Cortical inhibitory but not excitatory synaptic transmission and circuit refinement are altered after the deletion of NMDA receptors during early development

Deng, Rongkang; Chang, Minzi; Kao, Joseph P. Y.; Kanold, Patrick O.

doi:10.1038/s41598-023-27536-0

Download PDF

Article
Open access
Published: 12 January 2023

Cortical inhibitory but not excitatory synaptic transmission and circuit refinement are altered after the deletion of NMDA receptors during early development

Rongkang Deng^2,3,
Minzi Chang¹,
Joseph P. Y. Kao⁴ &
…
Patrick O. Kanold ORCID: orcid.org/0000-0002-7529-5435^1,2

Scientific Reports volume 13, Article number: 656 (2023) Cite this article

1768 Accesses
2 Altmetric
Metrics details

Subjects

Abstract

Neurons in the cerebral cortex form excitatory and inhibitory circuits with specific laminar locations. The mechanisms underlying the development of these spatially specific circuits is not fully understood. To test if postsynaptic N-methyl-D-aspartate (NMDA) receptors on excitatory neurons are required for the development of specific circuits to these neurons, we genetically ablated NMDA receptors from a subset of excitatory neurons in the temporal association cortex (TeA) through in utero electroporation and assessed the intracortical circuits connecting to L5 neurons through in vitro whole-cell patch clamp recordings coupled with laser-scanning photostimulation (LSPS). In NMDAR knockout neurons, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor-mediated connections were largely intact. In contrast both LSPS and mini-IPSC recordings revealed that γ-aminobutyric acid type A (GABA_A) receptor-mediated connections were impaired in NMDAR knockout neurons. These results suggest that postsynaptic NMDA receptors are important for the development of GABAergic circuits.

Developmental loss of ErbB4 in PV interneurons disrupts state-dependent cortical circuit dynamics

Article 17 April 2023

AMPA receptor deletion in developing MGE-derived hippocampal interneurons causes a redistribution of excitatory synapses and attenuates postnatal network oscillatory activity

Article Open access 28 January 2020

A critical period of neuronal activity results in aberrant neurogenesis rewiring hippocampal circuitry in a mouse model of epilepsy

Article Open access 03 March 2021

Introduction

Neurons form intricate and highly specific connections that are important for neural computation. For example, neurons in the neocortex form canonical circuits that are thought to underlie higher cognitive functions performed by the cortex^1,2. Sensory information processing involves sensorily evoked responses in both primary cortices as well as in higher order areas such as the temporal association area (TeA)^3,4,5,6. In particular, TeA is in close proximity to the primary auditory cortex (A1), receives dominant inputs from A1, and is involved in auditorily driven plasticity^7,8. Abnormal connections in the TeA during early development is a potential substrate of neurological disorder such as autism spectrum disorders^9,10. Yet the synaptic development in TeA has not been well studied.

NMDA receptors are critically involved in normal cortical development^11,12,13. Unlike mature neurons, in which ionotropic glutamatergic transmission is mediated by both AMPA and NMDA receptors (AMPAR and NMDAR), immature neurons have a higher level of NMDARs relative to AMPARs^14,15,16,17 or lack AMPARs completely^{18,19,20,21,22,23,24,25}. The NMDAR-only connections are thought to be ‘silent’, because activation of NMDARs requires membrane depolarization to relieve magnesium blockade of NMDAR channels^26,27,28. During development, AMPARs gradually increase and ‘unsilence’ the NMDAR-only connections and this process is thought to be mediated by the coordinated action of excitatory GABAergic synaptic transmission and activation of NMDARs in the immature neurons^29,30,31. Moreover, because activation of NMDARs and their downstream signaling is part of the mechanism for recruiting AMPARs into the postsynaptic sites during long-term potentiation³², NMDARs may play a similar role in shaping the development of AMPAR circuits^18,20.

Fast inhibitory synaptic communication in the cortex is mediated by ionotropic GABA_A receptors (GABA_AR) and malfunctioning of GABAergic synaptic transmission is thought to underlie various neurological disorders^33,34. Thus, the mechanisms regulating the development of GABAergic synaptic connections are important. NMDARs are crucial for the development of GABAergic synaptic connections^35,36,37. Genetic deletion of NMDARs leads to reduction of GABAergic synaptic connections that depends on the calcium signaling through the NMDARs^38,39,40. This evidence suggests that NMDAR signaling is indispensable for the development of GABAergic synaptic connections. Cortical neurons are embedded in microcircuits receiving glutamatergic and GABAergic inputs from specific presynaptic neurons located in specific layers¹, and these circuits emerge over development^{23,25,41,42,43,44}. To date, it is unknown if NMDAR signaling in postsynaptic cells contributes to the establishment of the specific laminar microcircuits impinging on them.

As disrupted NMDAR functions and abnormal neural circuits in the temporal cortex are both involved in autism spectrum disorders^9,10,45, we hypothesized that disrupting NMDAR functions could hinder the development of the glutamatergic and GABAergic circuits onto neurons in TeA. We here tested this hypothesis directly by creating a mosaic cortical deletion of the obligatory NMDAR1 (GluN1) subunit of the NMDA receptors in subsets of pyramidal neurons in TeA through in utero electroporation of Cre plasmid into the cerebral cortex of Grin1^flox mice. We subsequently investigated the development of AMPAR- and GABA_AR-mediated connections through in vitro electrophysiology in brain slices. We found that NMDARs are not required for the development of action potential-evoked and spontaneous synaptic transmission and laminar circuits mediated by AMPARs. In contrast, we observed a decrease in synaptic transmission mediated by GABA_ARs. Thus, postsynaptic NMDARs are required for the normal development of inhibitory but not excitatory synaptic transmission and circuits onto pyramidal cells.

Results

Here, we set out to test if NMDARs are required for the maturation of spatially specific AMPAR- and GABA_AR-mediated connections in TeA. We focused on early-developing excitatory neurons in layer 5 (L5).

Development of AMPAR-mediated connections onto layer 5 excitatory neurons

Cortical circuits undergo extensive changes during the first few postnatal weeks in mice, including an increase of AMPAR-mediated connections and maturation of the specific laminar connections^{23,25,41,43,44,46,47,48}. To verify that neurons in TeA followed a trajectory similar to primary areas⁴⁴, we first investigated the postnatal development of AMPAR-mediated connections onto L5 excitatory neurons. To identify the sources of functional excitatory and inhibitory inputs we combined whole-cell voltage clamp recording in brain slices with laser-scanning photostimulation (LSPS) to activate presynaptic neurons (Fig. 1A and B). AMPAR-mediated currents were recorded by holding cells at a resting membrane potential of −70 mV. When LSPS released glutamate near the soma of the recorded neuron, glutamate receptors on the recorded cells could be directly activated and the responses can mask the monosynaptically evoked EPSCs we want to study. We distinguished direct activation of the recorded cell from monosynaptically evoked EPSCs by the latency of the evoked currents.The identified monosynaptically evoked EPSCs were sensitive to TTX^25,41,44,49 (Fig. 1B). For each recorded neuron, we constructed a connection map showing stimulus locations that gave rise to an evoked EPSC (Fig. 1C). To identify the spatial pattern of connections over the population of recorded cells, we aligned the individual connection maps to the pia and ventricle locations and computed the fraction of cells that received inputs from each location, resulting in a spatial probability map or average input map (Fig. 1D). The average input maps for cells recorded at P6-7 and P12-14 showed that over development L5 excitatory neurons received more excitatory connections from locations further away from the recorded cells (Fig. 1E). Quantification of the numbers of inputs of each neuron indicated a significant increase in the total number of connected locations at the end of the second postnatal week (P12-14) (Fig. 1F left). To identify which laminar location contributed to the developmental change, we separated the cortex into 5 equal bins along the laminar direction and calculated the number of connected locations in each laminar group. The analysis showed that the developmental increase of AMPAR-mediated connections observed during the second postnatal week came largely from the lower layers (Fig. 1F right). Accordingly, the peak amplitude and transferred charge also increased for connections coming from lower layers (Fig. 1G and H). Additionally, the spiking responses of cortical neurons to our stimulation protocol are largely constant during postnatal development⁴⁴, suggesting that the observed changes in the average maps between P6-7 and P12-14 reflect functional circuit changes, not changes in the electrical properties of single neurons. Together, the results suggest a developmental increase of AMPAR-mediated connections onto L5 excitatory neurons.

Layer 5 neurons have NMDAR-only synapses at P7

Immature neurons in multiple brain regions have excitatory synapses with only NMDARs^23,24,25,41. We thus verified whether L5 excitatory neurons in the TeA also have NMDAR-only connections at young ages. We identified NMDAR-only connections using LSPS^25,41. AMPAR-mediated connections were recorded at a holding potential of -70 mV, while both AMPAR- and NMDAR-mediated connections were recorded at a holding potential of + 40 mV (Fig. 2A). Connections were classified as NMDAR-only if EPSCs were only detected at a holding potential of + 40 mV (Fig. 2B). We then computed the average input map for NMDAR-only connections for the population of cells at P6-7. This average input map showed that L5 neurons at P6-7 received NMDAR-only connections from stimulus locations that were located further away in the translaminar and intralaminar directions (Fig. 2C). Separating the laminar location of the connections revealed that lower layers tended to provide more NMDAR-only connections compared to AMPAR-mediated connections (Fig. 2D). Connections coming from lower layers had more NMDAR-only connections as quantified by the NMDAR-only to AMPAR-mediated connection index (Fig. 2E). By the end of the second postnatal week, more AMPAR-positive connections emerged from lower layers (Fig. 1E and F).

LSPS can only detect NMDAR-only connections that are outside of the regions with AMPAR-mediated connections, because we have to exclude regions with AMPAR-mediated connections. As an alternative, we used a minimal stimulation protocol to confirm the existence of NMDAR-only synapses in young animals (P6-P10). We performed minimal electrical stimulation of local axon terminals while recording EPSCs in young L5 neurons (Fig. 2F). We adjusted the stimulation strength such that around 70% of stimulation trials fails to evoke EPSCs in the recorded neurons¹⁹. Holding the neuron at + 40 mV revealed more synaptic connections, as evidenced by increased EPSC events and a decrease in the response failure rate (Fig. 2H). The increase in EPSCs recorded at + 40 mV was blocked by D-APV, an NMDAR blocker (Fig. 2G), suggesting that the newly revealed connections at + 40 mV holding potential were mediated by NMDAR-only synapses. Together, the results demonstrate that young L5 excitatory neurons have extensive NMDAR-only connections.

Absence of NMDAR-mediated currents in Grin1 knockout neurons

To investigate the role of NMDARs in the development of synaptic connections, we generated mosaic Grin1 (encoding the obligatory GluN1 subunit) knockout of excitatory neurons through in utero electroporation of Cdk5r-Cre-IRES-GFP plasmid in Grin1^flox pups at embryonic day (E)12.5 or E13.5. The Cdk5r promoter was used to restrict the deletion to postmitotic pyramidal neurons⁵⁰. After electroporation, the targeted region (i.e., TeA) contained a small population of transfected neurons (GFP-positive, GFP + , Grin1 knockout) and abundant non-transfected neurons (GFP-negative, GFP-, Grin1^flox control) (Fig. 3A). To test if GFP + neurons lacked NMDAR-mediated currents, we recorded EPSCs in response to focal glutamate uncaging onto both kinds of neurons at holding potentials of −70 mV and + 40 mV to delineate both AMPAR- and NMDAR-mediated currents (Fig. 3B and C). At P6-7, the GFP + , putative Grin1 knockout (KO) neurons had intact AMPAR-mediated responses (peak amplitude in Grin1^flox control: median 187 pA, IQR 152 pA, n = 13 cells; Grin1 KO: median 163 pA, IQR 240 pA, n = 8 cells. P = 1, Wilcoxon rank-sum test). As expected, GFP + neurons had greatly reduced responses at holding potential of + 40 mV compared to GFP- neurons in the same slice (Fig. 3D). In particular, long-lasting currents were abolished and remaining currents were similar to currents observed at −70 mV (Fig. 3D right). These results indicate a lack of NMDARs in GFP + neurons. We quantified the effect of electroporation by calculating the ratio of peak amplitudes between currents recorded at + 40 mV and −70 mV holding potentials. The peak current ratio is greatly reduced in GFP-positive neurons (Fig. 3E). Moreover, in GFP + neurons, the remaining current at + 40 mV holding potential was completely blocked by NBQX, the AMPAR-specific blocker (Fig. 3D right), indicating that the remaining current at + 40 mV was carried through AMPARs. Together, these data confirmed that our electroporation approach completely abolished NMDAR-mediated currents in GFP + neurons.

Development of AMPAR-mediated connections is largely unchanged in the absence of NMDARs

We next investigated if the development of translaminar connections mediated by AMPARs (Fig. 1E) requires NMDARs. We electroporated Grin1^flox pups at E12.5 or E13.5 and investigated the locations of AMPAR-mediated connections to L5 neurons in the TeA at P12-P14 through whole-cell voltage clamp recording and LSPS (Fig. 4A). Since we generated mosaic Grin1 KO slices through in utero electroporation, we recorded GFP + Grin1 KO neurons (Grin1 KO), as well as GFP- Grin1^flox control neurons (Grin1^flox control) in the vicinity of the Grin1 KO neurons. As over-expressing Cre and GFP might alter neuronal development, we performed an additional control by also electroporating C57BL/6 J wildtype pups with the same plasmid at E12.5 or E13.5 and then recording from GFP + neurons (WT control) (Fig. 4A). Thus, all three groups of cells were from pups that underwent the same surgical manipulations. Grin1 KO and WT control both expressed Cre and GFP (thus controlling for Cre and GFP expression). Both Grin1^flox control and WT control neurons had intact NMDAR currents, and were compared to Grin1 KO neurons to test the requirement of NMDARs during cortical development.

Proper interpretation of LSPS mapping experiments requires that the presynaptic response to glutamate uncaging remains similar between experimental conditions. Changes in a neuron’s response to glutamate uncaging may change the spatial resolution of the stimulus and confound the interpretation of the results. To rule out this possibility and to ensure that neurons were activated equally, we performed cell-attached recordings (Fig. 4B) from both transfected and non-transfected neurons across all layers in transfected slices from Grin1^flox and WT animals while performing glutamate uncaging. We divided neurons into two groups (i.e., upper and lower layers) based on their soma locations in relation to the middle of the cortex. To measure the spatial resolution of our stimulus, we quantified the total number of stimulus locations where action potentials were evoked by our stimulus (Fig. 4C) and the distance from the soma that encompasses 80% of those locations (Fig. 4D). We found no significant difference between Grin1 KO slices and WT control slices, suggesting that the spatial resolution of our stimulus did not change. Small changes were detected in the average number of evoked action potentials from neurons in the lower layers (Fig. 4E) and in the average first spike latency from neurons in the upper layers (Fig. 4F). The change in the number of evoked action potentials was < 0.5 spike and thus did not influence measures of evoked PSC strength. The increases in latency were small (< 5 ms) and APs still fell within our time window of counting synaptic EPSCs. Thus, these changes did not interfere with the spatial resolution of our LSPS technique.

We first investigated the laminar distribution of AMPAR-mediated connections for each cell and whether the removal of functional NMDARs would alter the distribution. To generate a laminar input profile for each cell we summed the AMPAR-mediated connections along the dorsal to lateral axis. We then generated a laminar input distribution plot for cells in each group (Fig. 5A). The laminar distribution of inputs revealed that individual cells in each group had diverse laminar distribution of the AMPAR-mediated inputs. AMPAR-mediated connections in all cell groups originate from both upper layers and lower layers of the cortex with no obvious differences among them. These results suggest that lack of functional NMDARs did not affect the gross laminar profile of AMPAR-mediated connections onto L5 excitatory neurons.

We next investigated if NMDAR removal altered the 2-dimensional pattern of connections. The average connection maps for cells in each group (Fig. 5B) were calculated as described in Fig. 1. Qualitative inspection of these average connection maps did not reveal gross differences between Grin1 KO neuron and both groups of control neurons–suggesting that lack of functional NMDARs did not affect the gross 2-dimensional topology of AMPAR-mediated connections onto L5 excitatory neurons.

We next quantified these qualitative observations. To compare the amount of AMPAR-mediated input among these groups, we measured the total number of stimulus locations that provided input to cells in each group (Fig. 5C left). This analysis showed that Grin1 KO and control cells received AMPAR-mediated connections from a similarly large area of cortex. We next quantified the laminar distribution of AMPAR-mediated connections. For each neuron we binned the cortical extent into 5 equal bins spanning from the pia to the bottom of L6. We then calculated the number of stimulus locations that provided input for each bin. This analysis revealed no significant difference in the laminar distribution of AMPAR-mediated connections among the three cell groups (Fig. 5C right), consistent with the qualitative observations. Together, these results suggest that the development of AMPAR-mediated inter-laminar connections does not require functional NMDARs.

Although the laminar analysis showed no significant differences, our laminar analysis emphasized comparison of the central tendency of the data but ignored the variability in the spatial connections at the individual cell level (Fig. 5A), which might be altered by knocking out NMDARs. To quantify the variability of the connection pattern between cells in the same group, we calculated the pairwise correlation of the AMPAR-mediated connection maps within each group (Fig. 5D). High correlation value indicates low variability in the connection patterns. There was a small but significant decrease of the correlation in Grin1 KO neurons (Fig. 5E), suggesting that the connection patterns were more diverse after knocking out NMDARs. Further comparing the pairwise correlation of cells from the same slice (intra-slice) to that of cells from different slices (inter-slice), we did see consistent change of pairwise correlation for cells from the same slice (Fig. S1). Meanwhile, the three groups of data had similar average number of recorded cells per slice. Thus, the change in map correlations is not due to higher correlations for intra-slice comparison.

Our analysis so far did not take into account the strength of the synaptic connections. To explore the possibility that NMDARs are involved in controlling the strength of AMPAR-mediated connections, we quantified the average peak amplitude and transferred charge of AMPAR-mediated EPSCs for each cell (Fig. 6A,B). This analysis revealed no significant changes among the three groups. Separately averaging peak amplitude and transferred charge for connections from different laminar locations (Fig. 6C) also showed similar distributions for all three groups. Thus, these analyses indicate that the strength of AMPAR-mediated connections is not changed in Grin1 KO neurons. Together, the results suggest that the development of AMPAR-mediated connections onto L5 excitatory neurons in the TeA is largely intact in neurons without NMDARs, with a small increase in the variability of the connection patterns.

Decreased GABA_AR-mediated connections in the absence of NMDARs

Removal of NMDARs in hippocampal and cortical neurons can alter the development of GABA_AR-mediated synaptic inputs^38,39,40, and NMDARs can also be found at GABAergic synapses^51,52. We therefore investigated whether the development of translaminar connections mediated by GABA_ARs requires NMDARs. We used LSPS glutamate uncaging to stimulate cortical neurons and isolated GABA_AR-mediated IPSCs by holding L5 excitatory neurons at 0 mV (Fig. 1B). We first analyzed the laminar distribution of GABA_AR-mediated connections for cells in each group by generating a laminar connection distribution plot (Fig. 7A). Similar to previous reports^48,49,53, the laminar connections plot revealed that L5 pyramidal neurons mostly received GABA_AR mediated connections from laminar locations close to the soma of the recorded L5 neurons, although connections can also come from distant locations. To reveal the 2-dimensional pattern of the locations of presynaptic cells, we generated the average input maps for GABA_AR-mediated connections (Fig. 7B). The average input maps from Grin1 KO and Grin1^flox control neurons were similar. However, both Grin1 KO and Grin1^flox control groups appeared to have fewer connections when compared to the WT control group. Quantification of the total number of stimulus locations that gave rise to GABA_AR-mediated connections (Fig. 7C) revealed that Grin1 KO and Grin1^flox control neurons had fewer number of effective stimulus locations than WT control neurons, which confirmed our qualitative observations. Separately analyzing the number of GABA_AR-mediated connections from different laminar locations revealed that the number of connected locations from the lower part of the cortex was significantly decreased for Grin1 KO and Grin1^flox control neurons (Fig. 7G left). Together these results indicate that the lack of functional NMDARs in a subpopulation of excitatory neurons resulted in decreased regional GABA_AR-mediated connections onto L5 excitatory neurons in the TeA.

To investigate if the variability of the GABA_AR-mediated connections pattern changes after knocking out NMDARs, we computed the correlation of the connection maps. There was no significant difference among the three groups (Fig. 7D), suggesting that the diversity of the connection pattern did not change after losing NMDARs.

We also quantified the connection strength of GABA_AR-mediated connections by calculating the average peak amplitude and transferred charge of GABA_AR-mediated IPSCs for each cell. Overall, we found no differences in peak amplitude between groups (Fig. 7E), but Grin1 KO neurons showed lower transferred charge compared to WT control neurons (Fig. 7F). Separately comparing peak amplitude and transferred charge of GABA_AR-mediated connections from different laminar locations revealed that Grin1 KO neurons had reduced peak amplitude and transferred charge for connections coming from the middle and lower part of the cortex (Fig. 7G middle and G right). Thus, this analysis suggests that the strength of inhibitory connections to Grin1 KO neurons is decreased. Since our cell-attached recordings did not show any differences between groups, the decrease of GABA_AR-mediated connections observed in Grin1 KO and Grin1^flox control groups were due to the change in the connected locations. Together, our results suggest that the development of regional inhibition onto L5 excitatory neurons in the TeA requires intact NMDAR-mediated signaling.

Altered excitation/inhibition balance after mosaic NMDARs knockout

In adults, neurons in the central nervous system normally receive balanced excitatory and inhibitory inputs. Although normal AMPAR-mediated connections were observed in Grin1 KO brain slices, the decrease of GABA_AR-mediated connections from lower layers indicate a shift in the balance between excitation and inhibition from specific laminar locations. To test our hypothesis, an excitation/inhibition balance index was calculated from the number of effective stimulus locations for AMPAR- and GABA_AR-mediated connections in the recorded neurons (Fig. 8). The overall excitation and inhibition index between Grin1 KO neuron and Grin1^flox control neurons was similar (Fig. 8A), but both Grin1 KO neuron and Grin1^flox control neurons showed an increase in excitation/inhibition index from lower layers of the cortex compared to WT control neurons (Fig. 8B). These results suggest that our mosaic NMDAR deletion caused an imbalance of excitation and inhibition arising from lower cortical areas in the TeA.

During LSPS glutamate uncaging, the glutamate receptors on the recorded neuron could be directly activated when the stimulation location was too close to the recorded neuron. Locations with AMPAR-mediated direct activations were not included in the analysis. Therefore, the most proximal AMPAR-mediated connections were masked by direct activations. Meanwhile, GABA_AR-mediated connections were not affected by direct activation. This uneven sampling of the origins of AMPAR- and GABA_AR- mediated connections might affect our results. To test this, we first separated GABA_AR-mediated connections into local and distal connections based on the direct activation area of AMPAR-mediated connections. GABA_AR-mediated connections within the direct activation area of AMPAR-mediated connections were labeled as local connections. For the local GABA_AR-mediated connections, the number of effective stimulus locations and average IPSC peak amplitude were unaffected by knocking out NMDARs (Fig. 9A left and middle). There was a small decrease of average IPSC transferred charge when comparing Grin1 KO and WT control neurons (Fig. 9A right). However, for distal connections, Grin1 KO and Grin1^flox control neurons had fewer number of effective stimulus locations and less average IPSC transferred charge than WT control neurons (Fig. 9B left and 9B right). There was a small decrease of average IPSC peak amplitude (Fig. 9B middle). The data suggest that GABA_AR-mediated connections from distal locations were preferentially affected by knocking out NMDA receptors. We further analyzed GABA_AR-mediated distal connections from different laminar locations. Separately comparing the number of effective stimulus locations, IPSC peak amplitude, and IPSC transferred charge of GABA_AR-mediated distal connections from different laminar locations revealed that Grin1 KO and Grin1^flox control neurons had fewer number of effective stimulus locations, Grin1 KO neurons had reduced peak amplitude and transferred charge in the middle and lower part of the cortex (Fig. 9D), which was similar to our results in Fig. 6. The data suggest that GABA_AR-mediated connections from lower layers of the cortex were affected after losing NMDARs. Calculating excitation/inhibition index using GABA_AR-mediated distal connections revealed altered excitation/inhibition balance for connections from lower layers of the cortex in Grin1 KO neurons (Fig. 9C and E), suggesting a change in the excitation/inhibition balance from lower layers of the cortex. Thus, our observation of reduced GABA_AR-mediated connections was not a result of including areas where the glutamate receptors on the recorded cell can be directly activated.

mIPSCS but not mEPSCs are altered in the absence of NMDARs leading to altered excitation/inhibition balance of spontaneous synaptic transmission

The synaptic connections revealed by LSPS are action-potential-evoked synaptic transmissions^25,41, and the glutamatergic connections from proximal neurons are masked by direct activation of the recorded cell by LSPS. Moreover, recent studies suggest that action-potential-evoked synaptic transmission and spontaneous synaptic transmission are generated through different mechanisms and can be regulated separately^54,55,56. To investigate if our results held for all inputs and if spontaneous synaptic transmission to L5 excitatory neurons is affected by lack of NMDARs, we performed voltage clamp recordings to measure the miniature excitatory postsynaptic currents (mEPSCs) and miniature inhibitory postsynaptic currents (mIPSCs) from Grin1 KO and control neurons.

AMPAR-mediated mEPSCs were recorded at a holding potential of −70 mV. We quantified the frequency, peak amplitude and decay time constant of the recorded mEPSCs (Fig. 10A–C). Grin1 KO neurons were similar to controls, suggesting that NMDARs are not required for the early development of spontaneous excitatory synaptic transmission to L5 excitatory neurons in TeA.

GABA_AR-mediated mIPSCs were recorded at 0 mV holding potential. Similar to other studies^38,39,40, we found that loss of NMDARs led to a significant decrease of mIPSC frequency (Fig. 10D). Electroporated WT neurons had significantly higher mIPSC peak amplitude than Grin1 KO neurons (Fig. 10E). The decay time constant did not change between the Grin1 KO and control neurons suggesting that the underlying subunit composition was similar (Fig. 10F). This suggests that the early development of spontaneous inhibitory synaptic transmission requires NMDARs.

The frequency and peak amplitude of mEPSCs and mIPSCs are related to the number and strength of the respective inputs. Thus, to quantify the excitation/inhibition balance of the spontaneous synaptic transmission, we used the frequency or the peak amplitude of mEPSC and mIPSC to calculate the excitation/inhibition index. The frequency-based excitation/inhibition index increased significantly in the Grin1 KO neurons (Fig. 10G), but there was only a small increase in the peak amplitude-based excitation/inhibition index when comparing Grin1 KO and WT control neurons (Fig. 10H). Together, our results suggest normal excitation/inhibition balance requires NMDARs.

Discussion

We tested the requirement of NMDARs for the development of AMPAR- and GABA_AR-mediated synaptic connections to L5 excitatory neurons in the temporal association cortex (TeA). Our results show that deletion of functional NMDARs from immature cortical excitatory neurons made the spatial pattern of AMPAR-mediated microcircuits more diverse without changing the amount of the connections. In contrast, NMDAR deletion in excitatory neurons caused a decrease in inhibitory connections. Action-potential-evoked and spontaneous inhibitory synaptic transmissions were differentially affected by NMDAR removal. For action-potential-evoked inhibitory synaptic transmission, both Grin1 KO and Grin1^flox control neurons had a decrease in regional inhibitory connections, leading to altered regional excitation and inhibition balance. For spontaneous inhibitory synaptic transmission, only Grin1 KO neurons had a decrease in inhibitory connections. Brain slices containing Grin1^flox control, Grin1 KO, and WT control neurons were all subject to electroporation. Additionally, WT control neurons were transfected with the same Cre plasmid as Grin1 KO neurons. Therefore, we can rule out that either electroporation or overexpression of Cre and GFP caused those changes. Thus, circuit maturation in the TeA can occur in the absence of NMDA receptors, but normal development of inhibition requires intact NMDAR signaling in these excitatory neurons.

Our results show that, as in the hippocampus, visual cortex, somatosensory cortex and auditory cortex, AMPAR-mediated connections onto L5 excitatory TeA neurons increase extensively during the first two postnatal weeks and that immature L5 excitatory neurons receive NMDAR-only connections^{19,20,21,23,24,25,57,58,59,60}. Together, the evidence suggests that NMDAR-only connections and the development of AMPAR-mediated connections are general features of the early development of glutamatergic synaptic transmissions³¹.

Because activation of NMDARs in mature neurons can promote the insertion of AMPARs into the postsynaptic sites during long-term potentiation, it has been hypothesized that neuronal activity in immature neurons activates NMDAR-only synapses and facilitates the development of AMPAR-containing synapses^59,61,62,63. Indeed, NMDAR-only connections are found to precede mature AMPAR-positive connections during development^23,25, and NMDAR-only connections can be converted into AMPAR-positive connections^23,64. However, several studies using genetic and pharmacological approaches to eliminate functional NMDARs have shown that functional AMPAR-containing synapses still develop in the absence of NMDARs^{13,65,66,67,68,69,70,71}. Furthermore, activation of NMDARs could suppress EPSCs mediated by AMPARs in young neurons⁶⁶, suggesting that NMDARs play a different role in regulating the developmental upregulation of AMPARs in immature neurons. Here, we investigated if NMDAR played a role in the early development of specific circuits. Our results show that neurons lacking NMDARs still have AMPAR-mediated EPSCs, consistent with prior observations^13,65,66,72. We did not observe significant increases in the amplitude of AMPAR-mediated connections in Grin1 KO neurons, possibly due to our probing of a variety of presynaptic sources or to the difference in the brain area under investigation⁷³. Moreover, we found that removal of functional NMDARs did not alter the development and laminar topology of functional AMPAR-mediated circuits onto L5 excitatory neurons. However, our results show that the connection pattern is slightly more variable in Grin1 KO neurons. Thus, together with prior studies, our results suggest that the development of AMPA-mediated circuits to L5 neurons is largely independent of postsynaptic NMDARs in these neurons.

We here use LSPS to probe neuronal connection. LSPS mapping methods cannot resolve the glutamatergic connections from nearby cells due to direct activation of dendrites of the recorded neurons. Thus, these local connections could be affected by the lack of NMDARs. However, our mEPSC recordings, which reflect all connections, were unaltered in excitatory neurons with NMDARs removed, suggesting that local connections are unaffected by lack of NMDARs. Additionally, we observed no difference in the decay time constant of the mEPSCs and mIPSCs. Since different subunit compositions underlie different kinetics of AMPA and GABA_A receptors^{74,75,76,77,78}, disrupting NMDARs did not affect the subunit composition of AMPA and GABA_A receptors.

Similar to previous reports^{38,39,40,79,80}, we observed a reduction of GABA_AR-mediated synaptic transmission and a consequent change of E/I balance after disrupting NMDARs. As proper E/I balance is thought to maintain normal network activity and cortical functions^80,81,82, it remains to be seen if the cortical network activity is changed in our Grin1 knockout model. Furthermore, the effect of disrupting NMDARs on GABA_AR-mediated synaptic transmission is not identical for action-potential-evoked and spontaneous synaptic transmissions. A decrease in action-potential-evoked GABA_AR-mediated transmission was observed in both Grin1 KO and Grin1^flox control neurons but a decrease in spontaneous GABA_AR-mediated transmission was only found in Grin1 KO neurons. Our results confirmed several prior observations of impaired GABA_AR-mediated activity when comparing Grin1 KO to Grin1^flox control neurons in the same brain slice^38,39,40,79, in particular decreased spontaneous mIPSC frequency. However, unlike the impairment of the spontaneous mIPSC, our LSPS experiments here show that action-potential-evoked GABA_AR-mediated IPSCs are similar between Grin1 KO and Grin1^flox control neurons. One possible explanation for this differential effect is that the spontaneous and evoked synaptic transmissions have different synaptic properties^54,55. The synaptic sites giving rise to spontaneous mIPSCs and evoked IPSCs are partially segregated at the postsynaptic sites in the hippocampus⁸³. Moreover, retinoic acid’s effect on the inhibitory synaptic transmissions in the somatosensory cortex is different for the spontaneous mIPSCs and evoked IPSCs⁵⁶, indicating that different groups of synapses might underlie both events. Thus, our manipulation of NMDARs might only affect the part of inhibitory synaptic transmission that is segregated from the evoked synaptic transmissions. An alternative interpretation is that LSPS could fail to reveal changes in connections from very proximal neurons, because glutamate release on the soma of recorded neuron could activate large amount of glutamate receptors and affect the accuracy of the recording by having a large change in the input resistance⁸⁴ and thus mask synaptic events. If this is the case, there could be an increase of proximal GABA_AR-mediated connections in Grin1^flox control neurons so that when assessed by miniature recordings, the total number of GABA_AR-mediated connections did not differ from the WT control neurons. For the action-potential-evoked inhibitory synaptic transmission, despite the reduced GABA_AR-mediated connections from lower layers onto Grin1 KO and Grin1^flox control neurons compared to WT control, GABA_AR-mediated connections from different cortical laminae still develop. Thus, the effect of NMDAR signaling on the development of translaminar inhibitory circuits is mostly restricted to connections from lower layers, which might be mediated by specific classes of interneurons.

Both Grin1^flox and WT pups were subject to the same electroporation procedures. Moreover, the Grin1^flox and WT mice we used share similar C57BL/6 genetic background. Thus, the difference we observed between the electroporated Grin1^flox and WT pups were not due to the surgical procedures or the genetic background of the mice. How would deletion of NMDARs in a subset of excitatory neurons affects the development of inhibition in the network? A large fraction of activity in developing cortical networks is mediated by NMDARs^85,86. Also, NMDAR removal reduces the correlation of spontaneous activity in L4 neurons of the somatosensory cortex⁸⁷. Since neural activity is crucial for the normal development of inhibitory synaptic transmission^88,89,90,91, we speculate that lack of NMDARs in a subset of excitatory neurons would alter network activity and the developmental trajectory of inhibitory connections. Alternatively, reduced network activity could lead to down regulation of inhibition through homeostatic plasticity^92,93.

A caveat of all manipulations of gene expression using Cre is that the effect depends on Cre activity. We here prevent NMDA receptor expression by utilizing in utero electroporation of Cre. While we did not detect any NMDAR mediated currents at P6 it is possible that some NMDA signaling might have been present at earlier developmental, e.g. embryonic, ages and that potentially this very early expression drove AMPAR upregulation. However, since most AMPAR upregulation occurs after P6 our results suggest that such early potential influence might be minor.

Prior studies showed that deleting Grin1 in all cortical excitatory neurons results in altered projections of whisker afferents and altered dendritic branching in mouse somatosensory thalamus and cortex^11,12,13, suggesting a role for NMDARs in the shaping these anatomical circuits. However, it is unclear from these studies where in the circuit NMDARs are required. Our manipulation removed NMDARs in a relatively small subset of excitatory neurons. When the acutely prepared brain slices were inspected under the fluorescence microscope, roughly 25–50 electroporated cells (GFP +) could be found in one square millimeter area. However, due to the thickness of the slice, the observation in the acutely prepared brain slices could be an underestimate. In few electroporated animals when the brains were processed with immunostaining, the GFP + neurons could be as high as 290 cells per square millimeter (around 1/10 of the total cells in one square millimeter area, without compensating for tissue shrinkage during fixation). Therefore, NMDARs were present in the majority of excitatory and all inhibitory neurons, potentially on their axon terminals^51,52,94,95. Given that we did not find significant changes in the amount of AMPAR-mediated connections between Grin1 KO and Grin1^flox control neurons, we speculate that presynaptic NMDAR-mediated signaling might play a role in the refinement of neural circuits, or that deletion of NMDARs in all neurons disrupts global network activity patterns leading to morphological changes. One feature of developing neuronal circuits is the presence of spontaneous and emergent sensory activity evident as bursts, many of which are slow and which can synchronize local neuronal populations^{96,97,98,99,100,101,102,103,104,105}. As the developing neurons are slightly depolarized, NMDARs with their relatively longer time constants might enhance participation of neurons in these activity bursts. In contrast, AMPAR-mediated connections integrate information on faster time scales. Thus, the presence of NMDAR-only synapses at early ages might be a developmental specialization to propagate slow activity patterns. It will be informative to monitor the cortical network activity after NMDAR deletion and test if the subsequent changes of synaptic connections are correlated with changes in the network activity. Since complete removal of NMDARs impairs development^11,12,13, the effect of removing NMDARs on the development of neural connections might be correlated with the number of cells lacking NMDARs. Since the number of electroporated cells is variable in our experiments, we separated the data based on the number of GFP + neurons observed from the acutely prepared brain slices. We did not find any correlations between the amount of electroporated neurons and the AMPAR- and GABA_AR-mediated connections from the LSPS experiments. Alternatively, the LSPS experiments might not be spatially sensitive enough to reveal potential subtle differences in circuit changes to the range of NMDAR deletion in our experiments.

To maintain consistency in the manipulated neurons, we performed electroporation in a small time-window (between E12.5 and E13.5). The electroporated neurons were distributed across different layers in the recorded region (TeA). This is likely because at this developmental stage the Tbr2 positive intermediate progenitor cells (IPCs) can give rise to both lower and upper cortical excitatory neurons^106,107,108. We did not observe a gross difference in the locations of the electroporated neurons between the electroporated Grin1^flox and WT brains slices. We tried to target the neurons in the middle of the cortex (corresponding to the upper layer 5 neurons in the primary sensory areas) for our recordings, as it is more consistent to locate electroporated neurons in this area in our experiments. Combining the electroporation time-window and the morphology of the recorded neurons, our manipulation is likely restricted in the excitatory neurons. But given the diversity of the cortical excitatory neurons^109,110,111, it is possible that the effect we observed is specific to the neurons we targeted. It will be informative to use other methods, such as layer-specific Cre lines or performing the Grin1 deletion at different developmental stages, to manipulate different populations of the cortical excitatory neurons and test if the results are different.

In summary, our results show that in excitatory cortical L5 neurons postsynaptic NMDA receptors are not required for the early assembly and refinement of AMPA-mediated microcircuits, but are required for the development of inhibitory synaptic transmission.

Materials and methods

All procedures were approved by the University of Maryland Institutional Animal Care and Use Committee to be in accordance with applicable guidelines and regulations. The study was carried out in compliance with the ARRIVE guidelines.

Animals

Grin1^flox mice (strain #005246, Jackson Laboratories) and C57BL/6J (wild type or WT) mice (strain #000664, Jackson Laboratories) were used to generate timed-pregnant mice for in utero electroporation. Mouse pups of both sexes from postnatal day 6 (P6) to P14 were used for in vitro electrophysiology experiments, including 19 electroporated Grin1^flox mice, 11 electroporated wild type mice, and 8 wild type mice.

In utero electroporation

Timed-pregnant mice were generated by pairing male and female mice for two days. Surgeries were performed on embryonic day 12.5 (E12.5) or E13.5 as described previously^112,113. We used electroporation of Cdk5r-Cre-IRES-GFP plasmid (courtesy from Dr. Paola Arlotta, Harvard University) to express Cre and GFP in postmitotic pyramidal neurons⁵⁰. For each embryo, 0.3–0.45 µl plasmid DNA solution (DNA concentration 1 ~ 1.5 µg/µl) was injected into the left ventricle. Five square voltage pulses (35–37 V, 50 ms duration, 1 Hz) were delivered through a needle electrode (Nepa Gene, CUY610P4-1). GFP signals in cortical neurons could be detected as early as 3 days after electroporation (data not shown). At P12-14, GFP + neurons were scattered across cortical layers, likely because intermediate neuronal progenitors can give rise to both upper layer and lower layer cortical neurons^106,107. The density of GFP + neurons in the recorded L5 area was around 25–50 cells/mm², when acutely prepared slices were imaged under the epifluorescence microscope. As a control we electroporated C57BL/6 mice. We did not see a gross difference in the distribution pattern of GFP + neurons between sections from Grin1^flox and WT mice. A small percentage of brains displayed abnormally enlarged lateral ventricles which were not correlated with the genotype of the mouse or the side of electroporation; these brains were excluded from experiments.

Immunohistochemistry and imaging

Electroporated animals at P14-15 were transcardially perfused with cold phosphate buffered saline (PBS) and 4% paraformaldehyde (Electron Microscopy Sciences) under deep anesthesia (Fluriso, VetOne). Perfused brains were fixed in 4% paraformaldehyde for 24 h before transferring to 30% phosphate-buffered sucrose solution for cryoprotection. Cryoprotected brains were sectioned with a freezing microtome (Leica) in the coronal plane at a thickness of 50 μm and stored in PBS until use. To identify GFP-immunopositive cells, immunohistochemistry was performed using standard protocols. Briefly, a blocking solution was prepared with 5% normal goat serum (Cat no. 5560-0007, SeraCare) and 0.3% Triton-X 100 (Sigma). Selected brain sections were incubated in the blocking solution at room temperature for 90 min before being incubated overnight in chicken-anti-GFP (1:1000, ab13970, Abcam) at 4 °C. The sections were washed three times before incubating in secondary antibodies (goat anti-chicken, labeled with Alexa Fluor 488, 1:500, A11039, Invitrogen). Sections were then mounted with antifade mounting medium containing DAPI (H-1200, Vectashield).

The sections were imaged with a confocal microscope (10 × lens, SP5 X, Leica); z-stacks were taken at 2048 × 2048 resolution in 3.0 μm steps, with the tile function. Images were then processed with Fiji software. Fluorescence intensities were modified for visualization purposes.

Electrophysiology

Acute brain slices were made as described previously⁴¹. In brief, after euthanasia, mouse brains were sectioned with a vibratome (VT1200, Leica) in ice cold sucrose artificial cerebrospinal fluid (ACSF) containing (in mM): 212.7 sucrose, 2.6 KCl, 1.23 NaH₂PO₄, 26 NaHCO₃, 10 glucose, 3 MgCl₂, 1 CaCl₂ (pH 7.35–7.4). 400 µm thick coronal slices were incubated at 30 ºC for 30 min then kept at room temperature in ACSF containing (in mM): 130 NaCl, 3 KCl, 1.25 NaHCO₃, 10 glucose, 1.3 MgSO₄ and 2.5 CaCl₂ (pH 7.35–7.4). ACSF was equilibrated with 95% O₂-5% CO₂. During LSPS experiments, the recording chamber was superfused with high-divalent ACSF to reduce polysynaptic transmission. High-divalent ACSF contains (in mM): 124 NaCl, 5 KCl, 1.23 NaH₂PO₄, 26 NaHCO₃, 10 glucose, 4 MgCl₂ and 4 CaCl₂. Electroporated cortical neurons expressing GFP were identified under the epifluorescence microscope. We recorded neurons in the middle layer (putative layer 5) of the temporal association area (TeA) because it is the most consistently electroporated area in our surgery. 1–2 adjacent slices containing the TeA were taken from each brain. Whole-cell voltage clamp recordings were performed at room temperature (21–24 °C) with 3–7 MΩ borosilicate recording electrodes filled with internal solution containing (in mM): 115 cesium methanesulfonate, 5 NaF, 10 EGTA, 9 CsCl, 3.5 MgATP, 0.3 NaGTP and 3 QX-314 (pH 7.25; 300 mOsm). Internal solution also contained 0.5% biocytin. We normally record from 2–3 cells per brain slice. Data were acquired with a voltage-clamp amplifier (Multiclamp 700B; Molecular Devices) and digitized using a DAQ board (NI PCI-6259, National Instruments) using EPHUS¹¹⁴ in MATLAB (Mathworks). Membrane potential was corrected for 10 mV of estimated liquid junction potential. Series resistance (R_s) was typically 20–40 MΩ and was not compensated during recording. To identify ‘silent’ synapses, minimal electrical stimulations were delivered through an extracellular bipolar electrode at the rate of 0.1 Hz¹⁹. AMPAR-mediated currents were recorded at −70 mV. AMPAR- plus NMDAR-mediated currents were recorded at + 40 mV. GABA_AR-mediated currents were blocked by picrotoxin (PTX, 100 µM) in the bath solution. 60 to 120 stimulation trials at each holding potential were used to calculate response failure rate. The spontaneous miniature EPSCs and IPSCs were recorded at the holding potential of -70 mV and 0 mV, respectively, in the presence of TTX (1.5 µM) at room temperature (21–24 °C). R_s was monitored during recording and cells were excluded from analysis when the change of R_s was greater than 20%. Loose seal cell attached (R_seal between 20 and 50 MΩ) recordings were used to record action potentials without washing out the intracellular content. Recording electrodes were filled with filtered ACSF during loosely cell-attached recordings.

Laser-scanning photostimulation (LSPS)

LSPS was performed as described previously^25,41,49,115. Caged glutamate [0.8 mM N-(6-nitro-7-coumarylmethyl)-L-glutamate]¹¹⁶ was added in the bath solution. Photostimulation (355 nm, 1-ms pulses; 3510–100, DPSS Lasers Inc.) was delivered through a 10 × water immersion objective (Olympus UMPLFLN10XW, numerical aperture 0.3). To map the locations of presynaptic cells, laser power on the specimen was ~ 24 mW. A rectangular array of up to 20 × 30 sites with 40 μm spacing covering the whole cortical length was stimulated once at 1 Hz in a pseudorandom order during whole-cell voltage clamp recording. This stimulation paradigm evokes action potentials in neurons at the stimulation sites with similar spatial resolution (~ 150 μm) during development⁴⁴. For identifying NMDAR-only connections, AMPAR- and NMDAR- mediated currents were recorded at holding potentials of -70 mV and + 40 mV in the presence of picrotoxin. Putative monosynaptic excitatory postsynaptic currents (EPSCs) were differentiated from direct activation of the glutamate receptors on the recorded cell based on the EPSCs’ latency. EPSCs in the 10 ms to 50 ms poststimulation time-window were classified as putative monosynaptic EPSCs (Fig. 1B). A minimal peak amplitude of 10 pA was used to reduce false positive PSCs caused by spontaneous PSCs or the noise during recording. We used the last 100 ms of the 1 s recording trace to estimate the false positive rate of our PSC detection criteria. The false positive rates were low and were comparable among the three groups (AMPAR-mediated: Grin1^flox control, median 0.0025 Hz, IQR 0.007; Grin1 KO, median 0.0055 Hz, IQR 0.0129; WT control, median 0.004 Hz, IQR 0.0086. GABA_AR- mediated: Grin1^flox control, median 0.0089 Hz, IQR 0.019; Grin1 KO, median 0.0029 Hz, IQR 0.007; WT control, median 0.0065 Hz, IQR 0.016. All P > 0.05. Kruskal–Wallis test was used and followed by Tukey's honest significant difference criterion for multi-group comparison). The classified direct activation and monosynaptic EPSCs have been verified by repeating the experiments in the presence of TTX in previous reports. Most of the classified direct responses (~ 85%) were TTX-resistant and most of the classified monosynaptic EPSCs (~ 90%) were blocked by TTX^25,41. Because the responses were highly repeatable (> 90%)²⁵, in most mapping experiments we only have one trial per recording condition per cell. For studying the AMPAR- and GABA_AR- mediated connections, EPSCs and inhibitory postsynaptic currents (IPSCs) were measured at −70 mV and 0 mV, respectively. For measuring the + 40 mV/-70 mV current ratio of EPSC responses to glutamate uncaging (Fig. 3), we used low laser power (5 – 10 mW) glutamate uncaging around the soma of the recorded cell. In these experiments GABA_AR-mediated currents were blocked by picrotoxin (100 µM) and action potentials in the brain slice were blocked by TTX (1 µM, Tocris). D-2-Amino-5-phosphonovalerate (D-APV, 50 µM, Tocris) was used to block NMDARs; 2,3-dioxo-6-nitro-1,2,3,4-tetrahydrobenzo[f]quinoxaline-7-sulfonamide (NBQX, 10 µM, Tocris) was used to block AMPARs. To verify that our stimulus resolution did not change between the electroporated Grin1^flox slices and the electroporated WT slices, loose-patch recordings were used to record action potentials from neurons across all layers of the cortex during glutamate uncaging. All drugs and chemicals were purchased from Sigma-Aldrich unless specified otherwise.

Data analysis and statistics

Data were analyzed using custom scripts written in MATLAB as previously described^25,41,49. Putative monosynaptic EPSCs and IPSCs were identified through the criteria above. AMPAR- and GABA_AR- mediated connection maps were constructed based on the photostimulation locations and the recorded synaptic currents. NMDAR-only connection maps were constructed by subtracting connected locations in AMPAR-mediated connection maps from connection maps recorded at + 40 mV. To reveal the general patterns of the connected locations, maps from different cells were aligned to the pia and ventricle locations to calculate the fraction of cells that received a connection from each stimulation location. Subsequently, data from AMPAR- and GABAAR-connection maps were further used for calculating laminar distribution plots, map correlations, and E/I balance. To give an overview of the laminar distribution of AMPAR- and GABA_AR-mediated inputs, we summed the number of input locations in individual input map along the direction parallel to the pia and plotted the laminar distribution of input locations for each cell. Traditional LSPS analysis divides the cortex into layers based on histological difference visible under IR illumination and analyzes inputs separately from each layer^{25,41,42,44,49,117,118}. Because TeA does not show clear layer boundaries under IR illumination and it is disputable if TeA has layer 4, we devised an alternative way to quantify the laminar distribution of the connections. We evenly divided the cortex into 5 areas (bins) along the pia-to-ventricle axis and quantified the number of presynaptic locations, average peak amplitude, average transferred charge of the inputs, and excitation-to-inhibition index in each area. Each area covered roughly 200 µm distance along the pia to ventricle axis. Roughly speaking, the bins 1–2 correspond to the layer 1 to 4 and bins 3–5 correspond to the layer 5 and 6 in the primary sensory area. We also repeated all of our analyses when aligning the cells to their soma location and got similar results, indicating that our results were not affected by how we align the individual connection maps from different cells. To quantify the balance of excitation and inhibition from the glutamate uncaging data, we used the number of locations with AMPAR- (nAMPAR) and GABA_AR-mediated (nGABA_AR) connections to calculate the excitation-to-inhibition (E/I) index (EII) as follows:

$$EII = \frac{{nAMPAR - nGABA_{A} R}}{{nAMPAR + nGABA_{A} R}}$$

Similarly, excitation-to-inhibition balance of spontaneous miniature recordings was calculated, where the frequency or the peak amplitude of miniature excitatory postsynaptic currents (mEPSCs) and miniature inhibitory postsynaptic currents (mIPSCs) were used to quantify excitation and inhibition, respectively. To quantify the laminar distribution of NMDAR-only/AMPAR-mediated connections ratio (N/A index) at P6-7, we used similar approach to calculate NMDAR-only/AMPAR-mediated connections index.

Miniature EPSC and IPSC events were identified in a semiautomatic fashion. At least 4 min of stable recordings at the same holding potential were used for each cell. Locations of the PSC peaks were identified from low-pass filtered data. PSC events were extracted and compared to the scaled average PSC. PSC events with abnormally different rise and decay kinetics were excluded. Detected PSC events were plotted on the original recording traces and manually inspected.

Median and interquartile range (IQR) were reported as the descriptive analysis. For two-group comparison, a two-sample t-test or a Wilcoxon rank-sum test was used to test the significance of the difference. For three-group comparison, Kruskal–Wallis test or one-way analysis of variance was used, then followed by Tukey's honest significant difference criterion for multi-group comparison. For paired samples, Wilcoxon signed rank test was used. Significance level was marked as P < 0.05 (*) or P < 0.01 (**); otherwise P > 0.05. Test methods, significant P values and the number of cells are detailed in the figure legends. All tests were performed in MATLAB (Mathworks). Effect size r¹¹⁹ is reported for all pairwise statistical tests (Table S1-7). r > 0.5 indicates a large effect; 0.3 < r < 0.5 indicates a medium effect; r < 0.3 indicates a small effect.

References

Douglas, R. J. & Martin, K. A. Neuronal circuits of the neocortex. Annu. Rev. Neurosci. 27, 419–451. https://doi.org/10.1146/annurev.neuro.27.070203.144152 (2004).
Article CAS Google Scholar
Harris, K. D. & Shepherd, G. M. The neocortical circuit: themes and variations. Nat. Neurosci. 18, 170–181. https://doi.org/10.1038/nn.3917 (2015).
Article CAS Google Scholar
Jung, J., Cloutman, L. L., Binney, R. J. & Lambon Ralph, M. A. The structural connectivity of higher order association cortices reflects human functional brain networks. Cortex 97, 221–239. https://doi.org/10.1016/j.cortex.2016.08.011 (2017).
Article Google Scholar
Goldman-Rakic, P. S. Topography of cognition: Parallel distributed networks in primate association cortex. Annu. Rev. Neurosci. 11, 137–156. https://doi.org/10.1146/annurev.ne.11.030188.001033 (1988).
Article CAS Google Scholar
Zingg, B. et al. Neural networks of the mouse neocortex. Cell 156, 1096–1111. https://doi.org/10.1016/j.cell.2014.02.023 (2014).
Article CAS Google Scholar
Nishio, N. et al. Higher visual responses in the temporal cortex of mice. Sci. Rep. 8, 11136. https://doi.org/10.1038/s41598-018-29530-3 (2018).
Article ADS CAS Google Scholar
da Costa, N. M., Martin, K. A. C. & Sägesser, F. D. A weighted graph of the projections to mouse auditory cortex. bioRxiv, https://doi.org/10.1101/228726 (2017).
Tasaka, G. I. et al. The temporal association cortex plays a key role in auditory-driven maternal plasticity. Neuron 107, 566-579 e567. https://doi.org/10.1016/j.neuron.2020.05.004 (2020).
Article CAS Google Scholar
Chomiak, T., Karnik, V., Block, E. & Hu, B. Altering the trajectory of early postnatal cortical development can lead to structural and behavioural features of autism. BMC Neurosci. 11, 102. https://doi.org/10.1186/1471-2202-11-102 (2010).
Article Google Scholar
Courchesne, E. et al. Mapping early brain development in autism. Neuron 56, 399–413. https://doi.org/10.1016/j.neuron.2007.10.016 (2007).
Article CAS Google Scholar
Lee, L. J., Lo, F. S. & Erzurumlu, R. S. NMDA receptor-dependent regulation of axonal and dendritic branching. J. Neurosci. 25, 2304–2311. https://doi.org/10.1523/JNEUROSCI.4902-04.2005 (2005).
Article CAS Google Scholar
Datwani, A., Iwasato, T., Itohara, S. & Erzurumlu, R. S. NMDA receptor-dependent pattern transfer from afferents to postsynaptic cells and dendritic differentiation in the barrel cortex. Mol. Cell. Neurosci. 21, 477–492 (2002).
Article CAS Google Scholar
Iwasato, T. et al. Cortex-restricted disruption of NMDAR1 impairs neuronal patterns in the barrel cortex. Nature 406, 726–731. https://doi.org/10.1038/35021059 (2000).
Article ADS CAS Google Scholar
Crair, M. C. & Malenka, R. C. A critical period for long-term potentiation at thalamocortical synapses. Nature 375, 325–328. https://doi.org/10.1038/375325a0 (1995).
Article ADS CAS Google Scholar
Hsia, A. Y., Malenka, R. C. & Nicoll, R. A. Development of excitatory circuitry in the hippocampus. J. Neurophysiol. 79, 2013–2024. https://doi.org/10.1152/jn.1998.79.4.2013 (1998).
Article CAS Google Scholar
Lu, H. C., Gonzalez, E. & Crair, M. C. Barrel cortex critical period plasticity is independent of changes in NMDA receptor subunit composition. Neuron 32, 619–634. https://doi.org/10.1016/s0896-6273(01)00501-3 (2001).
Article Google Scholar
Ye, G. L., Yi, S., Gamkrelidze, G., Pasternak, J. F. & Trommer, B. L. AMPA and NMDA receptor-mediated currents in developing dentate gyrus granule cells. Brain Res. Dev. Brain Res. 155, 26–32. https://doi.org/10.1016/j.devbrainres.2004.12.002 (2005).
Article CAS Google Scholar
Liao, D., Hessler, N. A. & Malinow, R. Activation of postsynaptically silent synapses during pairing-induced LTP in CA1 region of hippocampal slice. Nature 375, 400–404. https://doi.org/10.1038/375400a0 (1995).
Article ADS CAS Google Scholar
Isaac, J. T., Nicoll, R. A. & Malenka, R. C. Evidence for silent synapses: Implications for the expression of LTP. Neuron 15, 427–434 (1995).
Article CAS Google Scholar
Isaac, J. T., Crair, M. C., Nicoll, R. A. & Malenka, R. C. Silent synapses during development of thalamocortical inputs. Neuron 18, 269–280. https://doi.org/10.1016/s0896-6273(00)80267-6 (1997).
Article CAS Google Scholar
Malenka, R. C. & Nicoll, R. A. Silent synapses speak up. Neuron 19, 473–476 (1997).
Article CAS Google Scholar
Rumpel, S., Hatt, H. & Gottmann, K. Silent synapses in the developing rat visual cortex: Evidence for postsynaptic expression of synaptic plasticity. J. Neurosci. 18, 8863–8874 (1998).
Article CAS Google Scholar
Anastasiades, P. G. & Butt, S. J. A role for silent synapses in the development of the pathway from layer 2/3 to 5 pyramidal cells in the neocortex. J. Neurosci. 32, 13085–13099. https://doi.org/10.1523/jneurosci.1262-12.2012 (2012).
Article CAS Google Scholar
Hanse, E., Seth, H. & Riebe, I. AMPA-silent synapses in brain development and pathology. Nat. Rev. Neurosci. 14, 839–850. https://doi.org/10.1038/nrn3642 (2013).
Article CAS Google Scholar
Meng, X., Kao, J. P. & Kanold, P. O. Differential signaling to subplate neurons by spatially specific silent synapses in developing auditory cortex. J. Neurosci. 34, 8855–8864. https://doi.org/10.1523/JNEUROSCI.0233-14.2014 (2014).
Article CAS Google Scholar
Wollmuth, L. P., Kuner, T. & Sakmann, B. Intracellular Mg2+ interacts with structural determinants of the narrow constriction contributed by the NR1-subunit in the NMDA receptor channel. J. Physiol. 506(Pt 1), 33–52. https://doi.org/10.1111/j.1469-7793.1998.00033.x (1998).
Article CAS Google Scholar
Mayer, M. L., Westbrook, G. L. & Guthrie, P. B. Voltage-dependent block by Mg2+ of NMDA responses in spinal cord neurones. Nature 309, 261–263. https://doi.org/10.1038/309261a0 (1984).
Article ADS CAS Google Scholar
Nowak, L., Bregestovski, P., Ascher, P., Herbet, A. & Prochiantz, A. Magnesium gates glutamate-activated channels in mouse central neurones. Nature 307, 462–465. https://doi.org/10.1038/307462a0 (1984).
Article ADS CAS Google Scholar
Ben-Ari, Y., Khazipov, R., Leinekugel, X., Caillard, O. & Gaiarsa, J. L. GABAA, NMDA and AMPA receptors: A developmentally regulated “menage a trois”. Trends Neurosci. 20, 523–529. https://doi.org/10.1016/s0166-2236(97)01147-8 (1997).
Article CAS Google Scholar
Ben-Ari, Y., Gaiarsa, J. L., Tyzio, R. & Khazipov, R. GABA: A pioneer transmitter that excites immature neurons and generates primitive oscillations. Physiol. Rev. 87, 1215–1284. https://doi.org/10.1152/physrev.00017.2006 (2007).
Article CAS Google Scholar
Kanold, P. O., Deng, R. & Meng, X. The integrative function of silent synapses on subplate neurons in cortical development and dysfunction. Front. Neuroanat. 13, 41. https://doi.org/10.3389/fnana.2019.00041 (2019).
Article CAS Google Scholar
Malenka, R. C. & Bear, M. F. LTP and LTD: An embarrassment of riches. Neuron 44, 5–21. https://doi.org/10.1016/j.neuron.2004.09.012 (2004).
Article CAS Google Scholar
Markram, H. et al. Interneurons of the neocortical inhibitory system. Nat. Rev. Neurosci. 5, 793–807. https://doi.org/10.1038/nrn1519 (2004).
Article CAS Google Scholar
Le Magueresse, C. & Monyer, H. GABAergic interneurons shape the functional maturation of the cortex. Neuron 77, 388–405. https://doi.org/10.1016/j.neuron.2013.01.011S0896-6273(13)00055-X[pii] (2013).
Article Google Scholar
Aamodt, S. M., Shi, J., Colonnese, M. T., Veras, W. & Constantine-Paton, M. Chronic NMDA exposure accelerates development of GABAergic inhibition in the superior colliculus. J. Neurophysiol. 83, 1580–1591. https://doi.org/10.1152/jn.2000.83.3.1580 (2000).
Article CAS Google Scholar
Hartman, K. N., Pal, S. K., Burrone, J. & Murthy, V. N. Activity-dependent regulation of inhibitory synaptic transmission in hippocampal neurons. Nat. Neurosci. 9, 642–649. https://doi.org/10.1038/nn1677 (2006).
Article CAS Google Scholar
Henneberger, C. et al. GluR- and TrkB-mediated maturation of GABA receptor function during the period of eye opening. Eur. J. Neurosci. 21, 431–440. https://doi.org/10.1111/j.1460-9568.2005.03869.x (2005).
Article Google Scholar
Gu, X., Zhou, L. & Lu, W. An NMDA receptor-dependent mechanism underlies inhibitory synapse development. Cell Rep. 14, 471–478. https://doi.org/10.1016/j.celrep.2015.12.061 (2016).
Article CAS Google Scholar
Gu, X. & Lu, W. Genetic deletion of NMDA receptors suppresses GABAergic synaptic transmission in two distinct types of central neurons. Neurosci. Lett. 668, 147–153. https://doi.org/10.1016/j.neulet.2018.01.024 (2018).
Article CAS Google Scholar
Lu, W., Bushong, E. A., Shih, T. P., Ellisman, M. H. & Nicoll, R. A. The cell-autonomous role of excitatory synaptic transmission in the regulation of neuronal structure and function. Neuron 78, 433–439. https://doi.org/10.1016/j.neuron.2013.02.030 (2013).
Article CAS Google Scholar
Deng, R., Kao, J. P. Y. & Kanold, P. O. Distinct translaminar glutamatergic circuits to GABAergic interneurons in the neonatal auditory cortex. Cell Rep. 19, 1141–1150. https://doi.org/10.1016/j.celrep.2017.04.044 (2017).
Article CAS Google Scholar
Viswanathan, S., Bandyopadhyay, S., Kao, J. P. & Kanold, P. O. Changing microcircuits in the subplate of the developing cortex. J. Neurosci. 32, 1589–1601. https://doi.org/10.1523/JNEUROSCI.4748-11.2012 (2012).
Article CAS Google Scholar
Marques-Smith, A. et al. A transient translaminar GABAergic interneuron circuit connects thalamocortical recipient layers in neonatal somatosensory cortex. Neuron 89, 536–549. https://doi.org/10.1016/j.neuron.2016.01.015 (2016).
Article CAS Google Scholar
Meng, X. et al. Transient subgranular hyperconnectivity to L2/3 and enhanced pairwise correlations during the critical period in the mouse auditory cortex. Cereb. Cortex 30, 1914–1930. https://doi.org/10.1093/cercor/bhz213 (2020).
Article Google Scholar
Lee, E. J., Choi, S. Y. & Kim, E. NMDA receptor dysfunction in autism spectrum disorders. Curr. Opin. Pharmacol. 20, 8–13. https://doi.org/10.1016/j.coph.2014.10.007 (2015).
Article CAS Google Scholar
Anastasiades, P. G. et al. GABAergic interneurons form transient layer-specific circuits in early postnatal neocortex. Nat. Commun. 7, 10584. https://doi.org/10.1038/ncomms10584 (2016).
Article ADS CAS Google Scholar
Bureau, I., Shepherd, G. M. & Svoboda, K. Precise development of functional and anatomical columns in the neocortex. Neuron 42, 789–801. https://doi.org/10.1016/j.neuron.2004.05.002 (2004).
Article CAS Google Scholar
Shepherd, G. M., Pologruto, T. A. & Svoboda, K. Circuit analysis of experience-dependent plasticity in the developing rat barrel cortex. Neuron 38, 277–289. https://doi.org/10.1016/s0896-6273(03)00152-1 (2003).
Article CAS Google Scholar
Meng, X., Winkowski, D. E., Kao, J. P. Y. & Kanold, P. O. Sublaminar subdivision of mouse auditory cortex layer 2/3 based on functional translaminar connections. J. Neurosci. 37, 10200–10214. https://doi.org/10.1523/JNEUROSCI.1361-17.2017 (2017).
Article CAS Google Scholar
Rouaux, C. & Arlotta, P. Direct lineage reprogramming of post-mitotic callosal neurons into corticofugal neurons in vivo. Nat. Cell Biol. 15, 214–221. https://doi.org/10.1038/ncb2660 (2013).
Article CAS Google Scholar
Cserep, C. et al. NMDA receptors in GABAergic synapses during postnatal development. PLoS ONE 7, e37753. https://doi.org/10.1371/journal.pone.0037753 (2012).
Article ADS CAS Google Scholar
Mathew, S. S. & Hablitz, J. J. Presynaptic NMDA receptors mediate IPSC potentiation at GABAergic synapses in developing rat neocortex. PLoS ONE 6, e17311. https://doi.org/10.1371/journal.pone.0017311 (2011).
Article ADS CAS Google Scholar
Xu, X. & Callaway, E. M. Laminar specificity of functional input to distinct types of inhibitory cortical neurons. J. Neurosci. 29, 70–85. https://doi.org/10.1523/JNEUROSCI.4104-08.2009 (2009).
Article CAS Google Scholar
Kavalali, E. T. The mechanisms and functions of spontaneous neurotransmitter release. Nat. Rev. Neurosci. 16, 5–16. https://doi.org/10.1038/nrn3875 (2015).
Article CAS Google Scholar
Chanaday, N. L. & Kavalali, E. T. Presynaptic origins of distinct modes of neurotransmitter release. Curr. Opin. Neurobiol. 51, 119–126. https://doi.org/10.1016/j.conb.2018.03.005 (2018).
Article CAS Google Scholar
Yee, A. X. & Chen, L. Differential regulation of spontaneous and evoked inhibitory synaptic transmission in somatosensory cortex by retinoic acid. Synapse 70, 445–452. https://doi.org/10.1002/syn.21921 (2016).
Article CAS Google Scholar
Isaac, J. T., Nicoll, R. A. & Malenka, R. C. Silent glutamatergic synapses in the mammalian brain. Can. J. Physiol. Pharmacol. 77, 735–737 (1999).
Article CAS Google Scholar
Rumpel, S., Kattenstroth, G. & Gottmann, K. Silent synapses in the immature visual cortex: Layer-specific developmental regulation. J. Neurophysiol. 91, 1097–1101. https://doi.org/10.1152/jn.00443.2003 (2004).
Article Google Scholar
Isaac, J. T. Postsynaptic silent synapses: Evidence and mechanisms. Neuropharmacology 45, 450–460. https://doi.org/10.1016/s0028-3908(03)00229-6 (2003).
Article CAS Google Scholar
Ashby, M. C. & Isaac, J. T. Maturation of a recurrent excitatory neocortical circuit by experience-dependent unsilencing of newly formed dendritic spines. Neuron 70, 510–521. https://doi.org/10.1016/j.neuron.2011.02.057 (2011).
Article CAS Google Scholar
Kullmann, D. M. Silent synapses: what are they telling us about long-term potentiation?. Philos. Trans. R. Soc. Lond B. Biol. Sci. 358, 727–733. https://doi.org/10.1098/rstb.2002.1229 (2003).
Article CAS Google Scholar
Nicoll, R. A. Expression mechanisms underlying long-term potentiation: A postsynaptic view. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 358, 721–726. https://doi.org/10.1098/rstb.2002.1228 (2003).
Article CAS Google Scholar
Kerchner, G. A. & Nicoll, R. A. Silent synapses and the emergence of a postsynaptic mechanism for LTP. Nat. Rev. Neurosci. 9, 813–825. https://doi.org/10.1038/nrn2501 (2008).
Article CAS Google Scholar
Chancey, J. H. et al. GABA depolarization is required for experience-dependent synapse unsilencing in adult-born neurons. J. Neurosci. 33, 6614–6622. https://doi.org/10.1523/JNEUROSCI.0781-13.2013 (2013).
Article CAS Google Scholar
Ultanir, S. K. et al. Regulation of spine morphology and spine density by NMDA receptor signaling in vivo. Proc. Natl. Acad. Sci. USA 104, 19553–19558. https://doi.org/10.1073/pnas.0704031104 (2007).
Article ADS Google Scholar
Adesnik, H., Li, G., During, M. J., Pleasure, S. J. & Nicoll, R. A. NMDA receptors inhibit synapse unsilencing during brain development. Proc. Natl. Acad. Sci. USA 105, 5597–5602. https://doi.org/10.1073/pnas.0800946105 (2008).
Article ADS Google Scholar
Tsien, J. Z., Huerta, P. T. & Tonegawa, S. The essential role of hippocampal CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell 87, 1327–1338. https://doi.org/10.1016/s0092-8674(00)81827-9 (1996).
Article CAS Google Scholar
Friedman, H. V., Bresler, T., Garner, C. C. & Ziv, N. E. Assembly of new individual excitatory synapses: Time course and temporal order of synaptic molecule recruitment. Neuron 27, 57–69. https://doi.org/10.1016/s0896-6273(00)00009-x (2000).
Article CAS Google Scholar
Colonnese, M. T., Shi, J. & Constantine-Paton, M. Chronic NMDA receptor blockade from birth delays the maturation of NMDA currents, but does not affect AMPA/kainate currents. J. Neurophysiol. 89, 57–68. https://doi.org/10.1152/jn.00049.2002 (2003).
Article CAS Google Scholar
Gray, J. A. et al. Distinct modes of AMPA receptor suppression at developing synapses by GluN2A and GluN2B: Single-cell NMDA receptor subunit deletion in vivo. Neuron 71, 1085–1101. https://doi.org/10.1016/j.neuron.2011.08.007 (2011).
Article ADS CAS Google Scholar
Lu, W., Gray, J. A., Granger, A. J., During, M. J. & Nicoll, R. A. Potentiation of synaptic AMPA receptors induced by the deletion of NMDA receptors requires the GluA2 subunit. J. Neurophysiol. 105, 923–928. https://doi.org/10.1152/jn.00725.2010 (2011).
Article CAS Google Scholar
Li, Y., Erzurumlu, R. S., Chen, C., Jhaveri, S. & Tonegawa, S. Whisker-related neuronal patterns fail to develop in the trigeminal brainstem nuclei of NMDAR1 knockout mice. Cell 76, 427–437. https://doi.org/10.1016/0092-8674(94)90108-2 (1994).
Article CAS Google Scholar
Zhang, Z. W., Peterson, M. & Liu, H. Essential role of postsynaptic NMDA receptors in developmental refinement of excitatory synapses. Proc. Natl. Acad. Sci. USA 110, 1095–1100. https://doi.org/10.1073/pnas.1212971110 (2013).
Article ADS Google Scholar
Geiger, J. R. et al. Relative abundance of subunit mRNAs determines gating and Ca2+ permeability of AMPA receptors in principal neurons and interneurons in rat CNS. Neuron 15, 193–204. https://doi.org/10.1016/0896-6273(95)90076-4 (1995).
Article CAS Google Scholar
Labrakakis, C., Rudolph, U. & De Koninck, Y. The heterogeneity in GABAA receptor-mediated IPSC kinetics reflects heterogeneity of subunit composition among inhibitory and excitatory interneurons in spinal lamina II. Front. Cell Neurosci. 8, 424. https://doi.org/10.3389/fncel.2014.00424 (2014).
Article Google Scholar
Browne, S. H. et al. Kinetic and pharmacological properties of GABA(A) receptors in single thalamic neurons and GABA(A) subunit expression. J. Neurophysiol. 86, 2312–2322. https://doi.org/10.1152/jn.2001.86.5.2312 (2001).
Article CAS Google Scholar
Vicini, S. et al. GABA(A) receptor alpha1 subunit deletion prevents developmental changes of inhibitory synaptic currents in cerebellar neurons. J. Neurosci. 21, 3009–3016 (2001).
Article CAS Google Scholar
Bacci, A., Rudolph, U., Huguenard, J. R. & Prince, D. A. Major differences in inhibitory synaptic transmission onto two neocortical interneuron subclasses. J. Neurosci. 23, 9664–9674 (2003).
Article CAS Google Scholar
Jami, S. A. et al. Increased excitation-inhibition balance and loss of GABAergic synapses in the serine racemase knockout model of NMDA receptor hypofunction. J. Neurophysiol. 126, 11–27. https://doi.org/10.1152/jn.00661.2020 (2021).
Article CAS Google Scholar
Tatard-Leitman, V. M. et al. Pyramidal cell selective ablation of N-methyl-D-aspartate receptor 1 causes increase in cellular and network excitability. Biol. Psychiatry 77, 556–568. https://doi.org/10.1016/j.biopsych.2014.06.026 (2015).
Article CAS Google Scholar
Sohal, V. S. & Rubenstein, J. L. R. Excitation-inhibition balance as a framework for investigating mechanisms in neuropsychiatric disorders. Mol. Psychiatry 24, 1248–1257. https://doi.org/10.1038/s41380-019-0426-0 (2019).
Article Google Scholar
Aguilar, D. D. et al. Altered neural oscillations and behavior in a genetic mouse model of NMDA receptor hypofunction. Sci. Rep. 11, 9031. https://doi.org/10.1038/s41598-021-88428-9 (2021).
Article ADS CAS Google Scholar
Horvath, P. M., Piazza, M. K., Monteggia, L. M. & Kavalali, E. T. Spontaneous and evoked neurotransmission are partially segregated at inhibitory synapses. Elife https://doi.org/10.7554/eLife.52852 (2020).
Article Google Scholar
Brill, J., Mattis, J., Deisseroth, K. & Huguenard, J. R. LSPS/optogenetics to improve synaptic connectivity mapping: Unmasking the role of basket cell-mediated feedforward inhibition. eNeuro, https://doi.org/10.1523/ENEURO.0142-15.2016 (2016).
Allene, C. et al. Sequential generation of two distinct synapse-driven network patterns in developing neocortex. J. Neurosci. 28, 12851–12863. https://doi.org/10.1523/JNEUROSCI.3733-08.2008 (2008).
Article CAS Google Scholar
Allene, C. & Cossart, R. Early NMDA receptor-driven waves of activity in the developing neocortex: Physiological or pathological network oscillations?. J. Physiol. 588, 83–91. https://doi.org/10.1113/jphysiol.2009.178798 (2010).
Article CAS Google Scholar
Mizuno, H. et al. NMDA receptor enhances correlation of spontaneous activity in neonatal barrel cortex. J. Neurosci. 41, 1207–1217. https://doi.org/10.1523/JNEUROSCI.0527-20.2020 (2021).
Article CAS Google Scholar
Babij, R. & De Marco Garcia, N. Neuronal activity controls the development of interneurons in the somatosensory cortex. Front. Biol. 11, 459–470. https://doi.org/10.1007/s11515-016-1427-x (2016).
Article Google Scholar
De Marco Garcia, N. V., Karayannis, T. & Fishell, G. Neuronal activity is required for the development of specific cortical interneuron subtypes. Nature 472, 351–355. https://doi.org/10.1038/nature09865nature09865[pii] (2011).
Article ADS Google Scholar
De Marco Garcia, N. V., Priya, R., Tuncdemir, S. N., Fishell, G. & Karayannis, T. Sensory inputs control the integration of neurogliaform interneurons into cortical circuits. Nat. Neurosci. 18, 393–401. https://doi.org/10.1038/nn.3946 (2015).
Article CAS Google Scholar
Deng, R., Kao, J. P. Y. & Kanold, P. O. Aberrant development of excitatory circuits to inhibitory neurons in the primary visual cortex after neonatal binocular enucleation. Sci. Rep. 11, 3163. https://doi.org/10.1038/s41598-021-82679-2 (2021).
Article ADS CAS Google Scholar
Turrigiano, G. Too many cooks? Intrinsic and synaptic homeostatic mechanisms in cortical circuit refinement. Annu. Rev. Neurosci. 34, 89–103. https://doi.org/10.1146/annurev-neuro-060909-153238 (2011).
Article CAS Google Scholar
Turrigiano, G. G. & Nelson, S. B. Homeostatic plasticity in the developing nervous system. Nat. Rev. Neurosci. 5, 97–107. https://doi.org/10.1038/nrn1327 (2004).
Article CAS Google Scholar
Banerjee, A., Larsen, R. S., Philpot, B. D. & Paulsen, O. Roles of presynaptic NMDA receptors in neurotransmission and plasticity. Trends Neurosci. 39, 26–39. https://doi.org/10.1016/j.tins.2015.11.001 (2016).
Article CAS Google Scholar
Gill, I. et al. Presynaptic NMDA receptors–dynamics and distribution in developing axons in vitro and in vivo. J. Cell Sci. 128, 768–780. https://doi.org/10.1242/jcs.162362 (2015).
Article CAS Google Scholar
Khazipov, R. & Luhmann, H. J. Early patterns of electrical activity in the developing cerebral cortex of humans and rodents. Trends Neurosci. 29, 414–418. https://doi.org/10.1016/j.tins.2006.05.007 (2006).
Article CAS Google Scholar
Luhmann, H. J. & Khazipov, R. Neuronal activity patterns in the developing barrel cortex. Neuroscience 368, 256–267. https://doi.org/10.1016/j.neuroscience.2017.05.025 (2018).
Article CAS Google Scholar
Winnubst, J., Cheyne, J. E., Niculescu, D. & Lohmann, C. Spontaneous activity drives local synaptic plasticity in vivo. Neuron 87, 399–410. https://doi.org/10.1016/j.neuron.2015.06.029 (2015).
Article CAS Google Scholar
Katz, L. C. & Shatz, C. J. Synaptic activity and the construction of cortical circuits. Science 274, 1133–1138. https://doi.org/10.1126/science.274.5290.1133 (1996).
Article ADS CAS Google Scholar
Mooney, R., Penn, A. A., Gallego, R. & Shatz, C. J. Thalamic relay of spontaneous retinal activity prior to vision. Neuron 17, 863–874. https://doi.org/10.1016/s0896-6273(00)80218-4 (1996).
Article CAS Google Scholar
Wong, R. O., Meister, M. & Shatz, C. J. Transient period of correlated bursting activity during development of the mammalian retina. Neuron 11, 923–938. https://doi.org/10.1016/0896-6273(93)90122-8 (1993).
Article CAS Google Scholar
Anton-Bolanos, N. et al. Prenatal activity from thalamic neurons governs the emergence of functional cortical maps in mice. Science 364, 987–990. https://doi.org/10.1126/science.aav7617 (2019).
Article ADS CAS Google Scholar
Moreno-Juan, V. et al. Prenatal thalamic waves regulate cortical area size prior to sensory processing. Nat. Commun. 8, 14172. https://doi.org/10.1038/ncomms14172 (2017).
Article ADS CAS Google Scholar
Yamamoto, N. & Lopez-Bendito, G. Shaping brain connections through spontaneous neural activity. Eur. J. Neurosci. 35, 1595–1604. https://doi.org/10.1111/j.1460-9568.2012.08101.x (2012).
Article Google Scholar
Molnar, Z., Luhmann, H. J. & Kanold, P. O. Transient cortical circuits match spontaneous and sensory-driven activity during development. Science https://doi.org/10.1126/science.abb2153 (2020).
Article Google Scholar
Vasistha, N. A. et al. Cortical and clonal contribution of Tbr2 expressing progenitors in the developing mouse brain. Cereb. Cortex 25, 3290–3302. https://doi.org/10.1093/cercor/bhu125 (2015).
Article Google Scholar
Kowalczyk, T. et al. Intermediate neuronal progenitors (basal progenitors) produce pyramidal-projection neurons for all layers of cerebral cortex. Cereb. Cortex 19, 2439–2450. https://doi.org/10.1093/cercor/bhn260 (2009).
Article Google Scholar
Mihalas, A. B. et al. Intermediate progenitor cohorts differentially generate cortical layers and require Tbr2 for timely acquisition of neuronal subtype identity. Cell. Rep. 16, 92–105. https://doi.org/10.1016/j.celrep.2016.05.072 (2016).
Article CAS Google Scholar
Yuste, R. et al. A community-based transcriptomics classification and nomenclature of neocortical cell types. Nat. Neurosci. 23, 1456–1468. https://doi.org/10.1038/s41593-020-0685-8 (2020).
Article CAS Google Scholar
Gouwens, N. W. et al. Classification of electrophysiological and morphological neuron types in the mouse visual cortex. Nat. Neurosci. 22, 1182–1195. https://doi.org/10.1038/s41593-019-0417-0 (2019).
Article CAS Google Scholar
Hodge, R. D. et al. Conserved cell types with divergent features in human versus mouse cortex. Nature 573, 61–68. https://doi.org/10.1038/s41586-019-1506-7 (2019).
Article ADS CAS Google Scholar
Shimogori, T. Micro in utero electroporation for efficient gene targeting in mouse embryos. CSH Protoc. https://doi.org/10.1101/pdb.prot4447 (2006).
Article Google Scholar
Matsui, A., Yoshida, A. C., Kubota, M., Ogawa, M. & Shimogori, T. Mouse in utero electroporation: Controlled spatiotemporal gene transfection. J. Vis. Exp. https://doi.org/10.3791/3024 (2011).
Article Google Scholar
Suter, B. A. et al. Ephus: Multipurpose data acquisition software for neuroscience experiments. Front. Neural Circuits 4, 100. https://doi.org/10.3389/fncir.2010.00100 (2010).
Article Google Scholar
Zhao, C., Kao, J. P. & Kanold, P. O. Functional excitatory microcircuits in neonatal cortex connect thalamus and layer 4. J. Neurosci. 29, 15479–15488. https://doi.org/10.1523/JNEUROSCI.4471-09.2009 (2009).
Article CAS Google Scholar
Muralidharan, S. et al. Ncm, a photolabile group for preparation of caged molecules: Synthesis and biological application. PLoS ONE 11, e0163937. https://doi.org/10.1371/journal.pone.0163937 (2016).
Article CAS Google Scholar
Oviedo, H. V., Bureau, I., Svoboda, K. & Zador, A. M. The functional asymmetry of auditory cortex is reflected in the organization of local cortical circuits. Nat. Neurosci. 13, 1413–1420. https://doi.org/10.1038/nn.2659 (2010).
Article CAS Google Scholar
Meng, X., Mukherjee, D., Kao, J. P. Y. & Kanold, P. O. Early peripheral activity alters nascent subplate circuits in the auditory cortex. Sci. Adv. https://doi.org/10.1126/sciadv.abc9155 (2021).
Article Google Scholar
Rosenthal, R. In The handbook of research synthesis (ed Harris; Hedges Cooper, Larry V.) (Russell Sage Foundation, 1994).

Download references

Acknowledgements

The authors thank Dr. Tomomi Shimogori (RIKEN) for establishing IUE and Dr. Paola Arlotta (Harvard University) for sharing Cdk5r promoter plasmids. Supported by NIH RO1DC009607 (POK) and NIH R01 GM056481 (JPYK).

Author information

Authors and Affiliations

Department of Biomedical Engineering, School of Medicine, Johns Hopkins University, 733 N. Broadway Avenue / Miller 379, Baltimore, MD, 21205, USA
Minzi Chang & Patrick O. Kanold
Department of Biology, University of Maryland, College Park, MD, 20742, USA
Rongkang Deng & Patrick O. Kanold
Biological Sciences Graduate Program, University of Maryland, College Park, MD, 20742, USA
Rongkang Deng
Center for Biomedical Engineering and Technology, Department of Physiology, University of Maryland School of Medicine, Baltimore, MD, 21201, USA
Joseph P. Y. Kao

Authors

Rongkang Deng
View author publications
You can also search for this author in PubMed Google Scholar
Minzi Chang
View author publications
You can also search for this author in PubMed Google Scholar
Joseph P. Y. Kao
View author publications
You can also search for this author in PubMed Google Scholar
Patrick O. Kanold
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

R.D., and P.O.K. designed research. R.D. performed experiments and analyzed data. R.D. and P.O.K. discussed the results and wrote the manuscript. M.C. performed immunostaining, analyzed data and edited the manuscript. J.P.Y.K. contributed reagents and edited the manuscript.

Corresponding author

Correspondence to Patrick O. Kanold.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary Information.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Deng, R., Chang, M., Kao, J.P.Y. et al. Cortical inhibitory but not excitatory synaptic transmission and circuit refinement are altered after the deletion of NMDA receptors during early development. Sci Rep 13, 656 (2023). https://doi.org/10.1038/s41598-023-27536-0

Download citation

Received: 17 November 2021
Accepted: 04 January 2023
Published: 12 January 2023
DOI: https://doi.org/10.1038/s41598-023-27536-0

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.