CXCR4 signaling determines the fate of hematopoietic multipotent progenitors by stimulating mTOR activity and mitochondrial metabolism

Abstract

Both cell-intrinsic and niche-derived, cell-extrinsic cues drive the specification of hematopoietic multipotent progenitors (MPPs) in the bone marrow, which comprise multipotent MPP1 cells and lineage-restricted MPP2, MPP3, and MPP4 subsets. Patients with WHIM syndrome, a rare congenital immunodeficiency caused by mutations that prevent desensitization of the chemokine receptor CXCR4, have an excess of myeloid cells in the bone marrow. Here, we investigated the effects of increased CXCR4 signaling on the localization and fate of MPPs. Knock-in mice bearing a WHIM syndrome-associated CXCR4 mutation (CXCR4¹⁰¹³) phenocopied the myeloid skewing of bone marrow in patients. Whereas MPP4 cells in wild-type mice differentiated into lymphoid cells, MPP4s in CXCR4¹⁰¹³ knock-in mice differentiated into myeloid cells. This myeloid rewiring of MPP4s in CXCR4¹⁰¹³ knock-in mice was associated with enhanced signaling mediated by the kinase mTOR and increased oxidative phosphorylation (OXPHOS). MPP4s also localized further from arterioles in the bone marrow of knock-in mice compared with wild-type mice, suggesting that the loss of extrinsic cues from the perivascular niche may also contribute to their myeloid skewing. Chronic treatment with the CXCR4 antagonist AMD3100 or the mTOR inhibitor rapamycin restored the lymphoid potential of MPP4s in knock-in mice. Thus, CXCR4 desensitization drives the lymphoid potential of MPP4 cells by dampening the mTOR-dependent metabolic changes that promote myeloid differentiation.

Fig. 1.. The HSPC compartment is myeloid-skewed in the BM of WS mice and patients.

(A) Representative flow cytometric analyses show frequencies of MPP1, MPP2, MPP3, and MPP4 (gated on LSK and sorted on CD150, Flt3, CD48, and forward scatter (FSC)) as indicated, in the marrow fraction of WT, Cxcr4^+/1013 (+/1013), and Cxcr4^1013/1013 (1013/1013) mice. The bar graph shows the proportions of MPP1, MPP2, MPP3, and MPP4 in the marrow of WT and knockin mice. n= 8 independent experiments for a total of 22 WT, 22 +/1013 and 13 1013/1013 mice (MPP1) or n= 9 independent experiments for a total of 22 WT, 19 +/1013 and 13 1013/1013 mice (MPP2, MPP3, and MPP4). (B) Representative histograms show surface Cxcr4 abundance as determined by flow cytometry on MPP4 from the marrow fraction of WT and knockin mice. Background fluorescence is shown (isotype, gray histogram). Histograms are representative of n= 4 independent experiments for a total of 11 WT, 10 +/1013, and 8 1013/1013 mice. (C) Migration of WT and knockin MPP2 and MPP3 (MPP2–3) or MPP4 in response to Cxcl12 in the presence or absence of AMD3100 was assessed by flow cytometry. n= 3 independent experiments for a total of 9 WT, 6 +/1013 and 5 1013/1013 mice. (D) Representative histograms show the abundance of phosphorylated Akt (p-Akt) in MPP4 from WT and knockin mice stimulated with Cxcl12. Fluorescence in the absence of the p-Akt antibody (dashed line) or in absence of Cxcl12 stimulation (gray) are shown. The bar graphs show mean fluorescence intensity (MFI) for p-Akt determined by flow cytometry on MPP1, MPP2–3, and MPP4 from WT and knockin mice. n= 3 independent experiments for a total of 7 mice per group. (E) Absolute numbers of MPP1, MPP2, MPP3, and MPP4 in the marrow fraction of WT and knockin mice. n= 8 independent experiments for a total of 22 WT, 22 +/1013 and 13 1013/1013 mice (MPP1) or n= 9 independent experiments for a total of 22 WT, 19 +/1013 and 13 1013/1013 mice (MPP2–4). (F) Representative dot-plots obtained by flow cytometry show the frequencies of CD34⁺ cells in BM aspirates of n=4 healthy donors (HD) and n=2 WHIM Syndrome (WS) patients. (G) Pie chart representation of the contribution of each HSPC subset to the total CD34⁺ pool in representative HD and WS BM samples. (H) Representative dot-plots show the frequencies of CLPs (CD34⁺CD38⁺CD45RA⁺CD10⁺) among CD34⁺ cells and the absolute numbers of CLPs in BM aspirates of HD and WS patients. (I) Representative histogram shows the detection of CXCR4 on the surface of CD34⁺ cells from BM aspirates of HD and WS patients. Background fluorescence is shown (isotype, gray histogram). The graph shows the numbers of CXCR4⁺ cells within CD34⁺ cells from BM aspirates of HD and WS patients. Barplots indicate mean/SEM; horizontal lines correspond to median values. ^# P < 0.05; ^## P < 0.005; ^### P < 0.0005 for Kruskal–Wallis tests; * P < 0.05; ** P < 0.005; *** P < 0.0005 for Dunn’s test.

Fig. 2.. CXCR4 desensitization is required for efficient generation and maintenance of the lymphoid-biased MPP4 pool.

(A) Representative flow cytometric analyses comparing the frequencies of MPP3 and MPP4 (gated on LSK and sorted on CD150 and Flt3) 4 days after MPP1 from WT, +/1013, and 1013/1013 mice were cultured in the absence of AMD3100. Graphs show frequencies and absolute numbers of MPP3 and MPP4 in absence or presence of AMD3100. n= 6 independent experiments for a total of 10 WT, 7 +/1013, and 7 1013/1013 mice (no AMD3100) or 4 independent experiments for a total of 7 WT, 6 +/1013, and 3 1013/1013 mice (with AMD3100). (B) Heatmap of decreased ATAC-seq peaks in MPP1 from WT or knockin mice and gene ontology analysis with the genes associated with decreased ATAC-seq peaks in 1013/1013 vs WT MPP1. Data were generated using one replicate per group. Three mice of the same genotype were pooled to generate one replicate (3 mice for WT, 3 mice for +/1013, 3 mice for 1013/1013). (C) Highest enriched TF motifs in decreased ATAC-seq peaks in 1013/1013 vs WT MPP1. (D) Numbers of B cells after 14 days of coculture between WT or knockin MPP1 and OP9/IL-7 stromal cells. n= 3 independent experiments for a total of 3 mice per group. (E) Number of B cells (B220⁺CD11b⁻) after 7 days of coculture between MPP3 or MPP4 derived from WT or knockin MPP1 cultures and OP9/IL-7 stromal cells. n= 3 independent experiments for a total of 3 mice per group. (F) Heatmap of decreased ATAC-seq peaks in WT or knockin MPP4 and gene ontology analysis with the genes associated with decreased ATAC-seq peaks in 1013/1013 vs WT MPP4. Data were generated using two replicates per group. Three mice of the same genotype were pooled to generate the two replicates (3 mice for WT, 3 mice for +/1013, 3 mice for 1013/1013). (G) Highest enriched TF motifs in decreased ATAC-seq peaks in 1013/1013 vs WT MPP4. (H) Percentage of the indicated MPP subsets in G0 (DAPI^lowKi-67⁻), G1 (DAPI^lowKi-67⁺), S (DAPI⁺Ki-67⁺), and G2/M (DAPI^highKi-67⁺) phases. n= 5 independent experiments for a total of 15 WT, 15 +/1013, and 12 1013/1013 mice. Statistical significance compared to WT cells for each phase of the cell cycle is shown. ^# P < 0.05; ^## P < 0.005; for Kruskal–Wallis tests; * P < 0.05; ** P < 0.005 for Dunn’s test; ^£ P < 0.05; ^££ P < 0.005; ^£££ P < 0.0005 for Mann-Whitney’s test.

Fig. 3.. CXCR4 desensitization is required for MPP4 positioning near BM perivascular structures.

(A) Schematic diagram for the generation of CD45.2⁺->CD45.1⁺ short-term (3 weeks) and long-term (16 weeks) BM chimeras. (B) Absolute numbers of donor-derived CD45.2⁺ (WT, +/1013, or 1013/1013) MPP2, MPP3, and MPP4 recovered from the marrow of chimeras in CD45.1⁺ WT recipients 3 weeks or 16 weeks after marrow transplantation. n= 3 independent experiments for a total of 5 WT, 8 +/1013, and 8 1013/1013 donor mice (3-week chimeras) or 4 independent experiments for a total of 17 WT, 19 +/1013, and 15 1013/1013 donor mice (16-week chimeras). (C) Schematic diagram for the generation of CD45.1⁺ + CD45.2⁺-> CD45.1⁺/CD45.2⁺competitive BM chimeras. (D) Relative contributions of CD45.2⁺ and CD45.1⁺ cells to donor-derived MPP2, MPP3, MPP4, and CLPs in the BM of CD45.1⁺/CD45.2⁺ WT recipient mice 3 weeks after competitive BM transplantation. n=2 independent experiments for a total of 5 WT, 4 +/1013, and 4 1013/1013 donor mice. (E) Schematic diagram for the generation of CD45.1⁺->CD45.2⁺ short- and long-term BM chimeras. (F) Absolute numbers of donor CD45.1⁺ WT MPP2, MPP3, and MPP4 recovered from the marrow of chimeras in CD45.2⁺ (WT, +/1013, or 1013/1013) recipients 3 weeks or 16 weeks after marrow transplantation. n= 3 independent experiments for a total of 10 WT, 10 +/1013, and 5 1013/1013 recipient mice (3-week chimeras) or 4 independent experiments for a total of 12 WT, 11 +/1013, and 5 1013/1013 recipient mice (16-week chimeras). (G) Representative BM sections from WT and knockin mice stained to show laminin and nuclei (DAPI). MPP4 identified by RNA-scope and immunofluorescence staining (fig. S3, B and C) are indicated with yellow circles. Spatial distribution of centroids of MPP4 cells plot as yellow circles within points of DAPI detected cells (blue points). Scale bar, 800 μm. (H) Proportions of MPP4 (c-Kit⁺Sca-1⁺Flt3⁺) detected in the BM sections of WT and knockin mice after imaging (fig. S3, B and C). n= 3 independent determinations for a total of 6 WT, 4 +/1013, and 4 1013/1013 mice. (I) Proportions of Dsred⁺ cells within arteriolar cells (Lam⁺Sca-1⁺) in the BM sections of Cxcr4^+/+Cxcl12^Dsred, Cxcr4^+/1013Cxcl12^Dsred, and Cxcr4^1013/1013Cxcl12^Dsred mice. n= 3 independent determinations for a total of 3 mice per group. (J) Spatial distance analysis between each MPP4 centroid and the centroid of the nearest arteriolar cell within representative BM sections from WT and knockin mice. Dots indicate individual MPP4 analyzed for one representative mouse per group with >140 cells analyzed. Quantifications are representative of 3 independent determinations per group. Mann–Whitney U test was used to assess statistical significance (^££££, P < 0.0001 between WT and 1013/1013 mice). (K) Representative histograms for intracellular antibody-based detection of phosphorylated STAT5 (p-STAT5) in sorted, untreated MPP4 from the marrow fraction of WT and knockin mice. Fluorescence in absence of the p-STAT5–specific antibody (dashed line) is shown. MFI values for p-STAT5 were determined by flow cytometry on sorted MPP4 from the marrow fraction of WT and knockin mice. n= 3 independent experiments for a total of 3 WT, 2 +/1013 and 3 1013/1013 mice. ^# P < 0.05; ^## P < 0.005; ^### P < 0.0005 for Kruskal–Wallis tests; * P < 0.05; ** P < 0.005; *** P < 0.0005 for Dunn’s test compared with WT, and § for Dunn’s test comparing +/1013 and 1013/1013; ^££££ P < 0.0001 for Mann-Whitney’s test.

Fig. 4.. CXCR4 desensitization is required to maintain the molecular identity of MPP4.

(A) RNAseq data were generated from 3 × 10³ MPP1-4 sorted from the marrow fraction of WT, +/1013, and 1013/1013 mice. (B) Volcano plots of differentially expressed genes in +/1013 vs. WT MPP4 and in 1013/1013 vs. WT MPP4. In each graph, the knockin is shown on the left and WT on the right. Genes highly differentially expressed (log2FC(knockin vs WT MPP4) > 2; -log10(adjusted p-value) > 2) are shown in red. Data were generated by RNA-seq from 2 (1013/1013 MPP4) or 3 (WT and +/1013 MPP4) independent experiments (9 mice each for WT and +/1013 MPP4, 6 mice for 1013/1013 MPP4). Three mice of the same genotype were pooled to generate one replicate. (C) Examples of biological processes for which 1013/1013 and WT MPP4 display differential gene expression signatures. Biological processes overrepresented in increased and decreased transcripts in 1013/1013 MPP4 compared to WT MPP4 were determined by Gene Ontology (GO) analyses.

Fig. 5.. CXCR4 signaling termination is required to maintain the lymphoid potential in MPP4.

(A) RNAseq-based heatmap showing normalized dye intensity for the expression of genes involved in lymphocyte differentiation (GO:0030098) in WT, +/1013, and 1013/1013 MPP4. Data were generated from 2 (1013/1013 MPP4) or 3 (WT and +/1013 MPP4) independent experiments (9 mice each for WT and +/1013 MPP4, 6 mice for 1013/1013 MPP4). Three mice of the same genotype have been pooled to generate one replicate. (B) Biomark-based PCA of relative expression of selected genes involved in lympho-myeloid differentiation in MPP4. (C) The heatmap shows the relative quantification (RQ) for lymphoid transcripts normalized to Actb expression in each sample. The bar graph shows the RQ of the most decreased pro-lymphoid transcripts in MPP4. n = 4 WT, 3 +/1013, and 5 1013/1013 mice with samples from each mouse run in duplicate. (D) The heatmap shows the RQ for myeloid transcripts normalized to Actb expression in each sample. The bar graph shows the RQ of the most increased pro-GM transcripts in MPP4. n = 4 WT, 3 +/1013, and 5 1013/1013 mice with samples from each mouse run in duplicate. (E) Representative histograms for detection of Flt3 and Irf8 in MPP4. Fluorescence in absence of Irf8 antibody (dashed line) is shown. MFI values for surface Flt3 and proportions of Irf8^high cells were determined by flow cytometry in MPP4. n = 6 independent experiments for a total of 11 WT, 18 +/1013, and 12 1013/1013 mice (Flt3) or n= 3 independent experiments for a total of 6 mice per group (Irf8). (F) Representative flow cytometric analyses comparing the frequencies of B cells (B220⁺CD11b⁻) between WT and knockin MPP4 co-cultures. Proportions of B cells after co-culture between MPP4 and OP9/IL-7 stromal cells in the presence or absence of AMD3100 were determined after 7 days. n = 3 independent experiments for a total of 4 WT, 3 +/1013, and 4 1013/1013 mice. (G) Representative flow cytometric analyses comparing the frequencies of Lin⁻Sca-1⁻c-Kit⁺ [LK], Lin⁻Sca-1^lowc-Kit^low, Gr1⁻CD11b⁻ and Gr1⁺CD11b⁺ cells between WT and knockin MPP4 cultures at day 4 or day 7. Absolute numbers of GMPs were determined after 4 days of culture, whereas absolute numbers of Gr1⁺CD11b⁺ cells were assessed at day 7. Proportions of Lin⁻Sca-1^lowc-Kit^low cells were determined after 4 days of culture. n= 4 independent experiments for a total of 8 WT, 8 +/1013, and 7 1013/1013 mice (day 4) or n= 3 independent experiments for a total of 7 WT, 7 +/1013 and 6 1013/1013 mice (day 7). (H) Proportions of MPP4 and Flt3^low cells in WT and knockin MPP4 cultures after 4 days of differentiation in the presence or absence of AMD3100. n = 5 independent experiments for a total of 10 WT, 11 +/1013, and 8 1013/1013 mice (no AMD3100) or n= 3 independent experiments for a total of 4 WT, 4 +/1013, and 3 1013/1013 mice (with AMD3100). (I) Expression of Flt3 determined by flow cytometry on MPP4 after 4 days of culture in the presence or absence of AMD3100. n= 3 independent experiments for a total of 4 WT, 4 +/1013, and 3 1013/1013 mice. (J) Proportions of donor CD45.2⁺ myeloid or B cells recovered from the marrow of CD45.1⁺ recipients 2 weeks after the transplantation. n= 2 independent experiments for a total of 6 WT, 5 +/1013, and 5 1013/1013 donor mice. (K) Absolute numbers of GMPs were determined after 4 days of WT MPP4 culture in the presence or absence of Cxcl12 and AMD3100 as indicated. Absolute numbers of Gr1⁺CD11b⁺ cells were assessed at day 7. n = 3 independent experiments for a total of 3 mice per group (day 4) or n= 5 independent experiments for a total of 9 mice per group (day 7). ^# P < 0.05; ^## P < 0.005 for Kruskal–Wallis tests; * P < 0.05; ** P < 0.005 for Dunn’s test compared with WT, and § for Dunn’s test comparing +/1013 and 1013/1013; ^£ P < 0.05; ^£££ P < 0.0005 for Mann-Whitney’s test.

Fig. 6.. Overactive OXPHOS-driven metabolism in Cxcr4¹⁰¹³ knockin MPP4.

(A) Representative histograms for intracellular detection of phosphorylated mTOR (p-mTOR) and phosphorylated S6 (p-S6) in MPP4 of the indicated genotypes stimulated with Cxcl12. Fluorescence of WT MPP4 in the absence of p-mTOR or p-S6 antibody (dashed line), in the absence of Cxcl12 stimulation (gray histogram) or in the presence of Cxcl12 and AMD3100 (black histogram) are shown. MFI values for p-mTOR and p-S6 were determined by flow cytometry. n=3 independent experiments for a total of 4 (with AMD3100) or 5 (no AMD3100) WT, 6 +/1013, and 6 1013/1013 mice. (B) OXPHOS and glycolytic contents measured by oxygen consumption rate (OCR) and extracellular acidification rate (ECAR), respectively, were determined in freshly sorted WT and 1013/1013 MPP4 by Seahorse Mito Stress Test. n= 3 independent experiments for total of 9 WT, 9 +/1013, and 9 1013/1013 mice. Three mice of the same genotype were pooled for each independent experiment. Statistical significance compared with WT cells for each time point is shown. (C) Representative histograms and MFI values of TMRE and MTG staining in MPP4. n=3 independent experiments for a total of 10 WT, 10 +/1013, and 9 1013/1013 mice. (D) Representative images showing staining for TOMM20 and Hoechst 33342 in freshly isolated MPP4. Scale bar, 5 μm. n=3 independent determinations for a total of at least 50 cells analyzed per genotype. (E) MFI values of TOMM20 and number of mitochondrial fragments per WT MPP4 were determined using FIJI software. n=3 independent determinations for a total of at least 50 cells analyzed per condition (3 WT mice in total). (F) TMRE and MTG MFI values in WT MPP4 co-cultured with OP9/IL-7 cells. n= 3 independent experiments for a total of 3 mice per group. (G) Frequencies of TMRE^high/+ fraction A (FR. A) and TMRE^low/- fraction B (FR. B) within WT and knockin MPP4. n= 3 independent experiments for a total of 8 WT, 8 +/1013, and 6 1013/1013 mice. (H) Absolute numbers of GMPs were determined after 4 days of fraction A or fraction B MPP4 culture, whereas absolute numbers of mature myeloid cells were assessed at day 7. n= 3 independent experiments for a total of 6 WT, 6 +/1013, and 5 1013/1013 mice (day 4) or n= 3 independent experiments for a total of 7 WT, 10 +/1013, and 7 1013/1013 mice (day 7). * for Dunn’s test compared with WT cells, and £ for Mann-Whitney test comparing FR. A and FR. B. (I) Expression of pro-GM genes in single TMRE^low (fraction B) MPP4 as determined by Biomark. Proportions of WT and 1013/1013 TMRE^low MPP4 with detectable or undetectable expression of Mpo and Irf8. n= 2 independent experiments with > 50 cells per condition analyzed (2 mice of each genotype in total). (J) Proportions of donor CD45.2⁺ (fraction A or B) myeloid or B cells recovered from the marrow of CD45.1⁺ WT recipients 2 weeks after the transplantation. n= 2 independent experiments for a total of 3 mice per group. (K) Absolute numbers of mature myeloid cells in WT or knockin MPP4 cultures in the presence or absence of Mito-q or CCCP were assessed at day 7. n= 3 independent experiments for a total of 3 mice per group. ^# P < 0.05; ^## P < 0.005; ^### P < 0.0005 for Kruskal–Wallis tests; * P < 0.05; ** P < 0.005; **** P < 0.0001 for Dunn’s test; ^£ P < 0.05; ^££ P < 0.005; ^£££ P < 0.0005 for Mann-Whitney’s test.

Fig. 7.. Modulation of the Cxcr4-mTOR signaling axis normalizes the fate properties of Cxcr4¹⁰¹³ knockin MPP4.

(A) Absolute numbers of GMPs were determined after 4 days of fraction B (TMRE^low) MPP4 culture in the presence or absence of AMD3100 or rapamycin, and absolute numbers of mature myeloid cells (Gr1⁺CD11b⁺) were assessed at day 7. n= 4 independent experiments for a total of 4 WT, 4 +/1013, and 3 1013/1013 mice (day 4) or n= 3 independent experiments for a total of 3 mice per group (day 7). * for Dunn’s test comparing WT and 1013/1013, £ for Mann-Whitney test comparing PBS and AMD3100 or rapamycin. (B) Experimental procedure for in vivo AMD3100 treatment. At the end of the treatment, peripheral blood and BM were harvested 2 h after the last injection and analyzed by flow cytometry. (C)Absolute numbers of MPP4 were determined by flow cytometry in the marrow fraction of treated mice. n= 3 independent experiments for a total of 7 WT, 7 +/1013, and 5 1013/1013 mice (PBS condition) and a total of 10 WT, 10 +/1013, and 7 1013/1013 mice (AMD3100 condition). (D) Numbers of B cells (B220⁺) and T cells (CD3⁺) were determined in the peripheral blood of treated mice every 7 days. n= 3 independent experiments for a total of 7 WT, 7 +/1013, and 5 1013/1013 mice (PBS condition) and a total of 10 WT, 10 +/1013, and 7 1013/1013 mice (AMD3100 condition). Statistical significance compared to WT cells at day 21. (E) Representative histograms for TMRE staining in MPP4 from treated mice. The graph shows proportions of TMRE^low/- fraction B among WT or knockin MPP4. n= 3 independent experiments for a total of 5 WT, 6 +/1013, and 5 1013/1013 mice (PBS condition) and a total of 9 WT, 9 +/1013, and 8 1013/1013 mice (AMD3100 condition). (F) Representative histograms for intracellular detection of phospho-mTOR in TMRE^low MPP4. Fluorescence in absence of phospho-mTOR antibody (dashed line), in absence of Cxcl12 stimulation (gray histogram) or presence of Cxcl12 and AMD3100 (black histogram) are shown. MFI values for phosphorylated mTOR (p-mTOR) were determined by flow cytometry. n= 3 independent experiments for a total of 3 WT, 3 +/1013, and 4 1013/1013 mice. (G) Representative flow cytometric analyses comparing the frequencies of B cells (B220⁺CD11b⁻) between 1013/1013 MPP4/OP9 cocultures at day 7 in the absence or presence of rapamycin. Graph shows the proportions of B cells after coculturing MPP4 with OP9/IL-7 stromal cells in the absence or presence of rapamycin after 7 days. n= 3 independent experiments for a total of 3 mice per group. (H) Absolute numbers of MPP4 determined by flow cytometry in the marrow fraction of treated mice 2 h after the last injection of rapamycin. n= 3 independent experiments for a total of 5 WT, 6 +/1013, and 3 1013/1013 mice (PBS condition) and a total of 9 WT, 10 +/1013, and 5 1013/1013 mice (rapamycin condition). (I) TMRE staining profiles in MPP4 from treated mice. n= 3 independent experiments for a total of 5 WT, 6 +/1013, and 3 1013/1013 mice (PBS condition) and a total of 9 WT, 10 +/1013, and 5 1013/1013 mice (rapamycin condition). (J) Numbers of B and T cells determined in the peripheral blood of treated mice every 7 days. n= 3 independent experiments for a total of 5 WT, 6 +/1013, and 3 1013/1013 mice (PBS condition) and a total of 9 WT, 10 +/1013, and 5 1013/1013 mice (Rapamycin condition). ^# P < 0.05; ^## P < 0.005; ^### P < 0.0005 for Kruskal–Wallis tests; * P < 0.05; ** P < 0.005; for Dunn’s test; ^£ P < 0.05; ^££ P < 0.005; ^£££ P < 0.0005 for Mann-Whitney’s test.

Fig. 8.. Model for CXCR4-mediated control of MPP4 fate.

Collectively, our findings reveal an important role for the Cxcr4-mTOR signaling axis in regulating the fate of MPP4 and their mitochondrial machinery in mice. Cxcr4 desensitization is required for the efficient generation of MPP4 from MPP1 and seems also necessary for their positioning near perivascular niches, allowing their access to external cues such as IL-7. Proper Cxcr4 signaling dampens MPP4 cycling and mTOR-dependent metabolic changes that promote myeloid differentiation, thereby allowing for the maintenance of high mitochondrial membrane potential (TMRE^high) and the lymphoid potential of MPP4. In the absence of Cxcr4 desensitization in WS mice, MPP1 displayed impaired capacity to differentiate into MPP4, and MPP4 localized at greater distances from arteriolar cells, likely resulting in suboptimal access to critical niche factors. Aberrant Cxcr4 signaling in MPP4 leads to enhanced mTOR signaling, overactive OXPHOS metabolism, and accumulation of damaged mitochondria with low membrane potential (TMRE^low), thereby promoting myeloid differentiation at the expense of the lymphoid lineage. Modulation of the Cxcr4-mTOR signaling axis using AMD3100 or rapamycin rescued the fate of MPP4 and their mitochondrial machinery in WS mice.