Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/springer-journals/interneuronal-gluk1-kainate-receptors-control-maturation-of-gabaergic-ZORrejDuLP
- Publisher
- Springer Journals
- Copyright
- Copyright © The Author(s) 2023
- eISSN
- 1756-6606
- DOI
- 10.1186/s13041-023-01035-9
- Publisher site
- See Article on Publisher Site
Abstract
Kainate type glutamate receptors (KARs) are strongly expressed in GABAergic interneurons and have the capability of modulating their functions via ionotropic and G-protein coupled mechanisms. GABAergic interneurons are critical for generation of coordinated network activity in both neonatal and adult brain, yet the role of interneuronal KARs in network synchronization remains unclear. Here, we show that GABAergic neurotransmission and spontaneous net- work activity is perturbed in the hippocampus of neonatal mice lacking GluK1 KARs selectively in GABAergic neurons. Endogenous activity of interneuronal GluK1 KARs maintains the frequency and duration of spontaneous neonatal network bursts and restrains their propagation through the hippocampal network. In adult male mice, the absence of GluK1 in GABAergic neurons led to stronger hippocampal gamma oscillations and enhanced theta-gamma cross frequency coupling, coinciding with faster spatial relearning in the Barnes maze. In females, loss of interneuronal GluK1 resulted in shorter sharp wave ripple oscillations and slightly impaired abilities in flexible sequencing task. In addition, ablation of interneuronal GluK1 resulted in lower general activity and novel object avoidance, while causing only minor anxiety phenotype. These data indicate a critical role for GluK1 containing KARs in GABAergic interneurons in regulation of physiological network dynamics in the hippocampus at different stages of development. Keywords Glutamate receptor, Kainate receptor, GABAergic interneuron, Hippocampus, Network synchronization, Gamma oscillation, Cognitive flexibility Introduction Balanced interplay between GABAergic and glutamater- gic transmission is important for generation of coordi- nated network activity both during development and in the adult brain [1–4]. In the immature circuitry, the Simo Ojanen and Tatiana Kuznetsova have contributed equally to this work high-frequency oscillatory components of spontane- and co-first authors. ous network activity provide the functional templates *Correspondence: for activity-dependent synaptogenesis [2, 5–7, reviewed Sari E. Lauri in [8]] and constrain apoptosis [8, 9], thereby shaping the sari.lauri@helsinki.fi development and fine-tuning of the network connectiv - Tomi Taira tomi.taira@helsinki.fi ity. In the adult circuitry, GABAergic activity maintains Department of Veterinary Biosciences, Faculty of Veterinary Medicine, and paces the rhythmic oscillations suggested to underlie University of Helsinki, Helsinki, Finland various cognitive and motor functions [10]. HiLife Neuroscience Center, University of Helsinki, Helsinki, Finland Molecular and Integrative Biosciences Research Program, University Kainate receptors (KARs) are ionotropic glutamate of Helsinki, Helsinki, Finland receptors, also capable of metabotropic functions, that © The Author(s) 2023. 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:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Ojanen et al. Molecular Brain (2023) 16:43 Page 2 of 20 can profoundly influence the balance between inhibition Results and excitation in neuronal networks [11–14]. They con - Loss of GluK1 expression in the interneurons delays sist of five subunits, GluK1—GluK5, encoded by genes development of GABAergic synaptic transmission Grik1—Grik5 [12]. GluK1 in particular, has an interesting To monitor how loss of GluK1 expression selectively in developmental expression pattern in the hippocampus, GABAergic neurons affects synaptic transmission at dif - being expressed in both principal neurons and GABAe- ferent stages of development, we performed whole-cell rgic interneurons during the first postnatal weeks [15], patch clamp recordings from pyramidal cells located in after which its expression is mainly confined to interneu - the CA3 region in acute slices from control mice, with tmc1/tm1c rons [16–18]. The early GluK1 activity in principal neu - floxed Grik1 (Grik1 ), and mutant mice lacking rons plays a crucial role in regulating glutamate release, GluK1 expression in the GABAergic interneurons (Gad2- tm1d/tm1d synaptic plasticity and activity-dependent development Grik1 ). From here on, these mice are referred to −/− of functional neural networks both in the hippocam- as control and Gad-Grik1 , respectively. To be able to pus and in the amygdala [19–21]; however, the role of look at GABAergic and glutamatergic activity simultane- interneuronal GluK1 in development of behaviorally rel- ously, we used an intracellular solution with low chloride evant neural circuits is less clear. (2 mM) concentration in order to shift GABA-A rever- Pharmacological studies have indicated that activation sal potential [22] and clamped the membrane potential at of GluK1 KARs strongly recruits spontaneous GABAergic − 50 mV. Under these conditions, spontaneous inhibitory activity by direct ionotropic depolarization of interneu- GABA-A receptor mediated events (sIPSCs) appeared rons (e.g. [22–24]). In addition, endogenous activity of as an outward currents and excitatory glutamatergic GluK1 subunit containing KARs in the immature hip- events (sEPSCs) as inward currents (Fig. 1A). We did pocampus maintains high excitability of interneurons, not observe any difference in sEPSC frequency between −/− typical for early development, by regulating the function control and Gad-Grik1 groups across different devel - of calcium-activated potassium channels (SK channels) opmental stages (2-way ANOVA, p = 0.3; Fig. 1C). In in a G-protein dependent manner [42]. Both mechanisms both genotypes, sEPSCs were very rare in neonatal result in increased activation of GABAergic neurons, mice (0.265 ± 0.046 Hz and 0.153 ± 0.026 Hz; Fig. 1Ai), suggesting that the physiological role of interneuronal but became more frequent in juvenile (1.161 ± 0.250 Hz GluK1 KARs is to act as a feedback mechanism to main- and 1.377 ± 0.361 Hz; Fig. 1Aii) and adult groups tain the network excitability in a critical range during (0.752 ± 0.171 Hz and 1.469 ± 0.250 Hz, Fig. 1Aiii, for con- −/− developmental and activity-dependent fluctuations in the trol and Gad-Grik1 , respectively; age effect p < 0.0001; levels of endogenous glutamate [11]. However, activation Fig. 1C). On the contrary, the frequency of sIPSCs was −/− of GluK1 KARs can also result in depression of evoked different between control and Gad-Grik1 mice (2-way GABAergic transmission via G-protein dependent inhi- ANOVA, genotype effect p = 0.002), in particular in the bition of GABA release [23, 25], particularly at CCK/CB1 neonatal group, where it was dramatically lower in the −/− expressing interneurons [26], suggesting that interneu- Gad-Grik1 (1.036 ± 0.161 Hz) as compared to the con- ronal KARs might have more complex roles in shaping trol mice (4.243 ± 0.617 Hz; p < 0.000001, Mann–Whit- the network activity patterns towards adulthood. ney test; Fig. 1Ai, B). The reduced sIPSC frequency in the These data indicate that GluK1 KARs are closely mutants did not persist later in life and was comparable involved in functional regulation of the GABAergic net- to that in controls in both juvenile (3.655 ± 0.689 Hz, −/− work in the early postnatal hippocampus and suggest Gad-Grik1 vs 3.636 ± 1.113 Hz, control) and adult −/− that its malfunction might influence adult oscillatory net - (2.762 ± 0.711 Hz in Gad-Grik1 vs 4.666 ± 1.586 Hz in work activity and associated behaviors. However, surpris- control; Mann–Whitney test) mice (Fig. 1Aii, iii, B). We ingly little data exists on the roles of KARs in regulation did not observe any significant differences between gen - of physiological circuit activity and brain oscillations. otypes in sIPSC or sEPSC amplitudes (Additional file 1: Here we take the advantage of a mutant mouse line in Fig. S1A). Taken together, these results indicate that the which GluK1 expression has been selectively ablated in absence of GluK1 in the GABAergic interneurons results GABAergic interneurons, to understand the significance in impaired GABAergic transmission to CA3 pyramidal of interneuronal GluK1 for neonatal neurotransmission neurons in the neonatal hippocampus, but its effect is and network synchronization. Furthermore, the conse- developmentally restricted and levels out later in life. quences on adult hippocampal activity such as theta and gamma oscillations and sharp wave ripples are investi- gated in parallel with hippocampus dependent behaviors. O janen et al. Molecular Brain (2023) 16:43 Page 3 of 20 Fig. 1 Ablation of GluK1 in GABAergic interneurons attenuates GABAergic synaptic activity in the neonatal but not juvenile or adult CA3. A Example −/− traces of whole-cell patch-clamp recordings from neonatal (i), juvenile (ii) and adult (iii) control (left) and Gad-Grik1 (right) mice with sIPSCs appearing as outward and sEPSCs as inward currents. B, C Basal frequency of sIPSCs (B) and sEPSCs (C) in pyramidal CA3 cells from acute control −/− and Gad-Grik1 mice slices across different age groups (neonatal: n = 30 (21) and 24 (19); juvenile: n = 14 (12) and 13(11); adult: n = 11 (10) and 9 −/− (9), for control and Gad-Grik1 respectively; n refers to number of cells, followed by number of animals in parenthesis. Bars represent mean ± SEM. Frequencies were compared by 2-way ANOVA with multiple Mann–Whitney test as a post-hock to detect differences in genotype for each age group. **** p < 0.000001 (Mann–Whitney test) GABAergic and glutamatergic synaptic transmission different stages of development, we tested the effects of is regulated by distinct subpopulations of GluK1 KARs GluK1-selective agonist ATPA and antagonist ACET on in the neonatal hippocampus spontaneous synaptic activity in CA3 pyramidal neu- −/− To further understand the cell-type specific roles rons from control and Gad-Grik1 mice. of GluK1 in regulation of synaptic transmission at Ojanen et al. Molecular Brain (2023) 16:43 Page 4 of 20 ATPA (1 µm) application robustly increased sIPSC diminishing with age in accordance with the declin- frequency in neonatal (5.415 ± 0.983 times , p < 0.0001; ing expression of GluK1 towards the adulthood. In paired t-test) and juvenile (3.368 ± 0.598, p = 0.008) adult animals, ATPA increased sIPSC frequency, control mice (Fig. 2Ai, ii, B), confirming previous yet this effect did not reach statistical significance results [22, 23]. Notably, ATPA effect was gradually (2.589 ± 0.521 times, p = 0.081, paired t-test; Fig. 2Aiii, −/− Fig. 2 Pharmacological characterization of spontaneous synaptic activity in Gad-Grik1 mice. A Example traces of recordings from neonatal (i −/− and iv), juvenile (ii) and adult (iii) control (left columns) and Gad-Grik1 (right columns) slices, before (baseline) and during ATPA (i, ii, iii) or ACET −/− (iv) application. B Pooled data illustrating the effect of ATPA (1 µM) on sIPSC frequency in CA3 pyramidal cells from acute control and Gad-Grik1 slices at different stages of development (neonatal: n = 14 (10) and 10 (10); juvenile: n = 10 (10) and 8 (7); adult: n = 5 (5) and 6 (6), for control −/− and Gad-Grik1 , respectively). C Eec ff t of ATPA on sEPSC frequency, for the same cells as in B. D Eec ff t of ACET (200 nM) on the frequency of −/− −/− sEPSCs and sIPSCs in neonatal control and Gad-Grik1 slices (n = 13 (N = 10) and 7 (6), for control and Gad-Grik1 , respectively). Bars represent mean ± SEM. Frequency of events is normalized to the baseline (dashed line). Activity during ATPA / ACET application is compared to the baseline by paired t-test or Wilcoxon paired test. ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05 O janen et al. Molecular Brain (2023) 16:43 Page 5 of 20 −/− Immature‑type network activity in the hippocampus B). In Gad-Grik1 mice, ATPA had no effect on is regulated by GluK1 KARs expressed in both, sIPSC frequency neither in neonates nor in the juve- glutamatergic and GABAergic neurons nile animals, but surprisingly, caused a slight increase Pharmacological manipulation of GluK1 subunit con- in sIPSC frequency in adults (1.572 ± 0.258 times , taining KARs modulate the spontaneous network activity p = 0.029, paired t-test; Fig. 2 Ai-iii, B). These observa- in the neonatal hippocampus [22]. Since GluK1 recep- tions confirm that GluK1 activation in the hippocam- tors are expressed both in GABAergic and glutamatergic pal interneurons efficiently recruit GABAergic drive, neurons [15], the cell type mainly responsible for GluK1- particularly at earlier developmental stages. dependent regulation of the network excitability remains Consistent with previous data [22, 27], ATPA appli- unknown. cation significantly decreased sEPSC frequency in In the neonatal hippocampus, spontaneous net- neonatal (to 0.591 ± 0.063 from the baseline, p = 0.003, work bursts can be readily recorded from CA3 pyrami- paired t-test) and in young (to 0.590 ± 0.069, p = 0.004) dal neurons, and consist of a slow GABAergic current, control mice, but had no effect on sEPSCs in the adult intermixed with EPSCs (Fig. 3Ai) [22]. The frequency animals (Fig. 2Ai-iii, C). The strong effects of ATPA on of these network events was significantly lower in GABAergic transmission could result in attenuation −/− the Gad-Grik1 slices (0.025 ± 0.003 Hz), as com- of glutamatergic transmission indirectly, via changes pared to controls (0.034 ± 0.003 Hz; p = 0.016, unpaired in network excitability. However, ATPA application t-test; Fig. 3Ai, B), pointing out the important role of decreased sEPSC frequency also in the neonatal Gad- −/− interneuronal GluK1 in this form of network activity. Grik1 mice (to 0.678 ± 0.100, p < 0.006, Wilcoxon As described earlier [22], application of ATPA almost paired test), where no parallel changes in sIPSCs were completely abolished network bursts in the control ani- observed (Fig. 2B, C). ATPA had no effect on sEP- −/− mals (to 0.050 ± 0.031 of the baseline; p < 0.0001; paired SCs in the juvenile Gad-Grik1 slices, but caused a t-test), but had no effect on burst frequency in the Gad- slight reduction of sEPSC frequency in the adult stage −/− Grik1 mice (to 1.016 ± 0.081 of the baseline, p = 0.703, (to 0.801 ± 0.057, p = 0.030, paired t-test). These data paired t-test; Fig. 3Aii, C). Interestingly, ACET applica- are consistent with expression of presynaptic inhibi- tion reduced the frequency of network oscillations both tory GluK1 KARs in glutamatergic neurons in the neo- in slices from the control (to 0.749 ± 0.080, p = 0.018) nate hippocampus [19, 22, 27], while in the juvenile −/− and Gad-Grik1 (0.535 ± 0.125, p = 0.011) animals mice, sEPSC frequency appears to be regulated indi- (Fig. 3Aiii, C). These pharmacological data indicate that rectly, via the ATPA-induced changes in GABAergic in absence of GluK1 in the interneurons, spontaneous drive. The cellular basis underlying the small effects network bursts retain the sensitivity to GluK1 inhibition of ATPA on spontaneous transmission in the adult −/− (by ACET), suggesting that ongoing regulation of net- Gad-Grik1 mice remain unclear, but might repre- work activity depends on GluK1 KARs located in prin- sent some unspecific actions of ATPA on other KAR cipal neurons. Furthermore, loss of the ATPA dependent subtypes (e.g. [28, 29]), compensating for the loss of −/− regulation in Gad-Grik1 mice confirms that activation GluK1 in the mutants. of GluK1 receptors in interneurons impose a strong con- Inhibition of GluK1 containing KARs by ACET trol over the circuit excitability. (200 nM) triggered increase in sEPSC frequency in both control (2.022 ± 0.344 times ; p < 0.001, Wil- −/− coxon paired test) and Gad-Grik1 neonatal slices (1.503 ± 0.236 times; p = 0.031, paired t-test; Fig. 2Aiv, Interneuronal GluK1 KARs regulate synchronization D). At the same time, ACET had no effect on sIPSC and propagation of activity in cultured hippocampal slices frequency in either group. This further confirms that Spontaneous synchronous activity is observed in the GluK1 receptors, located in glutamatergic principal acute slices only during the first postnatal days. To fur - neurons, tonically inhibit glutamatergic transmission ther investigate the cell type-specific roles of GluK1 during early postnatal development [19, 22]. We did in network synchronization, we prepared organotypic not test ACET effect later on in development, since hippocampal cultures from P4-P5 mice and recorded endogenous activity of GluK1 receptors is limited the spontaneous network activity longitudinally, using to early postnatal age. In line with the previous data 8 × 8 electrode MEA probes at four time points, at DIV [22] neither ATPA nor ACET application (Additional 5, 7, 9 and 10. In the CA1 and DG regions, the spike −/− file 1: Fig. S1B, C, respectively) lead to any significant rates detected from Gad-Grik1 mice did not dif- changes in the sIPSC or sEPSC amplitudes. fer from controls (Fig. 4A, Bi, iii). In the CA3 region, Ojanen et al. Molecular Brain (2023) 16:43 Page 6 of 20 −/− Fig. 3 Pharmacological characterization of spontaneous network activity in neonatal Gad-Grik1 mice. A Example traces of spontaneous activity −/− in baseline (i) and upon ATPA (ii) or ACET (iii) application, recorded from CA3 pyramidal cells in acute slices from control (left) and Gad-Grik1 (right) mice (P4-6). Inserts on top represent network bursts (marked with asterisk) in expanded time scale. B Basal frequency of spontaneous −/− network bursts from acute control (n = 30 (21)) and Gad-Grik1 (n = 24 (19)) neonatal slices (P4-6). *p < 0.05, upaired t-test test. C Eec ff t of ATPA −/− (1 µM) and ACET (200 nM) on burst frequency in control and Gad-Grik1 groups. ATPA, n = 14(10) and 10(10); ACET n = 13(10) and 7(6) for control −/− and Gad-Grik1 , respectively. Frequency is normalized to the baseline (dashed line). ****p < 0.0001; **p < 0.01; *p < 0.05, paired t-test −/− however, slices from Gad-Grik1 mice showed data to look at the spread of activity in the network by reduced spike rates through the whole culturing period computing the maximum number of spatially adjacent (p = 0.0001, 2-way ANOVA; Fig. 4A, Bii). Furthermore, MEA channels displaying activity within 0.1 s time the frequency (p < 0.05, 2-way ANOVA) and dura- window after a spike detected in the dentate gyrus tion (p < 0.01, 2-way ANOVA) of large (minimum 10 (DG). The basal excitability, assessed by the average spikes within minimum 0.1 s time) bursts was signifi - number of clustered active channels, was lower in the −/− −/− cantly lower in Gad-Grik1 cultures as compared to Gad-Grik1 slices as compared to controls (Fig. 4E– controls (Fig. 4C, D). We then further analyzed these G). When the data was normalized to the first 20 ms (See figure on next page.) −/− Fig. 4 Longitudinal characterization of spontaneous network activity in organotypic hippocampal cultures from neonatal Gad-Grik1 mice. A −/− Example traces of the extracellular recordings across hippocampal regions in control and Gad-Grik1 cultures at DIV5. The inserts show typical bursts with the CA1/CA3–DG phase shift. Red arrowheads mark network bursts. B Frequency of spikes detected in the CA1 (i), CA3 (ii) and DG (iii) −/− regions of control (n = 12 (6)) and Gad-Grik1 cultures (n = 9 (6)). Both genotypes show equal spike rates in the CA1 and DG regions, but the −/− frequency of spikes in area CA3 is lower in the Gad-Grik1 as compared to controls (p = 0.0001, 2-way ANOVA). C Frequency of spontaneous −/− bursts in the CA1, CA3, DG regions of control (n = 11(6)) and Gad-Grik1 slices (n = 9 (6)) at indicated DIV (p < 0.05, 2-way ANOVA). D Duration −/− of spontaneous bursts in CA1, CA3, DG regions of control (n = 11(6)) and Gad-Grik1 (n = 9(6)) slices at indicated DIV (p < 0.01, 2-way ANOVA). E Image of the hippocampal organotypic culture on MEA probe (left) and example heat maps showing the spread of network activity following a −/− spike in the DG, in control and Gad-Grik1 cultures at 5 DIV. F Time course plot illustrating the mean number of spatially adjacent MEA channels −/− (out of total 64) displaying activity during a 100 ms time period after a DG spike in control (n = 10(6)) and Gad-Grik1 (n = 9(6)) cultures. The dashed blue line denotes the time point of DG spike. G Pooled data on the maximum number of spatially adjacent active channels during the first 20 ms following a DG spike (*p < 0.05, Mann–Whitney U-test). H The same data as in F, normalized to the first 4 time bins (20 ms) following the spike. (** p < 0.01, 2-way ANOVA). I Normalized number of spatially adjacent active channels 30–70 ms after the DG spike (****p < 0.0001, Mann– Whitney U-test.). Bars represent mean ± SEM. The n-values refer to number of cultures, followed by number of animals in parenthesis O janen et al. Molecular Brain (2023) 16:43 Page 7 of 20 Fig. 4 (See legend on previous page.) −/− after the spike (Fig. 4H, I). Consistent with the findings following the spike detected in DG, the Gad-Grik1 from acute neonatal slices, these data from organo- slices showed a greater increase in the number of clus- typic cultures indicates that ablation of GluK1 from the tered active channels (4.817 ± 0.6032 vs. 1.635 ± 0.1563 −/− interneurons attenuates synchronous network activ- in Gad-Grik1 and control, respectively, p < 0.0001, ity, observed as lower frequency of the spontaneous Mann–Whitney test) during the 30–70 ms period Ojanen et al. Molecular Brain (2023) 16:43 Page 8 of 20 bursting activity. At the same time, however, the spa- the gamma power in CA1 region was more strongly tial propagation of activity within the hippocampal net- modulated by theta phase than in the controls (Fig. 5Di, work is enhanced. p < 0.0001, quadratic regression, sum-of-squares F test). In females there were no differences between genotypes Ablation of interneuronal GluK1 leads to altered network in the theta phase to gamma power-cross frequency oscillations in the adult hippocampus in vivo modulation (Fig. 5Dii). Interneurons regulate all the main types of rhythms SWRs consists of the sharp wave detected as large occurring in the awake hippocampus: the theta amplitude negative deflection in the LFP in CA1 stratum (4–12 Hz), gamma (20–150 Hz) and sharp wave ripple radiatum, followed by a high frequency (150–200 Hz) (SWR) oscillations. In order to investigate the role of ripple oscillation in the CA1 pyramidal layer. We used interneuronal GluK1 in physiological oscillations in the the Kay ripple detection method [36] to identify ripple adult hippocampus, we performed a series of record- oscillations in the 150–200 Hz frequency range in the ings in awake, head-fixed mice, using multichannel CA1 pyramidal layer. Ripples were detected from epochs linear probes spanning over the CA1 and DG regions where the mouse was idle or resting, and the occurrence and separately, the CA3 region of the hippocampus rate of ripples was calculated across the full duration of (Fig. 5Ai and ii, respectively). Since the oscillatory such epochs. Ripples occurred as frequently in both gen- power in the hippocampus depends on locomotion and otypes, and were more frequent in females than in males exploratory activity [30, 31], the recorded signals were (2-way ANOVA gender effect p = 0.0393, Fig. 5E). How- divided into epochs of idle and running based on video ever, in female mice the duration of ripples was shorter analysis of the mouse activity. (p = 0.0180, two-tailed unpaired t-test, Fig. 5F) and the −/− The Gad-Grik1 mice did not differ from their lit - percentage of ripples exceeding 100 ms in duration was termate controls in oscillatory power in the theta fre- lower (p = 0.0047, two-tailed unpaired t-test, Fig. 5G) in −/− quency range (Fig. 5B) either during the resting state, Gad-Grik1 mice as compared to controls. nor while they were actively moving (Additional file 2: Fig. S2), with the exception of the high frequency Ablation of interneuronal GluK1 leads to reduced general (7–13 Hz) sub-band of theta power being higher in the activity, novel object avoidance and changes in cognitive −/− Gad-Grik1 male mice than controls when the mice flexibility were moving (Additional file 2: Fig. S2Di, p = 0.0167, The differences we found in excitability and synchronous Holm-Šídák). However, oscillatory power in the gamma network activity both in vitro and in vivo led us to inves- −/− range was elevated in the Gad-Grik1 males, in the tigate how knocking out interneuronal GluK1 affects the DG (p < 0.05, 2-way ANOVA, Tukey’s post hoc test) behavior of mice in adulthood. For this purpose, we per- independent on the behavioral state (Fig. 5C). On the formed a battery of behavioral tests on littermate control −/− contrary, in the female mice (n = 5) there were no dif- and Gad-Grik1 mice. First, we looked at the activity ferences between genotypes in any of the frequency and possible anxiety-like phenotypic features using the ranges we investigated (Fig. 5B, C). open field (OF) and elevated plus maze (EPM) tests. The −/− As theta-gamma coupling occurs in the hippocam- total distance traveled by Gad-Grik1 mice during both pus during running and REM-sleep in rodents [30, 32] the EPM (Fig. 6Ai) and OF tests (6Bi) was significantly and correlates with learning of spatial tasks [33–35], we lower as compared to controls in both males and females, next looked at how gamma power is modulated by theta indicating lower level of activity (p < 0.0001 for both −/− phase in two genotypes. In the Gad-Grik1 male mice tests, 2-way ANOVA). This phenotype was confirmed (See figure on next page.) Fig. 5 Physiological network oscillations are altered in the adult hippocampus in absence of GluK1 expression in the GABAergic neurons. A Image illustrating the localization of the fluorescently labelled recording electrode in a coronal section from the mouse hippocampus, in CA1 –DG (i) and CA3 (ii). The traces show an example of the LFP recording of oscillatory activity at different regions of the male control mouse hippocampus, and the heat map illustrates the color-coded voltage plots at identical time-scale after 350 Hz low pass filtering. B Oscillatory power in the theta (4–12 Hz) frequency range for channels located in the CA1 stratum moleculare (CA1 mol), CA1 stratum pyramidale (CA1 pyr) and dentate gyrus (DG), −/− for male and female control and Gad-Grik1 mice (n = 5 / group). C Oscillatory power in the gamma (20–90 Hz) frequency range, for the same recordings as in B. * p < 0.05, 2-way ANOVA, Tukey’s correction for multiple comparisons). D Oscillatory power in the gamma range, as function of the theta phase angle divided into 8 equal sized bins. The solid lines represent second order polynomial (quadratic) curve fit. Gamma power in the CA1 is more strongly modulated by theta phase (**** p < 0.0001, n = 5, quadratic regression, sum-of-squares F test) in male (i), but not in −/− female (ii) (n = 5, quadratic regression, sum-of-squares F test) Gad-Grik1 mice. E Rate of occurrence of ripple oscillations in the CA1 pyramidal layer. * p < 0.05, 2-way ANOVA. F Duration of ripple oscillations detected in the CA1 pyramidal layer. *p < 0.05, unpaired t-test. G Percentage of ripples lasting more than 100 ms. All data from idle or resting epochs, detected from videos recorded simultaneously with the electrophysiological recording. ** p < 0.005, unpaired t-test. Bars in all panels represent mean ± SEM O janen et al. Molecular Brain (2023) 16:43 Page 9 of 20 Fig. 5 (See legend on previous page.) Ojanen et al. Molecular Brain (2023) 16:43 Page 10 of 20 by automated monitoring home-cage activity in single- the escape box equally well as the controls. After the relo- −/− housed male control and Gad-Grik1 mice over 6 days. cation of the escape box (training trials 9–12) the male −/− −/− Gad-Grik1 mutants were less active as compared to Gad-Grik1 mice showed a steeper slope in relearning their littermate controls, particularly during the dark the correct target location (Fig. 6Ei, iii) as compared to (active) phase (full time period p = 0.0003, dark period littermate controls (p < 0.05, two-tailed unpaired t-test). p = 0.0008, light period p = 0.08; 2-way ANOVA; Fig. 6C). In females, there were no differences in the slope of Although GluK1 containing KARs, located in amyg- relearning the BM task (Fig. 6Eii, iii). dala interneurons have been implicated in anxiety-like To further assess the cognitive flexibility of the Gad- −/− behaviors in rodents (e.g. [37, 38]), only a minor features Grik1 mice, we used the flexible sequencing task of anxiety were observed in the present tests. There were in the IntelliCage [39]. In this test, female mice are no significant differences in the time spent in the center group housed in a home cage with automated tracking zone of the OF arena (6Bii), yet the number of entries and remotely controlled location of a reward. Follow- to the central zone was significantly lower in the Gad- ing habituation to the novel environment, the animal is −/− Grik1 mice as compared to controls (p = 0.0026, 2-way trained to obtain a reward by making a nose poke. In the ANOVA; Fig. 6Biii). In the EPM, the percentage of time learning test, the mouse obtains the reward by shuttling spent in the open arms was slightly but not significantly between two diagonally opposite corners of the cage, different between genotypes (p = 0.059, 2-way ANOVA; after which the pattern of reward locations is changed Fig. 6Aii), and the number of entries to the open arm for reversal learning. We found that control and Gad- −/− −/− was significantly lower in the Gad-Grik1 mice as com- Grik1 females were able to learn the task equally well. −/− pared to controls (p = 0.0285, 2-way ANOVA; Fig. 6Aiii). However, after the first two reversals, the Gad-Grik1 Interestingly, when a novel object was placed in the mid- females started showing less correct visits as compared −/− dle of the center zone in the OF test, Gad-Grik1 spent to the controls. Statistical testing indicated a significant significantly less time in the object zone as compared to genotype effect over the whole test (p = 0.0030, 2-way controls (p = 0.0057, 2-way ANOVA; Fig. 6D). This phe - ANOVA) and in reversals 3 and 4 (p = 0.0021 and notype was apparent in both male and female mice, but p = 0.0299, 2-way ANOVA), but not reversals 1 and 2, in post-hoc testing reached significance only in males individually (Fig. 6F). These data are consistent with (p = 0.0149, Holm-Šídák). alterations in cognitive flexibility, also in females. Since spontaneous network oscillations, especially theta-nested gamma oscillations are associated with Discussion spatial learning, memory functions and cognitive flex - KARs have established functions in modulation of synap- ibility [30, 34, 35], we next looked at the performance of tic transmission, plasticity and neuronal excitability, both −/− the Gad-Grik1 mice in the Barnes maze (BM) spatial during development and in the adult brain [12, 14, 40]. learning and memory task. The task consisted of 2 parts: However, how the various modulatory functions attrib- (1) training trials 1–8, where the animals learned to find uted to different types of KARs, expressed in distinct cell the target location, i.e. a hole leading to the escape box types and subcellular compartments, control the physi- and (2) training trials 9–12 which took place after the ological activity at the level of neuronal networks is less escape box had been relocated (paradigm shift). We well understood. To start to investigate this question, we ignored the first three training trials, as they are highly generated a mouse model lacking GluK1 subunit contain- affected by the environmental novelty the animals face. In ing KARs specifically in the GABAergic neurons. Elec - the training trials 4–8 we did not see differences between trophysiological characterization of these mice indicate the genotypes in either gender (Fig. 6E), indicating that a critical role for GluK1 KARs, expressed in GABAergic the mutant mice were able to learn the spatial location of interneurons, in maturation of GABAergic transmission (See figure on next page.) −/− Fig. 6 Behavioral phenotype in Gad-Grik1 shows reduced activity and altered cognitive flexibility. A EPM test. (i) total distance traveled (ii) time spent in the open arms (iii) entries to the open arms. * p < 0.05; ** p < 0.01; *** p < 0.001; Holm-Šídák posthoc test after 2-way ANOVA. B OF test. (i) total distance traveled (ii) time spent in the central zone (iii) entries to the central zone. * p < 0.05; ** p < 0.01; *** p < 0.001; Holm-Šídák posthoc test −/− after 2-way ANOVA. C Home-cage activity for single-housed male control and Gad-Grik1 mice. The activity score is plotted against time of the day, with 1 h bins. *** p < 0.001, 2-way ANOVA. D OF test, with a novel object placed in the center of the field. Example tracks traveled by a control −/− and Gad-Grik1 mouse, and pooled data for the time spent in the central zone. * p < 0.05; Holm-Šídák posthoc test after 2-way ANOVA. E Barnes maze test. Latency to the correct hole in the Barnes maze during training (trials 1–8) and after relocation of the escape box (trials 9–12) for male (i) −/− and female (ii) control and Gad-Grik1 mice. (iii) Pooled data on the slope of the learning curve between trials 9–12 (* p < 0.05, two-tailed unpaired −/− t-test). F Flexible sequencing task in the IntelliCage. Arrows denote location reversal times. Gad-Grik1 female mice show less correct visits and −/− more incorrect visits after the second reversal. ** p < 0.01, 2-way ANOVA. For all the behavioral data, controls, n = 10 and n = 21, Gad-Grik1 n = 11 −/− and n = 9, for males and females, respectively. As an exception, control, n = 15, Gad-Grik1 n = 9 for the data in panel F O janen et al. Molecular Brain (2023) 16:43 Page 11 of 20 Fig. 6 (See legend on previous page.) Ojanen et al. Molecular Brain (2023) 16:43 Page 12 of 20 and network synchronization during early postnatal Pharmacological characterization of the spontaneous −/− development. Furthermore, consistent with prolonged activity in Gad-Grik1 mice was performed to further interneuronal dysfunction, aberrant oscillatory activity understand the cell-type specific mechanisms involved. in the gamma frequency band as well as mild alterations Activation of GluK1 receptors by ATPA results in robust in behavioral tasks requiring cognitive flexibility were increase in sIPSC frequency, which has been attributed to detected in the adults. activation of somatodendritic GluK1-containing KARs in GABAergic neurons [22, 23]. This effect was completely −/− Cell‑type specific mechanisms involved in regulation abolished in the Gad-Grik1 mice, thus confirming of neonatal network activity by GluK1 the previous conclusion and also validating the mouse The physiological functions of GluK1 subunit containing model. ATPA application also results in depression of KARs have been comprehensively characterized in the glutamatergic transmission at the immature synapses [22, developing hippocampus, by using GluK1 selective phar- 27], via activation of inhibitory presynaptic GluK1 recep- macological tools [19, 22, 23, 41, 42]. These data suggest tors [27]. This effect persisted in neonatal but not juvenile −/− that activation level of GluK1 KARs in the neonatal hip- Gad-Grik1 mice, confirming that GluK1 receptors are pocampus is finely tuned to permit the typical rhythmic expressed in glutamatergic terminals in the hippocampus activity of the immature circuitry [11], yet the KAR sub- during the first postnatal week. populations involved in this regulation are not known. Blocking GluK1 KARs attenuated network bursts in −/− The frequency of spontaneous network bursts was sig - both control and Gad-Grik1 mice, indicating that −/− nificantly reduced in Gad-Grik1 mice, indicating a endogenously active GluK1 receptors in principal neu- critical role of interneuronal GluK1 receptors in regula- rons also contribute to synchronization of the neonatal tion of network synchronization. This phenotype as well network. In the adults, alterations in glutamatergic drive as the parallel decrease in spontaneous GABAergic syn- to interneurons modulate network oscillations [45–47]. −/− aptic transmission in Gad-Grik1 mice could be fully Similarly, ongoing GluK1 dependent regulation of glu- explained by lower firing rate of interneurons, due to loss tamate release to interneurons [22] may promote gen- of GluK1-dependent regulation of SK potassium channels eration of synchronous network bursts in the neonatal [42]. Decrease in the mean firing rate of GABAergic neu - hippocampus. rons is sufficient to elevate the threshold for generation Together, the present data confirm that network activ - of synchronous bursts in the immature network, shown ity in the neonatal hippocampus is regulated by distinct specifically for somatostatin-expressing interneurons subpopulations of GluK1-containing KARs, located in [43]. GluK1 coupling to SK-channels is developmentally GABAergic interneurons and glutamatergic principal downregulated and accordingly, there was no difference cells. GluK1 expressed in GABAergic neurons facilitates −/− in sIPSC frequency between Gad-Grik1 and control local synchronization of immature hippocampal net- mice during adolescence and adulthood. However, we work, presumably via regulation of interneuron excit- cannot rule out the possibility that various homeostatic ability and firing rate. In parallel, endogenous activity of compensatory mechanisms that can efficiently balance GluK1 KARs expressed in principal neurons maintains excitability of immature neuronal networks contribute glutamatergic drive to interneurons and facilitates net- to the loss of differences between the genotypes during work bursts. maturation of the circuitry. −/− MEA recordings from cultured Gad-Grik1 slices Loss of GluK1 in GABAergic interneurons associates indicated that in addition to lower occurrence of net- with small changes in hippocampal network dynamics work bursts, activity generated in the DG spread more and behavioral flexibility in adults efficiently across the hippocampus in the absence of Pharmacological studies have implicated GluK1 KARs in interneuronal GluK1. Previously, it has been shown that modulation of hippocampal theta oscillations in vivo [48] augmented inhibition in response to pharmacologi- and gamma oscillations in vitro [49–51], yet no previous cal activation of GluK1 receptors restricts propagation data on rhythmic network activity in mice lacking GluK1 of epileptiform activity from one hemisphere to other exists. Our recordings in awake mice indicated that abla- [44]. Our results extend these findings to more physi - tion of GluK1 in GABAergic interneurons increases ological context and suggest that activation of interneu- oscillatory power within the gamma frequency band ronal GluK1 by endogenous glutamate limits the size of in males. In addition, the theta-modulation of gamma −/− activated neuronal population by enhancing GABAe- oscillations was stronger in Gad-Grik1 males as com- rgic inhibition. Thus, in the immature hippocampus, pared to controls. Although interneuronal GluK1 has interneuronal GluK1 receptors promote local synchrony, been previously implicated in regulation of synaptic and but restrict spatial propagation of the activity. network activity in the theta frequency [48, 52, 53], no O janen et al. Molecular Brain (2023) 16:43 Page 13 of 20 significant differences in theta activity were detected in escape box in the Barnes maze slightly faster as com- −/− Gad-Grik1 mice. Interneuronal KARs are composed pared to controls, suggesting improved reversal learning heteromeric combinations of GluK2 and GluK1 subunits and cognitive flexibility. In contrast, females showed a [28, 29, 54], which can functionally compensate for each mild defect in cognitive flexibility in a task that requires other [54] and possibly explain the lack of some phe- the ability to relearn the correct spatial sequence sev- −/− notypes in the adult Gad-Grik1 mice. Alternatively, eral times over the course of several days. Interestingly, theta activity involves GluK1 receptors that are expressed suppression or disruption of ripple oscillations has been in other cell types than GABAergic interneurons. shown to impair spatial learning and memory [67, 68] Gamma oscillations are modulated by glutamate recep- while prolongation of ripple oscillations through optoge- tors in interneurons [46, 47, 55]. In particular, abla- netic stimulation improved performance in a spatial tion of NMDA receptors in parvalbumin (PV) positive memory task [69]. interneurons augments gamma oscillation power [46, While mice lacking GluK1 display changes in drug- 55]. GluK1 KARs modulate the excitability of PV neurons induced behavioral plasticity [70–72] and anxiety-like in the adult brain [37] providing a putative mechanism behaviors [37, 38], a role for GluK1 in reversal learning −/− for the altered gamma activity in the Gad-Grik1 male and cognitive flexibility has not been reported previ - mice. However, since previous work detected no effects ously. Interestingly, mice lacking GluK2, the KAR subunit of GluK1 antagonism on gamma oscillations in the highly expressed in GABAergic interneurons but also in hippocampus [48], it is also possible that the delayed many other cell types [54, 73], display reduced locomotor maturation of the GABAergic activity in the Gad- activity, faster spatial learning and impaired spatial rever- −/− Grik1 mice contributes to the phenotype, for example sal learning [74] phenotypes similar to those observed in −/− by causing long-lasting structural changes in the circuitry Gad-Grik1 . Since GluK1 and GluK2 form functional underlying gamma frequency synchronization. Interest- heteromeric receptors in GABAergic interneurons [28, ingly, no differences between genotypes were detected in 29, 54], it is possible that the shared behavioral pheno- −/− −/− the females. Instead, females had a distinctive phenotype types in GluK2 mice and Gad-Grik1 depend on in sharp wave ripple oscillations, which were shorter in interneuronal KARs. −/− length in Gad-Grik1 females. The hippocampal SWRs In the present study, we found only mild increase in are driven by strong excitatory drive from the CA3, com- anxiety-like behaviors in the standard open field and bined with synchronous local activation of interneurons EPM tests, suggesting that GluK1 expression in GABAe- in the CA1 region [56, 57]. Therefore, both the impaired rgic neurons is only partially responsible for this pheno- −/− gamma and high-frequency ripple oscillations could type. Interestingly, however, Gad-Grik1 mice showed be explained by lower excitability of GluK1 deficient strong avoidance of a novel object placed in the center interneurons, which reduces synchronization in the neo- of the open field. The novel object induces an approach- nate as well as in the adult circuitry [43, 47]. Why Gad- avoidance conflict, where avoidance is interpreted as an −/− Grik1 genotype results in different phenotype in terms indication of anxiety [e.g. [75]]. On the other hand, the of oscillatory activity in males and females is currently robust avoidance may also reflect fear of novelty, rather not clear. Interestingly, we could observe, slight, but not than lack of incentive motivation to explore novel stimuli. significant gender differences in control animals with ele - Both interpretations are consistent with previous find - vated gamma power in females, as previously described ings suggesting a role for interneuronal KARs in fear and [58, 59], and shorter SWRs in males. To our knowledge, anxiety—like behaviors [37, 38, 72]. even though SWRs are associated with features known to have gender specificity (as e.g. social memory [60], hypo - thalamic circuits [61], glucose metabolism [62]) there are Materials and methods no reports on SWRs gender differences in rodents. In Animals humans, however, sleep spindles (associated with SWRs Experiments were performed using the following mouse tm1c/tm1c tm1d/tm1d tm1a [63]) are shorter, less frequent and have lower power in lines: Grik1 and Gad2-Grik1 . Grik1 boys [64]. (KOMP)Mbp mice in C57BL/6N background were Theta and gamma oscillations as well as the phase obtained from KOMP repository (UC Davis) and crossed relationships between them are considered critical for with CAG-Flp transgenic line to produce a floxed con - tm1c/tm1c memory encoding [65, 66]. Despite the changes observed ditional knock-out mice (Grik1 ) [20]. The Grik - tmc1/tm1c tm2(cre)Zjh in gamma power and theta-gamma cross frequency cou- 1 mice were crossed with Gad2 /J mice −/− pling, Gad-Grik1 mice performed equally well as their (expressing Cre under the Gad2 promoter) to obtain the tm1d/tm1d −/− littermate controls in a spatial learning task. However, Gad2-Grik1 line (referred to as Gad-Grik1 ). −/− Gad-Grik1 males relearned the new location of the Homozygous animals were used for all the in vitro Ojanen et al. Molecular Brain (2023) 16:43 Page 14 of 20 tm1c/tm1c experiments. Grik1 mice heterozygous for the Preparation of organotypic hippocampal cultures on tm1c/tm1c Gad2-Cre, with littermate Grik1 controls were MEA probes. Before use, MEA probes (MED-P515A, used for the behavioral and in vivo experiments. All the Alpha MED Scientific, 8 × 8 electrodes; electrode size experiments were done in accordance with the University 50 × 50 μm; array size 1 × 1 mm; spacing 150 μm) were of Helsinki Animal Welfare Guidelines and approved by sterilized with 70% ethanol for one hour, washed with the Animal Experiment Board in Finland. sterile water and dried under UV-light for one hour. For adhesive coating, the probes were treated with steri- lized poly-L-lysine (Sigma-Aldrich) diluted 1:10 with In vitro electrophysiology MQ-water (1 ml dilution/probe) over night at RT. Next Preparation of acute slices. Acute sagittal sections day, before starting culturing, the probes were washed (300–400 μm) were prepared from brains of neona- 3 times with sterile MQ-water. tal (P4–P6) male or female mice, juvenile (P18–P21) or P4-P5 mice were quickly decapitated, the whole head adult (> P50) male mice using standard methods. Briefly, was briefly immersed in 70% EtOH and transferred mice were decapitated under isoflurane anesthesia, the into a laminar hood in Gey′s Balanced Salt Solution brains were extracted and immediately placed in carbo- (GBSS) (G9779, Sigma). The brain was extracted, the genated (95% O/5% CO ) ice-cold sucrose-based dis- 2 2 hemispheres were separated and placed on a stage, and section solution containing (in mM): 87 NaCl, 3 KCl, 7 covered with 1–2 ml of low melting point agarose gel. MgCl , 1.25 NaH PO , 0.5 CaCl , 50 sucrose, 25 glu- 2 2 4 2 350 µm coronal slices were cut using a tissue chop- cose, 25 NaHCO (in majority of the experiments) or in 2+ 2+ per (McIlwain), the slices were placed in cold GBSS low Ca–high Mg dissection solution, containing (in and the hippocampus was extracted from the slices. mM): 124 NaCl, 3 KCl, 1.25 NaH PO , 26 NaHCO , 15 2 4 3 The hippocampi were placed on Poly-L-lysine coated glucose, 10 MgSO , 1 CaCl (part of experiments in the 4 2 MEA probes with 280 µl preheated Neurobasal A neonatal mice). The hemispheres were separated and (Gibco) medium, supplemented with 2% B27-supple- slices were cut using a vibratome (Leica VT 1200S). Slices ment (Gibco), 2 µM L-glutamine and chloramphenicol containing the hippocampus were placed into a slice (NB-A medium). The medium was changed and the holder and incubated for 30 min in carbogenated, warm 2+ probes were put in a petri dish containing 1–2 ml sterile (34 °C) High-Mg ACSF (in mM): 124 NaCl, 3 KCl, 1.25 H 0, and placed in the humidified cell culture incuba - NaH PO , 26 N aHCO , 15 glucose, 3 M gSO *7H O, 2 2 4 3 4 2 tor (+ 37 °C, CO 5%), first 60 min without rocking and CaCl . Slices were then maintained at room temperature. then on a slowly moving plate rocker. The medium was Electrophysiological recordings from acute slices. After added or changed in the cultures according to the fol- 1–5 h of recovery, the slices were placed in a submerged lowing plan: Days in Vitro (DIV) 0: 280 µl NB-A, DIV 1: heated (32–34 °C) recording chamber and perfused add 20–60 µl NB-A, DIV 2: change ~ 70% NB-A, 3 DIV with standard ACSF, containing (in mM): 124 NaCl, 3 2: add 20–60 µl NB-A, DIV 4: change ~ 50% BrainPhys KCl, 1.25 N aH PO , 26 NaHCO , 15 glucose, 1 MgSO , 2 4 3 4 (BrainPhysTM Neuronal medium supplemented with 2 CaCl (95% O / 5% CO ) at the speed of 1–2 ml/min. 2 2 2 SM1 neuronal supplement (Stemcell Technologies), Whole-cell patch-clamp recordings were done from CA3 and 0.5 mM L-glutamine), followed by adding 20–60 µl principal neurons under visual guidance using patch elec- BrainPhys or changing ~ 70% BrainPhys on alternating trodes with resistance of 2–7 MΩ, filled by low-chloride days until 10 DIV. filling solution, containing (in mM): 135 K-gluconate, 10 Electrophysiological recordings from cultured organo- HEPES, 2 KCl, 2 Ca(OH) , 5 EGTA, 4 Mg-ATP, and 0.5 typic hippocampus slices. The spontaneous activity of Na-GTP, 280–285 mOsm (pH 7.2–7.35. Multiclamp 700B the slices was recorded at four time points, at DIV5, 7, amplifier (Molecular Devices), Digidata 1322 (Molecular 9 and 10. Before recording, on 5, 7, and 9 DIV 20–60 µl Devices) or NI USB-6341 A/D board (National Instru- fresh medium was added and 10 DIV 70% medium was ments) and WinLTP version 2.20 or pClamp 11.0 soft- changed, and the cultures were left for minimum one ware were used for data collection, with low pass filter hour in the incubator before starting the recording. (6 kHz) and sampling rate of 5 or 20 kHz. In all voltage- The slices on the probes were then imaged using Leica clamp recordings, uncompensated series resistance (Rs) MZ10F microscope (Leica Microsystems GmbH, Wet- was monitored by measuring the peak amplitude of the zlar, Germany) and Qimaging Rolera Bolt Scientific cam - fast whole-cell capacitance current in response to a 5 mV era (Teledyne QImaging, BC, Canada) to ensure healthy step. Only experiments where Rs < 30 MΩ, and with < 20% morphology and to confirm electrode positions. Slices change in Rs during the experiment, were included in that had detached from the probe or had clear, visu- analysis. The drugs were purchased from Tocris Biosci - ally detectable damage such as holes in the tissue were ence (ACET: 2728, ATPA: 1107). rejected. The probes were connected to the MED64 O janen et al. Molecular Brain (2023) 16:43 Page 15 of 20 system by the MED-C03 connector inside the recording Novartis). A grounding screw for connecting the ground incubator (+ 37 °C, CO 5%). The signal was monitored electrode was attached to the skull using dental cement for 30 min for stabilization, followed by 10–30 min of (3 M RelyX). A small circular craniotomy was drilled in recording using the MED64-amplifier (MEDA64HE1) the right hemisphere on top of the recording site (RC: and Mobius software (Alpha MED Scientific), with (− 2)–(− 2.1) mm, ML: 0.9–1.3 mm for CA1 and DG; 20 kHz sampling rate, 0.1 Hz high pass and 5 kHz low RC: (− 2.1)–(− 2.3) mm, ML: 2.3–2.6 mm for CA3) and pass filtering and 5 mV voltage range. covered with silicone adhesive (Kwik-Sil, World Precision Instruments). A bath from Kwik-Sil with ca 10 mm high In vivo electrophysiology walls was constructed on top of the head plate opening. Head plate implantation surgery. The mice were anes - The mice were let to recover in the home cage for at least thetized using ≥ 4% isoflurane. Carprofen (Rimadyl one hour prior the recording. vet, 5 mg/kg s.c.), dexamethasone (2 mg/kg s.c.) and For recording, the mice were head-fixed on top of a buprenorphine (Bupaq vet, 0.05 mg/kg s.c.) were injected rotating ball. The ground electrode was attached to the to reduce pain, inflammatory response and brain edema. grounding screw and a linear probe (Neuropixels 1.0) The mice were then placed on a 37 °C heating pad in the was placed on the surface of the brain and inserted hori- stereotaxic device. The depth of anesthesia was adjusted zontally into the desired depth (3.5 mm for CA1 and to 1.5–2% isoflurane and both eyes were covered with DG, 3.7–4.0 mm for CA3). Before insertion the probe ophthalmic ointment (Viscotears, Novartis) to protect was stained by the DiI (V22885, TermoFisher) drop- the cornea from dehydration. The hair on top of the head let, to allow post-experimental histological evaluation. was trimmed and the skin was disinfected by povidone- The bath on top of the head plate was filled with sterile iodine (Betadine). Lidocaine (0.5%, max 5 mg/kg) was filtered saline (0.9% NaCl). After 15 min stabilization injected under the scalp for local anesthesia, and scalp period, simultaneous recordings of local field potential and periosteum were removed. The surface of the skull (LFP) and multi-unit activity (MUA) were performed at was roughened using a large diameter drill bit and care- 30 kHz sampling rate (2500 Hz online down sampling fully cleaned. Tissue adhesive (3 M Vetbond) was used for the LFP signal) using the IMEC Neuropixels acqui- to seal the edges of the wound. The locations for crani - sition system and Spike GLX software. A custom made otomies were identified and marked using a drill. The Python application was used for video acquisition. After lightweight stainless-steel head plate with round 8.2 mm the recording, the craniotomy was covered with Kwik-Sil. opening (Neurotar model 5) was attached to the skull The recordings for CA1/DG and CA3 were done on con - using a small amount of super glue (Loctite precision), secutive days. Within one week from recording the mice and the sides of the head plate were covered with den- were transcardially perfused by PBS followed by 4% PFA tal cement (3 M RelyX). After operation the animals were and the brains were collected for histological analysis. allowed to recover on the heating pad, and placed in the home cage with some water-soaked soft food pellets. The Data analysis weight of the mice was monitored and carprofen (Rima- Analysis of the in vitro electrophysiological recordings dyl vet, 5 mg/kg s.c.), and buprenorphine (Bupaq vet, The frequency and amplitude spontaneous IPSCs (sIP - 0.05 mg/kg s.c.) were injected for two days after the sur- SCs), EPSCs (sEPSCs) and network bursts were analyzed gery for post-operative care. In case of weight loss, 0.1– using Minianalysis 6.0.0.3 (Synaptosoft). For the sIPSCs 0.3 ml glucose (20 mM; max: 20 ml/kg) solution in saline (outward current) and sEPSCs (inward current), the was i.p. injected in the first 3 days after operation. amplitude threshold was set two times the baseline RMS Habituation, craniotomy and electrophysiological noise level. Detected events were verified manually. Net - recording from awake, head-fixed mice. After recovery work bursts were identified on the base of slow outward from the head-plate surgery (at least 3 days), the mice current, with an amplitude and duration of at least 10 pA were habituated to being head-fixed for 5 min, 10 min, and 100 ms, respectively. For sIPCSs and sEPSCs, at least 15 min, 30 min and 1 h on consecutive days before the 5 min per condition or 200 events were analyzed, and the recordings. On the day of recording or one day before, network bursts were analyzed from the complete record- a small craniotomy was made for placing the recording ing. Cells with spontaneous event frequency less than probe. Briefly, the mice were anesthetized using ≥ 4% 0.008 Hz (0.5 event/min) were excluded. isoflurane. Carprofen (Rimadyl vet, 5 mg/kg s.c.) was For the MEA data analysis, representative channels injected, and the mice were placed on a 37 °C heat- were selected from the CA1, CA3 and DG regions based ing pad in the stereotaxic device. The depth of anes - on the images taken before data acquisition. Spikes and thesia was adjusted to 1.5–2% isoflurane and both eyes bursts (minimum number of spikes = 10, minimum dura- were covered with ophthalmic ointment (Viscotears, tion of burst = 0.1 s) from the MEA recordings were Ojanen et al. Molecular Brain (2023) 16:43 Page 16 of 20 detected using NeuroExplorer 5.205 functions Detect- the CA3 stratum pyramidale and CA3 stratum molecu- Spikes (using default parameters) and Bursts (using the lare were selected for analysis. The recorded signals in Surprise algorithm with default parameters). Dentate selected channels were divided into 3 s epochs of idle and spikes for MEA activity spread analysis were detected running based on video analysis of the mouse activity. using Spike2 (v. 9.03a Cambridge Electronic Design), For each sub region separately, complex Morlet wave- and cluster activity following the DG spikes was detected let convolution was performed using the MNE tfr_array_ using a custom made Matlab script. Before a calculation morlet-function, separately for theta (3–12 Hz) and of activity clusters traces were downsampled (5 kHz) and gamma (20–90 Hz) frequencies, and the mean power on demeaned (1 s window with 50% overlap). We defined each frequency band was extracted for all epochs. For the an activity cluster as a spatially adjacent (bwconncomp theta phase—gamma power intermodulation analysis, MATLAB function) group of channels in which local the theta phase at 7 Hz was extracted and divided into field potential is higher (for positive clusters) or lower eight equal sized phase angle bins. The average gamma (for negative clusters) than a threshold value (2 and − 2 (20–90 Hz) power corresponding to each theta phase bin standard deviation of the signal, respectively). Activity was then calculated. Ripple oscillation in the CA1 pyram- clusters were calculated in 21 time points near each spike idal layer were detected using the Kay_ripple_detector (100 ms window centered at the spike timestamp with method [36] from the ripple_detection package from 5 ms time lag). If there were several clusters at one time Eden-Kramer Lab [81]. point, only a cluster with a maximum size was used for further calculations. Behavioral tests Open field The behavioral apparatus consisted of four 50 cm × 50 cm arenas (light grey PVC, with wall height of Detection of epochs of rest and movement 40 cm) placed under camera for tracking the movement Deep Lab Cut 2.2.0.2 [76, 77] was used to track 7 points of animals by Ethovision XT15 (Noldus). The illumina - (the front paws, the mouth, and the tip, philtrum and tion was applied by indirect diffuse room light (20–25 lx). nares of the snout) in the videos recorded during the Each animal was released in one of the corners and moni- in vivo data acquisition. We used ResNet-50-based neu- tored for 10 min. The mice were then removed to the ral network [78, 79] to train the model for detection and holding cage and a 12 cm × 4 cm semi-transparent 50 ml tracking of the mouse body parts. The DLC tracking falcon tube was placed in the center of each arena. The data was passed to SimBA 1.3.12 [80] to detect epochs of animals were then released in the arena and observed for movement and rest. The SpikeGLX command line tool additional 10 min. CatGT (v. 2.4) was used to extract event times from the Elevated plus-maze (EPM) The maze consisted of two video TTL signal that was recorded to synchronize video open arms (30 × 5 cm) and two enclosed arms (30 × 5 cm, frames to the electrophysiological signals. A custom- inner diameter) connected by central platform (5 × 5 cm) made Python program was used to extract the electro- and elevated to 40 cm above the floor. The floor of each physiological data corresponding to each detected epoch. arm was light grey and the closed arms had transpar- ent (15 cm high) side- and end-walls. The illumination Analysis of oscillatory power and cross‑frequency coupling level in all arms was ~ 150 lx. The mouse was placed in The data was processed using custom-made Python the center of the maze facing one of the enclosed arms programs, with MNE (v. 0.24.0), Numpy (v. 1.19.5), rip- and observed for 5 min. The latency to the first open ple_detection (v.1.2.0) modules and readSGLX from the arm entry, number of open and closed arm entries (four SpikeGLX_Datafile_Tools package. The signals were paw criterion) and the time spent in different zones of low-pass filtered to 350 Hz and DC-components were the maze were measured. The number of faecal boli was removed by subtracting channel means. For each epoch, counted after trial. Distance travelled and time spent in a time–voltage–plot covering all channels was plotted, different areas (open, closed) was recorded with Ethovi - and epochs with artifacts or noise were manually dis- sion XT 10 tracking equipment (Noldus, Netherlands). carded based on the plotted figures. IntelliCage (IC) Mice were subcutaneously injected Channels for analysis were selected on the basis of with RFID transponders (Planet ID GmbH, Germany) post-mortem histological investigation (i.e. recording for individual identification. The IntelliCage by NewBe - sites of fluorescently-marked electrodes), LFP signal havior (TSE Systems, Germany) is an apparatus designed properties (i.e. amplitude, phase shift, CA1 ripples) and to fit inside a large cage (610 × 435 × 215 mm, Tecni- visually identified spike patterns. 10–20 channels repre - plast 2000P). The apparatus itself provides four record - senting (i) the CA1 stratum pyramidale, CA1 stratum ing chambers that fit into the corners of the housing moleculare and Dentate Gyrus (DG) regions and (ii) cage. Access into the chambers is provided via a tubular O janen et al. Molecular Brain (2023) 16:43 Page 17 of 20 antenna (50 mm outer and 30 mm inner diameter) read- escape box for 2–3 min. The bright light (500–600 lx on ing the transponder codes. The chamber contains two the platform) was used to motivate the mice to find and openings of 13 mm diameter (one on the left, one on the enter the escape box. The mice were trained to find the right) which give access to drinking bottles. These open - escape box in three (day 1–2) or two (day 3) training trails ings are crossed by photo beams recording nose-pokes of per day (inter-trial interval at least 60 min) over three the mice and the holes can be closed by motorized doors. days. The training trial ended when the mouse entered Four triangular red shelters (Tecniplast, Buguggiate, the escape box or after 3 min as cut-off time (in this case, Italy) were placed in the middle of the IntelliCage and the mouse was gently directed to the escape box). The used as sleeping quarters and as a stand to reach the food. memory test was carried out during the first trial on day The floor was covered with a thick (2–3 cm) layer of bed - 4 when the mice were monitored on the platform with- ding. The IntelliCage was controlled by a computer with out escape box for 90 s. Thereafter, reversal learning was dedicated software, executing preprogrammed experi- carried out, where the escape box was moved under the mental schedules and registering the number and dura- opposite hole and the mice received two training trials on tion of visits to the corner chambers, nose-pokes to the day 4 and 5. After the last training trial on day 5, the sec- door openings and lickings as behavioral measures for ond memory test was performed. Throughout the test - each mouse. In the beginning of the test, the mice were ing, the movement of animals was tracked by Ethovision released in the IntelliCage with all doors opened allowing XT15 (Noldus). unlimited access to the bottles (free adaptation). Adaptation to nose-poke All doors were closed at the beginning of experiment and mice were required to poke Statistical analysis into closed gates to reach drinking tubes. Only the first All data was transferred to GraphPad Prism (v. 9.3.1.471 nosepoke of the visit opened the door for 5 s (pre-defined or v. 9.4.1.681) for statistical analysis. The basal frequency −/− time). Animals had to start a new visit in order to get of sIPSCs and of sEPSCs in control and Gad-Grik1 access to water again. pyramidal CA3 cells across different age groups (neona - Adaptation to drinking sessions Doors were pro- tal, juvenile and adult) were analyzed by 2-way ANOVA. grammed to open after the first nose-poke only during To compare the event frequencies between the genotypes two 2-h periods, from 20:00 to 22:00 and from 04:00 to within each age group we used multiple Mann–Whitney 06:00. Drinking sessions were applied for increasing the test as a post-hoc, as the model residuals were not nor- motivation to visit the corners and thereby providing mally distributed. The basal frequencies of the spontane - −/− defined time windows for testing of learning. ous network bursts in control vs Gad-Grik1 neonatal Flexible sequencing task The animals were assigned two hippocampus were normally distributed and compared correct corners, which were rewarded alternately during by unpaired two-tailed t-test test. For testing of drug drinking sessions (task acquisition—after visiting a cor- effects (application vs baseline), two-tailed paired t-test rect corner, the next reward could be obtained in diago- or Wilcoxon match-pairs sign-rank test were used, for nally opposite corner, correct sequence of visits 1–3-1–3 normally or not normally distributed model residuals, etc., corners 2 and 4 were assigned as incorrect and never respectively. All the statistical tests were performed on rewarded). After 4 days (8 sessions) the sequence was raw data. For the graphical representation of the drug reversed, i.e. reward (water) was delivered in previously effects, the frequency during the drug application was incorrect corners for next 8 sessions. After first reversal normalized to that during the baseline recording. session, two more reversals were performed. The spike frequency, burst duration and frequency, −/− Home cage activity Male Control and Gad-Grik1 and activity spread clusters in the MEA data were ana- were housed individually in cages with an infrared sen- lyzed by 2-way ANOVA. To compare the activity spread sor (InfraMot; TSE-Systems) to monitor their activity during 0–20 ms and 30–70 ms after DG spike we used over 6 days (excluding the first night of adaptation) with the Mann–Whitney U-test. For the in vivo data, we used 12/12 h dark/light cycle. The mean hourly activity dur - 2-way ANOVA to compare the theta and gamma power, ing the dark/active phase of mice was averaged over the ripple frequencies and durations, and sum-of-squares 6 days. F-test to test the quadratic curve fit. We used 2-way Barnes maze (BM) The maze consists of a circular plat - ANOVA to compare the groups and Holm-Šídák’s test as form (100 cm diameter) with 20 holes (5 cm diameter) post-hoc for multiple pairwise comparisons in the behav- around the perimeter (Ugo Basile, Italy). One of the holes ioral experiments. We used two-tailed unpaired t-test for was connected with a dark chamber filled with bedding comparing the relearning slopes calculated for the Barnes material and two food pellets, the escape box. Two days Maze experiments. before the experiment, each animal was introduced to the Ojanen et al. Molecular Brain (2023) 16:43 Page 18 of 20 Consent for publication Supplementary Information Not applicable. The online version contains supplementary material available at https:// doi. org/ 10. 1186/ s13041- 023- 01035-9. Competing interests The authors declare no competing interests. Additional file 1: Figure S1. A. Basal amplitude of sEPSCs and sIPSCs in −/− CA3 pyramidal cells from acute control and Gad-Grik1 slices across different age groups (neonatal: n = 24(17) and 18(13); juvenile: n = 14(12) Received: 10 March 2023 Accepted: 7 May 2023 −/− and 12(10); adult: n = 10(8) and 10(9), for control and Gad-Grik1 respec- tively; n refers to number of cells, followed by number of animals in parenthesis. Bars represent mean ± SEM. Amplitudes were compared by 2-way ANOVA. B. Eec ff t of ATPA (1μM) on sEPSC and sIPSC amplitude in −/− CA3 pyramidal cells from acute control and Gad-Grik1 slices at different References stages of development (neonatal: n = 14(8) and 10(9); juvenile: n = 10(10) 1. Ben-Ari Y, Gaiarsa J-L, Tyzio R, Khazipov R. GABA: a pioneer transmitter −/− and 8(7); adult: n = 6(6) and 7(7), for control and Gad-Grik1 , respec- that excites immature neurons and generates primitive oscillations. tively). C. Eec ff t of ACET (200 nM) on the amplitude of sEPSCs and sIPSCs Physiol Rev. 2007;87(4):1215–84. −/− in neonatal control and Gad-Grik1 slices (n = 13(10) and 7(6), for control 2. Duan ZRS, Che A, Chu P, Modol L, Bollmann Y, Babij R, Fetcho RN, Otsuka −/− and Gad-Grik1 , respectively). Bars represent mean ± SEM. The ampli- T, Fuccillo MV, Liston C, Pisapia DJ, Cossart R, De Marco García NV. GABAe- tude of events is normalized to the baseline, and the amplitude during rgic restriction of network dynamics regulates interneuron survival in the ATPA / ACET application is compared to the baseline by paired t-test. developing cortex. Neuron. 2020;105(1):75-92.e75. 3. Lamsa K, Palva JM, Ruusuvuori E, Kaila K, Taira T. Synaptic GABA(A) activa- Additional file 2: Figure S2. A. Oscillatory power in the theta frequency tion inhibits AMPA-kainate receptor-mediated bursting in the newborn range for channels located in the CA3 stratum moleculare and CA3 stratum −/− (P0–P2) rat hippocampus. J Neurophysiol. 2000;83(1):359–66. pyramidale, for male and female control and Gad-Grik1 mice (n = 5 4. Wester JC, McBain CJ. Interneurons differentially contribute to spontane - / group). B. Oscillatory power in the gamma frequency range for chan- ous network activity in the developing hippocampus dependent on their nels located in the CA3 stratum moleculare and CA3 stratum pyramidale, −/− embryonic lineage. J Neurosci. 2016;36(9):2646–62. for male and female control and Gad-Grik1 mice (n = 5 / group). C. 5. Huupponen J, Molchanova SM, Lauri SE, Taira T. Ongoing intrinsic syn- Oscillatory power in the low theta (3-6 Hz) frequency range separately for −/− chronous activity is required for the functional maturation of CA3-CA1 epochs of resting or moving in male and female control and Gad-Grik1 glutamatergic synapses. Cereb Cortex. 2013;23(11):2754–64. mice. D. Oscillatory power in the high theta (7-13 Hz) frequency range 6. Leinekugel X, Khazipov R, Cannon R, Hirase H, Ben-Ari Y, Buzsáki G. Cor- separately for epochs of resting or moving in male and female control and −/− −/− related bursts of activity in the neonatal hippocampus in vivo. Science. Gad-Grik1 mice. High theta was elevated in the Gad-Grik1 mice were 2002;296(5575):2049–52. moving, * p = 0.0167, Holm-Šídák posthoc test after 2-way ANOVA. 7. Salmon CK, Pribiag H, Gizowski C, Farmer WT, Cameron S, Jones EV, Mahadevan V, Bourque CW, Stellwagen D, Woodin MA, Murai KK. Depo- larizing GABA transmission restrains activity-dependent glutamatergic Acknowledgements synapse formation in the developing hippocampal circuit. Front Cell The authors wish to acknowledge CSC—IT Center for Science, Finland, for Neurosci. 2020;14:36. computational resources. We acknowledge technical expertise and assistance 8. Warm D, Schroer J, Sinning A. Gabaergic interneurons in early brain by Nelli Koivisto and the Mouse Behavioural Phenotyping Facility, supported development: conducting and orchestrated by cortical network activity. by Helsinki Institute of Life Science and Biocenter Finland. Rodent in vivo Brain Front Mol Neurosci. 2022. https:// doi. org/ 10. 3389/ fnmol. 2021. 807969. Imaging and Electrophysiology unit, supported by Helsinki Institute of Life 9. Warm D, Bassetti D, Schroer J, Luhmann HJ, Sinning A. Spontaneous activ- Science and Neuroscience Center, is acknowledged for providing facilities, and ity predicts survival of developing cortical neurons. Front Cell Dev Biol. Marina Tibeykina for technical guidance and assistance. 2022. https:// doi. org/ 10. 3389/ fcell. 2022. 937761. 10. Buzsáki G. Theta rhythm of navigation: link between path integration and Author contributions landmark navigation, episodic and semantic memory. Hippocampus. SO and TK carried out all the in vitro and in vivo electrophysiological record- 2005;15(7):827–40. ings and data analysis. SO performed the experiments and data analysis on 11. Lauri SE, Taira T. Role of kainate receptors in network activity during organotypic cultures. ZK contributed to data analysis tools. VV supervised the development. In: Rodríguez-Moreno A, Sihra T, editors. Kainate receptors: behavioral testing. SEL and TT conceptualized the project, provided resources novel signaling insights. Springer-Verlag: Cham; 2011. p. 81–91. for the experimental work and supervised the project. The manuscript was 12. Lerma J, Marques JM. Kainate receptors in health and disease. Neuron. written by SO, TK and SEL. All authors read, commented and approved the 2013;80(2):292–311. final manuscript. 13. Mulle C, Crépel V. Regulation and dysregulation of neuronal circuits by KARs. Neuropharmacology. 2021;197: 108699. Funding 14. Pinheiro P, Mulle C. Kainate receptors. Cell Tissue Res. 2006;326(2):457–82. Open Access funding provided by University of Helsinki including Helsinki 15. Vesikansa A, Sakha P, Kuja-Panula J, Molchanova S, Rivera C, Huttunen University Central Hospital. This study was financially supported by the Acad- HJ, Rauvala H, Taira T, Lauri SE. Expression of GluK1c underlies the emy of Finland (grants 330710 (S.E.L.), 326045 ( T.T.)), Sigrid Juselius Foundation developmental switch in presynaptic kainate receptor function. Sci Rep. (S.E.L and T.T.), Finnish Cultural Foundation (S.E.L.) and Jane and Aatos Erkko 2012;2:310–310. Foundation ( V.V.) 16. Bahn S, Volk B, Wisden W. Kainate receptor gene expression in the devel- oping rat brain. J Neurosci. 1994;14(9):5525–47. Availability of data and materials 17. Ritter LM, Vazquez DM, Meador-Woodruff JH. Ontogeny of ionotropic The datasets and materials used and/or analyzed during the current study are glutamate receptor subunit expression in the rat hippocampus. Dev available from the corresponding author on reasonable request. Brain Res. 2002;139(2):227–36. 18. Wyeth MS, Pelkey KA, Yuan X, Vargish G, Johnston AD, Hunt S, Fang C, Declarations Abebe D, Mahadevan V, Fisahn A, Salter MW, McInnes RR, Chittajallu R, McBain CJ. Neto auxiliary subunits regulate interneuron somatodendritic Ethics approval and consent to participate and presynaptic kainate receptors to control network inhibition. Cell Rep. All the experiments were done in accordance with the University of Helsinki 2017;20(9):2156–68. Animal Welfare Guidelines and approved by the Animal Experiment Board 19. Lauri SE, Vesikansa A, Segerstråle M, Collingridge GL, Isaac JT, Taira T. Func- in Finland (License number. ESAVI-29384–2019, ESAVI-5031–2021 and tional maturation of CA1 synapses involves activity-dependent loss of KEK19-017). O janen et al. Molecular Brain (2023) 16:43 Page 19 of 20 tonic kainate receptor-mediated inhibition of glutamate release. Neuron. activation downregulates AMPA receptor surface expression and induces 2006;50(3):415–29. hippocampal LTD. iScience. 2021;24(9): 103029. 20. Ryazantseva M, Englund J, Shintyapina A, Huupponen J, Shteinikov V, 41. Clarke VRJ, Molchanova SM, Hirvonen T, Taira T, Lauri SE. Activity-depend- Pitkänen A, Partanen JM, Lauri SE. Kainate receptors regulate develop- ent upregulation of presynaptic kainate receptors at immature CA3-CA1 ment of glutamatergic synaptic circuitry in the rodent amygdala. Elife. synapses. J Neurosci. 2014;34(50):16902–16. 2020;9:e52798. https:// doi. org/ 10. 7554/ eLife. 52798. 42. Segerstråle M, Juuri J, Lanore F, Piepponen P, Lauri SE, Mulle C, Taira T. 21. Vesikansa A, Sallert M, Taira T, Lauri SE. Activation of kainate receptors High firing rate of neonatal hippocampal interneurons is caused by controls the number of functional glutamatergic synapses in the area attenuation of afterhyperpolarizing potassium currents by tonically active CA1 of rat hippocampus. J Physiol. 2007;583(1):145–57. kainate receptors. J Neurosci. 2010;30(19):6507–14. 22. Lauri SE, Segerstråle M, Vesikansa A, Maingret F, Mulle C, Collingridge GL, 43. Flossmann T, Kaas T, Rahmati V, Kiebel SJ, Witte OW, Holthoff K, Kirmse K. Isaac JT, Taira T. Endogenous activation of kainate receptors regulates Somatostatin interneurons promote neuronal synchrony in the neonatal glutamate release and network activity in the developing hippocampus. hippocampus. Cell Rep. 2019;26(12):3173-3182.e3175. J Neurosci. 2005;25(18):4473–84. 44. Khalilov I, Hirsch J, Cossart R, Ben-Ari Y. Paradoxical anti-epileptic effects 23. Maingret F, Lauri SE, Taira T, Isaac JT. Profound regulation of neonatal CA1 of a GluR5 agonist of kainate receptors. J Neurophysiol. 2002;88(1):523–7. rat hippocampal GABAergic transmission by functionally distinct kainate 45. Fuchs EC, Zivkovic AR, Cunningham MO, Middleton S, LeBeau FEN, Ban- receptor populations. J Physiol. 2005;567(Pt 1):131–42. nerman DM, Rozov A, Whittington MA, Traub RD, Rawlins JNP, Monyer 24. Orav E, Atanasova T, Shintyapina A, Kesaf S, Kokko M, Partanen J, Taira T, H. Recruitment of parvalbumin-positive interneurons determines hip- Lauri SE. NETO1 guides development of glutamatergic connectivity in pocampal function and associated behavior. Neuron. 2007;53(4):591–604. the hippocampus by regulating axonal kainate receptors. Eneuro. 2017; 46. Korotkova T, Fuchs EC, Ponomarenko A, von Engelhardt J, Monyer H. 4(3): ENEURO.0048-0017.2017. NMDA receptor ablation on parvalbumin-positive interneurons impairs 25. Rodríguez-Moreno A, Lerma J. Kainate receptor modulation of GABA hippocampal synchrony, spatial representations, and working memory. release involves a metabotropic function. Neuron. 1998;20(6):1211–8. Neuron. 2010;68(3):557–69. 26. Lourenço J, Cannich A, Carta M, Coussen F, Mulle C, Marsicano G. Synaptic 47. Mann EO, Mody I. Control of hippocampal gamma oscillation fre- activation of kainate receptors gates presynaptic CB1 signaling at GABAe- quency by tonic inhibition and excitation of interneurons. Nat Neurosci. rgic synapses. Nat Neurosci. 2010;13(2):197–204. 2010;13(2):205–12. 27. Sallert M, Malkki H, Segerstråle M, Taira T, Lauri SE. Eec ff ts of the kainate 48. Huxter JR, Zinyuk LE, Roloff EVL, Clarke VRJ, Dolman NP, More JCA, Jane receptor agonist ATPA on glutamatergic synaptic transmission and DE, Collingridge GL, Muller RU. Inhibition of kainate receptors reduces the plasticity during early postnatal development. Neuropharmacology. frequency of hippocampal theta oscillations. J Neurosci. 2007;27(9):2212. 2007;52(6):1354–65. 49. Brown JT, Teriakidis A, Randall AD. A pharmacological investigation of 28. Mulle C, Sailer A, Swanson GT, Brana C, O’Gorman S, Bettler B, Heinemann the role of GLUK5-containing receptors in kainate-driven hippocampal SF. Subunit composition of kainate receptors in hippocampal interneu- gamma band oscillations. Neuropharmacology. 2006;50(1):47–56. rons. Neuron. 2000;28(2):475–84. 50. Fisahn A, Contractor A, Traub RD, Buhl EH, Heinemann SF, McBain CJ. Dis- 29. Paternain AV, Herrera MAT, Nieto MA, Lerma J. GluR5 and GluR6 kainate tinct roles for the kainate receptor subunits GluR5 and GluR6 in kainate- receptor subunits coexist in hippocampal neurons and coassemble to induced hippocampal gamma oscillations. J Neurosci. 2004;24(43):9658. form functional receptors. J Neurosci. 2000;20(1):196. 51. Stanger HL, Alford R, Jane DE, Cunningham MO. The role of GLU K5 30. Bragin A, Jando G, Nadasdy Z, Hetke J, Wise K, Buzsaki G. Gamma (40–100 containing kainate receptors in entorhinal cortex gamma frequency Hz) oscillation in the hippocampus of the behaving rat. J Neurosci. oscillations. Neural Plast. 2008;2008: 401645. 1995;15(1):47. 52. Clarke VRJ, Collingridge GL, Lauri SE, Taira T. Synaptic kainate receptors in 31. Vanderwolf CH. Hippocampal electrical activity and voluntary movement CA1 interneurons gate the threshold of theta-frequency-induced long- in the rat. Electroencephalogr Clin Neurophysiol. 1969;26(4):407–18. term potentiation. J Neurosci. 2012;32(50):18215–26. 32. Belluscio MA, Mizuseki K, Schmidt R, Kempter R, Buzsáki G. Cross-fre- 53. Goldin M, Epsztein J, Jorquera I, Represa A, Ben-Ari Y, Crépel V, Cossart R. quency phase-phase coupling between theta and gamma oscillations in Synaptic kainate receptors tune oriens-lacunosum moleculare interneu- the hippocampus. J Neurosci. 2012;32(2):423. rons to operate at theta frequency. J Neurosci. 2007;27(36):9560. 33. Lopes-Dos-Santos V, van de Ven GM, Morley A, Trouche S, Campo-Urriza 54. Christensen JK, Paternain AV, Selak S, Ahring PK, Lerma J. A mosaic of N, Dupert D. Parsing hippocampal theta oscillations by nested spectral functional kainate receptors in hippocampal interneurons. J Neurosci. components during spatial exploration and memory-guided behavior. 2004;24(41):8986. Neuron. 2018;100(4):490-952.e497. 55. Carlén M, Meletis K, Siegle JH, Cardin JA, Futai K, Vierling-Claassen D, 34. Lisman J. The theta/gamma discrete phase code occuring during the Rühlmann C, Jones SR, Deisseroth K, Sheng M, Moore CI, Tsai LH. A critical hippocampal phase precession may be a more general brain coding role for NMDA receptors in parvalbumin interneurons for gamma rhythm scheme. Hippocampus. 2005;15(7):913–22. induction and behavior. Mol Psychiatry. 2012;17(5):537–48. 35. Lisman JE, Idiart MAP. Storage of 7 ± 2 short-term memories in oscillatory 56. Hájos N, Karlócai MR, Németh B, Ulbert I, Monyer H, Szabó G, Erdélyi F, subcycles. Science. 1995;267(5203):1512–5. Freund TF, Gulyás AI. Input-output features of anatomically identified 36. Kay K, Sosa M, Chung JE, Karlsson MP, Larkin MC, Frank LM. A hippocam- ca3 neurons during hippocampal sharp wave/ripple oscillation in vitro. J pal network for spatial coding during immobility and sleep. Nature. Neurosci. 2013;33(28):11677. 2016;531(7593):185–90. 57. Schlingloff D, Káli S, Freund TF, Hájos N, Gulyás AI. Mechanisms of sharp 37. Englund J, Haikonen J, Shteinikov V, Amarilla SP, Atanasova T, Shintyapina wave initiation and ripple generation. J Neurosci. 2014;34(34):11385. A, Ryazantseva M, Partanen J, Voikar V, Lauri SE. Downregulation of 58. Schroeder A, Nakamura JP, Hudson M, Jones NC, Du X, Sundram S, Hill RA. kainate receptors regulating GABAergic transmission in amygdala after Raloxifene recovers effects of prenatal immune activation on cognitive early life stress is associated with anxiety-like behavior in rodents. Transl task-induced gamma power. Psychoneuroendocrinology. 2019;110: Psychiatry. 2021;11(1):538. 104448. 38. Wu LJ, Ko SW, Toyoda H, Zhao MG, Xu H, Vadakkan KI, Ren M, Knifed E, 59. Vila-Merkle H, González-Martínez A, Campos-Jiménez R, Martínez-Ricós J, Shum F, Quan J, Zhang XH, Zhuo M. Increased anxiety-like behavior and Teruel-Martí V, Lloret A, Blasco-Serra A, Cervera-Ferri A. Sex differences in enhanced synaptic efficacy in the amygdala of GluR5 knockout mice. amygdalohippocampal oscillations and neuronal activation in a rodent PLoS ONE. 2007;2(1): e167. anxiety model and in response to infralimbic deep brain stimulation. 39. Endo T, Maekawa F, Võikar V, Haijima A, Uemura Y, Zhang Y, Miyazaki W, Front Behav Neurosci. 2023;17:1122163. Suyama S, Shimazaki K, Wolfer DP, Yada T, Tohyama C, Lipp H-P, Kakeyama 60. Oliva A, Fernández-Ruiz A, Leroy F, Siegelbaum SA. Hippocampal CA2 M. Automated test of behavioral flexibility in mice using a behavioral sharp-wave ripples reactivate and promote social memory. Nature. sequencing task in IntelliCage. Behav Brain Res. 2011;221(1):172–81. 2020;587(7833):264–9. 40. Nair JD, Braksator E, Yucel BP, Fletcher-Jones A, Seager R, Mellor JR, Bashir 61. Ishikawa D, Matsumoto N, Sakaguchi T, Matsuki N, Ikegaya Y. Operant ZI, Wilkinson KA, Henley JM. Sustained postsynaptic kainate receptor conditioning of synaptic and spiking activity patterns in single hip- pocampal neurons. J Neurosci. 2014;34(14):5044. Ojanen et al. Molecular Brain (2023) 16:43 Page 20 of 20 62. Tingley D, McClain K, Kaya E, Carpenter J, Buzsáki G. A metabolic function of the hippocampal sharp wave-ripple. Nature. 2021;597(7874):82–6. 63. Ngo H-V, Fell J, Staresina B. Sleep spindles mediate hippocampal-neocor- tical coupling during long-duration ripples. Elife. 2020;9: e57011. 64. Markovic A, Kaess M, Tarokh L. Gender differences in adolescent sleep neurophysiology: a high-density sleep EEG study. Sci Rep. 2020;10(1):15935. 65. Buzsáki G. Hippocampal sharp wave-ripple: a cognitive biomarker for episodic memory and planning. Hippocampus. 2015;25(10):1073–188. 66. Düzel E, Penny WD, Burgess N. Brain oscillations and memory. Curr Opin Neurobiol. 2010;20(2):143–9. 67. Ego-Stengel V, Wilson MA. Disruption of ripple-associated hippocampal activity during rest impairs spatial learning in the rat. Hippocampus. 2010;20(1):1–10. 68. Girardeau G, Benchenane K, Wiener SI, Buzsáki G, Zugaro MB. Selective suppression of hippocampal ripples impairs spatial memory. Nat Neuro- sci. 2009;12(10):1222–3. 69. Fernández-Ruiz A, Oliva A, Fermino de Oliveira E, Rocha-Almeida F, Ting- ley D, Buzsáki G. Long-duration hippocampal sharp wave ripples improve memory. Science. 2019;364(6445):1082–6. 70. Bogulavsky JJ, Gregus AM, Kim PT, Costa AC, Rajadhyaksha AM, Inturrisi CE. Deletion of the glutamate receptor 5 subunit of kainate receptors affects the development of morphine tolerance. J Pharmacol Exp Ther. 2009;328(2):579–87. 71. Gregus AM, Tropea TF, Wang Y, Hauck SCR, Costa ACS, Rajadhyaksha AM, Inturrisi CE. Deletion of the GluR5 subunit of kainate receptors affects cocaine sensitivity and preference. Neurosci Lett. 2010;468(3):186–9. 72. Ko S, Zhao M-G, Toyoda H, Qiu C-S, Zhuo M. Altered behavioral responses to noxious stimuli and fear in glutamate receptor 5 (GluR5)- or GluR6- deficient mice. J Neurosci. 2005;25(4):977. 73. Mulle C, Sailer A, Pérez-Otaño I, Dickinson-Anson H, Castillo PE, Bureau I, Maron C, Gage FH, Mann JR, Bettler B, Heinemann SF. Altered synaptic physiology and reduced susceptibility to kainate-induced seizures in GluR6-deficient mice. Nature. 1998;392(6676):601–5. 74. Micheau J, Vimeney A, Normand E, Mulle C, Riedel G. Impaired hippocam- pus-dependent spatial flexibility and sociability represent autism-like phenotypes in GluK2 mice. Hippocampus. 2014;24(9):1059–69. 75. Delini-Stula A, Hunn C. Differential effects of anxiolytics and β-receptor blocking drugs on novelty-oriented (“neophobic”) behavior in the rat. Pharmacopsychiatry. 1988;21(04):186–91. 76. Mathis A, Mamidanna P, Cury KM, Abe T, Murthy VN, Mathis MW, Bethge M. DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. Nat Neurosci. 2018;21(9):1281–9. 77. Nath T, Mathis A, Chen AC, Patel A, Bethge M, Mathis MW. Using Deep- LabCut for 3D markerless pose estimation across species and behaviors. Nat Protoc. 2019;14(7):2152–76. 78. He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 770–778, IEEE, Piscataway, NJ, USA, 2016. 79. Insafutdinov E, Pishchulin L, Andres B, Andriluka M, Schiele B. Deeper- Cut: a deeper, stronger, and faster multi-person pose estimation model. Cham: Springer International Publishing; 2016. 80. Nilsson SR, Goodwin NL, Choong JJ, Hwang S, Wright HR, Norville ZC, Tong X, Lin D, Bentzley BS, Eshel N, McLaughlin RJ, Golden SA. Simple Behavioral Analysis (SimBA)—an open source toolkit for computer clas- sification of complex social behaviors in experimental animals. bioRxiv. 2020; 2020.2004.2019.049452. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : 81. Denovellis E. Eden-Kramer-Lab/ripple_detection: v1.2.0. Zenodo. 2022. https:// doi. org/ 10. 5281/ zenodo. 70111 75 fast, convenient online submission thorough peer review by experienced researchers in your field Publisher’s Note rapid publication on acceptance Springer Nature remains neutral with regard to jurisdictional claims in pub- support for research data, including large and complex data types lished maps and institutional affiliations. • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions
Journal
Molecular Brain – Springer Journals
Published: May 20, 2023
Keywords: Glutamate receptor; Kainate receptor; GABAergic interneuron; Hippocampus; Network synchronization; Gamma oscillation; Cognitive flexibility