The Nogo Receptor Ligand LGI1 Regulates Synapse Number and Synaptic Activity in Hippocampal and Cortical Neurons

Exogenous application of LGI1 promotes synapse formation. A, Representative images from 18 DIV cultures hippocampal neuron labeled with Syn1 (red), PSD95 (green), and MAP2 (blue) in the absence and presence of exogenous LGI1 (6 DIV treatment). B, Magnification of sample images; the panel labeled “count” indicates colocalized particles of Syn1 and PSD95 considered to be synaptic puncta. C, Quantification of synaptic density indicated by colocalized puncta per µm of MAP2-labeled dendrites excluding the cell bodies. Data are shown as box plots with whiskers from 1^st to 99th percentile. Differences were analyzed by unpaired t tests, ***p = 0.0007; N = 34 WT and 29 treatment neurons from three separate cultures (8–14 images were included for each condition in each experiment).

NgR1 regulates synapse number in dissociated hippocampal neurons. Analysis of synapse density from cultured hippocampal neurons grown on coverslips for 15 or 18 DIV. A, Representative images from 18 DIV cultures labeled with Syn1 (red), PSD95 (green), and MAP2 (blue), genotypes are indicated. B, Magnification of sample images shown in the white box in A used for quantification. The panel labeled “count” indicates colocalized particles of Syn1 and PSD95 considered to be synaptic puncta. C, Quantification of synaptic density indicated by colocalized puncta per µm of MAP2-labeled dendrites excluding the cell bodies for 15 DIV cultures. D, Same as C but for 18 DIV cultures. Values in quantification were normalized to WT littermates within each culture. Data are shown as Box plots with whiskers from 1st to 99th percentile. NgR1^-/- neurons display significantly more synapses than NgR1^+/+ littermates. For 15 DIV cultures, NgR1^+/+ n = 29 neurons and NgR1^-/- n = 28 neurons from three cultures. For 18 DIV cultures, NgR1^+/+ n = 25 neurons and NgR1^-/- n = 31 neurons from three separate cultures. Differences were analyzed by unpaired t tests, ***p < 0.001.

LGI1 regulates synapse number in dissociated hippocampal neurons. Analysis of synapse density from hippocampal neurons dissociated from E16 embryos grown on coverslips for 15 or 18 DIV. Panels A, B show representative images from 18 DIV cultures labeled with Syn1 (red), PSD95 (green), and MAP2 (blue), genotypes are indicated. Sample images of LGI1^+/+ and LGI1^-/- neurons used in analysis of synaptic density, quantified from the number of colocalized Syn1 and PSD95 puncta per μm of MAP2-labeled dendrites excluding the cell bodies. B, Enlarged images from the box highlighted in A. The far right panel indicated “count” shows the colocalized puncta representing the synapse. Values in quantification were normalized to WT littermates within each culture. Data are shown as box plots with whiskers from 1st to 99th percentile. LGI1^-/- neurons have significantly fewer synapses than LGI1^+/+ littermates. C, Synapse density quantified from 15 DIV cultures. D, Synapse density quantified from 18 DIV cultures. LGI1^+/+ n = 40 neurons from four cultures, LGI1^-/- n = 40 from four cultures; ***p < 0.001 (unpaired t test).

NgR1 and LGI1 regulate synaptic proteins in cortical neurons in vitro. A, Twiss filter schematic showing culture system to coculture hippocampal neurons with astrocytes and separate neuronal processes from cell bodies. Hippocampal neurons seeded on filters with a pore size 1 µm that cell bodies will not pass through. Axons and dendrites grow on the filter tops and extend down onto the filter bottom. Astrocytes are seeded on the bottom of the well to provide growth factors. B, Time course of lysates from hippocampal neurons grown on filters suspended over an astrocyte feeder layer for the times indicated. The first lane in the left panel labeled E16 is a sample of hippocampal neurons lysed directly after dissociated before plating. Lysates from filter tops including cell bodies and processes are on the left. Lysates of the filter bottoms containing axons and dendrites but no cell bodies are on the right. Antibodies used to probe the lysates are indicated on the right. Histone-3 (H3), a structural protein found in chromatin and present only in the nucleus is detected only in the cell body lysates. C, Lysates from filter bottoms containing axons and dendrite but not cell bodies from LGI1^+/+ and LGI1^-/- littermates of cortical cultures grown for the indicated number of DIV. D, Quantification of PSD95 levels relative to actin levels and normalized to WT controls in LGI1 samples at 12, 15, and 18 DIV, n = 3 separate experiments. E, Western blottings of lysates from filter bottoms of NgR1^+/+ and NgR1^-/- cortical cultures harvested at 12, 15, or 18 DIV synaptic markers, Syn and PSD95. Actin and Tuj1 are loading controls. F, Quantification of PSD95 relative to actin levels and normalized to WT controls in NgR1, n = 4 separate experiments. Significant differences are indicated on the graphs analysis was performed by two-way ANOVA with Bonferroni post hoc tests, **p < 0.01, *p < 0.05.

LGI1 normally suppresses RhoA activity in cultured cortical neurons. A total of 15 DIV cortical neurons from LGI1^-/- mice have increased RhoA activation compared to WT cultures. A, Western blottings of input RhoA and active RhoA bound to GST-rhotekin beads (GST-RBD) that bind to GTP-RhoA. The lower panel corresponds to the membrane labeled with Ponceau showing total GST protein. B, Quantification of active GTP-RhoA relative to total RhoA input levels and normalized WT values for each of four experiments. Active GTP-RhoA is significantly higher in lysates from LGI1^-/- cortical neurons compared to LGI1^+/+ controls, 71% higher, p = 0.0257, analyzed by unpaired t test; N = 4 separate experiments.

LGI1 regulates a NgR1-TROY receptor complex. A, Sample images of COS7 cells transfected with control GFP only; NgR1; TROY; NgR1 + TROY: LGI1; NgR1 + LGI1; TROY + LGI1; NgR1 + TROY + LGI1, together labeled with rhodamine-WGA. COS7 cells transfected with the indicated plasmids and seeded at low density on glass coverslips. In all conditions GFP was cotransfected to follow the expression of the other proteins. Cell membranes were labeled with rhodamine-tagged WGA to visualize cells and measure cell area. B, Area of the cells was quantified using ImageJ Analyze Particles tools. The graph indicates the effect of expression of no receptors (GFP control), each receptor alone, or with LGI1 (indicated in red). Expression of LGI1, NgR1, or TROY separately in COS7 cells has no effect on cell size compared to GFP alone. However, NgR1 and TROY expression together decrease cell size compared to GFP expression alone, p < 0.05 indicated by # on the graph. Coexpression of LGI1 together with NgR1 and TROY rescues cell size. This condition is not significantly different from GFP alone. Analysis by two-way ANOVA followed by Bonferroni post hoc tests, ***p < 0.001; n = 5 where each n is the average area in one separate experiment. For each experiment the areas of 50–100 cells were quantified.

Dendritic spine density and morphology is regulated by LGI1 in vivo. A, Sample images of dendritic spines labeled with DiI from acute hippocampal slices from P10 LGI1^+/+ (left) and LGI1^-/- (right) mice. Segments were taken from the primary bifurcation point from CA1 pyramidal neurons shown in the upper images. The selected region if dendrite is shown below with arrows indicating mushroom spines (yellow), stubby spines (blue, thin arrowheads), and thin spines (green, wide arrowheads). Images were manually traced using the measure tool in ImageJ. B, Quantification of total spine-like protrusion density in spines per µm of dendrite. There is a non-significant trend toward fewer spine-link protrusions in LGI1^-/- mice and WT controls (unpaired t test, p = 0.17). C, Proportion of spines categorized based on ratios of the width of the spine head (Hw) to width of the spine neck (Nw) to the spine length, mushroom spines (Hw:Nw > 1.5), stubby spines (Hw:Nw < 1.5 and L:Nw < 1.5), and thin spines (Hw:Nw < 1.5 and L:Nw > 1.5). The proportion of mushroom spines is significantly less in LGI1^-/- slices compared to WT controls (p = 0.03), stubby spines trend toward an increase (p = 0.08), and the proportion of thin spines is unchanged, analyzed by two-way ANOVA with Bonferroni post hoc tests. D, Quantification of mushroom type protrusions per µm of dendrite. The density of mushroom type spines is significantly decreased in LGI1^-/- slices compared to WT controls, compared by Student’s t test (*p = 0.04). E, Measurements of spine length, neck width, and head width for total spines, significant differences indicated on the graph (*p = 0.031), analyzed by two-way ANOVA with Bonferroni post hoc tests. F, Measurements of spine length, neck width, and head width for mushroom type spines, significant differences indicated on the graph (*p = 0.034) analyzed by two-way ANOVA with Bonferroni post hoc tests. For spine analysis n = 6 animals for each genotype, littermates from three separate experiments, four to six spine segments per animal were quantified.