While the hippocampus has intrigued generations of neuroscientists for its contributions to cognitive and emotional processing, functional specialization along its longitudinal axis confers particular importance to the ventral hippocampus (vHPC) in affective regulation under normal and pathological conditions. In particular, vHPC is extensively linked to the encoding, expression, and extinction of fear memories, which mediate behavioral adaptation to environmental threats. Despite decades of research, however, many questions remain about precisely what is encoded among specific populations of vHPC neurons and what brain systems cooperate in processing this information during fear regulation. Furthermore, as insights accumulate into the function of discrete afferent projections of vHPC, an important area of focus is how vHPC circuitry might be organized to support different output patterns through synaptic integration. Here, we summarize the current understanding of these issues based on contemporary circuit-based approaches and highlight potential clues to the anatomical and functional organization of synaptic networks that may help reconceptualize vHPC as a system of interacting modules.

Introduction:Cumulative stress is a major risk factor for developing major depressive disorder (MDD), yet not everyone experiencing chronic stress develops MDD. In those who do not, it is unclear at what point, or by what mechanism, a trajectory of stable resiliency emerges. Methods:Utilizing a 10-day repeated social defeat stress model (RSDS) for MDD, we observed that a critical period between 7 and 10 daily defeats marks the phenotypical divergence of resilient from susceptible mice. Using cell-type selective electrophysiology, chemogenetics, optogenetics, fiber photometry and RNA quantification was employed to investigate the nature of stress effects on neuroadaptation in the oval nucleus of the bed nucleus of the stria terminalis (BNSTov) required to determine resilience. Results:In response to ongoing stress, corticotropin-releasing factor (CRF + , but not CRF - ) neurons of the (BNSTov) displayed a sustained increased firing rate in resilient, but not susceptible mice. This neurophysiological adaptation was self-sustaining, but only after 7 critical stress exposures, indicating that the process of developing resilience is dependent on stress history. Conclusion:Our study reveals a novel process by which individuals might persist in the face of adversity by way of stress-provoked activation, not inhibition of a key CRF limbic region that establishes a pathway to resilience.

Background:Cumulative stress is a major risk factor for developing major depressive disorder (MDD), but not everyone experiencing chronic stress develops MDD. In those who do not, it is unclear at what point or by what mechanism a trajectory of stable resiliency emerges. Methods:Utilizing a 10-day repeated social defeat stress (RSDS) model for MDD, we observed that a critical period between 7 and 10 daily defeats marks the phenotypical divergence of resilient from susceptible male mice. Cell-type selective electrophysiology, chemogenetics, optogenetics, and RNA quantification were used to investigate the nature of stress effects on neuroadaptation in the oval nucleus of the bed nucleus of the stria terminalis (BNSTov) required to determine resilience. Results:In response to ongoing stress, corticotropin-releasing factor (CRF+, but not CRF-) neurons of the BNSTov displayed a sustained increased firing rate in resilient but not susceptible mice. This neurophysiological adaptation was self-sustaining, but only after 7 critical stress exposures, indicating that the process of developing resilience is dependent on stress history. Conclusions:Our study reveals a novel process by which individuals may persist in the face of adversity by way of stress-provoked activation, not inhibition of a key CRF limbic region that establishes a pathway to resilience.

Contextual memories are critical for survival but must be extinguished when new conditions render them nonproductive. By most accounts, extinction forms a new memory that competes with the original association for control over behavior, but the underlying circuit mechanisms remain largely enigmatic. Here, we demonstrate that extinction of contextual fear conditioning recruits somatostatin interneurons (SST-INs) in the ventral hippocampus. Correspondingly, real-time activity of SST-INs correlates with transitions between immobility and movement, signaling exit from defensive freezing bouts. Optogenetic manipulation of SST-INs but not parvalbumin interneurons (PV-INs) elicits bidirectional changes in freezing that are specific to the context in which extinction was acquired. Finally, similar effects were obtained following extinction of sucrose-based appetitive conditioning, in which SST-IN inhibition triggers relapse to reward seeking. These data suggest that ventral hippocampal SST-INs play a fundamental role in extinction that is independent of affective valence and may be related to their disruption of spontaneous emotional responses.

The rodent medial prefrontal cortex (mPFC) is functionally organized across the dorsoventral axis, where dorsal and ventral subregions promote and suppress fear, respectively. As the ventral-most subregion, the dorsal peduncular cortex (DP) is hypothesized to function in fear suppression. However, this role has not been explicitly tested. Here, we demonstrate that the DP paradoxically functions as a fear-encoding brain region and plays a minimal role in fear suppression. By using multimodal analyses, we demonstrate that DP neurons exhibit fear-learning-related plasticity and acquire cue-associated activity across learning and memory retrieval and that DP neurons activated by fear memory acquisition are preferentially reactivated upon fear memory retrieval. Further, optogenetic activation and silencing of DP fear-related neural ensembles drive the promotion and suppression of freezing, respectively. Overall, our results suggest that the DP plays a role in fear memory encoding. Moreover, our findings redefine our understanding of the functional organization of the rodent mPFC.

Basolateral amygdala (BLA) is a key hub for affect in the brain,1,2,3 and dysfunction within this area contributes to a host of psychiatric disorders.4,5 BLA is extensively and reciprocally interconnected with frontal cortex,6,7,8,9 and some aspects of its function are evolutionarily conserved across rodents, anthropoid primates, and humans.10 Neuron density in BLA is substantially lower in primates compared to murine rodents,11 and frontal cortex (FC) is dramatically expanded in primates, particularly the more anterior granular and dysgranular areas.12,13,14 Yet, how these anatomical differences influence the projection patterns of single BLA neurons to frontal cortex across rodents and primates is unknown. Using a barcoded connectomic approach, we assessed the single BLA neuron connections to frontal cortex in mice and macaques. We found that BLA neurons are more likely to project to multiple distinct parts of FC in mice than in macaques. Further, while single BLA neuron projections to nucleus accumbens were similarly organized in mice and macaques, BLA-FC connections differed substantially. Notably, BLA connections to subcallosal anterior cingulate cortex (scACC) in macaques were least likely to branch to other medial frontal cortex areas compared to perigenual ACC (pgACC). This pattern of connections was reversed in the mouse homologues of these areas, infralimbic and prelimbic cortex (IL and PL), mirroring functional differences between rodents and non-human primates. Taken together, these results indicate that BLA connections to FC are not linearly scaled from mice to macaques and instead the organization of single-neuron BLA connections is distinct between these species.

Summary The rodent medial prefrontal cortex (mPFC) is a locus for both the promotion and suppression (e.g. extinction) of fear and is composed of four anatomically distinct subregions, including anterior cingulate 1 (Cg1), prelimbic (PL), infralimbic (IL), and the dorsal peduncular (DP) cortex. A vast majority of studies have focused on Cg1, PL, and IL. The Cg1 and PL have been implicated in the promotion of fear, while the IL has been linked to a role in the suppression, or extinction, of fear. Due to its anatomical location ventral to IL, the DP has been hypothesized to function as a fear-suppressing brain region however, no studies have explicitly tested its role in this function or in the regulation of memory generally. Moreover, some studies have pointed towards a dichotomous role for ventral mPFC in the dual suppression and promotion of fear, but the mechanisms underlying these opposing observations remains unclear. Here, we provide evidence that the DP paradoxically functions as a cued fear-encoding brain region and plays little to no role in fear memory extinction. By using a combination of cFos immunohistochemistry, whole-cell brain slice electrophysiology, fiber photometry, and activity-dependent neural tagging, we demonstrate that DP neurons exhibit learning-related plasticity, acquire cue-associated activity across learning and memory retrieval, and that DP neurons activated by learning are preferentially reactivated upon fear memory retrieval. Further, optogenetic activation and silencing of fear learning-related DP neural ensembles drives the promotion and suppression of freezing, respectively. Overall, these data suggest that the DP plays an unexpected role in fear memory encoding. More broadly, our results reveal new principles of organization across the dorsoventral axis of the mPFC.

With the prevalence of age-related cognitive deficits on the rise, it is essential to identify cellular and circuit alterations that contribute to age-related memory impairment. Increased intrinsic neuronal excitability after learning is important for memory consolidation, and changes to this process could underlie memory impairment in old age. Some studies find age-related deficits in hippocampal neuronal excitability that correlate with memory impairment but others do not, possibly due to selective changes only in activated neural ensembles. Thus, we tagged CA1 neurons activated during learning and recorded their intrinsic excitability 5 hours or 7 days post-training. Adult mice exhibited increased neuronal excitability 5 hours after learning, specifically in ensemble (learning-activated) CA1 neurons. As expected, ensemble excitability returned to baseline 7 days post-training. In aged mice, there was no ensemble-specific excitability increase after learning, which was associated with impaired hippocampal memory performance. These results suggest that CA1 may be susceptible to age-related impairments in post-learning ensemble excitability and underscore the need to selectively measure ensemble-specific changes in the brain.

The basolateral amygdala (BLA) projects widely across the macaque frontal cortex 1–4 , and amygdalo-frontal projections are critical for optimal emotional responding 5 and decision- making 6 . Yet, little is known about the single-neuron architecture of these projections: namely, whether single BLA neurons project to multiple parts of the frontal cortex. Here, we use MAPseq 7 to determine the projection patterns of over 3000 macaque BLA neurons. We found that one-third of BLA neurons have two or more distinct targets in parts of frontal cortex and of subcortical structures. Further, we reveal non-random structure within these branching patterns such that neurons with four targets are more frequently observed than those with two or three, indicative of widespread networks. Consequently, these multi-target single neurons form distinct networks within medial and ventral frontal cortex consistent with their known functions in regulating mood and decision-making. Additionally, we show that branching patterns of single neurons shape functional networks in the brain as assessed by fMRI-based functional connectivity. These results provide a neuroanatomical basis for the role of the BLA in coordinating brain-wide responses to valent stimuli 8 and highlight the importance of high- resolution neuroanatomical data for understanding functional networks in the brain.

Basolateral amygdala (BLA) projects widely across the macaque frontal cortex, and amygdalo-frontal projections are critical for appropriate emotional responding and decision making. While it is appreciated that single BLA neurons branch and project to multiple areas in frontal cortex, the organization and frequency of this branching has yet to be fully characterized. Here, we determined the projection patterns of more than 3,000 macaque BLA neurons. We found that one-third of BLA neurons had two or more distinct projection targets in frontal cortex and subcortical structures. The patterns of single BLA neuron projections to multiple areas were organized into repeating motifs that targeted distinct sets of areas in medial and ventral frontal cortex, indicative of separable BLA networks. Our findings begin to reveal the rich structure of single-neuron connections in the non-human primate brain, providing a neuroanatomical basis for the role of BLA in coordinating brain-wide responses to valent stimuli.

Neurons activated by learning have been ascribed the unique potential to encode memory, but the functional contribution of discrete cell types remains poorly understood. In particular, it is unclear whether learning engages specific GABAergic interneurons and, if so, whether they differ functionally from interneurons recruited by other experiences. Here, we show that fear conditioning activates a heterogeneous neuronal population in the medial prefrontal cortex (mPFC) that is largely comprised of somatostatin-expressing interneurons (SST-INs). Using intersectional genetic approaches, we demonstrate that fear-learning-activated SST-INs exhibit distinct circuit properties and are selectively reactivated to mediate cue-evoked memory expression. In contrast, an orthogonal population of SST-INs activated by morphine experience exerts opposing control over fear and supports reward-like motivational effects. These results outline an important role for discrete subsets of GABAergic cells in emotional learning and point to an unappreciated capacity for functional specialization among SST-INs.

Cumulative stress is a major risk factor for developing major depressive disorder (MDD), yet not everyone experiencing chronic stress develops MDD. In those who do not, it is unclear at what point, or by what mechanism, a trajectory of stable resiliency emerges. Utilizing a 10-day repeated social defeat stress model (RSDS) for MDD, we observed that a critical period between 7 and 10 daily defeats marks the phenotypical divergence of resilient from susceptible mice. In response to ongoing stress, corticotropin-releasing factor (CRF) neurons of the oval nucleus of the bed nucleus of the stria terminalis (BNSTov) display a sustained increased firing rate in resilient, but not susceptible mice. This neurophysiological adaptation was self-sustaining, but only after 7 critical stress exposures, indicating that the process of developing resilience is dependent on stress history. Our study reveals a novel process by which individuals might persist in the face of adversity by way of stress-provoked activation, not inhibition of a key CRF limbic region that establishes a pathway to resilience.

The paradigm of fear conditioning is largely responsible for our current understanding of how memories are encoded at the cellular level. Its most fundamental underlying mechanism is considered to be plasticity of synaptic connections between excitatory projection neurons (PNs). However, recent studies suggest that while PNs execute critical memory functions, their activity at key stages of learning and recall is extensively orchestrated by a diverse array of GABAergic interneurons (INs). Here we review the contributions of genetically-defined INs to processing of threat-related stimuli in fear conditioning, with a particular focus on how synaptic interactions within interconnected networks of INs modulates PN activity through both inhibition and disinhibition. Furthermore, we discuss accumulating evidence that GABAergic microcircuits are an important locus for synaptic plasticity during fear learning and therefore a viable substrate for long-term memory. These findings suggest that further investigation of INs could unlock unique conceptual insights into the organization and function of fear memory networks.

Neurons preferentially activated by learning have been ascribed the unique potential to encode memory. However, it remains unclear which genetically-defined cell types are recruited as part of such an ensemble, or what role discrete subpopulations play in behavior. Here we show that fear conditioning activates a heterogeneous neural ensemble in the medial prefrontal cortex (mPFC), comprised to a large degree of GABAergic interneurons immunoreactive for somatostatin (SST-INs). Using an intersectional genetic approach, we demonstrate that fear learning-activated SST-INs exhibit distinct circuit properties, are preferentially reactivated during memory retrieval, and mediate the expression of defensive freezing. We further show that a rewarding experience, morphine treatment, activates an orthogonal SST-IN population that exerts opposing control over fear. These results outline an important role for discrete GABAergic ensembles in fear memory encoding, and point to an unappreciated capacity for functional specialization among SST-INs.

Behaviors central to the procurement and consumption of food are among those most fundamental to survival, but their inappropriate expression can lead to overeating and obesity. Nevertheless, we have a poor understanding of circuits that promote feeding independent of physiological demand. Here we demonstrate that activation of basal forebrain (BF) GABAergic neurons results in consumption of food as well as non-food items in well-fed mice, and performance of fictive eating in the absence of ingestible materials. In addition, stimulation of these cells disrupts defensive threat responses and elicits reward-like motivational effects. Finally, BF GABAergic activity triggers skilled predatory attacks of live prey and prey-like objects, but not social targets. These effects were entirely recapitulated by selective stimulation of BF GABAergic projections to the periaqueductal gray (PAG). Our results outline a potent circuit mechanism for increased feeding through recruitment of distinct but synergistic behaviors, and add to growing evidence that PAG is an important integrator of feeding-related activity.

Frontal top-down cortical neurons projecting to sensory cortical regions are well-positioned to integrate long-range inputs with local circuitry in frontal cortex to implement top-down attentional control of sensory regions. How adolescence contributes to the maturation of top-down neurons and associated local/long-range input balance, and the establishment of attentional control is poorly understood. Here we combine projection-specific electrophysiological and rabies-mediated input mapping in mice to uncover adolescence as a developmental stage when frontal top-down neurons projecting from the anterior cingulate to visual cortex are highly functionally integrated into local excitatory circuitry and have heightened activity compared to adulthood. Chemogenetic suppression of top-down neuron activity selectively during adolescence, but not later periods, produces long-lasting visual attentional behavior deficits, and results in excessive loss of local excitatory inputs in adulthood. Our study reveals an adolescent sensitive period when top-down neurons integrate local circuits with long-range connectivity to produce attentional behavior.

Social isolation during the juvenile critical window is detrimental to proper functioning of the prefrontal cortex (PFC) and establishment of appropriate adult social behaviors. However, the specific circuits that undergo social experience-dependent maturation to regulate social behavior are poorly understood. We identify a specific activation pattern of parvalbumin-positive interneurons (PVIs) in dorsal-medial PFC (dmPFC) prior to an active bout, or a bout initiated by the focal mouse, but not during a passive bout when mice are explored by a stimulus mouse. Optogenetic and chemogenetic manipulation reveals that brief dmPFC-PVI activation triggers an active social approach to promote sociability. Juvenile social isolation decouples dmPFC-PVI activation from subsequent active social approach by freezing the functional maturation process of dmPFC-PVIs during the juvenile-to-adult transition. Chemogenetic activation of dmPFC-PVI activity in the adult animal mitigates juvenile isolation-induced social deficits. Therefore, social experience-dependent maturation of dmPFC-PVI is linked to long-term impacts on social behavior.

Heightened aggression is characteristic of multiple neuropsychiatric disorders and can have various negative effects on patients, their families and the public. Recent studies in humans and animals have implicated brain reward circuits in aggression and suggest that, in subsets of aggressive individuals, domination of subordinate social targets is reinforcing. In this study, we showed that, in male mice, orexin neurons in the lateral hypothalamus activated a small population of glutamic acid decarboxylase 2 (GAD2)-expressing neurons in the lateral habenula (LHb) via orexin receptor 2 (OxR2) and that activation of these GAD2 neurons promoted male-male aggression and conditioned place preference for aggression-paired contexts. Moreover, LHb GAD2 neurons were inhibitory within the LHb and dampened the activity of the LHb as a whole. These results suggest that the orexin system is important for the regulation of inter-male aggressive behavior and provide the first functional evidence of a local inhibitory circuit within the LHb.

Memory is a dynamic process that is continuously regulated by both synaptic and intrinsic neural mechanisms. While numerous studies have shown that synaptic plasticity is important in various types and phases of learning and memory, neuronal intrinsic excitability has received relatively less attention, especially regarding the dynamic nature of memory. In this review, we present evidence demonstrating the importance of intrinsic excitability in memory allocation, consolidation, and updating. We also consider the intricate interaction between intrinsic excitability and synaptic plasticity in shaping memory, supporting both memory stability and flexibility.

Article Figures and data Abstract Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Perturbed information processing in the amygdala has been implicated in developmentally originating neuropsychiatric disorders. However, little is known on the mechanisms that guide formation and refinement of intrinsic connections between amygdaloid nuclei. We demonstrate that in rodents the glutamatergic connection from basolateral to central amygdala (BLA-CeA) develops rapidly during the first 10 postnatal days, before external inputs underlying amygdala-dependent behaviors emerge. During this restricted period of synaptic development, kainate-type of ionotropic glutamate receptors (KARs) are highly expressed in the BLA and tonically activated to regulate glutamate release via a G-protein-dependent mechanism. Genetic manipulation of this endogenous KAR activity locally in the newborn LA perturbed development of glutamatergic input to CeA, identifying KARs as a physiological mechanism regulating formation of the glutamatergic circuitry in the amygdala. Introduction Kainate-type of ionotropic glutamate receptors (KARs) are expressed in different cell types in various parts of the brain. While physiological roles for KARs in modulation of glutamatergic and GABAergic transmission have been described in the mature neural networks (Lerma and Marques, 2013), increasing evidence suggests predominant functions for KARs in the developing circuitry. KARs modulate formation of glutamatergic contacts by regulating neurite growth (Ibarretxe et al., 2007; Joseph et al., 2011; Marques et al., 2013), filopodial dynamics (Chang and De Camilli, 2001; Tashiro et al., 2003) and synaptic differentiation (Marchal and Mulle, 2004; Lanore et al., 2012; Sakha et al., 2016). At immature hippocampal synapses, KARs control efficacy and short-term dynamics of transmission by tonically regulating glutamate release probability (Lauri et al., 2005; Lauri et al., 2006). Interestingly, many of the developmental functions of KARs are mediated by non-canonical G-protein-dependent signaling mechanisms (Lauri et al., 2006; Tashiro et al., 2003; Marques et al., 2013). In the absence of the early KAR activity, development of the glutamatergic connectivity in the hippocampus is perturbed (Vesikansa et al., 2012; Lanore et al., 2012; Orav et al., 2017). Thus, the existing data indicate KARs as key modulators of synaptic transmission and plasticity during the critical window of synapse development in the hippocampus. However, the applicability of these mechanisms to other areas of the developing brain remains largely unknown. Amygdala is part of the limbic system involved in memory, emotion and autonomic function. Basic architecture of the mammalian amygdala is present at birth, but similar to hippocampus, it undergoes structural and functional change across extended developmental period (Tottenham and Sheridan, 2009). External inputs to basolateral amygdala (BLA) develop postnatally, the thalamic inputs being detectable around P7 (Bouwmeester et al., 2002), while cortical innervation emerges gradually between P10 and P21 (Arruda-Carvalho et al., 2017). Both lateral (LA) and basal amygdala (BA) nuclei have strong glutamatergic projections to the central amygdala (CeA), where the principal cells are GABAergic (Sah et al., 2003; Pape and Pare, 2010). However, practically nothing is known on the development of the intrinsic connections in the amygdala. Here, we show that KAR subunits GluK1,2 and 5 are strongly expressed in the amygdala during early postnatal development, temporally coinciding with rapid development of functional glutamatergic synapses. In the newborn BLA, KARs are physiologically activated to regulate glutamate release probability via a G-protein-dependent mechanisms. Knockdown or overexpression of GluK1 expression locally in the newborn LA perturbed development of glutamatergic input to CeA, suggesting that endogenous KAR activity is critical for development of the LA-CeA connections. Interestingly, it was recently shown that enhanced activity of KARs in the adult amygdala, generated via GluK4 overexpression, leads to severe changes in the circuit excitability and amygdala dependent behaviors (Arora et al., 2018). Together, these data thus identify tonic KAR activity as a mechanism modulating behaviorally relevant synaptic circuitry in the amygdala. Results Kainate receptors are highly expressed in the BLA during the first postnatal week Based on the existing literature, at least subunits GluK1, GluK2 and GluK5 are expressed in the adult rat BLA (Bettler et al., 1990; Li et al., 2001) where they have been implicated in regulation of synaptic transmission and plasticity (Li et al., 2001; Ko et al., 2005; Shin et al., 2010). To characterize the developmental profile for expression of various KAR subunits, BLA was dissected from acute rat brain slices at three developmental stages (postnatal day (P)4, P14, and P50). Absolute quantitative RT-PCR analysis of Grik1-5 mRNAs, encoding for subunits GluK1-5, revealed that during early postnatal development (P4), the predominant KAR subunits in the BLA are Grik1, Grik2 and Grik5, while Grik3 and Grik4 mRNAs were detected at low levels. During development, Grik2 and Grik5 mRNA expression remained relatively high (at P14 and P50, Grik2: 67 ± 4% and 69 ± 10% of the level at P4, respectively, ANOVA F(2,21)=4.0, p=0.034; Grik5: 91 ± 8% and 163 ± 15%, ANOVA F(2,7)=4.5, p=0.055) while Grik1 and Grik4 mRNA levels were strongly downregulated already during the first two postnatal weeks (P14 and P50, Grik1: 32 ± 5% and 35 ± 9% of the level at P4, respectively, ANOVA F(2,16)=2.18, p<0.001; Grik4: 54 ± 6% and 42 ± 5%, ANOVA on ranks H(2)=19.39, p<0.001). Grik3 mRNA expression was undetectable at P14 and P50. Finally, the Neto1 and Neto2 mRNAs, encoding for KAR auxiliary subunits NETO1 and NETO2, were both detected in the neonatal BLA, Neto2 representing the predominant subtype (Figure 1A,B). Figure 1 Download asset Open asset Kainate receptors are highly expressed in the amygdala during the first postnatal week. (A) The predominant KAR subunits expressed in the newborn BLA are GluK1, GluK2 and GluK5 and the auxiliary subunit NETO2. RT-aqPCR data on Grik1-5, Neto1 and Neto2 mRNA expression in the BLA at P4 (n = 3–4 rats/group). (B) RT-qPCR analysis of Grik1-5, Neto1 and Neto2 mRNA expression in the BLA at different stages of development, expressed as a percentage of the level at P4 (n = 3–12 rats/group). Grik1 expression is strongly downregulated during early postnatal development. *p<0.05, **p<0.001 as compared to P4. (C) Western blot from neonatal (P4 and P14) BLA, using antibodies against GluK2/3 and GluK5. Beta-actin was used as a loading control. GluK2 and GluK5 protein was clearly detected in the BLA both at P4 and P14. Quantification of the signal indicates no significant developmental changes in the expression level. n = 3 rats/group. (D) Cell-type-specific expression pattern of KAR subunits in the amygdala during early postnatal development. Example images showing ISH staining (red) with antisense RNA probes against Grik1, Grik2 and Grik5 mRNA in the P4 mouse amygdala, in GAD67GFP positive (green) and negative (blue DAPI stain) cells. Scale bar, 100 µm. (E) Pooled data on the integrated density of the ISH signal, normalized to the number of cell bodies (identified by DAPI staining) within the analyzed region at P4 and P14. Grik1 mRNA expression is strongly downregulated in all amygdala subnuclei during first weeks of postnatal development. LA, Lateral amygdala; BA Basal Amygdala; CeA Central amygdala (data for each subunit is pooled from n = 3–4 mice and three rats). *p<0.05, **p<0.001. Figure 1—source data 1 Raw data for results shown in the Figure 1. https://cdn.elifesciences.org/articles/52798/elife-52798-fig1-data1-v2.xlsx Download elife-52798-fig1-data1-v2.xlsx To verify KAR expression at protein level, western blot was done from crude protein extracts from neonatal BLA using antibodies for GluK2/3 and GluK5. These antibodies have been previously validated using knockout tissue as a negative control (Ruiz et al., 2005; Wyeth et al., 2014). Unfortunately, selective specific antibodies for other KAR subunits are not available. Consistent with the qPCR data, GluK2/3 and GluK5 were clearly detected both at P4 and P14 BLA, at approximately the same levels (GluK2/3: ANOVA F(1,4)=5.48, p=0.079; GluK5: ANOVA F(1,4)=2.72, p=0.174; Figure 1C). KARs localize to both glutamatergic and GABAergic neurons in the neonatal amygdala In situ hybridization (ISH) using fluorescent probes was used to study the pattern of KAR subunit expression in the amygdala in more detail. ISH signal in the various nuclei of the amygdala was analyzed in rat brain sections, cut from P4 and P14. In addition, sections from GAD67-GFP knock-in (Gad1+/GFP) mice were analyzed to investigate the expression of various KAR subunits in GAD67GFP positive GABAergic neurons. In P4 Gad1+/GFP sections, GFP expression was detected in 10.8 ± 0.6%, 13.7 ± 0.9% and 99 ± 0.4% of DAPI stained cells in LA, BA and CeA, respectively. At P4, mRNAs encoding subunits GluK1, GluK2 and GluK5 were detected in all the amygdala nuclei analyzed (LA, BA and CeA) (Figure 1D,E). Grik1 mRNA was observed in both GAD67GFP positive (GABAergic) and negative (likely representing glutamatergic principal neurons) cells in the BLA (25 ± 5% of GAD67GFP+ neurons at P4; Figure 1D). In the CeA, Grik1 mRNA was detected in 30 ± 2% of the GAD67GFP-positive neurons. Consistent with the qPCR data, Grik1 mRNA was significantly downregulated in the amygdala nuclei during the first postnatal weeks (ISH signal at P14 as percentage of P4, LA: 17 ± 3%, ANOVA on ranks, H(1)=9.80, p<0.001; BA: 38 ± 6%, ANOVA F(1,12) = 19.93, p<0.001; CeA: 28 ± 4%, ANOVA F(1,12) = 60.87, p<0.001; Figure 1E). Grik2 mRNA was predominantly expressed in the GAD67GFP negative (glutamatergic) cells, and only detected in a minority of GABAergic neurons within the BLA (18 ± 3% of GAD67GFP+ neurons at P4; Figure 1D). In the CeA, Grik2 mRNA was detected in 42% of GAD67GFP+ neurons (Figure 1D). Grik2 mRNA expression was slightly downregulated during development, although this effect reached statistical significance only in the LA (ISH signal at P14 as percentage of P4, LA: 71 ± 10%, ANOVA, F(1,10)=14.7, p=0.003; BA: 66 ± 37%, ANOVA on ranks, H(1)=2.08, p=0.18; CeA: 31 ± 8%, ANOVA on ranks, H(1)=2.08, p=0.18; Figure 1E). Grik5 mRNA was detected in both GAD67GFP positive and negative cells (30 ± 2% of GAD67GFP+ neurons at P4) in the BLA. Grik5 mRNA was also strongly expressed in 43 ± 6% GAD67GFP+ neurons in the CeA at P4 (Figure 1D). At P14, the same overall pattern of staining remained, except for slight but not significant downregulation of Grik5 expression level in the CeA (ISH signal at P14 as percentage of P4, LA: 103 ± 15%, BA: 107 ± 12%; CeA: 77 ± 9%, ANOVA F(1,10)=2.06, p=0.18; Figure 1E). Early postnatal development of glutamatergic synaptic connectivity in the amygdala The high developmental expression of KAR subunits in the amygdala is similar to that in the hippocampus, where KARs have specific developmentally restricted functions related to maturation of the glutamatergic synaptic connectivity. However, the exact time course for development of glutamatergic synaptic inputs to various nuclei in the amygdala is not known. Therefore, we recorded miniature excitatory postsynaptic currents (mEPSCs) from LA, BA and CeA neurons to get a general view on the functional glutamatergic inputs at different stages of postnatal development. The recorded cells were filled with biocytin for post-hoc analysis of the spine density. Interestingly, the mean mEPSC frequency, reflecting the density of functional glutamatergic inputs to the recorded neuron, increased with a distinct developmental time course at different nuclei. In the LA and BA, the mEPSC frequency increased significantly from P4-7 to P9-11 (LA: 479 ± 88%, ANOVA on ranks, H(1)=10.42, p=0.001; BA: 493 ± 95%, ANOVA on ranks, H(1)=10.38, p=0.001), and from P9-11 to P14-16 (LA: 240 ± 50%, ANOVA F(1,15)=6.69, p=0.021; BA: 196 ± 35%, ANOVA F(1,16)=4.60, p=0.048), but reached a plateau at the second postnatal week (Figure 2A,B). In contrast, in the CeA the mEPSC frequency increased rapidly between P4-7 and P9-11 (353 ± 62%, ANOVA on ranks, H(1)=7.78, p=0.005), after which no significant changes in the mean frequency of the events were detected within the analyzed time window (Figure 2A,B). The mEPSC amplitudes remained stable throughout the early postnatal development (not shown). Figure 2 Download asset Open asset Development of glutamatergic synaptic connectivity in the amygdala. (A) Example traces of action-potential independent glutamatergic events (mEPSCs), recorded from LA, BA and CeA neurons at different stages of development. (B) Pooled data of mEPSC frequency in amygdala subnuclei at different stages of development (n = 7–11 from 4 to 9 rats/group). The mEPSC frequency increases with a different developmental time course in CeA as compared to BA and LA. *p<0.05, **p<0.001. (C) Confocal images illustrating dendritic spines in biocytin filled neurons from LA, BA and CeA at P5, P10, P15 and P21 (D) Pooled data on spine density in LA, BA and CeA neurons at different stages of development. Data is obtained from 6 to 10 dendrites from three rats/group. *p<0.05; **p<0.001. Figure 2—source data 1 Raw data for results shown in the Figure 2. https://cdn.elifesciences.org/articles/52798/elife-52798-fig2-data1-v2.xlsx Download elife-52798-fig2-data1-v2.xlsx Spine maturation takes place later than formation of a functional synapse. Visualization of dendritic spines of the biocytin filled neurons revealed that the mean spine density stabilized earlier in the CeA as compared to BLA (spine density, % increase P14-16 vs P20-21, LA: 149 ± 13%, ANOVA F(1,14)=5.30, p=0.037; BA: 166 ± 25%, ANOVA F(1,12)=5.19, p=0.042; CeA: 94 ± 7%, ANOVA F(1,14)=0.149, p=0.71; Figure 2C,D). This is consistent with the idea that the glutamatergic synapses to CeA form rapidly during the first two postnatal weeks in rats, while the inputs to BLA develop with an extended time course. Ionotropic KAR function is not detected in the BLA during the first postnatal week The developmental peak in KAR, and particularly GluK1 expression was associated with the time of rapid development of glutamatergic synapses in the BLA and CeA. To understand the physiological roles of KARs in the developing circuit, we mapped functional ionotropic KARs in the neonatal amygdala using agonist application during voltage clamp recordings. Surprisingly, application of kainate (KA, 2 µM) in the presence of selective antagonists for AMPA, NMDA and GABA receptors (2 µM NBQX, 50 µM APV and 100 µM PiTX, respectively) induced no or only a very small current (LA: 6 ± 1 pA; BA: 8 ± 1 pA) in the BLA neurons during the first postnatal week (P4-P7), while a marked inward current was observed later on in development (P14-P16) (LA: 81 ± 19 pA, BA 37 ± 5 pA) (Figure 3A,B). In the CeA, a small KA-induced inward current was detected already at P4-P6 (15 ± 2 pA), but the current was significantly larger at P14-16 (27 ± 3 pA, ANOVA on ranks, H(1)=8.00, p=0.005)(Figure 3A,B). Since kainate might cause rapid desensitization of the receptors, we also used another KAR agonist, domoate to confirm the result in the first postnatal week (P4-P7). Application of domoate (500 nM) induced no current in LA neurons (4 ± 1 pA), but in BA a small inward current was detected (17 ± 1 pA; Figure 3C,D). Thus, the high expression of the KAR subunits in the newborn LA does not correlate with ionotropic function of KARs, which emerges after the second postnatal week. Small KAR-dependent currents were detected in BA and in particular, in the CeA already at P4-P6, but in contrast to the declining mRNA expression profile, the ionotropic function increased towards the second postnatal week. Figure 3 Download asset Open asset Functional map of ionotropic KAR activity in the amygdala during the first postnatal week. (A) Example traces illustrating inward currents in response to application of 2 µM kainate (KA) in the presence of blockers for AMPA, NMDA and GABA-A receptors, in LA, BA and CeA neurons at P4-6 and P14-16. 33 mM KCl was applied as a positive control. In the neonatal BLA, no KA-induced currents are detected while a clear current is induced under similar conditions at P14-P16. (B) Pooled data on the effect of 2 µM kainate on holding current in LA, BA and CeA neurons at two different stages of development (P4-6 and P14-16; LA n = 10 [6] and 13[7]; BA: n = 14 [7] and 12 [6], CeA n = 12[10] and 12[8], respectively). (C) Example traces showing that application of 500 nM of domoate (DA) induced no current in neonate LA neurons, while a small inward current was detected in BA (P4-6). The recordings were made in the presence of blockers for AMPA, NMDA and GABA-A receptors. (D) Pooled data comparing the effect of 2 µM kainate (LA n = 10[4], BA n = 14[4]) and 500 nM domoate (LA: n = 10[4]; BA: n = 9[4]) on holding current in neonatal LA and BA neurons. The n numbers in this and following figure legends refer to the number of experiments, followed by the number of animals in brackets. Figure 3—source data 1 Raw data for results shown in the Figure 3. https://cdn.elifesciences.org/articles/52798/elife-52798-fig3-data1-v2.xlsx Download elife-52798-fig3-data1-v2.xlsx KARs regulate glutamatergic transmission to BA and CeA during the first postnatal week At immature CA3-CA1 synapses, presynaptic GluK1 subunit containing KARs tonically regulate glutamate release probability in a G-protein and PKC-dependent manner (Lauri et al., 2006; Sallert et al., 2007). To test whether KARs might have similar functions in the newborn amygdala, we studied the effect of 200 nM ACET, a selective antagonist for GluK1 subunit containing KARs (Dargan et al., 2009), on mEPSCs in LA, BA and CeA during first (P4-7) and second postnatal weeks (P10-16). In LA, ACET had no significant effect on mEPSCs at P4-6 (frequency 92 ± 14% of control, paired t-test t(6)=-0.80, p=0.45, amplitude 101 ± 8% of control) or at P14-P16 (frequency 91 ± 3%, paired t-test t(5)=2.56, p=0.051; amplitude 103 ± 4% of control; Figure 4A). In contrast, ACET application resulted in a small but significant increase in the mEPSC frequency in BA at P4-P6 (130 ± 4% of control, signed rank test, Z = 2.93, p=0.004), but not at P14-P16 (92 ± 7%, paired t-test, t(3)=1.11, p=0.35) (Figure 4B). ACET had no effect on mEPSC amplitude at either developmental time point (97 ± 1% and 100 ± 4% of control, respectively). The developmentally restricted effect of GluK1 KAR antagonism on mEPSC frequency in BA is similar to that previously characterized in the area CA1 of hippocampus (Lauri et al., 2006), indicative of tonic KAR-mediated inhibition of glutamate release. Figure 4 Download asset Open asset KARs regulate glutamatergic inputs to BA and CeA during the first postnatal week. (A) KAR antagonism does not affect action-potential independent glutamatergic events (mEPSCs) to LA during the first postnatal week. Example traces (P5) from the time points indicated (1-3) and time course plot of averaged data (± s.e.m.) illustrating little or no effect of ACET (200 nM) on mEPSC frequency in LA at P4-7 and at P14-16. The graphs on the right show the mEPSC frequency under control conditions (1), in the presence of ACET (2), and after wash-out of the drug (3), in individual experiments at two different stages of development (P4-7 and P14-16; n = 7 [4] and 6 [5], respectively). (B) Similar data for recordings in the BA. In newborn BA, ACET application reversibly increases mEPSC frequency (n = 11[8]). This effect is no longer observed at P14-P16 (n = 6[4]). *p<0.05. (C) Similar data for recordings in the CeA. In the first week of life (P4-6; n = 9[6]) but not at P9-11 (n = 6[4]), ACET application leads to robust depression of mEPSC frequency. **p<0.01. (D) Scatter plots illustrating the effect of ACET on mEPSC frequency in individual cells in various amygdala subnuclei in the first postnatal week. The P4-7 data from A-C is plotted as a function of the basal mEPSC frequency. Figure 4—source data 1 Raw data for results shown in the Figure 4. https://cdn.elifesciences.org/articles/52798/elife-52798-fig4-data1-v2.xlsx Download elife-52798-fig4-data1-v2.xlsx In CeA, ACET application resulted in a marked depression of mEPSC frequency in first postnatal week (P5-6) (73 ± 5%, paired t-test, t(10)=4.22, p=0.0018) without affecting their amplitude (99 ± 3%). Also this effect was developmentally restricted, as ACET had no effect of mEPSCs at P10-P11 CeA (frequency 101 ± 7%, amplitude 98 ± 4%; Figure 4C). These data show that GluK1 subunit containing KARs are physiologically activated to regulate glutamatergic transmission in the amygdala during the first postnatal week. Their effect is area and/or cell type specific, with KAR-mediated inhibition and facilitation of transmission at glutamatergic inputs to BA and CeA, respectively. Transmission at immature BLA – CeA synapses is tonically facilitated by presynaptic G-protein-coupled KARs The data on KAR expression and function at the neonatal amygdala so far show 1) high expression of GluK1, GluK2 and GluK5 in the BLA 2) low ionotropic KAR activity in the BLA and 3) KAR-dependent regulation of mEPSC frequency in the BA and CeA. One interpretation of these data is that GluK1 subunit containing KARs, expressed in the principal cells of LA (and BA), are targeted to axons to regulate transmitter release at emerging projections to BA and CeA via a mechanism that does not require ionotropic activity. To investigate this, we focused to study the pathway from BLA to CeA at P5-6, and recoded EPSCs from CeA in response to electrical stimulation of the LA. Consistent with hypothesis that presynaptic GluK1 subunit containing KARs regulate transmission at this pathway, ACET application resulted in significant reversible depression of the EPSC amplitude (62 ± 5% of control, paired t-test, t(7)=-3.586, p=0.009) that was associated with an increase in paired – pulse ratio (PPR) (150 ± 14%, paired t-test, t(8)=4.01, p=0.0039) (Figure 5A). Furthermore, EPSCs evoked by single pulse or paired pulses with 50 ms interval of BLA stimulation were completely blocked by the AMPA-receptor selective antagonist GYKI53655 (25 µM) (8 ± 2% of control, paired t-test, t(7)=-3.679, p=0.008), excluding the possibility that postsynaptic KARs significantly contributed to the eEPSC in this pathway (Figure 5B). Adult CeA neurons also receive glutamatergic input from extra-amygdalar brain regions, including sensory input from the auditory cortex and thalamus (Keifer et al., 2015). However, during the first postnatal week, we could not observe any postsynaptic responses in CeA neurons to stimulation of external and internal capsulae, the regions containing cortical and thalamic projections to amygdala (not shown). This suggests that external inputs to CeA are not present during the first postnatal week and thus do not contribute to the observed changes in mEPSC frequency in response to KAR antagonism in the neonate (P5-6) CeA. Figure 5 Download asset Open asset Transmission at immature BLA-CeA synapses is tonically facilitated by presynaptic G-protein-coupled KARs. (A) KAR antagonism leads to reversible depression of glutamatergic transmission from to BLA-CeA during first postnatal week. Example traces of EPSCs, recorded from CeA neurons at P5 in response to paired-pulse stimulation of the LA, under control conditions and in the presence of 200 nM ACET. Pooled data (± s.e.m.) showing that ACET application is associated with a reduction in the EPSC amplitude and an increase in paired pulse ratio. n = 9[5]; **p<0.001. (B) Postsynaptic KARs do not contribute to transmission at immature BLA-CeA synapses. Similar data as in A, showing that the AMPA receptor antagonist GYKI53655 fully blocks the EPSC. n = 8 (4); p<0.001. (C) Example traces from the time points indicated (1-3) and time course plot (average ± s.e.m.) depicting a transient increase mEPSC frequency in response to application of KA (500 nM). The graphs show mEPSC frequency under control conditions (1), in the presence of KA (2), and after wash-out of the drug (3) in individual cells in CeA (P4-6). n = 7 (6); *p<0.05. (D) Application of the GluK1 selective agonist ATPA leads to an increase in mEPSCs frequency, similar to that observed in response to 500 nM KA. Example traces and pooled data as in C. n = 7 (4); *p<0.05. (E) The tonic KAR activity regulating glutamatergic transmission to CeA acts via a G-protein-dependent mechanism. The effect of ACET on mEPSC frequency is fully blocked in slices that are pre-incubated in the presence pertussis toxin (PTX; n = 5[3]), but not in SHAM incubated controls (n = 5[3]). *p<0.05. Figure 5—source data 1 Raw data for results shown in the Figure 5. https://cdn.elifesciences.org/articles/52798/elife-52798-fig5-data1-v2.xlsx Download elife-52798-fig5-data1-v2.xlsx Pharmacological tools were used to further characterize the mechanism involved. Application of a broad KAR agonist KA (500 nM) and ATPA (1 µM), agonist selective for GluK1 subunit containing KARs, resulted in transient increase in mEPSC frequency in P5-P6 CeA (KA: 146 ± 27%, paired t-test, t(6)=4.24, p=0.0055; ATPA: 191 ± 21%, paired t-test t(6)=3,14, p=0.02; Figure 5C,D), with no effect on mEPSC amplitude (91 ± 8.8% for KA and 100 ± 13% for ATPA). The peak increase in mEPSC frequency in response to ATPA and KA was comparable in size or if anything, slightly larger in ATPA as compared to KA (ANOVA, F(1,12) = 4.50, p=0.055), suggesting that GluK1 subunit containing receptors can fully account for KAR regulation of transmission at this synapse without additional (GluK1 lacking) KAR populations involved. Finally, ACET had no effect on mEPSCs in P4-P6 CeA in slices that were preincubated in presence of pertussis toxin (PTX), a selective inhibitor of the Gαi/o signaling (mEPSC frequency in ACET, 108 ± 17% of control; paired t-test t(4)=1.243, p=0.282; Figure 5E). The incubation procedure itself had no effect on the regulation of transmission by KARs, as ACET application resulted in significant depression of mEPSC frequency in slices incubated under control conditions (72 ± 12%, paired t-test, t(4)=2.934, p=0.043). Together, these data indicate that GluK1-subunit containing KARs upregulate glutamate release at immature BLA-CeA synapses via a G-protein coupled signaling mechanism. Loss of GluK1 in the amygdala in a cKO mouse model impairs glutamatergic innervation and maturation of CeLA neurons Immature-type KAR activity is proposed to be critical for development and fine-tuning of the glutamatergic synapses (Vesikansa et al., 2007; Vesikansa et al., 2012; Orav et al., 2017). To investigate the significance of the early KAR activity on the development and maturation of glutamatergic transmission in the amygdala, we focused on studying the LA-CeA connections. To this end, we generated a conditional (floxed) mouse model for Grik1 (Grik1 cKO), which allowed us to inactivate GluK1 expression selectively in the neonate BLA in vivo with local injection of GFP-CRE encoding AAV virus at P2-P4. The viral transduction covered 79 ± 5% and 61 ± 8% of cells in LA and BA, respectively, and 10 ± 4% of cells in the CeA, based on green fluorescence signal in DAPI labeled cells at P21 (Figure 6A). Figure 6 with 3 supplements see all Download asset Open asset Loss of GluK1 in the amygdala in a cKO mouse model impairs glutamatergic innervation and maturation of CeLA neurons. (A) Image illustrating the GFP fluorescence in the BLA and CeA of Grik1 cKO mice (P21), after in vivo injection of GFP-CRE AAV virus at P2. The viral transduction mainly targeted LA and BA (79 ± 5% and 61 ± 8% of the DAPI labeled cell bodies (blue), respectively), but also some cells in the CeA (10 ± 4% of cells). Scale bar, 100 µm. (B) Example traces and pooled data showing that loss of GluK1 in the BLA has no effect on the mEPSCs at LA neurons but results in lower mEPSC frequency and amplitude in the CeLA as compared to controls. AAV viruses encoding for GFP-CRE or GFP (control) were injected to the BLA at P2-4, and mEPSCs were recorded from acute slices at P21. LA: GFP n = 13(5), GFP-CRE n = 12(4); CeLA: GFP n = 10(3), GFP-CRE n = 11(3); *p<0.05. (C) Post hoc morphological characterization of the biocytin filled CeLA neurons (P21). Density of dendritic spines was significantly lower in GFP-CRE slices as compared to controls (GFP-CRE, n = 11(4); GFP n = 13(4)). **p<0.01. Scale bar, 5 µm (D) Sholl analysis of the same neurons. The number of dendritic intersections and the total dendritic length is significantly lower in GFP-CRE slices as compared to controls. Pooled data (GFP-CRE, n = 11(4); GFP n = 10(4)) and confocal images (z-stack) illustrating the morphology of the dendritic tree. **p<0.01. Scale bar, 50 µm E) Local inactivation of Grik1 at a later developmental stage has no effect on mEPSCs and dendritic morphology of CeLA neurons. AAV viruses encoding for GFP-CRE or GFP (control) were injected to the BLA of the Grik1 cKO mice at P14 and ex vivo analysis was done at P35. The image panels show mEPSCs (E), dendritic spines (F), intersections and total dendritic length (G) for 7-9(3) neurons from both GFP and GFP-CRE slices. Figure 6—source data 1 Raw data for the results shown in the Figure 6. https://cdn.elifesciences.org/ar