Specification of haematopoietic stem cell fate via modulation of mitochondrial activity

Abstract

Haematopoietic stem cells (HSCs) differ from their committed progeny by relying primarily on anaerobic glycolysis rather than mitochondrial oxidative phosphorylation for energy production. However, whether this change in the metabolic program is the cause or the consequence of the unique function of HSCs remains unknown. Here we show that enforced modulation of energy metabolism impacts HSC self-renewal. Lowering the mitochondrial activity of HSCs by chemically uncoupling the electron transport chain drives self-renewal under culture conditions that normally induce rapid differentiation. We demonstrate that this metabolic specification of HSC fate occurs through the reversible decrease of mitochondrial mass by autophagy. Our data thus reveal a causal relationship between mitochondrial metabolism and fate choice of HSCs and also provide a valuable tool to expand HSCs outside of their native bone marrow niches.

Figure 1. Multi-lineage reconstitution capacity is restricted to the low mitochondrial activity cell fractions.

(a) Competitive transplantation strategy used to assess multi-lineage blood reconstitution levels from peripheral blood after 4, 8 and 16 weeks. (b,c) Within LKS, which contain all multipotent stem and progenitor cells in the BM, long-term multi-lineage HSC function is restricted to TMRM^low cells (LKS:TMRM^low) (n=8 for each condition; error bar: s.e.m.; t-student, ***P<0.001). (d,e) In the phenotypically defined ST-HSC compartment, stemness is restricted to TMRM^low cells (ST-HSC:TMRM^low) (n=9 for each condition; error bar: s.e.m.; t-student, **P<0.01). (d,f) In the phenotypically defined LT-HSC compartment, stemness is restricted to TMRM^low cells (LT-HSC:TMRM^low) (n=9 for each condition; error bar: s.e.m.; t-student, ***P<0.001).

Figure 2. A low mitochondrial activity marks self-renewing HSCs.

(a) Experimental paradigm to assess mitochondrial activity as a discriminator of self-renewing from differentiating LT-HSCs in culture. Freshly isolated LT-HSCs were labelled with CFSE. At the end of the culture, cells that divided one time were further sorted into TMRM^low and TMRM^high phenotypes and transplanted into lethally irradiated recipient mice together with helper cells. Blood reconstitution was assessed at 4, 8 and 16 weeks. (b) CFSE-labelled LT-HSCs were cultured under expansion conditions for 2 days and progeny that underwent one division were sorted into TMRM^low and TMRM^high phenotypes, and 100 cells of each population (along with 2 million helper cells) were transplanted into lethally irradiated recipients. (c) The TMRM^low fraction of the first generation of daughter cells (that is, dividing one time) exhibited strikingly higher long-term multi-lineage blood reconstitution efficiency compared with TMRM^high cells, providing evidence for self-renewing versus differentiating HSC divisions in culture (n=10 for each condition). Assessment of blood chimerism is shown for total blood (top panel) as well as the lymphoid and myeloid lineages (bottom panels; error bar: s.e.m.; t-student, ***P<0.001). (d) BM derived from each of the TMRM^low primary recipients (from c) was injected into four secondary recipient mice after 1 year of the primary transplant. Blood chimerism (average of four secondary recipients corresponding to each primary recipient) show long-term multi-lineage reconstitution in secondary transplants (error bar: s.e.m.).

Figure 3. Quiescent and cycling HSC populations in the native niche have comparable mitochondrial activity levels.

(a) Cell cycle analysis using Ki67 and Hoechst staining on freshly isolated HSCs (LKS CD150+ CD48− CD34−) indicate that more than 70% of the cells are in a quiescent state (G₀, red), with the remaining cells cycling (G₁+S/G₂-M, green) (n=3). (b) Flow cytometry analysis of quiescent and cycling HSCs based on mitochondrial activity labelled with MitoTracker Deep Red. Both, quiescent and cycling HSCs show overlapping mitochondrial activity profiles. The proportion of low and high mitochondrial activity cells within the quiescent and cycling HSC populations is similar (MitoTracker^low: P=0.44, MitoTracker^high: P=0.81; n=3; error bar: s.e.m., t-student). (c) IFN-α stimulation results in in vivo activation of HSCs as demonstrated by Hoechst/Ki67 staining (n=3). (d) Flow cytometry analysis shows overlapping MitoTracker profiles of quiescent and cycling HSCs in IFN-α condition. Similarly, the proportion of MitoTracker^low and MitoTracker^high cells in quiescent and cycling HSCs remains comparable (MitoTracker^low: P=0.32, MitoTracker^high: P=0.54; n=3; error bar: s.e.m., t-student), suggesting that mitochondrial activity is independent of HSC cell cycle state even under acute stress conditions.

Figure 4. Modulation of mitochondrial metabolism alters HSC fate through autophagy.

(a) LT-HSC:TMRM^low (LKS CD150+ CD34− TMRM^low) cultured for 5 days under differentiation conditions in the presence or absence of FCCP were transplanted in lethally irradiated recipient mice together with 2*10⁶ helper cells. (b) Cells cultured in the presence of FCCP show high levels of multi-lineage reconstitution, in contrast to controls lacking FCCP that result in rapid differentiation (error bar: s.e.m.; t-student, **P<0.01 and *P<0.05). (c) Mitochondrial mass (TOMM20, cyan) and autophagosomes (LC3B, red) in different haematopoietic stem/progenitor cell populations. Images represent maximum intensity projections of the corresponding Z-stacks (0.28 μm step) Scale bar: 5 μm. (d–f) Quantification of the mito-EGFP (d), LC3B (e) and mito-EGFP/LC3B ratio intensity-thresholded areas (f). (g) Gene expression analysis on key autophagosomal and mitophagic genes was performed on LT-HSCs treated over 5 days with FCCP (error bar: s.e.m.; t-student, ***P<0.001, **P<0.01 and *P<0.05).