Fig 2: TKT silencing results in the loss of CLL cell fitness and delay of leukemia progression in vivo. (A) Representative immunoblot of TKT in Cd4‐edited and Tkt‐edited TCL1 cells cultured for 4 days. (B) Total counts of live Cd4‐edited and Tkt‐edited TCL1 cells expanded in vitro in co‐cultures with irradiated CD40L‐m3T3 cells for 5 days. (C) Workflow of transfer experiments in which CRISPR/Cas9‐based edited TCL1 cells are injected into a healthy recipient and analyzed before and after their transfer. (D) Percentage of CD5+CD19+ leukemic cells in the blood of mice injected with Cd4‐edited (n = 8 mice) and Tkt‐edited TCL1 cells (n = 9 mice). Results display two independent experiments (left and right panels, respectively). Each dot represents one animal. (E) Measurements of spleen size in mice injected with Cd4‐edited (n = 8 mice) and Tkt‐edited TCL1 cells (n = 9 mice). Results display two independent experiments (left and right panels, respectively). Each dot represents one animal. (F) Representative immunoblot and (H) quantification of TKT levels in Cd4‐ and Tkt‐edited TCL1 cells before injection, and in the corresponding tumors. Each dot represents one animal. (I) Representative Sanger sequencing histograms of an amplified Tkt locus in edited TCL1 cells and resulting tumors (left panel). Indel frequency in the Tkt locus based on Inference of CRISPR editing (ICE) analysis (right panel). Each dot represents one animal. (H) Representative Sanger sequencing histograms of an amplified Cd4 locus in edited TCL1 cells and resulting tumors (left panel). Indel frequency in the Cd4 locus based on Inference of CRISPR editing (ICE) analysis (right panel). The Cas9 cutting site is indicated and the Tkt sgRNA sequence is underlined. Each dot represents one animal. Data in the figure represent mean ± SD of the indicated individual values. In panels (B, C, D, F, G, and H), P values were generated using the Mann–Whitney U test.

Fig 3: Upregulation of G6PD and TKT expression occurs via the IL‐21–STAT3 axis. (A) Representative immunoblots (left panel) and their quantification (right panel) for G6PD in primary CLL cells (n = 6) stimulated for 3 days with different cytokines. “CD40L” denotes hCD40L‐expressing 3T3 cells. (B) Representative immunoblots (left panel) and their quantification (right panel) for TKT in primary CLL cells (n = 7 donors) stimulated for 3 days with different cytokines. (C) RNA sequencing data analysis of STAT3 in proliferating CLL cells compared to resting controls (n = 5 donors). The P value was generated by DeSeq. 2. (D) Representative immunoblots and (E) their quantification for STAT3, G6PD, and TKT in primary CLL cells (n = 6–9) stimulated for 3 days with different cytokines after STAT3 silencing. CD4‐edited cells were used as controls. Details of the clinical features of CLL donors are reported in Table S3. Data are represented as mean ± SD of the indicated individual values. (F) Total counts of live cells (left panel) and percentage of proliferating cells (right panel) in CLL cell cultures (n = 5–9 donors). Cells were genetically modified as reported in panels (D, E) and cultured with irradiated CD40L‐3T3 cells, IL‐4, and IL‐21 for 5–6 days. P values were generated using the Mann–Whitney U test. (G) RNA sequencing data analysis of NFE2L2 (encoding NRF2) in proliferating CLL cells compared to resting controls (n = 5 donors). The P value was generated by DeSeq. 2. (H) Graphical summary of IL‐21‐mediated regulation of STAT3 levels and PPP enzyme expression in CLL cells. Immunoblots (A, B, E) are normalized to the condition stimulated with CD40L and IL‐21. Each dot represents one donor.

Fig 4: IL‐21 regulates the expression of G6PD and TKT in CLL cells but not in healthy B cells. Representative immunoblots (A) for G6PD and for (B) TKT in primary CLL cells (n = 9–13 donors) stimulated for 3 days as indicated. “CD40L” denotes hCD40L‐expressing 3T3 cells. (C, D) Quantification of immunoblots for G6PD and TKT in primary CLL cells (n = 9–13 donors) stimulated for 3 days as indicated. Data are normalized to the condition stimulated with CD40L, αIgM (BCR‐crosslinking antibodies), IL4, and IL‐21. (E) Analysis of IL21R expression from a published RNA sequencing analysis. 12 CLL cells were cultured for 48 h with 3T3 cells (Ctrl) or hCD40L‐3T3 cells (CD40L). The P value was generated using the Mann–Whitney U test. (F) Example of an immunoblot (left panel) and quantification of immunoblots (n = 6 donors, right panel) detecting G6PD protein levels stimulated for 3 days with different stimuli combinations. In the right panel, data were normalized to the condition stimulated with CD40L. (G) Example of an immunoblot (left panel) and quantification of immunoblots (n = 6 donors, right panel) detecting TKT protein levels stimulated for 3 days with different stimuli combinations. In the right panel, data were normalized to the condition stimulated with CD40L. HD, healthy donor. Details of the clinical features of CLL donors are reported in Table S3. Data are represented as mean ± SD of the indicated individual values. P values were generated using the Mann–Whitney U test. Each dot represents one donor.

Fig 5: PPP enzyme expression and activity are upregulated in proliferating CLL cells. (A) Schematic representation of glycolysis and glycolysis branching pathways. Solid arrows represent direct metabolic reactions; dashed arrows represent a series of reactions or the beginning of a glycolysis‐related pathway. Metabolite abbreviations: 3PG, 3‐phopshoglycerate; F6P, fructose 6‐phosphate; F1,6BP, fructose 1,6‐bisphosphate; G6P, glucose 6‐phosphate; PEP, phosphoenolpyruvate. Enzyme abbreviations: GPI, glucose 6‐phosphate isomerase; HK, hexokinase; PFK1, phosphofructokinase 1. (B) RNA sequencing data analysis showing enzymes of glycolysis and glycolysis‐branching pathways expressed in proliferating CLL cells compared to nonstimulated cells. Cells were cultured for 4 days in HPLM with 5 mM glucose, CD40L‐3T3 cells, αIgM (BCR‐crosslinking) antibodies, IL‐4, and IL‐21 (n = 4 donors). Gene expression is reported as log2 FPKM for genes detected as expressed (mean log2 FPKM > 0). Differentially expressed genes (Table S2) are marked with asterisks. Rate‐limiting enzymes of the pathways are highlighted in bold. (C) Expression values of PPP enzymes in n = 22 matched CLL cell samples sorted from the PB or the LN niche (LN). 14 Gene expression is reported as FPKM. Data are represented as box and whiskers depicting the median and min‐to‐max range of the individual values. P values were generated by DeSeq. 2. Rate‐limiting enzymes of PPP branches are highlighted in bold. (D) Schematic representation of oxidative and nonoxidative branches of PPP. Solid arrows represent direct metabolic reactions; dashed arrows represent a series of reactions. Metabolite abbreviations: F1,6BP, fructose 1,6‐biphosphate; F6P, fructose 6‐ phosphate; G3P, glyceraldehyde 3‐phosphate; G6P, glucose 6‐phosphate; R5P, ribose 5‐phosphate; Ru5P, ribulose 5‐phosphate; S7P, sedoheptulose 7‐phosphate; X5P, xylulose 5‐phosphate. Enzyme abbreviations: G6PD, glucose 6‐phosphate dehydrogenase; 6PGL (PGLS), 6‐phosphogluconolactonase; 6PGD (PDG), 6‐phosphogluconate dehydrogenase; RPE, ribulose 5‐phosphate epimerase; RPIA, ribulose 5‐phosphate isomerase; TALDO (TALDO1), transaldolase; TKT, transketolase. (E) Uniform Manifold and Projection (UMAP) plot for PBMCs of case 19 from the published study 15 based on the scRNA‐seq data colored by annotation. Only clusters expressing B cell markers are shown. (F) Expression of PPP genes in the two selected UMAP clusters identified in panel (E). (G) Kaplan–Meier analysis of the impact of PPP enzyme expression on failure‐free survival (FFS) in 502 treatment‐naive patient samples. 16 FFS was calculated for treatment‐naive patients as the time from the date of the sequenced sample to the date of first treatment, progression, or death, and censored at the last known event‐free date. An independent analysis was performed for each gene. P values were generated using the log‐rank test.