HSC cell cycling increases overall glycolytic flux, but not flux into mitochondria.

(A) Experimental design used for glucose isotope tracer analysis in HSCs from 5-FU- or PBS-treated mice. (B) Heat map of metabolite levels in HSCs derived from mice treated with PBS or 5-FU. (C-F) The semi-quantitative value (10-6 µM) of U-13C6-glucose-derived metabolites in glycolysis (C), the first round of TCA cycle (D), the PPP, and nucleotide synthesis (F) in HSCs from 5-FU- or PBS-treated mice (PBS group = 1.0); In (B)-(F), biological replicates from the PBS and 5-FU groups, obtained on three separate days, were pooled, analyzed by IC-MS, quantified based on calibration curve data for each metabolite (see “Ion chromatography mass spectrometry (IC-MS) analysis” section in “Methods” for details). (G–H) A Mito Stress test with the Seahorse flux analyzer on HSCs derived from mice treated with PBS or 5-FU; ECAR (G) and OCR (H) before and after oligomycin treatment. (Data were obtained from n = 7 technical replicates for PBS-treated HSCs and n = 6 for 5-FU-treated HSCs.) (I) Experimental schema of in vivo 2-NBDG analysis. (J) Representative histograms of 2-NBDG analysis (gray: no 2-NBDG, red: PBS group, blue: 5-FU group). (K) 2-NBDG positivity in each fraction; data represent four pooled biological replicates for the PBS group and three for the 5-FU group; MyP: myeloid progenitor. (L) EPCR expression and 2-NBDG positivity within HSC fractions. Data were extracted from each individual in (K).

Data are presented as mean ± SD. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 as determined by Student’s t-test (C–F, G–H when comparing the PBS and 5-FU groups, and K–L) or paired-samples t-test (G– H when comparing the conditions before and after exposure to oligomycin within the PBS/5-FU group). Abbreviations: G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; F1,6BP, fructose-1,6-bisphosphate; G3P, glycerol-3-phosphate; DHAP, dihydroxyacetone phosphate; 3PG, 3-phosphoglycerate; 2PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate; PYR, pyruvate; LAC, lactate; Ac-CoA; acetyl-CoA; CIT, citrate; ACO, cis-aconitic acid, isocitrate; 2OG, 2-oxoglutarate; SUC, succinate; FUM, fumarate; MAL, malate; OAA, oxaloacetate; 6PG, glucose-6-phosphate; Ru5P, ribulose-5-phosphate; Xu5P, xylulose-5-phosphate; R5P, ribose-5-phosphate; S7P, sedoheptulose-7-phosphate; E4P, erythrose-4-phosphate; PRPP, phosphoribosyl pyrophosphate; IMP, inosine monophosphate; ATP, adenosine triphosphate; GTP, guanine triphosphate; UMP, uridine monophosphate; UTP, uridine triphosphate; TTP, thymidine triphosphate. See also Fig. S1-3.

OXPHOS inhibition activates compensatory glycolysis in HSCs.

(A) Experimental design used for glucose isotope tracer analysis in HSCs treated with the OXPHOS inhibitor oligomycin. (B) Heat map of metabolite levels detected by in vitro tracer analysis of U-13C6-glucose in HSCs treated with DMSO or oligomycin (Oligo). (C-F) Relative amounts of U-13C6-glucose-derived metabolites in glycolysis (C), the first round of TCA cycle (D), the PPP(E), and nucleotide synthesis (F) in DMSO-(black) or oligomycin-treated (orange) HSCs; In (B)-(F), biological replicates of the DMSO and oligomycin groups obtained on four separate days were pooled, analyzed by IC-MS, and quantified based on calibration curve data for each metabolite (see “Ion chromatography mass spectrometry (IC-MS) analysis” section in “Methods” for details). (G-H) Mito Stress test on the Seahorse flux analyzer for HSC and MyPs; ECAR (G) and OCR (H) before and after oligomycin treatment. (Data were obtained from n = 3 technical replicates for HSCs and n = 10 technical replicates for MyPs.)

Data are shown as mean ± SD. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 as determined by paired-samples t-test (C-E and G–H). Abbreviations: G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; F1,6BP, fructose-1,6-bisphosphate; G3P, glycerol-3-phosphate; DHAP, dihydroxyacetone phosphate; 3PG, 3-phosphoglycerate; 2PG, 2-phosphoglycerate; PEP, phosphoenolpyruvate; PYR, pyruvate; LAC, lactate; Ac-CoA; acetyl-CoA; CIT, citrate; ACO, cis-aconitic acid, isocitrate; 2OG, 2-oxoglutarate; SUC, succinate; FUM, fumarate; MAL, malate; OAA, oxaloacetate; 6PG, glucose-6-phosphate; Ru5P, ribulose-5-phosphate; Xu5P, xylulose-5-phosphate; R5P, ribose-5-phosphate; S7P, sedoheptulose-7-phosphate; E4P, erythrose-4-phosphate; PRPP, phosphoribosyl pyrophosphate; IMP, inosine monophosphate; ATP, adenosine triphosphate; GTP, guanine triphosphate; UMP, uridine monophosphate; UTP, uridine triphosphate; TTP, thymidine triphosphate. See also Fig. S1–3.

Quantitative 13C-MFA of quiescent, proliferative, and stressed HSCs

(A-C) Overview of quantitative 13C-MFA of PBS-treated HSCs (A), 5-FU-treated HSCs (B), and OXPHOS-inhibited HSCs (C). The representative net flux for each reaction with glucose uptake as 100 is shown in the squares below the catalytic enzymes for each reaction listed in green letters. Red arrows indicate reactions with particularly elevated fluxes and blue arrows indicate reactions with particularly decreased fluxes. (D) Heatmap of the relative flux of each enzyme in the 5-FU or oligomycin groups compared to that in the quiescent (Ctl) HSC (The metabolic flux of each enzyme in the Ctl group was standardized as 100.). (E-J) Fluxes due to reactions with PFK (E, H), G6PD (F, I), and PDH (G, J). Fluxes of HSCs derived from mice treated with 5-FU (blue bars) or PBS (red bars) (D-F) and of HSCs treated with DMSO (black bars) or oligomycin (orange bars) (G-I) are shown. Data is obtained from 100 simulations in OpenMebius, and flux data for each enzyme is displayed (Table S4). (K-L) Ratio of fructose 1,6-bisphosphate (F1,6BP) to fructose-6-phosphate (F6P) calculated from tracer experiments shown in Figure 1B and Figure 2B. Effects of 5-FU administration (K) or mitochondrial inhibition by oligomycin (L) are summarized. Data are shown as mean ± SD. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 as determined by Student’s t-test (E-L). Abbreviations: HK, hexokinase; PGI, glucose-6-phosphate isomerase; PFK, phosphofructokinase; TPI, triose phosphate isomerase; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PGM, phosphoglycerate mutase; PK, pyruvate kinase; LDH, lactate dehydrogenase; PC, pyruvate carboxylase; PDH, pyruvate dehydrogenase; CS; citrate synthase; IDH, isocitrate dehydrogenase; αKGDH, α-ketoglutaric acid dehydrogenase; SDH, succinate dehydrogenase; G6PD, glucose-6-phosphate dehydrogenase; TAL, transaldolase. See also Fig. S4.

Pfkfb3 activates the glycolytic system in proliferating HSCs

(A) Experimental design used to conduct real-time ATP analysis of HSCs treated with 5-FU or PBS. PLFA medium containing mitochondrial substrates (pyruvate, lactate, fatty acids, and amino acids) but no glucose, was used for experiments with 2-DG; Ba-M containing neither mitochondrial substrates nor glucose was used for experiments with oligomycin, Pfkfb3 inhibitor, or AMPK inhibitor. (B-I) Results of real-time ATP analysis of PBS-(red) or 5-FU-treated (blue) HSCs after treatment with 2-DG (B, D), oligomycin (C, E), PFKFB3 inhibitor (F, H), or AMPK inhibitor (G, I). Bar graphs show corrected ATP concentrations for the last 2 min (D) of (B), 6–7 min (E) of (C), or the last 1 min (H, I) of (F, G) for PFKFB3 and AMPK inhibitors, respectively. Each group represents at least 60 cells. Data are representative results of pooled samples from three biological replicates. (see “Time-course analysis of FRET values” in “Methods” for details of the correction method used to calculate ATP concentration.) (J) Normalized mRNA counts of PFKFB isozymes based on the RNA sequencing of HSCs.

Data are presented as mean ± SD. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 as determined by Student’s t-test (D, E, H, and I) or a one-way ANOVA followed by Tukey’s test (J). See also Fig. S6–7.

PFKFB3 accelerates glycolysis in HSCs under OXPHOS inhibition in an AMPK-dependent manner

(A) Experimental design of real-time ATP analysis using GO-ATeam2 knock-in BMMNCs. Ba-M was used in experiments with oligomycin. For other experiments, PLFA medium was used. (B-C) Evaluation of factors affecting ATP concentration in HSCs (B) and GMPs (C) based on the GO-ATeam2 system. GO-ATeam2 knock-in BMMNCs were incubated with glucose, oligomycin, 2-DG, or glucose plus oligomycin, and the FRET/EGFP ratio was calculated. (D) ATP concentration in indicated stem/progenitor fractions in PLFA medium (red bars) alone or PLFA medium plus 2-DG (blue bars). ATP concentration for the last 2 min of the analysis time is shown. Data is summarized from Figure 5B-C and Supplemental Figure 5A-D. Each group represents at least 110 cells. Data are representative results of pooled samples from three biological replicates. (E) ATP concentration in indicated stem/progenitor fractions in Ba-M plus glucose (dark blue bars) or Ba-M plus glucose and oligomycin (orange bars). ATP concentration for the last 1 min of the analysis period is shown. Data is summarized from Figure 5B-C and Supplemental Figure 5A-D. Each group represents at least 43 cells. Data are representative results of pooled samples from three biological replicates. (F-I) Effects of PFKFB3 or AMPK inhibitors (PFKFB3i or AMPKi, respectively) on ATP concentration in HSCs from GO-ATeam2 mice in Ba-M plus glucose only (F) or Ba-M plus glucose and oligomycin (G). ATP concentrations for the last 1 min of the analysis period are shown in (H) and (I) for glucose only and glucose with oligomycin groups, respectively. Each group represents at least 90 cells. Data are representative results of pooled samples from three biological replicates. (J) Experimental schema for cell cycle assay and real-time ATP concentration analysis after overexpression of Pfkfb3. (K) Cell cycle status of Pfkfb3-overexpressing (Pfkfb3OE) and mock-transduced HSCs. (L-M) Effects of inhibitors on ATP concentration in Pfkfb3-overexpressing GO-ATeam2+ HSCs. Cells were exposed to vehicle or 2-DG (L), oligomycin in the presence or absence of glucose 12.5 mg/dL (M), and ATP concentrations for the last 2 min (L) or 1 min (M) of the analysis period were calculated. Data are representative results of pooled samples from three biological replicates.

Data are presented as mean ± SD. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 as determined by Student’s t-test (D, E, and K) or one-way ANOVA followed by Tukey’s test (H, I, L, and M). See also Fig. S8.

PFKFB3 methylation by PRMT1 enables ATP production by cell-cycling HSCs.

(A) Normalized Pfkfb3 mRNA counts based on RNA sequencing of PBS-treated (red) or 5-FU-treated (blue) HSCs. Data are representative results of pooled samples from three biological replicates. Data were extracted from the same pooled samples as in Figure 4J and Figure S9. (B) Quantification of mean fluorescent intensity (MFI) of PFKFB3 protein in PBS- or 5-FU-treated HSCs. The lower part of the graph shows representative images of immunocytochemistry of PFKFB3 in each group. n = 26–27 single HSCs for each group. The data are representative results from two independent experiments. (C) Quantification of MFI of phosphorylated-PFKFB3 (p-PFKFB3) protein in PBS- or 5-FU-treated HSCs. The lower part of the graph shows representative images of immunocytochemistry of p-PFKFB3 in each group. n = 27 single HSCs for each group. The data are representative results from two independent experiments. (D) Quantification of mean fluorescence intensity (MFI) of p-PFKFB3 in HSCs treated with glucose (200mg/dL); glucose plus oligomycin (1 µM); and glucose, oligomycin, and dorsomorphin (100 µM) for 5 min. The lower part of the graph shows representative images of immunocytochemistry of p-PFKFB3 in each group. n = 32–36 for each group. The data are representative results from two independent experiments. (E) Normalized Prmt1 mRNA counts based on RNA sequencing of PBS-treated (red) or 5-FU-treated (blue) HSCs. Data are representative results of pooled samples from three biological replicates. (F) MFI quantification of methylated-PFKFB3 (m-PFKFB3) in PBS- or 5-FU-treated HSCs. The lower part of the graph shows representative images of immunocytochemistry of m-PFKFB3 in each group. n = 23–41 for each group. The data are representative results from three independent experiments. (G) Quantification of MFI of m-PFKFB3 in PBS- or 5-FU-treated HSCs or 5-FU-treated HSCs after 15 min treatment with a PRMT1 inhibitor (90 μg/mL GSK3368715); n = 25–35 single HSCs for each group. The lower part of the graph shows representative images showing immunocytochemistry of m-PFKFB3. Data represent a single experiment. (H) Quantitation of m-PFKFB3 in NBDG-positive or -negative HSCs in mice treated with PBS or 5-FU. The lower part of the graph shows representative images of immunocytochemistry of m-PFKFB3 in each group. n = 28–41 for each group. The data are representative results from two independent experiments. (I) Corrected ATP levels in PBS-(red) or 5-FU-treated (blue) HSCs 15 min after treatment with vehicle or a PRMT1 inhibitor (90 µg/mL GSK3368715). Each group represents at least 101 cells. Data are representative results of pooled samples of two biological replicates. (see “Time-course analysis of FRET values” in “Methods” for details of the correction method used to calculate ATP concentration.) (J) ATP concentration in mock-transduced (Ctl) or PFKFB3-overexpressed (OE) HSCs after treatment with the PRMT1 inhibitor (90 µg/mL GSK3368715). ATP concentration for the last 1 min of the analysis period is shown. Data are presented as mean ± SD. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 as determined by Student’s t-test (A-C, E-F, and I-J) or one-way ANOVA followed by Tukey’s test (D, G, and H). See also Fig. S9.

PFKFB3 maintains HSC function under proliferative stress.

(A-C) Transplant analysis of Pfkfb3-KO or Pfkfb3CA-overexpressing HSCs. Experimental design (B). PB chimerism of donor-derived cells at 4 months post-transplant. Pfkfb3-KO group, n = 6; Rosa-KO group, n = 4; (B) Pfkfb3 group, n = 5; pMY-IRES-GFP group, n = 4. (C) The data are representative results from two independent experiments. (D-E) 5-FU administration after bone marrow reconstruction with Pfkfb3- or Rosa-KO HSPCs. Experimental schema (D). Behavior of the Pfkfb3- or Rosa-KO cells in PB after 5-FU administration (E). n = 5 for each group. (F-K) Cell cycle analysis and apoptosis assay of Pfkfb3- or Rosa-KO HSPCs on day 2 post-BMT. Experimental schema (F). Representative plots of Ki67/Hoechst33432 staining of Rosa-KO (G) or Pfkfb3-KO (H) HSPCs and summary of analysis (I); summary of in vivo BrdU labeling assay (J). Apoptosis assay results (K). n = 4–5 biological replicates for each group. (L-N) Cell cycle analysis of Pfkfb3CA or Mock-overexpressing HSPCs on day 2 after BMT. Experimental Schema (L). Representative plot of Ki67/Hoechst33432 staining for both groups (M) and summary of analysis (N). n = 5 biological replicates for each group. (O) Models showing ATP production and regulation in quiescent, OXPHOS-inhibited, and cell-cycling HSCs. Note that the GO-ATeam2 system identified plastic acceleration of glycolysis by PFKFB3 in response to different types of stress maintains ATP levels. Data are presented as mean ± SD. * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 as determined by Student’s t-test (B, C, E, H, J, K, and N). See also Fig. S10.