INTRODUCTION
Holoprosencephaly (HPE) is understood to be a failure in the generation of midline signals that normally instruct eye field and forebrain division (Roessler et al., 1996; Belloni et al., 1996; Chiang et al., 1996; Abramyan 2019). Five principal genes: SHH, ZIC2, SIX3 ,FGF8 and FGFR1 , are the most commonly detected genetic factors responsible for HPE (Roessler et al., 2018b). While de novo FGFR1 mutations match a specific syndromic set of brain and limb findings, called Hartsfield syndrome (Simonis et al., 2013; Hong et al., 2016), mutations in the first four HPE genes fit into a more typical non-syndromic pattern in probands with a normal chromosome analysis and microarray (Brown et al., 1998; Wallis et al., 1999; Hong et al., 2018; Roessler and Muenke, 2010). Each of these five driver genes is required in a conserved developmental program during gastrulation for midline specification. The roles of several infrequent minor genes, e.g. PTCH1, DISP1, GAS1, TGIF1, DLL1 , etc. are less well understood (Roessler et al., 2018a). However, a common pathogenetic HPE mechanism leading to diminished midline hedgehog signaling links the major driver mutations into a conserved gene regulatory network (GRN, Suppl. Fig.1).
Hedgehog (Hh) proteins secreted from the vertebrate prechordal plate and ventral diencephalic midline are lipid-modified morphogens that play a central role in patterning the early forebrain primordium (Ohkubo et al., 2002; Cooper et al., 2003; Wilson and Houart, 2004; Bertrand and Dahmane, 2006; Storm et al., 2006; Blassberg et al., 2016). Shh functions by binding and inactivating Patched (e.g. human PTCH1 ,PTCH2 ), the major sterol-sensing and sterol-transporter related receptor(s) for Hh ligands, which constitutively and non-stoichiometrically suppress Smo activity by depriving this GPCR receptor of cholesterol and/or related sterol ligands needed to stabilize its active conformation (Taipale et al., 2002; Ingham et al., 2011; Briscoe and Therond, 2013; Luchetti et al., 2016; Huang et al., 2018; Hu and Song, 2019; Deshpande et al., 2019). SMO activates a classical signal cascade that ultimately regulates the post-translational state of Gli family transcription factors, which in turn bind enhancers of target genes (Huangfu and Anderson, 2006). Hh morphogen binding to Patched decreases its presence in the cilium leading to its internalization and degradation. Activation of Smo in vertebrates is correlated with its enrichment in the cilium (Huangfu and Anderson, 2005; Corbit et al., 2005; Gigante et al., 2018) as well as ultimately altering the processing of Gli transcription factors from truncated repressors to full-length activators. The role of non-classical GPCR signaling via SMO in human pathologies is less clearly understood (Arensdorf et al., 2016; Qi et al., 2019). However, the heptahelical domain is the frequent target for recurrent activating mutations in cancer and somatic tumors (Xie et al., 1998; Lam et al., 1999; Taipale et al., 2000; Ayers and Therond, 2010; Twigg et al., 2016), while variation within the remaining domains characteristic of the GPCR superfamily are less frequently observed (see Figure 1). Mutagenesis screens in Drosophila and zebrafish have clearly demonstrated that there is a functional consequence to variation in, and cooperative interactions between, all of these domains (Nakano et al., 2004; Zhao et al., 2007; Aanstad et al., 2009; Nachtergaele et al., 2013). Furthermore, recent structural studies of various vertebrate Smo molecules emphasize that both agonist and antagonist binding can alter the three-dimensional structure(s) of Smo orthologs and shift the orientation of all constituent domains (Byrne et al., 2016; Luchetti et al., 2016; Huang et al., 2016; Huang et al., 2018; Byrne et al., 2018; Zhang et al., 2018). Therefore, for completeness we chose to analyze all variants detected in our series of HPE probands.
SHH signaling regulates the transcription of genes involved in the specification of the ventral forebrain primordium, the resolution of a primordial single eye field into paired optic vesicles, as well as optic nerve and retinal development (Li et al., 1997; England et al., 2006; Sanek et al., 2009; Xavier et al., 2016). This early function ofSHH in the splitting of the eye field into two lateral optic primordia followed by induction of optic stalk tissue at the expense of neural retina is key to our understanding of HPE and the dramatic cyclopic phenotypes seen in both animal models and humans (Chiang et al., 1996; Hammerschmidt et al., 1996; Hammerschmidt and McMahon, 1998; Vinothkumar et al., 2008; Pillai-Kastoori et al., 2015). Our understanding of these morphological and gene expression changes influenced our approach to SMO bioassay development.
Here we report our functional testing in zebrafish of seven non-synonymous variants of human SMO derived from HPE molecular genomic and clinical data. Our studies reveal some of the inherent difficulties of in vivo bioassay systems that depend on a lag in experimental intervention (e.g. injection of the SMO gene) and measurements of biological responses (e.g. marker gene changes, or morphogenetic phenotypes). Given the complex GRN involved in hedgehog signaling and its feedback wiring we now demonstrate that investigators need to incorporate a clear understanding of these regulatory relationships in assay development as well as in medical genetic interpretation.