Intercellular mitochondria trafficking highlighting the dual role of mesenchymal stem cells as both sensors and rescuers of tissue injury

ABSTRACT

Mitochondria are crucial organelles that not only regulate the energy metabolism, but also the survival and fate of eukaryotic cells. Mitochondria were recently discovered to be able to translocate from one cell to the other. This phenomenon was observed in vitro and in vivo, both in physiological and pathophysiological conditions including tissue injury and cancer. Mitochondria trafficking was found to exert prominent biological functions. In particular, several studies pointed out that this process governs some of the therapeutic effects of mesenchymal stem cells (MSCs). In this review, we give an overview of the current knowledge on MSC-dependent intercellular mitochondria trafficking and further discuss the recent findings on the intercellular mitochondria transfer between differentiated and mesenchymal stem cells, their biological significance and the mechanisms underlying this process.

Introduction

Mitochondria likely represent the most complex organelles found in the cytosol of eukaryotic cells, with regard to their structural organization and the diversity of their functions. One of their peculiarities is to originate from a bacterial ancestor, by an endosymbiosis that occurred more than 1.5 billion years ago [1]. From these ancient bacteria, mitochondria retained several structural features, notably their inner membrane. Indeed, mitochondria contain both an inner and an outer membrane. Whereas the mitochondria outer membrane, that separates the inside of the organelle from the rest of the cell, is a phospholipid bilayer membrane similar to that of the eukaryotic cell own membrane, the mitochondria inner membrane shares lipid components like cardiolipin with bacteria membranes, reminiscent of its prokaryotic origin. In addition, they contain a circular genome harboring a genetic code different from that of nuclear DNA [2]. As a result, they can replicate their DNA and divide, within the cells, independently of cell division. In addition, mitochondria exert a broad range of important functions in the cells. They are commonly considered as the powerhouse of the cells, devoted to convert nutrients into energy to fuel the cellular biological activities, as they produce most of the cell demands in adenosine triphosphate (ATP) through oxidative phosphorylation. Mitochondria are involved in numerous anabolic and catabolic processes. Overall, they regulate a broad range of cellular functions, including specific metabolic pathways activation, tissue temperature maintenance, calcium signaling and cell death induction. Therefore, they contribute to cell adaptation to physiological and pathological environmental changes. Understanding the complexity of mitochondria functions became a challenging area of research over the past few years, with implications in the fields of regenerative medicine, oncology and immunology. This stems from the discovery that the mitochondria-dependent metabolic reprogramming controls a wealth of different functions: the self-renewal and differentiation capacities of mesenchymal (MSCs) and embryonic (ES) stem cells [3–5], the secretion of inflammatory cytokines by immune cells such as macrophages [6–8] and dendritic cells [8–10] and also the malignant properties of cancer cells [11,12].

Mitochondria recently attracted a renewed attention from the scientific community, as either the whole organelles, the mitochondrial genome or other mitochondrial components were shown to translocate between cells, thus providing intercellular signaling cues able to (i) alert the recipient cells of a danger situation [13], (ii) restore their biological functions [14] or, in the case of cancer cells, (iii) modify their functional capacities and response to therapy [15,16]. In conditions of severe tissue damage, multiple mitochondrial elements, including mitochondrial DNA (mtDNA), N-formyl peptides, ATP or cardiolipin, were shown to be liberated from the dying cells to the surrounding tissue and to the bloodstream [17,18]. These mitochondrial products are recognized as damage associated molecular patterns (DAMPs) by specific receptors on immune cells and consequently trigger innate and adaptive inflammatory responses [19,20]. In addition, whole mitochondria can also be transferred between a broad diversity of cell types by the means of specialized structures including tunneling nanotubes (TNTs) [16,21] or microvesicles (MVs) [21,22]. A large part of the studies in this field focused on mesenchymal stem cells (MSCs) and their ability to communicate through organelle exchange with their surrounding environment following tissue injury or during cancer progression [23]. Here, we review the biological consequences of the mitochondria transfer between MSCs and differentiated cells, in damaged tissues as well as in tumors. We report both situations that is MSCs as either mitochondria donors or receivers. Finally, we discuss the mechanisms underlying this process, the fate of the transferred organelles in the recipient cells and the potential therapeutic applications of mitochondrial exchanges as means to repair damaged organs or treat mitochondrial inherited diseases and cancer.

Mitochondria, released by MSCs, as pro-survival effectors

Mesenchymal stem cells hold great promise for regenerative medicine due to their plasticity, their pro-angiogenic and anti-apoptotic functions and their immune-regulatory properties [24,25]. Although most of the beneficial effects exerted by MSCs have primarily been ascribed to paracrine mechanisms [26], this concept recently evolved with the discovery of the outstanding capacity of MSCs to share their mitochondria with target cells, resulting in the protection of the recipient cells from tissue injury. In this section, we will discuss to what extent the mitochondria delivery from MSCs to the recipient cells can be beneficial for wound healing processes and immune regulation and, on the other hand, be deleterious for the organism in the case of cancer.

MSC-mediated mitochondria transfer in cell rescue

Ten years ago, Spees and collaborators provided the first evidence that mitochondria or mtDNA could translocate from MSCs to mammalian cells harboring nonfunctional mitochondria, due to their lack of effective oxidative phosphorylation [27]. This pioneer study was revealing how, through such a phenomenon, MSCs could restore aerobic respiration in mtDNA-depleted human lung alveolar epithelial A549 ρ⁰ cells. In the following years, the mitochondria donor capacity of MSCs was confirmed by several laboratories worldwide as well as the physiological significance of this phenomenon. On the whole, these studies indicate that MSCs preferentially release their mitochondria to suffering or damaged cells and that this process results in the rescue of the recipient cells from cell death by preserving their energy metabolism. Through coculture settings, mitochondria transfer was observed to occur in vitro from MSCs to various kinds of differentiated cells, including cardiomyocytes [28,29], endothelial cells [30], bronchial epithelial cells [31], corneal epithelial cells [32] and neuronal cells [33]. These studies showed that the mitochondria transfer from MSCs conferred protection against apoptosis to damaged cells following exposure to several stressor stimuli such as ischemic/reperfusion injury [28,30,32,33], oxygen/glucose deprivation [30] and tobacco smoke exposure [31]. Mitochondria donation by MSCs was invariably found to improve the survival of the injured cells and to increase their respiratory function and ATP production. The rescue of the cellular bioenergetics of the differentiated cells required the delivery of functional respiring mitochondria by the MSCs as shown by the loss of the cytoprotective function of mtDNA-depleted MSCs (ρ⁰ cells) [29,30].

The transfer of mitochondria from MSCs to differentiated cells was also observed in animal models for tissue injuries such as ischemic heart [34], injured lung through exposure to LPS [35], rotenone [36] or cigarette-smoke [31] and rotenone-treated cornea [32]. These in vivo studies substantiated the initial in vitro coculture observations and confirmed that engrafted MSCs can transfer mitochondria to damaged cells, resulting for these cells in a pro-survival outcome through the OXPHOS-dependent restoration of their ATP production.

Finally, MSCs were demonstrated to have the capacity to reprogram fully differentiated mouse cardiomyocytes back to a cardiac progenitor-like state, in a process that relied on the mitochondrial transfer from MSCs [29]. In these settings, the mitochondria transfer from MSCs was observed to improve the survival of the mature cardiomyocytes in vitro, which is a prerequisite for their reprogramming back to a progenitor state.

MSC-mediated mitochondria transfer in inflammation

The therapeutic benefits of MSCs have been partly attributed to their immunosuppressive properties and their ability to regulate the functions of many cell types, from the innate and adaptive immune systems, such as dendritic cells [37–39], T-lymphocytes [40,41] and also macrophages [42–44], these latter cells playing a critical role in tissue repair by ensuring the clearance of dying cells and cell debris through phagocytosis. Two main groups of macrophages, differing by their pro- or anti-inflammatory phenotypes, co-exist within the healing wound. Following tissue damage, macrophages initially adopt a pro-inflammatory phenotype M1, that is then switched towards an anti-inflammatory pro-healing M2 phenotype at the time of resolution of inflammation [45–47]. MSCs are known to favor the macrophage differentiation towards an anti-inflammatory/pro-healing M2 phenotype [42–44], More recently, the transfer of mitochondria from MSCs to macrophages was observed to occur both in vitro and in vivo and shown to drive phenotypic changes in the macrophages [48–50]. In particular, Jackson and colleagues as well as Morrison and colleagues provided evidence that the mitochondria conveyed by MSCs, in the context of the Acute Respiratory Distress Syndrome (ARDS), increased the oxidative phosphorylation of the recipient macrophages and then stimulate their phagocytic activity [48,49] and their differentiation towards a M2 anti-inflammatory phenotype [50]. In addition, the inhibition of this mitochondria transfer was shown to abrogate the antimicrobial effects of MSCs following their engraftment in mice suffering from bacterial pneumonia (ARDS), supporting the importance of this process in the regulation of macrophage functions and bacteria clearance [49].

Interestingly, the transfer of mitochondria to macrophages does not solely occur from healthy but also from damaged MSCs [51]. In this latter context, this process was proposed as a mechanism allowing stem cells to get rid of their deleterious organelles to improve their own survival, although it could also be envisioned as a means of alerting macrophages of danger situations [51], as discussed in section II.

MSC-mediated mitochondria transfer in tumor progression

The recent research efforts to better understand the cross-talk between cancer cells and their microenvironment identified mitochondria transfer as a process contributing to the tumor development and progression. In a fashion comparable to that observed in the context of tissue repair, MSCs were shown to deliver mitochondria to various kinds of malignant cells, including those from breast and ovarian cancer, melanoma, acute myeloid leukemia and glioblastoma [52–55], resulting in induced invasiveness and resistance to chemotherapy.

The seminal work reporting on the horizontal mitochondria transfer was actually performed on A549 lung adenocarcinoma cells [27]. These mitochondria acceptor cells were ρ⁰ cells, harboring a defective mitochondrial DNA (mtDNA) after chronic ethidium bromide treatment and, as a consequence, having an inoperative respiratory chain and respiration. These ρ⁰ cells rely on glycolysis and are dependent of exogenous supplementation of pyruvate and uridine in the culture medium (auxotrophy). After the mitochondria transfer, evidenced by the detection in the acceptor cells of the mtDNA from the donor cells, the A549 cells recovered a respiratory function and an oxidative metabolism while they lost their auxotrophy [27]. Other ρ⁰ cells, including melanoma and breast solid tumor cells, have an increased tumor latency compared to the parental mitochondrial competent cells. It was nicely demonstrated that mitochondrial transfer from the tumor microenvironment toward these ρ⁰ cells could fully restore their respiration and invasiveness pattern [53]. Using C57BL/6Nsu9-DsRed2 mice that express a red fluorescent protein in their mitochondria, Neuzil and collaborators recently established the transfer of whole mitochondria from the host animal towards the injected B16 ρ⁰ mouse melanoma cells [56]. It is worth mentioning that the permanent recovery of the mitochondrial function of the ρ⁰ cells was achieved using donor and recipient cells either from the same murine species [53] or from different species (human and mouse) [27], suggesting a lack of species barrier for this particular phenomenon. However, long-lasting acquisition of exogenous mitochondria was not reported for other non-ρ⁰ cancer cell models. It will require further investigation to determine whether the auxotrophic status of the ρ⁰ cells constitutes a selection pressure and leads to mitochondria acquisition mechanisms different from those of the other cell systems described so far.

Independently of the auxotrophic issue of ρ⁰ cells, a higher tumorigenicity upon mitochondrial transfer was also observed for mitochondrial competent leukemic and bladder cancer cells [54,57]. The recipient cancer cells for the mitochondria exchange displayed a higher tumorigenic potential. Besides, the mitochondria recipient cells were consistently shown to display an increased oxidative phosphorylation (OXPHOS) phenotype in the cell systems independently investigated [30,35,54,58,59].

Mitochondria transferred to MSCs as sensors for tissue homeostasis

As detailed above, mitochondria trafficking from MSCs to differentiated cells and their biological outcomes have been reported by several laboratories. Conversely, mitochondria can also be transferred in the reverse orientation i.e. from fully differentiated cells to MSCs. Although this has been much less studied so far, current in vitro studies provide evidence that mitochondria released by differentiated cells can be captured by MSCs and that this process contributes to the maintenance of tissue homeostasis. In this section, we will provide an overview on how mitochondrial transfer promotes an adaptive response of recipient MSCs to face up to micro-environmental demands. We will distinguish the effects of mitochondria transferred from either “healthy” or “damaged” differentiated cells.

Mitochondria transferred from healthy differentiated cells

Several studies indicate that MSCs or progenitor cells co-cultured with fully differentiated cells can acquire the phenotype of these differentiated interacting cells, as shown for renal tubular cells and cardiomyocytes, and that this phenomenon is accompanied by mitochondrial transfer from the differentiated cells [60,61]. Although the actual role of the exogenous mitochondria in the acquired phenotype of the MSCs was not assessed in these studies, it is likely that the transfer of the external mitochondria provides to stem cells a means to regulate their metabolism, that was shown to be essential for their plasticity [62,63]. In particular, the exogenous supply of mitochondria is expected to favor the metabolic switch of the MSCs from glycolysis to oxidative phosphorylation, a process that was reported to be required for their differentiation [64–68]. In addition, mitochondria transferred from vascular smooth muscle cells (VSMCs) to MSCs were found to stimulate their proliferation in vitro. The actual role of the transferred VSMC mitochondria in promoting MSC proliferation was based on the fact that both the inhibition of this process and the transfer of mitochondria with impaired respiratory function, by long-term ethidium bromide treatment (ρ⁰), abrogated this effect [69]. Although the mechanisms underlying the proliferative effects of transferred mitochondria have not been addressed in this study, exogenous mitochondria are expected to modulate the bioenergetics of MSCs, known as a key regulator of their growth [70–72].

Mitochondria transferred from “damaged” differentiated cells

Beyond their role as energy producers and cell metabolism regulators, mitochondria are capable of sensing the cellular stress generated following tissue injury and of relaying danger signals, at both the intracellular and intercellular levels [73]. Following acute tissue injury, damaged cells release various mitochondrial components in the bloodstream. These components are recognized as damage associated molecular patterns (DAMPs) by innate immune cells (including neutrophils and monocytes/macrophages) through their interactions with specific pattern recognition receptors [17,74]. As a result, mitochondrial DAMPs elicit a sterile immune response (i.e. in the absence of any microorganisms). Mitochondrial DAMPs include mtDNA, N-formyl peptides, mitochondrial proteins, such as TFAM and cardiolipin, and ATP [14,74]. In addition, recent findings indicate that mitochondria, as whole organelles, can also act as DAMPs following acute injury [13] or inflammation [75]. In particular, our laboratory reported that the transfer of mitochondria from apoptotic endothelial or cardiac cells to MSCs constitute a signaling messenger that triggers the cytoprotective response of the MSCs, contributing to their supporting action for the injured cells [13]. We found that the conveyed organelles were degraded by MSCs through a mitophagy process involving the stress-inducible heme oxygenase-1 (HO-1) signaling pathway, associated with the stimulation of mitochondrial biogenesis. As a result, MSCs enhanced the donation of their own mitochondria towards the damaged cells to rescue them. Importantly, this cross-talk signaling was shown to depend on reactive oxygen species (ROS) produced by the damaged cells since a ROS scavenger abrogated both the mitochondrial transfer from the injured cells to the MSCs and the MSC rescuing function [13]. The molecular mechanisms whereby the mitochondrial ROS produced by injured cells activate the cytoprotective functions of MSCs remain to be carefully investigated. It is worth considering, though, that mitochondrial ROS are known to stimulate autophagy processes [76–78] and to regulate the HO-1 signaling pathway [76,77,79–81].

Mechanisms of intercellular mitochondria transfer

As described above, MSCs clearly demonstrate a capacity to exchange mitochondria with a diversity of cell types. For this, they can either form tunneling nanotubes (TNTs) and/or extracellular vesicles (EVs) as it will be discussed in this section. It is worth noting that this represents a novel property for MSCs that were, up to now, mostly characterized for their interactions with neighboring cells through secreted cytokines [82–84].

Mitochondrial transfer through tunneling nanotubes

Tunneling nanotubes were first described in 2004 by the two groups of Gerdes [85] and Davis [86]. Ever since, the interest in TNTs and the number of cell types shown to be involved in such intercellular connections has been steadily growing. TNTs are long tubular structures, connecting cells together, that can reach lengths of several hundreds of microns with diameters from 50 to 1 500 nm [87]. Actually, the fact that TNTs involve a continuity in plasma membrane and cytoplasm between the connected cells is radically changing our conception of the cell, limited by its own plasma membrane [88]. MSCs are particularly prone to engage in TNT connections and mitochondria donation to different cell types, including cardiomyocytes [29,59,89], lung epithelial cells [31,35,36], renal tubular cells [60], vascular smooth muscle cells [69], corneal epithelial cells [32] and macrophages [49] (Table 1). Apart from their role in tissue homeostasis, MSCs are also known to be recruited to tumors [90,91]. TNT-mediated mitochondria transfer was actually observed between MSCs and different cancer cells, including breast and ovarian cancers, melanoma, acute myeloid leukemia as well as glioblastoma as shown in Figure 1 [52–55].

Table 1.

Cell types involved in TNT-mediated mitochondria transfer with MSC and biological outcome.

References	Cell types involved in TNTs with MSCs	Cargo transferred	Biological Outcome
Spees et al. (2006)	A549 ρ0 lung epithelial cells	Mitochondria/mtDNA	Aerobic respiration restoration
Plotnikov et al. (2008)	Rat cardiomyocytes	Mitochondria	Restoration of cardiomyocytes bioenergetics
Plotnikov et al. (2010)	Rat renal tubular cells	Mitochondria	MSC differentiation into kidney tubular cells
Acquistapace et al. (2011)	Adult cardiomyocytes	Mitochondria	Conversion to progenitor-like state by metabolic reprogramming
Islam et al. (2012)	Damaged murine alveolar epithelial cells	Mitochondria	Tissue repair (ATP level restoration, pulmonary surfactant secretion)
Vallabhaneni et al. (2012)	Vascular smooth muscle cells	Mitochondria	Increase in MSC proliferation
Pasquier et al. (2013)	Human ovarian and breast cancer cell lines	Mitochondria	Increased doxorubicin chemoresistance
Ahmad et al. (2014)	Stressed murine lung epithelial cells	Mitochondria	Lung injury repair and mouse survival
Li et al. (2014)	Lung epithelial cells exposed with cigarette smoke	Mitochondria	Decrease in cigarette-smoke-induced alveolar damage
Liu et al. (2014)	Human umbilical vein endothelial cells (HUVEC)	Mitochondria	Injury rescue, increased oxygen consumption
Caicedo et al. (2015)	MDA-MB-231 breast cancer cells	Mitochondria	Increased OXPHOS, ATP production, invasion & proliferation
Han et al. (2015)	Ischemic H9c2 rat cardiomyocytes	Mitochondria	Apoptosis decrease, mitochondrial function restoration
Tan et al. (2015)	Murine B16 ρ0 melanoma and 4T1 ρ0 breast carcinoma	mtDNA	Restored respiratory functions, increased tumor-initiation
Jackson et al. (2016)	Macrophages	Mitochondria	Enhanced macrophage phagocytosis and improved bioenergetics
Jiang et al. (2016)	Corneal epithelial cells	Mitochondria	Protection against oxidative-stress-induced mitochondrial damage
Moschoi et al. (2016)	Acute Myeloid Leukemia cells	Mitochondria	Increased chemoresistance to ARA-C
Zhang et al. (2016)	Cardiomyocytes	Mitochondria	Rescue of anthracycline-induced myopathy
Nzigou Mombo et al. (2017)	Glioblastoma stem cells	Mitochondria
Sanchez et al. (2017)	Wharton's jelly mesenchymal stem cells	Mitochondria

Figure 1.

Tunneling nanotube (TNT) formation and mitochondria transfer from mesenchymal stem cells (MSC) to glioblastoma stem cells (GSC).

MSCs and GSCs were labeled beforehand, resp. with red MitoTracker CMXRos and green CellTracker CMFDA. The coculture was performed for 48h and fluorescence microscopy was performed on the fixed cells, using a Leica SPE confocal microscope. A TNT connection is observed between the two cell types. The lower frame shows a different MitoTracker and CellTracker exposure that allows the observation of MSC mitochondria (in red) observed inside the TNT and the glioblastoma cell. Scale bar = 50 μm.

Connexin 43 and Miro-1 (RHOT1) were shown to control TNT formation and the TNT-mediated mitochondria exchange, respectively [35,36]. Connexin 43 (Cx43), also known as GAP Junction Alpha-1 protein (GJA1), is a transmembrane protein that assembles as hexamers to form connexons. Two connexons from two neighboring cells can then dock end-to-end to establish the intercellular channel, called GAP junction, that allows the direct cell-to-cell transfer of small molecules (Goodenough and Paul, 2009). Cx43 was first described to be implicated in the formation of TNTs and the establishment of GAP junctions between the two connecting cells in a murine model of LPS-mediated lung injury [35]. In fact, while Cx43-expressing human MSCs were able to rescue the injured pulmonary alveolar epithelial cells via TNT formation, Cx43-depleted MSCs were unable to form TNTs and, consequently, could not reduce lung injury [35]. Miro-1, an outer mitochondrial membrane Rho-GTPase, was also shown to play an important role in mitochondria trafficking through TNTs. In neurons, Miro-1 is known to interact with the adaptor proteins TRAK1-2 and with mitofusins 1–2, resulting in the recruitment of kinesin motor proteins to the mitochondria and their shuttling along the axons and dendrites microtubules [92,93]. In models of either rotenone or ovalbumin-induced airway injury, MSCs overexpressing Miro-1 transferred higher amounts of mitochondria to the injured murine alveolar epithelial cells and induced more efficient tissue repair than the wild-type MSCs. Conversely, Miro-1-knocked-down MSCs lost their mitochondria transfer capacity and healing ability [36]. Besides, iPS-MSCs expressing high Miro-1 concentrations also demonstrated a higher rate of mitochondria trafficking, leading to the rescue of anthracycline-mediated cardiomyopathy [59].

Mitochondrial transfer through microvesicles

Intercellular trafficking can also rely on extracellular vesicles. These comprise microvesicles and exosomes. Microvesicles are circular fragments, with diameters from 0.1 to 1 μm, that are shed through blebbing and budding processes from lipid-raft-enriched cell membranes. Exosomes, on the other hand, are smaller fragments (40 to 150 nm in diameter) derived from endosomal cell membranes that originate from multivesicular bodies (MVBs). Whereas, due to their smaller size, exosomes can only load mitochondrial DNA, miRNAs, and small proteins like cytokines and chemokines, microvesicles do have the capacity to carry and transport whole mitochondria [94]. Mitochondria transfer through microvesicles has been less extensively described than through TNTs. It was first reported by Islam et al. in 2012 [35] for damaged murine lung alveolar epithelial cells. Interestingly, the authors reported that, in this cell system, the transfer of MSC mitochondria to the lung epithelial cells could occur through both TNTs and microvesicles. Other cell types were since demonstrated to acquire MSC mitochondria via microvesicles, notably macrophages [50,51] and renal tubular epithelial cells [95] (Table 2).

Table 2.

Cell types involved in microvesicle-mediated mitochondria transfer with MSC and biological outcome.

References	Cell types involved in microvesicles trafficking with MSCs	Cargo transferred	Outcome
Islam et al. (2012)	Damaged murine alveolar epithelial cells	Mitochondria	Increased ATP production, pulmonary surfactant secretion, lung injury repair
Phinney et al. (2015)	Macrophages	Mitochondria	Enhanced bioenergetics
Jackson et al. (2016)	Monocyte-derived macrophages	Mitochondria	Enhanced bioenergetics and macrophage phagocytosis
Morrison et al. (2017)	Macrophages	Mitochondria	Enhanced bioenergetics and macrophage phagocytosis, lung injury repair

It is remarkable that some cells, like lung epithelial cells and macrophages, can simultaneously employ both mechanisms of mitochondrial transfer, TNTs being apparently more potent in the case of macrophages [35,49]. In fact, following a cytochalasin B pre-treatment at concentrations that block TNT formation without affecting endocytosis, Jackson et al. reported that macrophages acquire MSC mitochondria simultaneously through a TNT-mediated, contact-dependent mechanism and, less intensively, through contact-independent mechanisms.

Transfer of free mitochondria

Beyond TNTs and microvesicles, cell extrusions could constitute an alternative process allowing the conveyance of mitochondria between cells. Although the occurrence of such a phenomenon is yet to be documented for intercellular mitochondria trafficking involving MSCs, several studies indicate that functional mitochondria can be released, during inflammation, as free organelles by different cell types such as platelets [96] to activate neutrophils, or by TNF-α-induced necroptotic cells to activate macrophages and dendritic cells [75]. Importantly, the occurrence of this phenomenon was also detected in vivo [96], pointing out the need to further study this mitochondria transfer mechanism and determine whether it is also relevant for MSC-related mitochondria exchange.

Cells have also been described to internalize free mitochondria, isolated in vitro beforehand. This phenomenon, dubbed “mitochondrial transformation”, was first described in 1982 by Clark and Shay. They observed that, by simple co-incubation, chloramphenicol-(CAP^s) and efrapeptin-(EF^s) sensitive mammalian cells were able to incorporate mitochondria, purified beforehand from CAP^r and EF^r resistant cells [97]. Ever since, many cells have been reported to undergo mitochondrial transformation [55,98–102]. This process is believed to occur via macropinocytosis. In fact, pre-treating cells with EIPA, a macropinocytosis inhibitor, impeded the mitochondria uptake in a dose-dependent manner whereas chlorpromazine, a clathrin-mediated endocytosis inhibitor, had no effect [98,100]. Mitochondrial transformation depends on the integrity of the outer mitochondrial membrane and the presence of intact fusogens that conduct effective mitochondrion-cell interactions [100,101]. For instance, Díaz-Carballo and colleagues reported that mitochondria acquisition in chemo-resistant U87^RETO glioblastoma cells is syncytin-mediated. In fact, the outer mitochondrial membrane harbors both HERV-WE₁ (syncytin-1) and HERV-FRD₁ (syncytin-2) and their respective receptors ASCT2 and MFSD2 in order to accomplish successful mitochondrial transformation, which was blocked by anti-syncytin 1–2 antibodies [101]. Mitochondria uptake was also described in vivo in a rabbit ischemia model. Cardiomyocytes were able to internalize free autologous mitochondria transplanted at the ischemia site, resulting in the reduction of the infarct size and the increase of contractility and ATP production [103]. Following the efficiency of mitochondria transplantation on cardioprotection in animal models, the first clinical application was conducted on five pediatric patients suffering from myocardial ischemia-reperfusion injury, leading to a significant improvement in their myocardial systolic function [104].

Unsolved questions and perspectives

As outlined above, the whole process of intercellular mitochondria trafficking, notably through TNTs, constitutes a novel biological concept [88]. It is of great interest both for the sake of a holistic comprehension of the cellular interactions taking place in tissues and for the possible clinical applications of this discovery. As this research domain is still in its early phases of study, a number of key questions obviously remain to be answered. For instance, it is unknown whether the mitochondria destined to be exchanged undergo a pre-selection process. It could be either the selection of “healthy” mitochondria originating from “healthy” cells and targeted to damaged cells, resulting in the repair of these cells or, conversely, mitochondria endowed with a lowered membrane potential released from cells of an injured tissue. Mitochondria are also known to undergo the dynamic process of fusion/fission [105,106]. So far, it is unclear to what extent this process interferes with intercellular mitochondria trafficking. However, mitochondria dynamics were reported to be affected by exogenous mitochondria delivery in vitro [13,107] and to be closely linked to mitophagy, mitochondria biogenesis and motility [13,81,108]. Consequently, this mitochondria fusion/fission process is predicted to play a critical role in the fate of the conveyed organelles and of their genome. Persistence of the mitochondria and their mtDNA, following their transfer from Wharton's Jelly MSCs (WJ MSCs), was recently reported by Chuang and colleagues, in cybrids harboring mitochondrial DNA mutations for the MERFF (Myoclonus epilepsy associated with ragged-red fibers) syndrome [107]. The authors reported that exogenous mitochondria correct the MERFF phenotype, in the long-term, through their fusion with the recipient mutated mitochondria. This led to mtDNA heterosplasmy and, as a consequence, to a reduction in the mutant mtDNA load in the MERRF cybrids [107].

A number of questions also arise about the cellular mechanisms enabling the mitochondria trafficking from one cell to the other. In particular, it still to be determined whether a given cell type exclusively uses TNTs, microvesicles or cell extrusion for mitochondrial transfer or whether it can accommodate all three. In addition, further investigation will be required to assess the effects of the mitochondria state per se or of more general factors, like the microenvironment conditions, in promoting either one of these mitochondria transfer mechanisms.

A robust knowledge of the processes that allow intercellular mitochondria transfers is expected to nurture clinical applications in a number of medical fields. First of all, the discovery of mitochondria transfer originating from MSCs is anticipated to open new avenues for the optimization of MSC-based therapies, notably for degenerative diseases associated with impaired mitochondrial functions. These include diseases affecting the central nervous system (i.e. Alzheimer and Parkinson diseases) [109,110], the cardiovascular system (myocardial infarction/ischemia) [29,30] and the lung (chronic obstructive pulmonary disease) [31].

Beyond the treatment of degenerative pathologies, the mitochondrial donor capacity of MSCs could be exploited for other clinical purposes. It could notably be extended to human diseases characterized by mitochondrial DNA mutations, as suggested by the recent findings on the MERRF syndrome [107]. This maternally inherited disease affects the nervous system and skeletal muscles. It is associated with a point mutation in the mitochondrial tRNA^Lys encoding gene, leading to severe defects in protein synthesis, to oxidative stress and impaired OXPHOS [111]. In their recent study, Chuang and al. reported that the delivery of healthy MSC mitochondria rescued the phenotype of the MERRF cybrids by reducing the oxidative stress and improving mitochondrial bioenergetics [107].

Finally, targeting the mitochondria transfer process, with the goal of inhibiting it rather than activating it, also appears as a promising adjuvant approach to combat cancer progression. It would be very instructive to be able to establish links between the occurrence of mitochondria transfers within tumors and clinical criteria such as cancer progression, relapse or overall patient survival. Nevertheless, it was established that mitochondria transfers lead to drug resistance, as shown for solid tumors and leukemia, therefore contributing to the general process of environment-mediated drug resistance (EM-DR) [52,54]. It is worth noting that this mitochondria-related EM-DR is further increased by chemotherapeutic agents like doxorubicine, cytarabine or etoposide [52,54,57,59,112] while other conventional chemotherapies, such as vincristine or cytochalasin that target the cytoskeleton, do not demonstrate these effects [54,112]. Overall, these data suggest the targeting of the intercellular mitochondria transfer as an attractive novel adjuvant approach to improve current anticancer regimens.

Conclusion

The mitochondria transfer occurring between cells, notably to and from MSCs, constitutes a fascinating and still largely unexplored process with expected pleiotropic effects on diseases including degenerative diseases, cancers and mitochondrial inherited pathologies. Novel insights on the mechanisms involved in these mitochondria exchanges will be valuable to pharmacologically stimulate or abrogate these organelles exchanges and, accordingly, to develop novel therapeutic approaches to regenerate damaged organs, treat defective mitochondria-related diseases and curb cancer progression.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank the Montpellier RIO imaging facility (MRI) for providing adequate environment for confocal microscopy. AMR and M.L.V are staff scientists from the French National Institute of Health and Medical Research (INSERM) and the National Center for scientific research (CNRS), respectively. J.N. is supported by a PhD fellowship from the French Ministry of Research. This work was supported by grants from the Ligue Contre le Cancer-Comité de l'Aude and Association pour la Recherche et l'Etude des Maladies Cardiovasculaires (AREMCAR).