Neuroligin-4 Regulates Excitatory Synaptic Transmission in Human Neurons

Abstract

The autism-associated synaptic-adhesion gene Neuroligin-4 (NLGN4) is poorly conserved evolutionarily, limiting conclusions from Nlgn4 mouse models for human cells. Here, we show that the cellular and subcellular expression of human and murine Neuroligin-4 differ, with human Neuroligin-4 primarily expressed in cerebral cortex and localized to excitatory synapses. Overexpression of NLGN4 in human embryonic stem cell-derived neurons resulted in an increase in excitatory synapse numbers but a remarkable decrease in synaptic strength. Human neurons carrying the syndromic autism mutation NLGN4-R704C also formed more excitatory synapses but with increased functional synaptic transmission due to a postsynaptic mechanism, while genetic loss of NLGN4 did not significantly affect synapses in the human neurons analyzed. Thus, the NLGN4-R704C mutation represents a change-of-function mutation. Our work reveals contrasting roles of NLGN4 in human and mouse neurons, suggesting that human evolution has impacted even fundamental cell biological processes generally assumed to be highly conserved.

Keywords: ASD; Neuroligin-4; autism; embryonic stem cells; induced neuronal (iN) cells; isogenic cell lines; synaptic transmission.

Figure 1.. Human NLGN4 is expressed in cortex and localizes to excitatory postsynaptic compartments.

(A) Amino acid conservation between Nlgns in mouse (Nlgn) and human (NLGN). (B) Analysis of NLGN4 protein (immunoblot, top) and RNA (qPCR, bottom) in the adult human brain. Bars represent technical replicates. (C) Expression of NLGN4 in tested areas shown in a representation of the human brain. Dark-red represents high, light-red low, blue not determined. (D) Excitatory neurons are generated by Ngn2; inhibitory neurons are generated by Ascl1 and Dlx2 (AD). (E) NLGN4 protein (top) and RNA (bottom) in Ngn2- and AD-iN cell cultures at day 42. (F) Schematics of gene targeting strategy used to generate NLGN4^HA ES cell line using rAAV vectors. SP: signal peptide. (G) Representative images of NLGN4-HA stain and (H) co-stain with synaptic markers Synapsin-1 and VGluT2. Arrowhead points to a NLGN4/Synapsin1/vGluT2 co-localization. (I) Quantification of HA signal overlapping Synapsin-1 signal. (J) Representative images of co-stain of NLGN4-HA with synaptic markers VGluT2, PSD-95, Homer-1, Gephyrin and VGAT. (K) Co-localization analysis of NLGN4-HA signal with synaptic markers. (L) Representative dSTORM images of single synapses co-stained with NLGN4-HA and Synapsin-1 (top) or PDS-95 (bottom). All images are from endogenously tagged NLGN4^HA knock-in AD iN cells mixed with WT Ngn2 iN cells cultured with mouse glia for five weeks. MAP2 (blue) is used to identify dendrites. Scale bars: 20μm panel G, 5μm for panel H,J and 100nm for panel L. Data are represented as mean ± SEM and N=3. See also Figure S1.

Figure 2.. NLGN4 overexpression in human neurons modulates excitatory synaptic transmission.

(A) Schematic of lentiviral vectors used and immunoblot of iN cells transduced with NLGN4 (NL4) or EGFP (CTRL). (B) Representative images and quantification of puncta density and size in dendrites labeled with Synapsin-1, Homer-1 and MAP2. The arrowhead shows an example of Synapsin1 and Homer1 colocalization. (C) Example traces (left) and quantification for EPSC amplitudes and frequencies (right). (D) Example traces and quantification of amplitude for evoked EPSCs. (E) Quantification of co-efficient of variation (CV) for evoked EPSCs recorded from Ngn2-iN cells with NL4 or CTRL. Analyses are performed at day 35. Scale bars: 20μm (upper panel in B) and 5μm (lower three panels in B). Numbers of neurons/ independent cultures analyzed are shown in the bars. Data are represented as mean ± SEM and N=3. (**, p < 0.01) (*, p < 0.05). See also Figure S2.

Figure 3.. The R704C mutation increases excitatory synapse formation.

(A) EGFP labeled WT, KO, and R704C AD-iN cells sparsely spiked into standardized cultures consisting of mixed Ngn2 and AD-iN cells to minimize variability. (B) Representative image of the cultures (labeled with EGFP and Synapsin-1). (C) Representative images and quantification of puncta density for excitatory (PSD-95) and (D) inhibitory synapses (VGAT) at day 42. Dendrites were labeled with MAP2 (blue). (E) Excitatory neurons expressing NLGN4 protein induced from ES cells by Ascl1 and Myt1l (AM). NLGN4 protein (immunoblot, left) and RNA (qPCR, right) levels in AM-iN cells at day 42. (F) Representative images and (G) quantification of morphology of excitatory AM-iN and inhibitory AD-iN cells differentiated from WT, KO and R704C sparsely transfected with EGFP for visualization of cellular processes belonging to a single cell in high density cultures. (H) Representative images and quantification of puncta density and size for Synapsin-1 in dendrites immunolabeled with MAP2 from WT, KO and R704C in AM-iN and AD-iN cells. Scale bars: 20μm for panels B, 50μm for panels F and 5pm for panels C,H. Numbers of total neurons/independent cultures analyzed are shown in the bars. Data are represented as mean ± SEM and N=3. (***, p < 0.001). See also Figure S3.

Figure 4.. Increased strength of excitatory synapses of NLGN4^R704C human neurons.

(A) Representative images of patch clamping of EGFP labeled AD-iN cells in mixed cultures. (B) Quantification of intrinsic properties of spiked-in AD-iN cells in mixed cultures differentiated from wild-type H1 (WT), NLGN4^KO (KO) and NLGN4^R704C (R704C) cells. (C) Example mPSCs traces recorded in the presence of TTX, (D) quantification of mIPSCs amplitudes and frequencies, (E) quantification of mEPSCs amplitudes and frequencies from spiked-in AD-iN cells WT, KO and R704C at day42. (F) Example traces and quantification of sEPSCs and (G) evoked EPSC amplitudes recorded from AM-iN cells WT, KO and R704C cells at day42. (H) Representative images of surface (live staining) and total HA signal (fixed staining) in day 28 Ngn2 iN cells expressing NGLN4-HA or NLGN4^R704C-HA. (I) Quantification of surface localized NLGN4 relative to total NLGN4 in Ngn2 iN cells (left) expressing NGLN4-HA (WT) or NLGN4^R704C-HA (R704C). Right panel, immunoblotting from the same cultures shows comparable levels of total NLGN4 protein in WT and R704C. (J) NLGN4 co-immunoprecipitates with GluA1 and PSD-95. The R704C mutation enhances co-immunoprecipitation of NLGN4 with GluA1. Protein lysates from Ngn2 iN cells expressing NLGN4 WT, R704C or EGFP (Ctrl) immunoprecipitated with HA antibodies and blotted for the AMPAR-GluA1 and for PSD-95. Left, quantification of the relative levels of GluA1 and PSD95 in the immunoprecipitates. Scale bars: 50μm for panel A and upper panel H and 10μm lower panel H. Data are represented as mean ± SEM and N=3. Numbers of neurons/ independent cultures analyzed are shown in the bars. (*, p < 0.05) (***, p < 0.001). See also Figure S4.