Accumulating evidence has demonstrated that drug resistance can be acquired in cancer through the repopulation of tumors by cancer stem cell (CSC) expansion. Here, we investigated mechanisms driving resistance and CSC repopulation in hepatocellular carcinoma (HCC) as a cancer model using two drug-resistant, patient-derived tumor xenografts that mimicked the development of acquired resistance to sorafenib or lenvatinib treatment observed in patients with HCC. RNA sequencing analysis revealed that cholesterol biosynthesis was most commonly enriched in the drug-resistant xenografts. Comparison of the genetic profiles of CD133+ stem cells and CD133 bulk cells from liver regeneration and HCC mouse models showed that the cholesterol pathway was preferentially upregulated in liver CSCs compared with normal liver stem cells. Consistently, SREBP2-mediated cholesterol biosynthesis was crucial for the augmentation of liver CSCs, and loss of SREBP2 conferred sensitivity to tyrosine kinase inhibitors, suggesting a role in regulation of acquired drug resistance in HCC. Similarly, exogenous cholesterol-treated HCC cells showed enhanced cancer stemness abilities and drug resistance. Mechanistically, caspase-3 (CASP3) mediated cleavage of SREBP2 from the endoplasmic reticulum to promote cholesterol biosynthesis, which consequently caused resistance to sorafenib/lenvatinib treatment by driving activation of the sonic hedgehog signaling pathway. Simvastatin, an FDA-approved cholesterol-lowering drug, not only suppressed HCC tumor growth but also sensitized HCC cells to sorafenib. These findings demonstrate that CSC populations in HCC expand via CASP3-dependent, SREBP2-mediated cholesterol biosynthesis in response to tyrosine kinase inhibitor therapy and that targeting cholesterol biosynthesis can overcome acquired drug resistance.

Significance:

This study finds that cholesterol biosynthesis supports the expansion of cancer stem cell populations to drive resistance to tyrosine kinase inhibitor therapy in hepatocellular carcinoma, identifying potential therapeutic approaches for improving cancer treatment.

The emergence of acquired drug resistance is an obstacle to effective cancer treatment. Resistance to molecularly targeted drugs, such as tyrosine kinase inhibitors (TKI), develops in most patients with cancer and limits their long-term survival. In the clinic, TKIs are administered in multiple cycles over a period of time, and consequently, residual surviving cancer cells repopulate tumors, resulting in tumor relapse. Accumulating evidence has emerged in support of a cancer stem cell (CSC) model in a wide range of solid tumors (1–4). CSCs are now regarded as the source of tumor recurrence and therapeutic resistance. Although we and others have shown that TKI-resistant tumors contain enhanced CSC populations (5–7), how tumors expand CSC populations in response to drug treatment remains largely unknown.

Hepatocellular carcinoma (HCC) treatment recently entered a new era with the development of molecular-targeted therapies, as sorafenib led to improved survival of patients with advanced HCC (8). However, the survival benefit in the sorafenib treatment arm was modest, with median survival times 2.8 and 2.3 months longer than that of the placebo arm in two large-scale trials (8, 9). Recently, lenvatinib was approved by the FDA as a first-line treatment for unresectable patients with advanced HCC. In a recent phase III clinical trial of lenvatinib for patients with HCC (NCT01761266), lenvatinib was noninferior to sorafenib in terms of overall survival in untreated advanced HCC (10). On the basis of these clinical observations, the unsatisfactory survival benefits of sorafenib/lenvatinib may be due to acquired drug resistance. Using HCC as a cancer model, we aimed to understand the molecular mechanism of how cancer cells acquire drug resistance by establishing sorafenib- and lenvatinib-resistant HCC patient-derived tumor xenografts (PDTX) in vivo via the administration of sorafenib and lenvatinib, respectively; this approach mimics the clinical situation in which acquired resistance develops in patients with HCC in response to sorafenib/lenvatinib treatment. Using RNA sequencing analysis, we compared the expression profiles between these two drug-resistant PDTXs and the corresponding mock controls. Upon ingenuity pathway analysis (IPA), we found that cholesterol biosynthesis was most commonly upregulated in sorafenib- and lenvatinib-resistant PDTXs. Consistent with this finding, we found that cholesterol deposition was increased in these xenografts and in drug-resistant HCC cell lines. This, together with the observation that SREBP2-mediated cholesterol biosynthesis was increased in enriched liver CSC populations, but not in normal liver stem cells, prompted us to investigate the potential role of SREBP2-mediated cholesterol biosynthesis in the regulation of drug resistance in HCC by augmenting liver CSCs. To date, cholesterol metabolism was reported to regulate cancer stemness and drug resistance in gastrointestinal cancers (11, 12).

Using CRISPR activation and shRNA knockdown approaches, we found that SREBP2-mediated cholesterol biosynthesis was critically involved in the regulation of liver CSCs and bears clinical significance. Strikingly, this process is a critical determinant of drug resistance in HCC cells. Exogenous cholesterol-treated and high cholesterol-utilizing HCC cells exhibited similar effects on cancer stemness and drug resistance. Specifically, molecularly targeted drugs, including sorafenib and lenvatinib, induced the activation of caspase-3 (CASP3), which subsequently induced the nuclear translocation of SREBP2 from the endoplasmic reticulum, resulting in activation of the cholesterol biosynthesis-driven sonic hedgehog signaling (SHH) pathway. Simvastatin, an FDA-approved cholesterol-lowering drug for the treatment of high cholesterol patients, sensitized HCC cells to sorafenib/lenvatinib by inhibiting liver CSC populations. Using PDTX models, we found that simvastatin at the clinically equivalent dose not only suppressed HCC tumor growth but also sensitized HCC cells to sorafenib. Collectively, CASP3-dependent SREBP2-mediated cholesterol biosynthesis regulates the drug resistance of HCC via augmentation of liver CSCs, and its inhibition sensitizes cells to conventional therapies, including sorafenib/lenvatinib.

Plasmids and reagents

Plasmids encoding wild-type and mutant (D468A) SREBP2 have been described in ref. 13. Lenvatinib was purchased from Selleckchem. Sorafenib was purchased from LC Laboratories. InSolution Simvastatin (#567022) was purchased from Calbiochem. GANT61 (HY-13901), Z-DEVD-FMK (HY-12466), Betulin (HY-N0083), and Erastin (HY-15763) were purchased in MedChemExpress. Lipoprotein-deficient serum (LPDS; #S5394) was purchased from Sigma-Aldrich.

Cell lines and cell culture

Human HCC cell lines MHCC-97L (Liver Cancer Institute, Fudan University), Hep3B (ATCC), HepG2 (ATCC), Huh7 and PLC/PRF/5 (Japan Cancer Research Bank), MIHA (a gift from Dr. J.R. Chowdhury, Albert Einstein College of Medicine, Bronx, NY), 293FT (Invitrogen), and HEK293T (ATCC) were maintained in DMEM with high glucose and l-glutamine (Gibco; Invitrogen) supplemented with 10% heat-inactivated FBS (Gibco, Invitrogen), 100 mg/mL penicillin G, and 50 μg/mL streptomycin (Gibco, Invitrogen) at 37°C in a humidified chamber containing 5% CO2. All cell lines used in this study were obtained between 2013 and 2016, regularly authenticated by morphologic observation and AuthentiFiler STR (Invitrogen) and tested for the absence of mycoplasma contamination (MycoAlert, Lonza). Experiments were performed within 20 passages after cell thawing.

Patient samples

Fifty archived paraffin-embedded pathologic specimens from primary patients with HCC were collected along with complete clinical and pathologic data at the Sun Yat Sen University Cancer Center. All samples were anonymous. This study was approved by the Institute Research Medical Ethics Committee. None of the patients had received radiotherapy or chemotherapy before surgery. Ninety-one HCC samples were randomly retrieved from patients with HCC who underwent curative resection followed by sorafenib treatment at the Eastern Hepatobiliary Surgery Hospital from December 2008 to May 2010. Postsurgical resection, sorafenib was given to the patients at a dose of 400 mg twice a day. Study was approved by the Ethical Committee of the Second Military Medical University. Venous blood samples were collected from 27 nonalcoholic fatty liver disease (NAFLD)-related patients with HCC, which were collected from patients with HCC before any therapeutic procedures were performed. Use of human samples was approved by the committee for ethical review of research involving human subjects at School of Public Health, Sun Yat Sen University. Tumors and paired nontumorous liver tissues from 17 patients diagnosed with HCC complicating nonalchoholic steatohepatitis were obtained from Prince of Wales Hospital, the Chinese University of Hong Kong and Zhongshan Hospital of Fudan University. This study was approved by the ethics committee of the Chinese University of Hong Kong and the clinical research ethics committee of Zhongshan Hospital of Fudan University. All the above samples were obtained from patients with their written informed consent.

Knockdown of CASP3 and SREBP2

SREBP2 shRNA expression vectors were cloned into the to pLKO.1 vector (Addgene). The scrambled shRNA nontarget control (NTC) was purchased from Sigma-Aldrich. CASP3 shRNA expression vectors were cloned into pLentivector (Vigene Biosciences). The scrambled shRNA control (shCTL) were purchased from Vigene. Clone ID of the two shRNAs directed against SREBP2 are as follow: TRCN0000020666 and TRCN0000020668. Product identification of the two shRNAs directed against CASP3 are as follow: pLenti-U6-CASP3-shRNA306-CMV-Puro and pLenti-U6-CASP3-shRNA524-CMV-Puro. Transduced cells were selected with 2 μg/mL puromycin. SREBP2 siRNA expression oligo duplex were purchased from Origene (#SR304580) with locus ID 6721, and consequently transfected at 20 nmol/L into sorafenib-resistant HCC cells. Supplementary Table S1 lists the sequences of the siSREBP2, shSREBP2, shCASP3, siCTL, NTC, and shCTL used.

Overexpression of SREBP2 by CRISPR activation

Gene activation of SREBP2 was performed using Edit-R transcriptional activation system (Dharmacon). Hep3B cells were first engineered for stably expression of dCas9-VP64-p65-Rta (dCas9-VPR) using blasticidin. Subsequent transduction of lentiviral CRISPR activation (CRISPRa) sgRNA targeting the promoter region of SREBP2 was performed in dCas9-VPR stable cells. Product identification of the two sgRNAs directed against SREBP2 are as follow: SVC18111602 1-B-03 and SVC18111602 1-B-04. Gene activation of SREBP2 for expression was selected by puromycin. Supplementary Table S1 lists the sequences of the SREBP2 and CTL used.

RNA extraction and qRT-PCR analysis

Total RNA was isolated using TRIzol reagent according to the manufacturer's protocol (Invitrogen). Complementary DNA (cDNA) was synthesized using PrimeScript RT Reagent Kit (Takara) according to the manufacturer's instructions and then subjected to qPCR with BrightGreen 2× qPCR Master mix (Applied Biological Materials) using QuantStudio 7 Flex Real-Time PCR System (Applied Biosystems) with primers specific to the sequences of genes of interest, which were provided in Supplementary Table S2. Relative expression differences were calculated using 2−ΔΔCTmethod with reference to β-ACTIN.

PDTXs

PDTX#1 and PY003 HCC tissues were obtained from patients undergoing hepatectomy at Queen Mary Hospital, Hong Kong and Pamela Youde Nethersole Eastern Hospital, Hong Kong respectively. Samples were collected from patients who had not received any previous local or systemic treatment prior to operation. Informed consent was obtained from all patients before the collection of liver specimens and the study was approved by the Ethics Committee of the University of Hong Kong. The use of human clinical specimens was approved by the Institutional Review Board (IRB) of the University of Hong Kong/Hospital Authority Hong Kong West Cluster. Consent from patients was obtained.

In vitro establishment of drug-resistant HCC cells

Sorafenib-resistant clones of PLC/PRF/5 and MHCC-97L were established by subjecting HCC cells to continuous administration of gradually increasing sorafenib concentrations and were trained up to 10 μmol/L. Same volume of DMSO was added to the cells as mock controls during establishment of these resistant cells. Lenvatinib-resistant clones of PLC/PRF/5 and MHCC-97L were established by subjecting HCC cells to continuous administration of gradually increasing lenvatinib concentrations and were trained up to 30 μmol/L.

In vivo establishment of drug-resistant HCC cells

Sorafenib-resistant PDTX#1 was established by administering sorafenib orally at 100 mg/kg/day in NOD/SCID mouse bearing PDTX#1 for 31 days. Successful establishment of sorafenib resistance was evidenced by an observation that there was no tumor suppression effect upon sorafenib treatment after two more rounds of sorafenib administration to secondary and third mouse recipient. Lenvatinib-resistant PY003 was established by administering lenvatinib orally at 30 mg/kg/day in NOD/SCID mouse bearing PY003 for 21 days. Same treatment protocol was applied to the secondary and third mouse recipient. The fourth mouse recipient was subjected to lenvatinib feeding for 16 days. Successful establishment of lenvatinib resistance was evidenced by an observation that there was no tumor suppression effect upon lenvatinib treatment after three more rounds of lenvatinib administration. The study protocol was approved by and performed in accordance with the guidelines for the Use of Live Animals in Teaching and Research at Hong Kong Polytechnic University.

In vivo drug treatment assay

A total of 1 × 106 PDTX PY003 and 0.5 × 106 sorafenib-resistant PDTX#1 cells, were prepared according to the cell dissociation protocol in Supplementary Information and were injected into the flanks of BALB/C nude mice. Once the tumors were established and reached approximately 7 mm × 7 mm (length × width), the mice were randomly divided into four groups: DMSO; DMSO and simvastatin (4 mg/kg; Calbiochem), sorafenib (30 mg/kg; LC Laboratories), and the combined treatment group. Simvastatin was dissolved in H2O. Sorafenib was dissolved in DMSO before diluting in water. The mice were given simvastatin/sorafenib orally on a daily basis. The tumor volume and body weight were measured every 3 days. The tumor volume was calculated using the following formula: volume (cm3) = L × W2 × 0.5. The mice were treated for 21 days before sacrifice, at which point, tumors were harvested for analysis. The study protocol was approved by and performed in accordance with the guidelines for the Use of Live Animals in Teaching and Research at Hong Kong Polytechnic University. No specific randomization method was used. Sample size of animals was chosen based on significant P values.

Statistical analysis

Statistical significance of qRT-PCR, limiting dilution assay, flow cytometry analysis, migration assay, and invasion assay results were determined by Student t test using Microsoft Office Excel software (Microsoft Corporation). The displayed results showed the means and the standard deviations, and those with P values less than 0.05 were considered statistically significant (*, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001). All tests are two-sided. Data point is excluded if it deviates from mean with more than three SDs. Investigators were not blinded to the group allocation during experiment and when assessing the outcome in all experiments including animal experiments. There is no estimate of variation within each group of data. Variance is similar between the groups that are being statistically compared. Chi square test was employed to examine the correlation of protein expression in HCC samples. Kaplan–Meier survival analysis was used to analyze overall survival and disease-free survival and a log-rank test was used to determine the statistical significance; these analyses were carried out using SPSS 20 software.

Data availability

More detailed methods are available in the Supplementary Materials and Methods. The data generated in this study are available within the article and its Supplementary data files. The sequencing data generated from this study are publicly available in Gene Expression Omnibus (GSE168783, GSE191224, and GSE192718).

Cholesterol biosynthesis was activated in drug-resistant HCC PDTXs

In an attempt to identify critical molecules/pathways involved in acquired drug resistance, we established sorafenib- and lenvatinib-resistant PDTXs derived from PDTX#1 (14) and PY003, which were sensitive to sorafenib/lenvatinib treatment, by administration of three rounds and four rounds of sorafenib or lenvatinib, respectively (Fig. 1A). Successful establishment of sorafenib and lenvatinib resistance was shown by the observation that there was no tumor suppression effect of the sorafenib/lenvatinib treatments (Fig. 1B). RNA-sequencing analysis coupled with ingenuity pathway analysis (IPA) showed that cholesterol biosynthesis was most significantly and commonly upregulated in sorafenib- and lenvatinib-resistant PDTXs (Fig. 1C). Heat-map representations of the differential expression of genes associated with cholesterol biosynthesis showed that these genes were deregulated in these two drug-resistant PDTXs (Fig. 1D). Among these genes, they are controlled by SREBP2 in PDTX#1 and PY003 with an activation Z-score of 2.979 and 3.002, respectively (Fig. 1E). Consistently, an increase in filipin-labeled free cholesterol was observed in these two drug-resistant PDTXs (Fig. 1F). Because SREBP1 activation was not significantly observed in these cells (Supplementary Fig. S1), we believe that SREBP2-mediated cholesterol biosynthesis is specifically activated in drug-resistant HCC PDTXs.

Figure 1.

SREBP2-mediated cholesterol biosynthesis was enriched in sorafenib- and lenvatinib-resistant PDTXs. A, Schematic diagram showing the workflow for the establishment of sorafenib- and lenvatinib-resistant PDTXs. Sorafenib-resistant PDTX#1 was established by administering sorafenib orally at 100 mg/kg for three rounds, whereas lenvatinib-resistant PDTX PY003 was established by administering lenvatinib orally at 30 mg/kg for four rounds. B, The dose–response curves showed that there was no tumor reduction upon sorafenib/lenvatinib administration when compared with their corresponding mock controls. C, Upon IPA analysis, top 10 most deregulated canonical pathways were shown in sorafenib- and lenvatinib-resistant PDTXs, among which, superpathway of cholesterol biosynthesis and cholesterol biosynthesis I was most commonly upregulated when compared with their corresponding mock controls. D, Heatmap analysis was performed on the basis of quantity of different genes in cholesterol biosynthesis. Level of gene expression is indicated by the color index. Color from blue to red indicates low to high expression in log2-fold change. E, Upstream regulator analysis showed that SREBP2 was activated with activation Z-scores of 2.979 and 3.002 in the sorafenib- and lenvatinib-resistant PDTX#1 and PY003, respectively. F, Representative filipin staining of tumors from sorafenib- and lenvatinib-resistant PDTX#1 and PY003 and their mock controls. Scale bar, 25 μm. Top, filipin staining (blue); middle, propidium iodide (red); bottom, merge. Intensity of filipin staining was quantified using Nikon NIS-Elements software. **, P < 0.01.

Figure 1.

SREBP2-mediated cholesterol biosynthesis was enriched in sorafenib- and lenvatinib-resistant PDTXs. A, Schematic diagram showing the workflow for the establishment of sorafenib- and lenvatinib-resistant PDTXs. Sorafenib-resistant PDTX#1 was established by administering sorafenib orally at 100 mg/kg for three rounds, whereas lenvatinib-resistant PDTX PY003 was established by administering lenvatinib orally at 30 mg/kg for four rounds. B, The dose–response curves showed that there was no tumor reduction upon sorafenib/lenvatinib administration when compared with their corresponding mock controls. C, Upon IPA analysis, top 10 most deregulated canonical pathways were shown in sorafenib- and lenvatinib-resistant PDTXs, among which, superpathway of cholesterol biosynthesis and cholesterol biosynthesis I was most commonly upregulated when compared with their corresponding mock controls. D, Heatmap analysis was performed on the basis of quantity of different genes in cholesterol biosynthesis. Level of gene expression is indicated by the color index. Color from blue to red indicates low to high expression in log2-fold change. E, Upstream regulator analysis showed that SREBP2 was activated with activation Z-scores of 2.979 and 3.002 in the sorafenib- and lenvatinib-resistant PDTX#1 and PY003, respectively. F, Representative filipin staining of tumors from sorafenib- and lenvatinib-resistant PDTX#1 and PY003 and their mock controls. Scale bar, 25 μm. Top, filipin staining (blue); middle, propidium iodide (red); bottom, merge. Intensity of filipin staining was quantified using Nikon NIS-Elements software. **, P < 0.01.

Close modal

Increased cholesterol biosynthesis was preferentially observed in enriched liver CSC populations

Drug-resistant cells showed enriched liver CSC populations (6). Therefore, we examined whether cholesterol biosynthesis was augmented in enhanced liver CSC populations by comparing the expression profiles between drug-resistant hepatospheres and differentiated counterparts derived from PLC/PRF/5 cells, which differ in self-renewal and tumorigenicity (15). Consistently, cholesterol biosynthesis was found to be the most upregulated (Fig. 2A), which is also controlled by SREBP2 with an activation Z-score of 4.318 using upstream regulator analysis (Fig. 2B). Consistently, an increase in the free cholesterol deposition, together with upregulation of cytoplasmic and nuclear SREBP2 were observed in hepatospheres by IF staining (Fig. 2C). Given the molecular similarities between liver CSCs and normal liver stem cells, we examined whether cholesterol biosynthesis is preferentially activated in liver CSCs. First, we sorted CD133+ cells and CD133 cells from a 0.1%DDC (3,5-diethoxycarbonyl-1,4-dihydrocollidine)-induced liver regeneration mouse model and two HCC mouse models, including (i) a mouse model with hydrodynamic injection of activated forms of NRAS/AKT (14) and (ii) a model with diethylnitrosamine (DEN)/carbon tetrachloride (CCl4)-induced fibrosis-related HCC (16). Next, we analyzed the enriched signaling pathways in CD133+ cells, using CD133 cells as corresponding baselines, in these three mouse models by RNA-sequencing analysis (Fig. 2D). When compared with that of regenerating liver, cholesterol homeostasis was found to be the most commonly upregulated pathway in CD133+ HCC cells from two HCC mouse models (Fig. 2E).

Figure 2.

Preferential activation of SREBP2-mediated cholesterol biosynthesis in liver CSCs. A, Upon IPA analysis, top 10 most deregulated canonical pathways were shown in enriched CSC populations (hepatospheres) when compared with their differentiated progenies. Superpathway of cholesterol biosynthesis and cholesterol biosynthesis I were most upregulated in enriched liver CSC populations. B, Upstream regulator analysis showed that SREBP2 was activated with activation Z-score of 4.318 in the CSC-enriched hepatospheres. C, IF images showed increase in nuclear and cytoplasmic SREBP2 expression and cholesterol deposition in enriched CSC subpopulations (hepatospheres from PLC/PRF/5) when compared with the differentiated progenies by filipin staining. Left, filipin staining (blue); SREBP2 staining (green); propidium iodide (red). Scale bar, 75 μm. Intensity of cytoplasmic and nuclear staining of SREBP2 was quantified using Nikon NIS-Elements software. D, The schematic diagram showing the workflow of genetic profiling and pathway analysis between CD133+ cells and CD133 from liver regeneration and HCC mouse models. CD133+ cells and CD133 (CD45/TER119) liver cells were isolated from mouse tissues from 0.1%DDC diet-treated mice and two HCC models including NRAS/AKT- and DEN/CCL4-induced HCC models by FACS approach. GSEA pathway analysis was performed to compare CD133+/CD133 in regenerating liver with two HCC models. E, When compared with regenerating livers, we found that cholesterol homeostasis was most enriched in CD133+ liver cells from two HCC models with a normalized enrichment score of 1.61 (FDR q value of 0.048). *, P < 0.05; ***, P < 0.001.

Figure 2.

Preferential activation of SREBP2-mediated cholesterol biosynthesis in liver CSCs. A, Upon IPA analysis, top 10 most deregulated canonical pathways were shown in enriched CSC populations (hepatospheres) when compared with their differentiated progenies. Superpathway of cholesterol biosynthesis and cholesterol biosynthesis I were most upregulated in enriched liver CSC populations. B, Upstream regulator analysis showed that SREBP2 was activated with activation Z-score of 4.318 in the CSC-enriched hepatospheres. C, IF images showed increase in nuclear and cytoplasmic SREBP2 expression and cholesterol deposition in enriched CSC subpopulations (hepatospheres from PLC/PRF/5) when compared with the differentiated progenies by filipin staining. Left, filipin staining (blue); SREBP2 staining (green); propidium iodide (red). Scale bar, 75 μm. Intensity of cytoplasmic and nuclear staining of SREBP2 was quantified using Nikon NIS-Elements software. D, The schematic diagram showing the workflow of genetic profiling and pathway analysis between CD133+ cells and CD133 from liver regeneration and HCC mouse models. CD133+ cells and CD133 (CD45/TER119) liver cells were isolated from mouse tissues from 0.1%DDC diet-treated mice and two HCC models including NRAS/AKT- and DEN/CCL4-induced HCC models by FACS approach. GSEA pathway analysis was performed to compare CD133+/CD133 in regenerating liver with two HCC models. E, When compared with regenerating livers, we found that cholesterol homeostasis was most enriched in CD133+ liver cells from two HCC models with a normalized enrichment score of 1.61 (FDR q value of 0.048). *, P < 0.05; ***, P < 0.001.

Close modal

SREBP2-mediated cholesterol biosynthesis regulates liver CSCs and bears clinical significance

To further examine whether SREBP2-mediated cholesterol biosynthesis functionally drives self-renewal and tumor formation, we performed SREBP2-knockdown and SREBP2-overexpression experiments using a lentiviral-based CRISPR activation and shRNA knockdown approaches. By Western blotting, we found high expression of SREBP2 in a panel of HCC cell lines, including PLC/PRF/5 and MHCC-97L cells (Supplementary Fig. S2A), with higher intracellular cholesterol levels in PLC/PRF/5 and MHCC-97L cells, when com-pared with those in Hep3B cells (Supplementary Fig. S2B). Therefore, these two cell lines were chosen for the knockdown experiment, whereas Hep3B cells were chosen for the overexpression experiment. Upon successful alteration of SREBP2 expression in HCC cells (Fig. 3A), we observed corresponding changes in the intracellular cholesterol levels and genes responsible for cholesterol biosynthesis in Hep3B, PLC/PRF/5, and MHCC-97L cells (Fig. 3B and C). SREBP2-mediated cholesterol biosynthesis was found to regulate self-renewal (Fig. 3D), tumorigenicity (Fig. 3E; Supplementary Tables S3A–S3C), expression of liver CSC markers (Fig. 3F), and invasiveness of HCC cells (Fig. 3G). To further confirm the role of cholesterol biosynthesis as the downstream effector of SREBP2 in mediating CSC function, we treated SREBP2 knockdown HCC cells with 5 μmol/L cholesterol solubilized in methyl-β-cyclodextrin (MβCD). Addition of exogenous cholesterol recovered the inhibitory effects of SREBP2 knockdown on stem cell frequency, CSC marker expression, cell invasiveness, and drug resistance of MHCC-97 L cells (Supplementary Figs. S3A–S3D). Analysis of publicly available dataset, GSE14520 and The Cancer Genome Atlas (TCGA) dataset revealed that SREBP2 mRNA was significantly upregulated in HCC tumor tissues compared with nontumor liver tissues (Fig. 3H). In a tissue microarray consisting of 50 HCC samples and the corresponding matched nontumor liver tissue samples, patients with high SREBP2 expression had shorter disease-free survival (P = 0.0002; log-rank test; Fig. 3I and J). In addition, we found that patients with high SREBP2 expression had a higher chance of HCC recurrence (P = 0.0006; χ2 test; Fig. 3J). Finally, we examined the role of endogenous cholesterol utilization in the regulation of liver CSCs. We first labeled PLC/PRF/5 cells with fluorescent BODIPY-conjugated-cholesterol and separated them into two populations based on their fluorescence intensities by a FACS approach (Supplementary Figs. S4A and S4B). PLC/PRF/5 cells with high uptake of BODIPY-cholesterol (BCHOLHigh) indicate high cholesterol utilization, whereas cells with low uptake of BODIPY-cholesterol (BCHOLLow) indicate low cholesterol utilization. Using in vitro and in vivo functional assays, BCHOLHigh PLC/PRF/5 cells showed enhanced stem cell frequency and in vivo tumorigenicity compared with their BCHOLLow counterparts (Supplementary Figs. S4C and S4D; Supplementary Table S3D).

Figure 3.

The crucial role of SREBP2-mediated cholesterol biosynthesis in regulation of liver CSCs. A, SREBP2 protein levels of NTC, shSREBP2 (#66 & #68) in PLC/PRF/5 and MHCC-97L cells and control (CTL), sgSREBP2 (#03 & #04) subclones derived from Hep3B were determined by Western blotting. B and C, Total cholesterol levels and genes responsible for cholesterol biosynthesis were decreased in shSREBP2 cells, whereas opposite effects were observed in SREBP2-overexpressing Hep3B cells. D,In vitro limiting dilution sphere analysis showed the role of SREBP2 in regulation of self-renewal ability. E, Left, knockdown of SREBP2 in PLC/PRF/5 and MHCC-97L cell lines suppressed tumorigenicity compared with that in NTC cells. Representative photos showing the injection of 500 and 1,000 cells derived from PLC/PRF/5 and MHCC-97L cells. Right, overexpression of SREBP2 led to increased tumorigenicity of Hep3B cells. Representative photographs showing the injection of 5 × 105 and 1 × 106 cells. Scale bar, 1 cm. F, Expression of liver CSC markers including CD47 and CD133 was measured by flow cytometry analysis. G, The migration and invasive abilities of HCC cell was evaluated by uncoated (top) and Matrigel-coated Transwell (bottom) assays, respectively. Scale bar, 250 μm. H,SREBP2 mRNA was significantly upregulated in HCC tumor tissues as compared with nontumor liver tissues in both GSE14520 and TCGA datasets. (GSE14520 dataset nontumor, n = 239; tumor, n = 247; TCGA dataset paired cases, n = 50). I, A tissue microarray consisting of 50 tumor tissues and corresponding nontumor liver tissues was subjected to IHC analysis. Case 49 showed low expression of SREBP2, while case 13 showed high expression of these proteins. Scale bar, 50 and 200 μm. J, Patients with high SREBP2 expression (n = 28) would have shorter disease-free survival than those with low expression (n = 22; P = 0.0002; log-rank test). Patients with high SREBP2 expression were significantly correlated with HCC recurrence (P = 0.0006; χ2 test). Error bars, mean ± SD. n = 3–7; *, P < 0.05; **, P < 0.01; ***, P < 0.001 from Student t test.

Figure 3.

The crucial role of SREBP2-mediated cholesterol biosynthesis in regulation of liver CSCs. A, SREBP2 protein levels of NTC, shSREBP2 (#66 & #68) in PLC/PRF/5 and MHCC-97L cells and control (CTL), sgSREBP2 (#03 & #04) subclones derived from Hep3B were determined by Western blotting. B and C, Total cholesterol levels and genes responsible for cholesterol biosynthesis were decreased in shSREBP2 cells, whereas opposite effects were observed in SREBP2-overexpressing Hep3B cells. D,In vitro limiting dilution sphere analysis showed the role of SREBP2 in regulation of self-renewal ability. E, Left, knockdown of SREBP2 in PLC/PRF/5 and MHCC-97L cell lines suppressed tumorigenicity compared with that in NTC cells. Representative photos showing the injection of 500 and 1,000 cells derived from PLC/PRF/5 and MHCC-97L cells. Right, overexpression of SREBP2 led to increased tumorigenicity of Hep3B cells. Representative photographs showing the injection of 5 × 105 and 1 × 106 cells. Scale bar, 1 cm. F, Expression of liver CSC markers including CD47 and CD133 was measured by flow cytometry analysis. G, The migration and invasive abilities of HCC cell was evaluated by uncoated (top) and Matrigel-coated Transwell (bottom) assays, respectively. Scale bar, 250 μm. H,SREBP2 mRNA was significantly upregulated in HCC tumor tissues as compared with nontumor liver tissues in both GSE14520 and TCGA datasets. (GSE14520 dataset nontumor, n = 239; tumor, n = 247; TCGA dataset paired cases, n = 50). I, A tissue microarray consisting of 50 tumor tissues and corresponding nontumor liver tissues was subjected to IHC analysis. Case 49 showed low expression of SREBP2, while case 13 showed high expression of these proteins. Scale bar, 50 and 200 μm. J, Patients with high SREBP2 expression (n = 28) would have shorter disease-free survival than those with low expression (n = 22; P = 0.0002; log-rank test). Patients with high SREBP2 expression were significantly correlated with HCC recurrence (P = 0.0006; χ2 test). Error bars, mean ± SD. n = 3–7; *, P < 0.05; **, P < 0.01; ***, P < 0.001 from Student t test.

Close modal

SREBP2-mediated cholesterol biosynthesis is a critical determinant of drug resistance in HCC cells

Consistent with the observation in drug-resistant HCC PDTXs, we found that sorafenib- and lenvatinib-resistant HCC cells derived from PLC/PRF/5 and MHCC-97L cells showed increased SREBP2 nuclear expression and cholesterol deposition (Supplementary Fig. S5). To further confirm that increased cholesterol deposition is due to de novo synthesis, we have performed stable isotope labeling of sorafenib-resistant MHCC-97L cells with 13C-glucose and compared the cholesterol 13C atom-percent excess (APE) with control mock cells at two different time points (24- and 72-hour incubations). Upon analysis at both time points, higher APE was observed in sorafenib-resistant MHCC-97L cells (Supplementary Fig. S6), which indicates that increased cholesterol deposition in drug-resistant HCC cells is at least in part via de novo cholesterol synthesis. Suppression of SREBP2 resulted in the sensitization of HCC cells to molecularly targeted drugs, including sorafenib and lenvatinib, whereas overexpression conferred drug resistance in Hep3B cells (Fig. 4A). Apart from stable knockdown of SREBP2, drug sensitization effect was consistently observed in sorafenib-resistant PLC/PRF/5 and MHCC-97L cells treated with siRNA against SREBP2 (Supplementary Fig. S7). To further understand the molecular mechanism for this acquired drug resistance, we treated HCC cells with sorafenib/lenvatinib at short time points. Upon treatment with the drugs up till 8 hours, the full-length form of SREBP2 was decreased, whereas the active nuclear-translocated form of SREBP2 was increased compared with mock control, which was further confirmed by IF analyses (Fig. 4B and C). To examine whether increased SREBP2 nuclear translocation is not due to uptake of sterols and lipids, we further examined the effect of sorafenib on SREBP2 nuclear translocation and cholesterol deposition in HCC cells supplemented with lipoprotein-deficient serum (LPDS). Similar to 10% FBS, SREBP2 activation and cholesterol deposition were also observed in HCC cells supplemented with LPDS (Supplementary Fig. S8), which further supports enhanced de novo cholesterol biosynthesis in drug-resistant HCC cells. SREBP2 was previously found to be activated by CASP3 via cleavage of premature form of SREBP2 (13, 17). Consistently, we found that CASP3 activity was elevated in sorafenib- and lenvatinib-resistant HCC cells (Supplementary Fig. S9A). This observation was also observed in HCC cells treated with sorafenib/lenvatinib (Fig. 4D). Next, we examined the effect of CASP3 activity suppression on nuclear translocation of SREBP2 in response to sorafenib/lenvatinib by treatment with Z-DEVD-FMK. We found that the nuclear translocation of SREBP2 was suppressed in sorafenib- and lenvatinib-treated HCC cells upon inhibition of CASP3 activity by Western blot analysis (Fig. 4E), which was accompanied with decrease in filipin deposition (Fig. 4F). Strikingly, HCC cells treated with Z-DEVD-FMK showed increased cell apoptosis after sorafenib/lenvatinib treatments (Fig. 4G). Using publicly available HCC datasets, we showed the crucial role of CASP3/SREBP2 signaling axis in HCC clinical samples, with evidence by increase in CASP3 expression, and a positive correlation with SREBP2 expression (Supplementary Figs. S9B and S9C). Furthermore, we observed a co-expression between CASP3 activity and SREBP2 expression in a cohort of 50 patients with HCC by IHC analysis (P = 0.0132; Supplementary Figs. S9D and S9E). To explore whether CASP3 is uniquely responsible for SREBP2 activity, we overexpressed mutant form of SREBP2 (SREBP2-D468A) with impaired CASP3 cleavage site and its wild-type form (SREBP2-WT) in Hep3B cells and examined the effect on SREBP2 expression and apoptosis with or without the presence of Z-DEVD-FMK (Supplementary Figs. S10A and S10B). When compared with SREBP2-WT, apoptotic enhancing effect of Z-DEVD-FMK was abolished in SREBP2-D468A transfected Hep3B cells. In addition, SREBP2 activation was suppressed in SREBP2-WT transfected Hep3B cells upon co-treatment of sorafenib with Z-DEVD-FMK; whereas no such phenomenon was observed in SREBP2-D468A transfectants (Supplementary Figs. S10C and S10D). These result show that SREBP2-mediated sorafenib resistance is dependent on CASP3 activity. Finally, we have examined the clinical relevance of SREBP2 expression in patients with HCC treated with sorafenib. In a tissue microarray consisting of 91 HCC samples from patients who had been treated with sorafenib, patients with high SREBP2 expression had shorter disease-free survival (P < 0.0001, log-rank test), and a higher chance of HCC recurrence after sorafenib treatment (P = 0.0002; χ2 test; Fig. 4H and I).

Figure 4.

Caspase-3-induced SREBP2 activation drives drug resistance of HCC cells. A, The apoptosis of shSREBP2 (PLC/PRF/5 and MHCC-97L) and sgSREBP2 (Hep3B) induced by either sorafenib (10 or 15 μmol/L) or lenvatinib (40 μmol/L) was evaluated by Annexin V staining. B and C, Expression of premature and mature SREBP2 in response to 2, 4, and 8 hours of sorafenib (20–30 μmol/L)/lenvatinib (30–40 μmol/L) treatment was evaluated by Western blot analysis. Increase in nuclear SREBP2 (indicated by arrows) was demonstrated in PLC/PRF/5 and MHCC-97L cells in response to 2, 4, and 8 hours of sorafenib/lenvatinib treatment. SREBP2 staining, green; DAPI staining, blue. Scale bar, 10 μm. D, Increase in CASP3 activity was observed in PLC/PRF/5 and MHCC-97L cells with treatment of sorafenib/lenvatinib. E, Expression of premature and mature SREBP2 in MHCC-97L cells treated with either sorafenib (30 μmol/L)/lenvatinib (40 μmol/L) or Z-DEVD-FMK (200 μmol/L) or in combination was evaluated by Western blot analysis. F, Decrease in filipin staining was observed in sorafenib/lenvatinib in combination with Z-DEVD-FMK. Filipin staining, blue; propidium iodide, red. Scale bar, 10 μm. Intensity of filipin staining was quantified using Nikon NIS-Elements software. G, The apoptosis of MHCC-97L cells treated with either sorafenib (30 μmol/L) /lenvatinib (40 μmol/L) or Z-DEVD-FMK (200 μmol/L) or in combination was evaluated by Annexin-V staining. H, A tissue microarray consisting of 91 tumor tissues with posttreatment of sorafenib was subjected to IHC analysis. Case 19 showed low expression of SREBP2, whereas case 68 showed high expression of this protein. Scale bar, 50 and 200 μm. I, Patients with high SREBP2 expression (n = 54) would have shorter disease-free survival than those with low expression (n = 37; P < 0.0001; log-rank test). Patients with high SREBP2 expression were significantly correlated with HCC recurrence (P = 0.0002; χ2 test). Error bars, mean ± SD. n = 3–5; *, P < 0.05; **, P < 0.01; ***, P < 0.001 from Student t test.

Figure 4.

Caspase-3-induced SREBP2 activation drives drug resistance of HCC cells. A, The apoptosis of shSREBP2 (PLC/PRF/5 and MHCC-97L) and sgSREBP2 (Hep3B) induced by either sorafenib (10 or 15 μmol/L) or lenvatinib (40 μmol/L) was evaluated by Annexin V staining. B and C, Expression of premature and mature SREBP2 in response to 2, 4, and 8 hours of sorafenib (20–30 μmol/L)/lenvatinib (30–40 μmol/L) treatment was evaluated by Western blot analysis. Increase in nuclear SREBP2 (indicated by arrows) was demonstrated in PLC/PRF/5 and MHCC-97L cells in response to 2, 4, and 8 hours of sorafenib/lenvatinib treatment. SREBP2 staining, green; DAPI staining, blue. Scale bar, 10 μm. D, Increase in CASP3 activity was observed in PLC/PRF/5 and MHCC-97L cells with treatment of sorafenib/lenvatinib. E, Expression of premature and mature SREBP2 in MHCC-97L cells treated with either sorafenib (30 μmol/L)/lenvatinib (40 μmol/L) or Z-DEVD-FMK (200 μmol/L) or in combination was evaluated by Western blot analysis. F, Decrease in filipin staining was observed in sorafenib/lenvatinib in combination with Z-DEVD-FMK. Filipin staining, blue; propidium iodide, red. Scale bar, 10 μm. Intensity of filipin staining was quantified using Nikon NIS-Elements software. G, The apoptosis of MHCC-97L cells treated with either sorafenib (30 μmol/L) /lenvatinib (40 μmol/L) or Z-DEVD-FMK (200 μmol/L) or in combination was evaluated by Annexin-V staining. H, A tissue microarray consisting of 91 tumor tissues with posttreatment of sorafenib was subjected to IHC analysis. Case 19 showed low expression of SREBP2, whereas case 68 showed high expression of this protein. Scale bar, 50 and 200 μm. I, Patients with high SREBP2 expression (n = 54) would have shorter disease-free survival than those with low expression (n = 37; P < 0.0001; log-rank test). Patients with high SREBP2 expression were significantly correlated with HCC recurrence (P = 0.0002; χ2 test). Error bars, mean ± SD. n = 3–5; *, P < 0.05; **, P < 0.01; ***, P < 0.001 from Student t test.

Close modal

SHH signaling cascade is the downstream effector of SREBP2-mediated CSC functions and drug resistance

To determine the major downstream effector of SREBP2, we employed RNA sequencing analysis to compare the gene expression profiles of shSREBP2 (#66) cells and nontarget control cells derived from MHCC-97L cells. GSEA showed enrichment of the SHH (Fig. 5A). By Western blot analysis, we found that expression of GLI-1, SUFU, SHH, PTCH1, and PTCH2 consistently altered in SREBP2 knockdown and overexpressing HCC cells (Fig. 5B).

Figure 5.

SREBP2-mediated cholesterol biosynthesis drives activation of SHH signaling pathway via 25-OHC. A, RNA sequencing was employed to compare the genetic profiles between shSREBP2 (#66) and NTC cells derived from MHCC-97L cells. GSEA analysis showed that SHH signaling was downregulated in shSREBP2 cells with a normalized enrichment score of −1.603 (FDR q value of 0.014). B, Western blot analysis validation of the key proteins involved in SHH signaling pathways including GLI-1, SUFU, SHH, PTCH1, and PTCH2. C, CH25H levels of shSREBP2 (#66 and #68) relative to NTC in PLC/PRF/5 and MHCC-97L cells and cholesterol-treated (Chol) at 5 μmol/L relative to methyl-β-cyclodextrin (MβCD) at 4 and 8 hours of treatments in Hep3B were evaluated by ELISA assay. D, Treatment of 25-OHC at 15 μmol/L for 48 hours mitigated the suppressive effects of shSREBP2 cells on expression of GLI-1, SUFU, SHH, PTCH1, and PTCH2. E, The addition of GANT61 at a dose of 3.125 μmol/L offset the enhancing effect of SREBP2 overexpression on the self-renewal ability of sgSREBP2 (#03) Hep3B cells. F, The addition of GANT61 at a dose of 5 μmol/L recovered the percentage of apoptotic sgSREBP2 (#03) Hep3B cells in response to 24 and 48 hours of sorafenib/lenvatinib treatment. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 from Student t test.

Figure 5.

SREBP2-mediated cholesterol biosynthesis drives activation of SHH signaling pathway via 25-OHC. A, RNA sequencing was employed to compare the genetic profiles between shSREBP2 (#66) and NTC cells derived from MHCC-97L cells. GSEA analysis showed that SHH signaling was downregulated in shSREBP2 cells with a normalized enrichment score of −1.603 (FDR q value of 0.014). B, Western blot analysis validation of the key proteins involved in SHH signaling pathways including GLI-1, SUFU, SHH, PTCH1, and PTCH2. C, CH25H levels of shSREBP2 (#66 and #68) relative to NTC in PLC/PRF/5 and MHCC-97L cells and cholesterol-treated (Chol) at 5 μmol/L relative to methyl-β-cyclodextrin (MβCD) at 4 and 8 hours of treatments in Hep3B were evaluated by ELISA assay. D, Treatment of 25-OHC at 15 μmol/L for 48 hours mitigated the suppressive effects of shSREBP2 cells on expression of GLI-1, SUFU, SHH, PTCH1, and PTCH2. E, The addition of GANT61 at a dose of 3.125 μmol/L offset the enhancing effect of SREBP2 overexpression on the self-renewal ability of sgSREBP2 (#03) Hep3B cells. F, The addition of GANT61 at a dose of 5 μmol/L recovered the percentage of apoptotic sgSREBP2 (#03) Hep3B cells in response to 24 and 48 hours of sorafenib/lenvatinib treatment. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 from Student t test.

Close modal

By IHC analysis, cleaved CASP3 was found to co-express with SREBP2 and GLI-1 (Supplementary Figs. S9D and S9E). In addition, GLI-1 was upregulated in sorafenib- and lenvatinib-resistant HCC cells, further confirming the role of SHH signaling as the downstream effector of CASP3/SREBP2 signaling in mediation of drug resistance (Supplementary Fig. S9F). Because intracellular cholesterol was previously reported to indirectly activate SHH pathway via its metabolite oxysterols (18), we examined whether SREBP2-mediated cholesterol synthesis leads to the activation of the SHH pathway by cholesterol metabolites. By ELISA, we found that cholesterol-25-hydroxylase (CH25H), an enzyme responsible for production of 25-hydroxycholesterol (25-OHC) was downregulated in shSREBP2 cells (Fig. 5C). Conversely, exogenous administration of cholesterol led to an increase in CH25H levels (Fig. 5C). Finally, administration of 25-OHC at 15 μmol/L in MHCC-97L cells rescued the suppression of the SHH pathway in shSREBP2 HCC cells (Fig. 5D). Finally, to further confirm the role of SHH as the downstream effector of CASP3/SREBP2 in mediating cancer stemness and drug resistance, we treated SREBP2-overexpressing Hep3B cells with GANT61, GLI-1 inhibitor, to investigate whether the effects of SREBP2 overexpression can be eliminated by suppression of GLI-1. We found that the addition of GANT61 offset the enhancing effect of SREBP2 overexpression on self-renewal ability as well as resistance to sorafenib/lenvatinib treatment (Fig. 5E and F).

Exogenous cholesterol drives cancer stemness, drug resistance, and HCC organoid expansion

De novo cholesterol synthesis and circulating cholesterol are two major sources of cholesterol in the body. Therefore, we examined whether exogenous cholesterol also exerts similar effects on cancer stemness and drug resistance. Administration of 5 μmol/L cholesterol led to an increase in cholesterol deposition in cholesterol-low Hep3B cells (∼1μmol/L intracellular cholesterol level derived from Supplementary Fig. S2B), according to filipin staining (Fig. 6A). By Western blot analysis, we found that expression of GLI-1, SUFU, SHH, PTCH1, and PTCH2 was increased in cholesterol-treated Hep3B cells (Fig. 6B). Interestingly, we found that 5 μmol/L cholesterol increased the liver CSC marker expression, augmented drug resistance and self-renewal abilities in Hep3B cells (Fig. 6CE). Using a CellTiter-Glo assay, we found that cholesterol administration at various doses (15 and 30 μmol/L) increased the size of organoids and their proliferation rate in a dose-dependent manner (Fig. 6F), suggesting that exogenous cholesterol acts as a mitogen for HCC growth. Recent report showed that exogenous uptake of cholesterol promotes NAFLD-induced HCC development (19). In mice fed with high fat with high cholesterol (HFHC) diets for 8 months (19), increased nuclear SREBP2 expression was observed in these livers, when compared with mice with normal diet (Supplementary Figs. S11A and S11B). Consistently, SREBP2 mRNA level was significantly upregulated in NAFLD-induced HCC clinical samples, when compared with normal counterparts (Supplementary Fig. S11C). In a cohort of 27 NAFLD-induced patients with HCC, we found that patients with high cholesterol levels showed a trend of poor survival (P = 0.1357, log-rank test; Fig. 6G).

Figure 6.

Exogenous cholesterol enhances CSC properties and expansion of HCC patient-derived organoid cultures. A, Administration of cholesterol at 5 μmol/L for 4 hours increased the cholesterol deposition in Hep3B compared with the cells treated with MβCD only. Top, filipin staining. Bottom, nuclei were counterstained with propidium iodide (PI). Scale bar, 25 μm. Intensity of filipin staining was quantified using Nikon NIS-Elements software. B, Western blot analysis of GLI-1, SUFU, SHH, PTCH1, and PTCH2 in Hep3B cells treated with cholesterol at 5 μmol/L. C, The effect of exogenous cholesterol on CD47 and CD24 was evaluated by flow cytometry analysis. D, The apoptosis of Hep3B cells induced by either sorafenib (15 μmol/L) or lenvatinib (40 μmol/L) was significantly offset by exogenous administration of cholesterol for 48 hours. E,In vitro limiting dilution sphere analysis showed the enhancing effect of cholesterol in self-renewal ability. F, The growth of two HCC patient-derived organoids (HK-HCC P1 and HK-HCC P2) was significantly enhanced upon administration of cholesterol at 15 and 30 μmol/L for 6 days using Cell-Titer-Glo assay. Representative images of organoids treated with MβCD (0 μmol/L) and cholesterol at 15 and 30 μmol/L on day 6 are shown. Scale bar, 100 μm. G, NAFLD-induced HCC Patients with high cholesterol level (n = 14) would have shorter disease-free survival than those with lower level (n = 13; P = 0.1357; cut-off by mean = 349.5 mg/dL). Error bars, mean ± SD. n = 3–5; *, P < 0.05; **, P < 0.01; ****, P < 0.0001 from Student t test.

Figure 6.

Exogenous cholesterol enhances CSC properties and expansion of HCC patient-derived organoid cultures. A, Administration of cholesterol at 5 μmol/L for 4 hours increased the cholesterol deposition in Hep3B compared with the cells treated with MβCD only. Top, filipin staining. Bottom, nuclei were counterstained with propidium iodide (PI). Scale bar, 25 μm. Intensity of filipin staining was quantified using Nikon NIS-Elements software. B, Western blot analysis of GLI-1, SUFU, SHH, PTCH1, and PTCH2 in Hep3B cells treated with cholesterol at 5 μmol/L. C, The effect of exogenous cholesterol on CD47 and CD24 was evaluated by flow cytometry analysis. D, The apoptosis of Hep3B cells induced by either sorafenib (15 μmol/L) or lenvatinib (40 μmol/L) was significantly offset by exogenous administration of cholesterol for 48 hours. E,In vitro limiting dilution sphere analysis showed the enhancing effect of cholesterol in self-renewal ability. F, The growth of two HCC patient-derived organoids (HK-HCC P1 and HK-HCC P2) was significantly enhanced upon administration of cholesterol at 15 and 30 μmol/L for 6 days using Cell-Titer-Glo assay. Representative images of organoids treated with MβCD (0 μmol/L) and cholesterol at 15 and 30 μmol/L on day 6 are shown. Scale bar, 100 μm. G, NAFLD-induced HCC Patients with high cholesterol level (n = 14) would have shorter disease-free survival than those with lower level (n = 13; P = 0.1357; cut-off by mean = 349.5 mg/dL). Error bars, mean ± SD. n = 3–5; *, P < 0.05; **, P < 0.01; ****, P < 0.0001 from Student t test.

Close modal

Simvastatin enhances the effect of drug treatment in organotypic ex vivo human HCC clinical samples and HCC PDTX models

By MTT assay, we found that simvastatin, an FDA-approved cholesterol-lowering drug, suppressed the growth of PLC/PRF/5 cells in a dose-dependent manner, with IC50 values of 34.3 μmol/L (Supplementary Fig. S12A). Simvastatin suppressed liver CSC properties (Supplementary Figs. S12B–S12D). These data prompted us to examine the combined effect of simvastatin with sorafenib/lenvatinib in the treatment of HCC. Using a CellTiter-Glo assay, we found that combination treatment resulted in the most significant reduction in the viability of HCC patient-derived organoids (HK-HCC P1) and that simvastatin treatment sensitized HCC cells to sorafenib/lenvatinib (Fig. 7A). Next, we examined the therapeutic effect of simvastatin alone at a dose of 4 mg/kg (an equivalent dose to 40 mg for the treatment of high cholesterol patients) and its combined effect with sorafenib in vivo using HCC xenografts derived from PY003. The tumors and their corresponding volumes after treatment for 21 days are shown in Fig. 7B to D. Simvastatin reduced the tumor volumes in a manner similar to that of sorafenib. In addition, simvastatin combined with sorafenib exerted maximal suppression of tumor growth compared with that of the control group. Strikingly, we found that this combination treatment markedly reduced the tumor volumes of PY003 by 37% relative to the original tumor volume on day 0 (Fig. 7D). During this experiment, no significant loss of body weight was observed in the animals undergoing combination treatment (Supplementary Fig. S12E). We have also evaluated the effects of these drug combinations in sorafenib-resistant PDTX#1 cells (6). Similarly, we found that simvastatin/sorafenib exerted the greatest tumor suppressive effects when compared with the single-agent treatment and mock controls (Fig. 7BD). By filipin staining, simvastatin alleviated cholesterol deposition in sorafenib-treated tumors, which is accompanied with decrease in SREBP2 and GLI-1 expression (Fig. 7E; Supplementary Fig. S13).

Figure 7.

The effect of Simvastatin and TKI treatment in suppressing tumor growth in HCC organoid culture and PDTX xenografts. A, Top, representative photos of HCC patient-derived organoid culture (HK-HCC P1), which was treated with DMSO, sorafenib (8 μmol/L)/lenvatinib (40 μmol/L), or combo for 6 days. Scale bar, 200 μm. Bottom, relative cell viability to mock (fold change) is presented. *, P < 0.05; **, P < 0.01 from Student t test. B, Response of PY003 and sorafenib-resistant PDTX#1 xenografts to treatment with simvastatin (4 mg/kg), sorafenib (30 mg/kg), or both drugs (simvastatin 4 mg/kg; sorafenib 30 mg/kg). The representative five tumors at the end of the treatment is shown. Scale bar, 1 cm. C, Graph showing the weight of tumors at the end of the treatment. D, Waterfall plot showing the response of each tumor after 21 days. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 from Mann–Whitney U test. E, Representative filipin staining of tumors from mock, sorafenib, simvastatin, and combo. Scale bar, 25 μm. Top, filipin staining (blue); middle, propidium iodide (red); bottom, merge. Intensity of filipin staining was quantified using Nikon NIS-Elements software. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 from Student t test. Error bars indicate mean ± SD (n = 3–10).

Figure 7.

The effect of Simvastatin and TKI treatment in suppressing tumor growth in HCC organoid culture and PDTX xenografts. A, Top, representative photos of HCC patient-derived organoid culture (HK-HCC P1), which was treated with DMSO, sorafenib (8 μmol/L)/lenvatinib (40 μmol/L), or combo for 6 days. Scale bar, 200 μm. Bottom, relative cell viability to mock (fold change) is presented. *, P < 0.05; **, P < 0.01 from Student t test. B, Response of PY003 and sorafenib-resistant PDTX#1 xenografts to treatment with simvastatin (4 mg/kg), sorafenib (30 mg/kg), or both drugs (simvastatin 4 mg/kg; sorafenib 30 mg/kg). The representative five tumors at the end of the treatment is shown. Scale bar, 1 cm. C, Graph showing the weight of tumors at the end of the treatment. D, Waterfall plot showing the response of each tumor after 21 days. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 from Mann–Whitney U test. E, Representative filipin staining of tumors from mock, sorafenib, simvastatin, and combo. Scale bar, 25 μm. Top, filipin staining (blue); middle, propidium iodide (red); bottom, merge. Intensity of filipin staining was quantified using Nikon NIS-Elements software. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 from Student t test. Error bars indicate mean ± SD (n = 3–10).

Close modal

The role of cholesterol dysregulation in cancer remains controversial. Certain cholesterol metabolites, such as dendrogenin A, were found to suppress the growth of melanoma and acute myeloid leukemia (20). In contrast, cholesterol metabolism alterations have been shown to lead to the formation of oncometabolites that support tumor growth in breast and prostate cancers (21). Specifically, in HCC, cholesterol and cholesterol ester were found to play a role in NAFLD-induced HCC formation (11, 22–23). These findings showed that the role of cholesterol in cancer may be both cancer-type and context-specific. In this study, we found that cholesterol biosynthesis is critical in the regulation of drug resistance using our established drug-resistant HCC xenograft models. In addition, we found that this process is crucial in the regulation of liver CSCs originating from mice and humans. This result is consistent with previous findings showing the role of cholesterol in the proliferation of hematopoietic stem cells and intestinal stem cells (24–26). Furthermore, we found that the role of cholesterol biosynthesis in drug resistance and cancer stemness is mediated by SREBP2.

SREBP2 is the master regulator of cholesterol homeostasis, which is synthesized as a precursor and preferentially regulates genes important for cholesterol synthesis, including HMG-CoA synthase, HMG-CoA reductase, LDL receptor (LDLR), and proprotein convertase subtilisin kexin type 9 (PCSK9). SREBP2 was found to be overexpressed in colon and prostate cancers (27, 28), and its overexpression is related to cell proliferation and drug resistance (29, 30). Recent report showed that SREBP2 induces transferrin in circulating melanoma cells and suppresses ferroptosis, leading to tumor metastasis (31). Using CRISPR activation and knockdown approaches, we found that SREBP2-mediated cholesterol biosynthesis regulates cancer stemness and drug sensitivity to sorafenib/lenvatinib treatment. Consistent with the in vivo observations, sorafenib- and lenvatinib-resistant HCC cells showed increased SREBP2 nuclear translocation and cholesterol deposition. Although recent publications showed increased lipid uptake/synthesis in TKI-resistant cells due to increased SREBP1 activity (32), we did not observe increase in SREBP1 staining in sorafenib- and lenvatinib-resistant HCC PDTXs. Nonetheless, we cannot exclude the possibility that TKI-resistant HCC cells also uptake neutral lipids and contribute to CSC and drug-resistant phenotypes. Next, we examined the regulatory mechanism of SREBP2 translocation. Previous reports showed that SREBP2 can be activated by CASP3 (13, 17), ER stress (33), and p53 deletion (34). In this study, we found that CASP3 plays an important role in regulating the translocation of SREBP2 in the process of acquired drug resistance development, as evidenced by the lack of apoptotic enhancing effect in SREBP2-D468A transfected HCC cells upon CASP3 inhibition. Further confirmation was evidenced by the observation that betulin does not prevent SREBP2 processing in sorafenib-treated cells (data not shown). Collectively, CASP3-driven activation of SREBP2 is a dominant mechanism that drives resistance to TKI treatment. Our current findings are in coherence with the recent report showing the role of CASP3 in driving the therapeutic resistance in HCC (35). We further showed that SREBP2-mediated cholesterol biosynthesis is a potential biomarker for the prediction of the sorafenib response in patients with HCC in a cohort of 91 post-sorafenib treatment patient tumor specimens. Previous reports also echoed this observation and showed the effects of free cholesterol and cholesterol sensor in driving sorafenib resistance in HCC (36, 37).

To determine the major downstream target of SREBP2 in mediating drug resistance, we compared the genetic profiles between SREBP2 knockdown cells and their control counterparts. GSEA showed enrichment of the SHH pathway, which was further confirmed by Western blot analysis, revealing alterations in the key proteins in this pathway upon SREBP2 overexpression and downregulation. This result is consistent with other studies showing the involvement of cholesterol in the activation of the SHH pathway either directly (38, 39) through the activation of smoothened (SMO) activity or indirectly via its metabolite oxysterols, such as 25-OHC (18). We found that SREBP2 regulates liver CSC functions via regulation of the SHH pathway, at least in part through 25-OHC regulation, as evidenced by the effect of 25-OHC in negating the suppressive effect of the SHH pathway in shSREBP2 HCC cells. 25-OHC was previously reported to promote HCC metastasis through upregulation of TLR4/FABP4 signaling cascade (40). In addition to 25-OHC, other oxysterols, such as 20-OHC and 24-OHC, were also reported to activate the SHH pathway (38). Besides, sorafenib was reported to inhibit the activity of the XCT transporter and can block the synthesis of glutathione peroxidase 4 (GPX4). Resistance to sorafenib was associated with the induction of FSP-mediated ferroptosis or engagement of other antioxidant pathways (41). However, our data showed that sorafenib exerts no obvious effect on XCT and GPX4 expression (Supplementary Fig. S14). In addition, sorafenib-resistant HCC cells were resistant to erastin treatment, suggesting that the redox balance associated with inhibition of XCT do not play a significant role in our cell-based drug-resistant models.

Simvastatin is an FDA-approved cholesterol-lowering drug used to prevent cardiovascular complications by blocking HMG-CoA reductase, a rate-limiting enzyme for cholesterol biosynthesis. Thus far, simvastatin has been proven to play a preventive role in HCC development (42). It would be crucial to further examine the effect of simvastatin on drug sensitization by inhibiting liver CSC populations. First, we found that simvastatin suppressed liver CSC properties. Using HCC patient-derived organoids, we found that simvastatin sensitized HCC cells to sorafenib/lenvatinib and suppressed tumor growth. Notably, the dose of simvastatin used in these in vitro studies is significantly above its IC50 for HMGCR, and thus the off-target effects of this drug cannot be underestimated. Using PDTX model, we examined the therapeutic efficacy of simvastatin in combination with sorafenib by oral gavage. Simvastatin/sorafenib treatment markedly reduced the tumor volume of PY003 by 37% compared with the original size at day 0. Interestingly, the cotreatment group also exhibited a maximal reduction in tumor volume in sorafenib-resistant PDTX#1. This data, together with the recent preclinical data (43), suggest that sorafenib/lenvatinib in combination of simvastatin may be a possible strategy for HCC treatment. However, a recent phase III clinical trial showed no additional efficacy for combination of sorafenib with pravastatin when compared with sorafenib alone (44). The discrepancy may be due to the fact that pravastatin is a hydrophilic statin with minimal post hepatic exposure while simvastatin is a hydrophobic drug. Furthermore, simvastatin has a better potency to produce cholesterol and lipid lowering effect when compared with pravastatin (45). In breast cancer, it was reported that there is local upregulation of HMGCR, allowing tumor cells to have a local (nonhepatic) source of cholesterol, that is likely only minimally impacted by the low post-hepatic circulating concentrations of statins (46). More studies are necessary to examine the efficacy of combinatory effect of sorafenib with simvastatin prior to clinical investigation.

In conclusion, we have identified the relationship of CASP3 and SREBP2 as the driving event that initiates acquired drug resistance through the promotion of the cholesterol biosynthesis-driven SHH signaling pathway (Supplementary Fig. S15), and its inhibition sensitized cells to conventional therapies, including sorafenib/lenvatinib. These findings reveal a link between SREBP2-mediated cholesterol biosynthesis and cancer stemness that modulates acquired drug resistance of cancer cells.

No author disclosures were reported.

E.H.K. Mok: Conceptualization, data curation, formal analysis, validation, methodology, writing–original draft. C.O.N. Leung: Conceptualization, data curation, formal analysis, validation, investigation, methodology, writing–original draft. L. Zhou: Conceptualization, resources, formal analysis, investigation, methodology. M.M.L. Lei: Data curation, software, formal analysis. H.W. Leung: Data curation, formal analysis, methodology. M. Tong: Conceptualization, formal analysis, investigation, methodology. T.L. Wong: Data curation, formal analysis, validation. E.Y.T. Lau: Conceptualization, resources, data curation. I.O.L. Ng: Conceptualization, resources, funding acquisition. J. Ding: Resources, data curation. J.P. Yun: Resources. J. Yu: Conceptualization, data curation, formal analysis. H.L. Zhu: Resources, data curation. C.H. Lin: Data curation, software, methodology. D. Lindholm: Conceptualization, resources. K.S. Leung: Investigation, methodology. J.D. Cybulski: Data curation, formal analysis, investigation, methodology. D.M. Baker: Conceptualization, formal analysis. S. Ma: Conceptualization, data curation, supervision, funding acquisition. T.K.W. Lee: Conceptualization, resources, data curation, formal analysis, supervision, funding acquisition, writing–original draft, project administration, writing–review and editing.

This study was supported by the RGC General Research Fund (15104119 to T.K. Lee), Collaborative Research Fund (C7026–18G), (C7050–18EF to SIRMS), Research Impact Fund (R5050–18F), and the Theme-based Research Scheme project (T12–704/16-R to T.K. Lee, S. Ma, and I.O.L. Ng). The authors thank the University Research Facility in Life Sciences at Hong Kong Polytechnic University for providing and maintaining the equipment and technical support needed for flow cytometry and imaging work as well as the Centralized Animal Facility at Hong Kong Polytechnic University and the Laboratory Animal Unit at University of Hong Kong for supporting our animal studies. The authors thank Dr. Xin Chen (University of California, San Francisco) for sharing of plasmids used for hydrodynamic tail vein injection. They also thank Centre for PanorOmic Sciences for providing RNA sequencing service and Department of Pathology at University of Hong Kong for frozen tissue sectioning service.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

1.
Ma
S
,
Chan
KW
,
Hu
L
,
Lee
TK
,
Wo
JY
,
Ng
IO
, et al
.
Identification and characterization of tumorigenic liver cancer stem/progenitor cells
.
Gastroenterology
2007
;
132
:
2542
56
.
2.
Yamashita
T
,
Budhu
A
,
Forgues
M
,
Wang
XW
.
Activation of hepatic stem cell marker EpCAM by Wnt-beta-catenin signaling in HCC
.
Cancer Res
2007
;
67
:
10831
9
.
3.
Yang
ZF
,
Ho
DW
,
Ng
MN
,
Lau
CK
,
Yu
WC
,
Ngai
P
, et al
.
Significance of CD90+ cancer stem cells in human liver cancer
.
Cancer Cell
2008
;
13
:
153
66
.
4.
Lee
TK
,
Castilho
A
,
Cheung
VC
,
Tang
KH
,
Ma
S
,
Ng
IO
.
CD24+ liver tumor-initiating cells drive self-renewal and tumor initiation through Stat3-mediated nanog regulation
.
Cell Stem Cell
2011
;
9
:
50
63
.
5.
Xin
HW
,
Ambe
CM
,
Hari
DM
,
Wiegand
GW
,
Miller
TC
,
Chen
JQ
, et al
.
Label-retaining liver cancer cells are relatively resistant to sorafenib
.
Gut
2013
;
62
:
1777
86
.
6.
Lo
J
,
Lau
EY
,
Ching
RH
,
Ma
MK
,
Cheng
BY
,
Ng
IO
, et al
.
NF-κB mediated CD47 upregulation promotes sorafenib resistance and its blockade synergizes the effect of sorafenib in hepatocellular carcinoma
.
Hepatology
2015
;
62
:
534
45
.
7.
Lee
TK
,
Guan
XY
,
Ma
S
.
Cancer stem cells in hepatocellular carcinoma: a review from origin to clinical implications
.
Nat Rev Gastroenterol Hepatol
2022
;
19
:
26
44
.
8.
Llovet
JM
,
Di Bisceglie
AM
,
Bruix
J
,
Kramer
BS
,
Lencioni
R
,
Zhu
AX
, et al
.
Design and endpoints of clinical trials in hepatocellular carcinoma
.
J Natl Cancer Inst
2008
;
100
:
698
711
.
9.
Cheng
AL
,
Kang
YK
,
Chen
Z
,
Tsao
CJ
,
Qin
S
,
Kim
JS
, et al
.
Efficacy and safety of sorafenib in patients in the Asia-pacific region with advanced hepatocellular carcinoma: a phase III randomised, double-blind, placebo-controlled trial
.
Lancet Oncol
2009
;
10
:
25
34
.
10.
Kudo
M
,
Finn
RS
,
Qin
S
,
Han
KH
,
Ikeda
K
,
Piscaglia
F
, et al
.
Lenvatinib versus sorafenib in first-line treatment of patients with unresectable hepatocellular carcinoma: a randomized phase 3 non-inferiority trial
.
Lancet
2018
;
391
:
1163
73
.
11.
Wang
Y
,
Wang
J
,
Li
X
,
Xiong
X
,
Wang
J
,
Zhou
Z
, et al
.
N1-methyladenosine methylation in tRNA drives liver tumourigenesis by regulating cholesterol metabolism
.
Nat Commun
2021
;
12
:
6314
.
12.
Gao
S
,
Soares
F
,
Wang
S
,
Wong
CC
,
Chen
H
,
Yang
Z
, et al
.
CRISPR screens identify cholesterol biosynthesis as a therapeutic target on stemness and drug resistance of colon cancer
.
Oncogene
2021
;
40
:
6601
13
.
13.
Pham
DD
,
Do
HT
,
Bruelle
C
,
Kukkonen
JP
,
Eriksson
O
,
Mogollón
I
, et al
.
p75 neurotrophin receptor signaling activates sterol regulatory element-binding Protein-2 in hepatocyte cells via p38 mitogen-activated protein kinase and caspase-3
.
J Bio Chem
2016
;
291
:
10747
58
.
14.
Leung
HW
,
Leung
CO
,
Lau
EY
,
Chung
KPS
,
Mok
EH
,
Lei
MML
, et al
.
EPHB2 activates β-catenin to enhance cancer stem cell properties and drive sorafenib resistance in hepatocellular carcinoma
.
Cancer Res
2021
;
81
:
3229
40
.
15.
Lee
TK
,
Cheung
VC
,
Lu
P
,
Lau
EY
,
Ma
S
,
Tang
KH
, et al
.
Blockade of CD47 mediated CTSS-PAR2 signaling provides a therapeutic target for hepatocellular carcinoma
.
Hepatology
2014
;
60
:
179
91
.
16.
Chan
LH
,
Zhou
L
,
Ng
KY
,
Wong
TL
,
Lee
TK
,
Sharma
R
, et al
.
Protein methyltransferasePRMT6 regulates cancer stemness through CRAF arginine methylation in hepatocellular carcinoma
.
Cell Rep
2018
;
25
:
690
701
.
17.
Wang
X
,
Zelenski
NG
,
Yang
J
,
Sakai
J
,
Brown
MS
,
Goldstein
JL
.
Cleavage of sterol regulatory element binding proteins (SREBPs) by CPP32 during apoptosis
.
EMBO J
1996
;
15
:
1012
20
.
18.
Corcoran
RB
,
Scott
MP
.
Oxysterols stimulate sonic hedgehog signal transduction and proliferation of medulloblastoma cells
.
Proc Natl Acad Sci U S A
2006
;
103
:
8408
13
.
19.
Liang
JQ
,
Teoh
N
,
Xu
L
,
Pok
S
,
Li
X
,
Chu
ESH
, et al
.
Dietary cholesterol promotes steatohepatitis related hepatocellular carcinoma through dysregulated metabolism and calcium signaling
.
Nat Commun
2018
;
9
:
4490
.
20.
Poirot
M
,
Silvente-Poirot
S
.
The tumor-suppressor cholesterol metabolite, dendrogenin A, is a new class of LXR modulator activating lethal autophagy in cancers
.
Biochem Pharmacol
2018
;
153
:
75
81
.
21.
Marwarha
G
,
Raza
S
,
Hammer
K
,
Ghribi
O
.
27-hydroxycholesterol: A novel player in molecular carcinogenesis of breast and prostate cancer
.
Chem Phys Lipids
2017
;
207
:
108
26
.
22.
Liu
D
,
Wong
CC
,
Fu
L
,
Chen
H
,
Zhao
L
,
Li
C
, et al
.
Squalene epoxidase drives NAFLD-induced hepatocellular carcinoma and is a pharmaceutic target
.
Sci Transl Med
2018
;
10
:
eaap9840
.
23.
Che
L
,
Chi
W
,
Qiao
Y
,
Zhang
J
,
Song
X
,
Liu
Y
, et al
.
Cholesterol biosynthesis supports the growth of hepatocarcinoma lesions depleted of fatty acid synthase in mice and humans
.
Gut
2020
;
69
:
177
86
.
24.
Westerterp
M
,
Gourion-Arsiquaud
S
,
Murphy
AJ
,
Shih
A
,
Cremers
S
,
Levine
RL
, et al
.
Regulation of hematopoietic stem and progenitor cell mobilization by cholesterol efflux pathways
.
Cell Stem Cell
2012
;
11
:
195
206
.
25.
Yvan-Charvet
L
,
Pagler
T
,
Gautier
EL
,
Avagyan
S
,
Siry
RL
,
Han
S
, et al
.
ATP-binding cassette transporters and HDL suppress hematopoietic stem cell proliferation
.
Science
2010
;
328
:
1689
93
.
26.
Wang
B
,
Rong
X
,
Palladino
END
,
Wang
J
,
Fogelman
AM
,
Martín
MG
, et al
.
Phospholipid remodeling and cholesterol availability regulate intestinal stemness and tumorigenesis
.
Cell Stem Cell
2018
;
22
:
206
20
.
27.
Jacobs
RJ
,
Voorneveld
PW
,
Kodach
LL
,
Hardwick
JC
.
Cholesterol metabolism and colorectal cancers
.
Curr Opin Pharmacol
2012
;
12
:
690
5
.
28.
Alfaqih
MA
,
Nelson
ER
,
Liu
W
,
Safi
R
,
Jasper
JS
,
Macias
E
, et al
.
CYP27A1 loss dysregulates cholesterol homeostasis in prostate cancer
.
Cancer Res
2017
;
77
:
1662
73
.
29.
Kondo
A
,
Yamamoto
S
,
Nakaki
R
,
Shimamura
T
,
Hamakubo
T
,
Sakai
J
, et al
.
Extracellular acidic pH activates the sterol regulatory element-binding protein 2 promote tumor progression
.
Cell Rep
2017
;
18
:
2228
42
.
30.
Zheng
L
,
Li
L
,
Lu
Y
,
Jiang
F
,
Yang
XA
.
SREBP2 contributes to cisplatin resistance in ovarian cancer cells
.
Exp Biol Med
2018
;
243
:
655
62
.
31.
Hong
X
,
Roh
W
,
Sullivan
RJ
,
Wong
KHK
,
Wittner
BS
,
Guo
H
, et al
.
The lipogenic regulator SREBP2 induces transferrin in circulating melanoma cells and suppresses ferroptosis
.
Cancer Discov
2021
;
11
:
678
95
.
32.
Yin
F
,
Feng
F
,
Wang
L
,
Wang
X
,
Li
Z
,
Cao
Y
.
SREBP-1 inhibitor Betulin enhances the antitumor effect of sorafenib on hepatocellular carcinoma via restricting cellular glycolytic activity
.
Cell Death Dis
2019
;
10
:
672
.
33.
Colgan
SM
,
Tang
D
,
Werstuck
GH
,
Austin
RC
.
Endoplasmic reticulum stress causes the activation of sterol regulatory element binding protein-2
.
Int J Biochem Cell Biol
2007
;
39
:
1843
51
.
34.
Moon
SH
,
Huang
CH
,
Houlihan
SL
,
Regunath
K
,
Freed-Pastor
WA
,
Morris
JP
4th
, et al
.
p53 represses the mevalonate pathway to mediate tumor suppression
.
Cell
2019
;
176
:
564
80
.
35.
Khatib
SA
,
Ma
L
,
Dang
H
,
Forgues
M
,
Chung
JY
,
Ylaya
K
, et al
.
Single-cell biology uncovers apoptotic cell death and its spatial organization as a potential modifier of tumor diversity in HCC
.
Hepatology
2022
.
36.
Suk
FM
,
Wang
YH
,
Chiu
WC
,
Liu
CF
,
Wu
CY
,
Chen
TL
, et al
.
Secretory NPC2 protein-mediated free cholesterol levels were correlated with the sorafenib response in hepatocellular carcinoma
.
Int J Mol Sci
2021
;
22
:
8567
.
37.
Li
D
,
Yao
Y
,
Rao
Y
,
Huang
X
,
Wei
L
,
You
Z
, et al
.
Cholesterol sensor SCAP contributes to sorafenib resistance by regulating autophagy in hepatocellular carcinoma
.
J Exp Clin Cancer Res
202241
:
116
.
38.
Huang
P
,
Nedelcu
D
,
Watanabe
M
,
Jao
C
,
Kim
Y
,
Liu
J
, et al
.
Cellular cholesterol directly activate smoothened in hedgehog signaling
.
Cell
2016
;
166
:
1176
87
.
39.
Luchetti
G
,
Sircar
R
,
Kong
JH
,
Nachtergaele
S
,
Sagner
A
,
Byrne
EF
, et al
.
Cholesterol activates the G-protein coupled receptor smoothened to promote hedgehog signaling
.
Elife
2016
;
5
:
e20304
.
40.
Wang
S
,
Yao
Y
,
Wang
X
,
Zheng
G
,
Ouyang
W
,
Chen
W
.
25-HC promotes hepatocellular carcinoma metastasis through up-regulation of TLR4 dependent FABP4
.
Am J Cancer Res
2019
;
9
:
2140
55
.
41.
Zheng
J
,
Sato
M
,
Mishima
E
,
Sato
H
,
Proneth
B
,
Conrad
M
.
Sorafenib fails to trigger ferroptosis across a wide range of cancer cell lines
.
Cell Death Dis
2021
;
12
:
698
.
42.
Mansourian
PG
,
Yoneda
M
,
Krishna Rao
M
,
Martinez
FJ
,
Thomas
E
,
Schiff
ER
.
Effects of statins on the risk of hepatocellular carcinoma
.
Gastroenterol. Hepatol (NY)
2014
;
10
:
417
26
.
43.
Feng
J
,
Dai
WQ
,
Mao
Y
,
Wu
L
,
Li
J
,
Chen
K
, et al
.
Simvastatin re-sensitizes hepatocellular carcinoma cells to sorafenib by inhibiting HIF-1α/PPAR-ϒ/PKM2-mediated glycolysis
.
J Exp Clin Cancer Res
2020
;
39
:
24
.
44.
Jouve
JL
,
Lecomte
T
,
Bouché
O
,
Barbier
E
,
Khemissa Akouz
F
,
Riachi
G
, et al
.
Pravastatin combination with sorafenib does not improve survival in advanced hepatocellular carcinoma
.
J Hepatol
2019
;
71
:
516
22
.
45.
Malini
PL
,
Ambrosioni
E
,
De Divitii
O
,
Somma
S Di
,
Rosiello
G
,
Trimarco
B
.
Simvastatin versus pravastatin: efficacy and tolerability in patients with primary hypercholesterolemia
.
Clin Ther
1991
;
13
:
500
10
.
46.
Bjarnadottir
O
,
Romero
Q
,
Bendahl
PO
,
Jirström
K
,
Rydén
L
,
Loman
N
, et al
.
Targeting HMG-CoA reductase with statins in a window-of-opportunity breast cancer trial
.
Breast cancer Res Treat
2013
;
138
:
499
508
.

Supplementary data