SREBP-1 upregulates SOAT1 to promote tumor growth by preventing lipotoxicity.

INTRODUCTION
Cholesterol is a vital lipid that plays an indispensable role in maintaining cellular structure, membrane dynamics, and signal transduction.1-3 Its rigid tetracyclic ring structure, hydroxyl group, and hydrocarbon tail enable cholesterol to intercalate into phospholipid bilayers, where it modulates membrane fluidity, integrity, and the function of membrane-associated proteins.2,4 In most mammalian cells, cholesterol constitutes approximately 30%–50% of total lipids in the plasma membrane, reflecting its essential role in preserving membrane stability.5,6 By contrast, intracellular organelles harbor much lower cholesterol concentrations; the endoplasmic reticulum (ER) contains only ~5%–10% cholesterol, and mitochondria are particularly cholesterol poor, containing less than ~1% cholesterol of their total membrane lipid content.7
Tumor cells, driven by uncontrolled proliferation and increased membrane biogenesis, exhibit a heightened demand for cholesterol to support cell growth, survival, and oncogenic signaling.1,8 Yet, how cancer cells maintain cholesterol homeostasis, especially given the strikingly asymmetric distribution of cholesterol among subcellular membranes, remains poorly understood. Deciphering the regulatory mechanisms that control cholesterol acquisition, trafficking, and storage is crucial for identifying metabolic vulnerabilities in malignancies.
We recently demonstrated that lung cancer and glioblastoma (GBM), the most aggressive form of brain cancer, sustain cholesterol sufficiency by upregulating sterol regulatory element-binding protein 1 (SREBP-1), a master transcriptional factor that orchestrates cholesterol uptake and de novo synthesis.9-13 Elevated SREBP-1 activity has since been reported in a wide range of other malignancies, including hepatocellular carcinoma, breast cancer, and gastric cancer.14-16 The SREBP family comprises three isoforms—SREBP-1a, SREBP-1c, and SREBP-2—initially identified in the 1990s.9-11 Among them, SREBP-1a robustly activates genes involved in cholesterol and fatty acid (FA) synthesis as well as cholesterol import; SREBP-1c is more specialized in regulating FA synthesis; and SREBP-2 primarily governs cholesterol uptake and synthesis.17 Our recent work revealed that SREBP-1a is the predominant isoform significantly upregulated in GBM, while SREBP-2 expression is not elevated and is slightly reduced compared to normal brain tissues.18
Paradoxically, recent studies from our group and others have shown that sterol O-acyltransferase 1 (SOAT1), an enzyme that converts free cholesterol into cholesteryl esters (CEs) for storage in lipid droplets (LDs), is highly upregulated in GBM, hepatocellular carcinoma, pancreatic cancer, prostate cancer, and several other malignancies.19-25 These tumors are characterized by an abundance of CE-rich LDs. This raises a fundamental question: why do rapidly growing tumors enhance SREBP-1-driven cholesterol acquisition while concurrently diverting cholesterol into inert storage pools? This strategy appears to be metabolically inefficient and counterproductive for tumor growth. Moreover, the regulatory mechanisms underlying SOAT1 upregulation and its functional significance in tumor cells remain unclear.
In this study, we address key gaps in understanding how cancer cells couple cholesterol acquisition and intracellular storage to preserve cholesterol homeostasis. We demonstrate that SREBP-1 activates SOAT1 transcription, establishing a coordinated regulatory axis that links cholesterol acquisition to its intracellular storage. This coupling allows tumor cells to meet the high demand for cholesterol while preventing lipotoxic stress from excess accumulation. Our findings identify SOAT1 as a critical downstream effector of SREBP-1 and a potential metabolic vulnerability in aggressive cancers with hyperactive lipid acquisition programs.

RESULTS

Human tumors concurrently upregulate SREBP-1 and SOAT1
We wondered whether there is an intrinsic connection between cholesterol acquisition and storage pathways. To test this, we first performed immunohistochemistry (IHC) analysis on GBM samples. The results showed that SREBP-1 and SOAT1 are markedly co-upregulated in GBM tumor tissues from the same patients, along with the presence of abundant LDs, reflected by immunofluorescence (IF) staining of TIP47, a specific LD membrane protein marker (Figure 1A).19 By contrast, all three markers were lower in low-grade glioma and undetectable in normal brain tissues (Figures 1A and S1A). To validate these findings, we further analyzed 91 glioma patient samples in a tissue microarray (TMA). Consistently, IHC staining demonstrated a strong correlation between high SREBP-1 and elevated SOAT1 expression across low- to high-grade gliomas (Figures 1B-1D and S1B). Kaplan-Meier survival analysis revealed that elevated SOAT1 expression was associated with poorer survival outcomes (Figure 1E). Consistently, analysis of both The Cancer Genome Atlas (TCGA) and Chinese Glioma Genome Atlas (CGGA) glioma databases revealed that patients with high SOAT1 gene expression had significantly shorter overall survival compared to those with low expression (Figures 1F and S1C-S1J), which is consistent with our previous report that high SREBP-1 levels are inversely correlated with overall survival in patients with GBM.26 We further analyzed the expression of SREBP-1 and SOAT1 using RNA sequencing data from TCGA compared to normal human tissues from the Genotype-Tissue Expression Project (GTEx) database, accessed through the Gene Expression Profiling Interactive Analysis (GEPIA2) resource.27 The analysis showed that the expression of SREBF1 (which encodes SREBP-1 protein) and SOAT1 is significantly upregulated in tumor tissues from patients with GBM (N = 163) compared to normal brain samples from non-diseased adult individuals (N = 207) (Figure S1D).
We next analyzed patient samples containing tumors and adjacent normal tissues, including GBM, non-small cell lung cancer (NSCLC) and triple-negative breast cancer (TNBC) as well as paired NSCLC samples containing paired tumor and normal lung tissues. IHC staining showed that nuclear SREBP-1 and SOAT1 expression were highly co-elevated in the same regions of tumor tissues (T), whereas both were low in adjacent tissues (Figures 1G and S1K).
We further investigated whether the co-upregulation of SREBP-1 and SOAT1 is a common feature across cancer types. Patient specimens from NSCLC, breast cancer, ovarian cancer (OV), and pancreatic ductal adenocarcinoma (PDAC) were analyzed by IHC. Across these cancer types, tumor tissues from the same patients exhibited strong SREBP-1 and SOAT1 staining (Figure 1H). Consistently, analysis of messenger RNA (mRNA) expression data from the TCGA and GTEx databases revealed that both SREBF1 and SOAT1 are co-upregulated across various cancer types compared to normal tissues (Figures 1I-1L). By contrast, expression of the SOAT2 isoform remains extremely low in both normal and tumor tissues across different cancer types, except for testicular germ cell tumors (TGCTs) (Figure S1L). Moreover, SREBF2 (which encodes SREBP-2 protein) is significantly downregulated in OV and GBM compared to normal tissues, while its expression shows no significant difference between normal and tumor tissues in most other cancer types (Figure S1L).
Together, these results demonstrate that SREBP-1 and SOAT1 are co-upregulated across diverse cancer types, suggesting that a mechanistic relationship between their co-elevation might exist.

Inhibition of SOAT1 preferentially reduces the viability of tumor cells with high SREBP-1 activity
We next examined whether SREBP-1 activation is linked to SOAT1 expression by analyzing multiple primary GBM and NSCLC cell lines. Consistent with our findings in patient samples (Figure 1), Western blot analyses demonstrated a strong correlation between SREBP-1 activation and SOAT1 expression, showing that both SREBP-1 and SOAT1 are markedly upregulated in GBM cell lines U251, U373, GBM30, and GBM528 as well as in the NSCLC cell lines H1299, Hcc15, A549, and Hcc827 (Figure 2A), whereas they were substantially lower in GBM26, GBM0866, H1703, and Hcc95 cells (Figure 2A). Similarly, confocal imaging showed that high-SREBP-1/SOAT1 GBM and NSCLC cell lines accumulated abundant LDs (green), whereas LD content was dramatically reduced in low-SREBP-1/SOAT1 lines (GBM26, GBM0866, H1703, and Hcc95) (Figure 2B).
We further investigated whether lipid storage within LDs is required for maintaining tumor cell viability. Pharmacological inhibition of SOAT1 using avasimibe (Ava), a phase II/III clinical trial-tested SOAT1 inhibitor for patients with elevated cholesterol,28 robustly suppressed LD formation in a dose-dependent manner (Figure 2C) and reduced the viability of high-SREBP-1/SOAT1 GBM (U251, GBM30, and GBM84) and NSCLC (H1299, and A549) cell lines by more than 60% relative to controls (Figure 2D). By contrast, Ava had minimal effects on low-SREBP-1/SOAT1 cell lines (GBM26, GBM0866, H1703, and Hcc95) (Figure 2D). Consistently, genetic inhibition of SOAT1 by lentivirus-mediated short hairpin RNA (shRNA) produced a similar selective inhibition in high SREBP-1/SOAT1 cell lines, with little impact on low-expressing lines (Figure 2E). We further examined the effects of pharmacological and genetic SOAT1 inhibition in normal cells, including human astrocytes, non-transformed lung epithelial cells (HBEC3-KT), human embryonic kidney cells (HEK293FT), and mouse adult neural stem cells (aNSCs). In contrast to the pronounced vulnerability observed in cancer cells with high baseline SREBP-1 activity, these normal cell models showed minimal reductions in viability following SOAT1 inhibition (Figure S2).
Together, these results strongly suggest that SOAT1-mediated lipid storage is critical for maintaining the viability of tumor cells with high SREBP-1 activity.

SREBP-1 binds the SOAT1 promoter to activate its expression in tumor cells
Notably, SREBP-1 is synthesized as an inactive precursor anchored in the ER membrane via two transmembrane domains.17 It forms a complex with SREBP cleavage-activating protein (SCAP) (Figure 3A). For activation, the SCAP/SREBP-1 complex must be transported by the coat protein complex II (COPII) vesicles to the Golgi apparatus, where SREBP-1 undergoes sequential cleavage by site 1 protease (S1P) and site 2 protease (S2P). This process releases the active N-terminal transcription factor domain of SREBP-1, which enters the nucleus to initiate the transcription of lipogenic genes and the low-density lipoprotein receptor (LDLR), which facilitates uptake of cholesterol from LDL particles (Figure 3A).17,29
In addition to its established role in promoting cholesterol uptake and synthesis, we wondered whether SREBP-1 potentially regulates SOAT1 expression (Figure 3A). To test this, we inhibited SREBP-1 activation using its small molecular inhibitor fatostatin, which blocks SCAP/SREBP-1 trafficking from the ER to the Golgi, as well as PF429242, a specific S1P inhibitor.30 Western blot analysis showed that inhibition of SREBP-1 activation by either compound markedly reduced SOAT1 expression across GBM (U251, T98, and GBM30) and NSCLC (H1299, and A549) cancer cell lines (Figures 3B and 3C). These results were accompanied by a pronounced decrease in LD formation (Figures 3D and 3E). Consistently, genetic silencing of SREBP-1 via lentivirus-mediated shRNA significantly reduced SOAT1 expression and LD formation in GBM30 and H1299 cancer cells (Figures 3F and 3G). IHC analysis of tumors from an orthotopic GBM model derived from primary GBM30 cells showed that SREBP-1 knockdown significantly reduced SOAT1 protein levels and led to a concomitant decrease in LDs, as visualized by TIP47 IF staining, consisting with findings from the SOAT1 knockdown xenograft model (Figure S3).
We next analyzed the SOAT1 promoter using the JASPAR online resource, which predicts potential transcription factor binding sites in gene promoters.31,32 We identified multiple sterol regulatory elements (SREs), known as SREBP-1 binding motifs, within the SOAT1 promoter (Figure 3H, top). Chromatin immunoprecipitation (ChIP) using a specific anti-SREBP-1 antibody, followed by real-time quantitative PCR (qPCR) analysis, confirmed direct binding of SREBP-1 to these promoter regions (Figure 3H, bottom). We further cloned the SOAT1 promoter into the pGL3-luciferase vector at multiple lengths and generated a series of pGL3-SOAT1-luciferase reporter constructs containing putative SREBP-1 binding sites (SREs) (Figure 3I). Luciferase assays revealed that both N-terminal SREBP-1a and SREBP-1c isoforms robustly activated SOAT1 promoter activity (Figure 3I). By contrast, mutating the SREBP-1 binding site within the SOAT1 promoter region (−404/−11 bp) completely abolished this transcriptional activation (Figure 3J). These results aligned with significantly increased SOAT1 mRNA expression in GBM30 cells and H1299 lung cancer cells upon adenoviral expression of the active N-terminal SREBP-1a and a modest increase with N-terminal SREBP-1c (Figure 3K). Consistently, western blotting showed that SREBP-1a robustly upregulated SOAT1 protein levels, along with the key lipogenic enzymes HMGCS1 (3-hydroxy-3-methylglutaryl-coenzyme A [CoA] synthase 1), HMGCR (3-hydroxy-3-methylglutaryl-CoA reductase), FASN (FA synthase), and LDLR (Figure 3L), and increased LD formation in both GBM and lung cancer cells (Figure 3M).
Together, these results demonstrate that SREBP-1 transcriptionally upregulates SOAT1 expression, leading to LD formation in cancer cells.

Genetic inhibition of SOAT1 under high SREBP-1 activity induces mitochondrial oxidative stress and apoptosis in tumor cells
SREBP-1 activation is negatively regulated by the increased cholesterol levels it induces; cholesterol binds to SCAP and blocks SCAP/SREBP-1 trafficking.33 We previously revealed that inhibiting SOAT1 to block cholesterol storage triggers this negative feedback loop, which suppresses SREBP-1 activation, leading to reduced FA synthesis and ultimately inhibiting GBM tumor growth.19 To decipher why tumor growth requires simultaneous upregulation of both cholesterol acquisition and storage pathways, we preserved SREBP-1 activation by expressing a FLAG-tagged, N-terminal active form of SREBP-1a in GBM cells, while silencing SOAT1 expression by lentivirus-mediated shRNA (Figure 4A). This approach bypasses the negative feedback inhibition on SREBP-1 triggered by SOAT1 suppression, allowing us to isolate and examine the consequence of sustaining the SREBP-1-driven cholesterol acquisition axis in tumor cells while disrupting the SOAT1-mediated cholesterol storage pathway (Figure 4A).
Indeed, genetic knockdown of SOAT1 via lentivirus-mediated shRNA in GBM cells reduced SREBP-1 activation and downregulated SREBP-1 targets, including the cholesterol synthesis enzymes HMGCS1 and HMGCR, LDLR, and the FA synthesis enzymes FASN and SCD1 (stearoyl-CoA desaturase 1) (Figures 4B and 4C), whereas these effects were reversed by adenoviral expression of FLAG-tagged active N-terminal SREBP-1a (Ad-nSREBP-1a) (Figures 4B and 4C). Fluorescence imaging demonstrated that expression of the active SREBP-1a markedly increased LD formation, which was abolished by genetic silencing of SOAT1 (Figures 4D and S4A). Cholesterol quantification demonstrated that free cholesterol levels were significantly elevated in GBM cells with expression of active SREBP-1a isoform under SOAT1 knockdown conditions (Figure 4E).
We next examined the effects of preserving high SREBP-1 activity in GBM cells under the condition of SOAT1 silencing. Unexpectedly, the expression of the N-terminal SREBP-1a in SOAT1 knockdown GBM cells resulted in a marked reduction in colony formation compared to either SREBP-1a expression or SOAT1 knockdown alone (Figures 4F and S4B). Western blot analysis revealed that sustained SREBP-1 activation under SOAT1 silencing conditions led to a significant increase in apoptosis markers, including cytosolic cytochrome c, cleaved caspase-3, cleaved caspase-9, and cleaved PARP (poly(ADP-ribose) polymerase) (Figures 4G and S4C). Consistently, cholesterol supplementation in SOAT1 knockdown GBM cells led to increased cell death compared to SOAT1 knockdown alone a significant increase in apoptosis markers, including cytosolic cytochrome c and cleaved caspase-3, cleaved caspase-9, and cleaved PARP (Figures 4H and 4I).
To investigate the underlying mechanisms by which sustained SREBP-1 activity leads to enhanced apoptosis in GBM cells under conditions of SOAT1 silencing, we examined reactive oxygen species (ROS) levels under these conditions. Using the ROS-sensitive fluorescent dye CellROX, which emits strong red fluorescence in the presence of high ROS levels, we observed a marked increase in ROS in GBM cells expressing the active N-terminal form of SREBP-1a following SOAT1 knockdown (Figure 4J). Moreover, we found that CellROX-stained ROS predominately localized to mitochondria, as demonstrated by co-staining with the mitochondrial fluorescent dye MitoTracker Green (Figures 4J and S4D). To confirm these findings, we used the mitochondrion-specific ROS probe MitoSOX, which consistently showed that the active form of SREBP-1a expression significantly elevated mitochondrial ROS levels in the context of SOAT1 knockdown (Figure 4K). Moreover, mitochondrial membrane potential, assessed by rhodamine 123 staining, was dramatically reduced under these conditions (Figure S4E).
To validate whether elevated mitochondrial ROS were responsible for inducing apoptosis, we supplemented GBM cells with two antioxidants, glutathione (GSH) and MitoTEMPO ((2-(2,2,6,6-Tetramethylpiperidin-1-oxyl-4-ylamino)-2-oxoethyl)triphenylphosphonium chloride; a specific scavenger of mitochondrial superoxide), to assess whether their additions could mitigate sustained SREBP-1a activation-induced apoptosis under SOAT1 knockdown conditions. Clearly, fluorescence imaging revealed that both GSH and MitoTEMPO markedly reduced mitochondrial ROS levels in GBM cells (Figures 4L-4M). Western blot analysis demonstrated that the apoptosis markers cytosolic cytochrome c, cleaved caspase-3, cleaved caspase-9, and cleaved PARP were all abolished by either GSH or MitoTEMPO antioxidants in SOAT1 knockdown GBM cells with the expression of the active N-terminal SREBP-1a isoform (Figure 4N).
Taken together, these findings demonstrate that sustaining SREBP-1 activation in the context of SOAT1 inhibition leads to elevated mitochondrial ROS and tumor cell apoptosis.

Sustained SREBP-1 activation enhances the antitumor efficacy of SOAT1 inhibition in GBM and lung cancer organoids and xenografts
We next examined GBM patient-derived organoids to assess the physiological relevance of co-upregulation of SREBP-1 and SOAT1 in tumor cells (Figure 5A). IHC and fluorescence imaging confirmed that both SREBP-1 and SOAT1 are co-upregulated in human GBM tumor tissues and their corresponding organoids, along with high levels of LDs and the proliferation marker Ki67 (Figure 5B). To evaluate the effects of sustaining SREBP-1 activation in organoids with SOAT1 knockdown, we infected GBM patient-derived organoids with a lentivirus expressing shRNA targeting SOAT1 overnight, followed by adenoviral expression of N-terminal SREBP-1a. After 3 days, organoids were co-stained with Calcein AM (for live cells), propidium iodide (PI; for dead cells), and Hoechst 33342 (for nucleus staining). Fluorescence imaging demonstrated that active N-terminal SREBP-1a expression under SOAT1 knockdown conditions dramatically induced tumor cell death (positive PI staining, red) (Figure 5C), significantly reduced organoid size, and decreased tumor cell viability compared to either shSOAT1 or Ad-nSREBP-1a expression alone in GBM patient-derived organoids (Figure 5D). We further confirmed these findings in NSCLC patient-derived organoids, where similar effects were observed upon expressing N-terminal active SREBP-1a following SOAT1 silencing by shRNA (Figures 5E-5H).
We next assessed the in vivo effects by expressing N-terminal active SREBP-1a in GBM30-shSOAT1 cells (designed GBM30-Ad-nSREBP-1a/shSOAT1) and implanting these cells into mouse brains to generate orthotopic xenograft GBM models (Figures 6A and 6B). Bioluminescence imaging on day 14 and histological staining on day 20 post implantation showed that SOAT1 knockdown alone reduced tumor growth compared to the control group (GBM30-shSOAT1 vs. GBM30-shControl) (Figures 6C and 6D). However, expression of N-terminal active SREBP-1a further impaired tumor growth when SOAT1 was silenced compared to SOAT1 knockdown alone (GBM30-Ad-nSREBP-1a/shSOAT1 vs. Ad-null/shSOAT1 group) (Figures 6C and 6D). Kaplan-Meier analysis demonstrated that active SREBP-1a expression significantly extended overall survival in mice implanted with GBM30-Ad-nSREBP-1a/shSOAT1 cells compared to the GBM30-Ad-null/shSOAT1 group (Figure 6E). Through IHC staining, we observed increased SOAT1 expression and LD formation in intracranial GBM tumor tissues with expression of the N-terminal SREBP-1a isoform (Figure 6F). These increases were abolished when SOAT1 was genetically knocked down (Figure 6F). This combination led to a significant reduction in Ki67-positive staining and a marked increase in the apoptotic marker cleaved caspase-3 in tumor tissues (Figure 6F).
We further validated the effects of sustaining SREBP-1 activation in lung cancer models with SOAT1 knockdown. We implanted H1299-shSOAT1 with the expression of N-terminal active SREBP-1a (designed H1299-Ad-nSREBP-1a/shSOAT1) via tail vein injection to target the lungs. Like the GBM models (Figures 6A-6D), bioluminescence imaging on day 40 post implantation revealed that the expression of N-terminal active SREBP-1a significantly reduced tumor growth in H1299-shSOAT cells (H1299-Ad-nSREBP-1a/shSOAT1) compared to the H1299-Ad-null/shSOAT1 group (Figure 6I). Macroscopic images of mouse lungs and histological analysis of lung sections on day 50 post implantation confirmed that N-terminal active SREBP-1a expression significantly suppressed tumor growth in lung cancer models with SOAT1 knockdown (Figures 6J and S5). Moreover, Kaplan-Meier analysis demonstrated significantly extended overall survival in mice implanted with H1299-Ad-nSREBP-1a/shSOAT1 cells compared to the H1299-Ad-null/shSOAT1 group (Figure 6K). IHC analysis showed that the expression of N-terminal SREBP-1a together with SOAT1 knockdown significantly reduced Ki67 staining and increased the apoptosis marker cleaved caspase-3 in the tumor tissues (Figure 6L).

DISCUSSION
Our study revealed that SREBP-1 and SOAT1 are concurrently upregulated across multiple cancer types, including GBM, lung, breast, ovarian, and pancreatic cancers. We demonstrated that SREBP-1 directly binds to the SOAT1 promoter, activating its transcription and promoting cholesterol storage in CEs/LDs. Genetic inhibition of SOAT1 under conditions of sustained SREBP-1 activity results in the accumulation of free cholesterol and a marked increase in mitochondrial ROS, leading to impaired tumor growth. These findings demonstrate that SREBP-1 coordinates cholesterol acquisition and storage to maintain cholesterol homeostasis and prevent lipotoxicity, thereby promoting tumor cell survival and proliferation (Figure 7). Selective targeting of SOAT1 in tumors with high SREBP-1 activity offers a promising therapeutic strategy for malignancies characterized by active lipid metabolism (Figure 7).
The distinct distribution of cholesterol across subcellular membranes is a defining feature of mammalian cells.34 For example, cholesterol comprises approximately 30%–50% of total membrane lipids in the plasma membrane but less than 1% in mitochondria.35 The plasma membrane relies on high cholesterol levels to ensure membrane stability and barrier function, while the ER, a hub for lipid and protein synthesis, demands higher membrane flexibility, containing only 5%–10% cholesterol.36 By contrast, mitochondria, as energy-producing organelles enriched in electron transport chain (ETC) complexes I–V, have uniquely low cholesterol levels to preserve the complicated but elegant spatial organization of ETC complexes that contain hundreds of proteins.37,38 Recent in situ cryoelectron microscopy (cryo-EM) studies have revealed that these ETC complexes form highly organized supercomplexes.37 We predict that mitochondrial supercomplexes require minimal cholesterol incorporation into their membranes to maintain proper mitochondrial architecture and support optimal function.
To sustain rapid proliferation, tumor cells must acquire large amounts of cholesterol for membrane biogenesis and metabolic functions. They achieve this by upregulating SREBP-1 to enhance both cholesterol update and de novo synthesis.39-41 However, maintaining cholesterol homeostasis across different subcellular organelles with diverse cholesterol requirements presents a major challenge. How tumor cells prevent excess cholesterol accumulation while preserving the distinct cholesterol levels among organelles remains poorly understood and is an important unanswered question in cancer biology.
Our present study provides key insights into this intriguing question by uncovering a mechanism through which tumor cells prevent SREBP-1-driven cholesterol acquisition from causing cholesterol accumulation and inducing lipotoxicity. We show that SREBP-1 simultaneously upregulates SOAT1, which converts excess free cholesterol into CEs for storage in LDs. This coordinated regulation ensures that cholesterol levels remain tightly controlled and matched to the demands of each organelle. Our experimental approach, blocking SOAT1 while maintaining high SREBP-1 activity, demonstrated that disruption of cholesterol storage leads to free cholesterol accumulation and elevated mitochondrial ROS, indicating mitochondrial vulnerability to cholesterol dysregulation. Given the inherently low cholesterol content of mitochondria, these organelles appear to be particularly sensitive to perturbations in cholesterol homeostasis. These findings support a model in which SREBP-1 couples cholesterol acquisition and storage to buffer intracellular cholesterol levels and suggest that selectively targeting cholesterol storage via SOAT1 inhibition may exploit a metabolic vulnerability in tumors with hyperactive SREBP-1 signaling.
In addition, beyond cholesterol acquisition and storage, our previous studies have revealed that SREBP-1 can activate autophagy to hydrolyze LDs and release stored cholesterol, supporting tumor cell survival under cholesterol reduction conditions.40,42 Hence, LDs serve not only as a storage depot but also as a dynamic reservoir that can be mobilized when cellular cholesterol is low. Collectively, our studies reveal an intrinsic, SREBP-1-driven regulatory network that integrates cholesterol acquisition, storage, and release to maintain cholesterol homeostasis. This axis enables tumor cells to adapt quickly to fluctuations in cholesterol availability in the tumor microenvironment, storing excess cholesterol via SOAT1, suppressing SREBP-1 activation through feedback inhibition when cholesterol is abundant, while reactivating SREBP-1 and autophagy to restore cholesterol levels when cholesterol is scarce. Disruption of this tightly controlled system, such as through inhibition of SOAT1, destabilizes cholesterol balance and compromises tumor cell viability, providing a compelling therapeutic approach for tumors with high SREBP-1 activity.
In future studies, it will be critical to explore how cholesterol is trafficked and communicated between plasma membrane and subcellular organelles, and to delineate the precise mechanisms that preserve the distinct cholesterol composition of various membrane compartments. A deeper understanding of these regulatory processes and the molecular machinery involved in maintaining compartmentalized cholesterol distribution may open a promising avenue for targeting cancer.

Limitations of the study
Our work identifies SOAT1 as a critical SREBP-1-regulated factor that protects cancer cells from cholesterol-induced lipotoxicity; however, several limitations should be acknowledged. First, our in vivo studies were conducted in immunodeficient mice, which may not fully recapitulate the complexity of the tumor microenvironment, including potential immune responses to SOAT1 inhibition. Second, although we show that SOAT1 inhibition leads to free cholesterol accumulation and mitochondrial stress, the broader consequences for other lipid species and interconnected metabolic pathways remain incompletely defined. Third, our studies focused on GBM and NSCLC, leaving open whether the SREBP-1–SOAT1 axis is equally critical in other tumor types with distinct lipid metabolic programs. Addressing these limitations will be an important direction for future work.

STAR★METHODS

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Patient sample collection and processing
The collection and analysis of human tissue were approved by The Ohio State University Institutional Review Board (IRB) under protocol numbers 2015C0067 (for GBM), 2017C0033 (for NSCLC, and ovarian cancer), 2016C0025 (for breast cancer), and 2016C0071 (for pancreatic cancer). A waiver of informed consent was granted for the use of clinical specimens, as all samples were de-identified prior to receipt by the investigators. Individual tumors and adjacent healthy tissues were from the Department of Pathology at The Ohio State University. All human tissues were collected from Ohio State University Hospitals under IRB and HIPPA-approved protocols and histologically confirmed. Glioma TMA with 91 tumors were from the University of Kentucky and IRB approval was obtained before study initiation. All samples were tested as negative for HIV and hepatitis B. TMA slides were stained using SREBP-1 or SOAT1 antibodies and then using biotinylated horse anti-mouse IgG antibodies. The slides were scanned using ScanScope and analyzed using ImageScope v.11 software (Aperio Technologies). The staining intensity of tissues was graded as 0, 1+, 2+ or 3+. The H score was calculated using the following formula H score = (1 × (% cells with 1+) + 2 × (% cells with 2+) + 3 × (% cells with 3+)) × 100. Those samples were diagnosed based on pathological assessment and detailed information could be found in Data S1.

Mouse models
Female athymic nude mice (NCr-nu/nu, Charles River Lab strain #553), female NSG mice (Preclinical Therapeutics Mouse Modeling Shared Resource (PTMMSR), OSU; Stock No. 005557, Strain Name: NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ, inbred). Mice aged 6–8 weeks were housed under a 12-h light/12-h dark cycle at 22°C with ad libitum access to water and food. All animals were maintained in barrier vivarium facilities under the highest level of sterility and were routinely monitored for health in accordance with OSU IACUC guidelines. Unless otherwise specified, mice didn’t receive any drug or test article. All animal experiments were performed under protocols approved by the Institutional Animal Care and Use Committee of Ohio State University College of Medicine (Reference No. 2011A00000064-R4).
Intracranial GBM Models: 5×104 patient-derived neurosphere GBM30 cells expressing luciferase (GBM30-luc) in 5 μL PBS were intracranially injected into 6-8-week-old female athymic nude mice, which are immune-compromised and lack T-cells, using a stereotactic system. Using OSU Small Animal Imaging Core, tumor growth was monitored on day 20 after injection. Mice were observed until they became moribund, at which point they were sacrificed. Survival until the onset of neurologic symptoms was applied for survival curves.
Lung cancer models: H1299-Luc cells were collected by trypsin digestion and washed with cold PBS before cell counting. Cells (1 × 106 cells per mouse in 0.1 mL of PBS) were injected into NSG female mice via tail vein. After 4 weeks, the mice were divided equally into different groups based on the luminescence signaling. Specifically, mice implanted with H1299-Luc cells were injected intraperitoneally with a luciferin solution (15 mg/mL in PBS, dose of 150 mg/kg). The bioluminescence images were acquired using the IVIS Lumina system and analyzed with Living Image software. Imaging experiments were conducted at the OSU Small Animal Imaging Core.

Cell lines and culture
Authenticated (short tandem repeat profiling) human GBM cell lines, U251 (Male), U373, T98 (CRL-1690) (Male) were cultured in Dulbecco’s modified Eagle’s medium (DMEM), supplemented with 5% HyClone fetal bovine serum (FBS), and 4 mM glutamine. Human NSCLC cell lines, H1299 (Male, large cell carcinoma), A549 (Male, adenocarcinoma), Hcc827 (Female, adenocarcinoma), H1703 (Male, squamous cell), and Hcc95 (Male, squamous cell), and H1299-luciferase were cultured in RPMI-1640 medium supplemented with 10% FBS, and 4 mM glutamine. Human astrocyte cell lines, either immortalized or primary cell cultures derived from human brain astrocytes, were maintained in CocKit Human Astrocyte Complete Medium Kit (Cat# NCC20-9PZ02). The HBEC3-KT cell line, an hTERT-immortalized epithelial cell established from a 65-year-old female patient, was cultured in Bronchial Epithelial Cell Basal Medium (BEBM, Lonza Bioscience, Cat# CC-3171) with BEGM SingleQuots (Lonza Bioscience, Cat# CC-4175).
Authenticated GBM26, GBM30, GBM0866, GBM84, and GBM528 are primary GBM patient-derived cell lines that were previously molecularly characterized and described.43,44 They were cultured in Advanced DMEM/F-12 Flex Media supplemented with 1×B-27 serum-free supplements, 2 mg/mL heparin, 2 mM glutamine, 50 ng/mL EGF, and 50 ng/mL fibroblast growth factor (FGF).44 GBM30-luc cells stably express luciferase (luc) and were previously described.43 All cells were maintained at 37°C with 5% CO2. Cell lines in this study have different morphologies and growth rates, and any signs of contamination were constantly monitored. All cell lines used in this study were free from mycoplasma contamination based on PCR detection and were regularly maintained with mycoplasma reagent.

METHOD DETAILS

Mouse adult neural stem cell (aNSC) isolation
aNSCs were isolated as previously described.45,46 Briefly, neural stem/progenitor cells were obtained from the hippocampus of 6-week-old female C57BL/6J WT mice (The Jackson Laboratory, Strain #000664), cultured in Advanced DMEM/F-12 Flex Media supplemented with B-27 serum-free supplements, 2 mg/mL heparin, 2 mM glutamine, 50 ng/mL EGF, and 50 ng/mL fibroblast growth factor (FGF), and subsequently subjected to mechanical and enzymatic dissociation, followed by enrichment based on Nestin and SOX2 expression.

Quantitative real-time PCR
Total RNA from tissue sample or cells was isolated using TRIzol according to the manufacturer’s protocol, and cDNA was synthesized using iScript cDNA Synthesis Kit. Quantitative real-time PCR was conducted with iQ SYBR Green Supermix on the Applied Biosystems (ABI) 7900HT Real-Time PCR System. Gene expression levels were normalized to the 36B4 housekeeping gene and analyzed using the comparative method (2–ΔΔCt). All primers used in this study were shown as below:
SOAT1 (ACAT1) forward: 5′-CGGGCTAACTGATGTCTACAAT-3′

reverse: 5′-GCATAAGCGTCCTGTTCATTTC-3′

SREBP-1a forward: 5′-TCAGCGAGGCGGCTTTGGAGCAG-3′

reverse: 5′-CATGTCTTCGATGTCGGTCAG-3′

SREBF1 forward: 5′-CGCTCCTCCATCAATGACA-3′

reverse: 5′-TGCGCAAGACAGCAGATTTA-3′

ACACA forward: 5′-GATGTGGATGATGGGCTACA-3′

reverse: 5′-TGAGGCCTTGATCATTACTGG-3′

FASN forward: 5′-GTTCACGGACATGGAGCAC-3′

reverse: 5′-GTGGCTCTTGATGATCAGGTC-3′

SCD1 forward: 5′-TGCGATATGCTGTGGTGCT-3′

reverse: 5′-GATGTGCCAGCGGTACTCA-3′

ACLY forward: 5′-GAAGGGAGTGACCATCATCG-3′

reverse: 5′-TTAAAGCACCCAGGCTTGAT-3′

HMGCR forward: 5′-ATAATCCTGGGGAAAATGCC-3′

reverse: 5′-TCTTCTTGGTGCAAGCTCCT-3′

HMGCS1 forward: 5′-AAAAAGATCCATGCCCAGTG-3′

reverse: 5′-TCAGCAACATCCGAGCTAGA-3′

LDLR forward: 5′-TCTTTACGTGTTCCAAGGGG-3′

reverse: 5′-TGCAGTTTCCATCAGAGCAC-3′

36B4 forward: 5′-AATGGCAGCATCTACAACCC-3′

reverse: 5′-TCGTTTGTACCCGTTGATGA-3.

The amplification conditions were as follows: 95°C for 10 min, then denatured at 95°C for 30 s, annealed at 60°C for 30 s, and extended at 72°C for 30 s with 40 cycles.

Preparation of cell membrane fractions
Membranes were isolated as previously described.33 Briefly, cells were washed once with PBS and harvested by scraping. Cells were resuspended in a buffer containing 10 mM HEPES-KOH (pH 7.6), 10 mM KCl, 1.5 mM MgCl2, and 1 mM sodium EDTA, 1 mM sodium EGTA, 250 mM sucrose and a mixture of protease inhibitors, 5 μg/mL pepstatin A, 10 μg/mL leupeptin, 0.5 mM Phenyl-methanesulfonyl fluoride (PMSF), 1 mM DTT (DL-Dithiothreitol), and 25 μg/mL ALLN (Calpain Inhibitor I) for 30 min on ice. Extracts were passed through a 22G × 1-1/2 inch needle 30 times and centrifuged at 900 × g at 4°C for 5 min to remove nuclei. The supernatants were centrifuged at 20,000 × g for 20 min at 4°C. For subsequent western blot analysis (for SOAT1 and CD71 protein), the pellet was dissolved in 0.1 mL of SDS lysis buffer (10 mM Tris-HCl pH 6.8, 100 mM NaCl, 1% (v/v) SDS, 1 mM sodium EDTA, and 1 mM sodium EGTA) and designated “membrane fraction”. The membrane fraction was incubated at 37°C for 30 min, and protein concentration was determined by reading SpectraMax Plus 384 (Molecular Devices). One μL of bromophenol blue solution (100×) was added before the samples were subjected to SDS-PAGE. All chemicals were purchased from Sigma-Aldrich.

Mitochondria and cytosol fractionation
The mitochondrial proteins were prepared using Qproteome Mitochondria Isolation Kit (QIAGEN) following the manufacturer’s instructions. Briefly, cells were harvested and washed with PBS and resuspended with Lysis buffer and incubated at 4°C for 10 min. The cells were centrifuged at 1,000 × g at 4°C for 10 min, and the supernatants were used as the cytosolic fractions. Pellets were resuspended in disruption buffer and disrupted by using a 21G needle and a syringe. Following a centrifugation at 1,000 × g at 4°C for 10 min, the supernatants were transferred to new tubes and centrifuged at 6,000 × g at 4°C for 10 min. The pellets containing mitochondria were resuspended in mitochondria storage buffer and centrifuged at 6,000 × g at 4°C for 20 min. Pellets were then resuspended in mitochondria storage buffer and protein concentration was determined.

Western blotting
Western blotting was performed as previously described.47 Cells were lysed using RIPA buffer containing a protease inhibitor cocktail and phosphatase inhibitor. The proteins were separated using 12% SDS-PAGE and transferred onto a Hybond ECL nitrocellulose membrane. After blocking for 1.5 h in 5% nonfat milk diluted in Tris-buffered saline containing 0.1% Tween 20, the membranes were incubated with various primary antibodies, followed by secondary antibodies conjugated to horseradish peroxidase. The immunoreactivity was revealed using an ECL kit.

Chromatin immunoprecipitations (ChIP)
ChIP was performed using SimpleChIP Plus Enzymatic Chromatin IP Kit (Cell Signaling) by following the manufacturer’s instructions. Briefly, cells in 15-cm dish were fixed with formaldehyde at final concentration 1% to crosslink proteins to DNA and then incubated with glycine. Remove media and wash cells two times with ice-cold 1 × PBS, and scape cells with ice-cold 1 × PBS containing Protease Inhibitor Cocktail following centrifuge at 2,000 × g for 5 min at 4°C. The pellet was used for nuclei preparation and chromatin digestion. Finally, 2 μg of purified mouse anti-SREBP-1 antibody or normal mouse IgG were used for chromatin immunoprecipitation. PCR primers used for analysis of SREBP-1 binding motifs in SOAT1 promoter are available as follows:
SRE1 forward: 5′-CTTTGGGAGGCCTAGTGG-3′

reverse: 5′-CTGCCAAGCCTGGCTAATCT-3′

SRE2 forward: 5′-CAGACAGCATGGTGGGTATG-3′

reverse: 5′-TCAAATTCCAGGCTCAGTCC-3′

SRE3/SRE4 forward: 5′-AGCCCTGATGGTCGAGATTA-3′

reverse: 5′-TAGGCCTACCCAGCACAAAC-3′

Negative site (NC) forward: 5′-GCTGTCCTGATGAACATAGTGG-3′

reverse: 5′-GACACCTGCCCTAGGAGAAA-3′

SOAT1 promoter-luciferase plasmid construction and mutagenesis
The human promoter region (2000kb surrounding the transcription start site; coordinates was amplified from genomic DNA using high-fidelity polymerase and cloned upstream of the firefly luciferase gene in pGL3 using restriction sites. After PCR, the fragment of gene promoter of SOAT1 (−1627/−11 bp), SOAT1 (−941/−11 bp), SOAT (−441/−11 bp) and SOAT1 (−404/−11 bp) were cloned into pGL3-basic vector at Nhel/XhoI site. Site-directed mutation of SOAT1 (−404/−11 bp) was introduced by PCR-based mutagenesis to produce single-site mutant. All constructs were sequence-verified across the cloned promoter region; exact mutated bases and primer sequences are provided as below:
Primer used to clone SOAT1 gene promoters:

SOAT1 (−1627/−11 bp) forward: 5′ -CTAGCTAGCGTGCAGTGGCTCACG-3′

SOAT1 (−941/−11 bp) forward: 5′- CTA GCTAGCAGCATGGTGGGTATG-3′

SOAT1 (−441/−11 bp) forward: 5′- CTA GCTAGCTGATGGTCGAGATTA-3′

SOAT1 (−404/−11 bp) forward: 5′- CTA GCTAGCTGCACTCCAGCTTGG-3′

reverse: 5′-CCGCTCGAGTCCCCAACCCCGCAC-3′

The mutation of SOAT1 (−404/−11bp):

forward: 5′-TGCACTCCAGCCCCCCCCCCAGAGTGAGACCTTGTCTCAAAGAAAAAAAA-3′

reverse:5′-GGGGGGGGGGCTGGAGTGCAGCTAGCACGCGTAAGAGCTCGGTACCTATC-3′

The restriction endonucleases Nhel and Xhol recognition sequences for SOAT1.

Promoter luciferase report assay
Promoter construct DNA (100 ng) and renilla plasmid (20 ng) (Promega) were transfected into U251 and H1299 cells by using X-tremeGENE HP DNA Transfection Reagent (Sigma) in a 12-well plate with 5% FBS full DMEM medium for 24 h then infected with adenovirus-mediated null, N-terminal SREBP-1a, -1c or -2 for another 24 h. Cells were lysed by Glo Lysis buffer and luciferase activity was measured by using Promega Renilla-Glo Luciferase Assay System according to the kit instruction, and the signal was detected by Promega GloMax Plate Reader.

Immunohistochemistry (IHC)
IHC was performed as previously described.18 Briefly, the paraffin-embedded tissue slides were placed in an oven at 60°C for 30 min and then deparaffinized in xylene followed by graded alcohols (100%, 95%, 80% and 70%). After that, the slides were washed with distilled water and immersed in 3% hydrogen peroxide for 10 min, transferred into pre-heated IHC antigen retrieval solution for 30 min, and then blocked with 3% BSA in PBS at room temperature for 1 h. The slides were then incubated with the primary antibody overnight at 4°C, followed by incubation with the secondary antibody at room temperature for 30 min. After incubating with avidin-biotin complex and staining with substrate kit (Peroxidase), the slides were washed with tap water, counterstained with hematoxylin, and briefly dipped in graded alcohols (70%, 80%, 95% and 100%), then in xylene twice for 5 min each. Finally, the slides were mounted and imaged.

Immunofluorescence (IF)
IF was performed as previously described.18 After the pre-treatment, the paraffin-embedded tissue slides were blocked by 3% bovine serum albumin (BAS) for 30 min at room temperature, then stained with primary antibodies overnight at 4°C, followed by incubation at room temperature for 2 h with the appropriate Alexa Fluor 568-labeled goat anti-rabbit IgG (H + L) secondary antibody. Tissues were washed three times with PBS in a dark chamber. The coverslips were washed as described above, inverted, mounted on slides using ProLong Gold Antifade Mountant with DAPI and examined with a Zeiss LSM510 Meta confocal microscopy.

Lipid droplets (LDs) staining and quantification
LDs were stained by incubating cells with 0.5 μM BODIPY 493/503 for 30 min and visualized by Zeiss LSM510 Meta confocal microscopy (63×/1.4 NA oil) and 1-μm-wide z-stacks acquired. At least 30 cells in each group were analyzed, and LDs numbers were quantified with the ImageJ software (NIH) in a 3D stack, as previously described.19,40

Mitochondrial membrane potential (MMP)
MMP in cells were analyzed by Rhodamine 123 according to the instructions of the manufacturer. Rhodamine 123 is a cell-permeant, cationic, green-fluorescent dye that is readily sequestered by active mitochondria without cytotoxic effects. In brief, 6 ×104 cells were plated in a glass bottom 35 mm cell culture dish and incubated with the indicated drugs. After treatment, Rhodamine 123 was added to a final concentration of 0.05 μg/mL for MMP detection for 30 min at 37°C. After washing twice with PBS, cells were then incubated with 1 μg/mL Hoechst 33342 Solution for 30 min before confocal imaging. Confocal images were taken using a Carl Zeiss LSM510 Meta (63×/1.4 NA oil). More than 30 cells were analyzed, and fluorescence was quantified by the ImageJ software.

Detection of reactive oxygen species (ROS)
ROS in cells were analyzed by use of the fluorogenic CellROX Deep Red reagent and MitoSOX Mitochondrial Superoxide Indicators according to the instructions of the manufacturer. The cell-permeant CellROX Deep Red reagent is non-fluorescent in its reduced state, produces bright near-infrared fluorescence upon oxidation by ROS, has been used to detect oxidative stress in cells, and MitoSOX superoxide indicators are fluorogenic dyes specifically targeted to mitochondria in live cells. 6 × 104 cells were plated in a glass bottom 35 mm cell culture dish and incubated with the indicated drugs. Upon treatment, CellROX Deep Red reagent or MitoSOX Mitochondrial Superoxide Indicators was added to a final concentration of 0.5 μM for ROS detection or co-stained with 50 nM MitoTracker Green FM Dye for 30 min at 37°C. After washing twice with PBS, cells were then incubated with 1 μg/mL Hoechst 33342 Solution for 30 min before confocal imaging. Confocal images were taken using a Zeiss LSM510 Meta confocal microscope. More than 30 cells were analyzed, and fluorescence was quantified by the ImageJ software.

Production and infection of lentivirus-shRNA
Mission pLKO.1-puro lentivirus vector containing shRNA for shSREBP1 (#1: TRCN0000414192; #2: TRCN0000422088), shSOAT1 (#1: TRCN0000234511; #2: TRCN0000234512) and the non-mammalian shRNA control (Cat# SHC002, Addgene) were purchased from Sigma. The shRNA vector and packing plasmids psPAX2 and the envelope plasmid pMD2.G were transfected into HEK293FT cells using polyethyleneimine. Supernatants were harvested at 48 h and 72 h and concentrated using the Lenti-X Concentrator. The virus titers were quantified by real-time PCR by using qPCR Lentivirus Titration Kit. The lentiviral transduction was performed according to Sigma’s MISSION protocol with polybrene (8 μg/mL). GBM or NSCLC cells were infected with the same multiplicity of infection (MOI) of shControl (shCtrl), shSREBP-1 or shSOAT1 lentivirus.

Colony formation assay
The colony growth ability of U251 cells was assessed following treatment with shSOAT1, Ad-nSREBP-1a virus or cholesterol treatment, or their combination. Cells were initially seeded at a density of 2000 cells per well in a 6-well plate and subjected to various treatments. The virus was added to the medium and incubated for 14 days, with medium changes every 3 days. Three replicates were performed for each condition. After 14 days of treatment, colonies were fixed and stained with 1% bromophenol blue sodium salt in 80% methanol for 1 h at room temperature. Samples were thoroughly washed with double distilled water and imaged using a GE Amersham Imager 600. Colony numbers were quantified using ImageJ.

Cholesterol measurement
3 × 105 U251 cells were seeded in 6-cm dish for 24 h, then treated with/without shSOAT1 and Ad-nSREBP-1a virus. After treatment, cells were washed with PBS twice and collected by scraping and centrifugation at 1,000 rpm for 10 min. The cell pellets were resuspended in the amount of Isopropanol/Triton X-100 (1% Triton X-100 in pure isopropanol) for 1 h at room temperature. After centrifugation at 12,000 rpm for 10 min, the supernatants were transferred into 12 × 75 mm ASTM Type 1, Borosilicate Glass Disposable Culture Tubes and dried under nitrogen. Cholesterol measurements were performed following the instruction manual of the Amplex Red Cholesterol Assay Kit.

Organoid culture
GBM or NSCLC patient tumors were dissociated using a human tumor dissociation kit (Cat# 130-095-929, Miltenyi Biotech) following the manufacturer’s protocol. Dissociated tissue was filtered via a 70 μm filter and subjected to isolation of human tumor cells. These tumor cells were cultured in ultra-low plate with SILAC Advanced DMEM/F-12 Flex Media supplemented with 1×B-27 serum-free supplements, 2 mg/mL heparin, 2 mM glutamine, 50 ng/mL EGF, and 50 ng/mL fibroblast growth factor (FGF) according to a previously published protocol.18,48,49 The diameter of the organoids was measured with ImageJ. To detect the degree of cell death in organoids, organoid medium was supplemented with Hoechst 33342 (10 μg/mL), Calcein AM (5μM) and PI (10μg/mL) and incubated for 1 h under the growth condition. Organoids were then imaged on an Echo Revolve fluorescence microscope. The intensity of PI staining in organoids was measured with ImageJ. The cell viability was measured by CellTiter-Glo luminescent cell viability assay according to manufacturer instructions.

Mice bioluminescence imaging
Analysis of mouse tumor by bioluminescence imaging was performed as previously described.33 Mice implanted with GBM30 or H1299 cells expressing luciferase were injected with luciferin (PerkinElmer) solution (15 mg/mL in PBS and dose of 150 mg/kg) through the intraperitoneal route. At about 3 min after luciferin injection, mice were anesthetized with isoflurane (2.0%–3.0% isoflurane). Mice were then placed on the imaging platform of the IVIS Lumina II. The imaging chamber was continuously infused with 1%–1.5% of isoflurane. Animals were imaged 10 min after luciferin injection. The imaging experiments were conducted at the OSU Small Animal Imaging Core.

Hematoxylin and eosin (H&E) staining
H&E staining was performed as previously described.19 Briefly, paraffin tissue sections were deparaffinized in xylene and rehydrated in decreasing concentrations of ethanol. Subsequently, slides were washed with dH2O and stained sequentially with hematoxylin and eosin (H&E) solution. After staining, slides underwent further dH2O washing. Finally, the sections were dehydrated in increasing concentrations of ethanol, immersed in xylene, and mounted in Permount.

QUANTIFICATION AND STATISTICAL ANALYSIS
For cell proliferation, quantification of LDs, quantification of ROS, and mitochondria membrane potential, data were analyzed using Student’s t test or one-way ANOVA. Mice were randomly assigned to groups. Mouse overall survival was analyzed by Log Rank Test. Data analysis was performed using SAS 9.4 (SAS; Cary, NC) or GraphPad Prism 9.41 statistical software. Sample sizes were based on results from our previous studies. p < 0.05 was considered biostatistical significance. All samples were included in the analysis.

Supplementary Material
123Supplemental information can be found online at https://doi.org/10.1016/j.celrep.2025.116896.