The alternative NF-κB RelB subunit is a novel critical sensor of lipid metabolism to promote tumor development

doi:10.21203/rs.3.rs-5233947/v1

Download PDF

Article

The alternative NF-κB RelB subunit is a novel critical sensor of lipid metabolism to promote tumor development

https://doi.org/10.21203/rs.3.rs-5233947/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Cancer cells reprogram their metabolism to fulfill their high energetic demand. Lipid metabolism is most often reprogrammed for cancer cell survival and tumor development. The role of alternative oncogenic NF-κB/RelB subunit in the reprogramming of lipid metabolism in cancer is unknown. Here we report that RelB plays a central role at the crossroads of lipid storage and liberation of fatty acids from the lipid droplets to feed the fatty acid oxidation (FAO) and mitochondrial energetic metabolism. High RelB expression defines a subset of hepatocellular carcinoma (HCC) patients and cell lines with a peculiar gene expression profile enriched in lipid catabolic-related genes, including lipases. Functional studies revealed that high RelB activation controls the expression of major lipolytic lipases including adipose triglyceride lipase (ATGL) and monoglyceride lipase (MAGL), and impacts on HCC cell survival, migration, and tumor development in vivo. Altogether, we uncovered that RelB is a central regulator of the lipid metabolism plasticity and an energy homeostasis sensor in cancer cells.

Biological sciences/Cancer/Cancer metabolism

Biological sciences/Cancer/Oncogenes

Cancer cells exhibit bioenergetics alterations in order to meet their high-energy requirement, and deregulation of cellular energetics has been recognized as one of the hallmarks of cancer¹. Lipids function as essential building blocks for membranes but also serve as fuel to drive energy-demanding processes and are regulators of numerous cellular functions². Lipid metabolism is most often reprogrammed for cancer cell survival and tumor development³ due to activated oncogenic pathways and environmental cues⁴.

Aberrant NF-κB activity has been observed in many cancers, including both solid and hematopoietic malignancies, and NF-kB transcription factors are major players in the control of cancer cell proliferation and survival, as well as in the pathophysiology of numerous cancers^5–10. The NF-κB transcription factor family consists of five members in mammals: RelA (p65), RelB, cRel (Rel), NF-κB1 (p50 and its precursor p105), and NF-κB2 (p52 and its precursor p100), and forms a collection of various homodimeric and heterodimeric complexes¹¹. The activity of NF-kB is regulated by two main pathways. The first one, known as the canonical NF-kB activation pathway, mainly applies to RelA- and/or cRel-containing complexes¹². The second one, the alternative NF-kB activation pathway, leads to the activation of RelB-containing dimers^7,13,14. How the alternative NF-kB/RelB subunit may impact on lipid metabolism in cancer cells, and its connection with the pathogenic process of cancer cells and poor patient outcome is currently unknown.

In the study presented here, we reveal a role for the alternative NF-kB/RelB subunit in the plasticity of lipid metabolism and energy homeostasis in cancer cells. RelB defines a new subset of hepatocellular carcinoma (HCC) cell lines and HCC patients with poor prognosis presenting a peculiar distinct gene expression profile enriched in lipases and lipase-related genes. Functional studies revealed that RelB drives HCC cancer cells to oxidize fatty acids (FA) for mitochondrial activity at the expense of lipid storage in lipid droplets. RelB-dependent control of FA oxidation (FAO) is linked to Adipose Triglyceride Lipase (ATGL) and Monoglyceride Lipase (MAGL) expression and impacts on HCC cell survival, migration, and tumor development in vivo. Ultimately, we show that RelB positivity is associated with high protein expression of ATGL in a small cohort of HCC patients. Altogether, our data highlight a previously unrecognized function for the alternative RelB NF-kB subunit as an energy sensor of cancer cell lipid metabolism and provide the framework for the development of new drugs targeting RelB to overcome resistance in HCC.

RelB controls lipases expression in HCC cancer cells

To assess how RelB may impact on the near-universal reprogramming of lipid metabolism in cancer cells, we developed different correlation matrixes using the Cancer Cell Lines Encyclopedia (CCLE) HCC cell lines dataset to evaluate the lipid metabolism features that correlated with RelB expression levels. A strong correlation was observed between RelB gene expression and a large set of lipases and lipase-related genes (Figure 1A and Supplemental Table 1). Further, we wanted to confirm the association between RelB and lipases by a second approach. In a first step, using the unsupervised K-means clustering, the HCC cell lines were separated into two different subsets (Figure 1B) that were significantly associated with RelB expression (Figure 1C). To go one step further, we then cataloged genes whose expression was associated with high levels of RelB expression. Interestingly, the high-RelB cluster was characterized by an enrichment of lipases (Figure 1D), thus further indicating the inter-relation between RelB and lipases.

Further in vivo evidence for a direct role for RelB in regulating lipase transcription was obtained by chromatin IP (ChIP) analysis. RelB is efficiently recruited to the promoter of various lipases, including the central lipolytic lipases ATGL, MAGL and the ATGL co-regulator ABHD5 for the release of FA from stored triacylglycerols (TAG) in lipid droplets (Figure 1E, F and Supplementary Figure 1).

RelB defines a new subset of HCC patients with a GEP enriched for lipase-related genes and associated with worse outcome

To go one step further and generate a molecular portrait of RelB activation in HCC patients, we further explored whether gene expression profile (GEP) could subdivide these patients based on their RelB expression levels. Kaplan-Meier method was first used to define a subset of high-RelB HCC patients with a poor prognosis impact on overall survival (OS) and disease specific survival (DSS) in the TCGA cohort of HCC patients (TCGA-LIHC, n= 365) (Figure 2A). There was no significant correlation between high-RelB expression and disease stages as evaluated by American Joint Committee on Cancer (AJCC 7^th edition) (Figure 2B). Interestingly, high RelB expression has also a strong prognosis impact at early stages of the disease (stage I and II) (supplementary Figure 2). Most importantly, the worse overall survival associated with the high-RelB HCC subset was validated in an independent cohort of patients with HCC (GSE 36376, n= 239) (Figure 2C). In a second step, a hierarchical clustering algorithm was used to group genes based on high-RelB cases vs low-RelB cases. We identified 872 genes associated with RelB expression with a fold change ≥2 (Figure D and Supplemental Table 2). Remarkably, a more detailed analysis of the high-RelB signature revealed increased expression of genes linked to lipid/FA metabolism (Figure 2E, F), along with immunity, inflammation and proliferation (Figure 2E). These observations confirm that the differences in lipid metabolism signature identified in HCC cell lines translate to primary tumor biopsies.

RelB is a sensor of lipid metabolism plasticity

The up-regulation of selected genes encoding for lipases in high-RelB expressing HCC cells predicts potential differences in lipid metabolism compared to low-RelB expressing HCC cells. To search for disturbance of this metabolic program, we first assessed the impact of RelB activation on lipid droplet content, the intracellular reservoir of lipids¹⁵. To do so, we generated a model in Huh7 cells (Huh7 belongs to the low-RelB cluster, see Figure 1B) that mimics high RelB expression in HCC cells by stable lentiviral transduction (Figure 3A and Supplemental Material). Importantly, Huh7 cells that ectopically express high RelB levels also resulted in a marked induction in RelB activity without affecting the canonical RelA activity (Figure 3B). To further characterize this cellular model, we performed a RNAseq analysis and showed that increased RelB activity in Huh7 cells induced a strong enrichment of the transcription of genes encoding for lipases (Figure 3C, n= 55 lipases, fold change >= 2), just as what was seen in the large panel of HCC cell lines (Figure 1). Altogether, these observations indicated that our high-RelB Huh7 cell model reflects RelB functions in naturally occurring high-RelB HCC cells. Remarkably, high RelB expression induced a marked decrease in lipid droplets (Figure 3D). Similarly, a decrease in intracellular TAG was also observed upon increased RelB expression in Huh7 cells (Figure 3E). In addition, using an untargeted NMR-based metabolomics approach, we monitored the metabolic changes occurring upon RelB activation in Huh7 cells. Ectopic expression of RelB led to a massive depletion of intracellular TAG and phospholipids (PL) (Figure 3F).

The RelB-dependent functional differences in neutral lipid storage predict differences in FA handling. We first evaluated the impact of RelB on lipid storage upon increasing amount of FA. In high-RelB Huh7 cells, lipid storage was significantly decreased in response to exogenously supplied monounsaturated FA (Oleic acid, Figure 3G, H) and saturated FA (Palmitic acid, Figure 3I) compared to what was seen in RelB-low expressing cells.

To provide independent and parallel evidence for a differential utilization of FA upon high RelB expression, a targeted ¹⁴C-oleic acid approach was undertaken and ¹⁴C enrichment was assessed in a defined set of metabolites derived from FA metabolism (Figure 3J). Very interestingly, the relative levels of ¹⁴C enrichment for TAG and DAG were significantly decreased upon high RelB expression, whereas acid-soluble metabolites (ASM) and CO₂ levels were strongly increased (Figure 3K). This is consistent with a greater entry of FA into FAO to promote mitochondrial activity, at the expense of esterification, the reaction that converts FA into neutral lipids, such as TAG, for lipid storage in lipid droplets. Altogether, these observations demonstrate the central role played by RelB into the metabolic reprogramming in HCC cells from lipid storage to FAO.

Further, to better understand the energetic consequences induced by high RelB activation, we evaluated how RelB impacts on the level of production of glycolytic vs mitochondrial ATP in HCC cells. Compared to low-RelB expressing Huh7 cells, high-RelB expressing cells derive a significantly higher portion of their total energy from mitochondria oxidative metabolism than glycolysis (Figure 3L). Altogether, these data indicate that RelB exerts a critical role in the energetic reprogramming of HCC cells and is a sensor of lipid metabolism.

RelB-dependent reprogramming of energy homeostasis in HCC cells impacts on tumor development

In a next step, we determined whether the RelB-dependent energy switch represented a mean for HCC cells to increase their survival upon metabolic stress. Nutrient deprivation-induced cell death was markedly decreased upon high RelB expression in Huh7 cells (Figure 4A). Remarkably, the RelB-dependent prosurvival function is abolished upon treatment with 4-BrCA, an inhibitor that irreversibly inhibits oxidation of long-, medium- and short-chain FA¹⁶ (Figure 4A), indicating that RelB promotes the ability of HCC cells to meet their metabolic demand during nutriment starvation via FAO, thus avoiding a metabolic crisis and cell death.

Next, we explored the functional relevance of RelB on cell migration which is important during the whole cascade of cancer development. A significant increase in wound closure was observed in high-RelB Huh7 cells as compared to what was seen in low-RelB Huh7 cells (Figure 4B). Further, since we observed that high RelB expression promotes transcription and protein expression of ATGL and MAGL (Figure 1D, F and Figure 4C, and Supplemental Material), the two major lipases controlling lipolysis¹⁷ (see scheme Figure 1E), we hypothesized that the pro-migration function of RelB could rely on ATGL and MAGL activity¹⁸. Remarkably, treatment with ATGL or MAGL inhibitors (ATGLstatin and JZL184, respectively) strikingly decreased the migration of high-RelB Huh7 cells (Figure 4B) with no effect on low-RelB Huh7 cells. Taken together, these results strongly suggest that high RelB expression promotes HCC cells migration via the control of expression and activity of central lipases (namely ATGL and MAGL), thereby leading to a profound rewiring of lipid metabolism from storage to usage as a major source of FA to feed FAO. To directly demonstrate that RelB-dependent FAO impacts on tumor growth, we took advantage of the well-recognized in vivo chorioallantoic membrane (CAM) model¹⁹ Although having little effect on basal tumor growth, high RelB activation renders tumors highly susceptible to FAO inhibition (Figure 4D). These findings indicate that boosting FA metabolism in established cancer cells through increased RelB expression promotes tumorigenesis in vivo. Ultimately, a significant accumulation of ATGL protein was observed in a cohort of HCC patients expressing high levels of RelB as compared to low expressing patients (Figure 4E and Supplemental Material). These observations indicate that the differences in lipid metabolism signature identified in HCC cell lines translate to primary tumor biopsies.

As summarized in Figure 5A, we have uncovered that RelB is a central sensor of lipid metabolism plasticity in HCC cancer cells. RelB controls the expression of central lipases (1) leading to increased lipolysis (2), thereby driving cancer cells to increased FAO for mitochondrial activity and energy production (3). The RelB-dependent rewiring of lipid metabolism towards FAO promotes tumor development (4).

Reprogramming of cancer cell metabolism is now a recognized hallmark of cancer. Recently, lipid metabolism has emerged as a central player in altered energetics in cancers affecting cancer cell survival and tumor development^3,4,20. Our study reveals that the alternative RelB NF-kB subunit plays a central role in the control of lipid metabolism plasticity at the crossroads of lipid storage and its use as a major source of FA to feed the process of FAO in the mitochondria. To our knowledge, this is the first report of a role for RelB in lipid metabolism reprogramming in cancer cells.

Few studies have reported the ability of the classical NF-kB pathway to promote either mitochondrial respiration^21–24 or Warburg effect²⁵ in cancer cells. However, an involvement of classical NF-kB in the control of lipid metabolism is currently unknown. In contrast, a role for the alternative NF-kB pathway in cancer cell metabolism remains almost undefined. The nuclear factor-kB-inducing kinase (NIK) is an upstream critical regulator of the alternative NF-kB pathway that leads mainly to RelB activation^7,26 . Although NIK was identified as a critical regulator of mitochondrial fission and metabolic adaptation of cancer cells^27–29, these effects were unexpectedly independent on NF-kB. We have been the first to report that RelB constitutive activation can reprogram diffuse large B-cell lymphoma (DLBCL) metabolism towards oxidative energy metabolism in DLBCL cells and protects DLBCL cells from antimetabolic drugs-induced apoptosis³⁰. A recent study suggested that RelB contributes to tamoxifen resistance in breast cancer cells via the upregulation of GPX4³¹.A preprint indicated that RelB overexpression in HepG2 cells under specific metabolic stress induced by chronic lipid-rich culture conditions induces a significant decrease in lipid accumulation³². However, how RelB impacts on lipid metabolism in cancer cells still remained a central unanswered question.

Our study reveals that RelB plays a central role at the crossroads of lipid storage and release of FA from the lipid droplets to feed FAO and mitochondrial energetic metabolism. First, we have shown that RelB expression correlates with the transcription of several lipases and lipase-related genes, and high RelB activity is linked to increased protein expression levels of ATGL and MAGL, the two central lipases involved in lipid droplet breakdown via lipolysis. Second, high RelB activation strongly decreases lipid droplet and TAG content. Third, high RelB activity rewires HCC cell metabolism to FAO at the expense of FA esterification and provides an energetic source to feed the mitochondrial respiration. Fourth, ATG and MAGL activities are required for increased HCC cell migration upon high RelB activation. Fifth, RelB-dependent increases in FAO protects HCC cells against nutrient deprivation-induced cell death and provides an essential signal for tumor growth. Sixth, high-RelB protein expression levels correlate with high ATGL expression in HCC patients.

Beyond HCC, abnormal RelB activity has been reported in several solid cancers, including glioma, breast, prostate, and colon cancer^33–38 as well as in hematopoietic malignancies such as multiple myeloma and DLBCL^39,40. Further abnormal RelB expression has been linked to increased several cancer aggressivity and poor patient outcome^{33,35,37,38,40–45}. Thus, it is likely that the RelB-dependent lipid/FA dynamics and associated pro-survival and tumorigenic effects may be generalized to other neoplasms, especially those addicted to alternative NF-kB.

In summary, we established that RelB is a central player in lipid metabolism plasticity and energy homeostasis in HCC cells. Our data are of great functional importance because they constitute a significant advance in the understanding of RelB function in cancer, and provide a strong rationale for the development of new molecules targeting RelB for therapeutic intervention in cancer.

Acknowledgments

We thank T. de Mont-Marin for his great help for teaching Python programming, and V. Lenoir, C. Klein, E. Benichou and B. Seffou for technical help. This work was supported by grants from Institut National du Cancer, Fondation ARC pour la Recherche sur le Cancer, 12 Rounds contre le Cancer, the integrated cancer research center "SiRIC InsiTu: Insights into cancer: From inflammation to tumor" (grant number INCa-DGOS-INSERM-ITMO Cancer_18008), INSERM, and Université Paris Cité (V.B.), and doctoral fundings from French Ministry « Ministère de l'Enseignement supérieur et de la Recherche » and « 12 Rounds contre le Cancer » (C.B., T.B).

Authorship

Contribution: V.B.: conceptualization, methodology, supervision, project administration; T.B., C.B., C.Br., B.E., D.B., L.T., C.R., M.-C.A.-G., C.C.: investigation; V.B., T.B., C.B., C.Br., B.E., L.T., A.M.,C.C., M.-C.A.-G., G.B., C.R., N.C., D.C., R.D., C.P.-B.: formal analysis; B.E., H.A., M.B., N.G., F.B., V.P.: ressources (reagents, materials, and analysis tools); V.B., T.B., C.B.: writing; V.B., T.B., C.B.: visualization ; V.P., M.B., R.D., C.P.-B.: critical reading of the manuscript.

Antibodies and Reagents

The antibodies were purchased from Santa Cruz Biotechnology (RelB, sc-226; RelA, sc-372; ATGL, sc-365278; MGL, sc-398942; β-actin, sc-47778; RelA-F6-X, sc-8008(X); RelB-C19X, sc-226(X); cRel-CX, sc-70(X), p105/p50-H119X, sc-7178X; p100/p52-K27X, sc-298X). Oleic acid (01008), palmitic acid (P0500), L-carnitine hydrochloride (C0283), 2-deoxy-D-glucose [2-DG] (D8375), Oil Red O (O0625) and mitomycin C (M4287) were purchased from Merck/Sigma Aldrich. 4’,6-diaminidino-2-phenylindole [DAPI] from Molecular Probes. 4-Bromobut-2-enoic acid [4-BrCa] (A698521) was purchased from Ambeed; Atglsitatin (HY-15859) and JZL184 (HY-15249) from MedChemExpress, and Matrigel (354234) from Corning.

Plasmid constructs

RelB cDNA lentiviral vector was previously described⁴⁶.

Cell lines and culture conditions

Huh7 cells were grown in MEM culture medium (41090-028 Gibco) supplemented with 10 % of fetal bovine serum [FBS], 100 U/mL penicillin, and 100 mg/mL streptomycin (15140-122 Gibco), 25 mM Hepes Buffer (15630-056 Gibco), 1 mM of sodium pyruvate (ref 11360-039 Gibco), 1 % (v/v) of non-essential amino-acids [NEAA] (ref 11140-050 Gibco), which stands for standard culture medium. For exogenous fatty acid treatment, exponentially growing Huh7 cells were serum-starved for 5 h, and then oleic or palmitic acid with 1% (w/v) of fatty acid-free bovine serum albumin (BSA) (A7030 Sigma Aldrich) were added for 16 h.

Lentiviral production and transduction

Production of infectious recombinant lentiviruses was performed by transient transfection of 293T cells as previously described⁴⁷. For infections, cells were incubated overnight with recombinant lentiviruses. An equal amount of fresh culture medium was added 24 h later and after 48 h, cells were washed and seeded in fresh culture medium. GFP positive cells were sorted with FACSAria™ sorter (Becton Dickinson).

Immunoblotting

Immunoblotting were performed as previously described⁴⁶

Electrophoretic Mobility Shift Assays (EMSA) for NF-kB

Protein cell extracts were prepared and analyzed for DNA binding activity using the HIV-LTR tandem kB oligonucleotide as kB probe⁴⁸. For supershift assays, protein extracts were incubated with specific antibodies for 30 min on ice before incubation with the labeled probe.

ChiPseq Atlas data analysis

RelB recruitment to the promoter regions of genes encoding for lipases was examined using ChIPseq Atlas database and available data from GSE 55105⁴⁹. Visualization using IGV browser.

Lipid droplet content

Fixed cells with formalin 10 % (v/v) and isopropanol 60 % (v/v) were stained with Oil Red O for 10 min at room temperature. Image acquisition was performed using optical microscopy (magnification: 20X), and quantification of Oil Red O staining was made using image J software (V1.53c) and expressed as lipid droplet area.

Total triglyceride content

Lipids were extracted using the Folch method using chloroform/methanol (2:1, v/v). Triglyceride content was quantified using a colorimetric diagnostic triglyceride assay kit (Triglycerides FS; Diasys), and further normalized to whole cell extract protein concentration as measured by BCA Protein Assay kit (23227, Pierce).

Fatty acid metabolic flux

[1-¹⁴C]Oleic acid was obtained from GE Healthcare. Huh7 cells were treated for 16 h with exogenous oleate (50 µM), and then oleate oxidation and esterification rates were determined during the last 3 h of culture in the presence of 300 µM of [1-¹⁴C] oleate (0.5 Ci/mol) bound to 1% (w/v) BSA and 1 mM carnitine. At the end of the 3 h-incubation period, the medium was collected to determine [¹⁴C]CO₂ and [¹⁴C]acid-soluble metabolites (ASMs) by scintillation counting⁵⁰ (Packard). ASMs essentially consist of acyl-carnitines, Krebs cycle intermediates and acetyl-CoA. Oleate oxidation rates were calculated from [¹⁴C]CO₂ and [¹⁴C]ASMs and expressed as nmoles of oleate oxidized over 3h per mg of protein. For esterification rates, cells were washed and scraped in PBS to determine [¹⁴C]TAG, [¹⁴C]phospholipids (PL) and [¹⁴C]diacylglycerols (DAG). Lipids were extracted with chloroform/methanol (2:1, v/v) and separated by thin-layer chromatography on silica-gels plates⁵⁰ (Merck). TAG, DAG and PL were identified using standard oils and coloration with iodine vapour. Specific bands for each fraction were cut and placed in scintillation liquid, and ¹⁴C-labelled PL, DAG and TAG products were quantified by radioactivity counting. Protein content was determined using the BCA protein assay (Thermo Fisher Scientific).

Measurement of intracellular ATP content

Intracellular ATP content was measured using an ATP luminescence detection assay kit (CellTiter-Glo^® Luminescent Cell Viability Assay, Promega, Madison, WI, USA) according to the manufacturer’s instructions. Briefly, around 4 x 10³ cells were seeded (96 well plates) in 100µL of cell culture medium (DMEM with no glucose, ref A1443001 supplemented with 10% FBS, glucose (1g/L), L-carnitine (1 mM), 100 U/mL penicillin, 100 mg/mL streptomycin, 25 mM Hepes Buffer) with or without exogenous fatty acid and 1% (w/v) fatty acid-free BSA for 16 h. Cells were treated for 1h 30 min with or without 20 mM 2-deoxy-D-glucose (2-DG) at 37°C followed by 30 min at room temperature to inhibit glycolysis. At the end of the incubation time, 100 µL of CellTiter-Glo solution were added, and cells were incubated for 10 min at room temperature. The luminescence signal (relative luminescence units (RLU)) was recorded with a microplate reader (Centro LB 960 microplate luminometer, Berthold Technologies, Bad Wildbach, Germany). The difference between total ATP content and ATP produced under 2-DG treatment results in glycolytic ATP.

Scratch-Wound Assays

Cells were grown to a confluent monolayer for 24 h in complete MEM medium with 10% (v/v) FBS, and then cells were grown in 0 % FBS for an additional 5 h before scratches were made using a 0.5-mm-diameter pipette tip to slow down proliferation. Two hours before the scratches, mitomycin C (14.9 nM) was added with or without Atglistatin (40 µM) or JZL184 (0.5 µM). After the scratch, the medium was replaced with MEM medium containing no FBS with or without Atglistatin or JZL184 at the same concentration. Scratch area was monitored over 48 h. Scratch assays were quantified using the MRI wound healing tool of ImageJ (V1.54i).

Cell viability assay

Cell viability was assessed using 4’, 6-diamidino-2-phenylindole (DAPI) 1 µg/mL staining for 5 min at room temperature with stained (non-viable) and unstained (viable) cells.

CAM model experiment

Fertilized chicken eggs (Gallus domesticus) were incubated at 37.7°C and 55% humidity laid vertically and turned 180° every hour for 8 days. On day 8 of embryonic development, an about 1 cm2 window on the eggshell was performed, careful not to disrupt the remaining structure, and then covered with tape for further incubation. High-RelB and low-RelB Huh7 cells growth in condition of exogeneous FA (see above, cell culture conditions) were harvested, resuspended in MEM/Matrigel 50/50 vol/vol (25 mL), and engrafted on the CAM (n = 10 to 12 eggs/condition). CAM presenting a tumor at day 12 were treated with or without 4-BrCA at days 12 and 14. At day 16, chick embryos were sacrificed by decapitation. Only embryos with no malformations at day 16 were included in the study. According to the French legislation, approval from an institutional or licensing committee is not required for experiments on the chorioallantoic membrane (CAM) terminated before embryonic developmental day 17.

Correlation matrix

RNA-seq reads data for hepatocellular carcinoma (HCC) cell lines were extracted and normalized from the CCLE dataset (https://portals.broadinstitute.org/ccle). The selected cell lines were as follow: JHH2 (ACH-000577), KMCH-1 (ACH-001853), SNU475 (ACH-000422), SNU-761 (ACH-000537), SNU-878 (ACH-000686), SNU886 (ACH-000316), SNU-398 (ACH-000221), SNU49 (ACH-000420), SNU182 (ACH-000483), SNU387 (ACH-000478), SNU423 (ACH-000493), PLC-PRF5 (ACH-001318), HEP3B217 (ACH-000625), SK-HEP1 (ACH-000361), HUH1 (ACH-000475), HUH7 (ACH-000480), HLF (ACH-000393), JHH4 (ACH-000476), JHH5 (ACH-000734), JHH6 (ACH-000217), JHH7 (ACH-000848), JHH1 (ACH-000620), and LI7 (ACH-000471). Python (version 3.11.9) was used for all data analysis. Pearson correlation coefficients were calculated to identify significant relationships of expression between RelB and other genes. Only correlations with an absolute value greater than or equal to 0.3 were considered significant. Genes correlated with RelB were functionally annotated using Gene Ontology (GO) terms obtained from the Gene Ontology database, focusing on terms related to «lipase».

K-means clustering

Using the CCLE dataset for HCC cell lines as above, unsupervised Principal Component Analysis (PCA) was performed (scikit-learn 1.5.1) with 2 outliers removed (SNU-761/ACH-000537 and PLC-PRF5/ACH-001318). HCC cell lines were clustered using the unsupervised K-Means (scikit-learn 1.5.1) method based on their «lipase» genes expression profiles, and two independent clusters were identified. Mann-Whitney U tests were conducted to compare the expression levels of selected genes between the two clusters identified by K-Means clustering. RelB expression was significantly different between the two clusters, named High-RelB vs Low-RelB expressing HCC cell line clusters. Statistical significance was assessed with a p-value threshold of <0.05.

Determination of RelB gene expression signature in HCC patients

Kaplan-Meier method using Python (version 3.11.9) and lifelines library were first used to define a subset of high-RelB HCC patients with a poor prognosis impact on disease specific survival (DSS) and overall survival (OS) from RNAseq dataset of the TCGA cohort of HCC patients (TCGA-LIHC, n= 365), and confirmed on a validation cohort (GSE 36376, n=239). Next, using the heatmaply v1.5.0 R package, heatmap was generated to compare gene expression levels defining two distinct subgroups corresponding to high-RelB and low-RelB expression HCC patients. The heatmap includes genes differentially expressed with a fold change ≥2 and a p value <0.05.

RNAseq analysis

RNA sequencing librairies and sequencing were performed by the GenomeEast platform, a member of the ‘France Génomique’ consortium (Illkirch, France). Briefly, RNA-Seq libraries were generated from 600 ng of total RNA using TruSeq Stranded mRNA LT Sample Preparation Kit (Illumina, San Diego, CA), according to manufacturer's instructions (TruSeq Stranded mRNA Sample Preparation Guide - PN 15031047: Rev.E Oct 2013). The final cDNA libraries were checked for quality and quantified using capillary electrophoresis with 2100 Bioanalyzer (Agilent Technologies France, Les Ulis) and further sequenced on a HiSeq 4000 sequencer (Illumina) as single-end 50 base reads following Illumina's instructions.

Reads were preprocessed using cutadapt 1.10 (Martin, 2011) in order to remove adaptors and low-quality sequences and reads shorter than 40 bp were removed for further analysis. Remaining reads were mapped to Homo sapiens rRNA and ERCC Spike-In sequences using bowtie 2.2.8 and reads mapped to those sequences were removed for further analysis. Remaining reads were aligned to hg38 assembly of Homo sapiens with STAR 2.5.3a. Gene quantification was performed with htseq-count 0.6.1p1, using “union” mode and Ensembl 94 annotations. Differential gene expression analysis was performed using DESeq2 1.16.1. Bioconductor R package on previously obtained counts (with default options). P-values were adjusted for multiple testing using the Benjamini and Hochberg method.

Metabolomics by NMR

NMR sample preparation. Exponentially growing high-RelB and low-RelB Huh7 cells were seeded in a total volume of 10 ml in 100 cm2 petri dish (total 5x10⁶ cells/condition). To address the reproducibility of the experiment, 10 individual samples of dry cell pellets were prepared. Dry cell pellets were treated with 400 µL of ice-cold methanol and 65 µL of ice-cold water. A volume of 200 µL of ice-cold chloroform, then 200 µL of ice-cold water and finally 200 µL of ice-cold chloroform were further added. The samples were then left on ice for 15 min, and centrifugated at 2000 rpm for 15 min. The upper methanol/water phase (containing polar metabolites) and the lower chloroform phase (containing lipophilic compounds) were transferred into separate glass vials. Solvents were removed under a stream of nitrogen for 30 min. The water phases were put in the freeze dryer one night. The dry water phase samples were resuspended in 580 µL of NMR buffer (100 mM sodium phosphate buffer pH 7.4, in D₂O, with 3.57 mM TSP and 6 mM NaN₃) to obtain the polar phase extract. The lipophilic phases were resuspended in 600 µL of deuterated NMR solvent (2:1 mixture of pure CDCl₃ and CDCl₃ containing 0.03 % of TMS, to obtain a final solution of 0.01 % of TMS) and then vortexed. After centrifugation (5 min), 580 µL of the supernatant was directly transferred into an NMR tube to prepare the organic phase extract.

NMR data acquisition. All samples were measured on a 500 MHz Bruker Avance III spectrometer equipped with a SampleXpress automation sample changer and a cryogenic 5mm ¹³C/¹H dual probe with Z-gradient. The endometabolome samples were acquired using the classical experiment used for the metabolomic analysis of cell or tissue extracts by NMR⁵¹. 1D-CPMG pulse sequence with presaturation and 1D 1H experiment were performed for the aqueous phase extract and the organic phase extract, respectively. The operating software was TopSpin 3.1 1 (Bruker BioSpin, France).

Metabolite data processing and metabolite identification. For each NMR spectrum, the standard processing protocol was applied, including Fourier transform, phase correction, baseline correction, and calibration. For metabolite identification, 1H NMR spectra were analyzed using Chenomx NMR Suite 8.0 (Chenomx Inc.). Metabolite identification was performed using the spectral reference library and fitting peak intensity and location to the observed signals. The identification of compounds found in the organic phase was based on the publication of Amiel et al⁵².

The processing of raw data was carried out using NMRProcflow, including baseline correction, chemical shift calibration, data alignment, and variable-sized bucketing. For all data sets, a threshold corresponding to a signal-to-noise ratio of 3 was applied.

Statistical Analysis. A constant sum normalization was applied for all samples. Subsequently, we used SIMCA 15 software (Umetrics) to perform principal component analysis (PCA) or orthogonal projections to latent structure discriminant analysis (OPLSDA) to discriminate cancer cell line classes. Pareto scaling was used in both analyses. MetaboAnalyst 6.0 was used to perform t-tests and determine p-values with FDR correction and to generate box plots (https://www.metaboanalyst.ca/).

Statistical analysis

Kaplan-Meier estimates of overall survival were performed by using the log-rank test (Mantel-Cox). All statistical analyses were performed with Python software (https://www.python.org/ , version 3.11.9) or GraphPad Prism software (version 10.1.2). Statistical analysis by student t test or Mann Whitney test. A value of p=0.05 was considered as statistically significant with the following degrees: *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001.

1. Hanahan, D. & Weinberg, R. A. Hallmarks of Cancer: The Next Generation. Cell 144, 646–674 (2011).

2. Olzmann, J. A. & Carvalho, P. Dynamics and functions of lipid droplets. Nat Rev Mol Cell Biol 20, 137–155 (2019).

3. Broadfield, L. A., Pane, A. A., Talebi, A., Swinnen, J. V. & Fendt, S.-M. Lipid metabolism in cancer: New perspectives and emerging mechanisms. Developmental Cell 56, 1363–1393 (2021).

4. Pavlova, N. N., Zhu, J. & Thompson, C. B. The hallmarks of cancer metabolism: Still emerging. Cell Metab 34, 355–377 (2022).

5. Baldwin, A. S. Regulation of cell death and autophagy by IKK and NF-κB: critical mechanisms in immune function and cancer. Immunol Rev 246, 327–345 (2012).

6. Naugler, W. E. & Karin, M. NF-kappaB and cancer-identifying targets and mechanisms. Curr Opin Genet Dev 18, 19–26 (2008).

7. Baud, V. & Karin, M. Is NF-kappaB a good target for cancer therapy? Hopes and pitfalls. Nat Rev Drug Discov 8, 33–40 (2009).

8. Ben-Neriah, Y. & Karin, M. Inflammation meets cancer, with NF-κB as the matchmaker. Nat Immunol 12, 715–723 (2011).

9. Eluard, B., Thieblemont, C. & Baud, V. NF-κB in the New Era of Cancer Therapy. Trends Cancer 6, 677–687 (2020).

10. Nakanishi, C. & Toi, M. Nuclear factor-kappaB inhibitors as sensitizers to anticancer drugs. Nat Rev Cancer 5, 297–309 (2005).

11. Oeckinghaus, A. & Ghosh, S. The NF-κB Family of Transcription Factors and Its Regulation. Cold Spring Harbor Perspectives in Biology 1, a000034–a000034 (2009).

12. Karin, M. & Ben-Neriah, Y. Phosphorylation meets ubiquitination: the control of NF-κB activity. Annu Rev Immunol 18, 621–663 (2000).

13. Sun, S.-C. The non-canonical NF-κB pathway in immunity and inflammation. Nat Rev Immunol 17, 545–558 (2017).

14. Baud, V. & Jacque, E. [The alternative NF-kB activation pathway and cancer: friend or foe?]. Med Sci (Paris) 24, 1083–1088 (2008).

15. Farese, R. V. & Walther, T. C. Lipid Droplets Finally Get a Little R-E-S-P-E-C-T. Cell 139, 855–860 (2009).

16. el-Aleem, S. A. & Schulz, H. Evaluation of inhibitors of fatty acid oxidation in rat myocytes. Biochem Pharmacol 36, 4307–4312 (1987).

17. Grabner, G. F., Xie, H., Schweiger, M. & Zechner, R. Lipolysis: cellular mechanisms for lipid mobilization from fat stores. Nat Metab 3, 1445–1465 (2021).

18. Nagarajan, S. R., Butler, L. M. & Hoy, A. J. The diversity and breadth of cancer cell fatty acid metabolism. Cancer Metab 9, 2 (2021).

19. Ranjan, R. A. et al. The Chorioallantoic Membrane Xenograft Assay as a Reliable Model for Investigating the Biology of Breast Cancer. Cancers (Basel) 15, 1704 (2023).

20. Currie, E., Schulze, A., Zechner, R., Walther, T. C. & Farese, R. V. Cellular fatty acid metabolism and cancer. Cell Metab 18, 153–161 (2013).

21. Mauro, C. et al. NF-κB controls energy homeostasis and metabolic adaptation by upregulating mitochondrial respiration. Nat Cell Biol 13, 1272–1279 (2011).

22. Capece, D. et al. NF-κB and mitochondria cross paths in cancer: mitochondrial metabolism and beyond. Semin Cell Dev Biol 98, 118–128 (2020).

23. Birkenmeier, K. et al. Hodgkin and Reed-Sternberg cells of classical Hodgkin lymphoma are highly dependent on oxidative phosphorylation. Int J Cancer 138, 2231–2246 (2016).

24. Johnson, R. F., Witzel, I.-I. & Perkins, N. D. p53-dependent regulation of mitochondrial energy production by the RelA subunit of NF-κB. Cancer Res 71, 5588–5597 (2011).

25. Kawauchi, K., Araki, K., Tobiume, K. & Tanaka, N. p53 regulates glucose metabolism through an IKK-NF-kappaB pathway and inhibits cell transformation. Nat Cell Biol 10, 611–618 (2008).

26. Baud, V. & Collares, D. Post-Translational Modifications of RelB NF-κB Subunit and Associated Functions. Cells 5, 22 (2016).

27. Pflug, K. M., Lee, D. W., Keeney, J. N. & Sitcheran, R. NF-κB-inducing kinase maintains mitochondrial efficiency and systemic metabolic homeostasis. Biochim Biophys Acta Mol Basis Dis 1869, 166682 (2023).

28. Jung, J.-U. et al. NIK/MAP3K14 Regulates Mitochondrial Dynamics and Trafficking to Promote Cell Invasion. Curr Biol 26, 3288–3302 (2016).

29. Kamradt, M. L. et al. NIK promotes metabolic adaptation of glioblastoma cells to bioenergetic stress. Cell Death Dis 12, 271 (2021).

30. Nuan-Aliman, S., Bordereaux, D., Thieblemont, C. & Baud, V. The Alternative RelB NF-kB Subunit Exerts a Critical Survival Function upon Metabolic Stress in Diffuse Large B-Cell Lymphoma-Derived Cells. Biomedicines 10, 348 (2022).

31. Xu, Z. et al. RelB-activated GPX4 inhibits ferroptosis and confers tamoxifen resistance in breast cancer. Redox Biology 68, 102952 (2023).

32. Tzouanas, C. N. et al. Chronic metabolic stress drives developmental programs and loss of tissue functions in non-transformed liver that mirror tumor states and stratify survival. Preprint at bioRxiv 2023.11.30.569407 (2023) doi:10.1101/2023.11.30.569407.

33. Scherr, A.-L. et al. Etiology-independent activation of the LTβ-LTβR-RELB axis drives aggressiveness and predicts poor prognosis in HCC. Hepatology 80, 278–294 (2024).

34. Chen, X., Zhou, Y., Li, Z. & Wang, Z. Mining Database for the Expression and Clinical Significance of NF-κB Family in Hepatocellular Carcinoma. J Oncol 2020, 2572048 (2020).

35. Lee, D. W. et al. The NF-κB RelB protein is an oncogenic driver of mesenchymal glioma. PLoS One 8, e57489 (2013).

36. Wang, M. et al. RelB sustains endocrine resistant malignancy: an insight of noncanonical NF-κB pathway into breast Cancer progression. Cell Commun Signal 18, 128 (2020).

37. Zhang, Y. et al. RelB upregulates PD-L1 and exacerbates prostate cancer immune evasion. J Exp Clin Cancer Res 41, 66 (2022).

38. Zhou, X. et al. RelB plays an oncogenic role and conveys chemo-resistance to DLD-1 colon cancer cells. Cancer Cell Int 18, 181 (2018).

39. Cormier, F. et al. Frequent engagement of RelB activation is critical for cell survival in multiple myeloma. PLoS One 8, e59127 (2013).

40. Eluard, B. et al. The alternative RelB NF-κB subunit is a novel critical player in diffuse large B-cell lymphoma. Blood 139, 384–398 (2022).

41. Mulligan, E. A. et al. Expression and Activity of the NF-κB Subunits in Chronic Lymphocytic Leukaemia: A Role for RelB and Non-Canonical Signalling. Cancers (Basel) 15, 4736 (2023).

42. Dimitrakopoulos, F.-I. D. et al. Expression Of Intracellular Components of the NF-κB Alternative Pathway (NF-κB2, RelB, NIK and Bcl3) is Associated With Clinical Outcome of NSCLC Patients. Sci Rep 9, 14299 (2019).

43. Giopanou, I., Lilis, I., Papadaki, H., Papadas, T. & Stathopoulos, G. T. A link between RelB expression and tumor progression in laryngeal cancer. Oncotarget 8, 114019–114030 (2017).

44. Tao, Y. et al. Alternative NF-κB signaling promotes colorectal tumorigenesis through transcriptionally upregulating Bcl-3. Oncogene 37, 5887–5900 (2018).

45. Wang, X. et al. Oestrogen signalling inhibits invasive phenotype by repressing RelB and its target BCL2. Nat Cell Biol 9, 470–478 (2007).

46. Authier, H. et al. IKK phosphorylates RelB to modulate its promoter specificity and promote fibroblast migration downstream of TNF receptors. Proc Natl Acad Sci U S A 111, 14794–14799 (2014).

47. Kieusseian, A. et al. Expression of Pitx2 in stromal cells is required for normal hematopoiesis. Blood 107, 492–500 (2006).

48. Jacque, E., Tchenio, T., Piton, G., Romeo, P.-H. & Baud, V. RelA repression of RelB activity induces selective gene activation downstream of TNF receptors. Proc Natl Acad Sci U S A 102, 14635–14640 (2005).

49. Zhao, B. et al. The NF-κB genomic landscape in lymphoblastoid B cells. Cell Rep 8, 1595–1606 (2014).

50. Akkaoui, M. et al. Modulation of the hepatic malonyl-CoA-carnitine palmitoyltransferase 1A partnership creates a metabolic switch allowing oxidation of de novo fatty acids. Biochem J 420, 429–438 (2009).

51. Beckonert, O. et al. Metabolic profiling, metabolomic and metabonomic procedures for NMR spectroscopy of urine, plasma, serum and tissue extracts. Nat Protoc 2, 2692–2703 (2007).

52. Amiel, A. et al. Proton NMR Enables the Absolute Quantification of Aqueous Metabolites and Lipid Classes in Unique Mouse Liver Samples. Metabolites 10, (2020).

There is NO Competing Interest.

SupplementalMaterial.pdf
Supplemental material
SupplementaryTable1V2.pdf
Supplemental Table 1
SupplementaryTable2V2.pdf
Supplemental Table 2
SupplementaryFig1V5.pdf
Supplemental Figure 1
SupplementaryFig2V5.pdf
Supplemental Figure 2

Download PDF

Version 1

posted

You are reading this latest preprint version

The alternative NF-κB RelB subunit is a novel critical sensor of lipid metabolism to promote tumor development

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Declarations

Methods

References

Additional Declarations

Supplementary Files

Status:

Version 1