Spatially Resolved Proteomic and Transcriptomic Analysis of the Neuroinflammatory Response to Intracortical Microelectrode Implantation

Introduction: Intracortical microelectrodes (IMEs) are devices implanted into the cortex of the brain that can record neural signals, which can be utilized in brain computer interfaces (BCIs) to operate devices such as prostheses or wheelchairs. However, IMEs currently fail at chronic timepoints, largely attributed to the neuroinflammatory response triggered post-implantation. Implantation breaks the blood brain barrier (BBB), damages neurons, and activates astrocytes and microglia, leading to formation of a glial scar as well as neurodegeneration. This causes recording quality to drastically decrease as well as causes damage to the brain. There is currently no successful method for alleviating post-implantation inflammation, in part because the nuances of the biological response are not yet fully understood.
Protein expression at the IME-tissue interface has been extensively evaluated with immunohistochemistry; however, this is limited by the number of proteins that can be quantified at once, usually single digits. mRNA expression in IME-implanted tissue has also been investigated, but due to post-transcriptional and post-translational modification, the ratio of mRNA to resulting protein is not one-to-one.
This project analyzes the differential expression of 62 neuroinflammatory proteins as well as the full mRNA transcriptome to gain a more complete picture of the biological response to IME implantation. Tissue was segmented into 0-90µm, 90-180µm, and 180-270µm from the implant to assess how protein and mRNA expression is affected by distance from the implant site. Through this multi-omics analysis, we aim to reveal biological pathways in IME-implanted tissue and illuminate potential therapeutic targets for improved IME performance.

Materials and

Methods: Male C57BL/6J were implanted with non-functional single-shank probes and sacrificed after 4 and 8 weeks via cardiac perfusion. The brains were extracted, cryosectioned into 5µm slices, and mounted on slides.
For proteomic analysis, slides were stained for GFAP for activated astrocytes, NeuN for neuronal nuclei, and fluorescent tags for proteins of interest. A custom ImageJ code was used to segment the tissue around the implant site into three 90µm wide concentric rings. The NanoString GeoMx and nCounter MAX/FLEX system was used to process the slides and count the fluorescent tags per ring. Counts were exported and a custom MATLAB code was used to normalize counts to housekeeping genes, calculate log2(fold change), and perform an unpaired two-tailed t-test for equal variance. A Benjamini-Hochberg correction with a false discovery rate of 0.05 was applied to account for random significance.
For transcriptomic analysis, slides were hybridized with Whole Transcriptome Atlas (WTA) probes for 21,000 protein-coding mRNA, followed by staining for GFAP and NeuN. A custom ImageJ code segmented the tissue, the GeoMx processed the slides, and the WTA probes were sequenced. The GeoMx was used to normalize counts, calculate log2(fold change), and perform a t-test similar to proteomics. No correction factor was applied, but significance threshold was adjusted to p=0.04 to account for random significance. Advaita’s iPathwayGuide was used to analyze transcriptomic data, with a focus on Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis.

Results, Conclusions, and Discussions: Our 4 week proteomics results is primarily characterized by the immune response, with upregulation of nineteen immune proteins in the inner ring, four in the middle, and two in the outer. The middle and outer rings also show downregulation of several immune proteins, potentially because immune proteins/cells are migrating to the inner ring where damage is greatest.
At 8 weeks, there is continued upregulation of eighteen immune proteins in the inner ring, three in the middle, and none in the outer, and no more immune protein downregulation. In all three rings, there is downregulation of autophagy and neuronal health proteins. Downregulation of autophagy, when autophagosomes clear out damaged and unnecessary cell parts, could indicate failure to maintain cellular health, resulting in apoptosis instead. Regarding neuronal health, while it has been previously reported that neuronal damage is contained within 50-200µm of the implant, our results show downregulation of 6 neuronal health proteins across all three rings, up to 270µm.
Our transcriptomics analysis of 21,000 mRNA reveals pathways not covered by the 62 protein panel. Specifically, many pathways for metabolism were differentially expressed in the 4 week inner and middle rings, indicating that IME implantation affects methods of energy conversion. The only KEGG pathway significant in all three rings was mRNA surveillance (KEGG:03015), a process that detects and breaks down faulty mRNA. Increased mRNA surveillance suggests cellular damage causing defective transcriptional processes.
At 8 weeks, pathways for cell signaling and death are differentially expressed in the inner ring, and neurodegeneration (KEGG:05022) is differentially expressed in all three rings. These results correlate with the 8 week downregulation of neuronal health proteins, with specifically severe damage in the inner ring.
Overall, limitations in the quantity of proteins that can be simultaneously analyzed can be supplemented with transcriptomics, which is capable of much higher throughput. However, since mRNA may not accurately predict protein expression, examining pathways instead of just individual genes is important. Our results reveal biological changes to IME implantation, including in the immune response, autophagy, neuronal health, metabolism, and transcription, informing future work in selection of molecular targets for improved IME performance.

Acknowledgements (Optional):