Search

Additional file 1: Supplemental figures S1-S9. Figure S1. Nlgn3−/y rats display reduced classic freezing behaviour in a contextual fear conditioning paradigm. Figure S2. Freezing analysed as “paw immobility response” (all four paws unmoving but allowing for movement of head and neck). Figure S3. WT and Nlgn3−/y rats show similar activity in an open field, rotational platform & marble interaction tasks. Figure S4. Effect of repeated footshocks & thermal stimulus on WT and Nlgn3−/y rats. Figure S5. Intrinsic properties of PAG cells recorded from WT and Nlgn3−/y rats. Figure S6. Hyperexcitability of dorsal, but not ventral PAG neurons in 8-10 week old Nlgn3−/y rats. Figure S7. PAG LFPs during fear recall are significantly shorter duration in Nlgn3−/y rats. Figure S8. Defensive reactions were not elicited by electrical stimulation of primary somatosensory cortex in WT or Nlgn3−/y rats. Figure S9. Western blots showing lack expression of NLGN3 in Nlgn3−/y rats both in sensory cortex and periaqueductal grey.

Additional file 2: Supplemental table 1. Summary of statistics for main figures 2–7. Supplemental table 2. Summary of statistics for supplemental figures 1–7.

Summary The GluN2 subtype (2A versus 2B) determines biophysical properties and signaling of forebrain NMDA receptors (NMDARs). During development, GluN2A becomes incorporated into previously GluN2B-dominated NMDARs. This “switch” is proposed to be driven by distinct features of GluN2 cytoplasmic C-terminal domains (CTDs), including a unique CaMKII interaction site in GluN2B that drives removal from the synapse. However, these models remain untested in the context of endogenous NMDARs. We show that, although mutating the endogenous GluN2B CaMKII site has secondary effects on GluN2B CTD phosphorylation, the developmental changes in NMDAR composition occur normally and measures of plasticity and synaptogenesis are unaffected. Moreover, the switch proceeds normally in mice that have the GluN2A CTD replaced by that of GluN2B and commences without an observable decline in GluN2B levels but is impaired by GluN2A haploinsufficiency. Thus, GluN2A expression levels, and not GluN2 subtype-specific CTD-driven events, are the overriding factor in the developmental switch in NMDAR composition., Graphical Abstract, Highlights • Mutating the GluN2B CaMKII site affects phosphorylation of its C-terminal domain • The developmental changes in NMDAR composition and synaptogenesis occur normally • Changes in NMDAR composition do not require distinct GluN2 C-terminal domains • Developmental changes in NMDAR composition are primarily sensitive to GluN2A levels, An important milestone in forebrain neuronal development is the switch in composition of the NMDA receptor: GluN2A becomes incorporated into previously GluN2B-dominated NMDARs. Using knockin mice, McKay et al. find that, contrary to earlier proposed models, the switch does not require GluN2A and GluN2B to possess distinct C-terminal domain sequences.

The human brain generally remains structurally and functionally sound for many decades, despite the post-mitotic and non-regenerative nature of neurons. This is testament to the brain’s profound capacity for homeostasis: both neurons and glia have in-built mechanisms that enable them to mount adaptive or protective responses to potentially challenging situations, ensuring that cellular viability and functionality is maintained. The high and variable metabolic and mitochondrial activity of neurons places several demands on the brain, including the task of neutralizing the associated reactive oxygen species (ROS) produced, to limit the accumulation of oxidative damage. Astrocytes play a key role in providing antioxidant support to nearby neurons, and redox regulation of the astrocytic Nrf2 pathway represents a powerful homeostatic regulator of the large cohort of Nrf2-regulated antioxidant genes that they express. In contrast, the Nrf2 pathway is weak in neurons, robbing them of this particular homeostatic device. However, many neuronal antioxidant genes are controlled by synaptic activity, enabling activity-dependent increases in ROS production to be offset by enhanced antioxidant capacity of both glutathione and thioredoxin-peroxiredoxin systems. These distinct homeostatic mechanisms in neurons and astrocytes together combine to promote neuronal resistance to oxidative insults. Future investigations into signaling between distinct cell types within the neuro-glial unit are likely to uncover further mechanisms underlying redox homeostasis in the brain., Highlights • Redox homeostasis is essential for brain health. • Homeostasis involves interactions within the neurogliovascular unit. • We describe cooperative contributions by neurons and astrocytes to redox homeostasis.

The human brain generally remains structurally and functionally sound for many decades, despite the post-mitotic and non-regenerative nature of neurons. This is testament to the brain's profound capacity for homeostasis: both neurons and glia have in-built mechanisms that enable them to mount adaptive or protective responses to potentially challenging situations, ensuring that cellular viability and functionality is maintained. The high and variable metabolic and mitochondrial activity of neurons places several demands on the brain, including the task of neutralizing the associated reactive oxygen species (ROS) produced, to limit the accumulation of oxidative damage. Astrocytes play a key role in providing antioxidant support to nearby neurons, and redox regulation of the astrocytic Nrf2 pathway represents a powerful homeostatic regulator of the large cohort of Nrf2-regulated antioxidant genes that they express. In contrast, the Nrf2 pathway is weak in neurons, robbing them of this particular homeostatic device. However, many neuronal antioxidant genes are controlled by synaptic activity, enabling activity-dependent increases in ROS production to be offset by enhanced antioxidant capacity of both glutathione and thioredoxin-peroxiredoxin systems. These distinct homeostatic mechanisms in neurons and astrocytes together combine to promote neuronal resistance to oxidative insults. Future investigations into signaling between distinct cell types within the neuro-glial unit are likely to uncover further mechanisms underlying redox homeostasis in the brain.

How the brain's antioxidant defenses adapt to changing demand is incompletely understood. Here we show that synaptic activity is coupled, via the NMDA receptor (NMDAR), to control of the glutathione antioxidant system. This tunes antioxidant capacity to reflect the elevated needs of an active neuron, guards against future increased demand and maintains redox balance in the brain. This control is mediated via a programme of gene expression changes that boosts the synthesis, recycling and utilization of glutathione, facilitating ROS detoxification and preventing Puma-dependent neuronal apoptosis. Of particular importance to the developing brain is the direct NMDAR-dependent transcriptional control of glutathione biosynthesis, disruption of which can lead to degeneration. Notably, these activity-dependent cell-autonomous mechanisms were found to cooperate with non-cell-autonomous Nrf2-driven support from astrocytes to maintain neuronal GSH levels in the face of oxidative insults. Thus, developmental NMDAR hypofunction and glutathione system deficits, separately implicated in several neurodevelopmental disorders, are mechanistically linked.