Tumor necrosis factor (TNF) is the prototypic pro-inflammatory cytokine. It is central to host defense and inflammatory responses but under certain circumstances also triggers cell death and tissue degeneration. Its pleiotropic effects often lead to opposing outcomes during the development of immune-mediated diseases, particularly those affecting the central nervous system (CNS). The reported contradictions may result from lack of precision in discussing TNF. TNF signaling comprises at minimum a two-ligand (soluble and transmembrane TNF) and two-receptor (TNFR1 and TNFR2) system, with ligands and receptors both differentially expressed and regulated on different cell types. The "functional multiplicity" this engenders is the focus of much research, but there is still no general consensus on functional outcomes of TNF signaling in general, let alone in the CNS. In this review, evidence showing the effects of TNF in the CNS under physiological and pathophysiological conditions is placed in the context of major advances in understanding of the cellular and molecular mechanisms that govern TNF function in general. Thus the roles of TNF signaling in the CNS shift from the conventional dichotomy of beneficial and deleterious, that mainly explain effects under pathological conditions, to incorporate a growing number of "essential" and "desirable" roles for TNF and its main cellular source in the CNS, microglia, under physiological conditions including regulation of neuronal activity and maintenance of myelin. An improved holistic view of TNF function in the CNS might better reconcile the expansive experimental data with stark clinical evidence that reduced functioning of TNF and its dominant pro-inflammatory receptor, TNFR1, are risk factors for the development of multiple sclerosis. It will also facilitate the safe translation of basic research findings from animal models to humans and propel the development of more selective anti-TNF therapies aimed at selectively inhibiting deleterious effects of this cytokine while maintaining its essential and desirable ones, in the periphery and the CNS.

CNS inflammation is a major driver of MS pathology. Differential immune responses, including the adaptive and the innate immune system, are observed at various stages of MS and drive disease development and progression. Next to these immune-mediated mechanisms, other mediators contribute to MS pathology. These include immune-independent cell death of oligodendrocytes and neurons as well as oxidative stress-induced tissue damage. In particular, the complex influence of oxidative stress on inflammation and vice versa makes therapeutic interference complex. All approved MS therapeutics work by modulating the autoimmune response. However, despite substantial developments in the treatment of the relapsing-remitting form of MS, approved therapies for the progressive forms of MS as well as for MS-associated concomitants are limited and much needed. Here, we summarize the contribution of inflammation and oxidative stress to MS pathology and discuss consequences for MS therapy development.

Background: During inflammatory demyelination, TNF receptor 1 (TNFR1) mediates detrimental proinflammatory effects of soluble TNF (solTNF), whereas TNFR2 mediates beneficial effects of transmembrane TNF (tmTNF) through oligodendroglia, microglia, and possibly other cell types. This model supports the use of selective inhibitors of solTNF/TNFR1 as anti-inflammatory drugs for central nervous system (CNS) diseases. A potential obstacle is the neuroprotective effect of solTNF pretreatment described in cultured neurons, but the relevance in vivo is unknown. Methods: To address this question, we generated mice with neuron-specific depletion of TNFR1, TNFR2, or inhibitor of NF-κB kinase subunit β (IKKβ), a main downstream mediator of TNFR signaling, and applied experimental models of inflammatory demyelination and acute and preconditioning glutamate excitotoxicity. We also investigated the molecular and cellular requirements of solTNF neuroprotection by generating astrocyte-neuron co-cultures with different combinations of wild-type (WT) and TNF and TNFR knockout cells and measuring N-methyl-d-aspartate (NMDA) excitotoxicity in vitro. Results: Neither neuronal TNFR1 nor TNFR2 protected mice during inflammatory demyelination. In fact, both neuronal TNFR1 and neuronal IKKβ promoted microglial responses and tissue injury, and TNFR1 was further required for oligodendrocyte loss and axonal damage in cuprizone-induced demyelination. In contrast, neuronal TNFR2 increased preconditioning protection in a kainic acid (KA) excitotoxicity model in mice and limited hippocampal neuron death. The protective effects of neuronal TNFR2 observed in vivo were further investigated in vitro. As previously described, pretreatment of astrocyte-neuron co-cultures with solTNF (and therefore TNFR1) protected them against NMDA excitotoxicity. However, protection was dependent on astrocyte, not neuronal TNFR1, on astrocyte tmTNF-neuronal TNFR2 interactions, and was reproduced by a TNFR2 agonist. Conclusions: These results demonstrate that neuronal TNF receptors perform fundamentally different roles in CNS pathology in vivo, with neuronal TNFR1 and IKKβ promoting microglial inflammation and neurotoxicity in demyelination, and neuronal TNFR2 mediating neuroprotection in excitotoxicity. They also reveal that previously described neuroprotective effects of solTNF against glutamate excitotoxicity in vitro are indirect and mediated via astrocyte tmTNF-neuron TNFR2 interactions. These results consolidate the concept that selective inhibition of solTNF/TNFR1 with maintenance of TNFR2 function would have combined anti-inflammatory and neuroprotective properties required for safe treatment of CNS diseases.

Multiple sclerosis (MS) is an inflammatory CNS demyelinating disease in which remyelination largely fails. Transmembrane TNF (tmTNF) and TNF receptor 2 are important for remyelination in experimental MS models, but it is unknown whether soluble TNF (solTNF), a major proinflammatory factor, is involved in regeneration processes. Here, we investigated the specific contribution of solTNF to demyelination and remyelination in the cuprizone model. Treatment with XPro1595, a selective inhibitor of solTNF that crosses the intact blood-brain barrier (BBB), in cuprizone-fed mice did not prevent toxin-induced oligodendrocyte loss and demyelination, but it permitted profound early remyelination due to improved phagocytosis of myelin debris by CNS macrophages and prevented disease-associated decline in motor performance. The beneficial effects of XPro1595 were absent in TNF-deficient mice and replicated in tmTNF-knockin mice, showing that tmTNF is sufficient for the maintenance of myelin and neuroprotection. These findings demonstrate that solTNF inhibits remyelination and repair in a cuprizone demyelination model and suggest that local production of solTNF in the CNS might be one reason why remyelination fails in MS. These findings also suggest that disinhibition of remyelination by selective inhibitors of solTNF that cross the BBB might represent a promising approach for treatment in progressive MS.

Mesenchymal stem cells (MSC) provide therapeutic effects in experimental CNS disease models and show promise as cell-based therapies for humans, but their modes of action are not well understood. We previously show that MSC protect rodent neurons against glutamate excitotoxicity in vitro, and in vivo in an epilepsy model. Neuroprotection is associated with reduced NMDA glutamate receptor (NMDAR) subunit expression and neuronal glutamate-induced calcium (Ca²⁺) responses, and increased expression of stem cell-associated genes. Here, to investigate whether MSC-secreted factors modulate neuronal AMPA glutamate receptors (AMPAR) and gene expression, we performed longitudinal studies of enriched mouse cortical neurons treated with MSC conditioned medium (CM). MSC CM did not alter total levels of GluR1 AMPAR subunit in neurons, but its distribution, reducing cell surface levels compared to non-treated neurons. Proportions of NeuN-positive neurons, and of GFAP- and NG2-positive glia, were equal in untreated and MSC CM-treated cultures over time suggesting that neurons, rather than differentially-expanded glia, account for the immature gene profile previously reported in MSC CM-treated cultures. Lastly, MSC CM contained measurable amounts of tumor necrosis factor (TNF) bioactivity and pre-treatment of MSC CM with the TNF inhibitor etanercept reduced its ability to protect neurons. Together these results indicate that MSC-mediated neuroprotection against glutamate excitotoxicity involves reduced NMDAR and GluR1-containing AMPAR function, and TNF-mediated neuroprotection.