Mathematical modeling of reactive gliosis using in silico Bayesian method

AUZMENDI J, ROSCIZEWSKI G, MOFFATT L, RAMOS AJ

Congreso; XXX Congreso de la Sociedad Argentina de Investigacion en Neurociencias (SAN),; 2015

Reactive astrogliosis (RA) is a general, graded, glial response to brain injury. Reactiveastrocytes are characterized by an increase of the cellular volume and secretion of proinflammatorymolecules. Until now it is unknown which are the signals thatinitiate/propagate RA, but it is proposed that diffusion of damage associated moleculespattern (DAMP) from the necrotic core; calcium waves propagated by glial gap junctions ordramatic changes in extracellular milieu (i.e. ATP levels) are likely to be involved. We hereevaluated two main paradigms to understand the progression of RA using mathematicalmodeling. Model I is a simple mechanism that contemplates the diffusion of DAMP fromthe necrotic core as the main responsible for the RA. Model II is more complex mechanisminvolves a subsequent signal carried out by soluble mediators secreted by the proximalastrocytes acting as amplifiers of the signal. Starting with GFAP-stained brain sections ofanimals subjected to cortical devascularization analyzed by Sholl analysis and confocal 3Dreconstruction, we studied the morphology of the cortical astrocytes at 3 and 7 days postischemic injury. Then, we applied a Bayesian Computational Modeling approach by buildinga parametric mathematical model using partial differential equation for diffusion of DAMPand soluble mediators coupled with a Markovian Model for the signal-triggered RA. Finally,we tested both models and found strong evidence for model II with soluble mediators.