Behavioural and neurobiological changes mediated by TNF-α signalling in the CNS

Background: TNF-α is a pleiotropic cytokine having a diverse range of functions extending from the immune system to the central nervous system (CNS). It is recruited into the CNS through the blood brain barrier and also produced by astrocytes and microglia, in the CNS. A majority of its actions are governed by its main receptors: TNF-R1 (expressed by nearly all cell types) and TNF-R2 (expressed by immune and endothelial cells). The general consensus over these receptors has been that TNF-R1 is primarily degenerative due to the presence of a death domain; while TNF-R2 is more neuroprotective due to the lack of said death domain. There is active involvement of these members of the TNF-α superfamily in mediating neurobiology and behaviour in the CNS; TNF-α mediates synaptic scaling and facilitates expression of neurotrophins, brain derived neurotrophic factor (BDNF) and nerve growth factor (NGF). However in peripheral or central inflammation, an up-regulation of TNF-α upsets this homeostatic balance triggering mechanistic changes in the CNS that can lead to behavioural impairments. Indeed TNF-α is a known biomarker of Alzheimer's disease, multiple sclerosis and major depressive disorder. It is also an active biomarker of rheumatoid arthritis and psoriasis that are comorbid with psychiatric symptoms.\ud \ud These factors make the need for understanding TNF-α neurobiology and developing targets to modulate TNF-α expression in inflammatory conditions essential. This body of work was therefore conducted to test the behavioural and neurobiological effects of genetic deletion and pharmacological blockade of TNF-α and its receptors in young and mid-adult mice.\ud \ud Methods: This body of work was carried out in three parts:\ud \ud Chapters 3 and 4 were run using similar methodologies in mice of different ages. Mice deficient for TNF-α (TNF⁻/⁻) and each of its receptors, TNF-R1⁻/⁻ and TNF-R2 ⁻/⁻ (n = 14/strain for both studies) were established on a C57BL/6 background. These mice were used to test the consequences of targeting TNF-α and its receptors on behaviour and neurobiology in young and mid-adult mice compared to C57BL/6 mice (WT: n = 21/chapter 3, n = 14/chapter 4). Mice underwent behavioural testing to measure locomotion, cognition-like, social, and emotion-like behaviour. Hippocampal tissue was collected from these mice to measure levels of NGF and BDNF by ELISA and immunohistochemistry performed to test for neurogenesis (chapters 3 and 4) and microglial numbers (chapter 4).\ud \ud For the third part of this study, only C57/BL6 mice were used and were divided in to 4 treatment groups. 2 groups were challenged with an acute dose of LPS (n = 20; IP administration) while 2 groups received saline (n = 20; IP administration). One group of LPS mice received etanercept via ICV administration (n = 10; LPS-Etan) and a control group received aCSF also via ICV (n = 10; LPS-aCSF). Similarly, one group of saline treated mice (n = 10) received etanercept through ICV (Sal-Etan) while the other group (n =10) received aCSF also through ICV (Sal-aCSF). 24 hr post-drug treatment, behaviour of mice was tested and hippocampal tissue was collected to measure the number of microglia and astrocytes in the dentate gyrus.\ud \ud Results: 3 months old TNF⁻/⁻ mice had impaired cognition, while TNF-R2⁻/⁻ mice showed good memory. All strains of mice had normal social behaviour and knockout strains had lower anxiety than WT mice. Analysis of neurotrophin levels revealed TNF⁻/⁻ and TNF-R2⁻/⁻ mice to have significantly lower levels of NGF compared to WT mice, suggesting that some of the behavioural changes could be linked to changes in levels of circulating neurotrophins.\ud \ud At 6 months age, TNF⁻/⁻ and TNF-α receptor knockout mice displayed similar learning to WT mice, but TNF-R1⁻/⁻ mice displayed better memory than all strains of mice. TNF⁻/⁻ mice had impaired social behaviour and exploratory behaviour in TNF⁻/⁻ and TNF-R1⁻/⁻ mice was decreased compared to WT mice. It is also interesting to note that both TNF⁻/⁻ and TNF-R2⁻/⁻ mice exhibited lower depression-like behaviour than WT mice. Additionally TNF⁻/⁻ and TNF-R1⁻/⁻ mice expressed significantly lower levels of BDNF than WT mice and TNF-R2⁻/⁻ mice displayed significantly higher levels of NGF compared to all strains of mice.\ud \ud Pharmacologically peripheral administration of LPS reduced locomotor activity, impaired cognition-like and increased anxiety-like behaviour. This change was associated with an increase in microglial numbers in the dentate gyrus. Centrally administered etanercept was able to significantly improve LPS induced anxiety-like behaviour, and prevent the increase of microglia, but had no significant effect on cognition-like behaviour.\ud \ud Discussion: Taken together, these findings demonstrate that TNF-α plays a complex role in mediating behavioural phenotypes. This role is further influenced by the inflammatory state and life stage of the individual.\ud \ud Specifically while normal signalling of TNF-α and its receptors is required for cognition in young mice, the absence of TNF-α and its receptors may improve cognition from mid-adulthood. So also from mid-adulthood, lack of TNF-α and its receptors may protect mice from depression-like behaviour, hinting at a possible target to counter depression in older individuals. Despite these positive features of lack of TNF-α signalling in mid-adulthood, it appears that TNF-α is required throughout a lifespan to maintain social and exploratory behaviour. This suggests that modulation of the cytokine rather than blockade may be the direction to follow.\ud \ud From a pharmacological perspective, peripheral LPS is effective in triggering behavioural and neurobiological changes as there appears to be an active exchange between the peripheral immune system and CNS. Furthermore, etanercept treatment appeared effective in reducing not only anxiety-like behaviour but also number of microglial cells induced by LPS, showing that even with non-specific inflammation, TNF-α is an active participant in CNS pathology. This reinforces the need for targeting TNF-α in inflammatory disorders with psychiatric disruptions.\ud \ud These results set a foundation to understand the complex role of TNF-α in mediating behavioural phenotypes over the course of an individual's life. Likewise TNF-α modulation appears to be the key to alleviating psychiatric conditions enhanced by this cytokine. Moving onwards, future work needs to look at the effects of targeting TNF-α and its receptors in aged animals to understand the consequences on behaviour and neurobiology specifically neurotrophin production and glial cell activity