Neuroprotection offered by mesenchymal stem cells in perinatal brain injury: Role of mitochondria, inflammation, and reactive oxygen species

Abstract

Preclinical studies have shown that mesenchymal stem cells have a positive effect in perinatal brain injury models. The mechanisms that cause these neurotherapeutic effects are not entirely intelligible. Mitochondrial damage, inflammation, and reactive oxygen species are considered to be critically involved in the development of injury. Mesenchymal stem cells have immunomodulatory action and exert mitoprotective effects which attenuate production of reactive oxygen species and promote restoration of tissue function and metabolism after perinatal insults. This review summarizes the present state, the underlying causes, challenges and possibilities for effective clinical translation of mesenchymal stem cell therapy.

Mesenchymal stem cells have a positive effect in perinatal brain injury models. Mitochondrial damage, inflammation, and reactive oxygen species are critically involved in the development of injury. Mesenchymal stem cells have immunomodulatory action and exert mitoprotective effects which attenuate production of reactive oxygen species and promote restoration of tissue function and metabolism after perinatal insults. This review summarizes the present state, the underlying causes, challenges and possibilities for effective clinical translation of mesenchymal stem cell therapy. The image illustrates transfer of mitochondria from mesenchymal stem cells to damaged neurons and inflamed microglia via tunneling nanotubes.

Abbreviations

AT‐MSCs

adipose tissue‐derived mesenchymal stem cells

ATP

adenosine triphosphate

BAG5

BCL2‐associated athanogene 5

BAX

Bcl‐2‐associated X protein

BDNF

brain‐derived neurotrophic factor

BMSCs

bone marrow‐derived stem cells

DC

dendritic cells

ES

embryonic stem cells

H₂O₂

hydrogen peroxide

HI

hypoxia‐ischemia

HIE

hypoxic‐ischemic encephalopathy

HIF‐1α

hypoxia‐inducible factor‐1α

IV

intravenous

IVH

intraventricular hemorrhage

LPS

lipopolysaccharide

MAPK

mitogen‐activated protein kinase

MCAO

middle cerebral artery occlusion

MLKL

mixed lineage kinase domain like protein

MSCs

mesenchymal stem cells

NE

neonatal encephalopathy

NK

natural killer

NLRs

NOD‐like receptors

O2•−

superoxide

OGD

oxygen glucose deprivation

OXPHOS

oxidative phosphorylation

RLRs

RIG‐like receptors

ROS

reactive oxygen species

TLR

Toll‐like receptors

TNTs

tunneling nanotubes

TSG‐6

TNF‐α‐induced protein

UCB

umbilical cord blood

1. INTRODUCTION

Intraventricular hemorrhage (IVH) in preterm neonates, stroke in preterm and term neonates, and neonatal encephalopathy (NE) in term neonates constitute three major destructive brain lesions observed during the perinatal period (Ferriero et al., 2019; Hill & Volpe, 1981). Treatment options for neonatal HI brain injury are limited to supportive care and therapeutic hypothermia (Hagberg et al., 2016; Jacobs et al., 2013). Hypothermia is only advised in term and late preterm infants with moderate‐to‐severe NE if identified before 6 hr of age (Jacobs et al., 2013). For cases of IVH unfortunately only supportive care can be provided. Pharmacological neuroprotective therapies have been thoroughly explored in multiple animal models of hypoxic‐ischemic encephalopathy (HIE) and IVH (Bel & Groenendaal, 2016). However, none have been successfully translated into a therapy to prevent or regenerate brain injury in newborn infants.

Inflammation is strongly recognized as a major cause leading to long‐term injury (Hagberg et al., 2015). Inflammation causes systemic up‐regulation of pro‐inflammatory cytokines and diffuse activation of microglia in the neonatal brain (Berger et al., 2012). Cytokine‐activated cells secrete toxic substances, such as ROS and toxic granules, including proteolytic enzymes and myeloperoxidase (Okazaki et al., 2006). This results in cell death causing white matter damage and long‐term neurological injury among pre‐term and term neonates (Leviton et al., 2005).

Cell death associated with HI and IVH can be classified as apoptotic, necrotic, or autophagic (Bobinger et al., 2018; Northington et al., 2011). Pro‐apoptotic proteins act directly (BAX and BAK) or indirectly (BAD, PUMA, BIM) as triggers of mitochondrial membrane permeabilization leading to activation of caspase activation and apoptosis‐inducing factor (AIF) translocation to nuclei which is believed to be crucial for injury to advance beyond the point of no return (Blomgren et al., 2001; Hagberg et al., 2009, 2014). Recruitment of mixed lineage kinase‐like protein (MLKL) to the necrosome results in mitochondria‐associated endoplasmic reticulum membranes triggering increased reactive oxygen species, fission, and necroptosis (Galluzzi et al., 2016). Death receptors trigger apoptotic death via caspase‐8 and necroptotic cell death through the formation of the necrosome, with both apoptotic and necrotic pathways converging at the mitochondria (Chavez‐Valdez et al., 2012, 2016; Galluzzi et al., 2011). Autophagy is increased in neuronal cells after neonatal HI and may represent a potential protective mechanism in the early stage of the brain injury (Carloni et al., 2008) but most data show that autophagy aggravates injury (Descloux et al., 2015; Xie et al., 2016).

Experimental results showing regeneration of the injured neonatal brain with stem cell‐based therapies are promising. Studies have shown that exogenous administration of stem cells significantly decreases brain injury in models of IVH (Ahn et al., 2013, 2015), HIE (Park et al., 2015), and neonatal stroke (Kim et al., 2012). These findings together with studies from adult neurological conditions (Caicedo et al., 2015) suggest that exogenous MSC transplantation might be a promising new therapy in the not too distant future for treating neonatal brain damage resulting from IVH, NE, or neonatal stroke.

2. MSCS—THE CELL OF CHOICE?

Major types of stem cells include embryonic stem cells, chemically induced pluripotent stem cells, neural stem cells, UCB cells, and MSCs. Embryonic stem (ES) cells have multiple characteristics such as self‐renewal, highly expandable, and possibility to derive neural stem cells and other neurogenic lineages (Svendsen & Smith, 1999) (Hynes & Rosenthal, 2000). Although ES seems an obvious choice for repairing brain injury, they can induce the formation of teratomas after transplantation (Bjorklund et al., 2002). Similarly, multipotent neural stem cells have the capacity to derive all neural lineages, but they also carry a significant risk of tumor formation (Comi et al., 2008; Fleiss et al., 2014). UCB contains MSCs, endothelial progenitor cells, and UCB‐mononuclear cells (Ingram et al., 2004). The advantages of UCB cells include easy access, limited ethical issues, and low immunogenicity. Administration of MSCs several days after neonatal HI markedly improves functional outcome, reduces infarct volume, and stimulates neurogenesis (Hawkins et al., 2018; van Velthoven et al., 2010a).

MSCs are characterized by their multi‐lineage differentiation capacity (Keating, 2012; Phinney et al., 2015; Pittenger et al., 1999), they express specific cell surface markers which include CD105 (Barry et al., 1999), CD73 (Barry et al., 2001), CD90 (Dominici et al., 2006), CD79 (Chu & Arber, 2001) and CD44 (L Ramos et al., 2016), but not CD79, CD19, CD14 (Wright et al., 1990), CD11b (Coxon et al., 1996), HLA‐DR (Dominici et al., 2006), and the hematopoietic marker CD45 (Irie‐Sasaki et al., 2001). They also lack co‐stimulatory proteins like CD80, CD40, and CD86 which makes them less immunogenic (Tse et al., 2003). Additionally, other non‐specific MSCs surface markers, like Stro‐1, SSEA‐4, CD271, and CD146, have also been proposed for positive characterization of MSCs (Lv et al., 2014). The International Society for Cellular Therapy has provided a consensus perspective on immune functional assays, functionally relevant surface markers etc., for MSCs as potency release criterion for clinical trials (Galipeau et al., 2016).

MSCs can be isolated from human placenta, umbilical cord (blood and Wharton's jelly), spleen, and adipose tissue‐derived mesenchymal stem cells (AT‐MSCs) (Gothard et al., 2014; Strioga et al., 2012) other than bone marrow. Other sites of MSC isolation include skin (Shih et al., 2005), dental pulp (Gronthos et al., 2000) and amniotic fluid (Nadri & Soleimani, 2007). Although these MSCs share common properties, they also exhibit differences in the expressed cell markers, in their differentiation potential and phenotypes (Gothard et al., 2014; Strioga et al., 2012).

Isolated MSCs from human submandibular skin tissues exhibit MSC characteristics and express a neural precursor marker, suggesting that human skin is a source of stem cells. However, the in vitro chemical neuronal induction of hMSC does not produce long‐lasting nerve cells and more studies are required before their use (Byun et al., 2012).

Recent studies have shown that tissue‐resident MSCs also express CD34, originally thought to be a hematopoietic marker as it is normally not found in culture‐expanded MSCs (Lin et al., 2012; Togarrati et al., 2017). CD146, a pericyte marker has been proposed to give rise to MSCs following blood vessel damage or inflammation (Bouacida et al., 2012; da Silva Meirelles et al., 2016). The secretome of MSCs has also been examined (although it varies by culture condition), and includes several growth factors such as vascular endothelial growth factor, brain‐derived neurotrophic factor, nerve growth factor, basic fibroblast growth factor as well as anti‐inflammatory cytokines (Qu et al., 2007). Growth factors secreted by MSCs mediates the proliferation and differentiation of progenitor cells (Ge et al., 2018; Martins et al., 2017).

MSCs have many favorable properties, they are non‐tumorigenic, have a positive safety profile, exhibit low immunogenicity and are easy to isolate/propagate reviewed in (Fleiss et al., 2014; Wagenaar et al., 2017). Also, MSCs have multiple therapeutic actions. They reduce inflammation and oxidative stress, restore energy failure, and decrease cell apoptosis (Li et al., 2012). MSCs boost the growth‐promoting environment for neural stem cells (Yoo et al., 2008). It has been shown that MSC mediated cell‐cell contacts provide additional signaling to improve adipogenic and osteogenic differentiation (Peng et al., 2012) through the increased gap junction or cadherin signaling (Lee et al., 2012; Tang et al., 2010). These effects of MSCs address many of the complex multifactorial factors driving tissue injury after severe IVH (Ballabh, 2010, 2014), or HIE (Aslam et al., 2019; Thornton et al., 2017; Wagenaar et al., 2018). A multifaceted therapeutic agent like MSCs may be suitable to improve outcomes of patients with these neonatal disorders (Park et al., 2018) (Figure 1).

FIGURE 1.

Mechanisms of neuroprotection offered by mesenchymal stem cells

3. EVIDENCE FOR MSC‐MEDIATED PROTECTION AND REPAIR OF PERINATAL BRAIN INJURY

The protective effects of MSCs, have been observed in various animal models of severe IVH (Ahn et al., 2013, 2015, 2017) and HIE (Gonzales‐Portillo et al., 2014; Pimentel‐Coelho & Mendez‐Otero, 2010). In an animal model of HIE, intranasal MSC transplantation at 10 days after the initial HI insult was still able to reduce brain damage (Donega et al., 2013), suggesting that MSCs also stimulate endogenous brain repair in addition to being solely neuroprotective early following injury (Donega et al., 2014; van Velthoven et al., 2013).

Ischemia‐reperfusion and/or infection‐inflammation are important factors in etiology of white matter injury (Hagberg et al., 2002). (Oppliger et al., 2016) showed that intranasal delivery of human Wharton's jelly‐MSC in rats prevented hypomyelination and microgliosis in a model of white matter injury in the premature rat pups. Following this another study showed the effect of human Wharton's jelly‐derived MSCs on white matter injury caused by intraperitoneal lipopolysaccharide (LPS) injection (0.1 mg/kg) in P3 rat pups followed by ligation of the left carotid artery and 40 min of hypoxia (8% O2) at P4. These animals then received intracranial MSC treatment at P11 which led to an improvement in locomotor activity, reduced myelin loss and astrocyte activation (Mueller et al., 2017). Several other studies have demonstrated that MSCs migrate to the ischemic boundary zone, induce changes in the brain environment, and support neurogenesis (van Velthoven et al., 2010a, 2011). It is also known that administration of MSCs after HI markedly induced cell proliferation in the hippocampus and cortex, stimulate neuronal cell differentiation, and formation of new neuroblasts within the subventricular zone (Chen et al., 2003; Donega et al., 2013; van Velthoven et al., 2010a; Yoo et al., 2008).

White matter involvement is recently realized as a common finding in mitochondrial disorders. Studies have shown that mitochondrial disorders play a principal role in the pathogenesis of white and grey matter injury (Galluzzi et al., 2009; Hagberg, 2004; Hagberg et al., 2014; Koning et al., 2019). Following HI, a primary phase of energy failure occurs because of ATP depletion, failure of mitochondrial respiration, increased excitotoxicity and associated influx of Ca²⁺ and Na⁺. Thereafter, a phase of recovery ensues usually referred to as the latent phase. This is followed by a secondary phase of energy depletion and tertiary phases of injury where inflammation, protein misfolding, increased ROS production, and disruption of mitochondrial integrity/function leading to cell death (Fleiss & Gressens, 2012; Hagberg et al., 2015).

4. MSC MECHANISMS OF ACTION

4.1. MSCs—neuroprotection via mitochondrial transfer

MSCs inhibit apoptosis and rejuvenate damaged cells by transferring mitochondria during reoxygenation after oxygen glucose deprivation (OGD) in vitro (Liu et al., 2014). This transfer of mitochondria was achieved through tunneling nanotubes (TNTs). Another in vitro study shows that besides mitochondria, TNTs also mediate cell‐to‐cell transfer of plasma membrane components, cytosolic molecules, and other cell organelles (Pasquier et al., 2013) (Figure 2). TNTs are membranous channels with a diameter of 50–200 nm and mediate continuity between the plasma membranes of connected cells (Gerdes et al., 2007; Rustom et al., 2004). Delivery of functional mitochondria via TNTs aid in the recovery of PC12 cells during early stages of apoptosis (Wang & Gerdes, 2015). A logical next step would be the development of a selective interference for TNT formation in vivo to unravel their functional role. TNTs are stated to be a potential defense response to stress since it is only stressed cells that develop TNTs (Zhang, 2011). In addition to OGD, described above, Wang et al. have shown that TNTs can be induced in rat hippocampal astrocytes and neurons by H₂O₂ or serum depletion (Wang et al., 2011).

FIGURE 2.

Mitochondrial transfer from mesenchymal stem cells via tunneling nanotubes for attenuation of brain inflammation and injury

Although the mechanisms associated with TNT formation have not been well studied, motor proteins, like calcium‐sensitive dynamin‐related Rho‐GTPases Miro1 and Miro2 (Ahmad et al., 2014; Fransson et al., 2006), KLF 5 kinesin motor protein (Falnikar et al., 2011), and accessory proteins like TRAK 1 and TRAK 2 (Glater et al., 2006; Stowers et al., 2002), and Myo 19 and Myo 10 (Bohil et al., 2006) facilitate efficient shipping of cargo between cells via an actin–myosin‐dependent mechanism (Nawaz & Fatima, 2017). Over‐expression of Miro1 in MSCs led to enhanced mitochondrial transfer and rescue of epithelial injury, while Miro1 knockdown led to reduced mitochondrial transfer (Ahmad et al., 2014). An important result of mitochondria transfer from MSCs is the restoration of cell functions such as mitochondrial respiration in the recipient cells (Caicedo et al., 2015). Mitochondria also have an innate ability to find their ways into a host cell. Isolated mitochondria can be incorporated in vitro into cells by a simple co‐incubation, without the need for transfection reagents or medium supplements. The incorporated mitochondria can increase ATP production and oxygen consumption inside the recipient cells (Kesner et al., 2016).

Mitochondria were isolated and transferred to target cells in a rabbit model of regional cardiac ischemia (Masuzawa et al., 2013). An injection of autologous mitochondria (~10⁷ ) at the site of ischemia lead to their internalization within 8 hr of their administration and resulted in reduced apoptosis and infarct size, as detected 4 weeks later (Masuzawa et al., 2013). Chang et al have shown that allogeneic/xenogeneic injection of peptide‐labeled mitochondria into the medial forebrain bundle of adult rats restored mitochondrial functions and attenuated 6‐hydroxydopamine‐induced neurotoxicity in a model of Parkinson's disease (Chang et al., 2016). In a clinical study, it has been shown that healthy autologous mitochondria from non‐ischemic skeletal muscle in pediatric patients can be safely injected into damaged myocardium after ischemic injury for improvement of ventricular function (Emani et al., 2017). Similar studies have not been reported in neonatal models of brain injury or performed in patients with neonatal brain injury.

Besides improving the metabolic situation, transfer of healthy mitochondria to injured cells via TNTs can increase the turnover of damaged mitochondria by mitophagy (Li et al., 2017). Mitophagy is the selective autophagy of damaged mitochondria through the autophagosomal‐lysosomal pathway (Rodolfo et al., 2018; Thornton et al., 2018). Mitophagy plays a vital role in maintaining cellular homeostasis and defending against oxidative stress because this process controls mitochondrial quality and quantity by eliminating dysfunctional or damaged mitochondria (Hagberg et al., 2014). HI‐induced brain damage increases mitophagy and the inhibition of mitophagy aggravates brain damage in neonatal rats (Li et al., 2017). In stroke, most studies support the hypothesis that mitophagy favors neurons adapted to the stress by removing impaired mitochondria and suppressing cell death signaling cascades. In permanent MCAO, mitophagy was triggered at 1 hr and in the reperfusion phase (Zhang et al., 2013) (Zuo et al., 2014). Rapamycin treatment improves mitochondrial function after experimental ischemic stroke (Li et al., 2014). Infusion of MSCs restored PTEN‐induced kinase (Pink1)/ PTEN‐induced kinase 1 (Parkin)‐mediated mitophagy, improved mitochondrial dysfunction and attenuated apoptosis in endothelial cells in diabetic rats (Zhu et al., 2018).

4.2. MSCs—neuroprotection by modulation of immune‐inflammatory response

Inflammation is another important risk factor for injury in the developing brain (Bokobza et al., 2019; Hagberg et al., 2012, 2015). Intranasal administration of UC‐MSCs in postnatal day 10 Sprague‐Dawley rat pups after HI significantly reduced markers of neuroinflammation and significantly decreased the number of activated microglia, the brains innate immune cells (McDonald et al. 2019).

In clinical cases of IVH, microglial density, particularly IBA1/CD68 microglia, was higher within the affected brain regions when compared to healthy infants. Severe cases also had increased TNFa expression (Supramaniam et al., 2013). Following IVH, bleeding may extend into the intraventricular CSF, with cell‐free hemoglobin proposed as an inducer of inflammation, and hemoglobin metabolites inducing pro‐inflammatory cytokine expression (Gram et al., 2013) causing inflammation of the choroid plexus and ventricular ependymal lining (Simard et al., 2011). Inflammation and scarring may also prevent the flow of CSF reducing drainage (Strahle et al., 2012). Clinical studies have shown that the CSF inflammatory response to IVH is directly proportional to the extension of bleeding (Fam et al., 2017). Injury to the germinal matrix and HI, both possibly mediated by cytotoxins from blood products, are linked to the mechanisms of injury after IVH (du Plessis, 1998). Interestingly, brain‐derived neurotrophic factor (BDNF) secreted by MSCs, after contact with thrombin is an important factor in MSC therapy for IVH. MSCs with siRNA‐induced BDNF knockdown lose their therapeutic capacity in severe IVH‐induced brain injury in newborn rats (Ahn et al., 2017).

MSCs are capable of homing from the circulation to sites of injury and inflammation (Rustad & Gurtner, 2012). This can result in the resolution of the glial scar surrounding the lesion, decrease in reactive astrocytes and microglia and polarization of microglia toward the anti‐inflammatory phenotype (Donega et al., 2014). Lin et al. have shown that infusion of MSCs in a global cerebral ischemia mouse model increased release of TNF‐α‐induced protein (TSG‐6) thereby inhibiting local inflammation and improving function of the nervous system (Lin et al., 2013). Van Velthoven et al have demonstrated in the damaged area of the brain after MSC treatment, the up‐regulation of gene expression profiles associated with cell proliferation (Spp1 and IL17), neurogenesis (NRCAm and NGF), migration (CXCR4), and neuronal survival (glial‐derived neurotrophic factor) whereas genes involved in inflammation (e.g. IL1‐ β) were down‐regulated.

Since the immunomodulatory properties of MSCs were first reported by Bartholomew and colleagues (Bartholomew et al., 2002), subsequent studies have shown that MSCs mediate immunosuppression via up‐regulation of miR‐155 (Xu et al., 2013). MSCs have the capacity to interact with immune cells like B cells, T cells, dendritic cells (DCs), natural killer (NK) cells, neutrophil, and macrophages (Wang et al., 2014) and the MSCs secretome alters both adaptive and innate immune responses (Aslam et al., 2009; Bruno et al., 2015).

MSCs induce persistent peripheral T‐cell tolerance in a preterm sheep model and reduced invasion of T cells into the brain following global HI (Jellema et al., 2013). Besides T‐cell modulation, MSCs inhibit B cell proliferation, neutrophil and monocyte function, and NK toxicity (Duffy et al., 2011; Gerdoni et al., 2007; Jellema et al., 2013). Although these modulatory effects are only partially understood, direct cell‐to‐cell contact and soluble factors appear both to be relevant (Rahmat et al., 2013). MSCs recognize activated NK cells and inhibit their proliferation by downregulating their activation via IL‐2 or IL‐15 (Spaggiari & Moretta, 2012). MSCs inhibit CD14 + monocyte differentiation toward dendritic cells (DC) (Jiang et al., 2005). MSCs also suppress TNFα and IFN‐γ secretion by CD4 + T cells while promoting IL‐4 secretion (Aggarwal & Pittenger, 2005). Artificial transfer of isolated MSC‐derived mitochondria also affects inflammation via actions on adaptive immune cells. Specifically, transfer of mitochondria from MSCs to T cells increases the expression of mRNA transcripts involved in T‐cell activation and T regulatory cell differentiation including FOXP3, IL2RA, CTLA4, and TGFβ1, leading to an increase in a highly suppressive CD25 + FoxP3+ population (Court et al., 2020).

Besides the actions mentioned above, mitochondria also act as central hubs in the innate immune system by regulating RIG‐I‐like receptors (RLRs), NOD‐like receptors (NLRs), and Toll‐like receptors (TLR). The incorporation of mitochondria into this arm of immunity can result in increased cellular energy output and significant metabolic reprogramming (West et al., 2011). Microglia are the immune cells of the brain and our previous study demonstrated that their inflammatory activation is linked to changes in mitochondrial function (Nair et al., 2019). Specifically, a short exposure to a low dose of lipopolysaccharide (LPS) causes a transient increase in oxidative phosphorylation (OXPHOS) whereas prolonged exposure to high dose LPS causes suppression of OXPHOS and mitochondrial respiration. This suppression of OXPHOS forces a metabolic shift toward glycolysis (Nair et al., 2019), a phenomenon similar to the Warburg effect (Kelly & O'Neill, 2015). Although glycolysis results in less ATP production overall, the preferential use of glycolysis over OXPHOS enables immune cells to produce ATP at a faster rate (Marelli‐Berg et al., 2012) and to produce intermediates for cytokine production (Chang et al., 2013) (Figure 2).

MSCs alleviate inflammation and the anti‐inflammatory effects include a shift from pro‐inflammatory cytokines such as IL‐1β, IL‐6, TNFα, IFNγ, IL‐1Rα, and prostaglandin E2 to anti‐inflammatory cytokines like IL‐10 (Bartosh et al., 2010; Chaubey et al., 2018; Chou et al., 2016; Ding et al., 2017; Nemeth et al., 2009; Ortiz et al., 2007; Park et al., 2013) (Figure 2). IL‐10 inhibits LPS‐induced glucose uptake and glycolysis and promotes OXPHOS in macrophages. IL‐10 also eliminates dysfunctional mitochondria by promoting mitophagy (Ip et al., 2017) which may drive them toward an anti‐inflammatory phenotype.

MSCs impact the metabolic phenotype of primary macrophages in co‐cultures (Vasandan et al., 2016). However, we do not know if MSC‐mediated factors directly lead to a metabolic shift in brain cells such as microglia. Islam et al. demonstrated in a comprehensive study that transfer of mitochondria via gap junctional channels from MSCs to pulmonary alveolar epithelial cells in a murine model of LPS‐induced acute lung injury is protective (Islam et al., 2012). They demonstrated that the neuroprotective effect was because of transferred mitochondria by including studies using gap‐junction channel‐deficient MSCs. Gap‐junction channels mediate shorter range cell‐to‐cell interactions, whereas TNTs allow long‐range communication and mediate the exchange of larger organelles and vesicles (Gerdes et al., 2013; Okafo et al., 2017).

4.3. MSCs—neuroprotection by regulation of ros production

During HI, oxidative stress induced by ROS ‘bursts’ plays a vital role in mitochondrial dysfunction and subsequent cell death (Forman et al., 2004; Piantadosi & Zhang, 1996; Singh‐Mallah et al., 2019). Elevated levels of intracellular ROS caused by dysfunctional mitochondria attenuate global protein synthesis (Topf et al., 2018) and will indiscriminately attack phospholipids, proteins and DNA (Singh‐Mallah et al., 2019). Oxidative stress is an early feature after cerebral ischemia and experimental studies targeting the formation of free radicals demonstrate various degrees of protection after perinatal insults (Singh‐Mallah et al., 2019). The premature brain, in particular, is vulnerable to ROS‐induced damage because of the deficient antioxidant stores due of their insufficient ability to synthesize antioxidant enzymes at birth and impaired up‐regulation in response to oxidative stress (Lee & Davis, 2011; Singh‐Mallah et al., 2019).

The main species of ROS which include O^2•− and H₂O₂ is produced from the dismutation of O^2•− by MnSOD within the mitochondria (Singh‐Mallah et al., 2019). Ischemia also results in succinate accumulation in the immature brain (Koning et al., 2019) which is rapidly oxidized in the first minutes of reperfusion, generating large amounts of O^2•(Sahni et al., 2018)⁻. The underdeveloped antioxidant system in the immature brain limits the inactivation of H₂O₂, which can cross lipid membranes and cause tissue damage. Mitochondrial ROS production and inflammation are increased after neonatal brain injury associated with altered Krebs cycle and succinate accumulation in the mitochondria (Koning et al., 2019). Excessive inflammation causes mitochondrial fragmentation and a metabolic switch in microglia cells, which is accompanied by increased ROS production (Nair et al., 2019).

MSCs reduce increased oxidative stress levels in pathophysiological conditions. Paliwal et al. (2017) investigated abilities of tissue‐specific MSCs (bone marrow, adipose tissue, dental pulp and umbilical cord), from young or old donors (<30 versus > 30 years of age) to reduce ROS and thereby mitigate cellular damage. They found that both donor's age and source of tissue affected the ability of MSCs to reduce increased ROS levels in injured cells, with young donor supplied dental pulp MSCs having the greatest efficacy (Paliwal et al., 2017). The importance of donor age in stem cell treatment has also been investigated in the use of UCB derived from term or preterm sheep, administered in a sheep model of HIE. Although both cell sources were protective, UCB from term (but not preterm) reduced markers of circulating oxidative stress (Li et al., 2018).

ROS levels can be efficiently reduced by stimulating mitophagy by rapamycin (Bin‐Umer et al., 2014; Kurihara et al., 2012). As damaged mitochondria are a primary source of ROS, failure to remove dysfunctional mitochondria results in increase ROS production. MSCs can ameliorate ROS overproduction by restoration of autophagic flux (Zhu et al., 2018). These findings indicate that MSCs are important regulators of oxidative stress production, enhance cell migration, and support their role to improve neuroprotection and brain recovery following injury.

5. HOW TO BUILD A BETTER MSC

Targeting metabolism is an important way to improve MSC function. Studies have shown that in a nutrient‐rich culture medium MSCs rapidly proliferate, and reconfigure their central energy metabolism to become significantly more dependent on oxidative phosphorylation (OXPHOS) (Pattappa et al., 2011). Whereas in their native (lower nutrient) environment MSCs exist in a quiescent state characterized by low proliferation and high multi‐potentiality (Yuan et al., 2019). In this state they are primarily glycolytic, with mitochondria maintained by active biogenesis‐mitophagy cycling (Katajisto et al., 2015; Sbrana et al., 2016). Dependence on OXPHOS‐fueled metabolic profile (in a nutrient‐rich environment) results in accumulation of reactive oxygen species (ROS) (Shyh‐Chang et al., 2013), that reduce basal autophagy and mitophagy rates while simultaneously increasing the fraction of senescent cells that have reduced clinical potency (Ho et al., 2017). Another important feature of MSCs for cell therapy is their immunomodulatory secretome—cytokines and cellular components such as microRNA and extracellular vesicles that affect their repair and regenerative capacities.

In vivo studies show that MSCs can home toward the ischemic lesion site via interaction between stromal cell‐derived factor‐1α and its receptor CXCR4 present on MSCs (Wang et al., 2008). The Stromal cell‐derived factor‐1α expression is upregulated near the lesion site and expressed for at least 14 d after induction of HI brain injury (Rosenkranz et al., 2010). The interaction of locally produced SDF‐1alpha and CXCR4 expressed on the MSC plays an important role in the migration of transplanted MSCs, suggests that it might be a potential approach to modulate the expression of the two molecules in order to promote the therapeutic effects of MSCs (Wang et al., 2008).

The characteristics of the secretome is significantly influenced by metabolism (Yuan et al., 2019), which itself is linked to MSC phenotype. MSC phenotype, is the state of the cell as a result of the interaction between the paracrine factors with receptors such as Toll‐like receptors (TLR) (Delarosa et al., 2012; Sangiorgi & Panepucci, 2016). (Waterman et al., 2010) has described two different MSC phenotypes distinguished as proinflammatory MSC phenotype (MSC1; induced by TLR4 activation) and anti‐inflammatory MSC phenotype (MSC2; induced by TLR3 activation).

Donald et al. have shown that MSCs manage intracellular oxidative stress by targeting depolarized mitochondria to the plasma membrane via arrestin domain‐containing protein 1‐mediated microvesicles (Phinney et al., 2015). Subsequently, these microvesicles are engulfed and re‐utilized via a process involving mitochondrial fusion by macrophages which results in enhanced bioenergetics (Phinney et al., 2015). In this process MSCs simultaneously shed micro RNA‐containing exosomes that inhibit macrophage activation by suppressing TLR signaling, thereby de‐sensitizing macrophages to the ingested mitochondria (Phinney et al., 2015).

Methods to manipulate MSCs to reduce ROS during their culture expansion phase and in the injured tissue microenvironment may also promote MSC engraftment and enhance tissue repair (Denu & Hematti, 2016). One potential method to reduce oxidative stress in MSCs is by modulating sirtuin expression and/or activity. Multiple studies have shown a protective role for SIRT1 in various models of neuronal injury and neurodegeneration (Della‐Morte et al., 2009). It is shown that SIRT1 over‐expression quenches age‐related MSCs senescence by enhancing TPP1 expression, telomerase activity, and reducing DNA damage. ROS levels in stem cells increase with age (Ito et al., 2006). Majority of ROS is produced in the mitochondria and SIRT3 is the major protein deacetylase in the mitochondria, where its actions decrease ROS production. In addition, SIRT3 initiates metabolic reprogramming toward more efficient electron transport and fuel usage away from carbohydrate catabolism, reducing excessive ROS (Denu, 2017). Data suggest that overexpressing SIRT3 (major mitochondrial deacetylase involved in reducing oxidative stress while preserving oxidative metabolism) might represent a strategy to increase the quality and quantity of MSCs that could be utilized for clinical applications (Denu, 2017).

6. CHALLENGES IN MSC‐BASED THERAPY

Researchers have made significant progress in enhancing the therapeutic properties of MSCs. Transplantation time and injected cell dose are key factors that determine the therapeutic effect of stem cell therapy (Wang et al., 2017). Studies with incomparable or conflicting results prevent the advancement of MSCs for clinical use. These discrepancies are mostly because of variabilities in dosage, administration route, timing, and the limited ability of animal models to mimic the extremely complex process of human physiology and etiology of encephalopathy. Possible routes for MSCs infusion in animals are, (i) systemic delivery [intra‐venous (IV) and intra‐arterial (IA) , intra‐peritoneal (IP), and intracardiac (IC) (ii) local/topical/regional delivery (cell‐spray, gel or subcutaneous injection with a carrier hydrogel intramuscular, intranasal, and intrathecal injection) or (iii) scaffold/bioengineered construct (cells embedded in a scaffold, such as vascular grafts and intra‐osseous injection), with each route having its advantages and disadvantages for specific disorders. Although intracerebral administration ensures direct and targeted delivery and a minimum loss of stem cells, it is an invasive procedure, especially as the sites for delivery are often deep in the brain. However, these approaches should not be considered inappropriate to develop, as invasive neurosurgical techniques have been developed for the treatment of severe IVH and these have had astounding positive effects on long‐term outcomes; a 23‐point increase in childhood IQ (Luyt et al., 2019). To avoid intracerebral injections, multiple studies have looked at systemic MSC administration routes: either intravenous (IV) or intra‐arterial or intranasal applications (Li et al., 2018, McDonald et al. 2019, Danielyan et al., 2009, van Velthoven et al., 2010b). IV injections by retro‐orbital injections, although may seem aesthetically distasteful, can be performed rapidly without requiring mechanical restriction or heat‐induced vasodilation. This route is ultimately more humane than alternative intravascular injection techniques in small animals (Leon‐Rico et al., 2015; Ukai et al., 2007; Yardeni et al., 2011).

Therapeutic modalities must not only be efficacious but also safe for the patient. While the majority of preclinical and clinical studies have found MSCs to be safe and well‐tolerated (Lalu et al., 2012), uncontrolled stem cell interventions have resulted in adverse incidents (Bauer et al., 2018; Moll et al., 2019). Systemic injections may be practical but they can activate the host innate immune cascade systems, such as complement and coagulation, termed the instant blood mediated inflammatory reaction (IBMIR) (Leibacher & Henschler, 2016; Moll et al., 2012, 2019). A meta‐analysis of preclinical studies of MSCs in ischemic stroke models found that although direct injection of MSCs provided the greatest benefit, although other methods of delivery such as IA and IV injections also demonstrated significant improvement in outcomes (Vu et al., 2014). Therefore, the present data, based on some preclinical studies and limited clinical trials, is not adequate to make any sound conclusions as to whether one delivery method is superior to another.

In addition to routes of administration, the delivery of sufficient MSCs is required to discern a significant therapeutic effect. Neuroprotective therapy for HIE such as hypothermia, needs to be started within 6 hr to be effective (Davidson et al., 2015). However, the therapeutic window of MSC treatment is probably wider in neonatal HI mice as it has been shown that MSCs improved functional outcome and lesion volume when administered at least until 10 days after induction of the insult (Donega et al., 2013). A single dose of 0.5 × 10⁶ MSCs administered intranasally was identified as the minimally effective dose with a therapeutic window of at least 10 days but less than 17 days post‐HI was sufficient for a marked beneficial effect in neonatal mice (Donega et al., 2013). Velthoven et al have shown that intracranial administration of 100,000 MSCs at 3 and 10 days after neonatal HI markedly improves functional outcome, reduces infarct volume, and stimulates neurogenesis, with each injection having distinct effects on regenerative processes. Briefly, they find that a single injection of MSCs increased cell proliferation and differentiation of newly divided cells into NeuN‐ and olig2‐positive cells and reduced the percentage of BrdU+‐microglia, whereas a second additional injection with MSCs induced axonal outgrowth and remodeling, myelination and synaptogenesis which led to improved functional outcome (van Velthoven et al., 2010a).

7. CONCLUSION and FUTURE DIRECTIONS

MSCs secrete many paracrine factors which can impart pleiotropic effects in response to microenvironmental cues in the injured brain. MSCs attenuate pro‐inflammatory responses and the adverse effects of excessive ROS production. MSC mediated mitochondria transplantation could be a game‐changer for treating perinatal and neonatal brain injury. Future meticulous studies should be performed to delineate the mitochondria‐mediated protective mechanisms specifically for neonatal brain injury. For instance, the development of a selective inhibitor for TNT formation can help unravel its functional role. Another example is accelerated mitochondrial transfer by Miro1 over‐expression, as this may be a way to rejuvenate injured cells. Scientists are investigating the role of artificial transfer of isolated mitochondria as it can emerge as a regenerative or immunomodulatory therapy.

CONFLICT OF INTEREST

We report no conflict of interest.

Nair S, Rocha‐Ferreira E, Fleiss B, et al. Neuroprotection offered by mesenchymal stem cells in perinatal brain injury: Role of mitochondria, inflammation, and reactive oxygen species. J Neurochem.2021;158:59–73. 10.1111/jnc.15267