Gene Expression Changes in Cultured Reactive Rat Astrocyte Models and Comparison to Device-Associated Effects in the Brain

Abstract

Implanted microelectrode arrays hold immense therapeutic potential for many neurodegenerative diseases. However, a foreign body response limits long-term device performance. Recent literature supports the role of astrocytes in the response to damage to the central nervous system (CNS) and suggests that reactive astrocytes exist on a spectrum of phenotypes, from beneficial to neurotoxic. The goal of our study was to gain insight into the subtypes of reactive astrocytes responding to electrodes implanted in the brain. In this study, we tested the transcriptomic profile of two reactive astrocyte culture models (cytokine cocktail or lipopolysaccharide, LPS) utilizing RNA sequencing, which we then compared to differential gene expression surrounding devices inserted into rat motor cortex via spatial transcriptomics. We interpreted changes in the genetic expression of the culture models to that of 24 hour, 1 week and 6 week rat tissue samples at multiple distances radiating from the injury site. We found overlapping expression of up to ~250 genes between in vitro models and in vivo effects, depending on duration of implantation. Cytokine-induced cells shared more genes in common with chronically implanted tissue (≥1 week) in comparison to LPS-exposed cells. We revealed localized expression of a subset of these intersecting genes (e.g., Serping1, Chi3l1, and Cyp7b1) in regions of device-encapsulating, glial fibrillary acidic protein (GFAP)-expressing astrocytes identified with immunohistochemistry. We applied a factorization approach to assess the strength of the relationship between reactivity markers and the spatial distribution of GFAP-expressing astrocytes in vivo. We also provide lists of hundreds of differentially expressed genes between reactive culture models and untreated controls, and we observed 311 shared genes between the cytokine induced model and the LPS-reaction induced control model. Our results show that comparisons of reactive astrocyte culture models with spatial transcriptomics data can reveal new biomarkers of the foreign body response to implantable neurotechnology. These comparisons also provide a strategy to assess the development of in vitro models of the tissue response to implanted electrodes.

Spatial transcriptomics at the brain-electrode interface in rat motor cortex and the relationship to recording quality

“…What are the implications of this observation? The literature available to interpret this effect is relatively limited, with a notable exception of a recently published overview of astrocytic barriers 65 . It is possible that the glia limitans and device-associated glial encapsulation might be functioning in coordination, essentially amplifying the inflammatory response.…”

“…Chi3L1 [61,62]), neurodegeneration(Hspb1, Lcn2) [59,60,63], intracerebral haemorrhage (Mt2a [64]), and neuroinflammation (C1, C3, Spp1 [65,66])). However, while analysing the whole transcriptome can reveal new and previously unreported potential biomarkers of device-tissue interaction, detailed discussions of these individual genes in isolation fails to identify those genes with a relationship with practical metrics of interest: namely, recorded signal quality.…”

Structural, Functional, and Genetic Changes Surrounding Electrodes Implanted in the Brain

“…The observation that the structural findings emerged by 1 week post-implantation (Figures and ), while functional effects were not observed until the later 6 week time point (Figures and ), implies an additional contributor to the functional effects, potentially implicating a role of device-reactive glia in functional remodeling at the electrode site . This hypothesis is supported by our repeated observation of the upregulation of genes (e.g., C3 and Serping1 ) linked to reactive astroglial-mediated synaptic loss. ,,, Further, reactive microglia and astrocytes can release cytokines, glutamate, and adenosine triphosphate (ATP) at the injury site, which could promote hyperpolarized postsynaptic contacts via opening of potassium and chlorine channels. ,, In summary, the neuronal gene expression results reinforce observations of structural and functional changes in the neurons, while the glial gene expression results point to potential underlying mechanisms.…”

“… 29 This hypothesis is supported by our repeated observation of the upregulation of genes (e.g., C3 and Serping1 ) linked to reactive astroglial-mediated synaptic loss. 30 , 34 , 61 , 62 Further, reactive microglia and astrocytes can release cytokines, glutamate, and adenosine triphosphate (ATP) at the injury site, which could promote hyperpolarized postsynaptic contacts via opening of potassium and chlorine channels. 29 , 63 , 64 In summary, the neuronal gene expression results reinforce observations of structural and functional changes in the neurons, while the glial gene expression results point to potential underlying mechanisms.…”