There is growing evidence for an age-dependent decline in axon growth, and regeneration potential, across a variety of model organisms. In aging zebrafish, axon regeneration has been shown to occur at a reduced speed and with increased latency (Graciarena et al., 2014). A similar decline in axon regeneration efficiency has been observed in C. elegans (Zou et al., 2013; Hammarlund and Jin, 2014) with both models putatively linked to altered neuron-intrinsic mechanisms.

In mammalian models, regrowth of aged peripheral axons is delayed, slower and less effective than that in younger animals (Verdú et al., 1995; Kerezoudi and Thomas, 1999; Kang and Lichtman, 2013). Pharmaceutical denervation also failed to elicit any growth response in aged (28 month old) rats (Pestronk et al., 1980). While the exact mechanisms and etiology of the decline of PNS regeneration with are unclear (Willcox and Scott, 2004), both neuron-intrinsic or extrinsic mechanisms seem to be at play. The processes of myelin clearance is also delayed in aging and is associated with decreases in fibers in the affected nerves (Vaughan, 1992; Kang and Lichtman, 2013). Adult DRG neurons in vitro present approximately 30% slower growth than their neonate counterparts (Lamoureux et al., 2010). The axonal atrophy observed in aged nerve fibers may be attributable to the reduced expression and transport of cytoskeletal proteins (Verdú et al., 2000), reduction in the rate of axonal transport (Stromska and Ochs, 1982; Kerezoudi and Thomas, 1999) as well as the decreased expression of nerve growth factor receptors (Parhad et al., 1995). Peripheral neuropathies resulting from these axonal changes with age are common in elderly populations (Cho et al., 2006). The age-related changes and decline are ambiguous, and do not progress linearly with age, exhibiting variation between studies (Verdú et al., 2000).

The relationship between age and axon regeneration in the CNS has received much less attention due to its already limited natural ability of CNS axons to regenerate. There is growing evidence for the same age-dependent decline that is seen in the PNS. Developmental studies have shown that changes in both the neuron-extrinsic environment of the spinal cord and intrinsic changes can reduce regeneration with age (Blackmore and Letourneau, 2006). In mammalian models of SCI, aging reduces locomotor recovery (Gwak et al., 2004) and is linked to changes in inflammation (von Leden et al., 2017) and myelination (Siegenthaler et al., 2008). Additionally, aging has varied effects on axon growth depending on the axon tract examined, with reduced rostral sprouting in the majority of major tracts at the lesion site (Jaerve et al., 2011). The neuronal deletion of phosphatase and tensin homolog (PTEN), a negative regulator of mammalian target of rapamycin (mTOR), has emerged as an effective target to promote the regeneration of the cortical spinal tract (CST) axons after an injury in young animals (Sun et al., 2011; Geoffroy et al., 2015). Recently, an age-dependent decline in mammalian CNS axon regeneration has been documented using PTEN-deletion strategies (Du et al., 2015; Geoffroy et al., 2016).

The regeneration, repair and regrowth of damaged axons is a complex process that relies on both internal mechanisms and responses to external signals. A balance of intrinsic and extrinsic cues determines the innate ability of injured CNS neurons to promote axon growth, and to elongate over distance. Both neuron-intrinsic and extrinsic mechanisms are altered in the normal aging process, resulting in a decline in the growth and regenerative potential of axons. There are a myriad of neuronal-intrinsic factors that interact and can impact axon growth after damage, including different pathways (PTEN/mTOR pathway, Park et al., 2008; Sun et al., 2011; Geoffroy et al., 2015), Suppressor of Cytokine Signaling 3 (SOCS3)/Signal Transducer and Activator of Transcription 3 (STAT3), Wnt/Ryk, insulin-like growth factor (IGF)-1, tumor suppressor p53, and Krüppel-like factors (KLFs), mitochondrial function, neuronal viscosity and axonal transport. These have been addressed in depth in our previous review focusing on the changes of neuron-intrinsic mechanisms with age (Geoffroy et al., 2017). The current review will only concentrate on the neuron-extrinsic factors influencing axon regeneration.

It is widely accepted that the pathophysiology of SCI consists of two distinct components (Kwon et al., 2004; Donnelly and Popovich, 2008). The primary injury results in the disruption of axons, the surrounding glial cells and blood vessels. The secondary injury is delayed and manifests in a broad spectrum of pathologies that exacerbate the injury and can continue for years after the primary trauma (Profyris et al., 2004; Rowland et al., 2008).

A range of cell signals and processes, which are intrinsically linked, influence the lesion microenvironment and the ongoing injury response (Figure 1). These include- microvascular disruption, hemorrhage, and disruption of cellular membranes resulting in the release of cytotoxic molecules, and excess ions and glutamate which causes damage to the surrounding cells by excitotoxicity (Brown et al., 2005), oxidative stress (Potts et al., 2006), lipid peroxidation of membranes (Shetye et al., 2014), disruption of intercellular ion balance (Potts et al., 2006), and induction of cell death signals. A hallmark of the secondary injury phase is a strong and persistent pro-inflammatory cascade, characterized by neutrophil and macrophage infiltration, microglial activation and astrogliosis (Profyris et al., 2004; Potts et al., 2006; Rowland et al., 2008). The injury environment is also influenced by contributions from fibroblasts, endothelial cells, pericytes and oligodendrocyte progenitor cells (OPCs) (Figure 2). This is characterized by degenerating myelin, increases in neuro-inhibitory molecules and oxidative stressors, alterations in the extracellular matrix (ECM), and fibrotic scar formation in the affected area.

In the inflammatory cascade occurring post-injury in the CNS, neutrophils are recruited, and endogenous microglia and astrocytes are activated. Blood monocytes migrate to the injury site and differentiate into macrophages nearly indistinguishable from activated microglia. There is a prolonged and dysregulated inflammatory response propagated by pro-inflammatory macrophages (Gensel and Zhang, 2015). Many elements of the inflammatory response have both neuroprotective and neurotoxic potential (Kwon et al., 2004). Infiltrating neutrophils produce oxidative and proteolytic enzymes that prepare the area for repair; however, the overwhelming numbers drawn to the lesion can be detrimental to the surrounding tissues. Macrophages and microglia are capable of producing factors that may help promote axonal growth, as well as molecules that are neurotoxic (Fleming et al., 2006). For in depth reviews see Hohlfeld et al. (2007).

Microglia play an important role in the development, monitoring and maintenance of the CNS. These cells, alongside peripheral monocytes/macrophages, also have a significant role in the post-injury environment (Fleming et al., 2006; Benowitz and Popovich, 2011). The dichotomy of macrophages and microglia stems from the heterogeneity of these cells. Their different activation phenotypes play different roles in the potential deterioration or repair of the surrounding tissue (Kigerl et al., 2009; Murray and Wynn, 2011), especially in relation to neuronal survival and axon growth. Activated microglia and macrophages have been suggested to both promote (Rabchevsky and Streit, 1997) and inhibit (Popovich et al., 2002) neurite outgrowth under different conditions (Gensel et al., 2009).

The functional states of macrophages are broadly categorized along a spectrum between classically activated pro-inflammatory (M1-like) and alternatively activated anti-inflammatory macrophages (M2-like) (Kigerl et al., 2009; Gensel and Zhang, 2015). Classically activated macrophages are efficient producers of potentially neurotoxic effector molecules and pro-inflammatory cytokines, while the alternatively activated phenotype is involved in dampening the inflammatory response and promoting tissue remodeling (Mantovani et al., 2013). While in the adult brain microglial density remains fairly stable over a lifetime, there is constant steady turn-over of microglia (cells surviving an average of 4.2 years) (Tan et al., 2020), allowing for the population to be refreshed and renewed multiple times in a lifetime (Askew et al., 2017).

Inflammation is viewed as an important driver of age-related CNS dysfunction. There is very little known about how the changes in the inflammatory response that occur with age influences the response to SCI. The literature in this area is sparse. Both microglial and astroglial responses are increased and more sustained in older animals, however, aged microglia may be less efficient at debris clearance (Li, 2013). Up to 30 days post-injury there is a greater increase in phagocytes after contusion SCI in middle aged rats, compared to their younger counterparts. Older animals exhibited a significant increase in CD68-positive cells and demonstrated a more pronounced sub-acute CD68 upregulation around the injury site (von Leden et al., 2017). In a thoracic SCI study greater locomotor deficits have been observed in aged (18 months) animals compared to their younger (3 months) counterparts (Hooshmand et al., 2014). The lesion size and amount of cell death was increased, with much lower tissue sparing in the aged animals. The acute localized inflammatory response was also significantly higher. Microglial markers and activation have been observed to increase with age, however, the function of these activated microglia is still largely unknown as is whether this increase is causative or reactive to neurodegeneration (Howcroft et al., 2013). The affect the aging inflammatory response has on recovery from SCI is an area that needs greater attention with the aging SCI population.

Age-related differences in inflammation have been explored to some extent in infants and younger adult animals, particularly in rodents (Sutherland et al., 2017) but have yet to be explored in middle-aged and aging animals. It is believed that the alternative activation of either microglia or macrophages to a M2a phenotype by interleukin (IL)-4/IL-13 may support neuronal growth and repair processes (Kigerl et al., 2009; Murray and Wynn, 2011). Aged (18–19 months old) mice have demonstrated impaired induction of the IL-4α receptor (IL-4Rα) in microglia, accompanied by attenuated downstream Arginase gene expression (anti-inflammatory associated), as well as decreased recruitment of IL-4 receptor expressing macrophages. Aged mice, both wild-type and IL-4Rα knockout, also displayed decreased functional recovery after contusion SCI (Fenn et al., 2014). Macrophages in 14 months old mice exhibited significantly reduced IL-10 expression and CD86/IL-10 (M2b) macrophages in the spinal cord after spinal contusion. This decrease in M2-like phenotype coincided with increased tissue damage and greater functional impairment (Zhang et al., 2015). In a contusion SCI study in mice significantly higher generation of ROS occurred in 14 months old mice compared to their 4 months old counterparts. This was accompanied by NOX2 increases and increased indications of lipid peroxidation. This suggests that macrophages and NOX contribute to SCI oxidative stress that is exacerbated in aging (Zhang B. et al., 2016). In middle-aged rats, both naïve and after a moderate contusion SCI, there were significant increases in markers of oxidative stress, microglial marker Iba1, and inflammatory and NOX2 component gene expression compared to their younger counterparts. These middle-aged rats also exhibited greater lesion volume and showed significantly reduced hind-limb motor scores after the contusion SCI (von Leden et al., 2017). These studies suggest the aged inflammatory state is more detrimental to the surrounding tissue and inhibitory to ongoing regeneration and repair. Both aging and spinal cord injury prime microglia, produce a pro-inflammatory environment and a detrimental oxidative environment, it follows that in combination these two factors will compound one another to worsen recovery after SCI.

Astrocytes have a broad range of roles in the development and maintenance of the CNS. In the mature CNS, they are involved in the cellular, molecular and metabolic homeostasis essential for normal neuronal function. Astrocytes monitor and maintain the extracellular ionic environment and pH, as well as clear extracellular glutamate and gamma aminobutyric acid (GABA). It has emerged that astrocytes also play a role in detoxifying free radicals to prevent excitotoxicity. These cells also have a role in synthesizing precursors for glutamate and GABA production. Furthermore, astrocytes have been shown to sequester and redistribute potassium (K⁺) for neuronal activity. These glia cells produce metabolic substrates for neuron function, including lactate and the production and release of ATP. Astrocytes store glycogen and can produce energy substrates to subsist on anaerobic ATP production. For in depth review see Ransom et al. (2003). Astroglia are also involved in the maintenance of the blood-brain barrier (BBB) and can release factors that modulate endothelial permeability (Abbott, 2002) allowing for the exchange of essential nutrients into the CNS. The processes of astrocytes are associated with capillaries and involved in maintaining cerebral blood flow to support neuronal activity (Mulligan and MacVicar, 2004). Additionally, astrocytes can regulate brain microvascular permeability via Ca²⁺ signaling (Erdő et al., 2017).

Astrocytes are also involved in the sculpting, maintenance and modulation of excitatory and inhibitory synapses and regulation of synaptogenesis (Ransom et al., 2003). They produce factors to induce synapse formation and maturation, and are involved in ‘synaptic pruning’ in the developing nervous system. In the functional synapse perisynaptic astrocytic processes house transporters and enzymes responsible for ion and transmitter homeostasis in the synaptic cleft, as well as for metabolic support (Pekny et al., 2016).

Astrocytes are vital to homeostasis, defense and regeneration of the CNS. In the aging brain, there are alterations in astrocyte morphology, and loss of functionality and reactivity. These changes are heterogeneous, disease-related, and region-specific (Clarke et al., 2018; Cunningham et al., 2019; Matias et al., 2019). In aging the brain, astrocytes play a role in the hypometabolic state that impinges on mitochondrial metabolism and neuronal survival (Jiang and Cadenas, 2014). With age, astrocytes exhibit an early stage of reactive astrogliosis, even in the absence of disease or trauma (Cotrina and Nedergaard, 2002; Boisvert et al., 2018). This is observed in cellular hypertrophy and variations in the expression of structural elements, such as glial fibrillary acidic protein (GFAP) and vimentin, and growth factors, such as ciliary neurotrophic factor (CNTF). An increase in oxidoreductive enzymes is also observed (Cotrina and Nedergaard, 2002). These are hallmarks of reactive astrogliosis that are seen in pathological conditions in the younger CNS. Similar astrocytic changes have been found in the aging spinal cord (Duan et al., 2009). There is evidence that astrocytes undergo cellular senescence with aging. The increase in neuroinflammatory function and loss of normal homeostatic function associated with astrocytic senescence may be an important component of neurodegenerative disorders prevalent in aging (Cohen and Torres, 2019). It has also been observed that the number of astrocytes in some brain regions, especially the cortex, increase significantly with aging (Cotrina and Nedergaard, 2002). Astrocytes in aging are extensively reviewed in Palmer and Ousman (2018).

The physiological changes in astrocytes with aging affect ability to perform their core functions, and can contribute to BBB disfunction (Erdő et al., 2017), alterations in antioxidant systems and increases in oxidative stress (Pertusa et al., 2007; Duan et al., 2009), an increased and altered inflammatory state (Clarke et al., 2018), and reduced neuronal support (Das and Svendsen, 2015). Age-dependent astrocytic alterations may affect synaptic efficacy in normal tissue and also hinder neuronal survival in pathological conditions. Aged astrocytes have been shown to upregulate genes that partially resemble reactive astrocytes, and also those associated with synapse elimination, that varies region to region within the brain. These alterations in aged astrocytes result in an environment susceptible to neuronal damage, a potential factor in age-related diseases and decline (Boisvert et al., 2018). Astrocytic differences in aging animals may impact the efficiency of post-SCI astrogliosis, specifically, the ability of reactive astrocytes to quickly sequester the inflammatory response and reduce its spread, which may result in worse outcomes for patients. One possibility is that a greater imbalance between the protective effects and the neuro-inhibitory effects of reactive astrogliosis develops with aging. Astrocytes undergo an increase in basal mRNA levels of activated astrocyte markers GFAP and S100B in the aged brain (Popa-Wagner et al., 2011). In the normal brain, the region-specific transcriptional identities of astrocytes change with age and astrocytes in the aged brain show the neuroinflammatory phenotype of A1-like reactive astrocytes (Clarke et al., 2018). The differences in astrocyte properties and reactive astrogliosis that occur with age is not a new question, with comparative studies as far back as the 1980s, however, it is still an area with many unanswered questions. In a cortical/hippocampal needle-puncture model, both 3 and 16–19 months old rats showed robust early astrocyte proliferation. This response was heterogenous and regional, with the older rats demonstrating greater proliferation in the cortex not in the hippocampus (Topp et al., 1989). After chemically induced excitotoxity in the striatum aged (22–24 months) rats exhibited an earlier onset and a more pronounced astrogliosis than young (3 months) rats. This is correlated with a delayed progression of neurodegeneration in the aged striatum, and reduced tissue injury and edema size (Castillo-Ruiz et al., 2007).

Reactive astrogliosis and astrogliopathy play an important role in diseases of the nervous system, especially neurodegenerative diseases, that increase in prevalence with age. In aging-prevalent diseases, such as AD, tauopathies, Lewy body diseases, Huntington’s disease (HD), amyotrophic lateral sclerosis (ALS) and Creutzfeldt-Jakob disease, astrogliopathy is often characterized by increased GFAP and abnormal protein aggregates (Ferrer, 2017). After cerebral ischemia, astrocytes form an astroglial scar around the infarct. In a stroke model, aged rats (18–24 months) demonstrated more severe functional impairment, accompanied by an accelerated glial response, compared to their young (2 months) counterparts. This suggests alteration in the temporal progression of the glial response with age may result in altered scar formation and an environment that is more inhibitory to axon growth (Badan et al., 2003). After stroke, levels of the immune modulator transforming growth factor (TGF)-β are increased in astrocytes, activated microglia and macrophages, suggesting that this increase may be associated with regulation of the astroglial scar formation and the immune response. TGF-β is involved in the maintenance of homeostasis, especially vascular homeostasis (Walshe et al., 2009), therefore, an increase in astrocytic TGF-β may suggest a deviation from homeostasis and vascular dysfunction. While TGF-β signaling is increased in the aged brain, it is yet to be correlated with infarct size or functional outcomes (Doyle et al., 2010). In a rodent model of ALS, an age-dependent acceleration in acquiring a senescent phenotype has been observed in astrocytes, accompanied by a reduction in their support of motor neurons (Das and Svendsen, 2015).

One of the hallmarks of trauma to the CNS is reactive astrogliosis, and the formation of a glial and fibrotic scar (Orr and Gensel, 2018). Reactive astrogliosis is the molecular and morphological changes that astrocytes undergo (Sofroniew, 2009). Astrocytes become reactive and undergo functional and physiological changes in trauma, disease and aging (Matias et al., 2019). These alterations include proliferation, hypertrophy of the soma and processes, branching, and upregulation of GFAP expression, and occur in a heterogeneous, context dependent manner regulated by different signaling events (Sofroniew and Vinters, 2010; Anderson et al., 2014). In severe cases reactive astrogliosis results in the formation of the astroglial scar. This is an important process after injury as these cells wall off the damaged areas and seal the lesion to protect the surrounding intact tissue from further damage (Faulkner et al., 2004). Indeed, removal of reactive astrocytes after neurotrauma results in failure of BBB repair, widespread tissue disruption, pronounced cellular degeneration, and severe motor deficits (Faulkner et al., 2004), indicating the role of reactive astrocytes in restoration of the BBB. However, there is still some contention as to whether reactive astrogliosis, and the formation of the astroglial scar, are beneficial or detrimental to functional recovery after SCI (Fawcett and Asher, 1999; Silver and Miller, 2004; Sofroniew, 2009). The role of astrocytes and microglia in regenerative failure in the CNS has been extensively reviewed previously (Silver et al., 2015).

There have been many studies demonstrating that reactive astrocytes have a substantial effect on axon regeneration after SCI. In vitro studies have shown that activated astrocytes produce both pro- and anti-inflammatory effector molecules, which can both help and hinder functional recovery (Sofroniew, 2009; Kawano et al., 2012; Pekny and Pekna, 2014). The astroglial scar is not only a physical barrier, obstructive to axon growth, it also produces a variety of potent growth inhibitory ECM molecules, and the lectican family of chondroitin sulphate proteoglycans (CSPGs) (McKeon et al., 1991, 1999; Li et al., 2013). For decades, the glial scar has been acknowledged as a source of molecules which can inhibit neurite outgrowth (Bovolenta et al., 1992). A 2011 study also suggested that reactive astrocytes also contribute to the failure of re-myelination by high expression of bone morphogenic protein (BMP) (Wang Y. et al., 2011). A study in adult mice demonstrated that neither complete prevention of the astroglial scar formation, attenuation of reactive astrocytes, or the deletion of chronic astroglial scars resulted in increased spontaneous growth of severed axons through a thoracic SCI. Even with the introduction of axon-specific growth factors, preventing the formation of the astroglial scar still significantly reduced the stimulated axon (Anderson et al., 2016).

Recently reactive astrocytes have been shown to be a highly heterogeneous population (Zamanian et al., 2012; Pekny and Pekna, 2014). Neuroinflammation and ischemia induce two different types of reactive astrocytes, referred to as ‘A1’ and ‘A2’ astrocytes respectively. A1 astrocytes demonstrate a more detrimental neuroinflammatory profile whereas A2 astrocytes upregulate many neurotrophic factors and appear more beneficial to injury resolution (Zamanian et al., 2012). A1 reactive astrocytes exhibit impaired ability to carry out normal astrocytic functions, and also release neurotoxic elements that may contribute to damage (Clarke et al., 2018). This heterogeneity of the reactive astrocyte population may go some way to explaining the dichotomy of results around reactive astrogliosis after CNS injury. In the field of SCI there is a large body of literature demonstrating both neuro-protective and neuro-inhibitory elements of post-traumatic reactive astrogliosis (Fawcett and Asher, 1999; Silver and Miller, 2004; Sofroniew, 2009). The duality of the astroglial scar and reactive astrogliosis is still a complicating factor in potential treatments for SCI, which is further impacted by changes in the aging CNS. This is an interesting path to explore in the context of the astrocytic response to SCI and also how this is altered with age.

In the astrocytes, several molecular pathways have been shown to be involved as regulator of reactive astrogliosis in trauma and disease. Studies have pointed to the Janus Kinase (JAK)/STAT3 signaling as an important pathway (Okada et al., 2006; Herrmann et al., 2008). In models of AD and HD the over-expression of SOCS3 in astrocytes inhibited the activation of the JAK/STAT3 pathway and attenuated astrocyte reactivity (Haim et al., 2015). Astrocytic STAT3 deletion reduces GFAP upregulation and scar formation after SCI, as well as increased spread of inflammation, tissue disruption and locomotor deficits (Herrmann et al., 2008). Conversely, astrocytic SOCS3 deletion, an endogenous inhibitor of STAT3, results in rapid astrocyte migration, deceased lesion area and better motor outcomes (Okada et al., 2006). Activation of the PI3K/Akt/mTOR signaling pathway is involved in astrogliosis. In culture, mTOR induces astrocyte proliferation, and its blockade suppresses proliferation, attenuates astrocytic migration, and reduces production of astrocytic-induced inflammatory factors (Li et al., 2015). Overexpression of PTEN in astrocytes attenuates astroglial scar formation, alters CSPG expression and improves axon regeneration into the lesion site (Chen et al., 2016). However, the role of mTOR signaling in both the astrocytic and inflammatory responses, is yet to be fully elucidated. There are mixed reports as how the MAPK pathways are involved in astrogliosis. After traumatic SCI, proliferation of reactive astrocytes in the lesion is accompanied by an increase in the expression and phosphorylation of mitogen-activated protein kinase kinase-extracellular signal-regulated kinases (MEK-ERK). A specific increase in the level of ERK phosphorylation was detected specifically in astrocytes after SCI (Ito et al., 2009). Inhibition of p38 MAPK or ERK has been demonstrated to be protective ischemia or brain injury models (Kang and Hébert, 2011). In another study, inhibition of p38 MAPK appeared to worsen ischemic damage (Lennmyr et al., 2003). Recently, leucine zipper-bearing kinase (LZK) has been shown to mediate astrogliosis, as its deletion in astrocytes reduces astrogliosis after SCI, subsequently increasing the lesion size, while its over-expression enhances astrogliosis and decreases the lesion size (Chen et al., 2018). Together, these suggest that the role of the MAPK signaling in astrocytes response still needs to be elucidated. Others pathways such as NFκB may also be involved in astrogliosis (Kang and Hébert, 2011). NFκB, mTOR, JNK/STAT3 and MAPK pathways are all linked to not just astrocytes and reactive astrogliosis, but are also involved in the inflammatory and neuronal response, demonstrating the complex interlinking of extrinsic factors as well as neuron-intrinsic factors after CNS trauma.

Astrocytes in aging exhibit a range of physiological changes and functional alterations that can impact their response to trauma. Variations in structural elements, growth factors, and oxidoreductive systems (Cotrina and Nedergaard, 2002) in naïve aged astrocytes can compound the existing neurotoxic elements of the injury micro-environment by increasing the detrimental oxidative state and decreasing support for neuron regeneration. The loss of antioxidant tolerance capacity and increased vulnerability to oxidative stress with age (Pertusa et al., 2007; Duan et al., 2009) may be a significant contributor to the increase in post-injury ROS, and oxidative stress after CNS trauma. Significantly, the increased inflammatory state, and skew toward pro-inflammatory phenotypes (Clarke et al., 2018) in aging astrocytes can have a significant impact on the response to trauma. In vitro aged astrocytes show amplified responses to inflammatory cytokines, IL-1β, and TNF-α (Jiang and Cadenas, 2014). The primed, early activation state of these cells (Cotrina and Nedergaard, 2002) may have a cumulative effect with the injury induced reactive astrogliosis, exacerbating the negative effects of the persistent pro-inflammatory response. This warrants investigation, especially in the field of SCI.

There have been few studies comparing younger and aged astrocytes responses to SCI. In primary spinal cord astrocytes aged in vitro (60 DIV), increased expression of ferritin and GFAP were observed along with decreases in glutamate transporter 1 and transferrin receptor 1, and in mitochondrial membrane potential (Duan et al., 2009). Coincidently, in the naïve middle-aged (13 months) mouse spinal cord, decrease in antioxidant enzymes heme oxygenase 1 and NAD(P)H/quinone oxidoreductase 1, and in nuclear factor E2-related factor 2 (Nrf2) which directs expression of these enzymes, were observed (Duan et al., 2009). In vitro, aged astrocytes display an age-dependent increase in mitochondrial oxidative metabolism. The activation of NFκB signaling in these cells results in increased expression of NOX2 and elevated nitric oxide production. This was also associated with an age-dependent increase in hydrogen peroxide generation (Jiang and Cadenas, 2014). Alterations in the antioxidant capabilities and mitochondrial function may have significant flow-on effects in injury scenarios where there is oxidative damage and a detrimental increase in ROS and RNS. In response to SCI in older animals, astrocytes may also present changes in the production profile in CSPGs, and in their role in maintaining blood-spinal cord barrier (BSCB) integrity, which could have far-reaching consequences through the entry of inflammatory and potentially neurotoxic elements into the CNS and which also impact the availability of nutrients and support for potentially regenerative tissue after trauma.

In the area of SCI, the differences in astrocyte reactivity that occur in aging may have significant and far-reaching consequences. Aging astrocytes have been observed to lose their capacity to support neurite outgrowth and their efficacy stimulating axon growth (Smith et al., 1990). In the aging spinal cord, astrocytes exhibit higher basal GFAP mRNA levels, however, their ability to respond to a non-invasive CNS injury was observed to be reduced compared to their younger counterparts (Kane et al., 1997). This is potentially significant in the context of SCI as astrogliosis is an important element of the post-injury response and the interaction between astrocytes and axons is highly influential for axon growth potential. Recently, neuronal PTEN deletion was shown to promote mTOR signaling and axon growth in both the rubrospinal and corticospinal tracts, regardless of the age of the mice, but the axonal regeneration was still greatly diminished in the older animals compared to their younger counterparts (Geoffroy et al., 2016). An increase in astrogliosis, as well as inflammatory marker CD68, was observed in older animals at sub-chronic time-points, suggesting that differences in neuron-extrinsic environmental influences could mediate the age-dependent decrease in axon regeneration. It would be interested to assess whether modification of the astroglial scar with neuronal PTEN deletion could enhance axon regeneration in middle-age animals.

Reactive astrogliosis and the formation of an astroglial scar are significant parts of the post-injury response in the CNS and can influence outcomes after injury. The glial response contributes significantly to the nature of the post-injury environment and may be an important extrinsic factor in regulating axon regeneration, however, this is an area that still needs to be examined in detail. Astrocytes exhibit changes in reactivity and response in the aging CNS that can have significant consequences for the progression of injury, and the development of CNS disorders. The significant alterations in astrocytic response with injury and aging are shown in Figure 4. This is potentially an important factor and one that we have little knowledge about. In the aging CNS these responses are robust but also irregular, which may contribute to the age-related decline in axon regeneration and also to worsening post-injury outcomes after SCI. This potential relationship between aging, the glial response and injury resolution is an area of great interest that requires a considerably more study.

After SCI, fibroblasts from different origins proliferate and form the fibrotic scar, which is isolated from the spare tissue by the reactive astrocytes. For detailed review see Dias and Göritz (2018). Fibroblasts are a major source of fibronectin, collagen and other ECM components known to inhibit regeneration. Reducing the fibrotic scar can enhance axon regeneration and increase functional recovery (Ruschel et al., 2015; Dias et al., 2018) in young animals. However, while most studies aim to reduce the fibrotic scar, ablation of fibroblasts may compromise tissue integrity after injury, elements of the fibrotic scar may play supporting roles, and removing them acutely after injury may have negative effects (Göritz et al., 2011; Bradbury and Burnside, 2019). In non-nervous tissue, aging is known to alter fibroblasts (Varani, 2010). Different pathways are altered in skin and cardiac fibroblasts with aging, including a decrease in ERK and Akt expression, an increase in JNK and TGF-β/Smad signaling, which can alter their proliferation and production of ECM proteins. Therefore, it is likely that age affects the fibrotic response after SCI. There is a need to understand how age specifically modify this response.

The vascular and nervous system are closely juxtaposed, and endothelial cells respond to similar patterning cues as growth cones (Adams and Eichmann, 2010). There is increasing interest in taking advantage of this interaction to control axon regeneration by controlling vascular patterning after SCI (Partyka et al., 2019; Tran et al., 2020). Endothelial cells secrete vascular endothelial growth factor (VEGF), which promotes cell survival and axonal outgrowth (Sondell et al., 1999). These cells can also secrete other neurotrophic factors promoting axon growth (Grasman and Kaplan, 2017). With age, the production of different factors by endothelial cells, such as VEGF, declines with age (Shetty et al., 2005), which can result in a decrease in cell survival and in axon growth with age. The production of endothelin-1 and nitric oxide by endothelial cells are also altered, resulting in the modification of vascular tone (vasoconstriction/vasodilation). Lack of control of the vascular tone can be result in inability to meet the metabolic demand of the tissue, increase of local hypoxia, and reduction in waste removal (Bosetti et al., 2016; Jacob et al., 2016). Alteration of this balance at the injury site is likely to have negative effect on axon regeneration. Finally, changes in vascular permeability and BBB leakiness with aging can result in infiltration of molecules into the CNS, alteration in the microenvironment, and increase in neurotoxicity. This demonstrates the need to better understand how SCI and age can impact endothelial cell function and their relation to axon growth potential.

Pericytes are closely associated with endothelial cells, and have projections encircling CNS vessels. These cells not only play a role in regulating and maintaining vascular barrier integrity (Sá-Pereira et al., 2012), they have a wide range of physiological functions (Muramatsu and Yamashita, 2014). CNS trauma often leads to vascular disruption and dysfunction, closely linked to neuronal dysfunction. One significant element of this is the reduction in trophic support for the CNS cells. After SCI, excess of monoamine receptors (5-HT1) on pericytes leads to local constriction of capillaries and chronic hypoxia. Inhibiting this receptor increases locomotor recovery (Picoli et al., 2019).

Pericytes have emerged as potential targets to increase functional recovery after SCI. For a detailed review, see Picoli et al. (2019). Pericytes in the spinal cord are heterogenous. After SCI, a subtype of pericyte gives rise to a population of stromal cells integral to the glial and fibrotic scars, and play a role in post-injury fibrosis. These cells were shown to be essential to the successful sealing of the lesion (Göritz et al., 2011). Inhibiting the proliferation of pericytes reduces the fibrotic scar tissue after SCI, and improves the regeneration of axon tracts, prompting improved sensorimotor functional recovery (Dias et al., 2018). After SCI, pericytes contribute a significant amount of collagen to the scar formation (Birbrair et al., 2014) and NG2+ pericytes are also involved with angiogenesis (Hesp et al., 2018).

To our knowledge, how the age factor impacts the response of pericytes in SCI models has not been characterized. Pericyte loss and neurovascular dysfunction have been linked to age-dependent neurodegeneration. Alteration of the BBB is observed in different neurogenerative conditions (ALS, HD) as well as neurotrauma (TBI and stroke). For review see Lendahl et al. (2019) and Li et al. (2019). Therefore, it is conceivable that SCI in older animals’ results in enhanced BBB breakdown, increase in serum proteins leakage to the spinal lesion, increase in neurotoxicity as well as worse functional outcome. Additionally, a poorly functioning vascular system will likely result in less nutrients reaching potentially growing axons. Modulating pericytes integrity and activity could be of therapeutic interest in the context of middle-age and age SCI.

Different cell types can express NG2 in the CNS. Here, we refer to NG2+ cells as OPCs. OPCs are present throughout the CNS. They react to injury by proliferating and migrating toward the lesion, and differentiating into oligodendrocytes and astrocytes (Nishiyama et al., 2009; Hackett et al., 2018). For review see Hackett and Lee (2016). Removing NG2+ cells (both OPCs and pericytes) results in astrocytic and fibrotic scar disruption, with reduction in scar density and less distinct borders (Hesp et al., 2018). This was surprisingly accompanied by axon growth into the lesion as well as persistent functional impairment. While previous studies show that NG2 proteoglycan inhibits axon growth, NG2+ cells themselves may facilitate growth (Hackett and Lee, 2016). NG2+ cells proliferation and differentiation is reduced (Sim et al., 2002; Psachoulia et al., 2009) and OPCs heterogenicity altered with age (Spitzer et al., 2019). NG2+ cells have been shown to maintain neuron survival, via reduction of IL-1β pro-inflammatory pathway in the hippocampus (Nakano et al., 2017). Additionally, intracellular pathways involved in OPCs differentiation and myelination include ERK/MAPK, PI3K/Akt/mTOR and Wnt/β-catenin (Gaesser and Fyffe-Maricich, 2016), all of which can be altered with aging. An attractive hypothesis would be that the reduction in NG2+ activity with age and injury participates to neuroinflammation and inducing neuron death, and the reduction in NG2 cells proliferation and differentiation results in an environment less favorable to axon growth.

The importance of the ECM and extracellular elements of the environment in an age-dependent decline in regeneration potential is highlighted in a study by Kawano et al. comparing postnatal 7, 14, and 21 with adult mice (P60-80) (Kawano et al., 2005). Transected axons regenerated across the lesion only in P7; in P14 and older, the axons stop at the lesion site. The fibrotic scar, containing type IV collagen, was not present in the P7 mice. Inhibition of collagen synthesis in older mice prevented the formation of this scar resulting in a greater number of axons extending across the lesion. These elements are intrinsically connected to the cellular elements and the formation of CNS scar tissue, as well as influencing the nature of the injury microenvironment. Overall, there has been very little done to elucidate the changes in extracellular components in aging and how it plays a role in the reduction of axon regeneration potential; the changes with age and traumatic spinal injury is still yet to be elucidated.

Studies have found multiple potent neurite growth inhibitory factors that are enriched in CNS myelin, such as Nogo-A, myelin-associated glycoprotein (MAG), oligodendrocyte/myelin glycoprotein (OMgp). After injury, myelin sheaths are damaged, and these factors are released into the extra-cellular environment. These molecules alter the regenerative potential after CNS and targeting them or their receptors can enhance axon growth after SCI, although, some conflicting results have been reported. Indeed, some groups reported a beneficial effect of genetic Nogo deletion on axon regeneration and sprouting and others only on sprouting (Cafferty et al., 2010; Lee et al., 2010). For review (Geoffroy and Zheng, 2014).

Few studies examined how age alters the expression of these proteins. Nogo-A levels are significantly decreased in aged animals (Trifunovski et al., 2006; Kumari and Thakur, 2014). A decrease in recovery is surprisingly observed in 12-month-old Nogo A/B deficient mice after TBI, compared to young adults, possibly related to hypomyelination (Marklund et al., 2009). Aged brains with stroke present a unique profile in growth-inhibitory molecule changes (Li and Carmichael, 2006). In aged brains OMgp is increased earlier than in young animals; MAG is increased in old injured brains but decreased in young; and Nogo-A is decreased in the aged while increased in young mice Nogo receptor 1 (NgR1) antagonist has been shown to increase OPC maturation and remyelination, and motor recovery after stroke in aged mice (Sozmen et al., 2016). While Nogo-A is not altered in 12-month-old mice, at least at 6 weeks post injury (Geoffroy et al., 2016), the exact age-related profile change of these molecules at different time point after SCI has not been characterized. Indeed, as discussed in Section “The Inflammatory Response Changes with Age: ‘Inflamm-Aging,” there are been reports of reduction in debris clearance by aged macrophages and monocytes (Neumann et al., 2008; Natrajan et al., 2015). Therefore, one would speculate that the inhibitory myelin proteins stay longer at the injury site of older animals, and reduce axon growth potential. Additionally, the expression of the MHC1 receptor, PirB, is increased with age in the hippocampus, localized primarily in neurons. Induction of the MHC1 pathway in neurons is potentially linked to increased neuro-immune signaling and altered synaptic homeostasis, contributing to age-related hippocampal dysfunction (VanGuilder Starkey et al., 2012). If a similar pattern occurs in spinal neurons, it may lead to greater inhibition in aged neurons than their younger counterparts with the same amount of myelin proteins. PirB is increased in the spinal cord after injury (Zhou et al., 2010), while how age alters spinal expression of PirB is still to be elucidated. Similarly, NgR1, and its ligands, are increased in the hippocampus with age (VanGuilder Starkey et al., 2013). These data suggest that, even absent alterations in myelin protein levels, changes in neuronal expression of receptors may increase the inhibitory effects of these proteins on axon regeneration.