Nonionotropic NMDA Receptor Signaling in Spine Shrinkage
Ivar S. Stein, Deborah K. Park, Juan C. Flores, Jennifer N. Jahncke, and Karen Zito
(see pages 3741–3750)
Neural circuits are continually modified by strengthening and weakening of synaptic strength, accompanied by enlargement and shrinkage of dendritic spines. Somewhat surprisingly, these opposing effects can both be driven by glutamate binding to NMDA receptors (NMDARs). If binding occurs with sufficient postsynaptic depolarization to remove the magnesium block from NMDAR channels, calcium rushes into cells and activates signaling pathways that drive long-term potentiation and spine growth. Without sufficient depolarization, however, intracellular calcium levels remain low and signaling pathways that drive long-term depression (LTD) and spine shrinkage are activated. Although LTD was originally thought to depend on low levels of calcium influx, recent studies have indicated that no ion flux is needed for NMDAR activation to produce LTD and spine shrinkage. Instead, glutamate binding to blocked NMDARs is thought to produce conformational changes that initiate signaling through p38 MAPK.
Stein et al. have now identified several other molecules that contribute to LTD and spine shrinkage after glutamate binds to NMDARs without evoking ion flux. High-frequency glutamate uncaging near single spines—a treatment that normally induces spine growth—induced spine shrinkage and LTD in the presence of drugs that block ion flow through NMDAR channels by binding to the receptor's glycine binding site. As expected, spine shrinkage was prevented by blocking glutamate binding to NMDARs and by inhibiting p38 MAPK. Spine shrinkage was also prevented by interfering with various molecules previously linked to NMDARs, p38 MAPK, and/or spine morphology. Specifically, spine shrinkage was prevented by: blocking neuronal nitric oxide synthase (nNOS); preventing nNOS interaction with the adaptor protein NOS1AP, which is required for NMDAR-dependent activation of p38 MAPK; inhibiting MAPK-activated protein kinase 2 (MK2), a substrate of p38 MAPK; knocking down cofilin, an actin-depolarizing protein; or inhibiting CAMKII.
These results provide additional evidence that when glutamate binds to NMDARs without evoking calcium influx, it induces spine shrinkage. This effect likely results from conformational changes in NMDARs that lead to interactions between nNOS and NOS1AP, local activation of p38 MAPK, downstream activation of MK2, and ultimately, activation of cofilin. Cofilin then depolymerizes local actin filaments, resulting in spine shrinkage. Future work will need to clarify the role of CAMKII and determine how MK2 leads to activation of cofilin.
When ion flow through NMDARs is blocked, high-frequency glutamate uncaging at single spines (left) leads to spine shrinkage within 30 min (right), in both the presence (bottom) and absence (top) of an AMPA receptor antagonist. See Stein et al. for details.
Role of Mouse Superior Colliculus in Visual Perception
Lupeng Wang, Kerry McAlonan, Sheridan Goldstein, Charles R. Gerfen, and Richard J. Krauzlis
(see pages 3768–3782)
The superior colliculus (SC) is a midbrain structure involved in orienting the head and eyes toward salient stimuli and in visual attention and target selection. Notably, nearly 90% of retinal ganglion cells project to the SC in mice, whereas only ∼10% of these cells project to the SC in nonhuman primates. This suggests that the SC contributes more to visual processing in mice than in primates. Consistent with this, many neurons in mouse SC have response properties similar to those found in primary visual cortex (V1) in other species. In particular, mouse SC neurons exhibit two properties required for image formation: orientation selectivity and a preference for high spatial frequencies (Cang et al., 2018, Ann Rev Vis Sci 4:239).
To elucidate the role of mouse SC in visual perception, Wang et al. trained mice to lick a spout to receive a water reward when the orientation of a drifting visual grating changed. Electrophysiological recordings showed that the firing rate of ∼43% of SC neurons increased within 50–150 ms after an orientation change in gratings presented in the contralateral visual field. Moreover, unilateral inhibition of SC (through activation of channelrhodopsin-expressing inhibitory neurons) during this 50–150 ms period significantly decreased detection of contralateral orientation shifts. In contrast, inhibiting SC <50 or >150 ms after the orientation shift had minimal effects on detection. Construction of psychometric curves confirmed that inhibiting SC decreased the detection threshold for orientation changes occurring in the contralateral visual field. Notably, however, the effects of SC inhibition were greatest when gratings were presented to both visual hemifields, consistent with previously described roles for SC in visual target selection and attention. In addition, SC inhibition increased guess rates and false-arm rates for stimuli presented either ipsilaterally or contralaterally, suggesting it increased response bias.
These results demonstrate that mouse SC is necessary for some forms of visual perception, as well as for visual target selection. This supports the hypothesis that mouse SC serves some of the functions performed by V1 in primates. Future work should examine the extent to which SC projections to V1 versus projections to other cortical and subcortical targets contribute to visual perception in mice.
Footnotes
This Week in The Journal was written by Teresa Esch, Ph.D.
