CAPRIN1 haploinsufficiency causes a neurodevelopmental disorder with language impairment, ADHD and ASD

Abstract

We describe an autosomal dominant disorder associated with loss-of-function variants in the Cell cycle associated protein 1 (CAPRIN1; MIM*601178). CAPRIN1 encodes a ubiquitous protein that regulates the transport and translation of neuronal mRNAs critical for synaptic plasticity, as well as mRNAs encoding proteins important for cell proliferation and migration in multiple cell types. We identified 12 cases with loss-of-function CAPRIN1 variants, and a neurodevelopmental phenotype characterized by language impairment/speech delay (100%), intellectual disability (83%), attention deficit hyperactivity disorder (82%) and autism spectrum disorder (67%). Affected individuals also had respiratory problems (50%), limb/skeletal anomalies (50%), developmental delay (42%) feeding difficulties (33%), seizures (33%) and ophthalmologic problems (33%). In patient-derived lymphoblasts and fibroblasts, we showed a monoallelic expression of the wild-type allele, and a reduction of the transcript and protein compatible with a half dose. To further study pathogenic mechanisms, we generated sCAPRIN1+/- human induced pluripotent stem cells via CRISPR-Cas9 mutagenesis and differentiated them into neuronal progenitor cells and cortical neurons. CAPRIN1 loss caused reduced neuronal processes, overall disruption of the neuronal organization and an increased neuronal degeneration. We also observed an alteration of mRNA translation in CAPRIN1+/- neurons, compatible with its suggested function as translational inhibitor. CAPRIN1+/- neurons also showed an impaired calcium signalling and increased oxidative stress, two mechanisms that may directly affect neuronal networks development, maintenance and function. According to what was previously observed in the mouse model, measurements of activity in CAPRIN1+/- neurons via micro-electrode arrays indicated lower spike rates and bursts, with an overall reduced activity. In conclusion, we demonstrate that CAPRIN1 haploinsufficiency causes a novel autosomal dominant neurodevelopmental disorder and identify morphological and functional alterations associated with this disorder in human neuronal models.

Keywords: ADHD; ASD; CAPRIN1; RNG105; neurodevelopment.

Figure 1

CAPRIN1-associated disease: variants and clinical features. (A) Schematic representation of a large ∼1.4 Mb deletion spanning the CAPRIN1 gene (top) and 11 CAPRIN1 (NM_005898.5) loss of function variants (bottom); 10 single nucleotide variants detected in this work and two variants reported in the literature^, are represented on a schematic view of the protein (created with BioRender.com). Overall, they include two splicing (c.279+1G>T; c.879G>A) and 10 nonsense/frameshift changes. CAPRIN1 structure was schematized according to Wu et al.: two highly conserved homologous regions (HR1, 56–248 amino acids and HR2, 352–685 amino acids) and a less conserved glutamate-rich (E-rich) region in between are shown. Moreover, C-terminal Arg-Gly-Gly (RGG) motifs—characteristic of some RNA-binding proteins—are indicated in purple. (B) Main neurological (top) and non-neurological (bottom) features of CAPRIN1-associated disease in the 12 cases described here. (C) Facial features of the 12 cases do not show peculiar dysmorphisms, although broad nasal tip, low-hanging columella and thin upper lip vermilion seemed recurrent. Facial features analysis by Face2Gene software (FDNA Inc., https://www.face2gene.com/) does not show a peculiar overlap. Further details are provided in Supplementary Fig. 2 and Supplementary Table 1.

Figure 2

Expression analysis of CAPRIN1 in patient-derived cells. (A) Expression of CAPRIN1 versus GAPDH was reduced in patients’ fibroblasts or PBMC, measured by RT-qPCR (Case 8, wt/Gln399Ter (in figure Q399*) = 0.59 ± 0.01, n = 3; Case 12, wt/Gln582Ter (in figure Q582*) = 0.41 ± 0.01, n = 3; CTRLs = 1.03 ± 0.08, n = 9; Case 1, wt/del11p13 = 0.34 ± 0.03, n = 9 CTRLs = 1.04 ± 0.09, n = 3; mean ± SEM. P = 0.0007 (Q399* versus CTRLs) and <0.0001 (Q582* versus CTRLs and del11p13 versus CTRLs). For each experiment, three different controls were used, and the experiments were performed in triplicate; controls were healthy young adults. (B) Sanger sequencing of the cDNA containing the c.1195C>T p.Gln399Ter and c.1744C>T p.Gln582Ter shows that the mutant allele is probably degraded by NMD; cDNA panels. Indeed, treating cells with CHX, an NMD blocker, prior to RNA extraction, partially rescued the expression of the mutant allele (cDNA + CHX panels). As comparison, the sequences obtained from genomic DNA (gDNA panels). (C) CAPRIN1 protein analysis in fibroblasts of the patients carrying the c.1195C>T p.Gln399Ter and c.1744C>T p.Gln582Ter variants. Western blot shows a protein reduction compatible with a half dose (Case 8 = 0.26 ± 0.03, n = 4; Case 12 = 0.36 ± 0.02, n = 4; CTRLs = 0.99 ± 0.06, n = 10. P < 0.0001). For every experiment, three controls were used from different healthy patients (young adults) and experiments were performed a minimum of three times. Vinculin was used as a normalizer, and MemCode stain (Thermo Fisher Scientific) to confirm the efficiency of protein transferring and the overall protein abundancy. (D) Analysis of the cDNA from LCLs of the c.279+1G>T (Case 2) and c.879G>A (Case 3) carriers. The c.279+1G>T change causes an in-frame 84 bp exon 3 skipping (left). The c.879G>A affects the last base of exon 8 and causes out-of-frame skipping of exon 8 and a degradation by NMD. Exon 8 skipping could be detected by CHX treatment (right). Control (CTRL) LCLs derived from a young adult sex-matched healthy patient.

Figure 3

Haploinsufficiency of CAPRIN1 causes impaired neuronal organization and increased degeneration of neuronal processes. (A) Representative images of neuronal rosettes derived from CTRL and CAPRIN1^+/− cells. CAPRIN1 protein (magenta) could be easily detected in these structures. No gross morphological differences were observed between CTRL and CAPRIN1^+/− rosettes. Images were acquired at ×40 magnification using Z-stack mode. The neuronal marker Tuj1 (green) is used to see the rosette morphology, and the DAPI (blue) to stain the nucleus. (B) CAPRIN1^+/− neurons 10 days post-ND showed an impaired neuronal organization (×20 magnification; blue = DAPI; magenta = CAPRIN1; green = Tuj1. Scale bar = 25 µm). (C) Quantification of neuronal processes length on ×20 brightfield images, Day 7 post-ND. Length was quantified using the Fiji/ImageJ tool. CTRL = 134 ± 5.62, n = 90; CAPRIN1^+/− = 74.8 ± 3.55, n = 90; mean ± SEM; P < 0.0001. Each dot represents a different neuronal process; images deriving from three independent differentiations were analysed. (D) Representative images of the neuronal death observed starting 18 days post-ND. Images were acquired 18 days post-ND (left: ×4 magnification, middle: ×10 magnification) and 21 days post-ND (right: ×20 magnification), using a brightfield microscope.

Figure 4

Calcium overload and increased oxidative stress are observed in CAPRIN1^+/− neurons. (A) Quantification of calcium signal in CAPRIN1^+/− neurons 10 and 14 days post-ND. Neurons were recorded for 2 min, and the integrated density (IntDen) signal of each cell was normalized on its lowest value during this time. Right: A representative image of the lowest (F₀) and maximum (F_max) fluorescence intensity for each cell; the lookup table ‘fire’ was used. Left: The Fluo-4 (green) signal, more appreciable in the image on the right (Day 10 post-ND: CTRL = 127 ± 0.25, n = 197; CAPRIN1^+/− = 131 ± 0.37, n = 197; Day 14 post-ND: CTRL = 125 ± 0.67, n = 354; CAPRIN1^+/− = 138 ± 0.45, n = 361; P < 0.0001). For every genotype and time point, three different coverslips were analysed, each of them deriving from an independent differentiation. (B) Analysis of CAPRIN1^+/− neurons (Day 18 post-ND) with CellROX showed a significant increase of ROS (CTRL = 1.73 × 10⁷ ± 1.80 × 10⁶, n = 697; CAPRIN1^+/− = 5.64 × 10⁷ ± 2.96 × 10⁶, n = 1596; P < 0.0001). Twenty-two images for every genotype have been analysed from three different coverslips, each of them deriving from an independent differentiation. Right: Example of the difference between CTRL and CAPRIN1^+/− neurons (×20 magnification; blue = DAPI; green = CellROX. Scale bar = 25 µm).

Figure 5

CAPRIN1 ^+/− neurons show increased global translation and slight alteration of local translation. (A) Western blot analysis of SUnSET assay on neurons 14 days post-ND treated with 1 µM puromycin for 30 min showed a significantly higher puromycin signal compared to CTRL. Each CTRLs and CAPRIN1^+/− lane (and dot on the graph) represents an independent ND (n = 3). (B) Immunofluorescence analysis of SUnSET assay on CAPRIN1^+/− neurons 4 days post-ND, treated with 1 µM puromycin for 30 min. Representative images for CTRL and CAPRIN1^+/− dendrites are reported (4 days post-ND, zoom of ×40 magnification, scale bar = 1 µm), using the ImageJ/Fiji lookup table ‘royal’, to visualize the strength of the signal better. Magnification of the dendrites show the puromycin fluorescence intensity from 0 to 80 µm of distance from the cell bodies; intensity of signal is maintained along the dendrite in CTRL neurons, while it decreases visibly in CAPRIN1^+/− ones. Enlarged views of the fragment from 15 to 30 µm of the dendrite are provided below (scale bar = 1 µm), showing reduced intensity of signal in CAPRIN1^+/− neurons. An image of the entire neuron is provided in Supplementary Fig. 8C. (C) Quantification of the puromycin signal, indicated as grey value, along the dendrite (n = 6). Both panels (left = 4 days post-ND; right = 14 days post-ND) show higher signal in proximity of the soma, that progressively decrease. However, the decrease is significantly higher in CAPRIN1^+/− neurons, particularly 4 days post-ND. Statistical significance was tested by two-way ANOVA followed by Bonferroni post hoc tests; column factor was considered. P-value < 0.0001; mean and error (SEM) are shown.

Figure 6

Neurons deficient for CAPRIN1 show abnormal firing properties. (A) MEA analyses in CTRL and CAPRIN1^+/− neurons at different stages of maturation (Days 4, 7, 10 and 14 post-ND). Each image is representative of three different wells, recorded for 2 min on 12 electrodes. Box and whisker plots (5–95th percentile), outliers are indicated as dots; mean is indicated with a plus symbol; numbers on each bar indicate the mean. Wells that never showed an activity throughout the time course have been excluded from the analysis (total CTRL n = 14 wells; total CAPRIN1^+/−n = 9 wells). (B) Quantification of MEA data, reporting the spike rate (Hz) along the 2 min of recording (Day 4 CTRL = 0.66 ± 0.14; CAPRIN1^+/− = 0.22 ± 0.09; mean ± SEM; P = 0.02; Day 7 CTRL = 0.78 ± 0.11; CAPRIN1^+/− = 0.36 ± 0.08; P = 0.006; Day 10 CTRL = 1.55 ± 0.28; CAPRIN1^+/− = 0.66 ± 0.13; P = 0.01; Day 14 CTRL = 1.82 ± 0.29; CAPRIN1^+/− = 1.03 ± 0.20; P = 0.04). (C) Quantification of MEA data, reporting the spikes count along the 2 min of recording (Day 4 CTRL = 78.7 ± 17.3; CAPRIN1^+/− = 26.7 ± 10.9; mean ± SEM; P = 0.02; Day 7 CTRL = 93.5 ± 13.6; CAPRIN1^+/− = 42.7 ± 9.06; P = 0.006; Day 10 CTRL = 186 ± 33.3; CAPRIN1^+/− = 78.8 ± 15.7; P = 0.01; Day 14 CTRL = 215 ± 35.3; CAPRIN1^+/− = 124 ± 24.3; P = 0.05). (D) Burst count along the 2 min of recording (Day 4 CTRL = 0.54 ± 0.25; CAPRIN1^+/− = 0.16 ± 0.10; mean ± SEM; P = 0.18; Day 7 CTRL = 0.50 ± 0.22; CAPRIN1^+/− = 0.30 ± 0.17; P = 0.48; Day 10 CTRL = 6.49 ± 1.05; CAPRIN1^+/− = 1.03 ± 0.47; P = 0.002; Day 14 CTRL = 7.62 ± 0.64; CAPRIN1^+/− = 3.30 ± 0.98; P = 0.002). (E) Burst duration (ms) along the 2 min of recording (Day 7 CTRL = 233 ± 36.7; CAPRIN1^+/− = 186 ± 25.3; Day 10 CTRL = 212 ± 8.76; CAPRIN1^+/− = 223 ± 9.51; Day 14 CTRL = 183 ± 18.5; CAPRIN1^+/− = 147 ± 4.49). The number of bursts in CTRL was higher compared to CAPRIN1^+/− , therefore the number of considered events (n) is reported. (F) Representative images of CTRL and CAPRIN1^+/− detected spikes.