ATP citrate lyase is an essential player in the metabolic rewiring induced by PTEN loss during T-ALL development

Abstract

Alterations inactivating the tumor suppressor gene PTEN drive the development of solid and hematologic cancers, such as T-cell acute lymphoblastic leukemia (T-ALL), in which phosphatase and tensin homolog (PTEN) loss defines poor-prognosis patients. We investigated the metabolic rewiring induced by PTEN loss in T-ALL, aiming to identify novel metabolic vulnerabilities. We showed that the enzyme adenosine triphosphate (ATP) citrate lyase (ACLY) is strictly required for the transformation of thymic immature progenitors and the growth of human T-ALL, which remain dependent on ACLY activity even upon transformation. Although Pten-mutant mice all died within 17 weeks, the concomitant Acly deletion prevented disease initiation in 70% of the animals. In these animals, ACLY promoted B-cell lymphoma (BCL-2) epigenetic upregulation and prevented the apoptosis of premalignant double-positive thymocytes. Transcriptomic and metabolic analysis of primary T-ALL cells next translated our findings to the human pathology, showing that PTEN-altered T-ALL cells activate ACLY and are sensitive to its genetic targeting. ACLY activation thus represents a metabolic vulnerability with therapeutic potential for high-risk patients with T-ALL.

Graphical abstract

Figure 1.

Metabolic analysis of preleukemic PTEN-mutant thymocytes. (A) Principal component analysis of 4 Pten-mutant and 5 Ptenflox2 thymi issued from preleukemic 8-week-old mice subjected to unsupervised metabolic analysis. (B-D) Heat maps of the 50 most deregulated metabolites (B), glycolytic intermediates, Krebs cycle intermediates, glutamine, glutamate, citrate, CoA, acetyl-CoA, and fatty acids (C), and pathway analysis in 4 preleukemic PTENΔ/Δ thymocytes or 5 flox control counterparts (D). (E) Phosphorylated ser455-ACLY in the thymi of preleukemic PTENΔ/Δ mutant or control animals in 3 independent experiments, with at least 2 per group. Adj, adjusted; aKG, α-ketoglutarate; CoA, coenzyme A; F1,6DP, fructose 1,6 diphosphate; FAS, fatty acid synthesis; FDR, false discovery rate; TCA, tricarboxylic acid.

Figure 1.

Metabolic analysis of preleukemic PTEN-mutant thymocytes. (A) Principal component analysis of 4 Pten-mutant and 5 Ptenflox2 thymi issued from preleukemic 8-week-old mice subjected to unsupervised metabolic analysis. (B-D) Heat maps of the 50 most deregulated metabolites (B), glycolytic intermediates, Krebs cycle intermediates, glutamine, glutamate, citrate, CoA, acetyl-CoA, and fatty acids (C), and pathway analysis in 4 preleukemic PTENΔ/Δ thymocytes or 5 flox control counterparts (D). (E) Phosphorylated ser455-ACLY in the thymi of preleukemic PTENΔ/Δ mutant or control animals in 3 independent experiments, with at least 2 per group. Adj, adjusted; aKG, α-ketoglutarate; CoA, coenzyme A; F1,6DP, fructose 1,6 diphosphate; FAS, fatty acid synthesis; FDR, false discovery rate; TCA, tricarboxylic acid.

Figure 2.

Metabolic analysis of thymi from PTENΔ/Δ, ACLYΔ/Δ, PTENΔ/Δ;ACLYΔ/Δ or PTENflox2;ACLYflox2 mice. (A) Principal component analysis of 4 PTENΔ/Δ mutant, 3 ACLYΔ/Δ mutant, 3 PTENΔ/Δ;ACLYΔ/Δ, and 5 Ptenflox2;Aclyflox2 thymi from preleukemic 8-week-old mice undergoing unsupervised metabolic analysis. (B-D) Heat maps of the 50 most deregulated metabolites (B), analysis of glycolytic intermediates, Krebs cycle intermediates, glutamine, glutamate, citrate, CoA, acetyl-CoA and fatty acids (C), and pathway analysis in 4 PTENΔ/Δ mutant, 3 ACLYΔ/Δ mutant, 3 PTENΔ/Δ;ACLYΔ/Δ, and Ptenflox2;Aclyflox2 thymi (D). (E-F) Extracellular acidification rate (ECAR) (E) and oxygen consumption rate (OCR) (F) of sorted DP thymocytes in 2 independent experiments, with at least 2 per group. Adj, adjusted; aKG, α-ketoglutarate; F1,6DP, fructose 1,6 diphosphate; FAS, fatty acid synthesis; FCCP, carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone; FDR, false discovery rate; TCA, tricarboxylic acid cycle; WT, wild-type.

Figure 2.

Metabolic analysis of thymi from PTENΔ/Δ, ACLYΔ/Δ, PTENΔ/Δ;ACLYΔ/Δ or PTENflox2;ACLYflox2 mice. (A) Principal component analysis of 4 PTENΔ/Δ mutant, 3 ACLYΔ/Δ mutant, 3 PTENΔ/Δ;ACLYΔ/Δ, and 5 Ptenflox2;Aclyflox2 thymi from preleukemic 8-week-old mice undergoing unsupervised metabolic analysis. (B-D) Heat maps of the 50 most deregulated metabolites (B), analysis of glycolytic intermediates, Krebs cycle intermediates, glutamine, glutamate, citrate, CoA, acetyl-CoA and fatty acids (C), and pathway analysis in 4 PTENΔ/Δ mutant, 3 ACLYΔ/Δ mutant, 3 PTENΔ/Δ;ACLYΔ/Δ, and Ptenflox2;Aclyflox2 thymi (D). (E-F) Extracellular acidification rate (ECAR) (E) and oxygen consumption rate (OCR) (F) of sorted DP thymocytes in 2 independent experiments, with at least 2 per group. Adj, adjusted; aKG, α-ketoglutarate; F1,6DP, fructose 1,6 diphosphate; FAS, fatty acid synthesis; FCCP, carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone; FDR, false discovery rate; TCA, tricarboxylic acid cycle; WT, wild-type.

Figure 3.

Genetic Acly ablation prevents PTEN loss–induced T-ALL/lymphoma development. (A) Kaplan-Meier survival curve for PTENΔ/Δ mutant mice (n = 43), ACLY Δ/Δ mutant (n = 16), PTENΔ/Δ;ACLY Δ/Δ double-mutant mice (n = 22), and floxed controls (n = 40). (B-C) TCRβ rearrangements (B) and phenotypic analysis of the thymus (C) in PTENΔ/Δ mutant, ACLYΔ/Δ mutant, PTENΔ/Δ;ACLYΔ/Δ animals, or flox controls. (D-F) CD4⁺ absolute numbers cells in the spleen (D), bone marrow (E), and lymph nodes (F) in 14-week-old PTENΔ/Δ mutant, ACLYΔ/Δ mutant, PTENΔ/Δ;ACLYΔ/Δ animals, or flox controls (including Ptenflox2, Aclyflox2 and Ptenflox2;Aclyflox2 mice). BM, bone marrow.

Figure 4.

Acly ablation abrogates apoptosis resistance in mutant DP progenitors. (A-B) Phospho ser473-AKT (A) and phospho ser455-ACLY (B) on DN, DP, SP4, and SP8 from preleukemic PTENΔ/Δ mutant or PTENflox2 controls. In mutant mice, carrying the ROSAYFPflox2 reporter, DN, DP, SP4, and SP8 cells are gated on YFP+ cells. Levels of BCL-XL (C-D) and BCL-2 (C,E) in DP cells from controls or mice lacking PTEN, ACLY, or both. (F-G) Bcl-2 relative expression to actin in DP cells (F) and DN (G) sorted from 4 PTENΔ/Δ mutant, 3 ACLYΔ/Δ mutant, 4 PTENΔ/Δ;ACLYΔ/Δ double-mutant mice, or 3 flox controls. (H) H3K27Ac enrichment at the Bcl-2 P1 promoter in 3 PTENΔ/Δ mutant, 3 ACLYΔ/Δ mutant, 3 PTENΔ/Δ;ACLYΔ/Δ double mutants, and flox controls. Data represent percentages of input normalized to the actin promoter. (I-J) Sensitivity of DP cells to dexamethasone ex vivo (I) or in vivo (2 mg/kg per day for 3 consecutive days) (J) in PTENΔ/Δ mutant, ACLYΔ/Δ mutant, PTENΔ/Δ;ACLYΔ/Δ animals or flox controls (including Ptenflox2, Aclyflox2, and Ptenflox2;Aclyflox2 mice). IgG, immunoglobulin G; n.s., not significant; PBS, phosphate-buffered saline.

Figure 4.

Acly ablation abrogates apoptosis resistance in mutant DP progenitors. (A-B) Phospho ser473-AKT (A) and phospho ser455-ACLY (B) on DN, DP, SP4, and SP8 from preleukemic PTENΔ/Δ mutant or PTENflox2 controls. In mutant mice, carrying the ROSAYFPflox2 reporter, DN, DP, SP4, and SP8 cells are gated on YFP+ cells. Levels of BCL-XL (C-D) and BCL-2 (C,E) in DP cells from controls or mice lacking PTEN, ACLY, or both. (F-G) Bcl-2 relative expression to actin in DP cells (F) and DN (G) sorted from 4 PTENΔ/Δ mutant, 3 ACLYΔ/Δ mutant, 4 PTENΔ/Δ;ACLYΔ/Δ double-mutant mice, or 3 flox controls. (H) H3K27Ac enrichment at the Bcl-2 P1 promoter in 3 PTENΔ/Δ mutant, 3 ACLYΔ/Δ mutant, 3 PTENΔ/Δ;ACLYΔ/Δ double mutants, and flox controls. Data represent percentages of input normalized to the actin promoter. (I-J) Sensitivity of DP cells to dexamethasone ex vivo (I) or in vivo (2 mg/kg per day for 3 consecutive days) (J) in PTENΔ/Δ mutant, ACLYΔ/Δ mutant, PTENΔ/Δ;ACLYΔ/Δ animals or flox controls (including Ptenflox2, Aclyflox2, and Ptenflox2;Aclyflox2 mice). IgG, immunoglobulin G; n.s., not significant; PBS, phosphate-buffered saline.

Figure 5.

PTEN loss activates ACLY in human T-ALL cells to sustain cell growth. (A-C) PTEN, pAKTser473, and pACLYser455 (A-B) and acetyl-CoA levels (C) in human T-ALL cell lines transduced with a lentiviral vector carrying an shRNA PTEN or a control shRNA (shRNA Ctrl A). (D-E) Relative abundance of unlabeled and labeled metabolites mapped to glycolysis (D) and TCA (E) in DND41 cells exposed to [¹³C] glucose for 1 hour, 4 hours, or 24 hours. (F) Growth of DND41, RPMI-8402, and HBP-ALL transduced with a shRNA PTEN, a shRNA ACLY, or the combination of both and relative control shRNA (shRNA Ctrl A as control for the shRNA targeting PTEN and shRNA Ctrl B as control for the shRNA targeting ACLY). (G) Growth of Jurkat cells transduced with an shRNA ACLY or control shRNA in the presence or in the absence of doxycycline (dox)–mediated PTEN re-expression. aKG, α-ketoglutarate; PBS, phosphate-buffered saline; TCA, tricarboxylic acid; Unlab, unlabeled.

Figure 5.

PTEN loss activates ACLY in human T-ALL cells to sustain cell growth. (A-C) PTEN, pAKTser473, and pACLYser455 (A-B) and acetyl-CoA levels (C) in human T-ALL cell lines transduced with a lentiviral vector carrying an shRNA PTEN or a control shRNA (shRNA Ctrl A). (D-E) Relative abundance of unlabeled and labeled metabolites mapped to glycolysis (D) and TCA (E) in DND41 cells exposed to [¹³C] glucose for 1 hour, 4 hours, or 24 hours. (F) Growth of DND41, RPMI-8402, and HBP-ALL transduced with a shRNA PTEN, a shRNA ACLY, or the combination of both and relative control shRNA (shRNA Ctrl A as control for the shRNA targeting PTEN and shRNA Ctrl B as control for the shRNA targeting ACLY). (G) Growth of Jurkat cells transduced with an shRNA ACLY or control shRNA in the presence or in the absence of doxycycline (dox)–mediated PTEN re-expression. aKG, α-ketoglutarate; PBS, phosphate-buffered saline; TCA, tricarboxylic acid; Unlab, unlabeled.

Figure 5.

PTEN loss activates ACLY in human T-ALL cells to sustain cell growth. (A-C) PTEN, pAKTser473, and pACLYser455 (A-B) and acetyl-CoA levels (C) in human T-ALL cell lines transduced with a lentiviral vector carrying an shRNA PTEN or a control shRNA (shRNA Ctrl A). (D-E) Relative abundance of unlabeled and labeled metabolites mapped to glycolysis (D) and TCA (E) in DND41 cells exposed to [¹³C] glucose for 1 hour, 4 hours, or 24 hours. (F) Growth of DND41, RPMI-8402, and HBP-ALL transduced with a shRNA PTEN, a shRNA ACLY, or the combination of both and relative control shRNA (shRNA Ctrl A as control for the shRNA targeting PTEN and shRNA Ctrl B as control for the shRNA targeting ACLY). (G) Growth of Jurkat cells transduced with an shRNA ACLY or control shRNA in the presence or in the absence of doxycycline (dox)–mediated PTEN re-expression. aKG, α-ketoglutarate; PBS, phosphate-buffered saline; TCA, tricarboxylic acid; Unlab, unlabeled.

Figure 6.

Targeted metabolic and transcriptomic analysis of patients-derived T-ALL cells. (A) Principal component analysis of 8 PDXs undergoing targeted metabolomic analysis. (B) Most upregulated and downregulated metabolites in PTEN-altered cases. (C-D) Heat maps of the most 50 deregulated metabolites (C) and pathways enriched in PTEN-altered patients (D). (E) Metabolic analysis of citrate, CoA, acetyl-CoA, glutamine, glutamate, α-ketoglutarate, palmitic acid, stearic acid, and oleic acid in the 8 PDXs shown in panel A. (F) Venn diagram for metabolites mostly changed in both PTEN-mutant mice and PTEN-altered PDX (P < .05). (G) Metabolic transcriptome-based clustering of 155 patients. (H) Pathway enrichment analyses based on the differential metabolic gene expression of PTEN-altered (n = 33) vs wild-type (WT; n = 122) patients determined using EnrichR. (I) PROGENy score computation from whole transcriptomics data from 155 patients. The median score of the series was used to stratify patients into high- and low-score groups. (J) Volcano plot depicting differentially expressed metabolic genes in high vs low PI3K score. (K-L) Enriched metabolic gene sets enrichment in high vs low PI3K score as determined by GSEA. (M) Intracellular levels of cholesterol in WT or altered PI3K signaling T-ALL PDX blasts. Adj, adjusted; ADP, adenosine diphosphate; ALT, altered; CDP, cytidine diphosphate; FDR, false discovery rate; GSEA, gene set enrichement analysis; GDP, guanosine diphosphate; GTP, guanosine triphosphate; ns, not significant; PIP, phosphatidylinositol; tRNA, transfer RNA; UDP, uridine diphosphate.

Figure 6.

Targeted metabolic and transcriptomic analysis of patients-derived T-ALL cells. (A) Principal component analysis of 8 PDXs undergoing targeted metabolomic analysis. (B) Most upregulated and downregulated metabolites in PTEN-altered cases. (C-D) Heat maps of the most 50 deregulated metabolites (C) and pathways enriched in PTEN-altered patients (D). (E) Metabolic analysis of citrate, CoA, acetyl-CoA, glutamine, glutamate, α-ketoglutarate, palmitic acid, stearic acid, and oleic acid in the 8 PDXs shown in panel A. (F) Venn diagram for metabolites mostly changed in both PTEN-mutant mice and PTEN-altered PDX (P < .05). (G) Metabolic transcriptome-based clustering of 155 patients. (H) Pathway enrichment analyses based on the differential metabolic gene expression of PTEN-altered (n = 33) vs wild-type (WT; n = 122) patients determined using EnrichR. (I) PROGENy score computation from whole transcriptomics data from 155 patients. The median score of the series was used to stratify patients into high- and low-score groups. (J) Volcano plot depicting differentially expressed metabolic genes in high vs low PI3K score. (K-L) Enriched metabolic gene sets enrichment in high vs low PI3K score as determined by GSEA. (M) Intracellular levels of cholesterol in WT or altered PI3K signaling T-ALL PDX blasts. Adj, adjusted; ADP, adenosine diphosphate; ALT, altered; CDP, cytidine diphosphate; FDR, false discovery rate; GSEA, gene set enrichement analysis; GDP, guanosine diphosphate; GTP, guanosine triphosphate; ns, not significant; PIP, phosphatidylinositol; tRNA, transfer RNA; UDP, uridine diphosphate.

Figure 6.

Targeted metabolic and transcriptomic analysis of patients-derived T-ALL cells. (A) Principal component analysis of 8 PDXs undergoing targeted metabolomic analysis. (B) Most upregulated and downregulated metabolites in PTEN-altered cases. (C-D) Heat maps of the most 50 deregulated metabolites (C) and pathways enriched in PTEN-altered patients (D). (E) Metabolic analysis of citrate, CoA, acetyl-CoA, glutamine, glutamate, α-ketoglutarate, palmitic acid, stearic acid, and oleic acid in the 8 PDXs shown in panel A. (F) Venn diagram for metabolites mostly changed in both PTEN-mutant mice and PTEN-altered PDX (P < .05). (G) Metabolic transcriptome-based clustering of 155 patients. (H) Pathway enrichment analyses based on the differential metabolic gene expression of PTEN-altered (n = 33) vs wild-type (WT; n = 122) patients determined using EnrichR. (I) PROGENy score computation from whole transcriptomics data from 155 patients. The median score of the series was used to stratify patients into high- and low-score groups. (J) Volcano plot depicting differentially expressed metabolic genes in high vs low PI3K score. (K-L) Enriched metabolic gene sets enrichment in high vs low PI3K score as determined by GSEA. (M) Intracellular levels of cholesterol in WT or altered PI3K signaling T-ALL PDX blasts. Adj, adjusted; ADP, adenosine diphosphate; ALT, altered; CDP, cytidine diphosphate; FDR, false discovery rate; GSEA, gene set enrichement analysis; GDP, guanosine diphosphate; GTP, guanosine triphosphate; ns, not significant; PIP, phosphatidylinositol; tRNA, transfer RNA; UDP, uridine diphosphate.

Figure 7.

ACLY silencing in primary PTEN-null T-ALL cells affects their viability. (A) pACLYser455 levels in 8 PTEN/PI3KCA altered and 6 WT PDX T-ALL blasts. (B) Representative FACS plot depicting pAktser473 and pACLYser455 in a PTEN/PI3KCA WT patient or a PTEN-deleted patient. (C) Percentages of GFP⁺PTEN–altered PDX blasts remaining viable 48 hours after transduction with lentivirus vectors carrying shRNA ACLY or a control construct. (D) Competitive transplantation assay, in which PDX cells transduced with a shRNA ACLY or a shRNA control competed with nontransduced cells in a 20:80 ratio following injection into NSG mice. The relative ratio of transduced and not transduced cells found in the bone marrow of NSG mice is shown. (E) Number of GFP⁺PTEN germ line PDX blasts remaining viable 3, 4, or 5 days after transduction. BM, bone marrow; FACS, fluorescence-activated cell sorting; GFP, green fluorescent protein.