Activating RAC1 variants in the switch II region cause a developmental syndrome and alter neuronal morphology

Abstract

RAC1 is a highly conserved Rho GTPase critical for many cellular and developmental processes. De novo missense RAC1 variants cause a highly variable neurodevelopmental disorder. Some of these variants have previously been shown to have a dominant negative effect. Most previously reported patients with this disorder have either severe microcephaly or severe macrocephaly. Here, we describe eight patients with pathogenic missense RAC1 variants affecting residues between Q61 and R68 within the switch II region of RAC1. These patients display variable combinations of developmental delay, intellectual disability, brain anomalies such as polymicrogyria and cardiovascular defects with normocephaly or relatively milder micro- or macrocephaly. Pulldown assays, NIH3T3 fibroblast spreading assays and staining for activated PAK1/2/3 and WAVE2 suggest that these variants increase RAC1 activity and over-activate downstream signalling targets. Axons of neurons isolated from Drosophila embryos expressing the most common of the activating variants are significantly shorter, with an increased density of filopodial protrusions. In vivo, these embryos exhibit frequent defects in axonal organization. Class IV dendritic arborization neurons expressing this variant exhibit a significant reduction in the total area of the dendritic arbour, increased branching and failure of self-avoidance. RNAi knock down of the WAVE regulatory complex component Cyfip significantly rescues these morphological defects. These results establish that activating substitutions affecting residues Q61-R68 within the switch II region of RAC1 cause a developmental syndrome. Our findings reveal that these variants cause altered downstream signalling, resulting in abnormal neuronal morphology and reveal the WAVE regulatory complex/Arp2/3 pathway as a possible therapeutic target for activating RAC1 variants. These insights also have the potential to inform the mechanism and therapy for other disorders caused by variants in genes encoding other Rho GTPases, their regulators and downstream effectors.

Keywords: RAC1; WAVE regulatory complex; intellectual disability; polymicrogyria; small GTPases.

Figure 1

RAC1 switch II variants cause a neurodevelopmental disorder. (A) Alignment of switch II region of RAC1 and related small GTPases. Residues in RAC1 affected by described variants are indicated in bold. (B) Structure of RAC1 bound to GTP analogue (Protein Database ID 1MH1). Switch II region indicated. Hydrogen bonds predicted to be formed between R68 and other residues in RAC1 are indicated by dashed lines. Image preparation and hydrogen bond prediction performed using UCSF ChimeraX. (C) T₁-weighted brain MRI images of Patient 3 with the Y64C variant illustrating bilateral perisylvian polymicrogyria (white quadrilaterals) and thin corpus callosum (arrow).

Figure 2

RAC1 switch II variants increase levels of GTP-bound RAC1 and alter fibroblast morphology. (A) Western blot of myc-RAC1-GTP (top panel) pulled down from lysates of HEK293 cells expressing indicated RAC1 variant using PAK-CRIB probed with anti-myc. Lower two panels show blots of raw lysates used in pulldowns probed with anti-myc to show total myc-RAC1 levels or anti-actin as loading control. Uncropped images of these blots are shown in Supplementary Fig. 1. (B) Quantitation of relative GTP-RAC1 levels for indicated RAC1 variants. Calculated by dividing GTP-RAC1 band intensity by total RAC1 intensity for each sample then normalizing to value obtained for constitutively active (Q61L) RAC1 in the same dataset. Bars indicate mean ± SEM. n = 7 independent experiments for all variants, except Y64C and R68G where n = 4. Data analysed in Graphpad Prism using mixed effects model. *P < 0.05. (C) Spreading NIH3T3 fibroblasts expressing indicated RAC1 variant stained with Alexa568-Phalloidin to label F-actin and anti-myc to label expressed RAC1 variant. Second row shows magnification of a section of the cell periphery in the above image. Arrows indicate localization of RAC1 variants to cell periphery. Scale bars in first column indicate 10 µm and apply to all images in the row. (D) Circularity of cells expressing indicated RAC1 variant. n > 50 cells pooled from three independent experiments. Line indicates mean value. Data statistically analysed using Kruskal–Wallis test with Dunn’s correction for multiple comparisons. ****P < 0.0001, ***P < 0.001, *P < 0.05, ns P > 0.05 relative to RAC1 WT. (E) Categorization of cell morphology based on predominant protrusion type. Cell scored as >50% lamellipodia or filopodia if this protrusion type occupies greater than 50% of cell periphery. n > 50 cells pooled from three independent experiments. CA = constitutively active; DN = dominant negative.

Figure 3

RAC1 switch II variants increase WRC and PAK activity. (A and B) Spreading NIH3T3 fibroblasts expressing indicated RAC1 variant stained for WAVE2 (A) or activated PAK1/2/3 (B). Top row: Staining for WAVE2 or activated PAK1/2/3 alone. Middle row: Merge of WAVE2 or activated PAK1/2/3 and myc-RAC1 channels. Bottom row: Magnification of a region of cell periphery from above merged images. Arrows indicate peripheral accumulations of WAVE2 or activated PAK1/2/3. Scale bars in first column indicate 10 µm and apply to all images in the row. (C) Activated PAK fluorescence intensity at cell periphery of NIH3T3 cells expressing indicated variant. n > 50 cells pooled from three independent experiments. Line indicates mean value. Data statistically analysed using Kruskal–Wallis test with Dunn’s correction for multiple comparisons. **** P < 0.0001, ns P > 0.05 relative to RAC1 WT. CA = constitutively active; DN = dominant negative.

Figure 4

Expression of Rac1-Y64D alters axon morphology and organization in Drosophila embryos. (A) Representative images of cultured Drosophila embryonic neurons extracted from stage 11 embryos of indicated genotype. Stained for tubulin, F-actin and DAPI. Arrows indicate filopodia. Scale bar = 10 µm and applies to all three images. (B and C) Quantitation of axon length of cultured Drosophila embryonic neurons and density of filopodia along axon shaft. Eighty-five neurons analysed for each genotype extracted from 24 stage 11 embryos. Lines indicate mean values. Data statistically analysed using Kruskal–Wallis test with Dunn’s correction for multiple comparisons. *P < 0.05, ****P < 0.0001, ns P > 0.05 relative to control. (D) Representative images of ventral nerve cord of stage 16 embryos with indicated genotype stained with anti-FasII to reveal axon fasciculation in developing CNS. Lower panels show magnified region from upper panels. Arrowheads show defasciculation and asterisks show fascicule breaks. Scale bars in first column = 10 µm and apply to all images in the row. (E) Quantitation of number of segments in which fasciculation defects are observed in ventral nerve cord of stage 16 FasII-stained embryos. At least 19 embryos analysed for each genotype. Lines indicate mean values. Data statistically analysed using Kruskal–Wallis with Dunn’s correction for multiple comparisons. ****P < 0.0001, ns P > 0.05 relative to control.

Figure 5

Expression of Rac1-Y64D alters dendritic branching in Drosophila sensory neurons. (A–C) Class IVda sensory neurons from dorsal surface of segments A1–A4 of L3 larvae of indicated genotype. Neurons visualized using CD8-mCherry expressing under the control of ppk-Gal4. Scale bar indicates 40 µm and applies to all three images. (D–F) Magnification of the area around the cell body for the cells shown in A–C. Scale bar = 40 µm and applies to all three images. (G) Sholl analysis of dendritic organization in which the number of times dendrites intercept a semicircle originating at the cell body is plotted against the radius of the semicircle. The semicircle comprises the region of the neuron that is dorsal to the cell body, corresponding to approximately the top half of the images shown in A–C. Graph shows mean ± SEM of ∼15 neurons for each data set. Statistical analysis by two-way ANOVA. ****P < 0.0001, ns P > 0.05 relative to control.