Variants in LRRC7 lead to intellectual disability, autism, aggression and abnormal eating behaviors

Abstract

Members of the leucine rich repeat (LRR) and PDZ domain (LAP) protein family are essential for animal development and histogenesis. Densin-180, encoded by LRRC7, is the only LAP protein selectively expressed in neurons. Densin-180 is a postsynaptic scaffold at glutamatergic synapses, linking cytoskeletal elements with signalling proteins such as the α-subunit of Ca²⁺/calmodulin-dependent protein kinase II. We have previously observed an association between high impact variants in LRRC7 and Intellectual Disability; also three individual cases with variants in LRRC7 had been described. We identify here 33 individuals (one of them previously described) with a dominant neurodevelopmental disorder due to heterozygous missense or loss-of-function variants in LRRC7. The clinical spectrum involves intellectual disability, autism, ADHD, aggression and, in several cases, hyperphagia-associated obesity. A PDZ domain variant interferes with synaptic targeting of Densin-180 in primary cultured neurons. Using in vitro systems (two hybrid, BioID, coimmunoprecipitation of tagged proteins from 293T cells) we identified new candidate interaction partners for the LRR domain, including protein phosphatase 1 (PP1), and observed that variants in the LRR reduced binding to these proteins. We conclude that LRRC7 encodes a major determinant of intellectual development and behaviour.

Fig. 1. Association of variants in LRRC7 with a neurodevelopmental disorder.

A Bevimed posterior probability of association (PPAs) > 0.95 with the Intellectual Disability (ID) disease class in the 100KGP. The associations are coloured by their level of supporting evidence in the PanelApp database of gene panels [accessed November 2022] (green: high, red: low, black: absent). B Left: Pedigrees A-I of 100KGP participants with probable loss-of-function (pLoF) variants in LRRC7. Individuals affected by ID according to 100KGP data are indicated in red; pedigree B stands out as two family members, shown in orange, are affected and assigned to the ‘Early onset and familial Parkinson’s disease’ (EOPD) 100KGP disease class. Right: Pedigrees J-# of the Genematcher/confirmation group of patients; affected individuals are indicated in blue. 1 indicates individuals heterozygous for a truncating variant, 0 indicates individuals homozygous for the allele in the reference genome. Amino acid positions refer to RefSeq entry NM_001370785.2. C Grid showing the human phenotype ontology (HPO) terms assigned to the phenotyped members of the nine pedigrees. The columns are grouped by pedigree. Only terms in at least two individuals are shown, and redundant rows have been removed.

Fig. 2. Variants in LRRC7 associated with a neurodevelopmental disorder.

A Domain structure of the Densin-180 protein encoded by LRRC7. Positions of missense, nonsense and frameshift variants are indicated, based on database entry NM_001370785.2. Variants from the 100KGP are listed in red, variants obtained through a GeneMatcher contact are shown in black. Known interaction partners of individual domains are indicated. No interaction partners have been published for the leucine rich repeat (LRR) and LAP specific domains (LAPsd) a and b (indicated by?). B Sequence alignment of the 17 LRRs of Densin-180, together with a consensus sequence. C 3D model of the Densin-180 LRR region, provided by the AlphaFold server. Residues altered by patient variants in LRRC7 are highlighted in (B) and indicated by space filling spheres in (C). Note that L65 and L296, as well as C106 and L221, are at equivalent positions in their respective repeats.

Fig. 3. let-413[L248P] embryos display an epidermal integrity defect.

A, C, E, G Single external focal plane images of AJM-1::GFP fluorescence. AJM-1 is an adherens junction component that is located at the apicolateral surface of epithelial cells in the (C). elegans embryo. B, D, F, H Nomarski light microscopy images of the corresponding fluorescent image above. let-413[L248L], let-413[L248P] homozygous and let-413[Del]/tmC3 heterozygous embryos were obtained 4.5 h after a 1 h egg lay at 25 ^oC, while the let-413[Del] homozygous embryo was 5.5 h after egg lay. Note WT continuous AJM-1::GFP staining surrounding let-413[L248L] epidermal cells (A) and let-413[Del]/tmC3 epidermal cells at a slightly higher focal plane (E), while let-413[L248P] and let-413[Del] show discontinuous staining, indicative of disruption of embryonic epithelial cell integrity. Older arrested embryos can show vacuolation and rupture. Scale bar: 10 µm for all images. For microscopic analysis, three separate experiments were performed, with about 50 embryos scored in each experiment with similar results.

Fig. 4. A variant in the PDZ domain interferes with binding to δ-catenin and targeting to postsynaptic clusters.

A 293T cells coexpressing GFP-tagged Densin-180 (full-length; WT or L1567Q mutant), or GFP alone, together with an HA-tagged δ-catenin were lysed (Input samples). GFP-tagged proteins were immunoprecipitated (IP) using the GFP-trap matrix. Input and precipitate samples were analysed by Western blotting using the epitope tag antibodies, as indicated. B Quantification of data from four independent transfections. Data are presented as the ratio of HA-δ-catenin signals in IP samples, divided by GFP-Densin signals, also in IP samples. Data are shown as mean values  ± SD. ****, significantly different; p < 0.0001; unpaired, two sided Student’s t test. C Hippocampal neurons were transfected with plasmids coding for Densin-180 WT or the L1567Q mutant, and stained for MAP2, Shank3 and the expressed protein. Cells were analysed by confocal microscopy. Left panel shows overview images of neurons (scale bar 20 μm) and right panel shows magnified dendritic segments (scale bar 10 μm). Three independent preparations were transfected and neurons were analysed, with similar results. Quantitative analysis showed that the number of primary dendrites (D), and the number of dendritic Densin-180 clusters (E), is not affected by the L1567Q variant. However, the number of postsynaptic clusters of Densin-180, defined by colocalization with Shank3, is reduced (F). Quantitative analysis was performed on 15 neurons (D) and on 45 dendrites of 15 neurons (E, F) obtained in three independent preparations. **, significantly different from WT; p = 0.009 paired, two sided t-Test. Data are shown as mean values  ± SD. Source data are provided as a Source Data file.

Fig. 5. Expression of Densin-180 LRR domain variants in cultured neurons.

A Primary cultured hippocampal neurons were transfected with plasmids coding for GFP-tagged Densin-180 (WT and variants) at day in vitro 7 (DIV7). After differentiation for a total of 14 days (DIV14), cells were fixed, stained for MAP2 and visualised by confocal microscopy (scale bar: 20 µm). Number of primary dendrites emerging from cell somata for variants shown in this Figure (B) and for variants shown in Supplementary Fig. 3 (C). Differences are non-significant upon analysis of n = 15 neurons from three independent transfections; one-way ANOVA, followed by Dunnett’s multiple comparisons test. Data are shown as mean values ± SD. Source data are provided as a Source Data file.

Fig. 6. Variants N94S and L296P affect targeting of Densin-180 to postsynaptic sites.

A High magnification micrographs of hippocampal neurons expressing GFP-Densin-180 WT and variants L65S, L221M and L296P (scale bar: 5 µm). Neurons were costained for the postsynaptic marker Shank3, and the somatodendritic marker MAP. For quantitative analysis, we counted the number of clusters per length of dendrite (B; p-values 0.0114 for L65S; 0.3448 for L221M; < 0.0001 for L296P), the number of Shank3-positive GFP-Densin clusters (C; p-values < 0.0001 for L65S; 0.9484 for L221M; < 0.0001 for L296P), and the spine-shaft ratio of GFP-fluorescence intensities (D; p-values < 0.0165 for L65S; 0.1007 for L221M; < 0.0001 for L296P). Data are shown as mean values ± SD. E–H The analysis was repeated for N94S, C106Y and A193P variants (scale bar: 5 µm). p-values for differences to WT are for (F): 0.0675 (N94S); 0.9903 (C106Y); 0.0208 (A193P); for (G): 0.0414 (N94S); 0.5894 (C106Y); 0.9950 (A193P); for (H): 0.9168 (N94S); 0.0531 (C106Y); 0.8678 (A193P). Data are shown as mean values  ± SD. *,**,***,**** indicate statistically significant differences from WT, p < 0.05, 0.01, 0.001, 0.0001, respectively. Analysis was performed on 45 dendrites (B, C, E, F) and 150 dendritic clusters (D, G) of 15 neurons from three independent transfections. One-way ANOVA, followed by Dunnett’s multiple comparisons test. Source data are provided as a Source Data file.

Fig. 7. Identification of novel interaction partners of the LRR domain of Densin-180.

A 293T cells were transfected with plasmids coding for GFP-tagged LRR of Densin-180, or GFP alone, in combination with DDX3, HA-tagged Erbin, or myc-tagged full-length Densin-180. After cell lysis, GFP-tagged proteins were immunoprecipitated (IP) using the GFP-trap matrix. Input and IP samples were analysed by Western Blot (immunoblot, IB) using the antibodies indicated. B 293T cells were transfected with plasmids coding for GFP-tagged Densin-180 (full length), GFP-tagged LRR domain, or GFP alone in combination with Flag-tagged PP1α. GFP-tagged proteins were immunoprecipitated as in (A) and analysed by Western blotting. Experiments were repeated three times, with similar results. Source data are provided as a Source Data file.

Fig. 8. Variants in LRRC7 disrupt binding to interaction partners of the LRR domains in 293T cells.

A 293T cells coexpressing GFP-tagged Densin-180 (full-length; WT or mutants), or GFP alone, together with an mRFP fusion protein of the C-terminus of MACF1 were lysed (Input samples). GFP-tagged proteins were immunoprecipitated (IP) using the GFP-trap matrix. Input and precipitate samples were analysed by Western blotting using the epitope tag antibodies, as indicated. B Quantification of the data shown in (A). Data are presented as the ratio of mRFP-MACF1 signals in IP samples, divided by GFP-Densin signals, also in IP samples. p-values for differences to WT are: 0.0298 (L65S); 0.0481 (N94S); 0.4864 (A193P); 0.6646 (L221M) and 0.0064 (L296P). Data are shown as mean values ± SD. C 293T cells coexpressing GFP-tagged LRR repeats of Densin-180 (WT or mutant) together with HA-tagged Erbin. Immunoprecipitation and analysis by Western Blot was performed as in (A). D Quantitative analysis of the data shown in (C). Data are presented as the ratio of HA-Erbin signals in IP samples, divided by GFP-LRR signals, also in IP samples. p-values for differences to WT are: < 0.0001 (L65S); <0.0001 (N94S); <0.0001 (C106Y); 0.0138 (A193P); 0.1276 (L221M) and <0.0001 (L296P). Data are shown as mean values ± SD. E Coexpression and coimmunoprecipitation analysis was performed with GFP-tagged LRRs of Densin-180, and Flag-tagged protein phosphatase 1α (PP1α). F Quantitative analysis of the data shown in E. Data are presented as the ratio of Flag-PP1α signals in IP samples, divided by GFP-LRR signals, also in IP samples. p-values for differences to WT are: <0.0001 (L65S); <0.0001 (N94S); <0.0001 (C106Y); 0.531 (A193P); 0.1004 (L221M) and <0.0001 (L296P). Data are shown as mean values ± SD. *, ****, indicate significant differences from WT; p < 0.05, 0.0001, respectively. Data were analysed by one-way ANOVA, followed by Dunnett’s multiple comparisons test (B, D, F). Data were from four independent transfections. Source data are provided as a Source Data file.