A restricted spectrum of missense KMT2D variants cause a multiple malformations disorder distinct from Kabuki syndrome

Abstract

Purpose: To investigate if specific exon 38 or 39 KMT2D missense variants (MVs) cause a condition distinct from Kabuki syndrome type 1 (KS1).

Methods: Multiple individuals, with MVs in exons 38 or 39 of KMT2D that encode a highly conserved region of 54 amino acids flanked by Val3527 and Lys3583, were identified and phenotyped. Functional tests were performed to study their pathogenicity and understand the disease mechanism.

Results: The consistent clinical features of the affected individuals, from seven unrelated families, included choanal atresia, athelia or hypoplastic nipples, branchial sinus abnormalities, neck pits, lacrimal duct anomalies, hearing loss, external ear malformations, and thyroid abnormalities. None of the individuals had intellectual disability. The frequency of clinical features, objective software-based facial analysis metrics, and genome-wide peripheral blood DNA methylation patterns in these patients were significantly different from that of KS1. Circular dichroism spectroscopy indicated that these MVs perturb KMT2D secondary structure through an increased disordered to ɑ-helical transition.

Conclusion: KMT2D MVs located in a specific region spanning exons 38 and 39 and affecting highly conserved residues cause a novel multiple malformations syndrome distinct from KS1. Unlike KMT2D haploinsufficiency in KS1, these MVs likely result in disease through a dominant negative mechanism.

Keywords: KMT2D; Kabuki syndrome; histone 3 lysine 4 methyltransferase; intrinsically disordered region; multiple congenital anomaly.

Fig. 1. Affected individuals have missense variants in parts of exons 38 or 39 of KMT2D.

We studied multiple affected individuals from seven families with missense KMT2D variants restricted to a region that encodes for 54 amino acids flanked by Val3527 and Lys3583 (ENST00000301067.7; NM_003482.3). (a) Schematic representation of KMT2D exons with each alternating exon represented in dark or light red shade (introns are not depicted). (b) Frequency of KMT2D missense variants (from gnomAD) in the general population is shown in yellow. Deeper troughs represent higher frequency at that particular location. (c) Green lollipop graph denoting the missense KMT2D variants in individuals with Kabuki syndrome from the published literature. The y-axis in this graph represents the frequency of the variant in the published literature. The x-axis is a schematic for the protein denoting the location of important domains and regions of KMT2D. Note that variants identified in this study are located in parts of exons 38 and 39 with high missense constraint but without any variants in individuals with Kabuki syndrome. (d) The region of interest of the KMT2D gene and protein in more detail. The red horizontal bar shows parts of exons 38 (amino acid 3503–3580) and 39 (amino acid 3581–4510). The blue vertical bars denote the coiled-coil regions. Red lines indicate the location of the variants identified in this study. Pedigree of each family is shown under the corresponding variant. Standard symbols are used to denote affected (filled symbols) and unaffected (unfilled) individuals. All individuals who were tested but not found to carry familial KMT2D variants are denoted by “N”. Father in family 4 (F4; I:1) was found to be likely mosaic and is denoted by gray square. In this family, genetic testing was not possible for the first born child (F4; II:1) but is shown as affected based on the clinical history.

Fig. 2. Missense KMT2D variants described in this study result in phenotype distinct from type 1 Kabuki and CHARGE syndromes.

(a) Photographs of individuals described here with missense KMT2D variants. Note the wide range of facial features. P2, proband from family 2, is shown at two different ages. Note facial asymmetry, hypertelorism, bilateral epicanthic folds, bulbous nasal tip, downturned corners of the mouth, microtia, and hypoplastic nipples. P3, proband from family 3, has a box-shaped head, bilateral microphthalmia, severely hypoplastic left pinna, and ectopic left external auditory canal. P5 and P7, probands from families 5 and 7 respectively, have prominent forehead, broad nasal root, flat midface, and thin upper lip. P7’s eyebrows are laterally flared. One individual with Kabuki syndrome type 1 (KS1) is shown for comparison. Note arched eyebrows, long palpebral fissures, eversion of the lateral part of the lower eyelids, large cupped ears, short columella, bulbous nasal tip, and pillowed lower lip. In individual with CHARGE syndrome note hypertelorism, bulbous and large nasal tip, and a repaired cleft lip. (b) Computerized tomography (CT) (i and ii) and T2-weighted magnetic resonance imaging (MRI) (iii) of P2 demonstrating absence of the posterior part of the semicircular canals (red arrows) and normal anatomy of the lateral and anterior semicircular canals. Brain MRI of P3 (iv) to show well-formed right and left middle ear cavities, with the right cavity being smaller than the left. Bilaterally the cochlea, semicircular canals, and inner auditory canals appear normal. T1-weighted MRI to demonstrate a small left optic globe with ballooning of the optic disc bilaterally (red arrows) and optic disc colobomata (v). CT imaging of P5 (vi, vii) demonstrating the presence of cysts in the lower jaw (red arrows). (c) Face2Gene analysis with confusion matrix showing that the system is able to predict correctly each group with a mean accuracy of 75.44%. (d) Receiver operating characteristic (ROC) graphs show the probability curve where the area under the curve (AUC) (0–1) represents the measure of separability between two groups. Score distributions show the distribution of those probabilities. The higher the AUC, the better the model is at distinguishing between two groups. (e) Principal component analysis (PCA) shows the four groups analyzed in the DNA methylation array clustering separately (p < 0.001). Importantly, the samples from individuals described in this study cluster together and separate from those with type 1 Kabuki syndrome. MV missense variant.

Fig. 3. Missense variants described in this study perturb protein secondary structure in KMT2D recombinant proteins.

(a) Central, highly conserved region of KMT2D containing a coiled-coil domains predicted by MARCOIL (blue trace). All KMT2D missense variants described here (red lines) occur within or close to the predicted coiled-coil domain (residues 3562–3614). KMT2D fusion proteins (residues 3503–3600; gray bar) contained the missense variants described here. (b) Heptad net plot view of a potential coiled-coil domain, predicted by MARCOIL between residues 3511 and 3559 of KMT2D, showing approximately seven heptad repeats of hydrophobic or nonpolar residues (black letters) and charged residues (blue or red letters). Missense variants of residues within or close to the predicted coiled-coil domain are indicated by red boxes. Positions of residues within the predicted heptad repeat sequences are labeled a to g. Residues at the “a” and “d” positions (gray boxes), which include Leu3528 and Leu3542, form the hydrophobic seam in a coiled coil. (c) Upper panel: expression of wild-type and mutant recombinant KMT2D fusion proteins, as indicated. Lower panel: circular dichroism (CD) spectroscopy traces of recombinant KMT2D wild-type and mutant proteins showing moderate levels of disordered secondary structure in the wild-type protein (black trace) with perturbed secondary structure and higher proportions of ɑ-helical structure in all recombinant KMT2D proteins with missense variants (purple, red, light blue, and green traces).