The ubiquitous second messenger PKA, one of the most well-investigated cytosolic kinases, is located on the mitochondrial surface and plays a regulatory role in maintaining mitochondrial activity, including mitochondrial dynamics (Feliciello et al., 2005; Cribbs and Strack, 2007). Drp1 relies on a GTP hydrolysis-dependent mechanism to form ring superstructures to contract and eventually incise mitochondria. Drp1 Serine 656 (Ser656) of Drp1 is a major PKA phosphorylation site in rat PC12 cells, this phosphorylation inhibits mitochondrial fission and reduces cellular sensitivity to apoptotic stimuli. Although Ser656 phosphorylation site is near the GTPase effector domain, there is no significant change in GTP hydrolysis among the Ser656 variants (Cribbs and Strack, 2007). However, in HeLa cells, active PKA suppresses Drp1 GTPase activity via Ser637 phosphorylation of Drp1, which may favor GTPase inactivation and Drp1 localization in the cytoplasm rather than recruitment to mitochondria, dampening mitochondrial fission (Chang and Blackstone, 2007; Yu et al., 2019). In contrast, phosphorylation of Drp1 at Ser600 is also dependent on adrenergic-stimulated PKA activation, inducing increased mitochondrial fragmentation and energy expenditure in brown adipocytes (Wikstrom et al., 2014). These findings suggest that cAMP/PKA controls mitochondrial fission principally via Drp1 modulation, and its effect on fission depends on phosphorylation sites and Drp1 transport. In fact, post-translational phosphorylation of Drp1 predominantly affects Drp1 activity. The different phosphorylated sites of Drp1 and they mediated-mitochondrial fission are demonstrated in Figure 1.

Some evidence has confirmed the role of the cAMP/PKA pathway in influencing mitochondrial fusion-related proteins. In myoblasts stimulated by reactive oxygen species (ROS), decreased mitochondrial cAMP/PKA signaling induces Sirt3 degradation/proteolysis, which in turn promotes the hyperacetylation of Opa1 and short Opa1 generation, consequently leading to hyperfragmentation of mitochondria and cell apoptosis (Signorile et al., 2017). Treatment with 8-Br-cAMP, an analog of cAMP, reverses the detrimental effect of Opa1 on mitochondrial dynamics, as well as cytochrome c release under oxidative stress in cardiac myoblast cells (Signorile et al., 2017). Although Zhou et al. (2010) verified that the specific PKA phosphorylation site Ser442 of Mfn2 effectively promotes Mfn2-mediated inhibition of vascular smooth muscle cell growth, these effects are thought to be independent of mitochondrial morphology and dynamics changes.

Mitogen-activated protein kinases are serine-threonine kinases that transmit signals driven by cytokines, hormones, and other factors from the cell surface to the nucleus. At least three MAPK families have been characterized, c-Jun N-terminal kinase (JNK), extracellular regulated kinase (ERK1/2), and p38, which are involved in a wide variety of cellular functions, such as proliferation, apoptosis, and differentiation (Widmann et al., 1999). The most well-characterized MAPK with a confirmed regulatory effect on mitochondrial dynamics is ERK1/2 (Serasinghe et al., 2015; Cook et al., 2017). Oncogenic Ras treatment can activate ERK2 to phosphorylate Drp1 at Ser616, resulting in an increase in mitochondrial fragmentation. Both activation of MEK activity and elevated ERK phosphorylation lead to consistent trends in the expression of Drp1 (Kashatus et al., 2015). ERK suppresses mitochondrial fusion by phosphorylating Mfn1 at threonine 562 (Thr562), which favors Mfn1 binding to BAK and promotes their oligomerization and activation, subsequently facilitating cytochrome c release and apoptosis. This indicates the significant role of ERK-targeted Mfn1 in modulating mitochondrial morphology and apoptosis for mouse embryonic fibroblasts (Pyakurel et al., 2015). ERK1/2 activation participates in mitochondrial fission mediation during the early stage of induced pluripotent stem cell (iPSC) reprogramming by phosphorylating Drp1 at Ser579. Additionally, MEK inhibitor treatment inhibits mitochondrial fragmentation, while mutations in Ser579 of Drp1 successfully rescue the fission inhibition caused by MEK inhibitor application (Prieto et al., 2016). Similarly, treatment with the MEK inhibitor, PD325901, results in dramatic decline in both ERK and Drp1616 phosphorylation, effectively reversing mitochondrial fragmented phenotype in oncogenic Ras-induced embryonic kidney cells (Kashatus et al., 2015).

JNK activation is connected to Mfn2 but not Mfn1 turnover (Chakraborty et al., 2018). Leboucher et al. demonstrated that JNK mediated the phosphorylation of Mfn2 at Ser27 of sarcoma U2OS cells in response to cellular stress, which contributed to ubiquitin-proteasome degradation in Mfn2 and enhanced apoptosis (Leboucher et al., 2012). Ko et al. (2017) showed that p38 MAPK promoted Drp1-dependent fission in MSCs, and treatment with SB203580, a p38 inhibitor, reversed Drp1 phosphorylation at Ser616, leading to a decline in fragmented mitochondria.

AMPK is a conserved, redox-activated cellular energy sensor and regulator that is sensitive to AMP/adenosine-5′-triphosphate (ATP) stimulation. This kinase is activated during ATP consumption in response to stresses, such as low glucose, hypoxia, and ischemia (Carling, 2004). Reduced ATP levels induced by, for example, iron overload (IO), can activate AMPK in bone marrow mesenchymal stem cells (BMSCs), followed by Mff phosphorylation, further resulting in Drp1 translocation to mitochondria and fission enhancement, which partly contributes to BMSC dysfunction derived from IO patients with myelodysplastic syndrome (Zheng et al., 2018). The expression of p-Drp1 was downregulated while Mfn2 was upregulated in BMSCs after they were treated with compound C, an AMPK inhibitor, and cell senescence was therefore increased. These results indicated that mitochondrial fission activated by AMPK protected BMSCs from senescence (Li et al., 2019). At low ATP levels, AMPK-mediated fission is promoted due to Drp1 transport to the mitochondrial membrane stimulated by phosphorylation of Mff (Ducommun et al., 2015; Toyama et al., 2016). Consistently, stimulation of U2OS osteosarcoma cells with either electron transport chain inhibitor complexes or AMPK agonists resulted in significantly increased mitochondrial fragmentation (Toyama et al., 2016). AMPK-mediated mitochondrial fission is widely involved in cellular reaction to energy stress, and these mitochondrial fission events may serve as a trigger to initiating mitophagy to remove damaged mitochondria (Youle and van der Bliek, 2012; Toyama et al., 2016).

In contrast, activation of AMPK by a specific AMPK activator reversed the reduction in drug-induced mitochondrial fusion via upregulation of Mfn1, Mfn2, and Opa1, thereby sustaining hepatocyte viability (Kang et al., 2016). AMPK activation by AICAR, a known activator of AMPK, induced fused mitochondria and interconnected networks when adipose-derived mesenchymal stem cells (ASCs) from pericardial adipose tissue were differentiated into adipocytes (Abdul-Rahman et al., 2016). In addition, AMPK activated by metformin is able to suppress Drp1-dependent mitochondrial fission and alleviate oxidative stress, thus ameliorating atherosclerosis in diabetic mice (Wang et al., 2017b). However, mitochondrial fusion activated by AMPK is not always related to positive consequences. TGF-1 induces mitochondrial fusion through the AMPK signaling pathway, which induces aging of vascular progenitor cells isolated from patients with Marfan syndrome (He et al., 2019).

These findings suggest that AMPK activation can have opposing effect on mitochondrial dynamics, which possibly depend on cell types, bioenergetic status, and oxidative stress levels. The positive situation is that active AMPK sustains well-quality mitochondria at least partially via triggering mitochondrial fission and subsequent autophagy. When damage or stress is alleviated or removed, AMPK may support mitochondrial function by strengthening fusion rather than fission. Although AMPK-mediated mitochondrial dynamics have been found to be involved in some cellular damage processes, their relationship and detailed mechanisms have not been well studied.

The Sirt family is composed of highly conserved nicotinamide adenine dinucleotide (NAD⁺)-dependent deacetylases (HDACs). There are seven mammalian sirtuin subtypes (SIRT1 to SIRT7) that have different subcellular localizations and regulate a variety of cellular functions through posttranslational modifications of target proteins (Huang et al., 2010). Among them, SIRT3, SIRT4, and SIRT5 are located on mitochondria and can mediate the activity of some proteins related to mitochondrial dynamics (Huang et al., 2010; Samant et al., 2014; Lang et al., 2017).

SIRT3-mediated deacetylation of Opa1 at lysine 926 and 931 enhances Opa1 GTPase activity and sustains mitochondrial morphology, protecting cardiomyocytes from doxorubicin-induced cell death (Samant et al., 2014). Accurately, OPA1 appears to work in two different isoforms. Long membrane-bound form of OPA1 is responsible for mitochondrial fusion. But cleavage of long OPA1 inhibits mitochondrial fusion accompanied with generation of fission or mitophagy-associated short, soluble forms (MacVicar and Langer, 2016). SIRT4 expression facilitates mitochondrial fusion with an incremental increase in mitochondrial mass by inducing long Opa1 instead of short Opa1, further contributing to decreased mitophagy in HEK293 cell lines (Lang et al., 2017). In addition to promoting fusion, Fu et al. (2017) showed that SIRT4 may modulate mitochondrial fission in lung cancer cell lines by reducing Drp1 phosphorylation and diminishing Drp1 recruitment to the mitochondrial membrane by modulating MEK/ERK signal and interacting with Fis1. SIRT5-overexpressing C₂C₁₂ cells contain large and lengthened mitochondria that are uniformly arranged in the cytoplasm partly due to SIRT5-mediated upregulation of Mfn2 and Opa1, whereas SIRT5-silenced cells display small, round mitochondria that are distributed around the nucleus (Polletta et al., 2015). Consistent with these findings, SIRT5-mediated fusion promotion in mouse embryonic fibroblasts was also reported by Guedouari, and the results showed that elongated mitochondria and depressed autophagy depended on SIRT5 under starvation conditions (Guedouari et al., 2017). SIRT5 deletion increased the expression of mitochondrial dynamic protein of 51 kDa, Fis1, and pDRP1-S637, subsequently causing Drp1 activation and its translocation to mitochondria, thus promoting mitochondrial fission (Guedouari et al., 2017).

The available evidence suggests that Sirt family members located on mitochondria generally promote mitochondrial fusion and/or inhibit mitochondrial fission, and this effect is largely accompanied by a reduction in mitophagy or autophagy. It is hypothesized that fused mitochondria modulated by Sirt subtypes on mitochondria help alleviate stress injury and avoid damage caused by mitophagy.

As previously mentioned, fragmented mitochondria meet the energy needs of immature MSCs, exerting a fundamental role on intrinsic function (Seo et al., 2018; Zhang et al., 2018). Mitochondrial fission is essential for maintaining stemness in MSCs, and inhibiting fission leads to a reduction in the expression of stemness markers and multidirectional differentiation potential (Feng et al., 2019). During somatic cell reprogramming to iPSCs, cells undergo mitochondrial reconstruction, usually from a mature network toward immature mitochondria, and a metabolic shift, usually from OXPHOS toward glycolysis. Drp1-dependent mitochondrial fission is also necessary for the acquisition of cellular pluripotency during the early stage of embryonic fibroblast induction to form iPSCs (Prieto et al., 2016). Moreover, depletion of fusion-related Mfn1 and Mfn2 activates hypoxia-inducible factor 1α signaling, a necessary mediator of the metabolic switch to glycolysis, thus facilitating the transition in metabolic pattern from OXPHOS to glycolysis (Son et al., 2015). These processes promote the conversion of somatic cells to a pluripotent state (Son et al., 2015). However, over fission induced by Mff overexpression impairs the pluripotency of ESCs and iPSCs, suggesting that active fission should be within a certain threshold to better maintain stem cell function (Zhong et al., 2019). In summary, MSCs prefer mitochondrial fission to mitochondrial fusion to sustain the cell state and function, and active to-fission promotes the self-renewal and pluripotency of stem cells by stimulating glycolysis.

In contrast to most stem cells, which have immature mitochondria, stem cells undergoing differentiation usually develop mature or specialized mitochondrial networks. MSCs undergoing differentiation display elongated and interconnected mitochondria accompanied by significantly decreased Drp1 and markedly increased OPA1 expression, indicating that mitochondrial fusion is conducive to MSC differentiation into adipocytes and osteocytes (Feng et al., 2019; Fujiwara et al., 2019). Similarly, Forni et al. induced osteogenesis and adipogenesis in MSCs and demonstrated enhanced mitochondrial biogenesis and network restructuring via mitochondrial fusion mediated by Mfn1 and Mfn2 during the early stages of induction (Forni et al., 2016). Furthermore, Mfn2 knockdown restrains respiratory activity, including basal ATP, maximal respiratory capacity, and H⁺-leak-related respiratory activity, leading to a loss of differentiation ability (Forni et al., 2016). The fused mitochondrial network facilitates mitochondrial energy production through OXPHOS, which corresponds to the energy required in these differentiated cells (Hsu et al., 2016; Li et al., 2017b). Consistently, during the differentiation of human iPSCs into cardiomyocytes, blocking Drp1 induces a metabolic transition from glycolysis toward OXPHOS, resulting in increased differentiation capacity in these cells (Hoque et al., 2018). Neuronal differentiation of iPSCs is closely related to upregulation of Mfn2 expression (Fang et al., 2016). Depletion of Mfn2 induces mitochondrial respiratory dysfunction by inhibiting complexes I and IV enzymatic activity and reducing ATP levels (Fang et al., 2016).

Interestingly, chondrogenesis in MSCs involves a fragmented mitochondrial phenotype at the beginning stage with increased expression of Drp1, Fis1, and Fis2 (Forni et al., 2016), which seems to contradict the finding that mitochondrial fusion is promoted during osteogenic and adipogenic differentiation of MSCs. Cells undergoing chondrogenesis do have low levels of basal respiration at the early commitment phase (Forni et al., 2016). Metabolomic analysis has revealed that catabolic processes occurring during the early stage of chondrogenesis are performed by both glycolysis and mitochondrial respiration (Kwon and Ohmiya, 2013). These metabolic changes differ from the enhanced mitochondrial OXPHOS that occurs during early adipogenic or osteogenic commitment (Chen et al., 2008; Li et al., 2017b), which may partly explain the increased fission to support glycolysis in MSCs during the early stage of chondrogenesis.

Mitochondrial dynamics clearly alter the differentiation fate of MSCs by regulating energy metabolism. However, mitochondrial dynamics-mediated regulation of MSC differentiation is not limited to energy metabolism regulation. Mitochondrial dynamics-dependent mitophagy may also be involved in modulating stem cell differentiation. Evidence has shown that mitochondria become fragmented before mitophagy enhancement (Forni et al., 2016; Marycz et al., 2016), demonstrating that mitochondrial fission is essential for mitophagy or autophagy (Gomes and Scorrano, 2008; Frank et al., 2012), which plays a critical role in MSC differentiation (Marycz et al., 2016; Vidoni et al., 2019). Additionally, although the relationship between mitochondrial dynamics and calcium homeostasis in MSCs has not been revealed, it has been suggested that mitochondrial dynamics can influence the differentiation of stem cells by regulating intracellular calcium balance. Zhong et al. (2019) reported that excess mitochondrial fission exacerbated cytosolic Ca²⁺ entry and CaMKII activity, resulting in the degradation of β-catenin and ultimately impairing the differentiation and embryonic development of iPSCs. Mfn2 negatively regulates calcineurin/NFAT activity by intracellular Ca²⁺ buffering, thereby maintaining HSCs with extensive lymphoid potential (Luchsinger et al., 2016). Kasahara et al. (2013) demonstrated that gene trapping of Mfn2 or OPA1 was sufficient to inhibit the differentiation of ESCs into cardiomyocytes. Mechanically, mitochondrial fusion orchestrates Ca²⁺ and calcineurin A to further affect Notch1-mediated suppression of the cardiomyocyte transition (Kasahara et al., 2013).

In short, although mitochondrial fission and mitochondrial fusion may exert pleiotropic effects during MSC differentiation depending on specific lineage commitment and different stages, we hypothesize that such mitochondrial dynamics undergo specific transformations to ensure the ever-changing energy demands and calcium balance, which is fundamental for the multidirectional differentiation of MSCs.

Long-term in vitro amplification is a common way that can induce senescence of MSCs (Kasper et al., 2009). After a certain number of cell divisions (7–12 passages), senescent cells increase, which is characterized by morphological abnormalities, enlargement, and increase of senescence-associated β-galactosidase positive cells. The long-term MSCs cultures (more than 100 passages) derived from rat have been found to exhibit increased susceptibility to senescence and have non-tumorigenic (Wagner et al., 2008; Geissler et al., 2012). Karyotype analysis in BMSCs reveals that aneuploidy chromosomal alterations may occurs during population doublings, but they became senescent without transformation features (Tarte et al., 2010).

Lengthened mitochondria often occur in various aging cells (Mai et al., 2010; Lin et al., 2015). Aged MSCs also exhibit a strong and complicated interconnected network that is distributed evenly in the cytoplasm, suggesting a potentiation of fusion processes (Geissler et al., 2012). p-Drp1 expression has been reported to be greatly downregulated, whereas Mfn2 expression is markedly upregulated in passage 12 (P12) BMSCs compared with those in P4 BMSCs, suggesting that these cells undergo aging accompanied by mitochondrial fusion (Li et al., 2019). Consistent with these observations, P7 ASCs have large tubular mitochondria forming an intertwined network that is regulated by Mfn1, Opa1, and Fis1 (Stab et al., 2016). In contrast, P2 ASCs show small tubular mitochondria forming a slightly interconnected network (Stab et al., 2016). Excessive mitochondrial fusion may adversely affect cells by altering ROS levels. Prolonged or giant mitochondria have been reported to augment ROS generation and weaken mitochondrial respiration activity in deferoxamine-induced senescent cells (Yoon et al., 2006). Furthermore, blocking mitochondrial fission, by overexpression of Drp1-K38A (active site is mutated in Drp1) and Fis1-ΔTM (transmembrane domain is deleted in Fis1), successfully leads to a senescent phenotype with ROS elevation in normal cells (Yoon et al., 2006). Additionally, the reduction in Drp1 levels during vascular aging exacerbates endothelial cell dysfunction by increasing mitochondrial ROS and suppressing autophagic flux, while the antioxidant N-acetyl-cysteine restores autophagosome clearance and improves angiogenesis in senescent endothelial cells (Lin et al., 2015).

Notably, increased mitochondrial fusion during aging is not always harmful (Stab et al., 2016). During fusion, depolarized mitochondria and normal mitochondria can join together to exchange their contents, further aiding in damaged mitochondrial repair and membrane potential maintenance (Twig et al., 2008b; Meyer et al., 2017). Of course, excessive fusion can be detrimental when depolarized and damaged mitochondria are overloaded, which is conducive to mitochondrial dysfunction during aging (Ikeda et al., 2015; Lang et al., 2017).

Conversely, cell apoptosis is usually accompanied by abnormal rupture of mitochondria in MSCs. Apoptotic ASCs isolated from equine metabolic syndrome (EMS) horses contain notably increased fragmented mitochondria (Kornicka et al., 2019). Analogously, prominently increased expression of Fis1 and decreased Mfn1 and Mfn2 expression have been reported in apoptotic human umbilical cord MSCs induced by monocrotophos exposure (Srivastava et al., 2018). Ma et al. (2019) demonstrated that dexamethasone increased mitochondrial fission and meanwhile diminished mitochondrial fusion. Mitochondrial fission was promoted by increased Fis1 and Mff expression, whereas the mitochondrial fusion was inhibited by decreased Mfn1 and Mfn2 expression. These mitochondrial dynamics alterations contributes to apoptosis enhancement and osteogenic suppression of BMSCs (Ma et al., 2019). Correspondingly, treatment with mdivi-1, an inhibitor of Drp1, dramatically ameliorated hydrogen peroxide (H₂O₂)-induced cellular apoptosis and death in human periodontal ligament stem cells (PDLSCs) and human W₈B₂₊ cardiac cells (Rosdah et al., 2017; He et al., 2018). Moreover, glucose/serum-deprived/hypoxia treatment-induced apoptosis in MSCs is reduced by increased glycolytic efficacy, which is modulated by the leptin/OPA1/SGLT1 signaling pathway (Yang et al., 2019).

The role of mitochondrial fission in modulating stem cell apoptosis is still unclear. Mitochondrial outer membrane permeabilization (MOMP) mediates the cascade conduction of many apoptotic signals. Arnoult et al. (2005) observed that Bax/Bak could facilitate the release of deafness dystonia protein 1 homolog, a member of mitochondrial intermembrane chaperone, into the cytoplasm, promoting its binding to the C-terminus of Drp1, this further results in Drp1 recruitment to mitochondria and Drp1-dependent mitochondrial fission. This Drp1-mediated mitochondrial fragmentation is vital for mitosis, which is involved in caspase-independent cell death (Arnoult et al., 2005). Earlier findings suggested that overexpression of Mfn1 reduces apoptotic HeLa cells induced by etoposide by impeding Bax transport to mitochondria and cytochrome c release (Sugioka et al., 2004). Intriguingly, other reports have shown that Drp1-mediated mitochondrial fission prevents cell apoptosis, especially Ca²⁺-related death (Szabadkai et al., 2004; Jahani-Asl and Slack, 2007). Specific upregulation of Drp1 leads to division of the mitochondrial network and damages the connectivity of the mitochondrial lumen, separating mitochondria from the endoplasmic reticulum, the source of calcium, which decreases Ca²⁺ absorption and ultimately abrogates Ca²⁺ overload-induced apoptosis (Szabadkai et al., 2004). However, these mitochondrial fission-dependent antiapoptotic events have not been demonstrated in stem cells.

Although current studies have shown that the apoptotic process of stem cells is dominated by mitochondrial fission, different apoptosis-inducing factors should be further studied to comprehensively understand the effect of mitochondrial dynamics on apoptosis. Overall, the interaction between mitochondrial dynamics and apoptotic signals is complicated. In response to different apoptotic or death pathway stimuli, mitochondrial dynamics may either support or restrain apoptosis. Increased apoptosis is usually attributed to enhanced mitochondrial fission since these processes act on Bcl-2 family dependent apoptotic pathways and related molecules. The effectiveness and mechanisms involved in mitochondrial dynamic-mediated alleviation of cell death caused by Ca²⁺ overload requires further confirmation.

Oxidative stress occurs when the balance of the oxidative stress and antioxidant systems breaks down, which subsequently causes cellular damage. During physiological or repair processes, stem cells inevitably suffer attacks caused by oxidative stress. In fact, oxidative stress is a well-explored mechanism in regulating stem cell fate (Ko et al., 2012; Tan and Suda, 2018). Overgeneration of ROS indicates the occurrence of oxidative stress in tissues (Tan and Suda, 2018; Vina et al., 2020). Under pathological conditions, excessive ROS can be produced by mitochondria, which in turn leads to the inactivation of mitochondrial components (Zorov et al., 2014). The damaged members involved in mitochondrial dynamics may destroy the mitochondrial morphology and structure, resulting in a vicious cycle of continuous ROS release (Ježek et al., 2018). This reflects the complex interaction between mitochondrial dynamics and ROS generation.

Oxidative stress with ROS evaluation induced directly by H₂O₂ treatment leads to mitochondrial fragmentation in hMSCs; moreover, the combination of N-acetylcysteine, a biologic antioxidant, and ascorbic acid 2-phosphate, an oxidation-resistant derivative of ascorbic acid, successfully inhibits mitochondrial fission, decreases ROS production, and stabilizes mitochondrial membrane potential (Li et al., 2015). In another similar study of oxidative stress conditions stimulated by serum deprivation and hypoxia treatment, BMSCs had increased mitochondrial fragmentation along with upregulation of p-Drp1 Ser616 expression and downregulation of Mfn2 expression (Deng et al., 2020). In addition, an in vitro study underscored that CoCl₂, a hypoxia mimetic, promoted mitochondrial fission in PDLSCs mediated by Drp1 elevation (He et al., 2018). Targeted inhibition of Drp1 markedly increased ATP levels, suppressed ROS generation, and eventually reduced cell apoptosis, indicating the important role of the ROS-Drp1-dependent mitochondrial pathway in CoCl₂-induced apoptosis in PDLSCs (He et al., 2018). These findings suggest that high ROS levels and oxidative stress generally lead to abnormal mitochondrial dynamics, especially excessive mitochondrial fission. Reducing ROS levels helps to restore normal mitochondrial dynamics. Moreover, the regulation of mitochondrial dynamics can also be beneficial for reversing ROS overgeneration.

Unlike the high level of ROS, which is always associated with cell damage and disease, low or normal ROS level has been shown to have a positive effect on cell homeostasis and function via participating in signal transduction and promoting mitophagy (Shadel and Horvath, 2015; Palmeira et al., 2019). Early outbreaks of transient oxidative phosphorylation and elevated ROS in somatic cells promote NRF2 transcription factor activity, which further initiates the hypoxia inducible factor α-mediated glycolytic shift in early reprogramming (Hawkins et al., 2016). During reprogramming toward iPSCs, mitochondria undergo reconstruction dominated by enhanced mitochondrial fission, gradually forming an immature state instead of a mature mitochondrial network (Vazquez-Martin et al., 2012; Prieto et al., 2016; Lisowski et al., 2018). Compared with somatic cells, stem cells including MSCs have low ROS levels and immature mitochondrial networks (Hsu et al., 2016; Lisowski et al., 2018). Therefore, it is speculated that the changes in mitochondrial dynamics associated with low ROS levels are conducive to mitochondrial remodeling and adaptive changes.

Studies on the effects of metabolic stress on mitochondrial dynamics mainly involve abnormalities in glucose and lipid metabolism. High levels of fatty acids alone or in combination with high glucose induce an increase in mitochondrial fragmentation (Molina et al., 2009). Dysfunctional ASCs isolated from patients with type 2 diabetes exhibit a similar trend in mitochondrial phenotype, in which overexpression of Fis1 causes fragmented, round mitochondria, and mitochondrial autophagy is also impaired, as indicated by reduced parkin RBR E3 ubiquitin protein ligase (PRKN, better known as Parkin) expression (Alicka et al., 2019). The Parkin is a crucial mitophagy regulator and its activation can build ubiquitin chains to ubiquitinating outer membrane protein on damaged mitochondria to label them for degradation in lysosomes (Bingol and Sheng, 2016). These findings are consistent with previous research by the same team, which showed that increased ASC apoptosis derived from EMS horses resulted in enhanced mitochondrial fission compared with the healthy control (Marycz et al., 2018). Similarly, metabolic syndrome impairs the swine ASC mitochondrial structure featured by an increase in mitochondrial fission and reduction of mitochondrial fusion (Farahani et al., 2020). Yu et al. (2006) have found that hyperglycemia-induced mitochondrial fission promotes ROS production and its periodic fluctuations, and either fission inhibition by Drp1 inhibition or fusion induction by Mfn2 overexpression prevents an increase in ROS induced by hyperglycemia.

Limited studies have indicated the potential impacts of physical stimuli or exogenous toxins on the mitochondrial dynamics of MSCs. Patten et al. (2019) observed that mitochondrial length in human BMSCs was slightly increased 4 h after exposure to 2 Gy radiation. They further demonstrated that Opa1 knockdown in mouse embryonic fibroblasts induced a decline in their adaptation to radiation, suggesting that mitochondrial networks may also be involved in the regulation of BMSC adaptation to radiation (Patten et al., 2019). Yin et al. (2017) revealed that low-level laser exposure facilitated mitochondrial biogenesis via upregulation of molecules associated with both mitochondrial fusion (Mfn1, Mfn2, and Opa-1) and mitochondrial fission (Fis1, Drp1, and MTP18), which contribute to elevated proliferation of BMSCs. In an in vitro study, methamphetamine exposure dampened the osteogenic differentiation of BMSCs due to abnormal OXPHOS and reduced ATP generation and mitochondrial membrane depolarization. These mitochondrial malfunctions were attributed to damaged mitochondrial biogenesis and mitochondrial fusion (Shen et al., 2018a). Similarly, treatment with carbon black Printex 90, a representative carbonaceous particle toxicant, induced mitochondrial dysfunction in BMSCs, which is closely related to suppressed mitochondrial biogenesis and mitochondrial dynamics, ultimately resulting in impaired osteogenic potential of BMSCs (Shen et al., 2018b).

According to the literature, stress often results in fragmentation but not elongation events in mitochondria in MSCs, which is attributed to dramatic enhanced mitochondrial fission with/without inhibited mitochondrial fusion (He et al., 2018; Marycz et al., 2018; Shen et al., 2018a; Ma et al., 2019). In addition, the suppression effect on both mitochondrial fission and fusion was observed in BMSCs when they were treated with high levels of ferric ammonium citrate, a commonly used agent to induce iron overload (Yao et al., 2019). In contrast, Marycz et al. (2019) suggested that ASCs derived from metabolic syndrome horses displayed abnormal dysregulated and mixed mitochondria due to hyperactive mitochondrial fission and fusion. Moreover, stress-induced mitochondrial hyperfusion has been proposed by Tondera et al. (2009) as a particular prosurvival response to stress, and their research confirmed that exposure to a low dose of UV irradiation or actinomycin D efficiently induces hyperfusion of mitochondria in mouse embryonic fibroblasts within 9 h. Similarly, mouse embryonic fibroblasts show elongated mitochondria after starvation treatment, which protects them from autophagy-induced mitochondrial clearance (Gomes et al., 2011). Jendrach et al. (2008) observed that short-term and low doses of H₂O₂ result in a transitory increase in mitochondrial fusion in human umbilical vein endothelial cells. This enhanced mitochondrial fusion is beneficial to damaged mitochondria, in which mitochondrial DNA and membrane potential are sustained and the energy supply capacity is maximized to enhance adaptation under stress (Tondera et al., 2009; Gomes et al., 2011; Meyer et al., 2017).

Therefore, it is reasonable to conclude that different stressors and stress intensities can lead to complex changes in mitochondrial dynamics in MSCs, which are involved in fission and/or fusion activity. Mitochondrial dynamics seem to be central to bridging the gap between external stress and MSC function (Figure 3). Fusion events may act as a compensatory reaction in response to weak stress to better promote survival. In many cases, mitochondrial elongation is beneficial for resisting stress (Agarwal et al., 2016; Yang et al., 2019). However, upregulated fusion does not always mean elevated function. For example, senescent cells usually have a highly fused mitochondrial network, but their functions are relatively poor (Lin et al., 2015; D’Amico et al., 2019). Likewise, mitochondrial fragmentation is not always the result of maladaptation. Under stress states, impaired mitochondria may cause damage by ROS or other overgenerated harmful by-products and abnormal accumulation of Ca²⁺ (Shaughnessy et al., 2014; Park et al., 2017; Dilberger et al., 2019). Accordingly, timely and proper clearance of damaged mitochondria is imperative to cell survival and function. Mitophagy is the main way to selectively eliminate damaged or old mitochondria (Twig et al., 2008a; Ashrafi and Schwarz, 2013). The suppression of mitophagy impairs cell adaptability to stress and has been implicated in diabetes and neurodegenerative and cardiovascular aging diseases (Wallace, 2005; Egan et al., 2011; Marycz et al., 2016; Baechler et al., 2019). Mitochondria rely on mitochondrial fission to produce different daughter units. The daughter units with healthier membrane potentials will continue to participate in the dynamic cycle and facilitate recovery, whereas depolarizing mitochondria can be removed by mitophagy (Twig et al., 2008a). Inhibition of mitochondrial fission impairs mitophagy, while transitory enhanced mitochondrial fission acts as a protective strategy via cooperation with mitophagy (Twig et al., 2008a; Youle and van der Bliek, 2012). Stress most often induces hyperfission or abnormal mitochondrial dynamics, thus stimulating or worsening cell damage or death. Taken together, these findings suggest that the coordination of mitochondrial fission and mitochondrial fusion is a pivotal cytoprotective mechanism for cellular renovation and homeostasis under stress.