Mitochondria: the key hub for hematopoietic stem cell homeostasis maintenance and fate determination

Abstract

Hematopoietic stem cells (HSCs) constitute the pivotal cellular subset sustaining long-term hematopoietic homeostasis, characterized by robust self‑renewal and multilineage differentiation potential. Under physiological conditions, HSCs undergo stepwise differentiation through rigorously controlled regulatory networks to produce a full repertoire of mature blood cells, fulfilling basal physiological demands. Upon exposure to stress or pathological insults (e.g., bone marrow niche dysregulation), HSCs rapidly activate emergency regenerative programs to reconstitute hematopoietic function and restore systemic homeostasis. As intracellular “powerhouses” and central hubs of metabolic regulation, mitochondria exert profound regulatory effects on HSCs fate determination. The dynamic balance of mitochondrial metabolism not only furnishes HSCs with sufficient bioenergy but also generates critical metabolic intermediates; meanwhile, the fine-tuning of oxidative stress and autophagic machinery ensures mitochondrial network integrity. These biological processes are intricately intertwined, forming a complex regulatory network that profoundly modulates HSCs self-renewal, lineage commitment, and long-term hematopoietic reconstitution potential. This review systematically dissects the multi-dimensional regulatory mechanisms by which mitochondria govern HSCs, elaborates on the synergistic interactions and antagonistic effects among distinct components of the regulatory circuitry, and defines the pivotal role of mitochondria in sustaining HSCs homeostasis and orchestrating their repair responses to cellular damage. This work establishes a novel theoretical framework for devising mitochondrial-targeted interventions to sustain metabolic homeostasis in HSCs. Furthermore, it lays a solid scientific foundation for the treatment of hematological diseases and the development of precision therapeutic strategies, offering new insights into the clinical management of hematopoietic disorders.

Introduction

Blood is a connective tissue composed of approximately 34% cells and 66% plasma, which transports nutrients, gases, and molecules throughout the body under physiological conditions. Hematopoiesis, the process of generating all blood cell lineages, is essential for maintaining hematopoietic function throughout life. This process involves the highly regulated differentiation and proliferation of hematopoietic stem cells (HSCs) [1]. As a unique subset of adult stem cells, HSCs possess self-renewal capacity and pluripotency—the potential to differentiate into all hematopoietic cell lineages. The origin of HSCs traces back to embryonic organogenesis [2]. Studies have confirmed that the initial wave of primitive hematopoiesis occurs in the blood islands of the extraembryonic yolk sac [3]. In adult mammals, HSCs reside within specialized bone marrow (BM) niches and predominantly remain in the G0 quiescent stage of the cell cycle. While quiescent, HSCs retain robust self-renewal potential, which sustains lifelong balanced turnover of all blood cell pools. Self-renewal happens through mitotic cell division, resulting in two daughter cells: one keeps the traits of HSCs (long-term hematopoietic stem cells, LT-HSCs), while the other becomes more mature and enters an active phase of the cell cycle (short-term hematopoietic stem cell, ST-HSCs). A range of mature blood cell populations are formed through lineage commitment [4, 5]. Distinct cell populations can be separated based on cell surface marker phenotypes, including erythrocytes, megakaryocytes/platelets, myeloid cells (monocytes/macrophages and granulocytes), mast cells, T and B lymphocytes, natural killer cells (NKs), and dendritic cells (DCs) (Fig. 1 right)—a process that reflects multipotency of HSCs [6]. The type of division described above is defined as asymmetric division, which produces two daughter cells with different functions and destinies.

Fig. 1.

Schematic overview of hematopoietic stem cell differentiation and mitochondrial structure/function. HSCs hematopoietic stem cells, LT-HSCs long-term hematopoietic stem cells, ST-HSCs short-term hematopoietic stem cells, MPP multipotent progenitor cells, CLP common lymphoid progenitor, CMP common myeloid progenitor, MEP megakaryocyte/erythrocyte progenitor, GMP granulocyte/macrophage progenitor, NK natural killer cell, AS alternative splicing, ATP adenosine triphosphate, GLS glutaminase, PDH pyruvate dehydrogenase, CPT1 carnitine palmitoyltransferase 1, OMM, outer mitochondrial membrane, IMS intermembrane space, IMM inner mitochondrial membrane, ROS reactive oxygen species, TCA cycle tricarboxylic acid cycle. This figure is originally drawn by the authors and does not require any reuse permissions

Mitochondria are considered to be descendants of ancient prokaryotes that underwent endosymbiotic events with early eukaryotes. As double-membrane-bound organelles measuring 0.5–1.0 μm in diameter and 1–2 μm in length, Mitochondria consist of the outer mitochondrial membrane (OMM), intermembrane space (IMS), and inner mitochondrial membrane (IMM). The IMM is the location for the positioning of respiratory chain complexes, which exist as monomers or supercomplexes [7]. Its highly folded structure significantly expands the membrane surface area and forms characteristic cristae-the hallmark morphological feature of mitochondria [8]. Distinct from other cytoplasmic organelles, mitochondria possess their own genome—a circular double-stranded DNA molecule of 16,569 base pairs (bp) encoding 37 genes (13 mRNAs for respiratory chain complexes, 22 tRNAs, and 2 rRNAs) [9].

The full mitochondrial proteome comprises approximately 1000–1500 proteins encoded jointly by both nuclear DNA (nDNA) and mitochondrial DNA (mtDNA). Only 13 oxidative phosphorylation (OXPHOS) related proteins are transcribed from mtDNA, all remaining mitochondrial proteins are encoded by nDNA, translated on cytosolic ribosomes, and subsequently imported into the mitochondrial compartment [10]. Functionally, mitochondria act as core hubs not only for aerobic respiration and cellular energy provision but also for signal transduction, cell proliferation and differentiation, biosynthesis of intermediates, reactive oxygen species (ROS) production, immune signaling, and apoptosis—making them indispensable for maintaining cellular and organismal homeostasis (Fig. 1, left) [11]. In addition, mitochondria are critical for preserving HSC quiescence, governing metabolic homeostasis, and regulating aging progression; their intact functionality directly affects HSCs fate and functional capacity. This review systematically summarizes the critical roles of mitochondria in HSC fate control and outlines recent advances in this field.

Mitochondrial energy metabolism in HSCs fate determination

Mitochondria serve as the foundation for the distinct metabolic processes of HSCs and hematopoietic progenitor cells (HPCs) [12]. LT-HSCs possess a sustained capacity for self-renewal, and their energy metabolism significantly differs from that of more differentiated progeny. LT-HSCs reside in the hypoxic microenvironment of the bone marrow and sustain ATP production primarily via glycolysis. Compared with oxidative phosphorylation, this metabolic strategy limits ROS production, protecting HSCs from oxidative stress and associated degenerative diseases while preserving their long-term functional integrity [13, 14]. Recent studies have indicated that the glycolytic pathway is activated in HSCs during the G1 phase, rather than in G0-phase HSCs [15]. DPY30, an epigenetic modulator that enhances H3K4 methylation, directly regulates the expression of multiple key glycolytic genes. Jiang et al. reported that HSCs lacking DPY30 and H3K4 methylation exhibit impaired glycolysis and compromised mitochondrial function, along with increased DNA damage. Additionally, they demonstrated that the CDK inhibitor p21 is upregulated in DPY30-deficient hematopoietic stem and progenitor cells (HSPCs), and that the loss of p21 mitigates functional deficits in these cells [16] (Fig. 2A). In the cellular metabolic regulatory network, hypoxia-inducible factor 1-alpha (Hif1α), pyruvate dehydrogenase kinase 2/4 (Pdk2/4), and meis homeobox 1 (Meis1) function cooperatively to balance cellular energy metabolism by dynamically regulating mitochondrial function and glycolytic flux. As a core transcription factor in hypoxia response, Hif1α inhibits PDH activity by inducing PDHK1 expression to restrain mitochondrial oxidative metabolism, while concurrently triggers mitochondrial autophagy to reduce ROS levels [17]. Pdk2/4 promotes the conversion of pyruvate to lactate by phosphorylating PDH, reinforcing the glycolytic pathway [18]. Meis1 maintains the quiescent state of HSCs by regulating glycolytic metabolic programs and the expression of metabolic regulators such as Hif1α and Pdk2/4; it also indirectly controls the expression of mitochondrial metabolic genes, establishing a metabolic mode that adapts glycolytic dominance to mitochondrial function [19]. This regulatory network provides metabolic support for cells to adapt to hypoxic environments and maintain stem cell homeostasis by precisely coordinating the dynamic balance between mitochondrial oxidative metabolism and glycolytic flux.

Fig. 2.

Molecular mechanism underlying mitochondrial control of HSCs fate. A Mitochondrial energy metabolism governing HSCs fate determination. B Mitochondrial-mediated regulation of HSCs redox status. C Mitochondrial characteristics of quiescent versus aged HSCs, alongside phenotypic impacts of dysregulated DNA methylation and mitochondrial dysfunction in Dnmt3a R878H/+ mutant HSCs. D Regulatory effect of mitochondrial signaling on HSCs fate. OXPHOS oxidative phosphorylation, MMP mitochondrial membrane potential (ΔΨm). This figure is originally drawn by the authors and does not require any reuse permissions

Upon activation, HSCs upregulate mitochondrial oxidative phosphorylation to meet the elevated energy demands associated with cell proliferation. Fumarate hydratase (Fh1) acts as a master regulatory factor governing the mitochondrial tricarboxylic acid (TCA) cycle and cytosolic succinate metabolism. Loss of Fh1 function elicits a specific phenotype in HSCs: although the intracellular succinate level remains normal, maximum respiratory capacity is significantly impaired, resulting in impaired self-renewal ability and aberrant lymphoid differentiation [20]. Conversely, increased expression of succinate dehydrogenase assembly factor 1 (SDHAF1) can enhances mitochondrial ATP production efficiency, enables aged stem cells to withstand metabolic stress, and consequently improve cell survival rate [21].

Fatty acid oxidation (FAO), an essential energy metabolic pathway for HSCs, is activated in early hematopoietic cells and gradually decreases with differentiation. FAO generates nicotinamide adenine dinucleotide phosphate (NADPH) and ATP. Beyond mitigating oxidative stress and restraining ROS production, recent work has uncovered an additional vital role for this pathway: FAO-derived NADPH drives cholesterol synthesis in HSCs, forming the FAO-NADPH-cholesterol metabolic axis, thereby regulating the biogenesis and release of extracellular vesicles (EVs). Functional assays demonstrate that blockade of EV signaling significantly impairs the self-renewal capacity of HSCs [22]. Additionally, the promyelocytic leukemia (PML)-peroxisome proliferator-activated receptor δ (PPAR-δ)-FAO signaling cascade is verified as a central axis sustaining HSCs stemness [23]. Carnitine palmitoyl transferase 1a (CPT1A), the key enzyme of FAO, maintains the metabolic balance between fatty acid β-oxidation and OXPHOS. It restrains aberrant mitochondrial overactivation and excessive ROS production, consequently ensuring HSCs quiescence and functional homeostasis [24]. Recent studies have revealed that HSPCs in young mice can compensate for FAO deficiency through metabolic plasticity, so CPT1A deletion has no impact on their function. In contrast, aging diminishes this metabolic plasticity, accordingly, CPT1A loss disrupts normal hematopoietic function [25].

Amino acid metabolism and translation regulation jointly govern HSCs homeostasis and proliferation. Quiescent HSCs activate the general control nonderepressible 2–eukaryotic translation initiation factor 2α (GCN2–eIF2α) axis via robust amino acid catabolism, which in turn suppresses protein synthesis. During proliferation, HSCs enhance OXPHOS energy production and reduce amino acid decomposition to accumulate amino acids. This shift inactivates the GCN2–eIF2α axis and accelerates protein synthesis, concurrent with elevated proteotoxic stress. Gcn2 deficiencydisrupts HSCs functionality, given that this signaling axis modulates mitochondrial OXPHOS by inhibiting AKT activation. The nicotinamide riboside (NR), an NAD precursor, promotes amino acid catabolism, triggers Gcn2 signaling, and sustains the functional integrity of LT-HSCs [26]. Studies have uncovered that serine exhibits distinctive metabolic characteristics in HSCs: it generates NADPH to sustain cellular redox equilibrium, lowering HSCs susceptibility to ferroptosis., Murine HSCs rely on extracellular serine and downstream mitochondrial catabolic process governed by the serine hydroxymethyltransferase 2–methylenetetrahydrofolate dehydrogenase 2 (SHMT2-MTHFD2) axis. Either dietary serine deprivation or pharmacological inhibition of this metabolic cascade renders HSCs prone to ferroptosis, ultimately compromising the maintenance of the HSCs reservoir. Conversely, exogenous serine supplementation mitigates ferroptosis by promoting this mitochondrial serine catabolic process [27].

N6-methyladenosine (m⁶A) represents the most prevalent internal modification on messenger RNA (mRNA), governing diverse cellular processes by regulating RNA stability and translation efficiency. In HSPCs, m⁶A modification dictates mRNA fate and primarily exerts its regulatory functions during the early stage of hematopoiesis. Yin et al. demonstrated that elevated expression of the m⁶A reader insulin-like growth factor 2 mRNA-binding protein (Igf2bp2) control the transcriptional status and maintenance of HSCs. Conversely, Igf2bp2 deficiency accelerates the decay of Bmi1 mRNA, increases the mitochondrial activity of HSCs, and relieves transcriptional repression of mitochondrial related genes [28]. Subsequent studies revealed that the m⁶A demethylase alkB homology 5 (Alkbh5) controls mitochondrial ATP production and adaptive responses of HSPCs in an m⁶A-dependent manner. Alkbh5 ablation elevates global RNA m⁶A methylation, resulting in decreased stability of oxoglutarate-dehydrogenase (Ogdh) mRNA and decreased protein expression. The reduction of Ogdh disrupts the TCA cycle, leading to the accumulation of α-ketoglutaric acid (α-KG) and its conversion to L-2-hydroxyglutaric acid (L-2-HG), which inhibits energy production in mouse and human hematopoietic cells in vitro. Impaired mitochondrial bioenergetics renders HSPCs at a competitive disadvantage and limits the clonogenic capacity of MLL-AF9-driven leukemia cells [29].

Regulation of redox status in HSCs by mitochondria

Mitochondria serve not only as the core organelle for cellular energy metabolism but also as a master signaling hub for regulating cellular redox homeostasis [30]. The production of ROS is closely associated with mitochondrial function, and those generated via electron leakage from mitochondrial OXPHOS complexes I and III account for approximately 0.1%–0.2% of total mitochondrial ROS output [31, 32]. The half-life of ROS ranges from nanoseconds to seconds, with concentrations in the picomolar to nanomolar range [33]. Dynamic levels of ROS significantly regulate the maintenance and fate determination of HSCs [13, 34–36]. Low ROS levels enhance the self-renewal capacity of HSCs, whereas excessive ROS accumulation induces oxidative stress damage [37–39]. Notably, previous studies have shown that ROS-mediated HSCs dysfunction is reversible to a certain extent [18].

Mechanistic target of rapamycin (mTOR) serves as a master upstream regulator of mitochondrial redox balance [40, 41] within the tuberous sclerosis complex (TSC)–mTOR signaling cascade, which maintains the quiescent state of HSCs by inhibiting ROS production. Conditional knockout of Tsc1 gene drives HSCs out of quiescence, accompanied by enhanced mitochondrial biogenesis and elevated ROS levels, ultimately resulting in significant impairment of HSCs self-renewal potential and hematopoietic reconstitution capacity [42] (Fig. 2B). Sterol regulatory element-binding transcription factor 1c (Srebf1c) inhibits mTOR signaling activation and excessive hyperactivity of mitochondrial metabolism by regulating Tsc1 expression. Aberrant expression of Srebf1c not only increase HSCs number and reduces the quiescent population but also significantly impairs HSCs function following competitive transplantation, chemotherapy, or radiotherapy [43]. Jumonji C-domain-containing protein 6 (Jmjd6), a 2-hydroxyglutarate-dependent protein hydroxylase belonging to the Jumonji family, sustains HSCs function integrity via suppression of mitochondrial OXPHOS, p53 stabilization, cell cycle checkpoint progression, and mTORC1 signaling. Antioxidant administration rescues the HSCs functional defects observed under Jmjd6 ablation [44]. In mutant knock-in mice lacking functional C-X-C chemokine receptor 4 (Cxcr4)- a critical mediator of hematopoietic cell migration and homing, uncontrolled mTOR kinase signaling drives compensatory upregulation of mitochondrial OXPHOS, which ultimately leads to myeloid differentiation bias in bone marrow-resident MPP4 cells [45].

Mitochondrial carrier 2 (MTCH2), a ligand of BH3-interacting domain death agonist (BID), is critical for regulating mitochondrial OXPHOS in HSCs. Its absence leads to elevated mitochondrial OXPHOS, driving HSCs and progenitor cells into the cell cycle. This metabolic shift also boosts mitochondrial mass, elevates ATP and ROS levels, and renders cells resistant to radiation-induced apoptosis. In contrast,, the BID phosphorylation-deficient mutant triggers comparable OXPHOS upregulation yet drives more robust ROS accumulation alongside reduced ATP abundance; this mutant phenotype correlates with heightened radiation sensitivity [46].

Delta-like non-canonical Notch ligand 1 (Dlk1), a paternally expressed imprinted gene within the mouse Dlk1-Gtl2 imprinted gene cluster, encodes a non-classical Notch ligand [47]. Studies have revealed that it is highly expressed in LT-HSCs of mice and gradually decreases during differentiation. Deletion of Dlk1 gene does not alter the total LT-HSCs pool size yet reduces their repopulation population potential. Mechanistically, Dlk1 knockout attenuates Notch signaling, elevates ribosome translation, enhances mitochondrial metabolism, increases ROS production and promotes cell cycle progression [48].

Mortalin (motility suppressor protein 2, also known as mtHsp70/GRP75) serves as an essential subunit of the IMM transport complex. Knockdown of mortalin in HSCs disrupts the redox balance, triggering excessive ROS production and accelerated cell cycle entry, leading to significant impairment of long-term repopulation potential. By contrast, mortalin overexpression enhances the stability of HSCs function by inhibiting ROS overproduction and reinforcing stem cell stemness.

Notably, deletion of Parkinson disease protein 7 (PARK7/DJ-1), a gene linked to Parkinson’s disease, leads to hematopoietic defects marked by HSCs depletion and excessive ROS accumulation, which further highlights the importance of mitochondrial homeostasis in regulating HSCs fate [49]. 3 phosphoinositide dependent protein kinase 1 (PDK1) regulates HSCs function by activating the PI3K/AKT signaling pathway. Functional assays reveal that PDK1 deficiency significantly reduces intracellular ROS levels in HSCs. Conversely, treatment with L-buthionine sulfoximine (BSO) to elevate ROS partially rescues the defective colony-forming capacity of PDK1-deficient HSCs [50].

Steroid receptor coactivator 3 (SRC-3/NCOA3) is abundantly expressed in HSCs. The deficiency of SRC-3 leads to a series of alterations: deletion of SRC-3 enhances mitochondrial metabolism and drives ROS overproduction, upregulates peroxisome proliferator activated receptor gamma coactivator 1α (PGC-1α) downstream target genes, suppresses the expression of general control of amino acid synthesis 5 (GCN5/KAT2A), and diminishes PGC-1α acetylation. These changes diminish HSCs quiescence and promote proliferation, sensitizing HSCs to radiotherapy and chemotherapy. Notably, GCN5 overexpression rescues aberrant hyperactive mitochondrial metabolism and mitigates excessive ROS production induced by SRC-3 loss [51].

Recent studies have revealed that NYN domain and retroviral integrase containing (Nynrin) blocks mitochondrial permeability transition pore (mPTP) opening, thereby protecting HSCs from radiation-induced damage. Nynrin is highly expressed in HSCs under both steady-state and stress conditions. Nynrin ablation leads to mitochondrial dysfunction, manifested by aberrant mPTP opening, mitochondrial swelling, and elevated levels of ROS [52].

Zinc finger E-box binding homeobox 1 (Zeb1) serves as a key regulator of epithelial-mesenchymal transformation (EMT). Studies have shown that Zeb1 maintains mitochondrial fission and low metabolic activity by suppressing the expression of mitochondrial fusion protein mitofusin 2 (Mfn2), thereby sustaining HSCs quiescence and stemness. However, Zeb1 deficiency leads to excessive mitochondrial fusion, elevated metabolic activity, and increased ROS levels, ultimately compromising HSCs function and long-term maintenance [53].

Mitochondrial complex II (MCII) is essential for HSCs homeostasis. Studies demonstrate that HSCs from mice harboring succinate dehydrogenase complex subunit C (Sdhc) mutations exhibit exacerbated ROS accumulation and DNA damage upon mitochondrial activation. Loss of MCII function triggers an aging‑like stress response, leading to myeloid‑biased leukopenia, macrocytic anemia, and thrombocytosis [54].

Fanconi anemia group D2 (FANCD2) protein interacts with ATAD3 and TUFM, two core components of the mitochondrial nucleoid complex. The deletion of the Fancd2 gene upregulates mitochondrial protein synthesis, causing an imbalance between mitochondrially encoded and nuclear-encoded mitochondrial proteins. This metabolic perturbation further elevates mitochondria mass, respiratory complex activity, and ROS levels in murine HSPCs. Interestingly, Fancd2-deficient HSPCs exhibit profound hypersensitivity to mitochondrial translation inhibitors, and their survival and proliferation rely on enhanced mitochondrial translation activity [55].

OCIA domain-containing protein 1 (Asrij/OCIAD1) localizes to both the endoplasmic reticulum and mitochondria. By inhibiting ROS accumulation and maintaining mitochondrial homeostasis, this protein delays the aging of HSCs. Asrij knockout in mice drives HSCs exit from quiescence, skews differentiation toward the myeloid lineage, reduces p53 abundance, and exacerbates DNA damage, resulting in the decline of organelle functions and accelerating the aging of HSCs [56].

Mitochondrial membrane potential: a key regulator of HSCs stemness, quiescence and aging

Mitochondrial signatures unique to quiescent HSCs represent a central research topic, as these organelle properties underpin the lifelong regenerative potential of HSCs. Studies have shown that mitochondrial membrane potential (MMP) distinguishes quiescent HSCs from primed HSCs—cells exiting quiescence that carry elevated MMP levels [57]. Conventionally, quiescent HSCs harbor a low mitochondrial number due to their low mitochondrial activity. Nevertheless, findings from a mouse model using the mitochondrial reporter gene mito-Dendra2 challenge this view, demonstrating that the mitochondrial content of quiescent HSCs is actually very high [58]. Despite high mitochondrial content, mitochondria within quiescent HSCs remain structurally and morphologically immature, featuring a punctate shape, poorly developed cristae, and diminished membrane potential (Fig. 2C). In addition, recent studies have shown that HSCs display greater mitochondrial mass and MMP compared with committed bone marrow populations, and these distinct mitochondrial characteristics are closely linked to their enhanced stemness and functional potency [58–60]. Apparent discrepancies across these datasets can be partially attributed to divergent experimental approaches for quantifying mitochondrial mass and MMP. One analysis of mitochondrial content via pyronin staining demonstrated approximately 90% of low-MMP HSCs resided in G0 phase, while only 55% of high-MMP HSCs remained in G0 phase, indicative of augmented cell cycle activity in the high-MMP subset. BrdU incorporation assays further confirmed that the proliferative capacity of the high-MMP HSCs markedly exceeds that of their low-MMP counterparts [57].

Notably, MMP serves as a critical indicator reflecting the transcriptional and functional status of young and aged HSCs. Although bulk aged HSCs exhibit reduced MMP, a subset retains MMP levels comparable to those of young HSCs, indicating that MMP is a key determinant of heterogeneity within aged HSCs. In addition, the study found that MMP was more accurate than chronological age in predicting the quantitative and qualitative transcription landscape of HSCs. Mitoquinol (Mito-Q) treatment significantly elevates MMP and global transcription rate, reduces intracellular ROS levels in aged HSCs, corrects aging-associated skewing of HSPCs compartments, contracts myeloid-biased subsets, and selectively promotes the proliferation of lymphoid-biased HSCs. RNA-sequencing analysis reveals that aged HSCs treated with Mito-Q exhibit a transcriptional profile closely resembling that of young HSCs, with MMP exerting a more dominant regulatory effect on transcription than chronological age. High MMP upregulates genes associated with stemness, self-renewal and DNA damage repair, markedly enhances DNA double-strand break (DSB) repair capacity in aged HSCs, and restores lymphoid differentiation programs. Collectively, these observations illustrate that elevating MMP partially rejuvenates aged HSCs at phenotypic, transcriptional and functional tiers [61] (Fig. 2C). These findings provide novel mechanistic insights into HSCs senescence and informs prospective intervention strategies.

HSCs are rich in mitochondria but have relatively low metabolic activity, which is more pronounced compared to MPP and more differentiated progeny. Studies have shown that the mitochondrial mass declines during the differentiation of HSCs into MPP, whereas downstream mature progenitor cells (including murine pre-colony-forming unit-erythrocytes (pre-CFU-E) and CFU-E) display progressive mitochondrial accumulation [62]. Upon metabolic activation, HSCs undergo irreversible mitochondrial network remodeling. Such damage cannot be repaired under quiescent conditions, which is distinctly different from HPCs. Even if cells return to a dormant state carrying damaged mitochondria, these mitochondria disrupt the coordination between cell cycle and metabolic regulation during subsequent divisions. This permanent alteration in mitochondrial morphology and division fidelity establishes a unique division memory, ultimately leading to HSCs functional decline [63]. In the DNA methyltransferase 3 alpha (Dnmt3a) mutant (Dnmt3a R878H/+) mouse model, mutant HSPCs gain a competitive advantage by sustaining high mitochondrial respiratory activity, characterized by global DNA hypomethylation, upregulated OXPHOS genes, elevated MMP, and strict reliance on mitochondrial respiration. Mito-Q selectively accumulates in mutant cells due to their high MMP. The compound triggers apoptosis by inhibiting mitochondrial respiration (Fig. 2C), thereby abolishing the competitive advantage of mutant HSPCs in aged mice and in a human Dnmt3a knockdown HSPCs model [64].

Regulatory effect of mitochondrial signaling on HSCs fate

Mitochondrial calcium (Ca²⁺) uptake acts as a core hub bridging intracellular signaling and MMP homeostasis. As a ubiquitous second messenger, Ca²⁺ not only modulates mitochondrial function through regulating MMP but also directly governs the maintenance of HSCs quiescence and stemness [60, 65]. 5-fluorouracil (5-FU)-induced myelosuppression elevates intracellular Ca²⁺ levels in HSCs, which in turn enhances MMP to drives HSCs division. Nevertheless, excessive activation of the mitochondrial Ca²⁺ signaling results in post-mitotic loss of HSCs function. Moderate pharmacological suppression of intracellular Ca²⁺ via exogenous calcium channel blockers prolongs HSCs division intervals, balancing robust proliferative output with sustained stem cell properties [66]. Twist family basic helix-loop-helix transcription factor 1 (Twist1) regulates HSCs function through the CACNA1B/Ca²⁺/mitochondrial axis: loss of Twist1 activates the voltage-gated calcium channel (VGCC) CACNA1B, leading to lymphoid-biased HSCs exhaustion, impaired hematopoietic recovery, and elevating mitochondrial Ca²⁺ load, metabolic flux, and reactive oxygen species production [67]. CD38 is an NAD⁺-dependent metabolic enzyme that catalyzes the hydrolysis of NAD⁺ into cyclic ADP-ribose (cADPR) and ADP-ribose (ADPR) [68]. Together with nicotinamide adenine dinucleotide phosphate (NAADP), these metabolites mobilize Ca²⁺ efflux from intracellular storage organelles into the cytosol, thereby coordinating calcium signal transduction. CD38 has long been implicated in HSCs regeneration; it promotes HSCs quiescence through restraining calcium signaling cascades, a function that is also conserved across multiple hematopoietic and immune cell subsets [69]. However, CD38 expression increases with age, and this upregulation is considered a major contributor to reduced cellular NAD⁺ levels. In young mice, CD38 facilitates HSCs proliferation by inducing mitochondrial Ca²⁺ influx and enhancing metabolism. In contrast, aged mice suffer from profound NAD⁺ metabolic imbalance and impaired mitochondrial stress tolerance, rendering CD38 upregulation detrimental and declined HSCs function. Notably, CD38 inactivation alleviates HSCs senescence and reverses age-related hematopoietic pathophysiological alterations in aged murine models [70]. S100 calcium-binding protein A6 (S100A6) is a canonical calcium sensor broadly expressed in various cell types, with established roles in cell cycle progression, apoptotic regulation, cytoskeletal remodeling, and cellular adaptive responses to diverse stress stimuli. Studies have shown that under cytokine stimulation, S100A6 mediates cytoplasmic and mitochondrial calcium buffering in HSCs. S100A6 deficiency drastically reduces phosphorylated Akt (p-Akt) and Heat shock protein 90 (HSP90), while simultaneously impairing mitochondrial respiratory efficiency and disrupting cellular calcium homeostasis. These changes ultimately result in a decreased total number of HSCs, reduced bone marrow output, increased apoptosis, and consequently severely compromised self-renewal and regenerative capabilities of HSCs [71].

Iron acts as a critical cofactor of mitochondrial respiratory chain complexes, and balanced iron metabolism is vital to sustain intact respiratory chain function. Conditional knockout of ferritin heavy chain 1 (Fth1) in HSCs depletes intracellular iron stores, disrupts the structural integrity of the respiratory chain, and drives robust suppression of MMP. This decrease in MMP directly triggers mitochondrial dysfunction, which not only markedly attenuates HSCs self-renewal capacity but also activates the cellular apoptotic program, ultimately resulting in reduced HSCs number and impaired hematopoietic regenerative potential. Exogenous iron supplementation effectively restores respiratory chain activity and MMP homeostasis. Furthermore, combinatorial treatment with antioxidants and apoptosis inhibitors substantially rescues the stemness-deficient phenotype triggered by Fth1 deficiency [72]. The ABC transporter protein ATP-binding cassette subfamily B member 10 (Abcb10), an ABC transporter localized to the inner mitochondrial membrane, serves as a crucial modulator of HSCs homeostasis and erythroid differentiation. Studies have found that Abcb10 maintains the activity of HSCs by regulating mitochondrial iron homeostasis and antioxidant defense. Abcb10 deficiency leads to mitochondrial iron overload and oxidative stress, but does not affect mitochondrial energy metabolism. In Abcb10 knockout mice, HSCs frequency and repopulating potential are diminished, accompanied by aberrant expansion of, erythroid progenitor cells [73].

Nitric oxide synthase (iNOS) and its catalytic product nitric oxide (NO) are key molecules in cellular signal transduction. Previous studies indicated that under physiological conditions, NO and iNOS levels in HSCs are maintained at relatively low levels. However, recent research demonstrate that upon 5-FU-induced hematopoietic stress, iNOS-deficient mice exhibit increased susceptibility to myelosuppression alongside delayed recovery of hematopoietic progenitors. Further observations reveal that HSCs in these mice exit quiescence and undergo severe mitochondrial stress evidenced by elevated MitoSOX levels and an increased proportion of MMP^high HSCs. These findings establish the iNOS-NO-mitochondrial signaling axis as an essential modulator governing HSPCs function and steady-state hematopoiesis, offering novel mechanistic insights into stress hematopoietic regulation [74].

Accumulating evidence indicates that precise regulation of mitochondrial signaling is critical for governing HSCs fate. Moderate activation of mitochondrial signaling promotes HSCs proliferation by elevating mitochondrial membrane potential, while excessive activation leads to the loss of stemness properties. These findings deepen our mechanistic understanding of mitochondria as a central hub for HSCs function and highlight the therapeutic potential of targeting mitochondria-related pathways, providing an innovative theoretical framework and promising translational prospects for future HSCs research.

Mitochondrial quality control regulates HSCs fate, quiescence, and regeneration

Mitophagy

Mitochondrial autophagy (Mitophagy) plays a crucial role in the maintenance and function of HSCs. By selectively eliminating damaged or dysfunctional mitochondria, mitophagy preserves the quality and functional integrity of mitochondria within HSCs, protecting them from oxidative injury and maintaining their viability and stem cell activity. Additionally, mitophagy sustains the quiescent state of HSCs by removing hyperactive mitochondria urb aberrant metabolic hyperactivation, thereby maintaining HSCs homeostasis [75].

Mitophagy exists through two primary pathways: the ubiquitin-dependent and the non-ubiquitin-dependent pathways. In the ubiquitin-dependent pathway, the PTEN-induced kinase 1 (PINK1)/Parkin pathway plays a central role: Upon loss of MMP, is lost (depolarized), PINK1 stably accumulates on the OMM and undergoes phosphorylation to activate Parkin. Through the ubiquitin cascade reaction, it marks the damaged mitochondria, and at the same time, PINK1 can independently recruit light chain 3 (LC3) adaptor proteins including OPTN/NDP52 to initiate autophagy [76, 77]. The non-ubiquitin-dependent pathway is driven by direct interactions between microtubule-associated protein LC3 and LC3-interacting region (LIR) residing on the OMM. Under stress stimuli such as hypoxia, receptor proteins NIX, BNIP3, and FUNDC1 orchestrate targeted mitophagic clearance of impaired mitochondria [78–80] (Fig. 3).

Fig. 3.

Mitochondrial quality control regulates HSCs fate, quiescence, and regeneration. This figure is originally drawn by the authors and does not require any reuse permissions

Mitophagy, a quality control system responsible for eliminating damaged mitochondria, acts as a core mechanism governing the maintenance and expansion of Tunica interna endothelial cell kinase 2(Tie2)-positive HSCs (Tie2⁺HSCs), and directly modulates their self-renewal capacity. Further studies have revealed that the activation of the peroxisome proliferator-activated receptor (PPAR) –fatty acid oxidation axis significantly boosts mitophagy by facilitating Parkin recruitment to damaged mitochondria, which in turn robustly elevates the expansion efficiency of HSCs. Notably, this metabolic regulatory cascade is highly conserved in human Tie2⁺HSCs. This conserved feature not only underscores the conserved functional properties of HSCs across species but also lays a theoretical foundation for regulating stem cells fate through metabolic intervention [81].

NK homeobox transcription factor 2–3 (NKX2-3) is a homeobox transcription factor that is preferentially expressed in HSCs with robust regenerative potential. It mediates the targeted clearance of hyperactivated mitochondria and sustains metabolic homeostasis in HSCs by directly regulating the transcription of the key mitophagy regulator ULK1 [82].

Under steady-state conditions, Forkhead box O (FOXO) transcription factors maintain the self-renewal and quiescence of HSCs via cell cycle control, apoptotic suppression, and enhanced oxidative stress resistance, which are crucial for their long-term regenerative potential [83]. FOXO3 specifically governs ROS homeostasis and quiescent state in HSCs, protecting HSCs from oxidative stress damage by regulating the expression of antioxidant genes and the autophagy process. Foxo3 deficiency leads to impaired long-term repopulation capacity of HSCs, mitochondrial dysfunction and autophagy defects, which in turn induces elevated ROS, accumulation of DNA damage and metabolic disorders. Although excess ROS contributes to the defective phenotype of Foxo3^−/− HSCs, their impaired reconstitution ability and metabolic abnormalities are also related to defective mitophagy and accumulation of damaged mitochondria, which further erode the long-term stem cell activity of HSCs [84].

Conditional knockout of the gene encoding ATPase family AAA domain-containing protein 3 A (Atad3a) specifically in mouse hematopoietic cells leads to excessive activation of mitophagy, which in turn reduces mouse survival, diminishes bone marrow cells, induces erythrocytic anemia, lowers B lymphocytes counts, drives skewed differentiation of HSPCs, and expands the HSCs compartment. Mechanistically, the deletion of Atad3a results in the accumulation of mitochondrial autophagy protein Pink1 and subsequent hyperactivation of mitophagy. Concurrent deletion of Pink1 in Atad3a-deficient mice markedly “rescue” excessive mitophagy, restores progenitor cell populations, and normalizes the HSCs pool [85]. GlcNAc transferase (OGT) is a primary enzyme that catalyzes the covalent attachment of n-acetylglucosamine (O-GlcNAc) to the serine or threonine residues of target proteins via O-glycosidic linkages. This post-translational modification, called O-GlcNAc modification, is a dynamic and reversible regulatory mechanism analogous to protein phosphorylation. It governs diverse biological processes including protein function, cell signaling, transcriptional activity, and metabolism. OGT is highly expressed in HSCs and is essential for maintaining the quiescence, mitochondrial quality and stress response of stem cells through modulating PINK1-dependent mitophagy pathway [86].

Autophagy-related protein 7 (Atg7) is an E1-like activating enzyme that promotes the elongation of autophagosome membranes by catalyzing the covalent conjugation of Atg12 to Atg5 and Atg8 family members (such as LC3) with phosphatidylethanolamine (PE). Studies have shown that Atg7 is an indispensable regulator for HSCs homeostasis. Hematopoietic-specific Atg7 ablation in mice impairs HSCs function, triggers severe myeloproliferative lesions, and causes lethality within several weeks. Atg7-deficient HSPCs also exhibit massive accumulation of damaged mitochondria and reactive oxygen species (ROS), alongside elevated proliferative activity and aggravated DNA damage [87]. The Atg12 gene encodes a ubiquitin-like protein and serves as one of the core mediators governing autophagy process. Autophagy constitutes an intracellular degradative pathway that generates double-membrane autophagosomes to sequester cytoplasmic cargo, which is subsequently delivered to lysosomes (or vacuoles in lower organisms) for degradation and material recycling. Loss of ATG12 abolishes basal autophagy in HSCs and drives the emergence of cellular senescence signatures [88].

Sirtuin 3 (Sirt3) gene encodes a core mitochondrial protein critical for sustaining hematopoietic homeostasis. It is highly expressed in HSCs and mainly localized in the mitochondrial matrix, exerting physiological functions as a core deacetylase. Studies have confirmed that Sirt3-knockout mice exhibit defective mitophagy in hematopoietic cells, whereas enforced SIRT3 expression robustly mitigates oxidative stress and attenuates HSCs aging [89–90].

Mitochondrial fusion and fission

Mitochondrial dynamic balance is coordinately governed by fission and fusion events. Dynamin-related protein 1 (Drp1) serves as the core mediator OF mitochondrial fission. Mitofusin 1 (Mfn1) and Mfn2 drive OMM fusion, whereas optic atrophy 1 (OPA1) facilitates IMM fusion. This fission–fusion equilibrium is essential to preserve HSCs stemness [77, 91].

Specifically, mitochondrial fission is mediated by DRP1, which drives the constriction of the OMM to complete mitochondrial fission. Existing studies have confirmed that mitochondria in quiescent HSCs are small, evenly distributed, and metabolically quiescent [92]. In HSCs with low MMP or those subjected to transplantation stress, DRP1 colocalizes far more extensively with mitochondrial fission sites. Hinge et al. (2020) reported that Drp1-deficient HSCs exhibit mitochondrial aggregation, decreased mitochondrial motility motility, and significantly impaired hematopoietic reconstitution ability after transplantation. Collectively, these findings establish that Drp1-mediated mitochondrial fission is essential tor maintain the physiological mitochondrial morphology and preserve the hematopoietic reconstitution potential of HSCs [57, 63].

UPR^mt-mediated mitochondrial quality control for HSCs homeostasis

Aspartyl-tRNA synthetase 2 (Dars2) mainly participates in regulating mitochondrial protein synthesis and the tissue-specific mitochondrial unfolded protein response (UPR^mt). Recent research indicates that Dars2 regulates the maintenance of HSPCs through a non-enzymatic mechanism. Mechanistic studies show that Dars2 deficiency leads to a significant reduction in the expression of splicing factors serine/arginine-rich splicing factor 2, 3, and 6 (Srsf2, Srsf3, and Srsf6), and simultaneously damages multiple mRNA alternative splicing (AS) events. The absence of Dars2 in HSPCs also shows abnormal splicing of mTOR and solute carrier family 22 organic cation transporter member 17 (Slc22a17), and inhibits the levels of unstable ferrous ions and iron-sulfur cluster proteins, which disturbs mitochondrial metabolic activity and DNA damage repair pathways in HSPCs [93].

As a member of the Sirtuin family, the sirtuin 7 (Sirt7) alleviates mitochondrial protein folding stress (PFS^mt) by inhibiting nuclear respiratory factor 1 (NRF1) activity and reducing the expression of mitochondrial translation-related proteins. Loss of Sirt7 reduces HSCs quiescence, increases PFS^mt levels, and impairs stem cell regenerative capacity [94].

Mitochondria and the endoplasmic reticulum jointly regulate HSCs

HSCs are activated and proliferate rapidly during regeneration to replenish the hematopoietic system. Studies have shown that this process is accompanied by substantial ER stress and marked upregulation of EVA 1 homolog A (EVA1A) protein. EVA1A is an ER membrane protein involved in autophagy and apoptosis, and its expression is induced by upregulation of C/EBP homologous protein (CHOP). EVA1A interacts with myeloid cell leukemia 1 (MCL1) to reinforce ER–mitochondria contact sites, which in turn triggers mitochondria-mediated apoptosis [95].

Discussion

As the core hub of cellular energy metabolism, mitochondria play a dominant role in shaping the fate decisions and maintaining the functions of HSCs through multi-dimensional regulatory mechanisms. This study systematically reveals the dynamic association between mitochondrial metabolic reprogramming and HSCs quiescence, and delineates the crucial role of the mitochondrial quality control in microenvironmental adaptation. Notably, HSCs and their differentiated progeny exhibit distinct metabolic signatures. Their reliance on anaerobic glycolysis represents either an adaptive response to the hypoxic bone marrow niche or an intrinsic prerequisite to preserve stemness properties within this low-oxygen milieu. This metabolic preference is closely related to the quiescent state of LT-HSCs. Suppression of mitochondrial OXPHOS efficiently mitigates ROS levels, thereby preserving genomic stability and maintaining long-term repopulation potential. Conversely, ST-HSCs and committed HPCs upregulate mitochondrial respiration to fuel proliferation; this metabolic transition constitutes a critical node for regulating hematopoietic cell fate [96, 97].

In addition, mitochondria exhibit high heterogeneity in multiple dimensions, including metabolic activity, morphology, membrane potential, mtDNA integrity, and dynamics, which is highly coupled with the maintenance of HSCs quiescence, stemness stability, differentiation initiation, and stress response. Mitochondria in quiescent HSCs are characterized by a small number, immature morphology, low metabolic activity, strong membrane potential heterogeneity, and low mtDNA mutation load, which lay an important foundation for maintaining their stemness and long-term reconstitution capacity. In contrast, excessive mitochondrial maturation, respiratory burst, membrane potential convergence, or mtDNA mutation accumulation usually indicates the loss of HSCs stemness, exit from quiescence, and initiation of differentiation. Notably, mitochondrial heterogeneity is not a random phenotypic variation but coordinately controlled by mitochondrial quality control systems including mitochondrial fission/fusion and mitophagy, and its dynamic changes directly affect HSCs function and cell fate. HSCs rely on mitophagy to clear damaged mitochondria, thereby tightly limiting ROS levels and suppressing the activation of programmed cell death. Genetic mutations in key mitophagy-related genes, such as PINK1, PRKN, and PARKIN, significantly impair mitophagy efficiency. These mutations are widely associated with neurological disorders including Parkinson’s disease, as well as tumors and heart failure [98].

Recent sequencing-based analyses of HSCs mtDNA have revealed that HSCs carry fewer mutations than differentiated cells, suggesting that subclonal mtDNA mutations accumulate gradually during cell differentiation. Based on this discovery, researchers have successfully constructed a human hematopoietic atlas combining clonal resolution and single-cell status data. This atlas clearly delineates the developmental origins and lineage relationships of various blood and immune cell populations, providing novel insights into the molecular mechanisms governing hematopoiesis [99].

Elucidating the molecular mechanisms by which mitochondria govern HSCs biology not only deepens our comprehension of intrinsic HSCs properties, but also provide an important theoretical basis for developing novel therapeutic strategies for hematological diseases. Mitochondrial regulators therefore represent promising molecular targets to boost HSCs functionality [100]. For instance, mitochondrial signaling promotes the differentiation of peripheral blood insulin-producing cells (PB-IPC) toward hematopoietic stem cell-like cells (miCD34⁺ HSCs) by regulating the Notch/HEY2 cascade. This finding uncovers an unreported cell reprogramming circuit and identifies a potential therapeutic target for hematological disorders [101]. Functional profiling of HSPCs from human manganese superoxide dismutase (MnSOD)-overexpressing transgenic mice demonstrates significantly enhanced physiological function. Administration of the MnSOD mimetic MnTnBuOE-2-PyP5+ (MnP) further increases the number of HSPCs. These findings suggest that MnP ameliorates pathological bone marrow injury by modulating HSPCs function, offering a feasible approach for expanding stem cells for bone marrow transplantation [102]. Recent advances have yielded major breakthroughs in functional HSCs screening strategies that utilize MMP as a discriminatory biomarker. Studies have revealed that even within the highly purified CD34⁺CD38⁻CD45RA⁻CD90⁺CD49f⁺ HSCs population, MMP displays marked heterogeneity [103].

The discovery of intercellular mitochondrial transfer has opened a new avenue to dissect the sophisticated molecular circuits governing stem cell fate. Accumulating evidence demonstrates that stem cells obtain mitochondria from donor cells via tunneling nanotubes (TNTs), gap junction channels (GJCs), and extracellular vesicles (EVs) [104]. This process is particularly prominent under stress conditions such as inflammation and hypoxia. Supplementation with exogenous mitochondria significantly enhances the proliferation, differentiation, and immunomodulatory properties of stem cells [105]. Preconditioning regimens for bone marrow transplantation damage the stromal microenvironment and impair the mitochondrial function of mesenchymal stromal cells (MSCs). Donor HSPCs deliver functional mitochondria to damaged MSCs through a Cx43‑mediated cell-cell contact mechanism. This process is triggered by elevated intracellular ATP concentrations, which subsequently activates the P2RX7–AMPK signaling cascade. Mitochondrial transfer not only promotes the metabolic recovery and stromal regeneration of MSCs, but also enhances the engraftment efficiency of donor HSPCs. These observations uncover a dual role of heathy HSPCs during bone marrow transplantation: reconstitution of the hematopoietic system and support of stromal tissue repair. Mechanistically, Cx43 deficiency delays the regeneration of mesenchymal and osteogenic tissues, whereas in vivo inhibition of AMPK accelerates stromal recovery and thereby improve hematopoietic reconstitution. This study identifies a promising molecular target to refine clinical bone marrow transplantation regimens and underscores that targeted regulation of mitochondrial transfer and stromal repair can substantially improve clinical transplantation outcomes [106]. A recent study using the Ndufs4⁻/⁻ mouse model of Leigh syndrome (LS) demonstrated that transplantation of wild-type bone marrow significantly alleviates neurological symptoms in model mice. The therapeutic effect stems from secretion of extracellular mitochondria and subsequent targeted mitochondrial delivery across multiple organs. Furthermore, injection of mitochondria derived from wild-type mice effectively prolongs survival, ameliorates neurological performance, and restores cellular energy metabolism in LS model mice. Notably, cross-species delivery of human mitochondria also ameliorates the LS phenotype. This study confirms the therapeutic potential of mitochondrial transfer-based interventions for mitochondrial diseases, providing a novel intervention strategy for hereditary metabolic disorders such as LS [107, 108].

In conclusion, mitochondria occupy a central position in the metabolic regulation, functional maintenance, and fate determination of HSCs. Dynamic homeostasis of mitochondrial heterogeneity is a key factor in sustaining HSCs quiescence and stemness. Mitochondrial dysfunction impairs HSCs function, thereby increasing the risk of hematological disorders. Future studies should further elucidate the molecular crosstalk between mitochondria and HSCs activity, which will facilitate the discovery of druggable therapeutic targets. Notably, mitochondrial transplantation represents a highly promising therapeutic strategy capable of rescuing defective mitochondrial function within HSCs, opening up an innovative translational route for the clinical management of mitochondrial and hematological diseases.

Abbreviations

BM

Bone marrow

HSC

Hematopoietic stem cell

HSPC

Hematopoietic stem and progenitor cell

LT-HSC

Long-term hematopoietic stem cell

ST-HSC

Short-term hematopoietic stem cell

MPP

Multipotent progenitor cells

CLP

Common lymphoid progenitor

CMP

Common myeloid progenitor

MEP

Megakaryocyte/erythrocyte progenitor

GMP

Granulocyte/macrophage progenitor

NK

Natural killer cell

DC

Dendritic cell

OMM

Outer mitochondrial membrane

IMS

Intermembrane space

IMM

Inner mitochondrial membrane

MMP

Mitochondrial membrane potential (Δψm)

ROS

Reactive oxygen species

TCA

Tricarboxylic acid cycle

FAO

Fatty acid oxidation

OXPHOS

Oxidative phosphorylation

Author contributions

Ran Xu wrote the main manuscript text, Yaowei Meng and Anyu Xu prepared figures, Changyan Li , Xiaoming Yang and Weiming Tian, responsible for reviewing and proofreading manuscript，All authors reviewed the manuscript.

Funding

This study was supported by grants from the Ministry of Science and Technology of the People’s Republic of China (Grant No. 2022YFA1103502).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.