NPC2 Regulates Glioblastoma Cholesterol Metabolism and Malignant Behavior via GPX4‐Linked Ferroptosis

ABSTRACT

Glioblastoma (GBM) is a highly aggressive intracranial malignancy originating from neuroepithelial glial progenitors. The prominent intra‐ and intertumoral heterogeneity of GBM is a major driver of therapeutic resistance and frequent tumor recurrence. Cholesterol is an essential structural component of mammalian cell membranes and plays crucial roles in membrane trafficking and transmembrane signal transduction. As a key intracellular cholesterol transporter, the Niemann‐Pick type C2 (NPC2) protein remains poorly characterized regarding its biological functions and clinical implications in GBM. In this study, immunohistochemical staining of clinical GBM specimens revealed that elevated NPC2 expression was significantly correlated with unfavorable patient prognosis. Consistent with clinical observations, genetic knockdown of NPC2 markedly suppressed the proliferation and invasion abilities of LN18 cells. Mechanistically, NPC2 silencing downregulated GPX4 expression at both transcriptional and translational levels, suggesting a potential regulatory relationship between NPC2 and the ferroptosis pathway. Collectively, these findings indicate that NPC2 may facilitate GBM malignant progression partially through modulating GPX4‐mediated ferroptosis, thereby representing a tentative prognostic biomarker and a potential therapeutic target for GBM that requires further validation.

Abbreviations

ANOVA

analysis of variance

CGGA

Chinese Glioma Genome Atlas

CREM‐tau

camp response element regulator tau

DAB

diaminobenzidine

DFS

disease‐free survival

EGR1

early growth reactive protein 1

Fer‐1

ferrostatin‐1

FFPE

formalin‐fixed paraffin‐embedded

GBM

GBM Glioblastoma

GPX4

glutathione peroxidase 4

GSEA

gene set enrichment analysis

GSH

glutathione

HMGCR

3‐Hydroxy‐3‐methylglutaryl‐CoA reductase

IHC

immunohistochemical

LDL

low‐density lipoprotein

LDLRs

low‐density lipoprotein receptors

LGG

Low‐grade GBM

LXR

liver X receptor

Lys05

Lys01 trihydrochloride

MDA

malondialdehyde

MVA

mevaleric acid

Nec‐1

Necrostatin‐1

NIH

National Institutes of Health

NPC

Niemann‐Pick Disease Type C

OS

overall survival

ROS

reactive oxygen species

RT

reverse transcription

SECIS

selenoccysteine insertion sequence

SIM

simvastatin

SP1

specificity protein 1

SREBP1

sterol regulatory binding element 1

STR

short tandem repeat

TC

serum total cholesterol

TCGA

The Cancer Genome Atlas

TEM

transmission electron microscopy

TFAP2C

transcription factor family activator protein 2

TMZ

temozolomide

VAD‐FMK

Z‐Val‐Ala‐Asp (OMe)‐FMK

1. Introduction

Glioblastoma (GBM) ranks among the most prevalent primary malignant central nervous system neoplasms in adults and derives from neuroepithelial cells (Ma et al. 2021). Distinguished by its extreme malignancy and poor curative prospect, GBM is associated with short overall survival and exceedingly high mortality (Wang et al. 2023). Characterized by rapid proliferation and aggressive infiltration into adjacent parenchyma, GBM progressively destroys peritumoral normal brain tissue; notably, GBM rarely spreads via lymphatic routes (Yang et al. 2023). Temozolomide (TMZ), a first‐line oral alkylating chemotherapeutic drug, is the standard clinical agent for GBM management, yet it only prolongs patients' median survival by approximately 2 months largely owing to frequent development of TMZ chemoresistance. Accordingly, uncovering the molecular pathogenesis of GBM to identify novel therapeutic candidates holds critical clinical significance (Tomar et al. 2021).

In GBM, analysis of the Cancer Genome Atlas (TCGA) database found that the expression of three key enzymes that regulate the de novo synthesis pathway of cholesterol and their mRNAs was significantly reduced compared to normal brain, suggesting that GBM cells use this tissue‐specific physiology to promote their growth, and they do not acquire cholesterol through the biosynthetic pathway but rely on the uptake of cholesterol produced and secreted by astrocytes via apolipoprotein (Fang et al. 2021). In addition, GBM cells increase cholesterol uptake by upregulating low‐density lipoprotein receptors (LDLRs) while inhibiting de novo synthesis of cholesterol and oxysterols (Wood et al. 2022). As a result, GBM cells bypass two key negative feedback mechanisms for cholesterol homeostasis: (1) feedback inhibition by 3‐Hydroxy‐3‐methylglutaryl‐CoA reductase (HMGCR), the rate‐limiting enzyme in cholesterol synthesis, and (2) oxysterol‐activated liver X receptor (LXR) inhibition of low‐density lipoprotein (LDL) uptake and induction of cholesterol efflux (Geng et al. 2023).

Statins are cholesterol‐lowering drugs with multiple pleiotropic effects, including impacts on cancer. In recent years, both basic and clinical scientists have focused their research on the potential role of statins in cancer treatment. Recent studies have shown that statins may exert beneficial effects on the response to chemotherapy in cancer patients. A recent study indicated that simvastatin(SIM) combined with TMZ can improve the survival rate of patients with GBM (Hajiahmadi et al. 2023). In addition, SIM inhibits the production of isopentenyl pyrophosphate in the mevalonate pathway, thereby suppressing the biosynthesis of selenoproteins such as glutathione peroxidase 4 (GPX4) and ultimately promoting ferroptosis in cells (Fan et al. 2026).

In‐depth investigation into Niemann‐Pick Disease Type C protein (NPC) has demonstrated that it functions as a regulator of cholesterol metabolism in the human body and serves as an indispensable cholesterol carrier in human cells (Davis et al. 2021). NPC protein is an important component involved in intracellular LDL cholesterol transport, and plasma LDL cholesterol enters the cell and reaches the early endosomes after being taken up by the LDLR‐mediated endocytosis pathway. As the pH of the endosome decreases, early endosomes gradually acquire various acidic hydrolases and eventually form late endosomes (Zhao et al. 2023). In late endosomes, LDLR is recycled back to the cell membrane and LDL is degraded. The cholesteryl esters in LDL particles are hydrolyzed to free cholesterol, which is then transported to the Golgi complex, endoplasmic reticulum and cell membrane for esterification and synthesis (Lim et al. 2019). NPC proteins are important proteins involved in intracellular cholesterol transport and lipid allocation (Hung et al. 2020). NPC1 and NPC2 are two common proteins in the NPC family. NPC1 protein is a product made of a hydrophobic integral membrane protein composed of 1278 amino acid residues is involved in the transport of cholesterol and other lipid molecules in somatic cells through a domain with a high affinity for cholesterol at its end. The NPC2 protein, also known as HE1, is a soluble glycoprotein with 151 amino acid residues and contains a 19‐amino acid signal peptide (Wheeler and Sillence 2020). NPC2 is an intracellular transport protein that affects the cholesterol homeostasis regulatory system. It first binds to cholesterol released from LDL, then interacts with specific domains of the NPC1 protein to transfer cholesterol to NPC1, which subsequently mediates the transport of cholesterol out of lysosomes (Qian et al. 2020). Further study of the function of proteins and their role in cholesterol transport is of great importance for further elucidation of the intracellular lipid transport mechanism. At the same time, it can also provide new methods and research ideas for targeted cholesterol metabolism in the treatment of tumors (Wei et al. 2021; Fukaura et al. 2021).

In recent years, a growing body of evidence has demonstrated the critical role of NPC2 in the development and progression of various human cancers. There is a strong association between NPC2 expression and tumors, with NPC2 characterized by overexpression in breast cancer, colon cancer, and lung cancer (Xu et al. 2019). Although some bioinformatics analyses have suggested that NPC2 may serve as a prognostic biomarker for GBM (Zheng et al. 2024), the clinical significance and prognostic value of NPC2 expression in GBM remain largely unknown. Here, we demonstrated that NPC2 expression was markedly elevated in GBM tissues relative to normal brain tissues, and its level positively correlated with tumor malignancy and adverse clinical prognosis in GBM patients. Functional experiments confirmed that NPC2 knockdown suppressed malignant phenotypes of GBM cells, including cell proliferation and invasion. Notably, unlike previous bioinformatics studies that only explored the oncogenic effects of NPC2, our study uncovered a novel regulatory cascade: downregulation of NPC2 reduced the expression of GPX4, a core regulator of ferroptosis. Mechanistically, we further verified that NPC2 modulates intracellular cholesterol metabolism, which in turn governs GPX4 expression and subsequently regulates ferroptosis in GBM.

Here, it was found that the expression of NPC2 in GBM is significantly higher than in normal brain tissue, and its expression is positively correlated with the malignancy of the tumor and the poor prognosis of GBM patients. Knocking down the expression of NPC2 will inhibit the proliferation and invasion of GBM cells and other malignant biological behaviors. It is worth noting that the experiment also found that knocking down NPC2 expression will lead to a decrease in the expression of the key factor in ferroptosis, GPX4, suggesting that the function of NPC2 in GBM may be related to the activation of the ferroptosis signaling.

2. Materials and Methods

2.1. Patient Samples

In this study, 11 untreated primary GBM and 7 Low‐Grade GBM (LGG) patients were recruited from the Hefei Cancer Hospital of the Chinese Academy of Sciences (Hefei, China; PJ‐KY2024‐011). Detailed information on the cohort is supplemented as follows: For cohort enrollment, patients met the following inclusion criteria: histopathologically confirmed primary GBM per the WHO 2021 classification, age between 18 and 75 years, no prior radiotherapy, chemotherapy, or targeted therapy for brain tumors, and availability of formalin‐fixed paraffin‐embedded (FFPE) tumor tissues for immunohistochemical (IHC) staining. Patients were excluded if they had a history of other malignant tumors within 5 years, severe systemic diseases (e.g., decompensated liver cirrhosis, end‐stage renal disease) that could affect treatment tolerance or follow‐up, incomplete pathological or clinical data, or refused to participate in the study or provide informed consent. The cohort included 7 males and 4 females, with a median age of 55 years (range: 41–70 years). All patients or their legal guardians provided written informed consent prior to study enrollment; the consent form clearly stated the study purpose, involved procedures, potential risks, and the right to withdraw at any time without affecting standard clinical care, and the study protocol (including the informed consent procedure) was approved by the Ethics Committee of Hefei Cancer Hospital of the Chinese Academy of Sciences (Approval No.: PJ‐KY2024‐011). Tissue samples were obtained from the pathology department, sectioned, and fixed tissue sections were subjected to IHC staining. Specifically, 4 μM‐thick FFPE tumor sections were used for NPC2 IHC staining, following standard protocols: deparaffinization with xylene, rehydration through gradient ethanol, antigen retrieval (citrate buffer, pH 6.0, 95°C for 20 min), blocking of endogenous peroxidase activity (3% H₂O₂ for 10 min), incubation with primary antibody (anti‐NPC2, Abcam, 1:500) at 4°C overnight, followed by secondary antibody incubation and DAB chromogenic reaction. Stained sections were scored using a combined “intensity + percentage” system: staining intensity was scored 0 (no staining), 1 (weak), 2 (moderate), or 3 (strong); positive cell percentage was scored 0 (0%), 1 (1%–25%), 2 (26%–50%), 3 (51%–75%), or 4 (76%–100%); the combined score was calculated as (staining intensity score × positive cell percentage score), with a total range of 0–12.

2.2. Prognostic Analysis

According to the clinicopathological data of the patients corresponding to the tumor specimens in the specimen database, the correlation between overall survival (OS) or disease‐free survival (DFS) and NPC2 expression was calculated according to the survival time and follow‐up status, and the test results were expressed by Kaplan‐Meier curve.

2.3. Immunohistochemistry

Tumor tissues from the transplanted mice were fixed in 4% neutral formalin solution, embedded in paraffin, and processed together with tissue cores for deparaffinization, antigen retrieval, antigen blocking, antibody incubation, DAB staining, nuclear staining, etc. The slides were then observed and analyzed under a brightfield microscope to determine the tissue localization of relevant proteins.

2.4. Cell Culture and Treatments

The GBM cell line LN18 and U87 used in this study was purchased from Cellcook in Guangzhou, China, and provided with an short tandem repeat (STR) certificate. All test results were negative during the entire experimental period (from cell thawing to assay completion), confirming the mycoplasma‐free status of the cell line. Experiments were conducted using LN18 or U87 cells at passage 6 to passage 12 cells beyond passage 15 were discarded to prevent phenotypic drift caused by excessive passaging, and only cells in the logarithmic growth phase (80%–90% confluence) were used for subsequent assays. Cells were cultured in DMEM medium (Gibco, USA) containing 10% fetal bovine serum (HyClone, USA).

2.5. Cell Viability

Cell viability assays were performed using the CCK‐8 kit (NCM Biotech, Suzhou, China, K0301). 6000 cells were seeded in each well of a 96‐well plate. Cells were treated with NPC2‐siRNA or TMZ and Simvastatin (SIM, APExBIO Technology LLC, 79902‐63‐9, Purity: 99.67%) for 48 h, then CCK‐8 reagent was added to each well and the absorbance at 450 nm was measured.

2.6. Cell Cloning Assay

0.5% agar containing 10% FBS/DMEM was used as the bottom layer and 0.35% agar containing transfected cell lines and 10% FBS/DMEM was used as the top layer. After solidification, an appropriate amount of 10% FBS/DMEM was added to cover the adherent surface and incubated for about 10 days at 37°C to detect clone formation.

2.7. RNA Interference Treatment

NPC2‐siRNA and control‐siRNA were purchased from GenePharma Corporation. Lipofectamine RNAiMAX was used to transfect NPC2‐siRNA and control‐siRNA into LN18 cells by following the manufacturer's instructions. Cultured with DMEM for 48 h, LN18 cells were collected for BODIPY 581/591 C11 assay (Beyotime, China, S0043S), RT‐qPCR and western blotting experients. The sequences were listed as following:

NPC2‐siRNA, (1) NPC2‐Homo‐272 5’‐GGUUGUAAGAGUGGAAUUATT‐3’ and

5'‐UAAUUCCACUCUUACAACCTT‐3', (2) NPC2‐Homo‐184

5'‐CAGCAAUAUUCAGUCUAAATT‐3' and

5'‐UUUAGACUGAAUAUUGCUGTT‐3', (3) NPC2‐Homo‐315

5'‐CCUAUAGCUACCUGAAUAATT‐3' and

5'‐UUAUUCAGGUAGCUAUAGGTT‐3',

control‐siRNA, 5’‐GCCACCAACUGGAUUGAAUTT‐3' and

5'‐AUUCAAUCCAGUUGGUGGCTT‐3'.

2.8. Plasmid Transfection for Overexpression

LN18 cells were seeded in culture plates and grown to 60%–70% confluence prior to transfection. The pLV3‐CMV‐NPC2‐3×FLAG‐CopGFP‐Puro plasmid was transfected into cells using Lipofectamine RNAiMAX. Briefly, the plasmid was mixed with transfection reagent in serum‐free medium, incubated at room temperature for 15–20 min to form complexes, and then added dropwise to the cell culture. After 4–6 h of incubation, the medium was replaced with fresh complete medium. At 48 h post‐transfection, the overexpression efficiency was verified by RT‐qPCR, and cells were subjected to subsequent experiments.

2.9. Detection of Lipid Peroxidation

Lipid peroxidation levels were assessed using the BODIPY 581/591 C11 fluorescent probe, a specific indicator for lipid oxidative damage. Treated LN 18cells were collected, adjusted to a concentration of 1 × 10⁶ cells/mL, and incubated with BODIPY 581/591 C11 at a final concentration of 5 μM for 30 min at 37°C in the dark. Cells were then washed twice with PBS to remove unbound probe, and fluorescence signals were detected using a flow cytometer or confocal microscope. The probe emits red fluorescence (λem ≈ 591 nm) in its non‐oxidized state; upon oxidation by lipid peroxides, its fluorescence peak shifts to the blue spectrum, emitting green fluorescence (λem ≈ 510 nm). The ratio of red to green fluorescence intensity was used to quantify lipid peroxidation, with a lower ratio indicating higher lipid peroxide accumulation.

2.10. RT‐qPCR

After RNA is extracted from LN18 cells, a reverse transcription (RT) reaction is first performed to transcribe the RNA into cDNA. Real‐time PCR is then combined with fluorescent probes or fluorescent dyes such as SYBR Green. The expression of mRNA was evaluated by comparative CT, and GAPDH was used as the internal control. The sequence of primers is as follows:

Transcription Factor Family Activator Protein 2 (TFAP2C)‐R:AGTAGGGTGGCGGCTGATATTC;

TFAP2C‐F:ACGAAGAGGACTGCGAGGATC,

Specificity Protein 1(SP1)‐R:GCCTCAGCCTCCCAAGTAGC; SP1‐F:GCAGGTGGATCACAAGGTCAAG,

NPC2‐R:AGAAGTGGCAGCAGGAAGTCTC;

NPC2‐F:CCCTTGGTCACTCTCATCACATAA,

GPX4‐R:TGCTGGGATTACAGGCGTGAG;

GPX4‐F:AGATGGAGAGTGAAGGCTGATGG.

2.11. Western Blotting

LN18 cells were lysed with RIPA lysis buffer (Solarbio, Beijing, China) containing 1% protease. Total protein in the lysate was quantified using the BCA protein detection kit (Beyotime, China). The primary antibodies used in this study are Srebp‐1c (Santa Cruz Bio Technology, Inc, sc‐365513, 1:1000), GPX4 (proteintech, China, 67763‐1‐Ig, 1:1000), NPC‐2(Abcam, USA, ab186829, 1:2000), β‐Tubulin (proteintech, China, 66240‐1‐Ig, 1:5000).

2.12. Co‐Culture With Death Inhibitors

After completing the 24 h SIM pretreatment, different types of death inhibitors were added to each well separately. The cells were then cultured for an additional 24 h, following which the Cell Counting Kit‐8 (CCK‐8) was used to detect cell viability. The death inhibitors used included Z‐Val‐Ala‐Asp (OMe)‐FMK (Z‐VAD‐FMK, MedChemEpress, HY‐100579, Purity: 99.71%), Necrostatin‐1 (Nec‐1, MedChemEpress, HY‐15760, Purity: 99.89%), Lys01 trihydrochloride (Lys05, MedChemEpress, HY‐12855A, Purity: 99.58%) and Ferrostatin‐1 (Fer‐1, MedChemEpress, HY‐100579, Purity: 99.71%).

2.13. Process for Screening NPC2‐Targeted Drugs

Differential analysis was performed on data from TCGA. Datasets from the differential analyses of TCGA and CGGA were used to screen for potential drugs for adjuvant therapy of GBM. Reverse screening was conducted based on the CMap website and L1000FWD website. For the differential analysis data, all downregulated genes were selected, while among the upregulated genes, the top 170 genes with the most significant differential expression were chosen. These two websites were used to perform reverse screening on the candidate drugs; the top 100 drugs with the highest similarity were selected from each website, and their intersection was taken. Finally, SIM was identified as a candidate drug.

2.14. Transmission Electron Microscopy (TEM)

Tumor tissues were fixed with a mixed solution containing 2% glutaraldehyde and 1% osmium tetroxide (volume ratio of glutaraldehyde solution to osmium tetroxide solution: 1:1) at 4°C for exactly 1 h. After fixation, the tissue fragments were embedded in epoxy resin, and ultra‐thin sections were prepared using an ultra‐microtome. Following staining, the sections were observed under a Hitachi transmission electron microscope to acquire images.

2.15. Mouse Xenograft Model

All animal experiments complied with NIH's Guide for the Care and Use of Laboratory Animals and animal ethics standards, and were approved by the Science and Technology Ethics Committee of Hefei Institute of Physical Science, Chinese Academy of Sciences (Approval No.: DWLL(P)‐2024‐21). Five‐week‐old female BALB/c immunodeficient nude mice were purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China), and housed in the Experimental Animal Center of Hefei Cancer Hospital, Chinese Academy of Sciences, under SPF conditions (22 ± 2°C, 50% ± 5% humidity, 12 h light/dark cycle). Sterilized food and autoclaved water were provided ad libitum for animal welfare.U87MG GBM cells (logarithmic growth phase) were resuspended in PBS to 1 × 10⁷ cells/mL; 100 μL of this suspension (1 × 10⁷ cells) was subcutaneously injected into the right upper flank of each mouse to establish xenograft models. When average tumor volume reached 100–200 mm³ (≈7–10 days post‐injection), mice were randomly divided into four groups (n = 5/group) via random number table, and intraperitoneally injected with: PBS, TMZ (20 mg/kg), SIM (2 mg/kg), or TMZ (20 mg/kg)+SIM (2 mg/kg) (Li et al. 2018; Shojaei et al. 2020). Injections were given four times/week for 3 weeks, with a standardized volume of 0.2 mL/mouse. Tumor volume (measured every 2 days via vernier caliper, formula: V = (L×W²)/2, L = longest diameter, W=shortest diameter) and body weight (toxicity indicator) were monitored every 2 days. Toxicity was assessed per NIH guidelines. Mice with severe toxicity or meeting humane endpoints (tumor volume > 2000 mm³, obvious ulceration, or immobility) were euthanized via cervical dislocation under isoflurane anesthesia. At the 3‐week treatment end or humane endpoint, surviving mice were euthanized similarly. Tumors were excised, weighed, and stored at −80°C.

2.16. Statistical Analysis

All experimental data were analyzed using GraphPadPrism version 9.0 and ImageJ version 1.8.0.112 and SPSS version 18.0 software. and if the P value was less than 0.05, the difference was considered significant. One‐way analysis of variance (ANOVA) was used for comparisons among multiple groups, and Duncan's test was applied for post‐hoc multiple comparisons between groups. If the p value was less than 0.05, the difference was considered significant.

3. Results

3.1. NPC2 Was Highly Expressed in GBM Patient Tissues and Cell Lines

To investigate whether there are differences in the expression levels of NPC2 in different tumors and adjacent tissues, and the relationship between NPC2 expression and prognosis in different cancers, we used the UCSCXenaShiny tool in Rstudio to perform a pan‐cancer analysis of NPC2 expression, and then divided 33 tumor types and patients into high, medium and low expression groups for prognostic analysis according to the median of NPC2 gene expression. The results showed that there were significant differences in the expression levels of NPC2 in different tumors and adjacent tissues, and the expression of NPC2 in most cancer tissues was higher than that in normal tissues. In the Chinese Glioma Genome Atlas (CGGA) dataset and TCGA dataset, we investigated whether there was a difference in the expression level of NPC2 in LGG and GBM patients, we found that the expression of NPC2 in GBM patients was significantly higher than that in LGG patients (Figure 1A), this finding is consistent with the results obtained from IHC sections of patients with human GBM (n = 11). (Figure 1B). Next, we analyzed the expression of NPC2 protein in different grades of GBM cells, and the expression level of NPC2 in GBM cells compared to LGG cells (Figure 1C). We further explored the prognostic value of NPC2 in independent datasets of CGGA (n = 1018), TCGA (n = 1681) and GSE4412 (n = 85), and the results showed that patients with high NPC2 expression had a worse prognosis than those with low expression (Figure 1D). These results suggested that NPC2 may play an important role in the development of GBM. However, considering the limited clinical sample size, the present finding should be regarded as preliminary.

Figure 1.

NPC2 was highly expressed in GBM patients and cell lines. (A) Pan‐cancer analysis of NPC2 expression was performed using the UCSCXenaShiny tool in Rstudio, showing significant differences in NPC2 expression between different tumor types and their adjacent normal tissues, with higher NPC2 expression in most cancer tissues than in normal tissues. In the CGGA (LGG = 625, GBM = 389) and TCGA (LGG = 515, GBM = 159) datasets, NPC2 expression was significantly higher in GBM patients than in LGG patients. (B) IHC staining of pathological tissues from GBM patients with different grades, showing that NPC2 expression gradually increased with increasing GBM grade. Representative images are shown (original magnification ×10; scale bar = 100 μM). n = 3 independent tissue samples per GBM grade.(C) Western blotting analysis of NPC2 protein expression in low‐grade and high‐grade GBM cell lines. Densitometric quantification of NPC2 protein levels is shown. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; *p < 0.05, **p < 0.01 versus the NC group. (D) Prognostic analysis of NPC2 expression in three independent datasets (CGGA, n = 1018; TCGA, n = 681; GSE4412, n = 85), showing that GBM patients with high NPC2 expression had worse prognosis than those with low NPC2 expression. Survival analysis was performed using the log‐rank test.

3.2. Knockdown of NPC2 Expression Could Inhibit GBM Cholesterol Metabolism and Tumor Migration Ability

To further explore the function of NPC2 in regulating the malignant phenotypes of GBM cells, we knocked down NPC2 expression via siRNA, achieving a knockdown efficiency of 60%. (Figure 2A), and after knockdown of NPC2, cell viability was reduced (Figure 2B). Total serum cholesterol (TC) is the sum of all cholesterol contained in all lipoproteins in the blood, and the NPC2 protein is an indispensable carrier of cholesterol in human cells. Compared with the control group, a lower expression level of NPC2 significantly inhibited intracellular cholesterol secretion, with the activity decreasing by approximately 40%. (Figure 2C) and could inhibit the rate of cell wound healing (Figure 2D) (Control stands for no treatment group, NC indicates transfection with blank si‐RNA). Consistently, knockdown of NPC2 in U87 cells also suppressed cell migration, as assessed by a wound healing assay (Supporting Information S1: Figure S1). These results indicted that NPC2 played an important regulatory role in GBM malignancy.

Figure 2.

Knockdown of NPC2 expression inhibits GBM cholesterol metabolism and tumor migration ability. (A) Three different siRNAs were used to downregulate NPC2 protein expression in the LN18 GBM cell line. Left: Representative Western blotting images of NPC2 protein levels; Right: Densitometric quantification of NPC2 protein levels. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; *p < 0.05, **p < 0.01 versus the NC group. (B) CCK‐8 showing decreased viability of LN18 cells after NPC2 knockdown. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; **p < 0.01 versus the NC group. (C) TC content assay showing reduced TC levels in LN18 cells after NPC2 knockdown. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; **p < 0.01 versus the NC group. (D) Cell migration assay (transwell or scratch) showing inhibited migration ability of LN18 cells after NPC2 knockdown. Representative images are shown (original magnification × 10; scale bar = 100 μM). Quantitative analysis of migration ability is presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; **p < 0.01 versus the NC group.

3.3. High Expression of NPC2 Was Closely Associated With the Occurrence of Temozolomide Resistance in Patients

TMZ is a first‐line chemotherapy drug for patients with GBM, and mutations in tumor tissue genes or long‐term use will lead to resistance to TMZ. The results showed that the expression of NPC2 was positively correlated with the expression of O⁶‐methylguanine‐DNA methyltransferase (MGMT), a gene associated with TMZ resistance (Figure 3A). In addition, analysis of cell resistance data also revealved that the expression of NPC2 was positively correlated with the IC50 of cells against TMZ. These results suggested that NPC2 was associated with the development of GBM malignancy and treatment resistance (Figure 3B). CCK‐8 assay results showed that inhibiting NPC2 expression significantly reduced cell viability, while TMZ treatment further enhanced this effect. Suggesting that NPC2 silencing can increase the sensitivity of GBM cells to TMZ (Figure 3C).

Figure 3.

The high expression of NPC2 is closely related to the occurrence of TMZ resistance in patients. (A) Prognostic analysis of NPC2 expression in GBM patients receiving chemoradiotherapy, showing that patients with high NPC2 expression had worse prognosis than those with low NPC2 expression; right: Correlation analysis showing a positive correlation between NPC2 expression and MGMT expression in patients receiving chemoradiotherapy. Patient sample size: n = 156 (from TCGA dataset). Survival analysis was performed using the log‐rank test; correlation analysis was performed using Pearson's correlation coefficient. (B) IC50 assay of TMZ in GBM cell lines with high or low NPC2 expression, showing significantly higher IC50 values in cell lines with high NPC2 expression; right: Correlation analysis showing a positive correlation between NPC2 expression and TMZ IC50 values in GBM cell lines. (C) Effect of different treatments on the viability of GBM cells detected by CCK‐8 assay. Both NPC2‐siRNA and TMZ treatment can inhibit the viability of GBM cells, and NPC2‐siRNA can enhance the sensitivity to TMZ. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; **p < 0.01 versus the Control group, ^## p < 0.01 versus the TMZ group.

3.4. SIM May Act as a Targeted Drug for NPC2 and Suppress the Malignant Behaviors of GBM

Transcriptome data of 51 GBM cell lines were downloaded from the CCLE database, and drug sensitivity data of 32 cell lines treated with TMZ were downloaded from the GDSC database, respectively. The intersection of cell lines from the two databases was obtained, resulting in 18 cell lines that simultaneously contained NPC2 expression data and TMZ‐treated IC50 values. Cell lines were divided into high‐expression and low‐expression groups based on the median NPC2 expression level. It was found that the IC50 values in the high‐expression group were significantly higher than those in the low‐expression group, and NPC2 expression was positively correlated with IC50 values. Cell lines were divided into high‐expression and low‐expression groups based on the median NPC2 expression level. It was found that the IC50 values in the high‐expression group were significantly higher than those in the low‐expression group, and NPC2 expression was positively correlated with IC50 values. Based on the NPC2 molecular interaction network and the CMap as well as L1000FWD databases, we initially screened out the small molecule compound SIM acting on NPC2‐related pathways. This conventional fat‐soluble lipid‐lowering drug can cross the blood‐brain barrier (Figure 4A,B). At the same time, SIM has also been shown to exert additional anti‐tumor effects in GBM (Fuentes‐Fayos et al. 2023). Further experiments demonstrated that SIM potentiated the cytotoxic effect of TMZ and reinforced its inhibitory action on cell migration, with significant differences observed across groups (Figure 4C–E). Additionally, SIM decreased the intracellular total cholesterol level (Figure 4F). The results of clonogenesis experiments showed that SIM could inhibit the proliferation of LN18 cells, and the inhibitory ability was stronger when combined with TMZ (Figure 4G). These results demonstrated the inhibitory effect of targeting NPC2 inhibitors on the malignant behavior of GBM, but the specific mechanism and structural basis of the effect need to be further elucidated.

Figure 4.

The drug SIM targeting NPC2 inhibits the malignant behavior of GBM. (A) Drug screening based on differential genes from the TCGA and CGGA datasets identified SIM as a potential adjuvant therapeutic drug for GBM. (B) CCK‐8 assay to detect the viability of LN18 cells treated with different concentrations of SIM, showing an IC50 value of 5.315 μM. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using nonlinear regression (for IC50 calculation). (C) Western blotting analysis of NPC2 protein levels in LN18 cells treated with SIM. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; **p < 0.01 versus the control group. (D) Scratch assay showing inhibited migration ability of LN18 cells after treatment with TMZ or SIM. Representative images are shown (original magnification × 10; scale bar = 100 μM). (E) Quantitative analysis of scratch healing rate, showing significantly inhibited cell migration ability in the TMZ + SIM combination group. Data are presented as mean ± S.D., n = 3. Statistical analysis was performed using one‐way ANOVA; **p < 0.01, ***p < 0.001 versus the control group. (F) Colony formation assay to verify the effect of NPC2 inhibition on LN18 cell proliferation, showing reduced colony formation ability after NPC2 inhibition. Representative colony images are shown; quantitative analysis of colony number is presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; **p < 0.01 versus the control group.

3.5. Inhibition of NPC2 Expression or Activity Might Inhibit the GPX4 Signaling Pathway

According to the KEGG database analysis, the pathways of NPC2 were significantly enriched, including cytokine receptor interaction, lysosomes, cholesterol metabolism, ferroptosis signaling pathway, etc. Gene Set Enrichment Analysis (GSEA) enrichment analysis was performed by differential genes and folds of differential expression of NPC2 in the high and low expression groups of NPC2, indicated that both iron homeostasis and iron transport were activated in the high expression group of NPC2, suggested that NPC2 may be a regulator of ferroptosis (Figure 5A). Further experiments showed that knockdown of NPC2 expression or the use of inhibitors significantly inhibited the expression of GPX4, an important factor in the ferroptosis pathway (Figure 5B,C). The TFAP2C gene is a sequence‐specific DNA‐binding transcription factor involved in the activation of multiple developmental genes, while SP1 is a member of the well‐known SP family of transcription factors, both of which are direct transcription factors of GPX4 (Dai et al. 2020). After inhibition of NPC2 expression levels, the mRNA expression levels of SP1 and TFAP2C were also reduced. However, it remains unclarified whether the downregulation of TFAP2C and SP1 stems from altered cholesterol compartmentalization or changed total intracellular cholesterol levels. To verify its specificity, we transfected LN18 cells with an NPC2 overexpression plasmid. The results showed that the NPC2 overexpression plasmid reversed the inhibitory effect induced by NPC2‐siRNA. The average overexpression efficiency of NPC2 reached 40%, and the mRNA levels of NPC2 and GPX4 were significantly upregulated in the overexpression group (Figure 5D). To further verify the type of cell death triggered by SIM, we applied a panel of cell death inhibitors (Z‐VAD‐FMK, Nec‐1, Lys05 and Fer‐1) to SIM‐treated GBM cells. CCK‐8 results showed that the ferroptosis inhibitor Fer‐1 exerted the most prominent protective effect on cell viability (Figure 5E). Intracellular lipid peroxidation was assessed via BODIPY 581/591 C11 staining. SIM led to robust lipid peroxide accumulation, and Fer‐1 effectively attenuated this phenotype (Figure 5F). We next analyzed key ferroptosis‐associated indicators. Exposure to SIM dramatically raised intracellular Fe²⁺ concentration, which was subsequently diminished by Fer‐1. In parallel, SIM treatment downregulated glutathione (GSH) and malondialdehyde (MDA), and Fer‐1 co‐treatment recovered their expression levels (Figure 5G). These results suggested that NPC2 may regulate GBM biology through GPX4‐mediated ferroptosis signaling, but the specific mechanism needs to be further elucidated.

Figure 5.

Inhibition of NPC2 expression or activity suppresses the GPX4 signaling pathway. (A) KEGG enrichment analysis of NPC2‐related signaling pathways, showing activated iron homeostasis and iron transport pathways in the high‐NPC2 expression group, suggesting that NPC2 may be a regulator of ferroptosis. (B) Correlation analysis showing a positive correlation between NPC2 expression and GPX4 expression in GBM, suggesting that NPC2 may be a negative regulator of ferroptosis. (C) Western blotting analysis of Srebp‐1c, GPX4, and NPC2 protein levels in LN18 cells treated with SIM or NPC2 siRNA. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; *p < 0.05, **p < 0.01 versus the control group. (D) RT‐qPCR analysis of GPX4, NPC2, SP1, and TFAP2C mRNA levels in LN18 cells. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using one‐way ANOVA; *p < 0.05, **p < 0.01 versus the control group, ^# p < 0.05, ^## p < 0.01 versus the control group. (E) CCK‐8 assay to detect the rescue effect of Z‐VAD‐FMK (apoptosis inhibitor), Nec‐1 (necroptosis inhibitor), Lys05 (autophagy inhibitor), and Fer‐1 (ferroptosis inhibitor) on SIM‐induced LN18 cell growth inhibition. Data are presented as mean ± SD, n = 3. Statistical analysis was performed using two‐way ANOVA; *p < 0.05, **p < 0.01 versus the SIM group. (F) Detection using the BODIPY 581/591 C11 probe revealed the accumulation of lipid peroxides in the LN18 cell line. (G) The levels of Fe²⁺, GSH and MDA in LN18 cell lines were measured after the addition of SIM and Fer‐1.

3.6. NPC2 Inhibitor Could Reduce Tumor Volume in GBM Xenograft Mice

To verify the effect of TMZ, SIM, and the combination of two drugs on GBM in vitro, we established a subcutaneous xenograft model. The dose administered for SIM was 2 mg/kg and the dose for TMZ was 20 mg/kg, which was the same as the drug used in cell experiments. The xenograft was seen to be the size of a soybean on Day 7, its volume was measured and drug treatment was given, then every 3 days until Day 21 of dosing. Until the end of the injection, all animals survived with a relatively good mental state, and no relevant adverse reactions were observed. The data showed that from day 6 onwards, the volume of tumors in the treated group was significantly smaller than that of the control group. At the end of the dosing cycle, the tumor volume was the smallest in the TMZ and SIM synergistic treatment group (Figure 6A). After peeling off the tumor tissue, rinsing with PBS, and visually inspecting the surface condition and volume of the tumor tissue, we found that the tumor size was significantly reduced in the drug treatment group, and the SIM‐assisted TMZ group had the best treatment effect (Figure 6B). Part of the tissues were collected and fixed in 4% neutral formalin solution, and thr tissue were sectioned after paraffin embedding. The immunohistochemistry results showed that the expression levels of NPC2, GPX4 and Ki‐67 were reduced in the drug treatment group compared with the control group, indicating that SIM‐assisted TMZ treatment of GBM could significantly inhibit tumor growth and proliferation of tumors (Figure 6C). The mice in all experimental groups exhibited no significant changes in body weight (Supporting Information S1: Figure S2). The WB results showed that the protein expression levels of NPC2 and GPX4 decreased after the use of SIM, and further decreased after the combination of TMZ (Figure 6D). Finally, TEM allowed us to observe changes in the mitochondria of cells. In particular, the addition of SIM led to a reduction in the size of the mitochondrial ridge and the disappearance of the mitochondrial ridge (Figure 6E).

Figure 6.

NPC2 inhibitor reduces tumor volume in GBM xenograft mice. (A) Tumor volume measurements of GBM xenograft mice (n = 5 per group: control, SIM, TMZ, TMZ + SIM) recorded every 3 days after drug administration. Data are presented as mean ± SD. Statistical analysis was performed using two‐way ANOVA; **p < 0.01, ***p < 0.001 versus the control group (B) Representative images of xenograft tumors from each group (n = 5 per group). (C) Representative IHC images showing the expression levels of GPX4, Ki‐67, and NPC2 in mouse tumor sections (original magnification × 10; scale bar = 100 μm). (D) Western blotting analysis of GPX4 and NPC2 protein levels in mouse tumor tissues treated with SIM or TMZ. Data are presented as mean ± SD, n = 3. One‐way ANOVA was used for statistical analysis; *p < 0.05, **p < 0.01 versus the control group. (E) TEM images showing mitochondrial changes in mouse tumor tissues from the control and drug‐treated groups (scale bar = 5 μm; original magnification × 1000). Biological repeats: n = 3 independent tumor tissue samples per group, with 2–3 TEM fields observed per sample.

4. Discussion

Over the past 20 years, GBMs have been the most common primary malignant brain tumors in adults, accounting for nearly 80% of such tumors. Their treatment has not improved significantly due to unrestricted cancer cell proliferation, apoptosis resistance, and metastasis (Weller et al. 2024). As GBM progression is primarily a molecular disease, identifying appropriate molecular targets is essential for its understanding and treatment (Yang et al. 2022). NPC2 is associated with gastric, liver, and thyroid cancer occurrence/metastasis, but its role in GBM remains unstudied (Yao et al. 2023; Wang et al. 2024). This study identifies NPC2 as a potential GBM diagnostic/treatment target, which may regulate GBM biology via GPX4‐mediated ferroptosis signaling.

Cholesterol and its metabolites are critical in cancer, regulating tumor processes like oncogenic signaling, ferroptosis, and the tumor microenvironment (Huang et al. 2020). Screening for key cholesterol metabolism molecules is a research focus (Xu et al. 2020). NPC proteins (lysosome‐localized) mediate extracellular cholesterol uptake via vesicular endocytosis; cholesterol metabolism reprogramming is a GBM hallmark (Guo et al. 2022). NPC deficiency causes lysosomal cholesterol accumulation, leading to lysosomal defects and dysfunction of autophagosomes, mitochondria, and peroxisomes, as well as impaired autophagic vesicle transport to lysosomes (Meng et al. 2020). Targeting NPC to regulate cholesterol may thus be a novel GBM treatment strategy. Exploratory studies show NPC2 expression is higher in GBM than normal brain tissue, positively correlated with tumor malignancy and poor prognosis; NPC2 knockdown inhibits GBM cell proliferation and invasion.

Notably, beyond the transcriptional regulation of GPX4, cellular cholesterol levels directly determine the intrinsic ferroptosis sensitivity of GBM cells. Cholesterol is a key component of cell membrane lipid bilayers; altered intracellular cholesterol content reshapes membrane lipid composition, affects the abundance of polyunsaturated fatty acids (PUFAs) on the cell membrane, and further modulates the accumulation of toxic lipid ROS, which is the core determinant of ferroptosis occurrence. Reduced intracellular cholesterol levels remodel membrane lipid profiles, enrich membrane PUFA deposition, and significantly increase the vulnerability of GBM cells to lipid peroxidation, thereby enhancing ferroptosis sensitivity. In contrast, elevated cholesterol content stabilizes membrane lipid structure, alleviates lipid oxidative damage, and renders GBM cells resistant to ferroptosis.

Programmed cell death, including apoptosis, pyroptosis, necroptosis, and ferroptosis, plays a vital role in maintaining tissue homeostasis and suppressing tumorigenesis (Griffioen and Nowak‐Sliwinska 2022). Ferroptosis is an iron‐dependent lipid peroxidation‐driven cell death modality, and GPX4 acts as the central inhibitory regulator of ferroptosis by maintaining cellular lipid homeostasis and eliminating toxic lipid ROS (Han et al. 2022; Taufani et al. 2023). Multiple transcription factors, including CREM‐tau, EGR1, and SREBP1(which stimulates SP1/TFAP2C transcription), are involved in the transcriptional regulation of GPX4, among which SREBP1 coordinates de novo lipid synthesis and lipid repair processes (Zhao et al. 2021; DeBose‐Boyd and Ye 2018). Structurally, the 3’ UTR of GPX4 mRNA contains a SECIS element that mediates selenocysteine incorporation via UGA codon recoding, rendering GPX4 expression dependent on intracellular selenium availability (Fang et al. 2006; Lammi and Qu 2018). Moreover, selenocysteine tRNA activation requires an isoamyl lipid group derived from the MVA pathway (Xu et al. 2010). Statins, as classic MVA pathway inhibitors, can block cholesterol synthesis, downregulate GPX4 expression, and promote lipid peroxidation and ferroptosis in GBM cells. Accumulating evidence also indicates that mesenchymal‐like tumor cells exhibit high sensitivity to GPX4 inhibition and ferroptotic stress (Forcina and Dixon 2019).

Further studying NPC protein function and cholesterol transport roles will clarify intracellular lipid transport mechanisms and provide new tumor treatment strategies targeting cholesterol metabolism. Exploratory studies confirm higher NPC2 expression in GBM than normal tissue, positive correlations with malignancy and poor prognosis (Ikliptikawati et al. 2023; Zheng et al. 2024). NPC2 knockdown inhibits GBM cell proliferation/invasion and reduces GPX4 expression, increasing ferroptosis hallmarks (partially reversed by Fer‐1). Similarly, SIM treatment inhibits NPC2 and GPX4 expression, disrupts intracellular cholesterol homeostasis, and triggers robust ferroptosis in GBM cells. Collectively, these findings demonstrate that NPC2‐dependent cholesterol homeostasis not only participates in the transcriptional regulation of GPX4 but also directly modulates membrane lipid peroxidation levels, thereby governing the ferroptosis sensitivity of GBM cells. Nevertheless, the causal relationship between NPC2‐mediated cholesterol metabolism and ferroptosis regulation remains preliminary and requires further in‐depth verification.

GO and KEGG enrichment analyses revealed that NPC2 expression is correlated with immune responses and drug resistance in GBM. Specifically, NPC2 expression showed a positive correlation with the expression of immune checkpoints and immune activators; in the TCGA and CGGA datasets, high NPC2 expression was associated with higher immune scores, lower tumor purity, and higher IPS (Immune Phenotype Score) scores, implying that GBM patients with high NPC2 expression might potentially benefit from immunotherapy (Supplementary Figure S3). Additionally, differential expression of NPC2 was found to affect the drug sensitivity of GBM cells, which provides preliminary reference for guiding medication strategies in clinical practice.

The therapeutic relevance of NPC2 inhibition needs further strengthening in future investigations. In particular, although our present data reveal that NPC2 suppression enhances TMZ‐induced ferroptosis in conventional GBM cells, we have not yet validated whether pharmacological or genetic NPC2 inhibition is capable of reversing TMZ resistance in established TMZ‐resistant GBM cell and animal models. Follow‐up research will focus on this critical issue: we will construct TMZ‐resistant GBM cell lines and corresponding drug‐resistant subcutaneous or orthotopic xenograft models to systematically explore whether blocking NPC2 expression restores TMZ chemosensitivity. Further relevant in vitro and in vivo validations will be performed to consolidate the preclinical therapeutic value of targeting NPC2 for overcoming TMZ resistance in GBM.

5. Conclusions

In conclusion, our experiments were the first to clarify the NPC2‐mediated cholesterol metabolism and its signaling pathway (GPX4), establish a molecular interaction network of cholesterol metabolism‐mediated signaling and transcriptional expression in the progression of GBM malignancy, and find the root cause of GBM malignancy from the aspect of tumor lipid metabolism regulation, so as to provide a theoretical basis for the prevention and treatment of GBM recurrence and deterioration. This study is helpful for the clinical application of translational medicine, and is expected to provide new ideas and targets for the treatment of related diseases. However, this study also has shortcomings, and it is necessary to construct a mouse model of orthotopic xenograft for further validation. In this study, the mechanism of the relevant molecules has not been deeply explored. In future studies, we will further investigate the vital role of NPC2 in GBM tumorigenesis.

Author Contributions

Wei Jiang: methodology, writing – original draft preparation. Wei Wang: data curation, software. Junping Wang: investigation. Fan Xia: visualization. Guoyu Wang: software. Hongzhi Wang: funding acquisition. Xueran Chen: conceptualization project administration, writing – review and editing. Zhiyou Fang: conceptualization, project administration, funding acquisition.

Ethics Statement

This study is strictly conducted on the basis of the Declaration of Helsinki, and was approved by the ethical committee of the Hefei Cancer Hospital of the Chinese Academy of Sciences (Hefei, China; PJ‐KY2024‐011).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1

Jiang, W. , Wang W., Wang J., et al. 2026. “NPC2 Regulates Glioblastoma Cholesterol Metabolism and Malignant Behavior via GPX4‐Linked Ferroptosis.” Cell Biology International 50: e70184. 10.1002/cbin.70184.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.