Versatile strategies for adult neurogenesis: avenues to repair the injured brain

Junyi Zhao ,Siyu Liu ,Xianyuan Xiang ,Xinzhou Zhu

Abstract Brain injuries due to trauma or stroke are major causes of adult death and disability.Unfortunately,few interventions are effective for post-injury repair of brain tissue.After a long debate on whether endogenous neurogenesis actually happens in the adult human brain,there is now substantial evidence to support its occurrence.Although neurogenesis is usually significantly stimulated by injury,the reparative potential of endogenous differentiation from neural stem/progenitor cells is usually insufficient.Alternatively,exogenous stem cell transplantation has shown promising results in animal models,but limitations such as poor long-term survival and inefficient neuronal differentiation make it still challenging for clinical use.Recently,a high focus was placed on glia-to-neuron conversion under single-factor regulation.Despite some inspiring results,the validity of this strategy is still controversial.In this review,we summarize historical findings and recent advances on neurogenesis strategies for neurorepair after brain injury.We also discuss their advantages and drawbacks,as to provide a comprehensive account of their potentials for further studies.

Key Words: adult neurogenesis;glia-to-neuron conversion;ischemic stroke;neurogenic niche;neuroinflammation;stem cell transplantation;traumatic brain injury

Introduction

Among all neurological disorders,acquired brain injury (ABI) possesses the most substantial threat to global health.ABI can be divided into two groups: traumatic and non-traumatic (Giustini et al.,2013).Traumatic brain injury (TBI) has the highest incidence and prevalence among all neurological disorders,affecting worldwide over 50 million people per year and leading to an annual global economic burden of around 400 billion US dollars (Maas et al.,2022).In turn,as the dominant type of non-TBI,stroke represents one of the leading causes of adult death and disability worldwide (GBD 2019 Stroke Collaborators,2021).The 2019 Global Disease Burden study reported that around 6.5 million deaths and 143 million disability-adjusted life years resulted from stroke in 2019 (GBD 2019 Stroke Collaborators,2021).

TBI is typically caused by blunt head trauma and leads to diffuse axonal injury (Blennow et al.,2016;Jang and Seo,2022).Some mild to moderate TBI patients may suffer from post-concussive syndrome in the chronic phase (Blennow et al.,2016).Unfortunately,most treatments for TBI patients are supportive,and limited success is often achieved by specific strategies,such as osmotic therapy or hypothermic treatment,used to reduce intracranial pressure and cerebral edema (Marehbian et al.,2017).Similar to TBI,targeted therapeutic methods for stroke patients are also extremely limited.Cardiovascular or cerebral ischemia account for nearly 90% of stroke cases (Hatleberg et al.,2019).Thrombolytic agents,including tissue plasminogen activator and surgical thrombectomy,are the only approved therapeutic approaches for ischemic stroke (Khatri,2022).However,they are effective within a narrow therapeutic window (up to～6 hours after stroke),and thereby many cases cannot benefit (Khatri,2022).Moreover,supportive treatments such as hyperosmolar,hypothermic,or thrombolytic therapies do not appear to directly contribute to neuroregeneration in the chronic phase of stroke (Bardutzky and Schwab,2007;Shafi and Kasner,2011;You et al.,2022).Thus,novel strategies to generate newborn neurons are eagerly sought to compensate for massive neuronal loss during ABI.

Patients who survive mild to moderate TBI or stroke may often recover in a couple of weeks (subacute to acute phase);then their recovery slows down,and is sometimes accompanied by diverse long-term deficits extending over months or even years (chronic phase) (Campbell and Khatri,2020;Maas et al.,2022).There are still many open questions about the cellular determinants of the recovery rate.Unlike glial cells,mature neurons cannot proliferate.Although neural plasticity minimizes to some extent the consequences of neuronal damage after ABI,considering that most such patients are adults a key question is whether it is possible to generate new neurons in the adult brain.

In this review,we aim to summarize the historical research and recent progress in three different aspects of adult neurogenesis: endogenous neurogenesis,exogenous stem cell transplantation,and direct glia-to-neuron conversion.Although the occurrence of adult neurogenesis in humans is still debated,many scientists and clinicians do believe that it is feasible to generate functional neurons after injury.We will thus comment on the advances in this field,highlight several limitations and controversial points,and discuss potential challenges of applying neurogenic approaches to brain injury therapy.

Literature Search

Peer-reviewed articles published until December 2022 were searched in PubMed database and Google Scholar.Keywords,including “adult neurogenesis”,“stem cell therapy/transplantation”,“mesenchymal stem/stromal cell”,“embryonic stem cell”,“induced pluripotent stem cell”,“neural stem cell”,“astrocyte to neuron”,” oligodendrocyte precursor cell to neuron” and “microglia to neuron” were used with or without combinations with “traumatic brain injury” or “ischemic stroke”.Search results were further filtered by titles and abstracts.

Endogenous Neurogenesis: Debate on the Basis of Neuroregeneration

In the early 20thcentury,the famous Spanish neuroscientist Ramon y Cajal claimed that neurogenesis occurs only during embryonic development (Ramon y Cajal and May,1928;Ming and Song,2005).For many decades it was therefore believed that new neurons could not emerge in the adult brain.Altman (1962) challenged this view by providing the first evidence that new neurons could be generated in the adult rodent brain.While subsequent evidence supported such possibility (Altman and Das,1965;Altman,1969),the corresponding autoradiographic and histological techniques were deemed controversial and for many years less attention was paid to this emerging field (Ming and Song,2005;Jurkowski et al.,2020).In the 1990s,when novel techniques such as 5-bromo-2′-deoxyuridine (BrdU) labeling and antibodybased biomarker detection methods were rapidly developing,accumulating evidence further supported the existence of neurogenesis in the adult rodent brain (Ming and Song,2005;Jurkowski et al.,2020).

The subventricular zone (SVZ) and the subgranular zone (SGZ) are two wellcharacterized neurogenic niches in the adult rodent brain (Jurkowski et al.,2020).However,whether neurogenesis occurs in adult humans is still controversial.Therefore,it is essential to consider the arguments favoring or refuting this important concept before addressing neurogenic strategies to improve ABI outcomes.

Classic neurogenic niche: SVZ-derived neurogenesis in olfactory bulb,striatum,and cortex

More than two decades ago,Alvarez-Buylla and colleagues,one of the pioneer groups in the field of adult neurogenesis,identified SVZ astrocytes as neural stem cells (NSCs) in the brains of adult rodents (Doetsch et al.,1999).NSCs in the SVZ migrate to the olfactory bulb via the rostral migratory stream (RMS) and generate neurons for odor discrimination throughout life (Gheusi et al.,2000).Later,researchers discovered that in both mice and rats,NSCs in the SVZ can rapidly respond to stroke-induced brain injury by migrating to the striatum or cortex and differentiating into neurons (Arvidsson et al.,2002;Parent et al.,2002;Yamashita et al.,2006).The inspiring results in animal studies convinced many scientists and clinicians about the feasibility of promoting post-stroke neurorepair in the striatum and cortex by stimulating endogenous neurogenesis in the SVZ (Lindvall and Kokaia,2015).

Compared to rodents,which have a highly developed olfactory bulb and rely mostly on smell for detection of food sources and social behaviors,humans have a less complex olfactory bulb and seem to lack a similar migration pathway for SVZ-derived NSCs.Moreover,it is still uncertain whether NSCs exist and exert functions in the SVZ of adult humans.In a 2011 report,Alvarez-Buylla’s group did not find evidence for migrating immature neurons in the adult human brain (Sanai et al.,2011).They observed instead that the SVZ and the RMS contained migrating immature neurons only in neonates <18 months old,with neurogenesis declining significantly in older children and becoming undetectable in adults (Sanai et al.,2011).In addition,they suggested a different RMS pattern in infants and young children,in which NSCs in the SVZ would migrate to the cerebral cortex and differentiate into neurons (Sanai et al.,2011).The differing patterns involved in SVZ-related neurogenesis thus seemed to reflect the inherent complexities of the olfactory bulb in rodents and the cortical architecture in humans.However,the evidence presented by Alvarez-Buylla’s group was exclusively based on immunofluorescent staining of neuroblast markers such as doublecortin (DCX) and polysialylated neural cell adhesion molecule (PSA-NCAM),and hence its validity was debated.Still,in 2014 a novel approach by Jonas Frisen’s group provided new evidence for the existence of neurogenesis in the adult human brain.Combining DCX and PSA-NCAM immunofluorescence,and 5-iodo-2′-deoxyuridine (IdU) and14C tracing,they reported the presence of adult-born neurons,probably derived from the migration of SVZ-NSCs,in the human striatum (Ernst et al.,2014).Notably,Ki67/DCX double staining had previously revealed that neurogenesis can occur in the ischemic penumbra surrounding cortical infarcts in stroke patients (Jin et al.,2006).Enticingly,besides the potential involvement of the SVZ-NSC population in this phenomenon,poststroke striatal or cortical neurogenesis might also result,as discussed further below,from direct glial-to-neuron conversion.

Classic neurogenic niche: SGZ-derived neurogenesis in the hippocampus

The hippocampus is the most important brain region for learning and memory.It is highly vulnerable to brain injury,and also affected in dementia,particularly Alzheimer’s disease.NSCs in the SGZ of the dentate gyrus may replenish hippocampal neurons for integration into existing neural circuits.In the 1990s,Fred Gage´s group reported age-dependent hippocampal neurogenesis in rats and mice (Kuhn et al.,1996;Kempermann et al.,1997).The neurogenic activity in rodents could be enhanced by an enriched environment and by running (van Praag et al.,1999).Whereas several studies later confirmed adult hippocampal neurogenesis in rodents (Sahay et al.,2011;Li et al.,2022),this phenomenon had been found to be strongly stimulated in monkeys following ischemic injury (Yamashima et al.,2004).

Based on the evolutionary anatomical and functional conservation of the hippocampus in mammals,an early study using BrdU tracing in humans provided,almost at the same time as in rodents,evidence for hippocampal neurogenesis in the adult human brain (Eriksson et al.,1998).This was one of the few published studies describing human neurogenesis through BrdU labeling,as this diagnostic method was later banned for clinical use.Reproducing the14C dating approach used to discover neurogenesis in the striatum,persistent adult neurogenesis in the hippocampus was later on reported by the Frisen´s group (Spalding et al.,2013).Shortly thereafter,Ki67 and DCX immunostaining provided evidence that neurogenesis occurs in both the SVZ and the SGZ,even in aged individuals (Dennis et al.,2016).However,these findings were again challenged in 2018,when Alvarez-Buylla’s group claimed,based on immunohistochemical analysis of a panel of neuroblast markers in brain samples from infants to aged individuals,that hippocampal neurogenesis was undetectable in the adult hippocampus of either humans or monkeys (Sorrells et al.,2018).Partially agreeing with Sorrells et al.article,another study simultaneously suggested that the SGZ represents a minor neurogenic niche in humans,while the subiculum may instead be a more important hippocampal neurogenic niche (Nogueira et al.,2018).Also in 2018,a report analyzing human brain samples at different ages with a strict stereological method pointed out the drawbacks of Sorrells et al.study,questioning the quality of brain samples and the lack of stereological analysis (Boldrini et al.,2018).Upon extensive immunofluorescent staining of neurogenic markers,the referred study concluded that neurogenesis persists in the adult hippocampus even until old age (Boldrini et al.,2018).Shortly after these findings,a consensus review was published which summarized the available evidence and commented on the limitations of current techniques to analyze adult neurogenesis in human samples;discussion topics included sample quality,differences between humans and rodents affecting selection of markers of neuronal maturation,and the importance of stereological analysis (Kempermann et al.,2018).Even though more sophisticated techniques are being deployed,the debate is,as of today,still ongoing.As RNA sequencing allows to observe transcriptome changes at single-cell resolution,this technique may overcome the shortcomings of traditional immunofluorescent staining of neurogenic markers when analyzing adult neurogenesis.Very recently,a comprehensive single-cell RNA sequencing analysis concluded that adult hippocampal neurogenesis occurs in mice,pigs,and monkeys,but not in humans (Franjic et al.,2022).But almost simultaneously,this view was again challenged by Ming and Song´s group,who performed single-cell and singlenucleus RNA sequencing and identified typical neurogenic genes in adult brain samples (Zhou et al.,2022).Hence,more studies with advanced techniques and high-quality brain samples are required to settle this dispute.

Novel neurogenic niches

Adult neurogenesis in rodents may not be restricted to the SVZ and SGZ,as it was also reported in other brain regions such as the hypothalamus,substantia nigra,and amygdala (Lie et al.,2002;Zhao et al.,2003;Paul et al.,2017;Jurkowski et al.,2020).Compared to the SVZ and SGZ,neurogenesis in these brain regions seems to show an age-dependent decline.As human research intensifies,it will be interesting to explore whether these novel potential neurogenic niches may have therapeutic relevance in post-injury recovery.

To sum up,adult neurogenesis in classic neurogenic niches is well-studied in rodents,but remains controversial in humans (Table 1).As detection techniques evolved from BrdU labeling to immunofluorescent staining of neurogenic markers,further aided by stereological methods,and were greatly expanded by single-cell transcriptomics analysis,there is a now great deal of information to be obtained about neurogenesis in healthy adults,and also in TBI or stroke patients,whenever high quality brain samples are available.The cortex,striatum,and hippocampus are key affected brain regions in many TBI and stroke cases.If adult human neurogenesis is confirmed in these regions,development of pharmaceutical strategies to enhance endogenous neurogenesis may be highly valuable to compensate for neuronal loss.However,since endogenous neurogenesis appears to be,at best,extremely limited in adult humans,relying on this mechanism might be not sufficient to tangibly improve patient outcomes.Therefore,supplementary strategies exploiting exogenous neurogenic sources need to be explored to restore neurological function in ABI cases.

Table 1｜Key studies on adult human neurogenesis in SVZ and SGZ

Exogenous Stem Cell Transplantation: Safe But Ineffective?

Since endogenous neurogenesis may be insufficient to repair brain damage,introducing exogenous stem cells arises as a promising complementary or alternative strategy.In recent years,stem cell transplantation has been widely explored in animal models to treat brain and spinal cord injuries,as well as neurodegenerative diseases (Politis and Lindvall,2012;Bjorklund et al.,2021).Although these studies demonstrated the strong neuroprotective and neurorepair capacity of different stem cell therapies in animal models,and prompted also several clinical trials,it remains unclear whether these strategies will benefit brain injury patients.

Source of stem cells

Common sources of stem cells include mesenchymal stem/stromal cells (MSCs),embryonic stem cells (ESCs),induced pluripotent stem cells (iPSCs),and NSCs,all of which can differentiate into mature neurons at leastin vitro(Politis and Lindvall,2012).

MSCs are a group of heterogenous multipotent precursor cells present in diverse adult tissues such as bone marrow,adipose tissue,kidney,liver,and pancreas (Nombela-Arrieta et al.,2011).Since these human cells are easily available,MSCs have been widely used in both animal experiments and clinical trials for brain injury,neurodegenerative diseases,cancer,and other major diseases.In the treatment of neuropathological conditions,the underlying principle is that MSCs would differentiate into neurons when transplanted into brain tissue.However,since multipotent MSCs are usually isolated from nonneural tissues,their capacity for neuronal differentiation is usually limited.In a comparative study,human adipose tissue-derived MSCs showed higher expression levels of neural markers and higher proliferation rates than MSCs isolated from bone marrow,skin,or umbilical cord;nevertheless,all of these MSCs exhibited a significantly lower neurogenic potential compared to NSCs (Urrutia et al.,2019).Although NSCs are the most reliable neurogenic stem cell type,primary human NSCs are extremely difficult to obtain,and genome instability remains a great clinical concern in relation to potential tumorigenic risk (Harrison,2012).

Compared with multipotent stem cells,pluripotent stem cells have a much broader capacity for differentiation and are more likely to generate functional neurons.Before iPSCs technology was developed,human ESCs were the main source of pluripotent stem cells used in stem cell therapy (Kimbrel and Lanza,2020).Over three decades ago,ESC-containing ventral mesencephalic tissues from aborted human fetuses were engrafted into the striata of two patients with severe Parkinson’s disease,leading to modest improvements in motor function (Lindvall et al.,1989).However,as human ESCs are mainly isolated from aborted embryos,ethical concerns and availability issues largely restricted their clinical applicability.The method to produce iPSCs from human skin fibroblasts,introduced in 2006 by Takahashi and Yamanaka,significantly increases the availability of pluripotent stem cells,overcomes ethical issues,and reduces transplant rejection.However,since genome instability in iPSCs remains a challenge for clinical therapy (Kimbrel and Lanza,2015),the technology has to be optimized to produce next-generation stem cells with the potential to differentiate into defined neuronal subtypes with high genomic stability (Kimbrel and Lanza,2020).The characteristics of different stem cells commonly used for stem cell therapy are summarized inTable 2.

Table 2｜Comparison of four types of human stem cells

Translational roadblocks: the gap between animal studies and clinical trials

Up to date,numerous efforts have been made to examine the effects of stem cell transplantation on brain injury models in rodents.Over two decades ago,intravenous administration of rat bone marrow-derived MSCs was used to treat cerebral ischemic injury in rats (Chen et al.,2001).To better adapt to potential clinical complications,later studies introduced non-human primate or even human stem cells into rodent models.For instance,transplantation of monkey-derived ECSs treated with retinoic acid to induce neural cells significantly improved motor functions in mice with experimental brain injury (Ikeda et al.,2005).Shortly thereafter,in a mouse model of stroke,monkey ESC-induced neuronal progenitors were reported to survive and expand until 4 weeks after transplantation (Hayashi et al.,2006).Likewise,it had been previously shown that human fetal NSCs differentiated into neurons and proliferated for at least 4 weeks after being transplanted to rats with ischemic stroke (Kelly et al.,2004).A more recent study applying the iPSC technique reported in turn that implantation of human fibroblast-derived iPSCs promoted stroke recovery in rat models (Oki et al.,2012).Ever since,numerous publications have demonstrated that stem cell therapies can efficiently improve neurological functions after experimental TBI or stroke in rodents (Li et al.,2021;Zhang et al.,2022).

These promising results accelerated the application of stem cell therapies in clinical trials of TBI,stroke,and other brain injuries.In a clinical trial involving 97 treated and 69 control TBI patients,no evident adverse effects and improved motor functions were reported after transplantation of autologous bone marrow-derived MSCs into the subarachnoid space by lumbar puncture (Tian et al.,2013).By means of a similar procedure,therapies employing umbilical cord-derived MSCs (Wang et al.,2013) and fetal nervous and hematopoietic cells (Seledtsov et al.,2005) also exhibited safety but somewhat limited efficacy in TBI recovery.

For stroke therapy,modified bone marrow-derived MSCs (SB623 cells),were considered to be safe and provide significant clinical improvement in a Phase I/IIa study that included 18 patients (Steinberg et al.,2016).A randomized,controlled Phase II study with 65 stem cell-treated and 58 placebo-treated patients (MultiStem Administration for Stroke Treatment and Enhanced Recovery Study,MASTERS) demonstrated that intravenous injection of multipotent adult progenitor cells passed the safety evaluation,but the primary endpoint of neurological recovery was not reached (Hess et al.,2017).In addition,a multicenter single-arm study with 23 stroke patients concluded that intracerebral implantation of fetal NSCs is feasible and may modestly improve upper limb function,with improvement in one patient after 3 months of treatment and in three patients after 6-12 months of treatment (Muir et al.,2020).

Despite wide inter-trial variability (i.e.cumulative transplanted cell doses varying from 1 x 106to 1 x 109cells),the safety and feasibility of stem cell transplantation has been confirmed in almost all clinical trials,with acceptable adverse effects and a low chance of tumorigenesis (Kawabori et al.,2020;Bonilla and Zurita,2021).However,it must be noted that almost all published clinical studies were Phase I/II trials with fewer than 100 patients in the treated group.Despite some hints of benefits for certain indications,the statistical power is inadequate to conclude significant improvement in neurological recovery after stem cell therapy,and the primary endpoint mainly shows little difference between treatment and control groups (Table 3).

Table 3｜ Key clinical trials of stem cell therapy in traumatic brain injury and ischemic stroke

Sidestepping stem cells’ short-term survival: exosomes as an alternative?

Aside from safety and feasibility issues,a major concern in stem cell therapy is whether the engrafted stem cells can survive,proliferate,and differentiate into neurons over a reasonable period of time as to repair the damaged brain.Some researchers argued that the long-term survival rate of exogenous stem cells was too low to generate a sufficient amount of neurons,and the observed therapeutic effects might rather be due to factors secreted by the transplanted cells (Hoang et al.,2022).Although it is still difficult to trace the destiny of transplanted stem cells in human studies,animal models provide indeed some evidence that the long-term effects of stem cells may be due to paracrine actions.This suggested,in turn,that therapeutic efficacy might be achieved by delivering,instead of whole stem cells,the corresponding molecular mediators by directly using stem cell-derived exosomes (Zhang et al.,2019;Zhou et al.,2019).

Exosomes are endosome-derived,membrane-bound vesicles that can be released by diverse cell types (Zhang et al.,2019).Exosomes from MSCs,NSCs,ESCs,or iPSCs contain abundant proteins and various types of RNAs and DNAs which may potentially facilitate the process of neurogenesis (Parfejevs et al.,2020).Meta-analysis was performed in a recent literature review to interpret the results of selected studies,performed in the last decade,on the efficacy of exosomal therapy in rodent stroke models (Cherian et al.,2023).After the delivery of MSC-,NSC-,or ESC-derived exosomes,significantly reduced infarction volume and improved neurological outcomes were observed in treated animals (Cherian et al.,2023;Liu et al.,2023).Although many animal studies have revealed the safety and efficacy of exosomal therapy,up to date there is only one ongoing clinical trial (NCT03384433) involving exosomal treatment for stroke patients,whose results have not yet been released.Accumulating evidence from animal studies will further promote randomized clinical trials to establish proofs of safety and efficacy of exosomal therapy as an alternative method of stem cell therapy for brain injuries.

In summary,although the safety of stem cell therapy has been extensively examined in animal models and clinical trials,the quality of patient data is still too weak to support its effectiveness in TBI and stroke patients.Clinical trials with iPSC-based cell therapy are still ongoing and may hopefully provide enticing results.However,it should be noted that quite diverse mechanisms may mediate the effects observed upon stem cell transplantation.As suggested above,the transplanted stem cells may be therapeutically active by secreting factors and vesicles to promote neuroprotection and endogenous neurogenesis,rather than differentiating into neurons for authentic stem cell-derived neurogenesis.Therefore,the efficacy of exogenous stem cell engraftment in brain injury should always be strictly evaluated longitudinally in clinical trials.Further attempts to enhance long-term survival of engrafted stem cells and promote their neuronal differentiation by modulating the microenvironment could be also valuable to optimize stem cell therapy for brain injury.

Direct Glia-to-Neuron Conversion: A New Era of Post-Injury Neurogenesis?

Mammals have a higher proportion of glial cells compared to other taxa.In the brains of invertebrate animal models likeCaenorhabditis elegansorDrosophila melanogaster,the proportion of glial cells is less than 25%.In comparison,around 65% of cells in the mouse brain are glia (Allen and Barres,2009).Early studies had suggested that the human brain contains about one trillion glial cells,with a glia:neuron ratio of 10:1 (Allen and Barres,2009;von Bartheld et al.,2016).However,more recent evidence showed that a 1:1 ratio of glia to neurons more accurately describes the adult human brain,with the estimation of glial cell number declining to 40-130 million (von Bartheld et al.,2016).Interestingly,in the adult brain (once mature neurons have lost their proliferative capacity),glial cells,including astrocytes,microglia,and oligodendrocyte precursors (OPCs),can constantly produce daughter cells.If the conversion from glia to neuron does actually occur,glial cells may represent an endogenous,virtually unlimited pool to replenish lost neurons after brain injury.With a focus on potential therapeutic applications in ABI,we summarize below recent advances in this “glia-to-neuron” cell trade.

Astrocyte-to-neuron reprogramming

Although both astrocytes and neurons initially originate from NSCs,those two cell types ultimately arise from specific,fate restricted (neuronal and glial) progenitors that exhibit distinct differentiation paths (Tang et al.,2017).Hence,the discovery of a direct conversion path from astrocytes to neurons by a simple but feasible reprogramming method is a remarkable breakthrough.

After injury,astrocytes become highly activated and display NSC properties (Escartin et al.,2021).The reactive astrocytes soon form what is known as a glial “scar” surrounding the injured area,which further induces neuroinflammation,secrets neuroinhibitory factors,and prevents further neuroregeneration (Adams and Gallo,2018;Bradbury and Burnside,2019).An early report in 2008 revealed that reactive astrocytes can represent an important source of multipotent cells in injured brains,but the specific mechanisms remained uncertain (Buffo et al.,2008).In the past decade,a burst of data has emerged in this field,as several research groups independently reported the potential to reprogram astrocytes into neurons by modulating a single factor or signaling pathway.Adding to evidence that transplanted astrocytes produce neuronsin vivo(Torper et al.,2013),Gotz’s group revealed that signaling though the sonic hedgehog pathway is a key mechanism providing NSC properties to astrocytes (Sirko et al.,2013).Simultaneously,Frisen’s group demonstrated that blocking Notch-1 signaling was also sufficient to drive a neurogenic program in cortical and striatal astrocytes after stroke (Magnusson et al.,2014).In the same year,another study reported that Sox2,a key transcription factor responsible for maintaining the pluripotency of stem cells,can also elicit a neurogenic program in astrocytes after spinal cord injury in adult animals (Su et al.,2014).

Notably,these studies indicated that astrocytes need to acquire NSC characteristics before entering a neurogenic program,as the target modulatory pathways and factors are all involved in stem cell pluripotency and multipotency.However,Chen and colleagues proposed that astrocytes can also be converted into neurons,without first dedifferentiating into NSCs or progenitor cells,through modulation of a single factor.They reported that retrovirus-mediated expression of NeuroD1,a transcription factor that promotes neuronal differentiation,can directly and efficiently convert reactive cortical mouse astrocytes into glutamatergic excitatory neurons after brain stab injury or modeling of Alzheimer’s disease (Guo et al.,2014).Although this finding was later on partially reproduced by another group,it was also challenged with issues such as conversion efficiency and region specificity (Brulet et al.,2017).Very recently,Zhang and colleagues published a study that strongly disputed the above conclusions.By using stringent lineagetracing techniques,they argued that the newborn neurons in question were actually derived from endogenous NSCs instead of astrocytes (Wang et al.,2021).In their opinion,Chen’s conclusion was mistaken mainly because a high adeno-associated virus titration was used,leading to leakage during neuronal tracing (Wang et al.,2021).As a rebuttal,recent research from Chen’s group using modified lineage tracing methods provided new evidence in support of their previous findings (Xiang et al.,2021).Still,in addition to fluorescent reporting in transgenic mice and virus-based tracing,improved detection techniques are required to clarify the origins of newborn neurons in studies examining glia-to-neuron conversion.

Indeed,a population of NSCs that express astrocyte markers such as nestin and glial fibrillary acidic protein (GFAP) has been identified in the SVZ of the adult mammalian brain.As these NSCs were shown to give raise,at least in mice,to reactive cortical astrocytes after stroke (Faiz et al.,2015),they are perhaps the most likely source of astrocyte-derived neurons in the adult brain.Although it is still difficult to distinguish in a definitive manner the specific NSCs that give rise to astrocytes (Escartin et al.,2021),there is much hope that targeted stimulation of specific astrocyte progenitors might result in more efficient neuronal repair.However,since all the studies so far conducted rely on modulation of transcription factors,the difficulties inherent to these approaches greatly challenge their clinical application.

OPC-to-neuron reprogramming

Mature myelinated oligodendrocytes are essential in safeguarding neural connectivity by providing myelin sheaths for large neuronal axons.However,their proliferative potential is thought to be extremely limited (Zhou et al.,2021).In contrast,a small pool of NG2-expressing OPCs,widely present in grey and white matter areas of the central nervous system (CNS),can readily proliferate and differentiate into mature oligodendrocytes (Zhou et al.,2021).Despite possessing stem cell properties,whether OPCs can also generate astrocytes and neurons remains controversial.In 2008,Rivers et al.(2008) reported that NG2+glia can mainly produce oligodendrocytes,and to a much fewer extent neurons,in the forebrain,but not astrocytes either in gray or white matter.However,in the same year Nishiyama’s group concluded that NG2+OPCs possess the capacity to simultaneously generate oligodendrocytes and astrocytes in gray matter (Zhu et al.,2008).Consistent with the Chen’s group’s report on direct astrocyte-to-neuron conversion,which had supported the notion that NG2+glial cells can differentiate into neurons (Guo et al.,2014),Nishiyama’s group recently identified distal-less homeobox 2 (Dlx2) as a key transcription factor driving OPCs conversion into GABAergic inhibitory neurons (Boshans et al.,2021).The latter report was challenged in turn by a previous study that suggested that NG2+glia are restricted to the oligodendrocyte lineage and cannot produce neurons after birth (Huang et al.,2019).Despite some pieces of evidence on the contrary,NG2+glial cells are widely believed to be restricted to the oligodendrocyte lineage.However,as NG2+OPCs remain to be well defined and characterized,their capacity for generating astrocytes and neurons is still uncertain.More solid evidence is thus required to convince scientists and clinicians of the feasibility ofin vivoreprogramming of NG2+glia.

Microglia-to-neuron reprogramming

Microglia are specialized CNS-resident macrophages that derive from primitive myeloid precursor cells in the embryonic yolk sac.In most studies,researchers focused on the phagocytic functions of microglia (Saijo and Glass,2011),like clearing debris from damaged neurons,and on their dual roles in post-stroke neuroinflammation as regulators of the pro-and antiinflammatory status (Hu et al.,2015).

Surprisingly,in 2019 the Nakashima’s group observed that NeuroD1-transduced microglia rearrange their transcriptional and epigenetic profiles and generate functional neurons in the mouse striatum (Matsuda et al.,2019).NeuroD1 has hence become a “star molecule” for its potential in converting two distinct glial groups into neurons;thus,maneuvers aimed at regulating its expression are believed to be worthwhile for treating neurodegenerative diseases (Trudler and Lipton,2019).In a following study employing a mouse stroke model,Nakashima’s group reported that microglia/macrophages in the striatum can convert into functional neurons that integrate into existing neural circuits (Irie et al.,2021).However,at the same time,Peng and colleagues challenged this mechanism by arguing an inappropriate use of lineage tracing strategy (Rao et al.,2021).Thus,interstudy variation in tracing techniques appear to be a critical issue determining opposite conclusions.

Due to the clearly distinct cell origins,the concept of microglia-to-neuron conversion is highly controversial,and definitive conclusions cannot be made until more convincing experimental data are published.Furthermore,as microglia are the key CNS cell population regulating neuroinflammatory responses,reprogramming of activated microglia after brain injury may introduce additional risk for patients.

The key findings of research attempts at direct glia-to-neuron reprogramming are summarized inTable 4.An interesting point is that in all types of CNS injuries,including stroke,TBI,and spinal cord injury,a glial scar will rapidly form and act as a barrier to initially prevent injury expansion.However,in the chronic recovery phase,the glial scar can inhibit neuroregeneration and induce neuroinflammatory responses (Wang et al.,2018).The glial scar is predominantly composed of reactive astrocytes,in addition to NG2+glia and microglia,which become activated and rapidly proliferate upon injury (Wang et al.,2018).As discussed above,all these three glia types have been reported to potentially possess the capacity of direct reprogramming into neurons.This suggests an enticing strategy to treat brain injury: repressing glial activation to limit the formation of the glial scar and attenuate excessive inflammation,while exploiting at the same time the proliferating capacity of glial cells to reprogram them into functional neurons.Nevertheless,it must again be noted that the different direct glia-to-neuron conversion methods so far used have been met with inconsistent results.Strict lineage tracing,including viral neuronal tracing,in combination with sequencing techniques in transgenic fluorescent reporter mice are required to ensure the authenticity of gliato-neuron reprogramming.Although glia-to-neuron conversion techniques have been tested with promising results in animal models,evaluation of such approaches in clinical trials is so far precluded by outstanding difficulties in modulating transcription factor expression in the human brain.

Table 4｜Key studies on glia-to-neuron conversion

Conclusion and Outlook

Neuronal damage characteristically occurs immediately after ABI,and the injury will be further exacerbated in the chronic stage to cause progressive neuronal death.To minimize or prevent this outcome,it is critical to confirm the existence of adult neurogenesis.However,considering that neurogenesis in the adult brain seems to be very limited,this mechanism may be insufficient even if it is stimulated rapidly after injury.Alternatively,exogenous stem cell transplantation has shown safety and feasibility in clinical trials;however,its efficacy remains to be validated in Phase III studies.Likewise,advances in glia-to-neuron conversion have opened a novel avenue for neurogenesis after brain injury,but its reliability is still uncertain.Current knowledge on the bases of adult neurogenesis and its potential for neurorepair via stem cell transplantation and glia-to-neuron conversion are summarized inFigure 1.

Figure 1｜Versatile strategies for post-injury neurogenesis.

It must be noted that the concept of neurogenesis,which is indeed relevant in both health and disease,differs from that of neuroregeneration,which was only marginally discussed in this review.Neuroregeneration entails axonal repairing,elongation,and remyelination,aimed at the reconstruction of functional neuronal networks.Nevertheless,since in therapeutic terms a clear overlap exists between both processes,additional research evaluating the impact of simultaneous modulation of potential endogenous and exogenous neuronal sources might greatly improve the therapy of brain injuries.

Continuous improvement of research techniques is steadily broadening our knowledge in the field of adult neurogenesis.By complementing traditional BrdU labeling and immunofluorescent staining approaches,state-of-theart techniques such as single-cell sequencing should provide more subtle evidence of adult neurogenesis from a systematic,molecular perspective.Next-generation stem cells with targeted gene editing may achieve longer survival and enhanced neurogenic capacity.Sophisticated tracking techniques with upgraded viral tracing and two-photon microscopy of lineage-specific transgenic mice will assist in revealing the authenticity of glia-to-neuron conversion.In sum,although substantial barriers remain,there is no doubt that an exciting new era has been ushered to exploit targeted neurogenesis therapies for brain repair.

Acknowledgments:We sincerely thank Professor Dr.Helmut Kettenmann (Max-Delbrück Center,Berlin,Germany,and Shenzhen Institute of Advanced Technology,Chinese Academy of Science) for useful suggestions on our manuscript.

Author contributions:Conceptualization: XZ and XX;manuscript writing: XZ;literature search: JZ and SL.All authors approved the final version of the manuscript.

Conflicts of interest:The authors declare no conflicts of interest in relation to this work.

Data availability statement:Not applicable.

Open access statement:This is an open access journal,and articles are distributed under the terms of the Creative Commons AttributionNonCommercial-ShareAlike 4.0 License,which allows others to remix,tweak,and build upon the work non-commercially,as long as appropriate credit is given and the new creations are licensed under the identical terms.