An Expanding Spectrum of Autism Models

Abstract

In this review, we contrast previous models of autism pathogenesis with newer models inspired by some recently appreciated and previously minimally considered pathological and clinical features of the disease. Autism has conventionally been viewed as an incurable behavioral disorder resulting solely from genetic defects impacting brain development. However, emerging evidence suggests that autism affects many organ systems beyond the brain and that some neuropathological and somatic pathophysiological processes are active even into adulthood. We incorporate these newer observations into a model of how this systemic and persistent disease process might impact brain function and ultimately impair behavior through potentially reversible mechanisms. In particular, observations of substantial transient and sometimes enduring increases in function and even losses of the diagnosis challenge researchers to identify pathophysiological mechanisms consistent with this dynamic course and potential plasticity. This broadened appreciation of the disease phenomenology and prognosis of autism calls for mechanistic models that encompass the full range of its features. To this end, our review contrasts several models of autism pathophysiology and lays out their differing underlying assumptions regarding mechanisms. First, we compare models of autism that are based on different underlying biological and experimental perspectives to addressing the question of autism pathogenesis. We contrast a “bottom-up, modular, genes–brain–behavior” model with a more inclusive “middle-out, multi-system biology” model. We then contrast different models that consider autism’s development over time. Beginning with a purely genetic prenatal brain development model, we expand the framework to include early environmental influences, epigenetics, later and ongoing environmental influences, and features consistent with chronic encephalopathy. The implications of these models, particularly the last, are spelled out through a discussion of the functional consequences of one prominent chronic feature, persistent immune activation, and its impact on neural–glial interactions. The implications of these newer models on potential treatments are also discussed.