Molly Godfrey, Michael A Levy, Christopher Campbell, Leigh Demain, Sarah Jenkinson, Sarah Hilton, Berta Almoguera, Meena Balasubramanian, Emilia K Bijlsma, Fiona Blanco-Kelly, Emma M M Burkitt Wright, Gerarda Cappuccio, Kate E Chandler, Koen Devriendt, Aranzazu Diaz de Bustamante, Maria K Haanpää, Carolin Hörnig, Elizabeth A Jones, Sinem Kocagil, Hannele Koillinen, Dhanya Lakshmi Narayanan, Emanuela Leonardi, Rajesh Madhu, Purvi Majethia, Alessandra Murgia, Elisabeth Rosser, Markus Schuelke, Anju Shukla, Emma Soengas-Gonda, Sarah Stewart, Yves Sznajer, Saoud Tahsin Swafiri, Maria Margarida Venancio, Renzo Guerrini, Martino Montomoli, Annalisa Vetro, Thomas Wright, David Gokhale, Bekim Sadikovic, Sofia Douzgou Houge, Siddharth Banka
Pathogenic CHD8 variants cause autosomal dominant 'Intellectual developmental disorder with autism and macrocephaly' (IDDAM) and are amongst the most common monogenic causes of autism. The clinical significance of CHD8 missense variants (MVs) frequently remains uncertain. Systematically applying ACGS/ACMG guidelines to 36 CHD8 MVs in 39 affected patients, only two variants were classified likely pathogenic (LP), with the remaining 34 classified variants of uncertain significance (VUS). We subclassified the variants according to posterior probability of pathogenicity (PPP), with 14 being at least tepid VUS (PPP ≥ 50%). Comprehensive phenotypic analysis revealed no discernible clinical differences between individuals carrying at least tepid VUS and others, offering no additional insight for variant classification. EpiSign™ testing revealed a CHD8-IDDAM episignature, as previously detected in patients with truncating/null variants, in 11 cases, allowing reclassification of 8 VUS (all previously classified at least tepid) as LP. Molecular modelling indicated that disease-causing (LP/P) CHD8 MVs are concentrated in structured and/or functional protein domains. Compared to truncating/null variants, disease-causing MVs were less often associated with attention issues and macrocephaly, but clinical features were otherwise similar. Additionally, three disease-causing MVs were inherited from unaffected/mildly affected parents. Overall, we show that determining the clinical significance of CHD8 MVs is challenging, even with detailed clinical information, but that incorporating episignature analysis increases diagnostic yield. Further, our results indicate that CHD8 MVs are likely to act via a loss-of-or reduced function mechanism. These findings reveal the importance and complexity of interpreting CHD8 MVs and will improve the diagnosis and understanding of CHD8-related disorders.