Abstract
Commentary
Insights into the impact of disrupting the delicate ratio of excitatory and inhibitory neurons came unexpectedly from mice with targeted mutations of the gene Plaur, encoding the urokinase plasminogen activator receptor (uPAR) (4,5). When uPAR binds its ligand, uPA, the resultant proteolytic cascade activates HGF/SF. Loss of uPAR expression in Plaur-null mice diminishes HGF/SF levels. These mice show regional deficits in cortical GABAergic interneurons, spontaneous seizures, lower thresholds for chemically induced convulsions, and anxiety-related behaviors. Plaur-null mice survive into adulthood, making this mouse model ideal for exploring the roles of HGF/SF and its receptor Met (Met proto-oncogene or hepatocyte growth factor receptor) in interneuron migration, as Hgf- or Met-null mice die during embryonic life (6). HGF was initially identified in the liver, where it promotes hepatocyte proliferation and regeneration by binding to MET, a tyrosine kinase receptor (7,8). Scatter factor is a molecule produced by fibroblasts that initiates cell dispersion in epithelial-derived cells.
One remarkable finding in the Plaur-null mice is that the attendant decrease in HGF/SF levels chiefly impacts interneurons expressing the calcium-buffering protein parvalbumin (PV), while leaving unscathed the calretinin or somatostatin-expressing interneurons. The mice show a 50% drop in PV-immunoreactive (PV+) interneurons in the parietal and cingulate cortices, while PV+ interneurons in other cortical areas are unaltered. The cortical and hippocampal PV+ interneurons comprise approximately 20% of the entire forebrain's GABAergic cells and are well described. Within the cerebrum and hippocampus, PV+ interneurons exhibit fast-spiking action potentials, form perisomatic synaptic baskets onto neighboring neurons, and are responsible for generating gamma oscillations in the hippocampus. Experimentally disrupting these interneurons impairs hippocampal-related tasks, such as spatial working memory and novel object recognition (9).
A second key observation by Bae et al. in Plaur mice is that the PV+ interneuron deficiency is limited to cingulate and parietal cortex (10). Several possible explanations are apparent. These areas might require higher levels of HGF/SF activity to guide migrating PV+ interneurons, or other types of interneurons might use different migratory cues. To explore the underlying mechanisms, Bae et al. tested whether HGF/SF overexpression could rescue the mutant phenotype of Plaur-null mice. A transgenic mouse line expressing human HGF/SF under the control of the mouse glial fibrillary acidic protein (GFAP) promoter (Gfap-HGF mice) was crossed to the Plaur-null mice (HGF/Plaur-null mice), to drive expression of HGF/SF postnatally. Their results demonstrate a striking rescue of GABAergic interneurons in cingulate and parietal cortices in HGF/Plaur-null mice.
Because PV is regulated in an activity-dependent manner, it was necessary to demonstrate that Plaur mice show a true deficit in PV+ interneurons, rather than reduced PV expression. The authors used an alternative method for identifying PV+ interneurons based on the presence of “perineuronal nets” that are fenestrated sheaths surrounding PV+ basket cell somas and initial axon segments. The nets contain extracellular matrix molecules with a special carbohydrate group, called N-acetylgalactosamine, which may stabilize synapses during postnatal critical periods. Staining brain sections with plant lectins conjugated to a fluorescent molecule or peroxidase allows visualization of these nets with light microscopy. Lectin staining of Plaur-null mouse brains verified that perineuronal nets are also absent. These experiments show that the apparent deficit in PV+ interneurons stems from an actual reduction in the interneurons, rather than a failure of these neurons to express detectable amounts of PV. Bae et al. then established that the perineuronal nets surrounding basket cells are normally formed in the HGF/Plaur-null mice, confirming that genetically increasing levels of HGF/SF not only establishes molecular guidance cues for PV+ interneurons but also allows them to develop extracellular matrix constituents that may be involved in synaptic stabilization.
In rodent epilepsy models, seizure susceptibility is often evaluated with a single threshold injection of pentylenetetrazole. Functional deficits in inhibitory networks can be unmasked with a subthreshold dose of pentylenetetrazole. While the majority of Plaur mice develop motor seizures after doses of pentylenetetrazole that are normally at subthreshold, less than 20% of the HGF or the HGF/Plaur mice showed seizures. Interestingly, Plaur mice exhibit abnormal baseline cortical activity, with episodes of high-amplitude spiking. While the significance of the abnormal interictal spikes is unknown, this form of augmented neuronal activity may be responsible for increasing seizure susceptibility. Although the EEGs are not completely normal in the HGF/Plaur mice, abnormal spikes are substantially reduced. These results suggest that postnatal supplementation of HFG/SF is sufficient to reduce seizure susceptibility and severity. Whether there are residual deficits in the ability of PV+ interneurons to orchestrate gamma oscillations is an open question raised by this work.
In addition to frequent seizures, Plaur mice demonstrate increased anxiety in behavioral tests. One test developed by Crawley and colleagues, measures time spent exploring a brightly lit arena or escaping into a darkened box (11). Higher anxiety is associated with spending more time in the safe dark box instead of exploring the arena. Wild-type mice show a slight preference for the dark box, avoiding the brightly lit areas 55% of the time. Plaur mice avoid the brightly lit areas more than 70% of the time. Notably, HGF/Plaur mice are indistinguishable from wild-type mice in this measure of anxiety. Similar findings were obtained following elevated plus-maze testing.
Taken together, the study by Bae et al., in combination with previous work, indicates that replenishing HGF/SF perinatally not only enables immature PV+ interneurons to complete their migrations in a Plaur mutant background, it also eliminates an anxiety-like phenotype and reduces EEG abnormalities. Because HGF/SF-MET and uPA signaling are important determinants in cell motility, future studies may provide mechanistic explanations for how interneurons migrate into specific cortical layers and areas. As additional experimental tools are developed to systematically manipulate putative guidance molecules for GABAergic interneurons, it can be expected that the unique roles played by different types of interneurons in mood, affect, and memory will be ascertained.
