Developmental regulation of GABAergic interneuron branching and synaptic development in the prefrontal cortex by soluble neural cell adhesion molecule

Abstract

Neural cell adhesion molecule, NCAM, is an important regulator of neuronal process outgrowth and synaptic plasticity. Transgenic mice that overexpress the soluble NCAM extracellular domain (NCAM-EC) have reduced GABAergic inhibitory and excitatory synapses, and altered behavioral phenotypes. Here, we examined the role of dysregulated NCAM shedding, modeled by overexpression of NCAM-EC, on development of GABAergic basket interneurons in the prefrontal cortex. NCAM-EC overexpression disrupted arborization of basket cells during the major period of axon/dendrite growth, resulting in decreased numbers of GAD65- and synaptophysin-positive perisomatic synapses. NCAM-EC transgenic protein interfered with interneuron branching during early postnatal stages when endogenous polysialylated (PSA) NCAM was converted to non-PSA isoforms. In cortical neuron cultures, soluble NCAM-EC acted as a dominant inhibitor of NCAM-dependent neurite branching and outgrowth. These findings suggested that excess soluble NCAM-EC reduces perisomatic innervation of cortical neurons by perturbing axonal/dendritic branching during cortical development.

Figure 1. Expression of different NCAM isoforms and cleavage products during development

A) NCAM, PSA-NCAM and cleavage fragments during development. Brain homogenates (75 μg) from E18.5 to P40 mice were separated by SDS-PAGE, and blotted using antibodies to PSA (top panel), NCAM-ICD (clone OB11; second panel), and NCAM-ECD (clone H300; third panel). Arrows indicate the size of PSA-NCAM, NCAM180, NCAM140, and cleavage fragments representing the NCAM extracellular domain (105, 110 kDa). NCAM120 expression (bottom panel). Homogenates were blotted using the NCAM-ECD antibody. Actin blots (fourth panel) were used as a loading control. Exposure times were: 1 min (NCAM-ICD, actin), 5 min (PSA, NCAM-ECD bottom panel), 1 h (NCAM-ECD third panel). B) NCAM-EC^tg expression and polysialylation. The left set of panels depicts Western blots of mouse brain lysates. For PSA detection, NCAM-EC^tg was immunoprecipitated using HA antibodies prior to SDS-PAGE and blots are depicted in the right set of panels. Western blots were performed using antibodies to HA (top panels, 1 min exposure), PSA (middle panels, 5 min exposure), and actin (loading control, bottom panels, 1 min exposure).

Figure 2. Development of parvalbumin-positive basket cells in layer II/III of the prefrontal cortex

Individual neurons from layer II/III of the PFC were imaged and reconstructed using confocal microscopy. Representative extended focus images of individual z-stacks from EGFP-labeled parvalbumin-positive basket cells of WT (GAD67-EGFP) and NCAM-EC transgenic (NCAM-EC/GAD67-EGFP) mice are shown from each time point (P10, P20, adult). Arrows indicate a dendrite in each image. Arrowheads indicate axons. Scale bar = 10 μm.

Figure 3. Proximal neuronal elaboration increases during development, but is stunted in NCAM-EC PFC

Four to five animals per genotype at P10, P20 and adult were analyzed, and 30−50 neurons per animal were imaged and traced. A) Mean branching index. B) Mean number of crossings per process at a Sholl distance of 20 μm for each time point. Student's t test, *p<0.05. C) Mean number of primary processes. D) Mean neuronal density. Five random fields in layer II/III were counted for each animal and averaged. At least 100 neurons were counted per animal. E) Mean somal area (μm²).

Figure 4. Perisomatic synapses increase during development, but are reduced in NCAM-ECPFC

A) Representative images from WT (GAD67-EGFP) and NCAM-EC transgenic (NCAM-EC/GAD67-EGFP) PFC, layer II/III. Scale bar = 10 μm. Arrows indicate the location of representative soma. B) Perisomatic synaptic puncta rings were traced for each image using ImageJ and pixel density measured as compared to the soma as a control. Four images were analyzed for each animal and 4−5 animals per genotype per stage were analyzed. *Student's t test p<0.05.

Figure 5. Synapses from other interneurons develop similarly to parvalbumin-positive basket cells, and are reduced in NCAM-EC mice

A) Representative images from layer II/III of the PFC of WT and NCAM-EC mice showing GAD65 immunofluorescence. Arrows indicate the location of representative soma. Nonimmune IgG control is shown as an inset. Scale bar = 10 μm. B) Perisomatic synaptic puncta rings were traced for each image using ImageJ and pixel density measured as compared to the soma and nonimmune IgG as a control. Four images were analyzed for each animal and 3 animals per genotype per stage were analyzed. Student's t test **p<0.005.

Figure 6. Excitatory synapse development is perturbed by NCAM-EC

A) Representative images of synaptophysin immunofluorescence from layer II/III of the PFC of WT and NCAM-EC transgenic mice at P10, P20 and adult stages. Arrows indicate the location of soma. Nonimmune Ig control is included as an inset. Scale bar = 10 μm. B) Pixel density of perisomatic synaptic puncta rings. Four images were analyzed for each animal and 3 animals per genotype per stage were analyzed. Student's t test *p<0.06.

Figure 7. Soluble NCAM extracellular domain inhibits neurite outgrowth and branching in culture

A representative experiment is shown. Four wells for each condition were analyzed and at least 150 neurons were analyzed per condition. A) Cortical neurons were plated onto monolayers of L-fibroblasts or L-fibroblasts expressing NCAM140 and neurites allowed to extend for 72 h in the presence (black bars) or absence (nonimmune Ig, white bars) of NCAM-Fc before being visualized using TUJ1 immunofluorescence. Arrows indicate branched neurites. Scale bar = 50 μm. B) The percent of neurons with branched neurites was calculated. Student's t test *p<0.05, **p<0.005. C) The cumulative percentage of neurons of a given length was determined by subtracting the percent of neurons with length X from 100 as described previously (Beggs 1994; Hinkle 2006).

Figure 8. Model for the mechanism of action of the shed NCAM extracellular domain in interneuronal development

Excess soluble NCAM-EC is depicted to interact with neuronal NCAM to inhibit NCAM-NCAM homophilic binding and heterophilic interactions with receptors such as the FGF receptor (FGFR) (Meiri 1998; Kiselyov 2003; Kiselyov 2005) or receptor protein tyrosine phosphatase alpha (RPTPα, (Bodrikov 2005)), thereby decreasing interneuron arborization to result in fewer perisomatic synapses. Alternatively, in other cell types where metalloprotease-induced soluble NCAM-EC promotes neurite growth (Meiri 1998; Hubschmann 2005; Kalus 2006), transgenic NCAM-EC might bind and inhibit this activity. In normal development, ERK MAP kinase, triggered by an external stimulus, activates an ADAM family protease to cleave NCAM and release NCAM-EC. This serves to limit interneuron branching promoted by NCAM homophilic or heterophilic binding.