DNA methylation analysis in patients with neurodevelopmental disorders improves variant interpretation and reveals complexity

Abstract

Analysis of genomic DNA methylation by generating epigenetic signature profiles (episignatures) is increasingly being implemented in genetic diagnosis. Here we report our experience using episignature analysis to resolve both uncomplicated and complex cases of neurodevelopmental disorders (NDDs). We analyzed 97 NDDs divided into (1) a validation cohort of 59 patients with likely pathogenic/pathogenic variants characterized by a known episignature and (2) a test cohort of 38 patients harboring variants of unknown significance or unidentified variants. The expected episignature was obtained in most cases with likely pathogenic/pathogenic variants (53/59 [90%]), a revealing exception being the overlapping profile of two SMARCB1 pathogenic variants with ARID1A/B:c.6200, confirmed by the overlapping clinical features. In the test cohort, five cases showed the expected episignature, including (1) novel pathogenic variants in ARID1B and BRWD3; (2) a deletion in ATRX causing MRXFH1 X-linked mental retardation; and (3) confirmed the clinical diagnosis of Cornelia de Lange (CdL) syndrome in mutation-negative CdL patients. Episignatures analysis of the in BAF complex components revealed novel functional protein interactions and common episignatures affecting homologous residues in highly conserved paralogous proteins (SMARCA2 M856V and SMARCA4 M866V). Finally, we also found sex-dependent episignatures in X-linked disorders. Implementation of episignature profiling is still in its early days, but with increasing utilization comes increasing awareness of the capacity of this methodology to help resolve the complex challenges of genetic diagnoses.

Keywords: ARID1A; ARID1B; BAFopathy; Coffin-Siris syndrome; DNA methylation; SMARCA2; SMARCA4; SMARCB1; episignatures; neurodevelopmental disorders.

Graphical abstract

Figure 1

Novel interpretations for discordant episignatures (A) Euclidean hierarchical clustering (heatmap) (left) and MDS plot (right) from two subjects with CSS4 that harbored likely pathogenic variants in SMARCA4: GM190941 [c.3068A>G p.(E1023G) and GM223379 (c.1646G>T p.(R549L)]. In the MDS, the DNAm methylation profiles of the CSS4 samples do not cluster with the BAFopathy episignature. The detected episignature is currently undefined and the two patients uncharacterized. Case GM160708 with ARID1B:c.2480C>T p.(A827V) had a BAFopathy EpiSign, supporting the diagnosis of a rare case of CSS1 due to a missense variant in ARID1B. (B) Family tree of patient NWM-024D (II.2), the second child of healthy parents. She had a de novo PHF6:c.890C>T p.(C297F) variant, strongly suggestive of BFLS. Note the coarse and wide face, low-set ears, bitemporal narrowing, hypertelorism, prominent supraorbital ridges, prominent eyebrows, synophrys, long philtrum, carpe-shaped nose, retrognathia, short neck, and brachydactyly (photo at 12 years of age). (C) Left shows the DNAm heatmap of two patients with BFLS, NWM-024D, and GDB1321, the latter being the only other female with BFLS so far analyzed, established BFLS cases and healthy controls. Right, the MDS plot shows clustering of NWM-024D and GDB1321 with controls (green) and not with BFLS cases (blue).

Figure 2

Episignature analysis suggests a diagnosis of CdLS in unsolved cases (A and B) Cases 150163 and NWM-236D of CdLS with no variants identified in CdLS genes by ES/CMA screening. Case 150163 was initially misleading as he had an ARID1B c.3220G>A p.(D1074N) de novo missense variant. This variant has now been reported in three cases in GnomAD (v.2.1.1), further supporting its likely benign role. (C and D) Heatmaps for cases 150163 (C, left) and NWM-236D (D, right). (E and F) MDS plots for the two patients showed that 150163 did not show a BAFopathy episignature (E, left MDS), whereas both cases clustered with the CdLS profile (F, right MDS).

Figure 3

Missense variants in the DNA-binding domain of the SMARCB1 protein reveal a novel rule for the CSS_c.6200 sub-episignature (A–C) MDS plots and heatmaps for two subjects with missense variants in SMARCB1 [GM223380 c.110G>A p.(R37H) and 11013846 c.31G>A p.(G11R)] show their profiles cluster with cases with the CSS_c.6200 sub-domain episignature, found in individuals with C-terminal variants in ARID1A [c.6232G>A p.(E2078K); c.6254T>G p.(L2085R)] and ARID1B (c.6133T>C p.(C2045R).. (D) Our two cases with SMARCB1 p.(R37H) and ARID1A p.(E2078L) variants show common facial features with the SMARCB1 p.(R37H) described patients.. (E) Schematic architecture of the human BAFopathy complex. All the variants associated with the CSS_c.6200 sub-domain EpiSign profile encode for amino acids in close spatial proximity of the DNA-binding domain of the SMARCB1 protein where the R37H and G11R reside. This suggests that the CSS_c.6200 sub-domain episignature depends on a specific alteration in BAF complex function.

Figure 4

Insights into the distribution of NCBRS/BIS-causative variants (A) Missense variants in SMARCA2 cause two different syndromes, depending on their location within the protein. The schematic structure of the SMARCA2 protein (figure modified from ref.²⁹) shows the five constituent domains with variants associated with NCBRS indicated above the protein and those associated with the BIS below. NCBRS variants cluster in the helicase ATP-binding or helicase C-terminal domain, whereas BIS variants are outside these regions. (B) Pedigree of case GM173400, who is carrier of a de novo SMARCA2 c.2566A>G p.(M856V) variant. The facial gestalt of GM173400 is compatible with a BIS phenotype. (C–E) Euclidean hierarchical clustering (heatmap) and MDS plots support the clinical finding showing that GM173400 has a typical BIS episignature, and not a broad BAFopathy one (C), BIS probe set presenting case GM173400 (red), BIS cases (purple), BAFopathy cases (blue), controls (green). (D) BAFopathy probe set presenting case GM173400 (red), BIS cases (purple), CSS4_c.2656 (black), BAFopathy cases (blue), controls (green). (E) BIS probe set presenting case GM173400 (red), BIS cases (purple), CSS4_c.2656 (black), BAFopathy cases (blue), controls (green).

Figure 5

The similar episignature are exerted by homologous missense changes in SMARCA2 and SMARCA4 (A) Alignment of SMARCA2 and SMARCA4 paralogous proteins showing a 19 amino acid tract of complete amino acid identity centered on the conserved residue Met856. (B) Superimposed 3D structures of SMARCA2 (red) and SMARCA4 (blue). On the right, zoom in on the region containing SMARCA2 Met856 and SMARCA4 Met886 show they are isopositional. Several other structurally homologous amino acids of the two proteins are shown.