Over the past decade, there has been rapid progress in identifying putative risk genes for autism spectrum disorder (ASD) and related neurodevelopmental disorders [Abrahams & Geschwind, 2008; Bill & Geschwind, 2009; El-Fishawy & State, 2010]. Thus far, however, much less is known about the mechanisms through which these genes impact brain development and, when disrupted, how they contribute to disorder pathogenesis. We identified a common functional polymorphism in the MET gene promoter, which leads to decreased transcription of MET and significantly increases the risk for developing ASD [Campbell et al., 2006]. This finding has been replicated by our own [Campbell, Li, Sutcliffe, Persico, & Levitt, 2008] and other groups [Jackson et al., 2009; Sousa et al., 2009]. The subsequent identification of copy number variants [Marshall et al., 2008] and rare functional mutations [Campbell et al., 2006] in the MET gene lends further support to the idea that alterations in MET signaling contribute to ASD risk. Consistent with these genetic data, MET transcript and MET protein expression are decreased in the temporal cortex of ASD cases compared to controls [Campbell et al., 2007].
MET is a receptor tyrosine kinase, which, when activated, can induce multiple cellular responses, including proliferation, migration, differentiation and survival, depending on the cell and environmental context [for e.g., Beilmann, Vande Woude, Dienes, & Schirmacher, 2000; Birchmeier, Birchmeier, Gherardi, & Vande Woude, 2003; Bladt, Riethmacher, Isenmann, Aguzzi, & Birchmeier, 1995; Ebens et al., 1996; Giacobini et al., 2007; Okunishi et al., 2005]. Binding of the only known endogenous ligand, hepatocyte growth factor (HGF), results in MET receptor dimerization and autophosphorylation that ultimately activates key intracellular signaling pathways, such as the PI3-kinase and ERK systems [Longati, Bardelli, Ponzetto, Naldini, & Comoglio, 1994; Ponzetto et al., 1994; Stefan et al., 2001; Xiao et al., 2001]. HGF is secreted as a single-chain pro-form that is devoid of signaling activity; pro-HGF requires proteolytic cleavage by a serine protease to acquire biological activity [Kirchhofer et al., 2004; Lokker et al., 1992]. Several such proteases have been reported to activate HGF in vitro, including the urokinase-like plasminogen activator (uPA), whose potency is increased when bound to the urokinase-like plasminogen activator receptor (uPAR) [Blasi, 1993; Mars, Zarnegar, & Michalopoulos, 1993; Naldini et al., 1992]. A recent study, however, suggests that uPA/uPAR may not contribute to HGF activation under physiological conditions [Owen et al., 2010]. Thus, our original hypothesis [Powell, Mars, & Levitt, 2001] connecting the bioavailability of activated HGF ligand through an intact uPA and uPAR system, which in turn modulates MET signaling in vivo, may not be accurate. Yet both uPA and uPAR are expressed in the forebrain [Bahi, Boyer, Kafri, & Dreyer, 2006; Del Bigio, Hosain, & Altumbabic, 1999; Dent, Sumi, Morris, & Seeley, 1993; Lahtinen et al., 2009; Masos & Miskin, 1996; Yoshida & Shiosaka, 1999] and genetic disruption leads to neurodevelopment perturbations [Del Bigio et al., 1999; Eagleson et al., 2010; Eagleson, Bonnin, & Levitt, 2005; Meiri, Masos, Rosenblum, Miskin, & Dudai, 1994; Powell et al., 2001, 2003a]. Moreover, functional polymorphisms in the plasminogen activator, urokinase receptor (PLAUR) gene, the human homolog of the mouse uPAR gene, like those found in MET, increase the risk for ASD [Campbell et al., 2008].
Animal model systems provide a useful experimental tool for investigating potential links between ASD susceptibility genes and alterations in brain architecture, including insights into spatial and temporal dynamics of altered developmental trajectories. In vitro model systems are also used to address important issues regarding the impact of altered signaling and/or environmental conditions on basic cellular processes. Our initial attempts to delineate the role of Met in cortical development utilized an in vitro model system. We demonstrated that HGF stimulates the migration of GABAergic interneurons that arise from cultures of explants of the ventral telencephalon (VTel, [Powell et al., 2001]), presumably through the activation of the Met receptor that is expressed on migrating VTel neurons in vitro. In vivo, consistent with our initial proposal, there is a disruption of cortical interneuron development in a region- and cell-type specific manner in adult uPAR−/− mice [Eagleson et al., 2005; Powell et al., 2001, 2003a]. Specifically, frontal and parietal cortical areas exhibit reduced numbers of parvalbumin (PV)-positive GABAergic neurons, while in the hippocampus, there is a loss of the somatostatin population of interneurons in the CA1 and dentate gyrus subfields. Based on these observations, we suggested that alterations in Met signaling directly impact interneuron migration, leading to alterations in the interneuron profile in the adult cortex.
More recent observations suggest a discrepancy between our initial reports and hypotheses regarding direct action of Met signaling on interneuron migration in the developing mouse forebrain. For example, detailed molecular neuroanatomical analysis of gene and protein expression indicates that the Met receptor protein is expressed at low levels in vivo in the developing VTel that includes the proliferative ganglionic eminences (GE) and the postmitotic neurons of the developing striatum, and that detection of protein by the western blot method is due to receptor transport by neurons projecting from the cortex into the striatum [Judson, Bergman, Campbell, Eagleson, & Levitt, 2009]. Furthermore, a recent study reported normal numbers of GABAergic neurons in the hippocampus when Met is deleted in cells arising from the GE, suggesting that interneuron migration to the hippocampus is unaffected in the absence of Met [Martins, Plachez, & Powell, 2007]. Finally, a recent report suggests that pro-HGF is not activated by uPA under biologically relevant conditions [Owen et al., 2010]. In the current study, we use conditional disruption of Met signaling to examine directly the influence of Met function on neocortical interneuron development. We also use radiolabel and fluorescent double-label in situ hybridization to clarify Met expression by cortical interneurons throughout key prenatal periods of development. In addition, we examine the influence of tissue culture environments on Met receptor expression to decipher possible differences in the roles for Met in vitro and in vivo.