High mitochondrial mass is associated with reconstitution capacity and quiescence of hematopoietic stem cells

Hematopoietic stem cells (HSCs) are capable of self-renewal and multilineage differentiation.^1,2  HSCs prefer glycolysis over mitochondrial oxidative phosphorylation for energy production.^3-5  This metabolic preference allows HSCs to limit reactive oxygen species (ROS) production and maintain HSC potential.^6,7  Although the function of mitochondria in HSCs is not fully understood, recent reports have suggested that they play important roles other than adenosine triphosphate production.^8,9  Contrary to the requirement to maintain low ROS levels, HSCs have recently been reported to exhibit higher mitochondrial mass (a major source of ROS) compared with more differentiated cells (Lin⁻Sca-1⁻c-Kit⁺ cells),¹⁰  which seemingly contradicts previous views that HSCs exhibit a low mitochondrial profile. However, whether mitochondrial mass varies among HSCs and is associated with cell cycle state and HSC function are largely unknown.

Here, we show, using superresolution imaging of HSCs from mito-Dendra2 (PhAM) mice¹¹  and droplet digital polymerase chain reaction (ddPCR), that mitochondrial mass varies within the HSC population. mito-Dendra2 fluorescence intensity in HSCs positively correlates with the expression of endothelial protein C receptor (EPCR).^12,13  Functionally, we show that HSCs with higher mitochondrial mass are more quiescent and have greater reconstitution capacity. Overall, this study clarifies the association between mitochondrial mass and HSC function and demonstrates that quiescent and potent HSCs are highly enriched within the mito-Dendra2^hi HSC fraction.

All experiments were performed according to protocols approved by the National University of Singapore Institution of Animal Care and Use Committee, and the Office of Safety, Health, and Environment. Detailed descriptions for all methods are described in the supplemental Methods.

To address whether mitochondrial mass varies among HSCs, we first measured the fluorescence intensity of bone marrow (BM) HSCs (CD150⁺CD48⁻ Lin⁻Sca-1⁺c-Kit⁺ [SLAM LSK] cells)¹⁴  obtained from mito-Dendra2 mice, which express Dendra2 specifically in mitochondria¹¹  (Figure 1A; supplemental Figure 1A). Because mito-Dendra2 fluorescence intensity varied among HSCs, we analyzed mito-Dendra2^lo and mito-Dendra2^hi HSCs, defined as HSCs with the bottom 10% and top 10% of mito-Dendra2 fluorescence intensity, respectively (Figure 1A). We quantitated the mitochondrial mass of mito-Dendra2^lo and mito-Dendra2^hi HSCs using superresolution imaging of mitochondria with z-stacks.¹⁵  In 3-dimensional imaging, optical aberrations due to refractive index mismatch between oil and mounting medium and imperfect spherical aberration correction lead to point spread function (PSF) enlargement (Figure 1B, left panel). Thus, we corrected PSF enlargement using images of fluorescent microspheres (Figure 1B, right panel) and quantified 3-dimensional mitochondrial mass in HSCs (Figure 1C). Notably, we observed higher mitochondrial mass in mito-Dendra2^hi HSCs compared with mito-Dendra2^lo HSCs (Figure 1C-D). To confirm this observation, we used ddPCR for absolute quantification of mitochondrial DNA copy number (Figure 1E) and found that mito-Dendra2^hi HSCs had higher mitochondrial DNA copy number than mito-Dendra2^lo HSCs (Figure 1F). These results indicate that mitochondrial mass varies among HSCs and confirm that mito-Dendra2^hi HSCs have higher mitochondrial mass than mito-Dendra2^lo HSCs.

In addition, the expression levels of electron transport chain–related genes were lower in mito-Dendra2^hi HSCs (supplemental Figure 1B), whereas the expression of glycolysis-related, mitochondrial fission–related, and mitochondrial fusion–related genes was comparable between mito-Dendra2^lo and mito-Dendra2^hi HSCs (supplemental Figures 1C and 2A). Consistent with the lower expression of electron transport chain–related genes, mito-Dendra2^hi HSCs also displayed lower intracellular ROS levels than mito-Dendra2^lo HSCs (Figure 1G-H), thereby suggesting that electron transport chain activity is lower in mito-Dendra2^hi HSCs.

To further characterize the mito-Dendra2^hi HSC fraction, we compared the expression of HSC-related genes between mito-Dendra2^lo and mito-Dendra2^hi HSCs (Figure 2A; supplemental Figure 3A). Interestingly, mito-Dendra2^hi HSCs expressed significantly lower levels of Cd34 and higher levels of Procr (encodes EPCR) compared with mito-Dendra2^lo HSCs. Because EPCR expression has been reported to strongly correlate with HSC function,¹²  we next compared the expression of EPCR between mito-Dendra2^lo and mito-Dendra2^hi HSCs. The frequency of EPCR^hi HSCs was higher within the mito-Dendra2^hi HSC fraction than within the mito-Dendra2^lo HSC fraction (Figure 2B-C). In addition, the frequency of EPCR^hi HSCs within the mito-Dendra2^hi HSC fraction was significantly higher than the frequency of EPCR^hi HSCs in the total HSC population (Figure 2C). We also compared the mito-Dendra2 fluorescence intensity between EPCR^lo/− and EPCR^hi HSCs; importantly, the median fluorescence intensity was higher in EPCR^hi HSCs (Figure 2D-E). When mito-Dendra2 fluorescence intensity was plotted against EPCR expression, it correlated positively with EPCR expression in HSCs (Figure 2F), thus suggesting that mito-Dendra2^hi HSCs contain HSCs with greater HSC potential.

We next studied the association between mitochondrial mass and cell cycle state and observed that mito-Dendra2^hi HSCs exhibited a higher frequency of quiescent (G0) cells than did mito-Dendra2^lo HSCs (Figure 2G-H). Correspondingly, mito-Dendra2^hi HSCs had lower expression of the cell cycle–associated genes Myc and Cdk6, but higher expression of the negative cell cycle regulator Cdkn1c, compared with mito-Dendra2^lo HSCs (supplemental Figure 3B). Under stress hematopoiesis induced by 5-fluorouracil (5-FU) injection, the number of Lin⁻EPCR^hiCD150⁺CD48⁻ HSCs⁹  decreased 2 days after 5-FU injection and recovered 7 days after 5-FU injection (supplemental Figure 4A). On the other hand, the number of mito-Dendra2^hi Lin⁻EPCR^hiCD150⁺CD48⁻ HSCs remained low 7 days after 5-FU injection (supplemental Figure 4B), which is consistent with cell cycle entry and proliferation of most HSCs 7 days after 5-FU injection.⁹  Therefore, these observations suggest that mito-Dendra2^hi HSCs are more quiescent than mito-Dendra2^lo HSCs.

To determine whether mitochondrial mass is associated with reconstitution capacity in HSCs, we conducted competitive BM transplantation of mito-Dendra2^lo or mito-Dendra2^hi HSCs with Ly5.1 competitor cells into lethally irradiated Ly5.1 recipients. Recipients transplanted with mito-Dendra2^hi HSCs exhibited significantly higher chimerism of PB mononuclear cells (MNCs) (supplemental Figure 5A-B) and BM HSCs (supplemental Figure 5C-D) compared with recipients transplanted with mito-Dendra2^lo HSCs. Secondary BM transplantation also revealed higher PB MNC and BM HSC chimerism in mice transplanted with mito-Dendra2^hi HSCs (supplemental Figure 5E-F), suggesting that long-term HSCs are enriched in the mito-Dendra2^hi HSC fraction. No difference in homing capacity was observed between mito-Dendra2^lo and mito-Dendra2^hi LSK cells (supplemental Figure 6A). Using additional gating of EPCR, we observed that recipients transplanted with mito-Dendra2^hi EPCR^hiCD150⁺CD48⁻ (ESLAM)¹⁶  LSK cells exhibited higher PB MNC chimerism than recipients transplanted with mito-Dendra2^lo ESLAM LSK cells over 4 months posttransplantation, whereas PB chimerism in recipients with mito-Dendra2^lo ESLAM LSK cells progressively increased (Figure 2I). Moreover, analysis of HSC chimerism in BM 4 months posttransplantation showed higher chimerism in recipients transplanted with mito-Dendra2^hi ESLAM LSK cells compared with those transplanted with mito-Dendra2^lo ESLAM LSK cells which also exhibited significant chimerism of HSCs (Figure 2J). Taken together, these results indicate that mito-Dendra2^hi HSCs have a greater reconstitution capacity than mito-Dendra2^lo HSCs.

In this study, we identified a significant variation in mitochondrial mass among HSCs. HSCs with higher mitochondrial mass were more quiescent and exhibited greater reconstitution capacity. Thus, high mito-Dendra2 fluorescence intensity enriches quiescent and potent HSCs. Because mitochondrial membrane potential has been reported to be low in quiescent HSCs,⁹  our study may suggest a mass/function paradox regarding mitochondria in HSCs. The quality and function of mitochondria in quiescent HSCs should be clarified in the future.

The authors thank Yang Chong, Jessica L. Berlier, Michelle Meng Huang Mok, Tamie Endoh, A’Qilah Banu Binte Abdul Majeed, Muddineni Siva Sai Naga Anurag, Nur Humaira Binte Fua’Di, Samuel Ko, Aled O’Neill, Malini Rethnam, and Shu Ying Lee for advice, help, and discussions regarding experiments.

This work was supported by the National Research Foundation of Singapore and the Singapore Ministry of Education under its Research Centres of Excellence initiative, the Singapore Ministry of Health’s National Medical Research Council (NMRC) Singapore Translational Research (STaR) Investigator Award (NMRC/STaR/0019/2014) (T.S.) and Clinician Scientist–Individual Research Grant–New Investigator Grant (NMRC/CNIG/1181/2018) (Y.T.), the Uehara Memorial Foundation Overseas Research Fellowship (Y.T.), and the Japan Ministry of Education, Culture, Sports, Science and Technology (MEXT) KAKEN Grant-in-Aid for Scientific Research (S) (26221309) (T.S.).

Contribution: Y.T. and A.N.-I. designed the study and wrote the first draft of manuscript; D.Q.T. and M.F. revised the manuscript; and all authors analyzed and interpreted the results and read and approved the final version of the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Toshio Suda, Cancer Science Institute of Singapore, Centre for Translational Medicine, National University of Singapore, 14 Medical Dr, #12-01, Singapore 117599, Singapore; e-mail: csits@nus.edu.sg.