Mitochondrial metabolism in hematopoietic stem cells requires functional FOXO3

Abstract

Hematopoietic stem cells (HSC) are primarily dormant but have the potential to become highly active on demand to reconstitute blood. This requires a swift metabolic switch from glycolysis to mitochondrial oxidative phosphorylation. Maintenance of low levels of reactive oxygen species (ROS), a by-product of mitochondrial metabolism, is also necessary for sustaining HSC dormancy. Little is known about mechanisms that integrate energy metabolism with hematopoietic stem cell homeostasis. Here, we identify the transcription factor FOXO3 as a new regulator of metabolic adaptation of HSC. ROS are elevated in Foxo3(-/-) HSC that are defective in their activity. We show that Foxo3(-/-) HSC are impaired in mitochondrial metabolism independent of ROS levels. These defects are associated with altered expression of mitochondrial/metabolic genes in Foxo3(-/-) hematopoietic stem and progenitor cells (HSPC). We further show that defects of Foxo3(-/-) HSC long-term repopulation activity are independent of ROS or mTOR signaling. Our results point to FOXO3 as a potential node that couples mitochondrial metabolism with HSC homeostasis. These findings have critical implications for mechanisms that promote malignant transformation and aging of blood stem and progenitor cells.

Keywords: FOXO3; HSC; ROS; metabolism; mitochondria.

ROS levels and mitochondrial membrane potential in HSPC Endogenous ROS levels were measured in WT and Foxo3^−/− LSK cells by flow cytometry.

1. A, B Representative FACS histogram (left panel) and geometric mean fluorescence intensity normalized with WT (right panel) of ROS levels by chloromethyl dichlorodihydrofluorescein diacetate, CM-H2DCFDA (A), and mitoSOX Red (B) fluorescence are shown. One representative of three independent experiments (n = 3 mice per genotype) is shown.

2. C, D Representative FACS histograms of DiIC1(5) (C) and contour plots of JC-1 red and green fluorescence (D) in WT LSK treated (CCCP) or not (WT) with a ΔΨ_m inhibitor CCCP (carbonyl cyanide 3-chlorophenylhydrazone) 50 mM in vitro for 20 min.

3. E Histogram of TMRE fluorescence displaying shifts in fluorescence intensity after treatment with either CCCP or oligomycin in BM cells.

Figure 1

Mitochondrial dysfunction in Foxo3^−/− HSC

1. A–C Mitochondrial parameters as ATP level (A), oxygen consumption (B), and lactate production (C) were measured in WT and Foxo3^−/− LT-HSC (LSKCD34⁻Flk2⁻) (n = 10 mice in each group), and experiments were performed in triplicate.

2. D–F Mitochondrial mass (D) and membrane potential (E, F) were measured in freshly isolated primitive hematopoietic stem and progenitor cells. (D) One representative FACS plot of the mitochondrial mass measured by the geometric mean fluorescence intensity of MitoTracker Green of 2 (LT-HSC, LSKCD48⁺CD150⁻) or 3 (LSK and c-Kit⁺) independent experiments (n = 3 mice per genotype) is shown: LT-HSC, P = 0.324; LSK, P = 0.021; and c-Kit⁺, P = 0.092. (E) One representative FACS plot of the mitochondria membrane potential measured by the geometric mean fluorescence intensity of 1,1′,3,3,3′,3′-hexamethylindodicarbo-cyanine iodide [DiIC1(5)] of 2 (LT-HSC, LSKCD48⁺CD150⁻) or 3 (LSK and c-Kit⁺) independent experiments (n = 3 mice per genotype) is shown: LT-HSC, P = 0.046; LSK, P = 0.042; and c-Kit⁺, P = 0.478. (F) Mitochondria membrane potential was also measured using 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolyl-carbocyanine iodide (JC-1) probe by flow cytometry. Monomeric JC-1 has a green fluorescent emission spectrum while its aggregated form has a red fluorescent emission spectrum. As JC-1 probe accumulates, its aggregates and shifts fluorescent color. Representative FACS plots of JC-1 red and green fluorescence and relative ΔΨ_m measured by the red/green fluorescence ratio are shown for each population (n = 3 mice per genotype for LT-HSC and n = 6 mice per genotype for LSK and c-Kit⁺ cells).

3. G Histogram (top) and quantification (bottom) of TMRE fluorescence intensity comparing mitochondrial membrane potential between WT and Foxo3^−/− LT-HSC. TMRE florescence normalized to WT TMRE levels in LT-HSC (n = 3 mice per group).

Data information: All data are expressed as mean ± SEM (Student’s t-test, *P < 0.05).

Mitochondrial morphology in Foxo3^−/− LSK cells

1. FACS-sorted LSK cells (10,000 cells) cytospun onto slides and stained with TOM20 (mitochondria-specific antibody) and a secondary conjugated to Alexa Fluor 488 and mounted with DAPI (60×). Images analyzed using ImageJ and the Mito-Morphology macro . Representative images are displayed with nucleus in blue and mitochondria in red. Outlines of mitochondrial networks based on TOM20 fluorescence intensity set at a threshold ≥ 170 a.u. are shown.

2. Quantification of perimeter and area of distinct mitochondrial networks using the Mito-Morphology macro. The average mitochondrion perimeter (left) or area (right) per cell analyzed and error bars represent SEM. *P < 0.05, Student’s t-test.

3. The circularity (left) of mitochondria was quantified, with higher values representing more circular mitochondria. Quantification of mitochondrial network interconnectivity (right) based on a ratio of the area to the perimeter, normalized to the circularity. Bars represent mean ± SEM. *P < 0.05, Student’s t-test.

Figure 2

Inhibition of ROS does not rescue Foxo3^−/− LT-HSC phenotype

1. WT and Foxo3^−/− mice were injected intraperitoneally daily with 100 mg/kg body weight of N-acetyl-L-cysteine (NAC) or vehicle control for 2 weeks (left panel). Total number of bone marrow (BM) cells isolated from one femur and one tibia (n = 6 mice per genotype) (right panel).

2. ROS levels were measured by flow cytometry in WT and Foxo3^−/− LSK cells isolated after 2 weeks. Fold change of geometric mean normalized with WT of DCF fluorescence (n = 6 mice per genotype).

3. Total number of LSK cells in the BM (n = 6 mice per genotype).

4. Total number of LSKCD48⁻CD150⁺ (LT-HSC) cells in the BM (n = 6 mice per genotype).

5. Contribution of LT-HSC (CD45.1) from (A) to the peripheral blood (PB) of recipient mice (CD45.2) in a long-term competitive repopulation assay (n = 5 in each group).

6. Contribution of LSKCD48⁻CD150⁺ (LT-HSC) (CD45.1) isolated from NAC- or vehicle control-treated mice for 4 weeks to long-term competitive repopulation of lethally irradiated mice treated with NAC or vehicle control during post-transplantation (P.T.). One of two experiments is shown.

Data information: Data are expressed as mean ± SEM (n = 5 mice transplanted per group); NS, not significant; *P < 0.05, Student’s t-test.

Oxyblot of NAC-treated BM cells (Top) WT and Foxo3^−/− mice treated with either PBS or NAC were sacrificed, and protein isolated from total BM cells was subjected to protein oxidation analysis. Blots comparing extent of protein oxidation in PBS-treated animals (left blot) and NAC-treated animals (right blot). Each number represents a single mouse with DNPH (+)-treated and DNPH (−)-negative controls in adjacent lanes. (Bottom) Quantification of oxidized proteins detected by oxyblot comparing WT and Foxo3^−/− mice treated with vehicle (PBS) or NAC. Values were normalized to DNPH (−) controls. Bars represent mean ± SEM. n.s., not significant; *P < 0.05, Student’s t-test.

Effects of NAC on HSPC ROS and mitochondrial membrane potential

1. Quantification of ROS levels by DCF probe in LT-HSC after 4-week treatment with NAC or vehicle control prior to transplantation. Bars represent DCF fluorescence (mean ± SEM,n = 3). *P < 0.05, Student’s t-test.

2. Quantification of ROS levels in donor-derived peripheral blood cells at 16 weeks post-transplantation comparing all combinations of post- and pre-transplantation treatment with NAC. 16-week PB donor cells (ROS), post-transplant vehicle: WT-C = 4, KO-C = 4, WT-NAC = 5, KO-NAC=3; post-transplant NAC: WT-C = 4, KO-C = 3, WT-NAC = 3, KO-NAC = 1. Bars represent mean ± SEM.

3. Mitochondrial membrane potential quantification with TMRE in LSK cells. TMRE fluorescence (mean ± SEM), n = 3. n.s., not significant; *P < 0.05, Student’s t-test.

4. Quantification of ROS levels by DCF probe in LSK cells after treatment with NAC or vehicle control. Bars represent DCF fluorescence (mean ± SEM). *P < 0.05, Student’s t-test.

Figure 3

Reducing ROS levels does not improve mitochondrial dysfunction in Foxo3^−/− HSPC

1. Mice were treated with NAC or vehicle control, and mitochondrial membrane potential was measured in freshly isolated cells after 2 weeks. Fold change of geometric mean of DilC1(5) fluorescence measured in LSK cells (n = 3 mice per genotype) normalized with vehicle-treated (WT).

2. Relative ΔΨ_m measured in LSK cells by the red/green fluorescence ratio of JC-1 probe (n = 6 mice per genotype).

3. TMRE quantification of mitochondrial membrane potential in LT-HSC isolated from mice treated with NAC or vehicle control (n = 3 mice).

Data information: All data are expressed as mean ± SEM; NS, not significant; *P < 0.05, Student’s t-test. Controls are the same as in Fig1G.

Figure 4

Alteration of mitochondrial and metabolic gene expression in Foxo3^−/− HSPC

1. qRT–PCR analysis of genes whose products are implicated in mitochondrial metabolism. Results are relative to WT set to one in each population.

2. qRT–PCR analysis of glycolytic genes. Results are relative to WT set to one in each population.

Data information: Results are from three independent experiments each based on three replicates of one cDNA generated from a pool of cells isolated from three mice. All data are expressed as mean ± SEM (Student’s t-test, *P < 0.05).

Figure 5

Inhibition of mTOR signaling normalizes ROS levels and mitochondrial parameters in Foxo3^−/− HSPC

1. WT and Foxo3^−/− mice were injected intraperitoneally with 4 mg/kg body weight of rapamycin on 5 consecutive days/week daily for 2 weeks (left upper panel). Analysis of phosphorylated form of S6K1 target ribosomal protein S6 (pS6) by flow cytometry in LSK cells isolated from WT and Foxo3^−/− mice treated with rapamycin or vehicle control by flow cytometry measuring fold change of geometric mean of pS6 compared to negative control (Neg cont: anti-rabbit IgG-PE antibody only; right upper panel) and the percentage of pS6-positive cells (right lower panel). Representative FACS plots of pS6 expression (left lower panel). The anti-rabbit IgG-PE antibody only (Neg cont) and rapamycin treatment (3 μM) for 1 h in vitro before the LSK staining were used as negative controls of pS6 antibody. One representative of three independent experiments (n = 3 mice per group) is shown.

2. Fold change of geometric mean normalized with WT of DilC1(5) fluorescence measured in LSK cells isolated in (A) (right panel) (n = 6 mice per genotype).

3. Relative ΔΨ_m measured in LSK cells isolated in (A) by the red/green fluorescence ratio of JC-1 probe (n = 6 mice per genotype).

4. Fold change of geometric mean normalized with WT of DCF fluorescence (n = 6 mice per genotype).

Data information: All data are expressed as mean ± SEM; NS, not significant; *P < 0.05, Student’s t-test.

Figure 6

Inhibition of mTOR does not rescue the Foxo3^−/− HSC long-term repopulation activity

1. Total number of bone marrow (BM) cells isolated from one femur and one tibia from WT and Foxo3^−/− mice treated with rapamycin or vehicle control (n = 10 mice per group).

2. Total number of LSK cells in the BM of mice from (A) (n = 10 mice per genotype).

3. Frequency of Lin⁻Sca-1⁺c-Kit⁺ cells (n = 10 mice per genotype).

4. Total number of LSKCD48⁻CD150⁺ (LT-HSC) cells in the BM of mice from (A) (n = 6 mice per genotype).

5. The contribution of transplanted LT-HSC (CD45.1) cells isolated from (A) to the PB of recipient mice (CD45.2) in a long-term competitive repopulation assay (n = 8 in each group).

6. Mean values of cell cycle distribution measured by Ki-67/DAPI staining of LSK cells isolated from (A) (n = 6 mice per genotype).

Data information: All data are expressed as mean ± SEM; NS, not significant; *P < 0.05, Student’s t-test.

Figure 7

Model of FOXO3 regulation of mitochondrial function in HSC Loss of FOXO3 leads to mitochondrial defects that might be implicated in the loss of Foxo3-deficient HSC long-term competitive repopulation ability.