A Logic Model of Neuronal-Glial Interaction Suggests Altered Homeostatic Regulation in the Perpetuation of Neuroinflammation

Abstract

Aberrant inflammatory signaling between neuronal and glial cells can develop into a persistent sickness behavior-related disorders, negatively impacting learning, memory, and neurogenesis. While there is an abundance of literature describing these interactions, there still lacks a comprehensive mathematical model describing the complex feed-forward and feedback mechanisms of neural-glial interaction. Here we compile molecular and cellular signaling information from various studies and reviews in the literature to create a logically-consistent, theoretical model of neural-glial interaction in the brain to explore the role of neuron-glia homeostatic regulation in the perpetuation of neuroinflammation. Logic rules are applied to this connectivity diagram to predict the system's homeostatic behavior. We validate our model predicted homeostatic profiles against RNAseq gene expression profiles in a mouse model of stress primed neuroinflammation. A meta-analysis was used to calculate the significance of similarity between the inflammatory profiles of mice exposed to diisopropyl fluorophostphate (DFP) [with and without prior priming by the glucocorticoid stress hormone corticosterone (CORT)], with the equilibrium states predicted by the model, and to provide estimates of the degree of the neuroinflammatory response. Beyond normal homeostatic regulation, our model predicts an alternate self-perpetuating condition consistent with chronic neuroinflammation. RNAseq gene expression profiles from the cortex of mice exposed to DFP and CORT+DFP align with this predicted state of neuroinflammation, whereas the alignment to CORT alone was negligible. Simulations of putative treatment strategies post-exposure were shown to be theoretically capable of returning the system to a state of typically healthy regulation with broad-acting anti-inflammatory agents showing the highest probability of success. The results support a role for the brain's own homeostatic drive in perpetuating the chronic neuroinflammation associated with exposure to the organophosphate DFP, with and without CORT priming. The deviation of illness profiles from exact model predictions suggests the presence of additional factors or of lasting changes to the brain's regulatory circuitry specific to each exposure.

Keywords: homeostatic regulation; logical modeling; mouse models; neural glial interaction; neuroinflammation; regulatory biology; systems neuroscience; treatment course prediction.

Figure 1

Simple neuroimmune signaling network. Connections with green arrow terminators represent stimulatory effects while connections with red circle terminators represent inhibitory effects.

Figure 2

Steady states predicted by the neuro-immune model. White—nominal state (0); Green—high state (1); Red—low state (−1). The steady states are named according to their profile: Typical Health, and Neuroinflammation. Typical Health state is shown for reference.

Figure 3

Alignment of exposure conditions compared to saline controls with model predicted stable states. Sammon projection in two dimensions for gene expression obtained in cortex. Black dots represent the model predicted stable states (SS0, and SS1). Red dots represent the aggregated gene expression data for CORT, DFP, and CORT + DFP exposure conditions vs. saline controls. Axes represent arbitrary units such that the relative distance between points approximates the aggregated P-values between all points.

Figure 4

Gene expression changes in mouse cortex for exposure conditions compared to control. Fold change in marker gene expression levels from RNAseq of samples collected at 6 h relative to resting saline control for data obtained in cortex for the subset of 12 markers represented in the logic circuit model.

Figure 5

Simulation of putative treatment strategies. Percent of Monte Carlo simulations of putative treatment strategies that return the neuroimmune system to typically healthy regulation. GRB—glucocorticoid receptor blockade; IST—immunosuppressive therapy; BAA—broad acting anti-inflammatory. Transition to yellow indicates point of greater than 50% of trials returning to health.