New ex vivo demyelination/remyelination models to defeat multiple sclerosis and neuromyelitis optica

Multiple sclerosis (MS) and neuromyelitis optica (NMO) are inflammatory diseases of the central nervous system (CNS) resulting in CNS inflammation, infiltration of peripheral immune cells, loss of myelin and oligodendrocytes, interruption of axonal communication, and neurologic deficits. Following oligodendrocyte injury, newly generated myelinating oligodendrocytes derived from oligodendrocyte progenitors (OPCs) may produce new myelin sheaths around denuded axons (remyelination) restoring neuronal function (Verden and Macklin, 2016). While remyelination is apparent in MS lesions, the process is often inefficient; in NMO, remyelination is even more limited. Currently, there are no restorative therapies for MS and NMO. Understanding the mechanisms underlying driving disease-specific myelin damage and repair is critical to identify remyelination barriers and develop new treatments promoting remyelination. Current studies to understand the molecular and cellular events regulating demyelination and remyelination processes have employed developmental, toxic, and caustic models of oligodendrocyte injury. These models cannot replicate the inflammatory pathology of demyelination in MS and NMO, and fail to replicate the milieu that may inhibit repair (Plemel et al., 2017).

Immunoglobulin G (IgG) deposition is a histopathologic feature of active MS and NMO lesions. In MS, oligoclonal cerebrospinal fluid (CSF)-specific IgGs remain stable over time and are not affected by pharmacological therapies (Krumbholz and Meinl, 2014). In NMO, aquaporin-4 autoantibodies accessing the CNS cause astrocyte destruction and secondary myelinolysis and neuronal loss. We have produced IgG1 monoclonal recombinant antibodies (rAbs) from clonally-expanded CSF plasmablasts recovered from MS and NMO patients (Bennett et al., 2009; Owens et al., 2009). Using these disease-specific rAbs to initiate complement-dependent cytotoxicity in ex vivo cerebellar slices, we have developed novel experimental models of MS and NMO lesions. MS myelin-specific rAbs bind to discrete surface domains on oligodendrocyte processes and myelinating axons, causing robust oligodendrocyte loss, rapid demyelination and microglia activation; astrocytes, OPCs and neurons remain unaffected. In contrast, NMO aquaporin-4-specific rAb results in complement-dependent astrocyte destruction, followed by oligodendrocyte loss, demyelination, microglia activation, and neuronal death (Liu et al., 2017). Our rAb-slice models recapitulate some of the reported pathologic features of active MS and NMO lesions and provide strong evidence that antibodies produced by B cell populations expanded within the CNS compartment contribute to NMO and MS damage. The distinct patterns of injury induced by the MS and NMO rAbs in the presence of complement indicate that the target of complement-dependent cytotoxicity, and not activation of the complement cascade itself, is important for delineating the spectrum of glial and neuronal injury (Liu et al., 2017).

Using our NMO- and MS-specific rAbs to generate disease-specific injuries, we are primed to evaluate the recovery of cerebellar tissue and characterize distinct patterns of glial responses that may determine their disparate capacities for remyelination. Oligodendrocytes repopulate after both MS and NMO rAb-mediated injury; however, oligodendrocytes only mature into functional myelinating cells after exposure to MS myelin-specific rAb and complement. Remyelination from MS rAb-induced damage is accompanied by pronounced microglial activation. In contrast, oligodendrocyte maturation and remyelination fail following NMO rAb-mediated injury despite the rapid restoration of astrocytes and the early preservation of axons. Deficient remyelination following NMO rAb-mediated injury is associated with progressive axonal loss and the return of microglia to a resting state (Liu et al., 2018).

Comparing the distinct patterns of damage and repair discovered in rAb-slice models of MS and NMO (Figure 1)reveals critical steps in remyelination and potential therapeutic strategies for facilitating remyelination in these inflammatory neurological disorders.

Figure 1.

Distinct glia responses in MS and NMO rAb-slice models during damage and recovery.

Cerebellar slice cultures are prepared from mice at postnatal day 10 and cultured for 7–10 days prior to treatment. Slices are treated with MS myelin-specific rAb or NMO aquaporin-4 + rAb (human IgG1) at 20 µg/mL in the presence of 10% (vol/vol) normal human serum as a source of complement for 24–48 hours to induce damage. Then the treatment is removed and slices are cultured in medium for additional 7–14 days for recovery. Distinct patterns of rAb-induced complement-dependent cytotoxicity contribute to demyelination injury. Different glial responses are coupled with disparate capacities for remyelination after treatment withdrawal. MS: Multiple sclerosis; NMO: neuromyelitis optica; N: neuron; AST: astrocytes; OL: oligodendrocyte; MG: microglia; rAb: recombinant antibody.

Remyelination failure in MS and NMO: Similar to other models of demyelination, our MS rAb-slice model demonstrates spontaneous and robust remyelination following the cessation of injury. In contrast, remyelination in MS patients is highly variable: considerable in some cases, and minimal to virtually absent in others. MS patients exhibiting robust remyelination demonstrate lower levels of disability, providing optimism that remyelination therapies are possible and can facilitate the restoration of neurological function (Louapre et al., 2015). Human studies and experimental animal models have indicated that causes of remyelination failure in MS include the presence of extrinsic inhibitors, insufficient pro-regenerative factors, and deficient regenerative capacity within oligodendrocyte lineage cells (Plemel et al., 2017). Many of these experimental models of myelin injury are induced by focal and transit injuries and may not faithfully reproduce the chronic inflammatory environment of CNS inflammatory disorders. For instance, in MS patients, the intrathecal synthesis of CSF-specific IgGs results in chronic CNS exposure to IgG over the course of the disease (Krumbholz and Meinl, 2014). Whether the persistent presence of MS specific intrathecal antibodies contribute to remyelination inhibition is unclear.

Remyelination has received less attention in NMO. The combination of astrocyte, oligodendroglial, and neuronal pathology likely limit remyelination through diverse mechanisms: (i) inhibition of oligodendroglial differentiation; (ii) inhibition of oligodendrocyte progenitor migration due to blood-brain barrier injury; and (iii) impaired myelin wrapping secondary to irreversible axonal destruction (Weber et al., 2018). Nevertheless, pathology in NMO cerebellar slice model suggests similar axonal preservation in early NMO and MS lesions (Liu et al., 2017). It has been reported that the approved drug clobetasol promotes remyelination in a mouse model of NMO, providing proof-of-concept for the potential utility of a remyelinating agents in the treatment of NMO (Yao et al., 2016).

Glial responses for effective remyelination: Remyelination failure results from impaired recruitment of OPCs into the lesion site and inhibition of OPC differentiation into new mature remyelinating oligodendrocytes. Therefore, current approaches to enhance remyelination have been focused on promoting the recruitment or stimulating the proliferation and differentiation of OPCs (Plemel et al., 2017). In our NMO rAb-slice model, no defect in OPC differentiation is observed. In contrast, regenerated early stage oligodendrocytes fail to mature to the myelinating stage and result in remyelination failure despite the restoration of astrocyte and preservation of axons in early lesions (Liu et al., 2018). This finding suggests that oligodendrocyte maturation represents a crucial checkpoint for effective myelin regeneration and deserves further and more extensive evaluation as a remyelination strategy.

Although we have observed a rapid repopulation of astrocytes in cerebellar slices recovering from NMO rAb-mediated damage, the regenerated cells may not function equivalently to nascent astrocytes (Liu et al., 2018). Abundant evidence has shown that astrocytes actively participate in both MS development and repair. In addition to driving inflammatory neurotoxicity and contributing to neuroprotection during glial scar formation, reactive astrocytes suppress remyelination (Ponath et al., 2018). Further characterization of regenerating astrocytes in NMO lesions may elucidate a role in remyelination suppression.

We have observed a significant difference in microglia activation following MS and NMO rAb-mediated injury (Liu et al., 2018). Microglia cells are the principal resident immune cells in the CNS and play versatile roles in CNS development, maintenance, repair and pathology. Although microglia-induced neuroinflammation participate in the occurrence and progression of many neurological disease models, they also exhibit protective and regenerative properties. Microglial activation accompanies demyelination induced by MS and NMO rAbs in cerebellar slices (Liu et al., 2017). A second increase in microglial numbers and reactivity occurs in slices recovering from MS rAb-induced damage, whereas microglial activation continues to decline during recovery from NMO rAb demyelination, suggesting that microglial activation promotes effective remyelination (Liu et al., 2018). Increasing evidence in other demyelination models is consistent with this observation. Transcriptomics and functional assays implicate that microglia function in remyelination likely involves phagocytosis of myelin debris, the secretion of growth factors and remodeling of the extracellular matrix to recruit OPCs and promote oligodendrocyte regeneration (Lloyd et al., 2017). However, without any characterization of the regenerated oligodendrocytes in these studies, whether microglia promotes OPC differentiation or oligodendrocyte maturation from the early myelinating to mature myelinating stage remains unclear. Results from our rAb-slice models demonstrate that following NMO rAb-mediated demyelination, there is limited microglial activation and remyelination despite normal OPC differentiation (Liu et al., 2018). Hence, microglia activation may play an important role in advancing oligodendrocytes to a mature myelinating stage. Understanding the role of microglial activation in the later stage of functional remyelination will contribute to identifying the factors that promote oligodendrocyte differentiation.

Further development of MS and NMO experimental models: The absence of a peripheral immune compartment in the cerebellar slice culture model allows the effect of CNS resident cells to be clearly delineated. Unfortunately, the absence of peripheral immune cells may mask the effects of cell-mediated injury. Further modification of the cerebellar slice culture model to include mononuclear cells, phagocytes, and inflammatory cytokines may help to dissect the complex relationship between the immune response and CNS myelin damage and repair. In addition, due to the nature of the model system, the cerebellar slices can only be cultured from postnatal 10–12 mice. The development of in vivo models of MS and NMO rAb injury will further elucidate the complex mechanisms governing myelin damage and repair in MS and NMO.

Conclusion and future perspective: We have developed novel experimental ex vivo models of rAb-mediated demyelination/remyelination for MS and NMO. These models recapitulate some of the seminal pathologic features of MS and NMO lesions. With further modifications, they can serve as new models to dissect disease-specific mechanisms, such as the complex interactions of the inflammatory response with CNS glia damage/repair and the remyelination obstacles. Furthermore, these MS and NMO rAb-slice models provide an efficient system to identity and validate potential therapies to overcome remyelination inhibition and effectively improve myelin regeneration in affected patients.

This work was supported by National Multiple Sclerosis Society, NIH and the Guthy-Jackson Charitable Foundation.

I would like to thank Dr. Wendy Macklin (Department of Cell & Developmental Biology, School of Medicine, University of Colorado), and Drs. Gregory Owens and Jeffrey Bennett (Department of Neurology, School of Medicine, University of Colorado) for their helpful comments on this manuscript and their collaboration and support in the development of the ex vivo models.