Mitochondrial pyruvate metabolism and glutaminolysis toggle steady-state and emergency myelopoiesis

Abstract

To define the metabolic requirements of hematopoiesis, we examined blood lineages in mice conditionally deficient in genes required for long-chain fatty acid oxidation (Cpt2), glutaminolysis (Gls), or mitochondrial pyruvate import (Mpc2). Genetic ablation of Cpt2 or Gls minimally impacted most blood lineages. In contrast, deletion of Mpc2 led to a sharp decline in mature myeloid cells and a slower reduction in T cells, whereas other hematopoietic lineages were unaffected. Yet MPC2-deficient monocytes and neutrophils rapidly recovered due to a transient and specific increase in myeloid progenitor proliferation. Competitive bone marrow chimera and stable isotope tracing experiments demonstrated that this proliferative burst was progenitor intrinsic and accompanied by a metabolic switch to glutaminolysis. Myeloid recovery after loss of MPC2 or cyclophosphamide treatment was delayed in the absence of GLS. Reciprocally, MPC2 was not required for myeloid recovery after cyclophosphamide treatment. Thus, mitochondrial pyruvate metabolism maintains myelopoiesis under steady-state conditions, while glutaminolysis in progenitors promotes emergency myelopoiesis.

Graphical abstract

Figure 1.

Long-chain fatty acid oxidation and glutaminolysis are dispensable for most hematopoietic lineages. (A) Schematic representation of relevant metabolic pathways. Pyruvate enters the mitochondria through the MPC, composed of subunits MPC1 and MPC2. Long-chain fatty acids enter the mitochondria via carnitine palmitoyltransferase 1 and 2 (CPT1 and CPT2) and carnitine acylcarnitine translocase (CAT). Within the mitochondria, GLS hydrolyzes glutamine into glutamate. (B) Schematic representation of mixed bone marrow chimera experiments to assess hematopoietic requirements for Cpt2 control (Cpt2^fl/fl; ROSA26 CreER^−/−) or KO (Cpt2^fl/fl; ROSA26 CreER^+/+) cells and Gls control (Gls^fl/fl; ROSA26 CreER^−/−) or KO (Gls^fl/fl; ROSA26 CreER^+/+) cells. (C) CD45.2 chimerism was normalized to either pre-tamoxifen peripheral blood chimerism levels for the spleen or HSC chimerism for the bone marrow of Cpt2 chimeras. Data are shown for B cells, DCs, Gr1^hi and Gr1^lo myeloid cells, NK cells, and T cells for the spleen. Data are shown for HSCs, MPPs, CMPs, GMPs, MEPs, and CLPs in the bone marrow. Each symbol represents an individual mouse. Mean values ± SEM are shown, and data are pooled from three independent experiments. P values >0.05 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test are not depicted. (D) CD45.2⁺ GFP⁻ chimerism was normalized as stated in Fig. 1 C for Gls chimeras. Cell populations assessed are the same as in Fig. 1 C with the addition of platelets in the blood. Each symbol represents an individual mouse. Mean values ± SEM are shown, and data are pooled from two independent experiments. *, P < 0.05 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test.

Figure S1.

Gating strategy and quantification of deletion of Cpt2 and Gls. (A) Representative flow cytometry gating strategy for splenic populations. (B) Representative flow cytometry gating strategy for bone marrow populations. (C) Quantitative PCR analysis of genomic Cpt2 deletion. CD45.2⁺ cells from the bone marrow or spleen were sorted for genomic DNA extraction. DNA quantification within loxP sites was quantified relative to WT cells and GAPDH. Mean values ± SEM are shown. Data are pooled from two independent experiments. ****, P < 0.0001 by Student’s two-tailed t test. (D) Representative flow cytometry gating strategy for platelets in the blood. (E) Quantitative PCR analysis of genomic Gls deletion. CD45.2⁺ GFP⁻ cells from the bone marrow or spleen were sorted for genomic DNA extraction. DNA quantification within loxP sites was quantified relative to WT cells and GAPDH. Mean values ± SEM are shown. Data are pooled from two independent experiments. ****, P < 0.0001 by Student’s two-tailed t test.

Figure S2.

Raw chimerism frequencies for Cpt2 and Gls chimeras. (A) Raw chimerism values for Cpt2 chimeras shown in Fig. 1 C. Each symbol represents an individual mouse. Mean values ± SEM are shown, and data are pooled from three independent experiments. P values >0.05 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test are not depicted. (B) Raw chimerism values for Gls chimeras shown in Fig. 1 D. Each symbol represents an individual mouse. Mean values ± SEM are shown, and data are pooled from two independent experiments. P values >0.05 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test are not depicted.

Figure 2.

Mitochondrial pyruvate is transiently required by both myeloid progenitors and mature cells. (A) Schematic representation of mixed bone marrow chimeras to assess the hematopoietic requirement for Mpc2 (Mpc2^fl/fl; ROSA26 CreER^−/− or ROSA26 CreER^+/−), both immediately following deletion and after prolonged deletion (10+ wk after tamoxifen). (B) Representative flow cytometry plots of CD45.2 chimerism of Ly6C^hi monocytes and neutrophils for Mpc2 control, recent KO, and prolonged KO chimeras. (C) CD45.2 chimerism of splenic immune lineages at early or prolonged timepoints after deletion of Mpc2. Each symbol represents an individual mouse. Mean values ± SEM are shown. Data are pooled from four independent experiments. *, P < 0.05; ***, P < 0.001; and ****, P < 0.0001 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (D) Representative flow cytometry plots of CD45.2 chimerism of GMPs and CMoPs. (E) CD45.2 chimerism of bone marrow progenitors described previously with the addition of CMoPs and CDPs at early or prolonged timepoints after deletion of Mpc2. Data are normalized to HSC CD45.2 chimerism. Each symbol represents an individual mouse. Mean values ± SEM are shown, and data are pooled from three independent experiments. *, P < 0.05; **, P < 0.01; ****, P < 0.0001 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test.

Figure S3.

Gating strategy, quantification of deletion of Mpc2, and Mpc2 retroviral rescue of MPC2-deficient myeloid cells. (A) Representative flow cytometry gating strategies for splenic Ly6C^hi and Ly6C^lo monocytes and neutrophils. (B) Schematic representation of Mpc2 retroviral bone marrow chimera experiments to assess ability to rescue Mpc2 KO (+tam) or control (no tam) Ly6C^hi monocytes and neutrophils. C-kit⁺ cells were enriched from the bone marrow of Mpc2^fl/fl; ROSA26 CreER^+/− mice and transduced with Mpc2 retrovirus prior to transplantation alongside bone marrow from WT mice into irradiated recipients. (C) Representative flow cytometry plots of Mpc2 retrovirus (RV) expression in HSCs, Ly6C^hi monocytes, and neutrophils of an individual control or KO mouse. (D) Mpc2 transduction in HSCs paired with Ly6C^hi monocytes (top) or neutrophils (bottom) from the same mouse (n = 6 for untreated and n = 4 for tamoxifen). Data are pooled from two independent experiments. *, P < 0.05; **, P < 0.01 by Student’s two-tailed paired t test. (E) Quantitative PCR analysis of genomic Mpc2 deletion in Mpc2 chimeras. CD45.2⁺ cells from the bone marrow or spleen were sorted for genomic DNA extraction. DNA quantification within loxP sites was quantified relative to WT cells and GAPDH. Mean values ± SEM are shown. Data are pooled from two independent experiments. ****, P < 0.0001 by one-way ANOVA with post-hoc Tukey’s multiple comparisons test. (F) Raw chimerism values for Mpc2 chimeras shown in Fig. 2 E. Each symbol represents an individual mouse. Mean values ± SEM are shown, and data are pooled from three independent experiments. ****, P < 0.0001 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (G) Representative flow cytometry gating strategies for CDPs and CMoPs in the bone marrow.

Figure 3.

Mature myeloid cells diminish then rapidly recover after Mpc2 deletion. (A) CD45.2 peripheral blood chimerism of mature cell populations was assessed every 2–3 wk in Mpc2 chimeras. Values are normalized to pre-tamoxifen chimerism of each cell type. Data are pooled from three independent experiments. Mean values ± SEM are shown. ****, P < 0.0001 by paired two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (B and C) CD45.2 peripheral blood chimerism of mature cell populations was assessed every 2 wk in Cpt2 (B) or Gls (C) chimeras. Values are normalized to pre-tamoxifen chimerism of each cell type. Mean values ± SEM are shown. Data are pooled from two independent experiments. P values >0.05 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test are not depicted.

Figure S4.

Proliferation, survival, and recovery of MPC2-deficient cells. (A) Raw chimerism values for peripheral myeloid cells of Mpc2 chimeras in Fig. 3 A. Data are pooled from three independent experiments. Mean values ± SEM are shown. ****, P < 0.0001 by paired two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (B) BrdU incorporation was measured in both CD45.1⁺ (WT) and CD45.2⁺ (WT, recent Mpc2 deletion, or prolonged Mpc2 deletion) HSCs, MPPs, CLPs, Ly6C^hi monocytes, and neutrophils following a 1-h pulse of BrdU. Each line connects CD45.1⁺ and CD45.2⁺ cells within the same mouse (n = 11 for WT chimeras and n = 19 for recent and prolonged KO chimeras). Data are pooled from two independent experiments. P values >0.05 by paired two-way ANOVA with post-hoc Sidak’s multiple comparisons test are not depicted. (C) Mpc2 chimeras were setup at a ratio of 90% WT to 10% Mpc2^fl/fl; ROSA26 CreER^−/− or CreER^+/− bone marrow cells (as opposed to the 1:1 ratio in Fig. 3). CD45.2 peripheral blood chimerism of Ly6C^hi monocytes and neutrophils was assessed every 2 wk in Mpc2 chimeras. Values are normalized to pre-tamoxifen chimerism of each cell type. Mean values ± SEM are shown. ****, P < 0.0001 by paired two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (D) BrdU incorporation was measured in peripheral CD45.2⁺ B and T cells for 3 d following a week of BrdU water administration. Mean values ± SEM are shown. Data are pooled from two independent experiments. P values >0.05 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test are not depicted.

Figure 4.

GMPs and CMoPs proliferate rapidly immediately following Mpc2 deletion. (A) Schematic representation of BrdU pulse experiment using Mpc2 mixed bone marrow chimeras. Chimeric mice were injected with BrdU 1 h prior to sacrifice. (B) Representative histograms of BrdU staining in GMPs of Mpc2 control, recent KO, and prolonged KO chimeras. Gray lines represent staining of GMPs from a mouse that was not injected with BrdU. Blue and red lines represent CD45.1⁺ (WT) and CD45.2⁺ (WT, recent KO, or prolonged KO) GMPs respectively from the same mouse injected with BrdU. (C) BrdU incorporation was measured in both CD45.1⁺ (WT) and CD45.2⁺ (WT, recent Mpc2 deletion, or prolonged Mpc2 deletion) CMPs, GMPs, and CMoPs following a 1-h pulse of BrdU. Each line connects CD45.1⁺ and CD45.2⁺ cells within the same mouse (n = 11 for WT chimeras and n = 19 for recent and prolonged KO chimeras). Data are pooled from two independent experiments. ****, P < 0.0001 by paired two-way ANOVA with post-hoc Sidak’s multiple comparisons test. (D) Schematic representation of BrdU pulse-chase experiment using Mpc2 mixed bone marrow chimeras to assess survival and turnover. Chimeras were administered BrdU water for 1 wk and then were switched back to normal drinking water, at which point BrdU incorporation was assessed in peripheral blood at the time of removal from BrdU and three successive days afterwards. (E) BrdU incorporation was measured in peripheral CD45.2⁺ Ly6C^hi monocytes and neutrophils following a 3-d chase after a week of BrdU water administration. Mean values ± SEM are shown. Data are pooled from two independent experiments. P values >0.05 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test are not depicted.

Figure 5.

GMPs differentially express numerous genes immediately following Mpc2 deletion but few persist following myeloid recovery. (A and B) Volcano plots of gene expression fold changes between Mpc2 recent KO and WT GMPs (A) or Mpc2 prolonged KO and WT GMPs (B). Following RNA-seq, adjusted P values were calculated using DESeq2. Each dot represents a gene. Recent Mpc2 KO mice were pooled into groups of two mice to obtain a sufficient number of cells. A total of 10 mice were used but were processed as five samples. For all other groups, five mice were analyzed without pooling for each genotype. (C) Venn diagram analysis depicting genes differentially expressed by only recent Mpc2 KO GMPs, prolonged Mpc2 KO GMPs, or both relative to WT GMPs. Lists of genes include all hits or the top 10 hits not including pseudogenes or predicted genes. Red up arrows indicate genes upregulated in KOs relative to WT, while blue down arrows indicate genes downregulated in KOs relative to WT. All data are available in the NCBI GEO (recent KO accession number GSE225578 and prolonged KO accession number GSE184548).

Figure 6.

In vitro GMPs do not require the MPC for survival. (A) WT GMPs were cultured with DMSO or 10 μM UK5099 for 7 d. On each day of the culture, cells were harvested, and live cells were counted. One experiment with 10 technical replicates from two mice per group is depicted as a representative of four independent experiments. P values >0.05 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test are not depicted. (B) Pyruvate excretion rate in the media of GMPs cultured with DMSO or UK5099 over a 48-h culture. ****, P < 0.0001 by Student’s two-tailed t test. Mean values ± SEM are shown for four replicates from two independent experiments. (C) GMPs from Mpc2^fl/fl; ROSA26 CreER^+/− or ROSA26 CreER^+/− mice were treated with 0.5 μM 4-OHT. On each day of the culture, cells were harvested, and live cells were counted. One experiment with 10 technical replicates from two mice per group is depicted as a representative of two independent experiments. P values >0.05 by Student’s two-tailed t test are not depicted. (D) Quantitative PCR analysis of genomic Mpc2 deletion in GMPs from Mpc2^fl/fl; ROSA26 CreER^+/− or ROSA26 CreER^+/− mice. GMPs were treated with 4-OHT for 2 d prior to DNA extraction. DNA quantification within loxP sites was quantified relative to WT cells and GAPDH. Mean values ± SEM are shown for four replicates from two independent experiments. ****, P < 0.0001 by one-way ANOVA with post-hoc Tukey’s multiple comparisons test.

Figure 7.

GMPs use glutamine in vitro to generate TCA cycle intermediates. LC/MS analysis of ¹³C incorporation into TCA cycle intermediates following a 24-h culture of WT GMPs with uniformly ¹³C-labeled glucose, BCAAs (leucine [leu]/isoleucine [ile]/valine [val]), glutamine (gln), palmitate, pyruvate, lactate, alanine (ala), acetate, asparagine (asn), lysine (lys)/phenylalanine (phe)/tryptophan (trp)/tyrosine (tyr)/cysteine (cys)/glycine (gly)/serine (ser), or glutamate (glu)/arginine (arg)/histidine (his)/proline (pro)/methionine (met)/threonine (thr)/phenylalanine (phe)/tyrosine (tyr)/aspartate (asp). Labeling data were corrected for natural-abundance ¹³C. Mean values ± SEM are shown for three to five replicates from two independent experiments for each ¹³C-labeled carbon source. Cells were pooled from five mice per each carbon source for each experiment.

Figure 8.

Mature myeloid cells fail to fully recover after deletion of both Mpc2 and Gls. (A) Peripheral blood CD45.2 chimerism of mature cells was assessed every 2 wk. Values are normalized to pre-tamoxifen chimerism of each cell type. Each line represents longitudinal analysis of individual mice with a symbol at each time point measured. Data are pooled from two independent experiments. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 by paired two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (B) CD45.2 chimerism of bone marrow progenitors normalized to HSC chimerism immediately following 2 wk of tamoxifen administration. Each symbol represents an individual mouse. Mean values ± SEM are shown. Data are pooled from two independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (C) BrdU incorporation was measured in both CD45.1⁺ (WT) and CD45.2⁺ (WT or recent deletion of Gls and Mpc2) CMPs, GMPs, and CMoPs following a 1-h pulse of BrdU. Each line connects CD45.1⁺ and CD45.2⁺ cells within the same mouse (n = 9 for WT chimeras and n = 6 for recent DKO chimeras). Mice in which Mpc2 was <90% reduced relative to WT controls were excluded. Data are pooled from two independent experiments. P values >0.05 by paired two-way ANOVA with post-hoc Sidak’s multiple comparisons test are not depicted.

Figure S5.

Quantification of efficiency of Gls and Mpc2 deletion. (A) Raw chimerism frequencies for Gls Mpc2 chimeras in Fig. 8 B. Each symbol represents an individual mouse. Mean values ± SEM are shown. Data are pooled from two independent experiments. ***, P < 0.001; ****, P < 0.0001 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (B) Quantitative PCR analysis of genomic Gls and Mpc2 deletion in recent Gls Mpc2 KO chimeras. CD45.2⁺ cells from the bone marrow or spleen were sorted for genomic DNA extraction. DNA quantification within loxP sites was quantified relative to WT cells and GAPDH. Mean values ± SEM are shown. Data are pooled from two independent experiments. ****, P < 0.0001 by Student’s two-tailed t test. (C) RNA-seq reads across exons 1–3 of Gls (top) and 2–5 of Mpc2 (bottom) of CD45.2⁺ GMPs from ROSA26 CreER^+/− (WT), Mpc2 KO, Gls KO, or Gls Mpc2 DKO chimeras about 10 wk after tamoxifen treatment. Exon 1 of Gls and exon 3 of Mpc2 are floxed in the appropriate genotyped mice. One representative trace for each genotype is shown using Integrative Genomics Viewer (left). TPMs were quantified for the floxed exon of Gls (top right) or Mpc2 (bottom right). TPMs were calculated from reads obtained using DEXSeq. Each data point represents a mouse (n = 5 for each genotype). Mean values ± SEM are shown. ***, P < 0.001; ****, P < 0.0001 by one-way ANOVA with post-hoc Tukey’s multiple comparisons test.

Figure 9.

Deletion of Gls delays emergency myelopoiesis. (A) Schematic representation of cyclophosphamide experiment using Gls KO, Mpc2 KO, and Cre only control mixed bone marrow chimeras. Chimeric mice were injected with cyclophosphamide and then were bled on days 2 through 7 after treatment. (B) Peripheral blood cell counts of both CD45.1⁺ (WT) and CD45.2⁺ (WT, Gls KO, or Mpc2 KO) Ly6C^hi monocytes, neutrophils, B cells, and T cells from cyclophosphamide-treated chimeras. Values are normalized to pretreatment counts of each cell type. Data are pooled from two independent experiments. Mean values ± SEM are shown. *, P < 0.05; **, P < 0.01 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test. (C) GMPs from WT or Gls KO chimeras were cultured with DMSO or 10 μM UK5099 in the presence of IL-3 and IL-6 for 7 d. On each day of the culture, cells were harvested and CD45.2 chimerism (left) and live cell counts (right) were obtained. Chimerism is normalized to the chimerism of the sorted GMPs. Cell counts are normalized to the total number of CD45.2⁺ cells plated on day 0. Data are pooled from two independent experiments with a total of 14 replicates from four mice per group. Mean values ± SEM are shown. **, P < 0.01; ***, P < 0.001 by two-way ANOVA with post-hoc Tukey’s multiple comparisons test.

Figure 10.

Full chimeras do not survive deletion of Mpc2 and/or Gls. Full chimeras were generated with bone marrow from WT (ROSA26 CreER^+/−), Mpc2 KO (Mpc2^fl/fl; ROSA26 CreER^+/−), Gls KO (Gls^fl/fl; ROSA26 CreER^+/−), or Gls Mpc2 DKO (Gls^fl/fl Mpc2^fl/fl; ROSA26 CreER^+/−) donors. Chimeras were provided tamoxifen chow for 14 d, and severe moribundity and survival were monitored over time. The single WT mouse that died was found in a flooded cage, suggesting an unrelated death. Data are pooled from two independent experiments. **, P < 0.01; ****, P < 0.0001 by log-rank test.