A Therapeutic Link between Astrogliosis and Remyelination in a Mouse Model of Multiple Sclerosis

Injuries to the mammalian CNS often provoke astrogliosis, a process in which reactive astrocytes surround the site of injury to restrict tissue damage (Voskuhl et al., 2009; Hamby and Sofroniew, 2010). Formation of the glial scar has historically been viewed as a short-term solution that creates long-term problems: while astrogliosis effectively limits the spread of inflammation and infection, it also creates a physical and chemical barrier to the regeneration and remyelination of axons (Cregg et al., 2014). Many therapies for CNS injuries have thus sought to remove this barrier by blocking astrogliosis and restricting or neutralizing the glial scar. However, in certain contexts, blocking reactive astrocyte formation has been found to impede functional recovery (Anderson et al., 2016), suggesting that some astrocytes may exert neuroprotective effects (Faulkner, 2004). Therefore, rather than prohibit astrogliosis entirely, alternative therapies have sought to specifically target molecules produced by reactive astrocytes that inhibit axon regeneration and remyelination.

Among these molecules, substantial attention has been paid to chondroitin sulfate proteoglycans (CSPGs), which are extracellular matrix proteins that inhibit neuronal growth in vivo and in vitro and have been linked to remyelination failure (Galtrey and Fawcett, 2007). Therapies that selectively remove the inhibitory glycosaminoglycan (GAG) chains of CSPGs have been vital to studies of axon regeneration and remyelination in many models of CNS injury and disease (Soleman et al., 2013). Given the complexity of extracellular responses to acute and chronic damage, it is crucial to develop a clear understanding of how therapeutic interventions modify the inhibitory microenvironment to ensure their success.

One disease that might benefit from treatments targeting CSPGs associated with astrogliosis is multiple sclerosis (MS), an inflammatory disease characterized by multiple focal regions of demyelination and axonal damage, known as plaques, in the brain and spinal cord (Compston and Coles, 2008). At the time of diagnosis, most patients exhibit a relapsing–relapsing remitting disease course, in which periods of disability are followed by periods of recovery; however, many patients eventually enter a secondary progressive phase in which disabilities accumulate without recovery (Compston and Coles, 2008). Whereas demyelinated axons in MS plaques may spontaneously remyelinate in the relapsing–remitting phase, the failure of remyelination in the progressive phase leads to extensive neurodegeneration, resulting in severe dysfunction and disability (Franklin, 2002). Curiously, these demyelinated plaques are not typically depleted of myelin-forming oligodendrocytes (Chang et al., 2002).

The inability of oligodendrocytes in MS lesions to successfully remyelinate damaged axons suggests that some aspect of the lesion environment, such as astrogliosis, inhibits the remyelination process. Astrocytes in the white matter of MS lesions associate with CSPGs, and remyelination tends to be more robust in areas with fewer reactive astrocytes and lower levels of CSPGs, in conjunction with higher numbers of oligodendrocyte precursor cells (OPCs; Chang et al., 2012). Furthermore, CSPGs inhibit OPC growth and differentiation in vitro, and enzymatic digestion of their GAG chains reverses this inhibition (Siebert and Osterhout, 2011; Lau et al., 2012; Pendleton et al., 2013). Moreover, in vivo therapies that block CSPG synthesis (Lau et al., 2012; Keough et al., 2016), digest GAG chains (Bartus et al., 2014), or disrupt CSPG receptors (Lang et al., 2015) have all been shown to improve axon regeneration and remyelination, providing convincing evidence that CSPGs antagonize the remyelination process. Together, these findings point to CSPGs as an important target for MS therapies.

In an article published in The Journal of Neuroscience, Feliú et al., (2017) describe a pharmacological treatment that reduces CSPG deposition in demyelinating lesions of the spinal cord and leads to improved motor function in a mouse model of MS, illustrating an important new link between CSPGs and remyelination (Feliú et al., 2017). To study changes in the extracellular environment of demyelinated lesions, the researchers examined mice inoculated with Theiler's murine encephalomyelitis virus (TMEV), which stimulates a progressive model of MS in which the chronic phase is accompanied by astrogliosis (Feliú et al., 2015). The spinal cords of TMEV-infected mice were analyzed at 85 d postinjury (dpi), during the chronic phase of the disease, and 105 dpi, a well established time point for assessing early remyelination (Warrington et al., 2000). Astrogliosis was evaluated by immunostaining for GFAP and vimentin. CSPG levels were measured using immunostaining with an antibody (CS-56) specific to the GAG chains of CSPGs, as well as with RT-PCR and Western blot. By 85 dpi, a robust increase in reactive astrocytes was observed in conjunction with CSPG levels and increased expression of CSPG-related mRNA and proteins in demyelinated spinal cord lesions. CSPGs were closely associated with reactive astrocytes. Given the evidence that astrocytes produce CSPGs following CNS injury (McKeon et al., 1999), this observation suggests that the CSPGs in TMEV mice were likely produced by reactive astrocytes. Encouragingly, treating mice with xyloside, which inhibits CSPG synthesis, reduced markers of astrogliosis and CSPGs, and also rescued motor function as measured by spontaneous motor activity (Feliú et al., 2017).

Having confirmed that TMEV produces demyelinated lesions characterized by reactive astrocytes and elevated CSPG deposition, Feliú et al. (2017) sought to determine whether the endogenous cannabinoid 2-arachidonoylglycerol (2-AG) affects astrogliosis. 2-AG plays a central role in promoting oligodendrocyte proliferation and differentiation (Gomez et al., 2011; Gomez et al., 2015). Previous work showed that a single injection of 2-AG reduced white matter damage after spinal cord injury in rats (Arevalo-Martin et al., 2010) and that sustained delivery over 14 d delayed disease onset in mice with experimental autoimmune encephalomyelitis, another commonly used model of MS (Lourbopoulos et al., 2011). However, the effects of 2-AG on astrogliosis had not yet been explored. To produce elevated 2-AG levels in vivo, Feliú et al. (2017) pharmacologically inhibited the hydrolysis of 2-AG by monoacylglycerol lipase (MAGL) using a reversible MAGL inhibitor, UCM03025. After 10 d of treatment, the MAGL inhibitor yielded significant functional improvements in spontaneous motor activity and reduced immunostaining for markers of reactive astrocytes and CSPGs (Feliú et al., 2017). This indicates that the previously demonstrated therapeutic effects of 2-AG administration may be due in part to its reduction of astrocyte-derived CSPGs and that these beneficial effects can be achieved by pharmacologically inhibiting 2-AG hydrolysis.

In addition to reducing inhibitory components of the glial scar, inhibiting 2-AG hydrolysis increased the number of OPCs and mature oligodendrocytes in the demyelinated lesions of TMEV mice at 85 dpi. Western blot analysis indicated that levels of myelin basic protein were restored to sham levels in UCM03025-treated mice. This suggests that suppressing MAGL and elevating 2-AG enabled OPCs to migrate to the lesion and differentiate into mature oligodendrocytes, thereby restoring the endogenous mechanism of remyelination. This observation was extended by assessing motor activity at 105 dpi without further UCM03025 treatment. CSPG immunostaining and protein levels remained lower in treated mice than in controls, even at this later time point. Furthermore, lesions examined by transmission electron microscopy at 105 dpi revealed newly formed myelin sheaths and an increase in myelinated axons in the spinal cords of mice treated with the MAGL inhibitor, supporting the conclusion that remyelination was indeed taking place. Interestingly, UCM03025 administration led to reduced immunostaining for a marker of microglial activation and lowered levels of mRNA for the immune mediator IL-1β, indicating that 2-AG may also play an immunomodulatory role.

Understanding the mechanisms that prevent remyelination in MS is vital to developing effective therapies, particularly for patients in the progressive phase of the disease. The discovery that enhancing endogenous 2-AG levels also reduces CSPG accumulation in demyelinated lesions is a significant step toward establishing a new therapeutic target. Our knowledge of the role of the glial scar in tissue remodeling after CNS injury has become increasingly nuanced, with evidence that astrogliosis is not uniformly inhibitory to repair. Astrocytic scar formation was found to support axon regeneration after spinal cord injury in mice (Anderson et al., 2016), and it was recently discovered that reactive astrocytes can be classified into at least two subtypes, only one of which inhibits the repair of neuronal pathways (Liddelow et al., 2017). This dichotomy extends to the molecules produced by reactive astrocytes: emerging evidence shows that even CSPGs themselves, long considered an unequivocal barrier to axonal growth, can undertake both permissive and inhibitory functions depending on their structural features. For instance, it has been demonstrated that the different sulfation patterns of CSPG GAG chains influence how neurons respond to these molecules, with cultured neurons avoiding 4-sulfated GAGs yet growing readily over 6-sulfated GAGs (Wang et al., 2008). Shifts in the ratio of 4-sulfated GAGs to 6-sulfated GAGs in the brain were recently linked to age-related declines in plasticity in mice, suggesting that selective targeting of inhibitory GAG chain motifs may make for more effective therapies (Foscarin et al., 2017).

More research is required to untangle the complex interactions between neurons and their environment after injury or disease. It would be valuable, for instance, to study whether CSPG GAG chain sulfation changes after inhibiting 2-AG hydrolysis in TMEV mice or whether selectively targeting inhibitory sulfate groups without directly reducing CSPG levels might improve functional outcomes in a similar manner. The relationship between CSPGs, OPCs, and inflammatory cells also deserves further investigation. Feliú et al., (2017) found that inhibiting 2-AG hydrolysis reduces microglial activation and IL-1β mRNA levels, suggesting that this treatment also modulates the inflammatory response. Isolating the contributions of these various effects may prove challenging due to their interdependence. As the mechanisms underlying axon regeneration and remyelination become better described, it is increasingly vital to elucidate how the extracellular matrix contributes to these injury response processes and to develop therapeutic interventions that selectively and effectively promote neuronal repair. The findings reported by Feliú et al., (2017) expand our understanding of how the glial scar contributes to the failure of remyelination in chronic disease and suggest promising future avenues of investigation and potential treatments for demyelinating conditions like MS.