independent samples, each derived from approximately 10 cells. Scale bar = 50 μm. d Measuring ER cholesterol levels after treatment with JTC801 (3.5 µM, 24 h). e Measuring total cholesterol levels after treatment with JTC801 (3.5 µM, 24 h) in control and CYP51A1-knockdown cells. f Western blot analysis of CYP51A1 protein expression in indicated PDAC cells following treatment with JTC801 (3.5 µM) in the absence or presence of cholesterol (15 µM) and/or MβCD (1 mM) for 24 h. g Western blot analysis of protein expression in indicated PDAC cells following treatment with JTC801 (3.5 µM) for 24 h. Data in (b, d, and f, g) represent three independent experiments and are presented as mean ± SD. Statistical significance for panels (d, f, g) was determined using one-way ANOVA, and for (b) using two-way ANOVA, both followed by Tukey’s multiple comparisons test. Data in (c–e) are derived from 3 or more biologically independent samples (d, n = 5) and presented as mean ± SD. Statistical significance was assessed using a two-sided unpaired t-test. Source data are available in the accompanying source data file.

A key trigger for SREBF2 activation is the reduction of ER cholesterol levels^23,24. Quantitative assays revealed that treatment with JTC801 led to a decrease in ER cholesterol levels (Fig. 2d), while intracellular total cholesterol levels remained unaltered (Fig. 2e). The cholesterol scavenger MβCD²⁵ also resulted in the upregulation of CYP51A1 protein (Fig. 2f). In contrast, supplementation with cholesterol inhibited the protein expression of CYP51A1 (Fig. 2f), further suggesting that CYP51A1 expression is changed by cholesterol levels.

To determine whether SREBF2 is required for JTC801-induced CYP51A1 expression, we performed RNAi experiments with two specific shRNAs targeting SREBF2 or SREBF1 in two PDAC cell lines (Supplementary Fig. 3a, b). The suppression of SREBF2 expression blocked JTC801-induced CYP51A1 protein upregulation (Fig. 2g). A luciferase reporter gene assay further confirmed that CYP51A1 is a direct target gene of SREBF2 (Supplementary Fig. 3c). In contrast, the knockdown of SREBF1 had no effect on JTC801-induced CYP51A1 protein upregulation (Supplementary Fig. 3d). Thus SREBF2, rather than SREBF1, is required for the inducible expression of CYP51A1 in PDAC cells following JTC801 treatment.

SREBF2-dependent CYP51A1 expression inhibits JTC801-induced cell death

To investigate the role of upregulated CYP51A1 in JTC801-induced cell death, we used shRNA or siRNA to suppress CYP51A1 expression in SW1990, MIAPaCa2, and PANC1 cells (Supplementary Fig. 4a, b). This resulted in increased JTC801-induced growth inhibition compared to control groups (Fig. 3a and Supplementary Fig. 4c), supporting the notion that CYP51A1 functions as a repressor of JTC801-induced cytotoxicity. Propidium iodide (PI) staining of cell membrane damage as well as colony formation assays in CYP51A1-knockdown PDAC cells further confirmed this role (Fig. 3b, c and Supplementary Fig. 4d, e). In contrast, the overexpression of CYP51A1 through gene transfection in SW1990 and MIAPaCa2 cells reduced JTC801-induced growth inhibition and cell death (Fig. 3d, e and Supplementary Fig. 4f, g). Cell cloning experiments validated that CYP51A1 overexpression reduced JTC801-induced cytotoxicity (Supplementary Fig. 5a).

Fig. 3. SREBF2-dependent CYP51A1 expression inhibits cell death.

a Cell viability analysis of indicated PDAC cells after treatment with JTC801 for 24 h. b Cell death assay of indicated PDAC cells following treatment with JTC801 (3.5 µM) for 24 h. c Colony formation analysis of indicated PDAC cells after treatment with JTC801 (3 µM) for 24 h. d Cell viability analysis of indicated PDAC cells after treatment with JTC801 for 24 h. e Cell death assay of indicated PDAC cells following treatment with JTC801 (3.5 µM) for 24 h. f, g Intracellular pH in indicated PDAC cells after treatment with JTC801 (3.5 µM) for 24 h. n = 10. h Cell viability analysis of indicated MIAPaCa2 and SW1990 cells following treatment with JTC801 (3.5 μM), RSL3 (1 μM), STS (500 nM), and H₂O₂ (500 μM) for 24 h. Ferrostatin-1 (1 μM) was used as a positive control to inhibit RSL3-induced growth inhibition. n = 6. i Cell viability analysis of indicated PDAC cells after treatment with JTC801 for 24 h. j Intracellular pH in indicated PDAC cells after treatment with JTC801 (3.5 µM) for 24 h. n = 8. k Cell viability assay of indicated PDAC cells following treatment with JTC801 for 24 h. Data in panels a-k are derived from three or more biologically independent samples (f, g, n = 10; h, n = 6; j, n = 8), presented as mean ± SD. Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons test. Source data are available in the accompanying Source Data file.

Intracellular pH assays demonstrated that CYP51A1 expression inhibits JTC801-induced intracellular alkalinization (Fig. 3f, g and Supplementary Fig. 6a, b). In contrast, CYP51A1-knockdown cells showed no significant effects on sensitivity to ferroptosis, necroptosis, or apoptosis (Fig. 3h). Thus, CYP51A1 serves as a specific regulator of JTC801-induced cell death in PDAC cells.

We further investigated the relationship between CYP51A1 and SREBF2 in controlling JTC801-induced cytotoxicity. The knockdown of SREBF2 not only increased the growth inhibition induced by JTC801 but also enhanced intracellular alkalinization (Fig. 3i, j). The enhanced sensitivity to JTC801 observed in SREBF2-deficient cells was reversed upon knock-in of CYP51A1 (Fig. 3k and Supplementary Fig. 5b). These findings support that CYP51A1 acts as a downstream effector of SREBF2-mediated cell death resistance.

CYP51A1 prevents the accumulation of cholesterol in lysosomes

Under normal conditions, the movement of cholesterol between lysosomes and the ER maintains cholesterol balance²³. To visualize cholesterol distribution, we used filipin, a fluorescent antibiotic that binds to cholesterol but not esterified sterols²⁶. Filipin staining revealed that inhibiting CYP51A1 resulted in cholesterol accumulation (Fig. 4a and Supplementary Fig. 7a), similar to the effect observed with U18666A (a lysosomal cholesterol transport inhibitor) (Supplementary Fig. 7b)²⁷. Conversely, CYP51A1 overexpression reduced cholesterol accumulation (Fig. 4b and Supplementary Fig. 7b). The loss of CYP51A1 resulted in lysosomal cholesterol accumulation (Fig. 4c and Supplementary Fig. 7c).

Fig. 4. CYP51A1 prevents the accumulation of cholesterol in lysosomes.

a, b The indicated PDAC cells were treated with JTC801 (3.5 µM) for 24 h following a 2 h incubation with 0.5 mg/mL of filipin. Fluorescence microscopy was used to visualize free cholesterol in blue without excitation of the light field, and Image J software was utilized to analyze the average fluorescence intensity. Representative images are shown in the Supplementary Figs. c Fluorescence confocal microscopy demonstrated the co-localization of lysosomes with cholesterol, and Image J software was used for fluorescence co-localization analysis. Representative images are shown in the Supplementary Figs. d The indicated PDAC cells were treated with JTC801 (3.5 µM) in the absence or presence of bafilomycin A1 (Baf A1; 100 nM) or MβCD (1 mM) for 24 h. Fluorescence confocal microscopy showed co-localization of lysosome with cholesterol, and Image J software was used for fluorescence co-localization analysis. Representative images are shown in the Supplementary Figs. e Cell viability assay of indicated PDAC cells following treatment with JTC801 in the absence or presence of bafilomycin A1 (Baf A1; 100 nM) for 24 h. f Cell viability assay of indicated PDAC cells following treatment with JTC801 in the absence or presence of MβCD (1 mM) for 24 h. g Cell viability assay of indicated PDAC cells following treatment with JTC801 in the absence or presence of water-soluble cholesterol (15 µM) for 24 h. h Cell death assay of indicated PDAC cells following treatment with JTC801 (3.5 µM) in the absence or presence of water-soluble cholesterol (15 µM) for 24 h. Representative images are shown in the Supplementary Figs. Data in (a–h) are derived from three biologically independent samples and presented as mean ± SD. Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons test. Source data are available in the accompanying Source Data file.

Functionally, treatment with the cholesterol scavenger MβCD or bafilomycin A1 (a lysosomal acidification inhibitor) prevented lysosomal cholesterol buildup and reversed JTC801-induced growth inhibition (Fig. 4d–f and Supplementary Fig. 7d). Adding cholesterol, in contrast, increased JTC801-induced growth inhibition and cell death (Fig. 4g, h, and Supplementary Fig. 7e). Pre-complexing MβCD and cholesterol at a 1:1 ratio is a common method for delivering cholesterol to membranes^25,28, and cell viability assays showed a similar increase in JTC801-induced cell growth inhibition in both the CYP51A1 knockdown and overexpression groups (Supplementary Fig. 7f).

We further explored if CYP51A1 deficiency-induced cholesterol accumulation relates to changes in cholesterol transporters. Western blot analysis revealed no impact on known lysosomal cholesterol transporters, including NPC intracellular cholesterol transporter 1 (NPC1)²⁹, oxysterol binding protein (OSBP)³⁰, and solute carrier family 38 member 9 (SLC38A9)²⁵ (Supplementary Fig. 8a–d). However, our results suggest that JTC801-induced lysosomal cholesterol accumulation results from pH imbalances, as bafilomycin A1 limited this accumulation (Fig. 4d). Consistently, knockdown of CYP51A1 resulted in upregulation of ATP6V0D1 (Supplementary Fig. 8e), further indicating a link between lysosomal pH and cholesterol homeostasis. Additionally, transmembrane protein 55B (TMEM55B) is crucial for maintaining lysosomal lipid homeostasis and lysosomal function³¹. Western blot results showed that knockdown of CYP51A1 reduced the protein level of TMEM55B (Supplementary Fig. 8f). Collectively, these findings indicate that impaired lysosomal pH and TMEM55B expression may contribute to CYP51A1-dependent inhibition of lysosomal cholesterol accumulation.

Furthermore, the knockdown of CYP51A1 reduced the expression level of GRAM domain-containing 1 C (GRAMD1C), a cholesterol transporter involved in the transfer between the ER and plasma membrane (Supplementary Fig. 8g, h), indicating that JTC801 not only induces lysosomal cholesterol accumulation, but also destroys plasma membrane cholesterol homeostasis, which was consistent with the results that adding to plasma membrane cholesterol also increased JTC801-induced cell viability inhibition (Supplementary Fig. 7f). Mechanistic target of rapamycin (MTOR) kinase is a multifunctional protein that plays a crucial role in lysosomal biosynthesis. Although TMEM55B contributes to amino acid-induced MTOR activation³², JTC801 failed to affect MTOR or its phosphorylation in the presence or absence of CYP51A1 knockdown (Supplementary Fig. 9a). Thus, this process mediated by CYP51A1 is MTOR-independent, although it can affect TMEM55B expression in some cases.

While ferroptosis primarily relies on membrane lipid peroxidation as a cell death mechanism, N-acetylcysteine inhibited both ferroptosis and alkaliptosis^4,8. However, MβCD failed to prevent RSL3-induced ferroptosis and, conversely, impeded the inhibitory effect of JTC801 (Fig. 5a). Additionally, the ferroptosis inducer RSL3 inhibited GPX4 protein expression in a dose-dependent manner without affecting CYP51A1 (Fig. 5b), aligning with previous findings that GPX4 degrades during ferroptosis^33,34. This observation further underscores the differences between cell death induced by JTC801 and RSL3.

Fig. 5. Impact of CYP51A1 on various cell death pathways.

a Cell viability analysis of indicated MIAPaCa2 and SW1990 cells following treatment with RSL3 (1 μM) or JTC801 (3.5 μM) in the presence or absence of MβCD (1 mM) or ferrostatin-1 (Ferr-1; 1 μM) for 24 h. n = 6. b Western blot analysis of indicated protein expression in PDAC cells treated with RSL3 (0, 1, 1.25, and 1.5 μM) for 24 h. c Cell viability assay of indicated PDAC cells following treatment with JTC801 in the absence or presence of lovastatin (15 μM) for 24 h. Data in (a–c) are derived from three or more biologically independent samples (a, n = 6) and are presented as mean ± SD. Statistical significance was determined using two-way ANOVA with Tukey’s multiple comparisons test. Source data are available in the accompanying Source Data file.

The ER synthesizes cholesterol and other lipids, which are then distributed to various cellular compartments, including lysosomes. Next, bioinformatics analysis indicated a positive correlation between lysosomal-associated membrane protein 1 (LAMP1)/LAMP2 and CYP51A1 (Supplementary Fig. 9b, c). Furthermore, combining lovastatin, a 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR) inhibitor that disrupts cholesterol biosynthesis³⁵, with JTC801 treatment did not enhance growth inhibition in the CYP51A1-knockdown cells compared to control cells (Fig. 5c).

In summary, these findings indicate that CYP51A1 inhibits anticancer activity of JTC801 primarily by accumulating cholesterol in lysosomes.

TMEM175-mediated lysosomal proton efflux inhibits cell death

Given that CYP51A1 is not a direct lysosomal proton transporter, we hypothesize that CYP51A1 inhibition may be linked to lysosomal cholesterol accumulation. This accumulation could potentially lead to the blockage of lysosomal proton efflux channels, such as TMEM175^36–38. Western blot analysis demonstrated that the knockdown of CYP51A1 led to a suppression of TMEM175 expression, whereas the overexpression of CYP51A1 increased TMEM175 expression in response to JTC801 (Supplementary Fig. 10a, b). Furthermore, when cholesterol levels were replenished, which inhibited CYP51A1 expression, there was a decrease in TMEM175 expression. In contrast, cholesterol removal using MβCD restored TMEM175 expression (Fig. 6a and Supplementary Fig. 10c). In addition, western blot results showed that the addition of oleic acid did not restore the protein expression of TMEM175 in CYP51A1-knockdown cells, indicating the role of cholesterol, rather than fatty acids, in regulating TMEM175 expression (Supplementary Fig. 10d).

Fig. 6. TMEM175-mediated lysosomal proton efflux inhibits cell death.

a Western blot analysis of protein expression in indicated PDAC cells following treatment with JTC801 (3.5 µM) in the absence or presence of cholesterol (15 µM) and/or MβCD (1 mM) for 24 h. n = 6. b Gas chromatography-mass spectrometer quantification of dihydrolanosterol (Dhl) and lanosterol levels in MIAPaCa2 cell after JTC801 (3.5 µM) for 24 h. c Western blot analysis of protein expression in indicated PDAC cells following treatment with JTC801 (3.5 µM) in the absence or presence of lanosterol (20 µM) and/or MβCD (1 mM) for 24 h. d Cell viability assay of indicated PDAC cells following treatment with JTC801 in the absence or presence of lanosterol (20 µM) for 24 h. e Cell viability assay of indicated PDAC cells following treatment with JTC801 for 24 h. f Cell death assay of indicated PDAC cells following treatment with JTC801 for 24 h. g Cell viability assay of indicated PDAC cells following treatment with JTC801 for 24 h. Data in (a–g) are derived from three biologically independent samples and presented as mean ± SD. Statistical significance for (a, c, d–g) was determined using two-way ANOVA with Tukey’s multiple comparisons test, while (b) used a two-sided unpaired t-test. Source data are available in the accompanying Source Data file.

CYP51A1 catalyzes lanosterol as a substrate for subsequent reactions. As expected, liquid chromatography-mass spectrometry confirmed that JTC801 inhibited intracellular lanosterol levels (Fig. 6b). Similar to cholesterol, lanosterol inhibited TMEM175 expression and induced cell growth inhibition (Fig. 6c, d). However, unlike cholesterol, the association with MβCD did not inhibit TMEM175 expression (Fig. 6c, d). To further elucidate the role of lanosterol in cell death, we knocked down the upstream enzyme lanosterol synthase (LSS) and the downstream enzyme transmembrane 7 superfamily member 2 (TM7SF2) of CYP51A1, respectively (Supplementary Fig. 10e). Cell viability assays revealed that knockdown of LSS partially inhibited JTC801-induced cytotoxicity, whereas knockdown of TM7SF2 increased it (Supplementary Fig. 10f). In addition, individual inhibition of TM7SF2 and LSS had little effect on cholesterol levels (Supplementary Fig. 10g), supporting previous findings that cholesterol metabolism regulation involves multiple pathways³⁹.

A qPCR assay indicated that the mRNA expression of TMEM175 was not affected by lanosterol (Supplementary Fig. 10h). Subsequent protein degradation assays, in conjunction with the ubiquitin–proteasome system (UPS) inhibitor MG132 and autophagy inhibitor hydroxychloroquine, indicated that lanosterol inhibits TMEM175 protein expression through ubiquitin-proteasomal degradation (Supplementary Fig. 10i, j). However, treatment with lanosterol did not appear to have direct effects on proteasome activity in pancreatic cancer cells (Supplementary Fig. 10k), indicating that lanosterol-induced TMEM175 proteasomal degradation may occur through increased E3 ligase-mediated TMEM175 ubiquitination.

Functionally, the overexpression of TMEM175 in CYP51A1- or SREBF2-knockdown PDAC cells inhibited JTC801-induced cell death and growth inhibition (Fig. 6e–g and Supplementary Fig. 11a–d). Ectopic expression of TMEM175 in CYP51A1-knockdown cells also alleviated pH alkalinization (Supplementary Fig. 11e). To further investigate JTC801-induced pH changes, we utilized a pH-sensing protein, SMAD family member 5 (SMAD5), which translocates from the nucleus to the cytoplasm during intracellular alkalization^40,41. Western blot analysis showed that JTC801 (6–12 h) induced elevated cytosolic levels of SMAD5 (Supplementary Fig. 11f, g), and immunofluorescence confirmed a similar effect (Supplementary Fig. 12a). However, JTC801 treatment for 1 h had little effect on pH and cytosolic SMAD5 levels (Supplementary Fig. 12b, c). Intracellular pH changes were assessed using the pH-sensitive fluorescent dye BCECF-AM. First, we performed intracellular calibration of BCECF-AM, establishing the relationship between intracellular pH and the normalized fluorescence ratio at pH 7.0, based on the emitted light intensities at 490 nm and 440 nm excitation wavelengths (with emission at 528 nm) (Supplementary Fig. 13a). Our results revealed an initial decrease in intracellular pH at 6 h, followed by an increase at 16 h of JTC801 treatment (Supplementary Fig. 13a). Additionally, PI staining revealed that JTC801 did not cause cell membrane rupture at 6 h but induced significant membrane rupture at 12–24 h (Supplementary Fig. 13b). These findings suggest that JTC801 triggers distinct intracellular pH changes at early and late stages, contributing to its mechanism of action. Furthermore, JTC801 treatment enhanced lysosomal acidification, as assessed using the pH-sensitive LysoSensor Yellow/Blue DND-160 (Supplementary Fig. 13c). This effect was more pronounced in CYP51A1 knockdown cells (Supplementary Fig. 13c). Lysosomal acidification was partially reversed following the ectopic expression of TMEM175 in CYP51A1 knockdown cells (Supplementary Fig. 13c).

Consistent with our previous studies⁸, N-acetyl alanine (NAA), a potential acidic agent, inhibited JTC801-induced growth inhibition (Supplementary Fig. 13d). LC-MS/MS analysis revealed minimal accumulation of JTC801 in isolated lysosomes compared to its total levels in whole cells (Supplementary Fig. 13e). Furthermore, NAA did not affect JTC801’s lysosomal accumulation (Supplementary Fig. 13e) or upstream processes, including CYP51A1 gene upregulation and changes in lanosterol levels observed at 1 and 6 h (Supplementary Fig. 13f and 13g). These findings suggest that early lysosomal accumulation of JTC801 is not essential for its cytotoxic activity. Additionally, attempts to couple JTC801 to the fluorescent moiety Cyanine5 to assess cellular distribution were unsuccessful. The low quinoline amino activity of JTC801 led to poor coupling efficiency, resulting in the failure of various coupling methods with Cyanine5 acid. (Supplementary Fig. 14a, b and Supplementary Fig. 15).

Collectively, these results indicate that TMEM175 functions as a downstream effector of SREBF2-dependent CYP51A1 expression, playing a role in inhibiting JTC801-induced cell death by facilitating lysosomal proton efflux.

Targeting the CYP51A1 pathway increases anticancer activity of JTC801 in vivo

To investigate whether targeting the CYP51A1 pathways enhances the anticancer activity of JTC801, we initially employed human xenograft models. PANC1 cells with either control or CYP51A1-knockdown were subcutaneously injected into immunocompromised NOD-SCID mice for 10 days, followed by oral administration of JTC801 at doses of 5 or 20 mg/kg (Fig. 7a). JTC801 at 20 mg/kg almost completely inhibited tumor growth, irrespective of whether the cells were control or CYP51A1-knockdown (Fig. 7b). In contrast, at a lower dose of 5 mg/kg, JTC801 displayed enhanced anticancer activity in the CYP51A1- knockdown group compared to the control group (Fig. 7b). Furthermore, ketoconazole, a potential CYP51A1 inhibitor, enhanced the anticancer activity of JTC801 at 5 mg/kg in the PANC1 xenograft model (Fig. 7c).

Fig. 7. Targeting the CYP51A1 pathway increases the anticancer activity of JTC801 in vivo.

a A schematic diagram illustrating the experimental setup involving the subcutaneous implantation of human WT or CYP51A1-KD PANC1 cells in immunocompromised NOD-SCID mice, followed by a 3-week treatment protocol. b Tumor growth curves of WT and KD PANC1 cells implanted subcutaneously in NOD-SCID mice (n = 5 per group) following the treatment regimen described in (a). c Tumor growth curves of WT PANC1 cells implanted subcutaneously in NOD-SCID mice (n = 5 per group) subjected to JTC801 treatment (5 mg/kg, administered orally once daily for 5 days per week) in the presence or absence of ketoconazole (30 mg/kg, administered orally once daily for 5 days per week). d A schematic diagram illustrating the experimental setup involving the orthotopic implantation of WT or CYP51A1-KD KPC cells in C57BL6/J mice, followed by a 3-week treatment protocol. e Survival curves of indicated mice treated with vehicle, JTC801, and/or ketoconazole (n = 10 mice/group). f–l Analysis of pancreatic weight, metastases to the liver and lung, and indicated tumor-related markers in the pancreas of the syngeneic orthotopic model (n = 5–10 mice/group). Scale bar = 100 μm. m The relationship between CYP51A1 mRNA expression and disease-free survival in pancreatic cancer patients was assessed using the GEPIA online tool (http://gepia.cancer-pku.cn/), which utilizes the TCGA database. The selected cut-off was the optimal choice. n A schematic diagram illustrating the experimental setup involving the subcutaneous implantation of PDXs in immunocompromised NOD-SCID mice, followed by a 3-week treatment protocol. o Western blot analysis of CYP51A1 expression in PDAC tumors exhibiting either high or low levels of expression. p Tumor growth curves of indicated PDXs implanted subcutaneously into NOD-SCID mice (n = 5 mice per group). The P values were calculated by two-way ANOVA (b, c, j, f, i–l) or log-rank tests (e, f, g). Source data are available in the accompanying source data file.

Subsequently, we examined whether the loss of Cyp51a1 also increased JTC801-induced tumor suppression in syngeneic orthotopic models. In these models, wild-type and Cyp51a1-knockdown mouse KPC cells were implanted into the pancreas of immune-intact C57BL/6 J mice (Fig. 7d). KPC cell lines, derived from pancreatic tumors of Kras^G12D;Tp53^R172H;Pdx1-Cre mice, serve as a widely-used model for studying drug responses⁴². JTC801 displayed increased efficacy in extending animal survival in the Cyp51a1-knockdown group compared to the control group (Fig. 7e). In subsequent assays, we measured pancreatic weight (Fig. 7f), tumor metastasis to liver and lung (Fig. 7g), intracellular pH levels (Fig. 7h), proliferation marker Ki67 (Fig. 7i), CD8⁺ T cell infiltration (Fig. 7j), alpha-smooth muscle actin (α-SMA)⁺ cancer-associated fibroblasts (CAFs) (Fig. 7k), and performed Trichrome staining (Fig. 7l) in the pancreas of the syngeneic orthotopic model. These assays confirmed that suppressing CYP51A1 enhances the anticancer efficacy of JTC801 in vivo.

An analysis of the TCGA database revealed a trend toward lower survival time in patients with high CYP51A1 expression, although the difference was not statistically significant (P = 0.13; Fig. 7m). To investigate whether the expression of CYP51A1 is linked to the response to JTC801 in PDAC patients, we further utilized patient-derived xenografts (PDXs; Fig. 7n)⁴³. In PDX models, PDAC tissues with lower CYP51A1 expression showed a more positive response to JTC801 when compared to those with higher CYP51A1 expression (Fig. 7o, p).

Collectively, these findings from various mouse PDAC models provide in vivo evidence that both genetic and pharmacological inhibition of CYP51A1 enhances JTC801-induced tumor suppression.

Discussion

Pancreatic cancer presents formidable medical challenges with limited therapeutic options. While the precise causes of its low survival rates are multifaceted, metabolic reprogramming stands as a well-recognized hallmark of pancreatic cancer, influencing both tumor development and responses to treatment⁴⁴. The induction of pH-dependent form of alkaliptosis has emerged as a strategy for anticancer therapy, particularly within the context of PDAC⁴⁵. In this study, we unveiled that CYP51A1 acts as a negative regulator of JTC801-induced intracellular alkalization and cell death in PDAC by modulating TMEM175-mediated lysosomal proton efflux (Supplementary Fig. 16). These findings expose a mechanism that bridges lysosomal cholesterol metabolism to cellular susceptibility to cell death and furnish preclinical evidence that targeting the CYP51A1 pathway could enhance the effectiveness of JTC801-induced tumor suppression.

The reversal of the pH gradient is a feature of cancer cells, mainly driven by various proton transporters and channels responsible for excluding excess intracellular acid produced during aerobic glycolysis⁴⁶. This reversed pH gradient not only supports cancer cell growth but also creates a suitable metabolic environment for it⁴⁷. Conversely, changes in metabolic pathways can influence pH levels, resulting in a complex interplay between metabolism and pH regulation⁴⁸. Targeting the pH gradient within tumors is an emerging strategy in cancer therapy. Our findings suggest that JTC801-induced disruption of lysosomal pH leads to the accumulation of lysosomal lipids, subsequently activating CYP51A1-dependent TMEM175-mediated lysosomal proton release. The inhibition of CYP51A1 enhances JTC801-induced cell death due to the accumulation of lanosterol, which in turn inhibits TMEM175.

SREBF2, a transcription factor crucial for maintaining cellular cholesterol homeostasis, serves as an oncogenic driver in tumor progression, including in pancreatic cancer^49,50. Enhanced SREBP2 activation contributes to chemotherapy resistance through multiple mechanisms. For instance, SREBP2 has been implicated in cisplatin resistance in ovarian cancer cells by inhibiting apoptosis⁵¹. Our study reveals that the upregulation of CYP51A1 by JTC801 is reliant on SREBF2. Enhanced sensitivity to JTC801-induced cytotoxicity in SREBF2-deficient cells is mitigated by ectopic expression of CYP51A1, highlighting its role as a target gene mediating SREBF2-dependent effects. Therefore, targeting SREBF2-dependent CYP51A1 expression may potentiate the response of pancreatic cancer cells to JTC801 treatment.

Previous studies have demonstrated that deleting Cyp51a1 in mice upregulates multiple SREBF2-related cholesterol synthesis genes⁵². We observed that JTC801 increases the nuclear translocation of SREBF2, and knockdown of CYP51A1 also increases SREBF2 levels. These findings indicate a potential negative feedback pathway between SREBF2 and CYP51A1. We also found that knockdown of SREBF2 exhibits a phenotype similar to CYP51A1 inhibition, while knockdown of upstream (e.g., LSS) or downstream (e.g., TM7SF2) enzymes of CYP51A1 only partially affects JTC801 sensitivity. These findings suggest the need for further investigation into the involvement of other cholesterol intermediate metabolites, lipids or non-classical functions of CYP51A1 in the regulation of cell death. Since JTC801 also directly binds to CYP51A1, it suggests a dual pathway—both transcription-dependent and transcription-independent—regulating CYP51A1 expression and function during cell death. Moreover, pH-dependent lysosomal cholesterol levels may regulate the expression of pro- and anti-cell death proteins involved in cellular homeostasis, such as CYP51A1 in coordination with ATP6V0D1 expression.

Cholesterol metabolism and fatty acid metabolism are two essential lipid pathways that are intricately interconnected, often compensating for each other when one is disrupted⁵³. Our research indicates that CYP51A1 primarily operates within the cholesterol metabolic pathway. CYP51A1 prevents the abnormal accumulation of lanosterol, while lanosterol inhibits TMEM175-dependent lysosomal proton efflux, leading to cell death. However, it remains uncertain whether the inhibition of CYP51A1 activates the fatty acid pathway or if other cholesterol pathways, such as low-density lipoprotein receptor-mediated lipoprotein uptake, play a role in regulating JTC801-induced cytotoxicity.

Lanosterol has been reported to influence protein degradation in a context-dependent manner. For instance, lanosterol can stimulate the ubiquitination and subsequent degradation of HMG-CoA reductase⁵⁴. Additionally, lanosterol regulates proteostasis by solubilizing cell membrane sequestosomes/aggresome-like inducible structures⁵⁵. Our study demonstrated that the UPS system is involved in the degradation of TMEM175 by lanosterol. However, lanosterol did not affect the activity of the 20S proteasome. Further investigation is required to determine if lanosterol regulates the expression of specific E3 ligases or deubiquitinating enzymes for TMEM175. Moreover, while classical lysosomal cholesterol transport proteins (e.g., NPC1, OSBP, and SLC38A9) may not be involved in JTC801-mediated lysosomal cholesterol accumulation, other proteins such as autophagy-related 2 (ATG2), fatty acid-binding proteins (FABP), Wnt family member 5 A (WNT5A), and scavenger receptor class B member 2 (SCARB2) are implicated in lysosomal cholesterol transport^30,56–58. The role of these proteins in the regulation of JTC801-induced cytotoxicity remains unclear and warrants further study.

TMEM175 is a unique lysosomal potassium channel with little sequence similarity to other potassium channels³⁷. While it is primarily implicated in brain diseases like Parkinson’s, its role in cancer biology is poorly understood⁵⁹. We found that TMEM175 overexpression in CYP51A1-knockdown cells reversed JTC801-induced cell death in PDAC, suggesting that TMEM175 could be a potential target for pH-related tumor therapy. Ectopic TMEM175 expression only partially rescued JTC801-induced cell death, implying the involvement of other mechanisms. Indeed, CYP51A1 inhibition also affected the expression of GRAMD1C, a protein that mediates cholesterol transport between the plasma membrane and the ER. A deeper understanding of the relationship between cholesterol transport and lysosomal ion channels is crucial for comprehending lysosomal pH homeostasis in regulating cell survival and death.

Intracellular pH changes are rapid and sensitive, making current commercial pH kits insufficient for accurately representing real-time cellular pH values. Although the intranuclear distribution of SMAD5 can assess intracellular pH changes, factors such as temperature and osmotic pressure also influence its distribution. CA9 functions as an H⁺ efflux transporter, and its inhibition should theoretically lead to intracellular acidification⁶⁰. However, the downregulation of CA9 by JTC801 has been shown to increase intracellular pH⁸, indicating the activation of additional pH regulators as an adaptive mechanism⁶¹. Therefore, pH imbalance during cell death is regulated by various transporters and proteins. Consequently, there is a need to develop more accurate pH assay methods to monitor the network of different regulators involved in cell death.

Our results suggest that CYP51A1 does not affect RSL3-induced ferroptosis sensitivity in PDAC cells. However, a recent study demonstrated that inhibition of CYP51A1 increased ferroptosis sensitivity in HEK293T cells³⁹. This discrepancy may be attributed to the heterogeneity of tumor cells, which could influence ferroptosis sensitivity. The inhibition of CYP51A1 did not affect total cellular cholesterol content during ferroptosis³⁹. We demonstrated that CYP51A1 can influence the accumulation of cholesterol in different organelles following JTC801 treatment. We also demonstrated that cholesterol and lanosterol did not affect RSL3 sensitivity, consistent with previous reports on HEK293T cells³⁹. Although the lysosome is a final station of many cell death pathways⁶², we found that CYP51A1 is a repressor of JTC801-induced cytotoxicity by preventing lysosomal cholesterol accumulation and pH imbalance. Further understanding the profile of lysosomal lipid and signals in different types of cell death could lead to new anticancer strategies.

We observed that JTC801 treatment led to an early decrease in intracellular pH without evidence of cell membrane damage. This early reduction in cytosolic pH may represent a cellular defense mechanism in response to JTC801-induced stress. As this defense system becomes compromised, the cytosolic pH subsequently rises, ultimately triggering cell death. This pattern mirrors that of oxidative stress-related cell death, where an initial antioxidant response is followed by a later increase in reactive oxygen species (ROS). Lysosomes, as critical regulators of cellular pH homeostasis, play a role in various forms of cell death^10,62–65. Understanding the dynamic interplay between early defense responses, subsequent cellular injury, and the involvement of lysosomes and other organelles is crucial for developing targeted strategies to induce controlled cell death.

One limitation of this study is that the intracellular distribution of JTC801 remains unknown. Although we demonstrated that JTC801 cannot be labeled with Cy5, further studies are warranted to explore other fluorescent moieties. Additionally, each pH-sensing probe or protein used in this study has limitations regarding the evaluated pH range, which may impact data interpretation. Despite these challenges, identifying the function of different JTC801-binding proteins in pH regulation and cell death is crucial.

In summary, our findings indicate that the CYP51A1 pathway acts as a negative feedback mechanism that mitigates JTC801-induced cell death in PDAC cells. Considering the crucial roles of cholesterol metabolism and lysosomal dysfunction in tumor progression, further investigation into their complex interactions promises to enhance our understanding of cell death mechanisms and aid in developing targeted therapies.

Methods

The reagents are described in Supplementary Table 1.

Cell culture

Human PDAC cell lines, including SW1990 (CRL-2172), PANC1 (CRL-1469), and MIAPaCa2 (CRL-1420), were acquired from the American Type Culture Collection. These cells underwent cultivation in Dulbecco’s modified Eagle’s medium (DMEM; Thermo Fisher Scientific, 11995073) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, A3840001) and 1% streptomycin/penicillin (Yeasen, 60162ES76). Cell cultures were maintained within a controlled incubator environment set at 37 °C, with a relative humidity of 95% and a CO₂ concentration of 5%.

Regular screening for mycoplasma contamination was conducted using short tandem repeat assays to confirm the absence of mycoplasma in all cell lines. In cases where dimethyl sulfoxide (DMSO) was employed as a drug solubilization agent, the final concentration of DMSO in the working solution was carefully regulated to remain below 0.01%. Furthermore, 0.01% DMSO was utilized as a carrier control in corresponding cell assays.

Western blot

Cytosolic and nuclear proteins were isolated from cells using a commercial kit (Solarbio Life Science, R0050) following the manufacturer’s protocol. Whole-cell proteins were extracted using 1× cell lysis buffer (Biosharp, BL509A) supplemented with protease inhibitor (ROCHE, 11836153001) on ice for 30 min. Following centrifugation at 15,000 × g for 10 min at 4 °C, the resulting supernatants were collected and quantified using a bicinchoninic acid assay (BCA; Thermo Fisher Scientific, 23225).

For protein separation, an appropriate amount (typically 30 μg) of the protein sample was resolved on either a 10% or 12.5% polyacrylamide gel electrophoresis gel (Epizyme, PG112) and subsequently transferred onto a polyvinylidene fluoride membrane (Millipore, IPVH00010). After blocking in TBST containing 5% skim milk for 1 h, the membrane was incubated with various primary antibodies (diluted 1:500–1:1000) overnight at 4 °C. Following primary antibody incubation, the membrane was treated with peroxidase-conjugated secondary antibodies (diluted 1:1000) for 1 h at room temperature, followed by washing with TBST five times for 5 min each. The protein signals were visualized using enhanced chemiluminescence (Thermo Fisher Scientific, 34095). Blots were subsequently analyzed using the ChemiDoc Touch Imaging System (Bio-Rad) and Image Lab Software (Bio-Rad).

The qPCR analysis

Total RNA was extracted using the RNeasy Plus Micro Kit (QIAGEN, 74034) following the manufacturer’s instructions. Briefly, cells were lysed and homogenized in highly denatured Buffer RLT Plus containing guanidine isothiocyanate. The cell lysate passed through a gDNA Eliminator column to eliminate double-stranded DNA, and total RNA was subsequently purified using a RNeasy MinElute centrifuge column.

To synthesize first-strand cDNA, 1 µg of RNA was used with PrimeScript RT Master Mix (Takara, RR036A). The resulting cDNA was then used in qPCR reactions, which consisted of 4 μl of PrimeScript RT Master reaction mix, 2 μl of gene-specific enhancer solution, 1 μl of reverse transcriptase, 1 μl of gene-specific assay pool (20×, 2 μM), and 12 μl of RNA diluted in RNase-, DNase-, and genomic DNA-free water. The qPCR was performed using SsoFast EvaGreen Supermix (Bio-Rad, 172-5204) on the C1000 Touch Thermocycler CFX96 Real-Time System (Bio-Rad) according to the manufacturer’s protocol. Data analysis was carried out using Bio-Rad CFX Manager software 3.1 (Bio-Rad).

The data were normalized to GAPDH RNA levels, and the fold change was calculated using the 2^−ΔΔCt method. Relative mRNA concentrations were expressed in arbitrary units relative to the untreated group, which was assigned a value of 1. The primers used for GAPDH were 5′-ATCACCATCTTCCAGGAGCGA-3′ and 5′-CCTTCTCCATGGTGGTGAAGAC-3′. The primers for CYP51A1 were 5’-TGTAAAACGACGGCCAGT and 5’-CAGGAAACAGCTATGACC. The primers for TMEM175 were 5’-CAACGCATGCTCAGCTTCAG and 5’-ATATCTTCGCAGGGCCACAC. The primers for LSS were 5’- GACGACCGATTCACCAAGAGCA and 5’- AGACATGCTCCTGGAAGGCAGT. The primers for TM7SF2 were 5’- GGTCAATGGCTTCCAGTTGCTC and 5’- AACGCCAGCATGAAGCCAAACC.

RNAi and gene transfection

Human CYP51A1-shRNA-1 (5′-CCGTTACAAACGAAGATCAAA-3′), CYP51A1-shRNA-2 (5′-CGCCTGGACTTTAATCCTGAT-3′), SREBF1-shRNA-1 (5’-CCAGAAACTCAAGCAGGAGAA-3’), SREBF1-shRNA-2 (5’-CCCTGTGCTGACGGAAGCCAA-3’), SREBF2-shRNA-1 (5’-CCTGAGTTTCTCTCTCCTGAA-3’), SREBF2-shRNA-2 (5’-CCTCAGATCATCAAGACAGAT-3’), and human CYP51A1-cDNA developed by our laboratory were used in this study. We used 293FT cells (Thermo Fisher Scientific, R70007) to produce high-titer lentiviral particles. In addition, 293FT cells were incubated with Opti-MEM I reduced serum medium (Gibco, 31985070) for 2 h and then co-cultured with a mixture (shRNA [1200 ng], pSPAX2 [1600 ng], and pMD2G [400 ng]) for 12 h. After changing to DMEM medium, the lentiviral particles were harvested 48 h later. Finally, cells were transduced with lentiviral particles containing indicated shRNA for 24–48 h, and then cells were selected with puromycin (2 μg/ml; YEASEN, 60210ES72) for 7 days.

The control siRNA, CYP51A1 siRNA, LSS siRNA, and TM7SF2 siRNA were purchased from GenePharma and transfected into indicated cells using Lipofectamine RNA iMAX (ThermoFisher Scientific, 13778500) according to the manufacturer’s instructions. The sequences are as follows: control siRNA sense (5’-UUCUCCGAACGUGUCACGUTT-3’) and antisense (5’-ACGUGACACGUUCGGAGAATT-3’); CYP51A1-siRNA-1 sense (5’-GCUCUUUCUGAGCUCAUAATT) and antisense (5’-UUAUGAGCUCAGAAAGAGCTT); CYP51A1-siRNA-2 sense (5’-GCACAGCUGUAUGCAGAUUTT) and antisense (5’-AAUCUGCAUACAGCUGUGCTT); LSS-siRNA-1 sense (5’-CCGGAACAUUCUUCAUAAGAATT) and antisense (5’-UUCUUGUGAAGAAUGUUCCGGTT); LSS-siRNA-2 sense (5’-GCACAAGCUGUAUGAACACAUTT) and antisense (5’-AUGUGUUCAUACAGCUUCUGCTT); TM7SF2-siRNA-1 sense (5’-AGUUGCUCUACGUGGGUGAUGTT) and antisense (5’-CAUCACCCACGUAGAGCAACUTT); and TM7SF2-siRNA-2 sense (5’-AUAUCACACAUGACGGGUUUGTT) and antisense (5’-CAAACCCGUCAUGUGUGAUAUTT).

Cell viability assay

Cell viability was assessed using the Cell Counting Kit-8 (CCK-8; YEASEN, 40203ES80) following the manufacturer’s instructions. In brief, cells in the logarithmic growth phase were seeded into 96-well plates at a density of 5–10 × 10⁴ cells per well and allowed to adhere fully. They were then treated with JTC801 for 24 h. After treatment, the culture medium was replaced with a 10% CCK-8 working solution, and the plates were incubated at 37 °C for 30–60 min. The absorbance at 450 nm was measured, and this absorbance value was directly proportional to the number of living cells in the culture. Cell viability was expressed as a relative level, with 100% cell viability corresponding to a value of 1.

Colony formation assay

Long-term cell survival was assessed using a colony formation assay. Cells in the logarithmic growth stage were initially seeded in 6-well plates at a density of 1500 cells per well. After cells had fully adhered, the experimental groups were treated with JTC801 for 24 h. Following treatment, the medium was replaced, and cell growth was allowed to continue for 10–14 days to facilitate colony formation. Colonies were subsequently visualized by staining with 4% crystal violet (Solarbio, C8470). Quantitative analysis of colony formation was conducted using Image J software, with untreated controls serving as the reference (set to 100% or 1). The percentage of cell clones was determined by calculating the ratio of the number of cell colonies in each well to that in the control wells.

Thioflavin T staining

For thioflavin T staining (Selleck Chemicals, S6873), cells were initially seeded in 6-well plates at a density of 3 × 10⁶ cells per well and allowed to grow until fully confluent. In addition to the control group, cells were treated with JTC801 for 24 h. Thioflavin T solution was prepared by diluting it with PBS (1:3000), and the cells were then incubated with this solution in the incubator for 30 min. Subsequently, the cells were observed under a fluorescence microscope. Fluorescence intensity was quantified using Image J software, with the control group used as the reference (set to a value of 1).

Propidium iodide/Hoechst 33342 staining

Cells in the logarithmic growth phase were seeded in 6-well plates at a density of 3 × 10⁶ cells per well and allowed to grow until fully confluent. After treatment with the specified drug reagents for the designated duration, cells were stained with propidium iodide (PI) and Hoechst 33342 for 30–40 min in a 5% CO₂ cell incubator. Morphological changes were examined using a fluorescence microscope at ×100 magnification. For quantitative analysis using Image J software, the percentage of cell death was determined by calculating the ratio of PI-positive cells to Hoechst 33342-stained cells in each well.

Luciferase assay

WT and knockdown cells were cultured at a density of 5 × 10⁴ cells per well in 24-well plates. Each well was transfected with 0.05 μg of Renilla luciferase plasmid (Synbio Technologies) and 0.5 μg of pGL3 basic luciferase vector containing the human CYP51A1 promoter, using Lipofectamine RNAiMAX (ThermoFisher Scientific, 13778500) as the transfection reagent. After 48 h of post-transfection incubation, luciferase activity was measured using a fluorometer and the Dual Luciferase Reporter Analysis System (Beyotime; RG029S). Firefly luciferase activity was normalized to Renilla luciferase activity.

Drug coupling

DMSO-solubilized JTC801 was fully dissolved with an appropriate dose of Cyanine5 acid, 1-ethyl-(3-dimethylaminopropyl) carbodiimide (EDC), p-hydroxybenzonitrile (HOBt), and N-methylmorpholine. The reaction was conducted overnight at 40 °C under nitrogen protection. Alternatively, CY5-CHO dissolved in DMSO was reacted with JTC801 and sodium triacetoxyborohydride overnight at room temperature. Another method involved dissolving Cyanine5 acid in thionyl chloride, reacting at 60 °C under nitrogen protection, removing thionyl chloride by spinning under reduced pressure, and then adding DMSO-dissolved JTC801 and triethylamine in an ice bath. The target products were detected using mass spectrometry.

Proteasome 20S activity

The assay was conducted using a fluorometric proteasome 20S activity assay kit (ApexBio; K2242). Briefly, cells in the logarithmic growth phase were seeded in 96-well plates and treated with the appropriate drug for the required duration. The proteasome substrate LLVY-R110 and the reaction solution were prepared as a working solution in a 1:400 ratio. This working solution was added to the treated cells in a 1:1 ratio, and the cells were incubated in a cell culture incubator for over 1 h, depending on the experimental conditions. Fluorescence was measured using a multifunctional microplate reader.

Detection of dihydrolanostanol and lanostanol

Treated cells were collected by scraping (avoiding trypsin), washed twice with pre-cooled PBS, and snap-frozen in liquid nitrogen. Targeted metabolites were detected using gas chromatography (Agilent 7890 GC) and mass spectrometry (Agilent 5975) in positive ion mode. Deuterostyrene internal standard solution (CAS: 19361-62-7) was prepared in methanol. Samples were mixed with the internal standard solution, steel beads and centrifuged. The supernatant was collected after adding anhydrous sodium sulfate.

Chromatographic conditions were as follows: column 19091S-433 HP-5MS (30 m × 0.25 mm × 0.25 μm), high-purity helium (99.999%) as the carrier gas at 1.0 mL/min flow rate, with a split injection ratio of 50:1. The temperature program was 30 °C/min from 50 °C to 320 °C, maintained at 320 °C for 5 min. The GC/MS interface temperature was 270 °C. Mass spectrometry conditions included an EI ion source with 70 V electron energy, ion polarity positive, ion source temperature at 240 °C, quadrupole temperature at 160 °C, and SIM mode scanning mass ions at 112.1, 413, 411 with a scanning speed of 1562 u/s and frequency of 2.7 scans/s. Peaks were aligned and quantified using Qualitative Workflows B.08.00 software, with internal standard correction and quantification based on standard concentrations.

Measurement of pH

Intracellular pH was measured in living cells using the pHrodo Green AM fluorescent probe (Thermo Fisher Scientific, P35373). The fluorescence of the pHrodo Green probe is directly correlated with acidic pH levels and vice versa. In brief, treated cells were first washed once with a live cell imaging solution (LCIS). The cells were then incubated for 30 min at 37 °C in a 5% CO₂ cell culture incubator with a pH dye working solution (prepared by mixing LCIS, PowerLoad concentrate, and pHrodo Green AM in a ratio of 100:10:1). After incubation, the cells were washed twice with LCIS to remove excess dye. Cell fluorescence intensity was measured using a fluorescence microscope. Subsequently, the cells were incubated with standard buffers of different pH values in a cell incubator for 15 min, and the fluorescence intensity was detected either by fluorescence microscopy or a multifunctional microplate reader. Cell fluorescence intensity at different pH buffers was analyzed using Image J software, and a standard curve was generated to calculate the pH value of the indicated cells.

Lysosomal pH was measured using LysoSensor Yellow/Blue DND-160 (Yeasen; 40768ES50). Cells were incubated with pre-warmed LysoSensor Yellow/Blue DND-160 solution (3 µM) at 37 °C for 15 min in a humidified cell incubator. Following incubation, the supernatant was aspirated, and cells were washed and replaced with fresh phenol red-free medium. Images were captured using a confocal microscope with excitation and emission wavelengths set to 352 nm/461 nm for optimal signal detection. To calibrate the fluorescence signal to actual pH values, cells were equilibrated with pH calibration buffers containing 130 mM KCl, 1 mM MgCl₂, 15 mM HEPES, 15 mM MES, 10 μM nigericin, and 10 μM valinomycin. Calibration was performed using buffers adjusted to pH values of 3, 4, 5, and 6 for 10 min. After treatment, fluorescence intensity data were plotted against pH to construct a standard calibration curve. The lysosomal pH in JTC801-treated cells was determined using the Boltzmann sigmoid equation⁶⁶. Representative images of cells treated with standard pH buffers were also included for comparison.

Intracellular pH was measured using BCECF-AM (MedChemExpress, HY-101883) according to the manufacturer’s instructions. Cells were incubated with BCECF-AM at a final concentration of 3 µM for 30 min at 37 °C in the dark. During this incubation, intracellular esterases cleave the AM ester groups of BCECF-AM, trapping the dye inside the cells and allowing it to respond to intracellular pH changes. After incubation, the cells were washed with a live cell imaging solution from the kit. Fluorescence intensities were alternately excited at 440 nm (pH-insensitive) and 490 nm (pH-sensitive), and emission was collected at 528 nm using a confocal microscopy. A standard curve was generated by equilibrating cells in pH calibration buffers containing 130 mM KCl, 1 mM MgCl₂, 15 mM HEPES, 15 mM MES, 10 μM nigericin, and 10 μM valinomycin at pH values of 6.5, 7.0, 7.5, 8.0, and 8.5 for 10 min. The fluorescence ratios at each pH were then measured to construct the calibration curve. The fluorescence ratios of JTC801-treated samples were compared against this calibration curve to determine precise intracellular pH. The calibration equation used for calculating intracellular pH from fluorescence intensity ratios is based on the Henderson-Hasselbalch equation⁶⁷. Representative images of cells treated with standard pH buffers were also included.

Filipin staining

The filipin III-dependent Cholesterol Assay Kit (Abcam, ab133116) is a widely used tool for detecting sterols in biological membranes. Interaction with cholesterol causes changes in the absorption and fluorescence spectra of filipin III, enabling fluorescence microscopy with excitation at 340–380 nm and emission at 385–470 nm. Briefly, cells were seeded in either 96-well plates or confocal dishes and allowed to grow overnight. The cells were then treated with the specified drugs for a duration of 24 h. Following treatment, the culture medium was removed, and cells were washed with a cholesterol detection wash buffer. Filipin III stock solution (diluted at a 1:100 ratio) was added to the cells, which were then incubated in a cell incubator for 60 min. After incubation, cells were washed again, and fluorescence microscopy was performed. As a positive control for lysosomal cholesterol transport blockade, cells treated with 1.25 μM U18666A (Abcam, ab133116) for 48 h were used. Quantitative analysis of fluorescence was conducted using Image J software, with the fluorescence of the untreated control group set to 1.

For co-localization analysis of filipin III with lysosomes, cells were pretreated as described above. The culture medium was discarded, and cells were incubated for 40 min at 37 °C with pre-warmed LysoTracker Green DND-26 (Yeasen, 40738ES50) at a 1:2000 dilution. The subsequent procedure was the same as described earlier. Co-localization experiments were conducted using confocal microscopy. Image J software was used for quantitative analysis.

Lysosomal isolation

Lysosomes were isolated using the Lysosome Enrichment Kit for Cultured Cells (Thermo Fisher Scientific, 89839), following the manufacturer’s guidelines. Briefly, cells were homogenized in the provided lysis buffer supplemented with protease inhibitors, and the homogenate was centrifuged at low speed to remove nuclei and cellular debris. The resulting supernatant was layered onto a density gradient prepared with the kit’s OptiPrep medium and subjected to ultracentrifugation at 150,000 × g for 2 h at 4 °C. Lysosome-enriched fractions were collected, washed to remove contaminants, and validated for purity using lysosomal markers, such as LAMP1, by western blot analysis.

Total cholesterol and ER cholesterol detection

To measure total cholesterol levels, we utilized a cholesterol assay kit (Sigma- Aldrich, MAK043). Cells were initially seeded in 6-well plates at a density of 3 × 10⁶ cells per well and treated with JTC801 for a 24 h period. Following treatment, the cells were detached using trypsin. We then added an appropriate mixture of chloroform, isopropanol, and IGEPAL (in a ratio of 7:11:0.1) to facilitate efficient lipid extraction. The samples were centrifuged at 13,000 × g for 10 min to remove insoluble components. The organic phase, containing lipids, was carefully transferred and subjected to air drying at 50 °C to eliminate chloroform and other organic solvents. The dried lipids were then fully dissolved using cholesterol assay buffer. We prepared the assay reagents according to the kit instructions and selected either the colorimetric or fluorometric method for cholesterol quantification.

To detect ER cholesterol, cells (1 × 10⁸) were used for ER isolation following the manufacturer’s instructions (Sigma, ER0100). Briefly, cells were resuspended in 1× hypotonic extraction buffer (3× cell volume, PCV) and incubated at 4 °C for 20 min to allow cell swelling. After centrifugation at 600 × g for 5 min, the supernatant was discarded, and the pellet was resuspended in 1× isotonic extraction buffer (2× PCV). Cells were homogenized using a Dounce homogenizer (10 strokes) at 4 °C and centrifuged at 1000 × g for 10 min. The supernatant was then centrifuged at 12,000 × g for 15 min at 4 °C. The resulting supernatant was subjected to ultracentrifugation at 100,000 × g for 60 min at 4 °C to obtain the microsomal fraction. The pellet was resuspended in 1× isotonic extraction buffer, layered between 30% and 15% Optiprep, and ultracentrifuged on a fixed-angle rotor. The fractions were collected from the top to the bottom of the gradient. ER purity was confirmed by western blot analysis using specific marker antibodies, and cholesterol levels were quantified in lysates using the Sigma-Aldrich cholesterol assay kit (MAK043), normalized to equal sample amounts.

LC-MS/MS for intracellular drug concentration detection

Intracellular drug concentrations were measured by LC-MS/MS as previously described⁶⁸. Briefly, cells were harvested, thawed on ice, and subjected to three freeze-thaw cycles with intermittent vortexing to ensure complete cell disruption. A 100 µL aliquot of the sample was combined with 10 µL of internal standard, vortexed for 10 s, and 500 µL of ethyl acetate was added. The mixture was vortexed for 30 s and centrifuged at 14,000 × g for 5 min. Subsequently, 400 µL of the supernatant was collected for LC-MS/MS analysis. Quantitative measurements were performed using a NexeraX2 LC-30AD chromatography system (Shimadzu) coupled with a 3200 QTRAP mass spectrometer (Sciex). Separation was achieved using an Atlantis T3 C18 column (50 mm × 2.1 mm, 3.5 µm particle size) maintained at 40 °C. The mobile phase consisted of solvent A (water with 0.1% formic acid) and solvent B (acetonitrile with 0.1% formic acid). The gradient elution program was as follows: 0-1 min, 5% B; 1–1.2 min, 5% B; 1.2–3.0 min, 95% B; 3.0–3.2 min, 95% B; 3.2–4.0 min, 5% B for column re-equilibration.

Immunofluorescence

To prepare cells for experimentation, they were initially cultured in a confocal petri dish until they reached the logarithmic growth phase. Subsequently, the cells were subjected to the required drug treatments. After drug treatment, the culture medium was carefully removed, and the cells were washed twice with PBS. Fixation of the cells was carried out using 4% paraformaldehyde for a duration of 15 min. Following fixation, the cells were rinsed with PBS three times. A 0.5% Triton X-100 solution was applied for 15 min to facilitate permeabilization. Blocking was performed using 5% BSA at room temperature for 1 h. The cells were then incubated overnight at 4 °C with the primary antibody of interest. After incubation with the primary antibody, the cells were washed with PBST (PBS containing Tween-20) three times. Subsequently, the cells were incubated with a fluorescent secondary antibody at room temperature for 1 h. Finally, fluorescence confocal microscopy was employed for observation, and quantitative analysis was conducted using Image J software. It is worth noting that for fluorescence localization studies involving lysosomes and associated antibodies, a lysosomal labeling reagent (Yeasen, 40738ES50) should be added before fixation to facilitate staining.

Transcriptome analysis

Total RNA was extracted using the RNeasy Mini Kit (Qiagen, Germany). Subsequently, paired-end libraries were prepared following the TruSeq RNA Sample Preparation Guide (Illumina). The library preparation process involved the following: 1) isolation of poly-A containing mRNA molecules using poly-T oligo-attached magnetic beads, 2) fragmentation of mRNA into smaller pieces through exposure to divalent cations at 94 °C for 8 min, 3) synthesis of NO-strand cDNA from the cleaved RNA fragments using reverse transcriptase and random primers, 4) generation of second-strand cDNA through the use of DNA polymerase I and RNase H, 5) end repair of the resulting cDNA fragments, addition of a single A base, and ligation of adapters, and 6) purification and enrichment of the cDNA library via PCR to create the final library. The quantification of purified libraries was performed using a Qubit 2.0 Fluorometer (Life Technologies). Validation was conducted using the Agilent 2100 Bioanalyzer (Agilent Technologies) to confirm the insert size and calculate the molar concentration. Subsequently, cluster generation was accomplished using cBot with the library diluted to 10 pM, followed by sequencing on the Illumina NovaSeq 6000.

Lipidomics analysis

In the logarithmic growth phase, cells were harvested, and the culture medium was removed. Cells were then washed twice with ice-cold PBS followed by two washes with ice-cold saline (0.9% NaCl). A small amount of PBS was added to the cells, which were subsequently scraped off using a cell spatula. To prevent trypsin digestion, the cells were placed in 1.5 mL centrifuge tubes, centrifuged at 1000 × g, and snap-frozen in liquid nitrogen. Three independent replicate groups were collected for each sample, each containing >1 × 10⁷ cells.

For lipid extraction, approximately 25 mg of cells were weighed into a 2 mL thick-walled centrifuge tube. Two small magnetic beads and two small steel beads were added to each tube. Subsequently, 800 µL of pre-chilled dichloromethane/methanol (3:1, V/V) precipitant and 10 µL of the prepared internal standard 2 were added to each sample. The samples were then processed using TissueLyser, followed by sonication in an ice bath. After overnight storage in a −20 °C refrigerator, the supernatant was extracted via ultracentrifugation and then air-dried. The dried samples were reconstituted in an appropriate volume of lipid complex solution (isopropanol:acetonitrile:water = 2:1:1), sonicated, and centrifuged before analysis by ultra-performance liquid chromatography-mass spectrometry.

For data analysis, LipidSearch v.4.1 software (Thermo Fisher BGI Co., Ltd.) was employed. This software facilitated mass spectrometry data analysis, resulting in the identification of lipid-containing molecules and the generation of a data matrix containing quantitative results and other relevant information. Key parameters used in LipidSearch included library search mass deviation (5 ppm), response threshold (5%), peak mass deviation (5 ppm), M-score (5), and c-score (2).

Bioinformatics analysis encompassed data preprocessing, data quality control, overall analysis, differential analysis between groups in comparison groups, and differential analysis across multiple comparison groups.

Animal models

We conducted all animal care and experiments in accordance with the Association for Assessment and Accreditation of Laboratory Animal Care guidelines (http://www.aaalac.org) and with approval from our institutional animal care and use committee (University of Texas Southwestern Medical Center [102605], Guangzhou Medical University [S2023-786], and Jinlin University [KT20240266]). All experimental and control animals were matched on sex and age. None were excluded from analysis at the time of harvest. The tumor volume in mice was constrained to approximately 2000 mm³, and the tumor burden did not exceed 10% of the animal’s body weight.

To develop murine subcutaneous tumors, we injected 5 × 10⁶ PANC1 cells into the flanks of female NOD-SCID recipient mice. Treatment began when the tumors reached a size ranging from 150 to 200 mm³, typically around day 10 post-inoculation. The treatment regimens consisted of administering one of the following options: 1) a control vehicle, 2) JTC801 alone (5 or 20 mg/kg orally, once daily, 5 days per week), 3) ketoconazole alone (30 mg/kg orally, once daily, 5 days per week), or 4) JTC801 (5 mg/kg orally, once daily, 5 days per week) in combination with ketoconazole (30 mg/kg orally, once daily, 5 days per week). These treatment regimens spanned a period of 3 weeks. We measured tumor volumes weekly using the formula length × width² × π/6. Additionally, we generated patient-derived organoids from PDAC patient liver metastases and subsequently expanded them as PDXs in NOD-SCID mice⁶⁹. The use of patient samples was obtained through written informed consent and approved by the Institutional Review Board of Jilin University (2024020708).

To establish orthotopic tumors, male or female (1:1) C57BL/6 J mice underwent surgical implantation of either 5 × 10⁵ KPC cells into the tail of the pancreas^3,8. Ten days after implantation, we randomly assigned the mice to different treatment groups and initiated a 3-week treatment regimen. During this period, the mice received one of the following treatments: 1) a control vehicle and 2) JTC801 alone (5 or 20 mg/kg orally, once daily, 5 days per week). We monitored the survival of the animals on a weekly basis.

Analysis of tissue samples

Formalin-fixed, paraffin-embedded tumor tissue sections (5 μm) were tested with antibodies against CD8α (98941) and α-SMA (56856) from Cell Signaling Technology, following standard procedures⁷⁰. Briefly, slides were deparaffinized in xylene, rehydrated through ethanol gradients, and subjected to antigen retrieval in sodium citrate buffer (pH 6.0) via microwave heating (2 min) followed by subboiling incubation (95°–98°) for 10 min. Endogenous peroxidases were quenched with 3% hydrogen peroxide for 10 min, and endogenous avidin and biotin were blocked using 1× Animal-Free Blocking Solution (15019, Cell Signaling Technology). Primary antibodies (1:200) were applied overnight at 4 °C. The next day, slides were incubated with SignalStain Boost IHC Detection Reagent (HRP, Rabbit, 8114, Cell Signaling Technology) for 30 min, followed by SignalStain DAB (8059, Cell Signaling Technology) for 2–5 min. Slides were counterstained with hematoxylin (14166, Cell Signaling Technology), dehydrated, and mounted with SignalStain Mounting Medium (14177, Cell Signaling Technology). For assessing tumor stroma, the Masson’s trichrome stain kit (HT15, Sigma-Aldrich) was used. In addition, Ki67 levels in tissues were assayed by immunofluorescence using an antibody (9129) from Cell Signaling Technology. Images were captured and quantified from 5 fields using an EVOS microscope (Invitrogen). The levels of these markers were expressed by the relative signal in each field (×400).

Statistical analysis

Data collection and statistical analysis were conducted using GraphPad Prism 8.02 software. An unpaired t-test was employed to assess differences between the means of two groups. For comparisons among multiple groups, a one-way or two-way analysis of variance (ANOVA) followed by Tukey’s comparisons test was utilized. Log-rank tests were used to compare differences in mortality rates between groups. Results are primarily presented as mean ± standard deviation, with n representing the number of independent replicates. Statistical significance was defined as a P value of less than 0.05.

Author contributions

D.T. and J.L. conceived and planned the experiments. F.C., H.T., C.L., R.K., and J.L. carried out the simulations and sample preparation and analyzed the data. F.C. and D.T. wrote the paper. R.K. assisted in data interpretation and edited the manuscript.

Peer review

Peer review information

Nature Communications thanks Ming Chen, Marja Jäättelä and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The authors declare that all data supporting the findings of this study are available within the article and its supplementary information. The RNA-seq data is deposited in the Sequence Read Archive under the accession number PRJNA1056128. The metabolic data is deposited in MetaboLights database under accession codes MTBLS9283 and MTBLS9288. The proteins that bind JTC801 are listed in our previous publication¹⁰.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-57583-2.