Consolidation of the clinical and genetic definition of a SOX4- related neurodevelopmental syndrome

Abstract

Background: A neurodevelopmental syndrome was recently reported in four patients with SOX4 heterozygous missense variants in the high-mobility-group (HMG) DNA-binding domain. The present study aimed to consolidate clinical and genetic knowledge of this syndrome.

Methods: We newly identified 17 patients with SOX4 variants, predicted variant pathogenicity using in silico tests and in vitro functional assays and analysed the patients' phenotypes.

Results: All variants were novel, distinct and heterozygous. Seven HMG-domain missense and five stop-gain variants were classified as pathogenic or likely pathogenic variant (L/PV) as they precluded SOX4 transcriptional activity in vitro. Five HMG-domain and non-HMG-domain missense variants were classified as of uncertain significance (VUS) due to negative results from functional tests. When known, inheritance was de novo or from a mosaic unaffected or non-mosaic affected parent for patients with L/PV, and from a non-mosaic asymptomatic or affected parent for patients with VUS. All patients had neurodevelopmental, neurological and dysmorphic features, and at least one cardiovascular, ophthalmological, musculoskeletal or other somatic anomaly. Patients with L/PV were overall more affected than patients with VUS. They resembled patients with other neurodevelopmental diseases, including the SOX11-related and Coffin-Siris (CSS) syndromes, but lacked the most specific features of CSS.

Conclusion: These findings consolidate evidence of a fairly non-specific neurodevelopmental syndrome due to SOX4 haploinsufficiency in neurogenesis and multiple other developmental processes.

Keywords: abnormalities; congenital; gene expression regulation; genetic variation; hereditary; neonatal diseases.

Figure 1.. In silico analysis of SOX4 variants at the protein level.

(A) Schematic of the human SOX4 protein showing the location of the patients’ variants. Numbers below the schematic indicate amino acid numbers. Frameshift and nonsense variants (blue) and missense variants (red) are indicated with superscripts referring to the numbers assigned to patients in the Supplemental Table 4. (B) MetaDome plot showing the tolerance of SOX4 protein residues to missense mutations. Red arrows, location of the patients’ missense variants. (C) Swiss-Model rendering of the human SOX4 HMG domain–DNA complex (template 3u2b). The HMG domain (colored) forms three α-helices (H1, H2 and H3) that interact with DNA (grey shade) in the minor groove and induce a strong bend of the DNA helix. The N- and C-termini of the HMG domain are indicated (N and C, respectively). The residues altered in patients are indicated, with their side chains depicted in grey (carbons), red (oxygen) and blue (amine group). (D) Alignment of the HMG-domain sequences of various SOX4 vertebrate orthologues. SOX4 missense variants described in this study (red), described previously (pink) and reported in gnomAD (purple) are shown above sequences. Amino acids matching variants are similarly colored in the sequences. Above the sequences, symbols denote fully conserved (asterisks) and semi-conserved (dots) amino acids. Below the sequences, residues important for DNA binding and bending are shown with open blue triangles and green-colored triangles, respectively. Brown brackets demarcate the H1, H2, and H3 α-helices. Key amino acids in the N-terminal and C-terminal nuclear localization signal sequences (NLS) and nuclear export signal sequence (NES) are shown with continued lines and linked with dotted lines. (E) Alignment of the sequences encompassing Ala316 in SOX4 vertebrate orthologues. (F) Alignment of the C-terminal TAD sequences of SOX4 vertebrate orthologues. (G) Alignment of the HMG-domain sequences of all human SOX proteins. Proteins are listed based on group (A to H) classification. (H) Alignment of the sequences encompassing Ala316 in the human SOXC proteins. (I) Number of missense variants in the Ala316 region reported in gnomAD. (J) Alignment of the C-terminal TAD sequences of the human SOXC proteins. (K) Number of missense variants in the TAD C-terminus reported in gnomAD.

Figure 2.. Tests of the stability and intracellular distribution of SOX4 variant proteins.

(A) Western blots of cytoplasmic and nuclear extracts from Neuro-2a cells transfected with plasmids encoding 3FLAG-tagged SOX4 wild-type (WT) and variant (1 to 17) proteins. Cells were treated without or with lactacystin for the last 6 h of a 20–24 h transfection period. SOX4 proteins were detected using a FLAG antibody. Signals obtained for P84 (nuclear) and GAPDH (cytoplasmic) demonstrate the quality of samples and their fairly even loading amounts. The Mr of protein standards is indicated. Images are representative of duplicate samples tested in each of at least two independent experiments. Variants are numbered according to the patients in which they were detected (online supplemental table S4). (B) Quantification of SOX4 protein levels visualized in western blots prepared as in panel a. The top and middle graphs show the nuclear and cytoplasmic levels of the various SOX4 protein types relative to wild-type SOX4 in the absence of lactacystin. Bottom graphs show the ratios of nuclear versus cytoplasmic SOX4 protein levels. Individual values are shown for four samples corresponding to duplicate cultures (same symbols) in two independent experiments (circles and triangles). The bars show the mean ± standard deviation for all four values. Two-tailed paired Student’s T-tests were used to calculate the statistical significance of differences recorded between wild-type SOX4 and each variant, and between values obtained without and with lactacystin treatment for each protein type, as indicated with brackets (*, p ≤0.05; **, p ≥ 0.01; and ***, p ≥ 0.001; ns, not significant).

Figure 3.. Functional tests of SOX4 variant proteins

(A) EMSA using a Tubb3 DNA probe and whole-cell extracts from COS-1 cells transfected with expression plasmids for SOX4 wild-type and HMG missense variants. Top panel, picture of the gel showing the formation of complexes between the probe and SOX4 or a non-specific (ns) protein (the entire gel is shown in the online supplemental figure S2A). Bottom panel, western blot showing that the extracts contained fairly even amounts of SOX4 wild-type (WT) and variant proteins. The proteins were truncated at residue 214 and thus ran with an apparent Mr of 45k. Images are representative of data obtained with protein extracts from three independent experiments. Variants are numbered according to the patients in which they were detected (online supplemental table S4). (B-G) Tests of the abilities of SOX4 wild-type and variant proteins to transactivate pTubb3-Luc or 6FXO-p89-Luc reporters. Neuro-2a cells were transfected with either reporter, a Nanoluc control reporter, and empty (−), SOX4 and/or BRN2 expression plasmids, as indicated. Reporter activities were normalized for transfection efficiency and are presented as fold change increase relative to the activity of the reporter tested with the empty expression plasmid. Individual values are shown for triplicate cultures in three experiments (with squares, circles and triangles used to differentiate the experiments). The mean ± standard deviation of all data points are shown as bars and brackets. Average values are written. Asterisks indicate that a p value ≤0.05 was obtained in a Student’s T-test comparing values for the SOX4 wild-type and variant proteins. (H) Test of the abilities of SOX4 variants to interfere with wild-type SOX4 activity. Neuro-2a cells were transfected with 6FXO-p89-Luc, Nanoluc, and either 200 ng of empty expression plasmid (−) or 100 ng of SOX4 expression plasmid and 100 ng of SOX4 WT or variant expression plasmid, as indicated. Reporter activities were calculated and are presented as in (b to g). Asterisks indicate p values ≤0.05 in a Student’s T-test comparing values obtained when the SOX4 wild-type plasmid was tested in one dose and when it was tested along with another dose of itself or a SOX4 variant expression plasmid. (I) Test of the abilities of SOX4 wild-type and variant proteins to enhance the expression of neuronal markers in Neuro-2a cells. Cells were transfected with empty or SOX4 expression plasmids, as indicated. Total RNA was extracted the next day and used in qRT-PCR to measure the expression levels of Tubb3 and Hes5 relative to that of Gapdh. Data are presented as fold change increases obtained in cells transfected with a SOX4 plasmid compared to cells transfected with an empty plasmid. Individual values are shown for duplicate cultures in two independent experiments (with light and dark circles differentiating the experiments). The mean ± standard deviation obtained for all four values are shown as bars and brackets. Average values are written. Asterisks indicate p values ≤0.05 in a Student’s T-test comparing values obtained for SOX4 wild-type or variant proteins to values obtained without protein.