BCL11A Haploinsufficiency Causes an Intellectual Disability Syndrome and Dysregulates Transcription

Abstract

Intellectual disability (ID) is a common condition with considerable genetic heterogeneity. Next-generation sequencing of large cohorts has identified an increasing number of genes implicated in ID, but their roles in neurodevelopment remain largely unexplored. Here we report an ID syndrome caused by de novo heterozygous missense, nonsense, and frameshift mutations in BCL11A, encoding a transcription factor that is a putative member of the BAF swi/snf chromatin-remodeling complex. Using a comprehensive integrated approach to ID disease modeling, involving human cellular analyses coupled to mouse behavioral, neuroanatomical, and molecular phenotyping, we provide multiple lines of functional evidence for phenotypic effects. The etiological missense variants cluster in the amino-terminal region of human BCL11A, and we demonstrate that they all disrupt its localization, dimerization, and transcriptional regulatory activity, consistent with a loss of function. We show that Bcl11a haploinsufficiency in mice causes impaired cognition, abnormal social behavior, and microcephaly in accordance with the human phenotype. Furthermore, we identify shared aberrant transcriptional profiles in the cortex and hippocampus of these mouse models. Thus, our work implicates BCL11A haploinsufficiency in neurodevelopmental disorders and defines additional targets regulated by this gene, with broad relevance for our understanding of ID and related syndromes.

Figure 1

Clinical and Molecular Features of Individuals with BCL11A Mutations (A) Schematic representation of the three major isoforms of BCL11A (GenBank and Ensembl transcript identifiers indicated): BCL11A-XL (NM_022893.3, ENST00000335712), BCL11A-L (NM_018014.3, ENST00000356842), and BCL11A-S (NM_138559, ENST00000359629). Mutations are represented on and annotated according to the BCL11A-XL isoform (NM_022893.3); predicted protein variants are represented as follows: orange triangles, missense variants; red stars, truncating variants identified in the current study; gray stars, variants in previously reported individuals with autism spectrum disorder. C2H2 DNA binding zinc finger domains are represented in red; non-DNA binding zinc finger is shaded brown. The red line indicates the putative dimerization region. Putative region required for SUMO1 recruitment is shaded gray. Proline (Pro)- and glutamate (Glu)-rich regions are shaded blue. Abbreviation is as follows: aa, amino acids. (B–D) Clinical features of individuals 1, 2, 4, 5, 7, 8, and 9. Individual number and predicted protein variant are indicated below the respective photograph. Various and partly overlapping facial features of individuals with missense (B) and loss-of-function (C) mutations in BCL11A, including strabismus, downslanting palpebral fissures, synophrys, flat midface, thin upper lip, and full lower lip. Shared core clinical features of all affected individuals (detailed in Table 1) are summarized in (D). The heatmap colors represent the frequency of the feature seen, from 100% (red) to 30% (blue); frequency is calculated based on information available for each feature. Features that are present predominantly (microcephaly, abnormal external ears) or exclusively (blue sclerae in infancy) in individuals with LoF mutations are indicated by a red star. Abbreviations are as follows: DD, developmental delay; ID, intellectual disability; ASD, autism spectrum disorder.

Figure 2

BCL11A Missense Substitutions Disrupt Protein Localization (A) Fluorescence micrographs of HEK293 cells transfected with wild-type (WT) or mutant short isoform of BCL11A (BCL11A-S; GenBank: NM_138559, Ensembl: ENST00000359629) fused to mCherry (red). Nuclei were stained with Hoechst 33342 (blue). (B) Fluorescence micrographs of HEK293 cells transfected with wild-type or mutant long isoform of BCL11A (BCL11A-L; NM_018014.3, ENST00000356842) fused to mCherry (red). Nuclei were stained with Hoechst 33342 (blue). (C) Fluorescence micrographs of cells transfected with wild-type or mutant BCL11A-L fused to mCherry (red) and NONO fused to YFP (green). Nuclei were stained with Hoechst 33342 (blue). Scale bars represent 10 μm. (D) Bioluminescence resonance energy transfer (BRET) assay for interaction of wild-type or mutant BCL11A-L with NONO. HEK293 cells were transfected with BCL11A-L fused to Renilla luciferase, and NONO fused to YFP. Values are mean corrected BRET ratios ± SEM (n = 3, ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, one-way ANOVA followed by Bonferroni post hoc correction).

Figure 3

BCL11A Missense Substitutions Disrupt Protein Dimerization and Transcriptional Regulation (A) BRET assay for interaction of mutant BCL11A-L (GenBank: NM_018014.3, Ensembl: ENST00000356842) with wild-type (WT) BCL11A-L. HEK293 cells were transfected with wild-type or mutant BCL11A-L fused to Renilla luciferase (donor) and wild-type BCL11A-L fused to YFP (acceptor). The control donor protein is a nuclear-targeted luciferase. Values are mean corrected BRET ratios ± SEM (n = 3, ^∗∗∗p < 0.001, one-way ANOVA followed by Bonferroni post hoc correction). (B) BRET assay for interaction of mutant BCL11A-L with wild-type BCL11A-S (NM_138559, ENST00000359629). HEK293 cells were transfected with wild-type or mutant BCL11A-L fused to Renilla luciferase (donor) and wild-type BCL11A-S fused to YFP (acceptor). Values are mean corrected BRET ratios ± SEM (n = 3, ^∗p < 0.05, ^∗∗∗p < 0.001; one-way ANOVA followed by Bonferroni post hoc correction). (C) Fluorescence micrographs of HEK293 cells transfected with wild-type or mutant BCL11A-L fused to YFP (green), together with wild-type BCL11A-S fused to mCherry (red). Nuclei were stained with Hoechst 33342 (blue). Scale bar represents 10 μm. (D) Mammalian one-hybrid assay for BCL11A-L transcriptional regulatory activity. HEK293 cells were transfected with wild-type or mutant BCL11A-L fused to the DNA-binding domain of GAL4, together with a firefly luciferase reporter plasmid containing GAL4 binding sites, and a Renilla luciferase normalization plasmid. Values are mean firefly luciferase activity normalized to Renilla luciferase activity ± SEM (n = 3), expressed relative to the control (^∗∗∗p < 0.001, one-way ANOVA followed by Bonferroni post hoc correction).

Figure 4

Neuroanatomical Imaging of Mouse Brain Reveals Microcephaly Abbreviations are as follows: +/+, wild-type; +/−, Bcl11a^+/−· (A) Total brain volume in mm³ is significantly decreased in mutant mice (mean volumes: 457 mm³ in wild-type; 384 mm³ in Bcl11a^+/− mice; p = 1.3 × 10⁻⁵ two-tailed F test corrected for multiple comparisons by controlling the false-discovery rate at q < 0.05). (B) 3D reconstructions of average MRI volumes for wild-type (blue) and Bcl11a^+/− (pink) mice. (C) 3D reconstruction of heterozygous brain MRI; regions depicted in pink are significantly smaller in Bcl11a^+/− mice after normalization for total brain volume. Arrows indicate orientation: blue, dorsal; green, rostral; black, ventral. (D) 3D reconstruction of representative wild-type and Bcl11a^+/− skulls demonstrating normal cranial morphology and reduced cranial size. (E) Cranial measurements in mm are decreased in heterozygous mice (Mann-Whitney test, ^∗p < 0.05; ^∗∗p < 0.01). Boxes indicate 25^th, mean, and 75^th percentiles; whiskers indicate minimum and maximum values. MRI: Bcl11a^+/+, n = 11; Bcl11a^+/−, n = 5. μCT: Bcl11a^+/+, n = 5; Bcl11a^+/−, n = 5.

Figure 5

Bcl11a^+/− Mice Present with a Cognitive and Behavioral Phenotype (A) Social recognition assay: diagram of test arenas. (B) On day 1, mice were tested for habituation-dishabituation to a conspecific (Bcl11a^+/+, n = 11; Bcl11a^+/−, n = 12). Both genotypes show decline in investigation of the same stimulus mouse (black) over four trials, with recovery on presentation of a novel stimulus mouse on trial 5 (pink): two-way ANOVA for trial 4 versus 5, ^∗∗∗p < 0.0001 for trial, with no significant interaction or difference per genotype (p = 0.112, ns). No significant difference between genotypes on post hoc t test (Bcl11a^+/+ versus Bcl11a^+/−, p = 0.637). Values are mean time in seconds (s) ± SEM. (C) On a discrimination test 24 hr later, Bcl11a^+/− animals are unable to discriminate between the familiar (presented on trials 1–4 of day 1) and novel unfamiliar (green) stimulus, unlike wild-types (2-way ANOVA, significant interaction, p = 0.007; post hoc t test per genotype: Bcl11a^+/+, ^∗∗∗p = 0.0005; Bcl11a^+/−, p = 0.71, ns). (D and E) Three-chamber social behavior test. Heatmaps of automated tracking recordings of a representative Bcl11a^+/+ and a Bcl11a^+/− mouse; the color gradient represents time in that location: red (maximum) to blue (minimum). (D) On habituation to empty chambers (5 min in center + 5 min in all three), there is no preference for left or right chamber in either genotype (p = 0.414). (E) Mice were then tested for preference for a novel object (cylinder) with a conspecific versus novel object only. The dashed white circles indicate the novel objects introduced: one empty cylinder, one containing a conspecific mouse represented in light gray. (F) Bcl11a^+/− mice show decreased preference for the chamber containing the object with a conspecific compared to wild-type. Boxplot of the preference index (PI) shows significantly less preference for the novel object with the conspecific in Bcl11a^+/− mice. Unpaired t test, ^∗∗∗p = 0.0005; Bcl11a^+/+, n = 12; Bcl11a^+/−, n = 16. For box and whisker plots (C and F), boxes indicate 25^th, median, and 75^th percentiles; whiskers indicate minimum and maximum values.

Figure 6

Bcl11a^+/− Mice Have Altered Gene Expression Profiles in Cortex and Hippocampus (A and B) MA plots of differential gene expression between Bcl11a^+/− and Bcl11a^+/+ cortex (A) and hippocampus (B). The x-axis is the log₁₀ average expression of all samples (normalized counts); the y-axis is DESeq2’s shrinkage estimation of log₂ fold changes between genotypes. Each gene is represented as a dot; significantly differentially expressed genes (BH-adjusted p value < 0.1) are highlighted in red. n = 9 per genotype for cortex; n = 6 per genotype for hippocampus. (C) Comparison of DESeq2’s shrinkage estimation of log₂ fold changes of genes in both tissues. Red dots represent genes with significant differential expression in both tissues; yellow dots represent genes differentially expressed in the cortex only, and blue in the hippocampus only; gray dots represent genes not differentially expressed in either tissue (BH-adjusted p value < 0.1).