Cytokine expression patterns predict suppression of vulnerable neural circuits in a mouse model of Alzheimer’s disease

Introduction: Alzheimer’s disease is a neurodegenerative disorder characterized by progressive amyloid plaque accumulation, tau tangle formation, neuroimmune dysregulation, synapse an neuron loss, and changes in neural circuit activation that lead to cognitive decline and dementia. Early molecular and cellular disease-instigating events occur 20 or more years prior to presentation of symptoms, making them difficult to study, and for many years amyloid-β, the aggregating peptide seeding amyloid plaques, was thought to be the toxic factor responsible for cognitive deficit. However, strategies targeting amyloid-β aggregation and deposition have largely failed to produce safe and effective therapies, and amyloid plaque levels poorly correlate with cognitive outcomes. Despite these failures, a role still exists for amyloid-β in the variation in an individual’s immune response to early, soluble forms of aggregates, and the downstream consequences of this immune response for aberrant cellular behaviors and creation of a detrimental tissue environment that harms neuron health and causes changes in neural circuit activation.

Materials and

Methods: Here, we perform resting state functional magnetic resonance imaging of awake, unanesthetized Alzheimer’s disease (5xFAD) mice to map changes in functional connectivity over the course of disease progression, in comparison to wild-type littermates. In these same individual animals, we spatiotemporally profile the immune milieu by measuring cytokines, chemokines, and growth factors across various brain regions and over the course of disease progression from pre-pathology through established cognitive deficit, using bead-based multiplexing immunoassays. We use graph theoretical analyses to identify vulnerable neural circuits in Alzheimer’s disease vs. wild-type mice, as well as decline in global brain communication with advancing disease. We quantify spatiotemporal changes in immune activation across brain regions and with advancing disease by creating regional molecular signatures at multiple timepoints using the supervised machine learning tool partial least squares.

Results, Conclusions, and Discussions: We find that Alzheimer’s disease mice first exhibit local suppression of intra-hippocampal functional connectivity before the hippocampus loses inter-regional connectivity to thalamus and cortical regions. Global brain networks become weaker, less integrated, and less efficient with disease progression, as evidenced by graph theoretical metrics of network structure (Fig. 1). At the same time, we find that advancing molecular signatures indicating microglial activation and chemotaxis progress from the hippocampus to the parietal (Fig. 2) and frontal cortices, and last to the temporal cortex, following patterns of amyloid deposition and corresponding to a period of hyperactivity and then decreases in intra- and inter-regional functional connectivity in disease-relevant brain regions. Microglia actively modulate neuronal activity, and previous evidence in the literature suggests that their transition between cellular sub-states may negatively impact this crucial part of their homeostatic function. In previous works, we and others have found microglia-neuron communications to be dysregulated in Alzheimer’s disease. Our future studies will aim to establish causal effects, which may be bi-directional, of the distribution of local microglial cellular states on spatiotemporal regulation of vulnerable neural circuits in Alzheimer’s disease mice.

Acknowledgements (Optional): This work was supported by R21AG068532 from the National Institute on Aging (EAP/NZ), grant TSF41000095613 from the Pennsylvania Department of Health using Tobacco CURE Funds (EAP/NZ), and start-up funds from Penn State College of Medicine Departments of Neurosurgery and Pharmacology (EAP). DCC and MKK were supported in part by training fellowship T32NS115667 from the National Neurological Disorders and Stroke. All funding agencies disclaim responsibility for any analyses, interpretations, or conclusions.
The co-authors would like to acknowledge the Huck Institutes’ High Field Magnetic Resonance Imaging Core Facility (RRID:SCR_024461) and facility director Thomas Neuberger for helpful discussions in design of the imaging protocol.
The authors acknowledge help and advice from Samuel Cramer in restrainer design and testing, Said Hayreddin in building the surface imaging coil, Nikki Beloate in design of imaging post-processing procedure, and Kevin Turner in training for the headpost animal surgery.