Metabolomic insights into pathogenesis and therapeutic potential in adult acute lymphoblastic leukemia

Abstract

Acute lymphoblastic leukemia (ALL) poses challenges in adult patients, considering its heterogeneous nature and often suboptimal treatment outcomes. Here, we performed a study on 201 newly diagnosed adult ALL cases (age ≥ 15 y) to generate intracellular and dynamic serum metabolomic profiles. Our findings revealed a predominant increase in bile acid (BA) metabolites in serum, alongside metabolic rewiring that supported highly proliferative states and actively metabolic signaling, such as enriched nucleotide metabolism in leukemic blasts. By integrating intracellular metabolomics and transcriptomics, we constructed the Comprehensive Metabolic Information Dataset (CMID), which facilitated the development of a clustering system to supplement current risk stratification. Furthermore, we explored potential metabolic interventions targeting the serum BA profile and energy metabolism in blasts. The combined use of simvastatin with vincristine and dexamethasone regimen demonstrated a synergistic therapeutic effect in a murine ALL model, effectively lowering key BA levels in serum and suppressing the infiltration of leukemic blasts in the liver. In light of the enhanced intracellular redox metabolism, combining FK866 (a nicotinamide phosphoribosyltransferase inhibitor) and venetoclax significantly prolonged survival in a patient-derived xenograft ALL model. Our findings, along with the resulting resources (http://www.genetictargets.com/MALL), provide a framework for the metabolism-centered management of ALL.

Keywords: adult acute lymphoblastic leukemia; comprehensive metabolic information dataset; disease risk stratification.

Fig. 1.

Construction of a multiomics landscape of the adult ALL cohort centering on the metabolome. (A) Overview of the experimental design and the number of samples for metabolome and RNA-seq analyses. * Among the 106 samples subject to intracellular metabolome investigation on D0 of induction therapy, BMMC samples were used in 100 cases and PBMC samples were available in 6 cases. (B) Number of annotated metabolites in serum and intracellular samples, grouped by metabolites functional category, mainly based on KEGG. (C) (Middle Panel) The heatmap provides a comprehensive overview of the metabolome in BCP-ALL and T-ALL patients at specified time points, further grouped by D28 status. The entire metabolome pattern is categorized into nine clusters, with significantly enriched pathways for clustered metabolites annotated on the Right (P < 0.05). Each row represents a metabolite, each column corresponds to a sample, and each score reflects its relative abundance compared to HC levels, calculated as the median within each subgroup. (Left Panel) The correlation between each cluster and clinical parameters assessed on Day 0 (D0) is displayed (P-value < 0.01). PI: postinduction phase; UA: uric acid; CREA: creatine; HDL: high-density lipoprotein; PAB: Prealbumin; AST: Aspartate aminotransferase; ALP: Alkaline phosphatase; WBC: White blood cell count; TBIL: Total bilirubin; GGT: Gamma-glutamyl transferase; BUN: Blood urea nitrogen; PLT: Platelet count; DBIL: Direct bilirubin; TP: Total protein; ALB: Albumin; CHOL: Cholesterol; GLU: Glucose; Hb: Hemoglobin.

Fig. 2.

Intracellular metabolome reveals aberrant metabolic profiles underlying the pathogenesis of ALL. (A) PCA of intracellular samples from all groups (n = 127) reveals a plane (depicted schematically in gray) that distinguishes most samples of the ALL-D0 groups from the control (HC PBMC and ALL BMMC-MRD-) groups. (B) The DA score reveals the enrichment of metabolism pathways (each row) linked to ALL-associated metabolites, and captures the average gross changes for all metabolites in a pathway. A score of 1 or −1 indicates that all measured metabolites in the pathway are increased or decreased in the ALL compared to HC samples. Activity levels of pathway are color coded: pink for upregulated, green for downregulated, otherwise colored in brown. (C) The intracellular metabolome landscape reveals that most disturbed metabolites were upregulated in ALL. The columns represent samples, and rows represent metabolites. ETP: Early T cell precursor; FMN: Flavin Mononucleotide; DHAP: Dihydroxyacetone Phosphate; NMN: Nicotinamide Mononucleotide; F-1,6-BP: Fructose 1,6-bisphosphate; G-1-P: Glucose 1-phosphate; sn-Gly-3-P: sn-Glycero-3-phosphate. (D) A simplified schematic plot illustrates the globally upregulated metabolism pathways in ALL blasts. (E) A specific comparison between BCP-ALL and T-ALL elucidates the pathways enriched with significantly differential metabolites. The colors represent metabolic pathway categories and the size of each circle reflects the number of annotated metabolites in indicated pathway.

Fig. 3.

Relationship between intracellular metabolome, transcriptome, and genomic features in BCP-ALL and T-ALL. (A) Gene set variation analysis (GSVA) of enrichment score for metabolic pathways in individual BCP-ALL and T-ALL samples at the RNA-seq level. Each subtype exhibits both individuality and potential commonality. (B) The heatmap illustrates the associations between the metabolites abundance (column) of ALL BMMC and the presence of mutations within the indicated genes (row). The mutations include high frequency somatic mutations within reported ALL-related genes. T statistics are calculated by a linear regression model to represent the correlation level. (C) Correlations between IDH1/2 mutations and 2-HG (Left panel), and ETV6 mutations and glutamine (Right panel). All samples were ordered based on the abundance (y-axis) of indicated metabolites, and the ones with corresponding mutations were highlighted in red and indicated by the corresponding lines displayed in the x-axis. (D) The heatmap illustrates the associations between the abundance of metabolites and the presence of indicated well-known gene fusions. (E) Correlations between the BCR::ABL1 fusion and the abundance of the indicated metabolites are shown. Fusion genes were shown as lines and samples were ordered based on the abundance (y-axis) of indicated metabolites.

Fig. 4.

Metabolic information yields an independent prognostic clustering classification for BCP-ALL and T-ALL. (A) The schematic workflow following SNF integration analysis. The BMMC omics data were inputted to generate a fused similarity network, which was subsequently utilized for sample clustering in this study. (B) The discrimination of three BCP-ALL clusters (BC1, BC2, and BC3) generated from SNF method. (C) The association of the three metabolic clusters with clinical OS in 75 BCP-ALL patients. (D) The multivariate Cox analysis reveals that the metabolic cluster bore an independent prognostic value for OS and EFS, respectively. The WBC was stratified by 30 × 10⁹/L. In the genetic-based risk stratification, BCR::ABL1, ETV6::RUNX1, DUX4::IGH, and TCF3::PBX1 were categorized into standard risk subgroup, otherwise were into poor risk subgroup. (E) A core metabolism-based subnetwork that best explains the difference between BC3 and BC1/2. In this network, each circular node represents a metabolite, labeled with its KEGG ID, and each square node represents a gene. Node colors indicate logFC values: red and green signify different gene expression levels, while yellow and blue signify varying metabolite levels. The highlighted labels correspond to pathways in the cholesterol metabolism and nicotinate and nicotinamide metabolism. (F) The discrimination of two T-ALL clusters (TC1 and TC2) based on SNF method. (G) The association of the two metabolic clusters with clinical OS outcomes in 24 T-ALL patients. The result suggests TC1 is of significantly poorer prognosis than TC2. (H) A core metabolism-based subnetwork that best explains the difference between TC2 and TC1. The highlighted labels correspond to pathways in the TCA and OXPHOS. PC refers to a group of phosphatidylcholines without specific KEGG IDs. The P-value was calculated from the log rank test.

Fig. 5.

Enhanced survival in MH/N ALL with simvastatin and VCR/DEX combination therapy. (A) Illustration of MH/N ALL mouse model and experimental design. (B) The differential metabolism pathways between serum of MH/N ALL mice and normal C57BL/6 J controls. (C) Levels of taurocholic acid (T-CA), tauroursodeoxycholic acid (T-UDCA), and glycocholic acid (G-CA) in serum collected from MH/N ALL mice (Day 14 post transplantation) and normal C57BL/6 J counterparts were determined. (D) Survival curves of MH/N ALL mice treated with solvent (vehicle), VCR [0.15 mg/kg, intraperitoneal injection (IP)] plus DEX (1 mg/kg, IP) (VCR/DEX), and the combination of T-CA [100 mg/kg, oral gavage (PO)] and VCR/DEX, respectively. The median survival was 13, 20, and 18 d, respectively. (E) Survival curves of MH/N ALL mice treated with solvent (vehicle), simvastatin (Simva) (50 mg/kg, PO), VCR/DEX, and the combination of Simva+VCR/DEX, respectively. The median survival was 14, 15, 19.5, and 25 d, respectively. (F) Levels of T-CA, T-UDCA, and G-CA in serum collected from MH/N ALL mice on Day 14 post transplantation were determined. (G) Representative Hematoxylin and eosin-stained liver sections from MH/N ALL mice treated with indicated drugs. Scale bar, 500 µm and 50 µm, respectively. (H) The percentage of liver weight (Left), and spleen weight (Right) relative to the whole-body weight of each MH/N ALL mouse from the indicated groups. (I) Engraftment of GFP-positive MH/N ALL cells in PB, liver, spleen, and BM was assessed on Day 18. (J and K) The frequencies of infiltrated CD3e+ T cells and CD11b+ myeloid cells in the CD45+ cells of the BM (J), and liver (K) are shown, respectively. (L) Survival curves of KMT2A rearranged (KMT2Ar) ALL patient-derived xenograft (PDX) mice treated with indicated treatment for 3 cycles of 5-d-on/2-d-off. Data are shown as mean ± SEM; the two-tailed Mann–Whitney test was performed for continuous variables, and the log rank test was performed for survival analysis. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 and ns denotes nonsignificant comparison.

Fig. 6.

FK866 with venetoclax significantly improved the survival of KMT2Ar ALL PDX mice. (A) Seven NAMPT inhibitors targeting nicotinamide metabolism were screened out as hits for exceeding 90% cell viability suppression in most of ALL cell lines at single concentration of 1 µM. (B) The sensitivity of each NAMPTi on ten ALL cell lines was validated using normalized IC₅₀ values interpreted by drug sensitivity assay. (C) Schematic illustration of FK866 (an NAMPTi) capable of specifically inhibiting the NAD+ biogenesis and subsequently suppressing the TCA cycle, OXPHOS, and ETC biological processes. (D) Dose–response curves of SEM (Left) and NALM6 (Right) cells showing cell viability following treatment with FK866 or FK866 in combination with nicotinamide mononucleotide (NMN, 10 µM) for 72 h. (E) The ratio of NAD+/NADH levels in SEM (Left) and NALM6 (Right) cells treated with FK866 (10 nM) or FK866 with nicotinic acid (NA, 10 µM) for 24 h. (F) Volcano plot showing represented differential metabolites of SEM treated with DMSO or FK866 (100 nM) for 24 h. (G) Event-free survival curves of KMT2Ar ALL PDX mice treated with solvent (vehicle), VCR/DEX, FK866 (30 mg/kg, IP), venetoclax (Ven) (100 mg/kg, PO), and the combination of FK866 and Ven, respectively, for 4 cycles of 5-d-on/2-d-off. The median survival was 40, 46, 53, 49, and 67 d, respectively. (H) hCD45+ cells of PB WBC counts in the KMT2Ar ALL PDX mice with indicated treatment at the end of treatment.