Signaling mechanisms underlying activity-dependent integration of adult-born neurons in the mouse olfactory bulb

Abstract

Adult neurogenesis has fascinated the field of neuroscience for decades given the prospects of harnessing mechanisms that facilitate the rewiring and/or replacement of adult brain tissue. The subgranular zone of the hippocampus and the subventricular zone of the lateral ventricle are the two main areas in the brain that exhibit ongoing neurogenesis. Of these, adult-born neurons within the olfactory bulb have proven to be a powerful model for studying circuit plasticity, providing a broad and accessible avenue into neuron development, migration, and continued circuit integration within adult brain tissue. This review focuses on some of the recognized molecular and signaling mechanisms underlying activity-dependent adult-born neuron development. Notably, olfactory activity and behavioral states contribute to adult-born neuron plasticity through sensory and centrifugal inputs, in which calcium-dependent transcriptional programs, local translation, and neuropeptide signaling play important roles. This review also highlights areas of needed continued investigation to better understand the remarkable phenomenon of adult-born neuron integration.

1. Introduction: olfactory bulb circuitry, adult-neurogenesis, and activity-dependent synaptic integration

Olfaction is a critical sensory modality through which animals directly sample their surrounding environments via the detection of volatile molecular compounds (i.e., odorants) with high sensitivity. As such, odorants convey a plethora of information, ranging from the presence of food or poison, to approaching enemies, and social cues. In this way, olfactory information guides and informs behaviors and internal states to maximize fitness and survival. Rodents are ideal models for the study of olfaction since they heavily rely on olfaction to guide most of their behaviors. Additionally, rodents afford a diverse experimental toolbox that allows for genetic tractability and detailed olfactory-related behavioral analysis. Therefore, this review draws heavily from research accomplished in the mouse model, and is focused on what has been learned towards mechanisms that facilitate and guide adult-born neuron circuit integration in the olfactory system.

The process of olfaction in rodents begins in the olfactory epithelium of the nose. Here, olfactory sensory neurons (OSNs) transduce volatile chemical information into neural activity through activation of G-protein coupled olfactory receptors (Gaillard et al., 2004). OSNs send axonal projections to the main olfactory bulb (OB) of the forebrain, which serves as the first node of olfactory processing. More specifically, OSNs make connections in the glomerular layer (GL) of the OB, where they synapse with apical dendrites of mitral and tufted cells (MTCs) (Shepherd, 1972). As the primary excitatory neurons of the OB, MTCs also serve as the predominant output of the bulb, conveying information regarding the dynamic composition of the odorant landscape to higher order brain regions in the olfactory cortex, which use this olfactory information to guide behavior and modulate physiological and internal states.

Within the OB, MTC activity is tuned by local inhibitory interneurons, including periglomerular cells (PGCs), superficial short axon cells (sSACs), external plexiform layer (EPL) interneurons, and granule cells (GCs) (Shepherd, 1972) (Figure 1). GCs are axon-less GABAergic interneurons that communicate via dendro-dendritic connections, and alongside periglomerular cells represent the main cell types generated through ongoing adult neurogenesis. Although most GC somas reside within the granule cell layer (GCL), there is also a subpopulation located within the mitral cell layer (MCL) (Imamura et al., 2006; Merkle et al., 2014). GCs have a unique neuronal morphology, with apical dendrites that extend from the GCL to the EPL. The main dendritic shafts of GCs branch extensively, with their distal dendrites forming specialized dendrodendritic synapses with the lateral dendrites of MTCs (Rall et al., 1966). Through these reciprocal dendrodendritic synapses, GCs receive excitatory glutamatergic input from MTCs and, in turn, exert inhibitory GABAergic output onto the same and other interconnected MTCs. Such wiring leads to recurrent or lateral inhibition, which optimizes olfactory detection, discrimination, and contrast enhancement (Arnson & Strowbridge, 2017; Malvaut & Saghatelyan, 2016; Nunes & Kuner, 2015; Schoppa, 2006). In this way, GCs facilitate olfactory processing of OB microcircuit by tuning MTC activity. In addition to resident GCs that are generated during embryonic development, new GCs are continually born postnatally and throughout adulthood. These adult-born granule cells (abGCs) provide an additional layer of plasticity since their wiring into existing olfactory circuitry is guided by relevant sensory experience (Lledo et al., 2006; Song et al., 2016; Yoshihara et al., 2015). This phenomenon of adult-neurogenesis happens in two primary niches of the adult rodent brain: the dentate gyrus of hippocampus, and the subventricular zone (SVZ) of the olfactory system (Lepousez et al., 2015; Lledo et al., 2006), the latter being the topic of this review.

Figure 1. Olfactory bulb adult-born granule cell development, inputs, and local circuitry.

A: adult-born granule cell (abGC) maturation biomarkers throughout development. B: local and centrifugal (including feedback) inputs which synapse with mature, integrated granule cells (blue arrow: excitatory input, red arrow: inhibitory input, purple arrow: neuromodulatory input). C: Adult neurogenesis begins in the subventricular zone (SVZ). Astrocyte-like stem cells proliferate creating neural progenitors which migrate tangentially through the rostral migratory stream (RMS) and reach the OB. From here, they undergo radial migration to reach their final destination. Both local activity information (bottom-up) or centrifugal (top-down) inputs, such as the hippocampus (CA1/EC), the horizontal limb of the diagonal band of Broca (HDB), and piriform cortex (PC), molecularly guide abGCs to either survive and integrate into surrounding circuitry or to under apoptosis. D: main olfactory bulb circuit diagram, organized by layer with discrete cell types color coded (OE: olfactory epithelium; ONL: olfactory nerve layer; GL: glomerular layer; EPL: external plexiform layer; MCL: mitral cell layer; IPL: internal plexiform layer; GCL: granule cell layer; OSNs: olfactory sensory neurons; PG cells: periglomerular cells; sSA cells: superficial short axon cells). (Created with BioRender.com)

Shortly after birth, SVZ-born neurons leave the lateral ventricle and migrate tangentially through the rostral migratory stream (RMS) into the OB. Once in the bulb, they exit the RMS and undergo radial migration to reach their final location. Most adult-born neurons integrate into the GCL as adult-born granule cells (abGCs), while a minor population continue into the glomerular layer (GL) and take final residence as adult-born periglomerular cells (abPGCs). By six-weeks after birth in the lateral ventricle, around half of all abGCs undergo apoptosis, while the rest become long-term components of the OB circuitry with a steady-state turnover (Mandairon et al., 2006; Petreanu & Alvarez-Buylla, 2002; Winner et al., 2002). In the deep region of the GCL, the majority of older GCs are replaced by new abGCs over the course of 12 months (Imayoshi et al., 2008). Also, the addition of abGCs causes the gradual expansion of the adult OB volume over age (Platel et al., 2019). Understanding the mechanistic programs that govern adult-born neuron circuit integration has been a long-standing challenge, and one of the main goals in studies of adult neurogenesis.

Interestingly, survival and circuit integration of adult-born neurons is strongly affected by activity input. A common notion is that ongoing adult neurogenesis continually provides OB circuitry with more candidate neurons than the spots needed for circuit optimization. During this “internship”, these newborn candidates initially establish naive connections that alter their activity. Here, we divide the types of activity-dependent inputs that alter newborn neuron activity into two categories: bottom-up and top-down. Here we refer to bottom-up connections as related to peripheral sensory afferents that are bringing information “up” (from sensory organs) to the central nervous system; whereas we reference top-down connections as relaying integrated information from higher-order brain regions associated with processing past experience and knowledge. In the context of olfaction, bottom-up activity changes occur as a direct result of odorant sensation, while top-down activity changes may be associated with learning, or goal-related (e.g., an odorant-based task).

Based on the level and nature of such input activity, adult-born neurons are molecularly instructed to either succumb to programmed cell death, or to survive and adjust their connectivity. As a result, activity experience guides abGC survival and connectivity to optimize olfactory processing and learning (Belnoue et al., 2011; Chow et al., 2012; Lepousez et al., 2013; Magavi et al., 2005; Sakamoto et al., 2014). Revealing the molecular programs underlying this activity-dependent plasticity may provide mechanistic insight that can be clinically leveraged to deal with neurodegenerative diseases in other parts of the brain. To build a unified understanding of the work that has already been done, this review will attempts to summarize what is known about activity-dependent abGC circuit integration in the rodent OB, with emphasis on the plasticity changes following tangential abGC migration.

First, we will introduce what is known about the developmental trajectory of abGCs and the various inputs that affect abGC activity. Second, we will discuss different forms of activity manipulations that have provided insight into the activity-dependent nature of abGC survival, morphogenesis, and synaptic plasticity. We will include a discussion of how cortical feedback and centrifugal projections relay physiological and brain state information, and how this serves as a potent source of activity input. We will then focus on the influences of different signaling molecules, ranging from neurotransmitters to neuropeptides, and the related pathways that mediate downstream plasticity programs. Lastly, we will envision future directions that may lead to a greater understanding of the activity-dependent mechanisms underlying adult-born circuit integration.

2. Diverse activity inputs guide abGC development and circuit integration

Adult-born granule cell (abGC) development begins with radial glial-like neural stem cells within the adult SVZ, where the programs that guide adult neurogenesis are influenced by age, health, and activity (Conover & Todd 2017; Ma et al. 2009; Ming & Song 2011). These cells give rise to transit-amplifying cells, which then go on to generate neuroblasts (Ming & Song, 2011). As they differentiate, neuroblasts migrate tangentially through the RMS towards the OB. During this time, they express unclustered GABA_A receptors, and a mosaic of extra-synaptic glutamate receptors (e.g. AMPA, NMDA, mGluR5) capable of sensing ambient GABA and glutamate (Carleton et al., 2003; Young et al., 2011). By the first week after their birth, most immature adult-born neurons arrive at the core of the OB, where they then switch from a pattern of tangential to radial migration. (Petreanu & Alvarez-Buylla, 2002). The majority of these newborn neurons differentiate into abGCs and ultimately take residence within the granule cell layer (GCL), where they undergo rapid dendritic outgrowth around the second week of their life (Petreanu & Alvarez-Buylla, 2002). These elaborating dendrites extensively branch throughout the external plexiform layer (EPL), where they establish multiple connections with diverse cell types (Song et al., 2016). Once reaching the GCL, abGCs receive functional GABAergic and glutamatergic synapses on their soma and proximal dendrites (Kelsch et al. 2008; Panzanelli et al. 2009; Katagiri et al. 2011; Belluzzi et al. 2003), which arise from local MTCs, deep short-axon cells (dSACs), top-down projections from the olfactory cortex (including the anterior olfactory nucleus and anterior piriform cortex), and other cortical, limbic, and subcortical projections (Arenkiel et al., 2011; Balu et al., 2007; Deshpande et al., 2013; Hanson et al., 2020; Komano-Inoue et al., 2014; Lepousez et al., 2014; Padmanabhan et al., 2018). Notably, abGCs also receive various non-synaptic neuromodulatory inputs, such as acetylcholine (ACh) from the horizontal limb of the diagonal band of Broca (HDB), serotonin (5-HT) from the dorsal raphe (DR), noradrenaline (NE) from the locus coeruleus (LC) (Lepousez et al., 2015), and oxytocin from the paraventricular nucleus of the hypothalamus (PVN) (Pekarek et al., 2022). GABAergic input initially predominates abGC input at early stages. Eventually, significant glutamatergic input is made onto abGCs (Panzanelli et al., 2009). During the next two weeks of this developmental program, abGCs undergo significant morphological changes. Through spinogenesis and synaptogenesis, abGCs form dendritic spines and the distal apical dendrites to establish reciprocal dendrodendritic synapses with MTCs within the EPL. Once integrated, abGCs receive functional synaptic glutamatergic input, and consequently begin to release synaptic GABA (Bardy et al., 2010; Kelsch et al., 2008; Panzanelli et al., 2009; Whitman & Greer, 2007). In this way, and through this newly established MTC circuitry, olfactory input is capable of activating abGCs, and this activity promotes abGC survival and long-term integration into existing olfactory circuitry (Song et al., 2016).

Notably, this activity-dependent circuit integration of abGCs is not simply driven by input alone, but depends on a series of developmentally-linked processes. First, abCGs must migrate into their proper residential location within the OB, as well as undergo elaborate morphogenesis to form the appropriate dendritic arborization and spine structures. As abGCs mature, they assemble synaptic machinery to make functional synapses with MTCs and other partners. Even following functional integration of abGCs into the residential olfactory circuitry, experience constantly remodels abGC morphology and functionality through the remodeling of dendritic spines, the pruning of connections, and the strengthening or weakening of existing synapses. Therefore, in a broad sense, the complete scope of circuit integration includes stereotypical developmental events that span proliferation, migration, survival, dendritogenesis, spinogenesis, synaptogenesis, remodeling, and removal. Through numerous studies, these developmental stages have been molecularly characterized, and have been broadly demarcated by defined biomarkers (Figure 1) (Lepousez et al., 2015; Lledo et al., 2006; Ming & Song, 2011). Here we focus mainly on the connectivity-related aspects of this developmental program. Paying particular attention to how activity affects abGC synaptic integration following tangential migration, when abGCs take residence within the OB and are subject to a rich context of activity inputs.

3. Global context-dependent and cell-intrinsic activities influence abGC integration

Between 2 and 4 weeks after abGC birth, sensory input activity strongly influences their survival, morphology, and functional integration into OB circuitry. This time period in abGC development has been referred to as “the critical period”. Many events crucial for abGC circuit integration, such as distal dendritic morphogenesis, spine formation, and functional synaptogenesis, peak during this window (Mouret et al., 2008; Petreanu & Alvarez-Buylla, 2002).

The critical period was first appreciated by studies that globally manipulated olfactory circuit activity via olfactory deprivation or olfactory enrichment in conjunction with BrdU-based birthdating methods. For example, unilateral olfactory sensory deprivation via naris occlusion reduces not only the gross number of abGCs, but also abGC dendritic arborization and spine density in sensory-deprived animals (Mandairon et al., 2006; Saghatelyan et al., 2005; Yamaguchi & Mori, 2005). Moreover, sensory deprivation has been shown to induce marked changes in the number of input and output synapses on abGCs with a resultant decrease in synapse density on the distal domains of apical dendrites (Kelsch et al., 2009). Conversely, enriched odor experience increases the survival of abGCs and improves odor memory (Bovetti et al., 2009; Mandairon et al., 2006; Rochefort et al., 2002). Furthermore, olfactory learning increases spine density in a dendritic-region-restricted manner by promoting the remodeling of both excitatory and inhibitory synapses in deep dendritic domains, which are known to receive top-down input from the olfactory cortex (Lepousez et al., 2014).

The known activity influences on abGCs have both cell-intrinsic and cell-extrinsic components, both of which directly affect abGC survival and integration into OB circuitry. Accordingly, activity manipulations that have sought to better understand the mechanisms that guide abGC integration can be divided into two main categories. One strategy has been to manipulate the membrane properties of abGCs by introducing exogenous ion channels, or knockdown/knockout of endogenous ionotropic receptors. Depending on the ion channels being manipulated, abGC baseline activity (e.g. resting membrane potential or firing rates) can be either increased or decreased. Alternatively, signaling mediated by certain neurotransmitters can be post-synaptically manipulated. The other general strategy has been to manipulate external inputs that alter overall abGC activity. One such approach has been to globally change olfactory circuit activity via olfactory enrichment or olfactory deprivation. Interestingly, however, more recent approaches have included more precise strategies that genetically target given neuronal sources of abGC inputs for activation or silencing in a selective way. Such targeted manipulations include modifying abGC inputs from local interneurons, centrifugal projections from the olfactory cortex, or state-dependent neuromodulatory centers. Practically, these diverse approaches have helped elucidate the role of different inputs on activity-dependent programs that impact abGC integration. The following section will elaborate on findings using these strategies, and how long-range projections convey brain states, and modulate context-dependent abGC integration.

3.1. Centrifugal inputs convey context-dependent information and influence abGC plasticity

Olfactory processing in the OB is not limited to the local microcircuitry where MTCs, GCs, and other interneurons communicate. It is also extensively regulated by inputs from distant brain areas, including the olfactory cortex, piriform cortex, entorhinal cortex, and other neuromodulatory centers. The nature of these centrifugal inputs is diverse, consisting of excitatory, inhibitory, neuromodulatory, and peptidergic projections (Brunert & Rothermel, 2021). It is thought that such projections into the OB relay physiological and brain state information (Otazu et al., 2015; Rothermel & Wachowiak, 2014), and subsequently influence not only abGC firing and circuit dynamics (Balu et al., 2007), but also abGC integration and plasticity (Magavi et al., 2005), as well as olfactory-related behaviors (Peretto & Paredes, 2014). In this sense, activity input from outside of the OB local circuitry is considered “context-dependent” signaling, conveying aspects of behavioral state or bodily physiology, which can thus engage and impact abGC integration to optimize olfactory performance.

This is best exemplified by experiments that couple olfactory manipulations to behavioral tasks that involve learning, as abstraction and learning require top-down communication. For example, it has been shown that adult neurogenesis is necessary for olfactory perceptual learning (Belnoue et al., 2011), and that odor discrimination learning during the critical period promotes abGC survival within the deep portion of the granule cell layer (GCL) (Mouret et al., 2008). These deep regions of the OB contain a high density of centrifugal inputs from cortical, limbic, and subcortical projections (Lepousez et al., 2014). Interestingly, top-down cholinergic signaling to the OB has been shown to be critical for social behaviors in rodents, such as recognizing known conspecifics (Suyama et al., 2021). Furthermore, both olfactory discrimination learning and olfactory perceptual learning significantly increase the survival rates of newborn neurons (Alonso et al., 2006; Moreno et al., 2009), while simple exposure to olfactory stimuli doesn’t generate the same pro-survival effect (Alonso et al., 2006). Consistent with this, silencing centrifugal feedback by unilateral lesioning of the olfactory peduncle leads to decreased changes in oscillation activity in the corresponding OB during olfactory learning tasks (Kiselycznyk et al., 2006). Although prior investigations have largely focused on such changes in oscillatory responses, this may indeed result from the reduction in survival of abGCs following the silencing of centrifugal inputs. Nonetheless, collectively these findings support a model whereby learning-induced plasticity is dependent on centrifugal inputs (Kiselycznyk et al., 2006; C. Martin et al., 2004). Additionally, pyramidal cells in hippocampal CA1 and the entorhinal cortex project to the GCL of the OB, further suggesting that areas involved in learning, memory, and stress can also directly modify olfactory processing (Padmanabhan et al., 2018). Taken together, these results suggest that both learning-evoked and sensory-evoked activation of abGCs are important in promoting abGC survival, and that centrifugal inputs play an important role in modulating abGC circuit integration in the adult brain (Alonso et al., 2006).

Given that olfactory learning tasks demand olfactory cortex (OC) involvement (Fletcher & Chen, 2010), it is important to consider how inputs from the OC modulate abGC integration. Towards this, selective optogenetic stimulation of OC projections to the OB has shown that learning promotes input-specific synaptic plasticity in abGCs, indicating that top-down feedback regulates abGC plasticity in an activity-dependent manner (Lepousez et al., 2014). It has also been shown that olfactory discrimination learning increases odor responsiveness and apical dendritic spine density of young abGCs, whereas such plasticity does not occur in abGCs from passive odor exposure alone. Furthermore, neural activity recordings show elevated activity in the piriform cortex during learning. Finally, inactivation of piriform feedback blocked abGC plasticity during learning, while targeted activation of these same piriform feedback projections during passive experience induced learning-like plasticity of abGCs (Wu et al., 2020). Taken together, results from these experiments suggest that feedback activity from the olfactory cortex directly impacts abGC integration to optimize olfactory circuitry for enhanced processing of behaviorally-relevant odor information.

Though most centrifugal inputs to the OB (including those previously discussed) are excitatory, inhibitory centrifugal inputs also influence abGC activity, circuit integration, and plasticity. Notably, GABAergic projections from the horizontal limb of the diagonal band of broca (HDB) in the basal forebrain innervate the GCL of the OB (Gracia-Llanes et al., 2010; Niedworok et al., 2012). These connections synapse onto abGCs and strengthen throughout the critical period to promote the survival of abGCs (Hanson et al., 2020). However, this is not the only source of GABAergic input onto GCs. The magnocellular preoptic nucleus (MCPO) of the basal forebrain also innervates the OB (Paolini & McKenzie, 1996, 1997). Originating from a subpopulation of GABAergic neurons in the anterior olfactory cortex, such projections innervate both output and local OB neurons (including GCs) to provide feedback control of sensory processing (Mazo et al., 2022).

Given various feedback and neuromodulatory projections identified to regulate GC firing properties and OB circuit dynamics, it is promising to study their effect on the structural and functional plasticity of abGC integration.

3.2. Different modalities of activity regulate abGC integration

Olfactory manipulation and centrifugal projection activation/silencing show that global and state-dependent circuit activity plays important roles in guiding abGC integration. In contrast, targeted genetic manipulations of abGCs provide insight into how different modalities of activity contribute to abGC integration. Common cell-autonomous manipulations have included two general methods, or variations of such. One method has been to introduce exogenous ion channels into abGCs to change intrinsic cell excitability (i.e., the likelihood a cell will fire given a stimulus). The second method has employed targeted knock-down or knock-out of endogenously expressed receptors to disrupt neurotransmitter signaling.

Using the former approach, resting membrane potential and basal firing rate can be increased or decreased depending on the modifications to the channels being introduced. Dampening the excitability of adult-born progenitor cells by delivering Kir2.1 potassium channels into adult rat SVZ decreases abGC survival, whereas enhancing abGC excitability by introducing bacterial sodium channels (NaChBac) increases abGC survival (Lin et al., 2010). Similarly, knock-down of voltage-gated sodium channels and naris occlusion both resulted in decreased abGC excitability and reduced dendritic arborization and spine number (Dahlen et al., 2011). Interestingly, the dendritic region where spine reduction occurs depends on the type of inhibitory manipulation. Naris occlusion, which reduces activity input from global circuitry, decreases synapse number indiscriminately throughout all dendritic domains. However, spine reduction induced by sodium channel knock-down is mainly localized to the proximal part of the apical dendrite (Dahlen et al., 2011). Together, these results suggest that activity input modulates the number and localization of abGC connections, and the nature of this input has differential effects on abGC development.

Besides assaying abGC integration at terminal timepoints, two-photon time-lapse imaging of cultured brain slices (ex vivo) or direct visualization of intact OB layers through cranial windows (in vivo) has allowed for longitudinal tracking of adult-born neuron integration. Expression and monitoring of the genetically encoded calcium indicator GCaMP has revealed a striking increase of intrinsic calcium activity in abGC precursors when they switch from tangential to radial migration in the OB (Bugeon et al., 2021). Decreasing abGC precursor excitability affects their normal positioning and survival, whereas decreasing activity of adult-born periglomerular cell precursors does not show either of these effects (Bugeon et al., 2021). Interestingly, in contrast to the positive correlation of abGC activity with integration outcome, in vivo imaging has shown that abPGCs with low neuronal activity have a higher likelihood of survival (Su et al., 2023). These results suggest that there is a cell type-specific role that neuronal excitability plays in guiding different types of adult-born OB interneuron development and circuit integration.

Physiologically relevant stimuli are often encoded by specific temporal patterns. The advent of light-inducible ion channels such as Channelrhodopsin-2 (ChR2) with rapid kinetics has enabled emulation of the temporal patterning of endogenous neural activity. For example, photoactivation of ChR2-expressing abGCs enabled improved odor learning and odor memory (Alonso et al., 2012). Importantly, simultaneous photoactivation and odor presentation are necessary for this observed olfactory improvement, indicating a need for specific temporal activity to enhance circuit plasticity. In addition, the temporal frequency of ChR2 stimulation influences the extent of olfactory learning improvement, as well as enhancement of GABAergic inhibition onto mitral cells (Alonso et al., 2012), further supporting that timing and frequency pattern of abGC activity is important for functional olfactory circuit plasticity.

Another common approach of cell-autonomous manipulations to disrupt neurotransmitter signaling mechanisms is through knock-out of endogenous ionotropic receptors. One study used this approach to selectively knock out NMDA receptors, which have been shown to function as important mediators of long-term synaptic plasticity. In this manipulation, loss of NMDARs significantly decreased abGC survival throughout the critical period (Lin et al., 2010). Interestingly, increasing cell-intrinsic activity through targeted expression of NachBac completely rescued cell death associated with NMDA receptor deletion, suggesting that overall activity levels govern abGC survival, and that specific firing patterns and activation domains are not necessarily required for abGC survival (Lin et al., 2010). Furthermore, genetically raising intrinsic abGC excitability also rescued reduced glutamatergic input observed with sensory deprivation (Kelsch et al., 2009). Similarly, deletion of GluN2B, the subunit expressed predominantly during early stages of abGC development impaired the maturation of glutamatergic synaptic input, decreased abGC responses to novel odors, and reduced abGC survival. Interestingly, in this scenario reduced abGC survival could not be rescued by GluN2A, the subunit which normally dominates mature NMDAR function. Together, these results indicate a subunit-specific role for GluN2B-containing NMDA receptors in promoting synapse activation and the subsequent wiring of abGCs into circuits with correlated synaptic activity (Kelsch et al., 2012).

Alongside excitatory drive, GABAergic signaling also plays an important role in the structural and functional integration of abGCs. Ablation of the α2-subunit of the GABA_A receptor from abGCs dramatically delayed maturation, with reduced dendritic branching, diminished spine density, and decreased synaptic integration (Pallotto et al., 2012). Moreover, molecular studies have found that these effects depend on the formation and stabilization of the GABAergic postsynaptic structure organized by the scaffolding protein gephyrin (Deprez et al., 2015).

Taken together, these global, input-specific and cell-intrinsic manipulations show that activity from olfactory circuitry and other brain regions in the form of both glutamatergic or GABAergic signaling contributes to the plasticity, survival, and function of abGC circuit integration. Collectively, and in a broad sense, these studies have revealed the functional aspects of abGC circuit integration programs. However, to more fully explore the comprehensive set of integrated signaling mechanisms that guide the development and integration of abGCs into functional circuits, this review will further discuss the intracellular components and processes that mediate these activity-dependent changes in abGCs.

4. Interrelated signaling mechanisms contribute to abGC plasticity

4.1. Ca-dependent transcriptional programs

Diverse forms of neuronal activity induce various structural and functional plasticity changes in abGCs. Such activity-dependent changes can affect dendritic arborization, spine growth, turnover, pruning, and remodeling. In general, these activity-dependent processes work towards fine-tuning abGC connectivity for optimized odor processing and behavioral output. As might be expected, many well-known signaling pathways and genes are involved in activity-regulated programs, linking local neurite activity (i.e., where activity input actually occurs) to the nucleus (where transcriptional responses manifest). One such well-studied pathway that couples local activity to transcriptional modulation is NMDA-receptor mediated Ca²⁺-dependent signaling (Figure 2).

Figure 2. Interrelated signaling pathways contributing to adult-born granule cell plasticity.

Activated MTCs and projecting neurons release glutamate from their pre-synaptic terminals, which activates post-synaptic AMPA receptors on adult-born granule cell (abGC) dendritic spines with resultant depolarization. Upon coincident detection of membrane depolarization and glutamate binding, NMDA receptors allow the influx of Ca²⁺, a critical second messenger. From here, neural activity is transduced through intracellular pathways mediated by Ca-dependent kinase and phosphatase, including CaMK, MAPK (e.g., ERK5), and calcineurin. Additionally, secreted neuropeptides activate G protein-coupled receptors (GPCRs) to trigger signaling cascades through corresponding G proteins. For example, CRF signals through Gs to activate PKA, while oxytocin signals through Gq to activate downstream MAPK and PKC. Activity-activated transcription factors (TFs) in abGCs (e.g., CREB) integrate signaling from multiple pathways, most importantly through phosphorylation, and transduce signals into the nucleus to induce transcription of immediate early genes (IEGs), including Fos and Npas4. Together, activity-activated TFs and IEG TFs epigenetically modify and regulate the transcription of a large number of effector genes. For example, Npas4 regulates cytoskeletal remodeling by downregulating E3 protein ligase Mdm2 to free microtubule-associated protein Dcx from degradation. Synaptic activity also regulates local protein synthesis by modulating dendritic mRNA transport and translation efficiency. For example, mGluR activation can release FMRP-mediated repression of its mRNA substrates (e.g., CaMKII mRNA) to increase local translation. (Created with BioRender.com)

The NMDA receptor complex serves as a detector for coincident presynaptic activity input and postsynaptic depolarization. One well-established ligand that drives activity-dependent abGC plasticity by meeting the above two conditions is the excitatory neurotransmitter glutamate. Glutamate released by presynaptic MTCs or centrifugal projections binds to ionotropic AMPA receptors at the post-synaptic membrane of GC dendritic spines, leading to depolarization, and subsequent Ca²⁺ influx (Halabisky et al., 2000; Sassoè-Pognetto & Ottersen, 2000). Subsequent activation of local voltage-gated sodium channels (NaVs) at the dendritic spines of GCs can also boost Ca²⁺ entry into spines through voltage activated calcium channels (Bywalez et al., 2015; Lage-Rupprecht et al., 2020; Nunes & Kuner, 2018). Notably, intracellular Ca²⁺ serves as a critical second messenger capable of signaling to the nucleus via multiple downstream pathways, including Ca-dependent kinases (e.g. CaMK) and phosphatases (Calcineurin), which each induce diverse downstream transcriptional changes (Bito et al., 1996; Chawla et al., 1998; Sheng et al., 1991; Xing et al., 1996; Yap & Greenberg, 2018). Another key downstream pathway recruited via Ca²⁺ signaling is the MAPK cascade. Research has shown that odor-induced activation of MAPK promotes abGC survival in the OB (Miwa & Storm, 2005). Furthermore, activation of MAPK in cultured GCs protects them against apoptotic induction signals (Miwa & Storm, 2005). These data suggest that stimulation of MAPK triggered by certain forms of odorant exposures may contribute to the survival of abGCs in an activity-dependent manner. Finally, targeted deletion of extracellular signal-regulated kinase 5 (ERK5), a type of MAPK, has also been shown to impair abGC differentiation, migration, and survival in the OB (Li et al., 2013). While kinases and phosphatases are key upstream players that impact intracellular cascades related to plasticity, their downstream transcription factor responses are ultimately responsible for enacting such activity-dependent alterations.

The molecular players associated with Ca²⁺-mediated activity-dependent transcription pathways can be broadly divided into three interrelated categories. The first is pre-existing transcription factors (TFs) that rapidly respond to Ca-dependent signaling through post-translational modifications. These types of TFs serve as critical hubs by integrating upstream signaling from multiple enzymatic pathways (i.e., not just NMDA-receptor Ca²⁺ signaling) to generate a cumulative output. Such “activity-activated” TFs are already present in neurons, and do not need to be transcriptionally induced by activity. With Ca²⁺ influx, pre-existing TFs translocate into the nucleus and form complexes with other existing transcription cofactors. Two critical activity-activated TFs that belong to this category include CREB (McLean et al., 1999; Merz et al., 2011) and members of the MEF2 family (Chen et al., 2020; Mao & Wiedmann, 1999; Okamoto et al., 2000). Both of these examples play important roles in activity-dependent transcription in many different types of neurons. As such, binding of activity-activated TFs to their corresponding cis-regulatory elements (i.e. activity-regulated enhancers) triggers transcription of immediate early genes (IEGs), which constitute the second type of molecular players responding to activity-dependent transcriptional programs.

Unlike activity-activated TFs, IEGs are induced by de novo transcription. Importantly, many of the IEGs themselves are also TFs, which facilitate further downstream transcriptional changes. The most familiar IEG examples include Fos, Npas4, and Egr1, whose mRNA and/or protein detection can be used as an approximation of neuronal activity (Benito & Barco, 2015; Yap & Greenberg, 2018). For example, Fos heterodimerizes with its partner Jun to form the activating protein 1 (AP-1) complex (Sheng & Greenberg, 1990), which recruits chromatin remodeling factors (e.g. BAF complex) to facilitate their binding to AP-1 binding sites (Malik et al., 2014; Vierbuchen et al., 2017). Similarly, together with their respective cofactors, IEG TFs access specific enhancer elements to confer long-term gene expression changes. As such, collectively the composite responses via activity-activated TFs and IEG TFs induce the transcription of numerous downstream late-response genes (LRGs), which represent the last type of players in the above-mentioned activity-dependent transcriptional programs. LRGs encode a variety of effector proteins that function in various molecular programs, cellular processes, and within distinct neuronal domains to govern aspects of neuronal survival and synaptic plasticity regulation (Benito & Barco, 2015; Leslie & Nedivi, 2011; West & Greenberg, 2011; Yap & Greenberg, 2018). In this way, activity-activated TFs, IEGs, and LRGs perform complementary yet distinct roles to mediate the diverse array of activity-dependent plasticity programs operating in the abGCs.

cAMP response element binding protein CREB is an activity-activated transcription factor that is activated upon phosphorylation of Ser133 (denoted as phosphorylated-CREB or pCREB) in response to diverse extracellular stimuli. Canonically, this occurs upon increased intracellular levels of the second messenger cAMP and PKA signaling (Gonzalez & Montminy, 1989). However, CREB has many alternative means of activation. For example, the well-known Ca²⁺ dependent kinases CaMK and MAPK can both phosphorylate and activate CREB (Lonze & Ginty, 2002). As such, CREB itself can serve as an activity-dependent integration hub of multiple signaling pathways to drive distinct transcriptional programs underlying abGC development (Merz et al., 2011).

In the SVZ/OB, CREB regulates specific phases of adult neurogenesis. Interestingly, CREB activity parallels SVZ adult-born neuron differentiation, increasing during the late phase of tangential migration, and decreasing after dendrite elongation and spine formation (Giachino et al., 2005). In vitro, inhibition of CREB function impairs morphological differentiation of SVZ-derived neuroblasts (Giachino et al., 2005). Experimentally, olfactory sensory deprivation via OSN afferent denervation leads to downregulation of CREB phosphorylation in neuroblasts, impairing their maturation and survival, suggesting that CREB regulates differentiation and survival of newborn neurons in the OB in an activity-dependent manner (Giachino et al., 2005). Consistent with this, ablation of CREB activity in the SVZ-OB neurogenic lineage leads to reduced survival and impaired dendritic development in post-migratory abGCs, again supporting that active CREB signaling is critical to the proper development of abGCs (Herold et al., 2011). Interestingly, pCREB expression in the neonate rat OB is selectively and transiently increased immediately after odor preference conditioned training, suggesting a potential link of pCREB to olfactory memory related plasticity (McLean et al., 1999). Together, these results substantiate the role of CREB signaling for the normal development of abGCs, similar to what has been described in other areas of developing brain tissue.

Working alongside CREB, the activity-dependent transcription factor Npas4 plays an important role in the programs associated with inhibitory synapse development. It does so through transcriptional regulation of downstream effector genes that regulate the number of GABA-releasing synapses that form onto excitatory neurons. Unlike other activity-dependent transcription factors such as CREB and c-Fos, neuronal Npas4 transcription is selectively induced by Ca²⁺ influx, but not by neurotrophic factors, growth factors, or PKA (Lin et al., 2008). Interestingly, olfactory stimulation rapidly and robustly induces the transcription of Npas4 in the GCL (Bepari et al., 2012). In fact, studies have indeed shown that Npas4 is both necessary and sufficient for increased synaptogenesis in abGCs associated with sensory input. Notably, Npas4 overexpression in abGCs increases dendritic spine density, even under conditions of sensory deprivation. Loss of Npas4 upregulates the expression of the E3-ubiquitin ligase Mdm2, which then ubiquitinates the microtubule-associated protein Dcx for degradation. In turn, this reduces dendritic spine density in abGCs (Yoshihara et al., 2014). Taken together, these results suggest that Npas4-mediated transcription directly regulates abGC spine development following olfactory experience.

4.2. Local protein synthesis mediates abGC plasticity

Coupling neural activity events to changes in neural plasticity frequently requires transcriptional regulation. However, this is not always the case. Local Ca²⁺-dependent pathways can also act on pre-existing mRNA and proteins within post-synaptic compartments to elicit fast, localized effects (Holt & Schuman, 2013; Martin & Zukin, 2006) (Figure 2). Notably, such local Ca²⁺ influxes (transients) can drive local mRNA translation and cytoskeletal stabilization (Fischer et al., 2000; Schubert & Dotti, 2007; Sharma et al., 2006; Star et al., 2002). Furthermore, they can also modify, mobilize, and mediate the degradation of local synaptic proteins. For example, local Ca²⁺ can facilitate dynamic trafficking events, including the insertion and removal of AMPA receptors from the postsynaptic membrane (Chowdhury et al., 2006; Rial Verde et al., 2006; Sharma et al., 2006).

How this form of activity-regulated plasticity affects abGCs is not well understood. However, evidence suggests that local translation of dendritic mRNA has the strong potential to modulate abGC plasticity. GCs contain synaptically localized CaMKIIα mRNA, whose dendritic transport and synaptic localization is regulated by olfactory activity via NMDA receptor signaling (Néant-Fery et al., 2012). CaMKII is a calcium/calmodulin sensitive kinase localized in the postsynaptic domain associated with NMDARs (Leonard et al., 1999). It senses Ca²⁺ dynamics and phosphorylates a cast of different downstream effectors, such as AMPAR, and calcium and potassium channels (Colbran & Brown, 2004; Gardoni et al., 2001). Functionally, dendritic translation of CaMKIIα in abGCs is necessary for olfactory associative learning, where it has been shown that when localization of CaMKIIα mRNA to the dendritic compartment is disrupted, olfactory associative learning is severely impaired. (Daroles et al., 2016; Néant-Fery et al., 2012).

As might be expected, such activity-induced local protein synthesis is tightly regulated. The dendritic localization and translational efficiency of CaMKIIα mRNA is regulated through its 3’ UTR (Subramanian et al. 2011; Mayford et al. 1996). Furthermore, the fragile X mental retardation protein (FMRP) is necessary for activity-dependent increases of CaMKIIα dendritic translation underlying learning-induced abGC structural plasticity (Daroles et al., 2016). FMRP plays an important role in regulating local protein synthesis by binding to and inhibiting basal translation of particular mRNA substrates. However, in response to group 1 metabotropic glutamate receptor (mGluR1, mGluR5) activation, this “translational brake” imposed by FMRP is released, allowing local translation of dendritic mRNA (Bassell & Warren, 2008). In abGCs, it has been shown that FMRP downregulates spinogenesis to limit spine overgrowth. As such, knock-down of FMRP from abGCs leads to both increased spine density and spine length. Also, FMRP is necessary for activity-dependent dendritic remodeling of abGCs (Scotto-Lomassese et al., 2011). Underlying this, it’s thought that CaMKIIα mRNA likely represents a primary mRNA target of FMRP involved in the structural plasticity changes associated with olfactory learning (Daroles et al., 2016). However, target mRNAs of FMRP encode many other important synaptic proteins, such as PSD-95, Arc, and GluR1/2, which all impact adult-born neuron plasticity in the hippocampus, and can be regulated by mGluR stimulation at the level of mRNA transport, stability, or translation (Bassell & Warren, 2008). Taken together, these findings support the intricate interplay between NMDAR and mGluR signaling, which cooperatively regulate dendritic mRNA, and contribute to activity-dependent abGC plasticity.

4.3. Spine dynamic serves as an activity-dependent substrate for abGC plasticity

Dynamic expression and localization of structural proteins in the synapse underlies mechanisms of plasticity. Additionally, the localization and remodeling of spines themselves serve as important substrates for abGC plasticity. GCs in the olfactory bulb continue to show dynamic spine turnover throughout life. As such, coordinated synaptic structural plasticity between GC and MTC synapses serves as a major mode of functional plasticity underlying adaptability to changing sensory inputs (Sailor et al., 2016). Evidence suggests that the dynamics of the dendritic spines represent a highly malleable form of structural plasticity underlying abGC integration, facilitated via continuous filopodia formation and retraction on the distal dendrites of immature abGCs (Breton-Provencher et al., 2014). Interestingly, patterned ChR2 stimulation of mitral cells designed to mimic odor responses promotes NMDAR-dependent remodeling of filopodia during early stages of abGC circuit integration (Breton-Provencher et al., 2014). Furthermore, evidence shows that mature spines of abGCs can relocate in an activity-dependent manner through reciprocal mitral cell (MC) activity. More specifically, glutamate secreted from MCs regulates motility of spine head filopodia through AMPAR signaling, while activity-dependent release of MC-derived BDNF induces directional spine relocation (Breton-Provencher et al., 2016). OB network modeling suggests that this form of spine relocation promotes fast synchronization of MCs with functional consequences for odor information processing, which may function within the bulbar network to adapt to rapid changes in odor environment with shorter time scales than synaptogenesis (Breton-Provencher et al., 2016; Migliore et al., 2015). Considering this form of structural plasticity responds to the directionality of secreted molecules to guide local spine dynamics, it likely involves local signaling pathways downstream of AMPA, NMDA, and BDNF receptors, which in turn modulate cytoskeleton-related effectors.

5. Neuropeptidergic regulation of abGC integration

Unlike neurotransmitters which are fast acting and immediately impact neural activity, neuropeptides elicit slower, and more prolonged effects on their targets (Russo, 2017). Through complementary yet distinct means, both types of signaling molecules play cooperative and important roles towards activity-dependent abGC integration. Often, upon binding to their cognate receptors, neuropeptides elicit second messenger signaling cascades to influence transcription, translation, and post-translational modifications. Therefore, their effects are long-lasting and diverse (Russo, 2017).

Neuropeptides originate from either local sources or long-range projections. Such long-range projections frequently communicate physiologic and brain states to the bulb, since they often arise from distant nuclei that transmit organismal state such as arousal, attention, hunger, or sleep. Consequently, overall physiological state represents another prominent mechanism that contextually orchestrates olfactory sensory processing, and thus abGC development and circuit integration (Athanassi et al., 2023). In the following section, we review the roles of two different neuropeptides, corticotropin releasing factor and oxytocin, in abGC maturation and integration into OB circuitry.

5.1. Corticotropin Releasing Factor (CRF)

CRF has historically been studied as a hypothalamic neurohormone that coordinates central and peripheral stress responses (Smith & Vale, 2006). However, more recent investigations have delineated a variety of neuronal populations that use CRF as a signaling neuropeptide (Dedic et al., 2018). For example, abGCs receive CRF locally from EPL interneurons in the OB towards synapse formation (Garcia et al., 2014). abGCs dynamically express the CRF Receptor 1 (CRFR1), which is a G-coupled protein receptor using predominantly G(s) as a downstream signaling mechanism. Once CRFR1 is activated, cAMP accumulates and activates downstream effector cascades. One critical downstream effector is phosphorylated CREB (pCREB), which is a transcription factor that alters the transcriptional rate of downstream genes (Lonze & Ginty, 2002).

Importantly, endogenous upregulation of CRFR1 expression coincides with the critical period of abGC development, strongly suggesting a role of CRF in directing abGC activity-dependent maturation (Garcia et al., 2014). Loss of function experiments have revealed that CRF signaling is critical for abGC survival. In contrast, gain of CRFR1 function in abGCs increases dendritic arborization, spine density, synaptic machinery, and functional circuit integration. These gain of function studies also revealed that CRF signaling decreases both AMPA receptor expression and miniature excitatory post-synaptic current (mEPSCs) amplitudes (Garcia et al., 2014). Lastly, optogenetic activation of CRF+ interneurons promotes both pCREB and synaptic protein expression in the OB. Taken together, these results strongly support a model whereby in vivo activity-dependent CRF release from EPL neurons promotes abGC synaptogenesis to form de novo synapses (Garcia et al., 2014).

The diverse effects of CRF signaling indicate that pCREB is only one of many transcription factors that mediate abGC maturation. Other work has revealed POU6f1 as another transcription factor upregulated by CRF in abGCs, where loss of POU6f1 in CRFR1+ neurons leads to decreased dendritic complexity and synaptic connectivity (McClard et al., 2018). In contrast, overexpression of POU6f1 in CRFR1+ neurons promotes dendritic outgrowth and branching, and alters synaptic function. Together, these findings suggest that POU6f1-mediated transcriptional programs act downstream of local CRF signaling in abGCs to influence activity-dependent circuit integration (McClard & Arenkiel, 2018).

The types of stimuli here that lead to CRF release in vivo go beyond the typical hypothalamic-pituitary axis response. Therefore, understanding the ethological role of CRF signaling in olfactory circuits remains open. However, based on what is known about the wiring of olfactory circuits, it is possible to speculate on their naturalistic function. For example, EPL CRF interneurons are reciprocally connected to the MCs. In other words, CRF neurons are stimulated by MC glutamatergic signaling, but they promote the wiring of abGCs to inhibit MCs. Therefore, activity-dependent CRF modulation of abGC may exert local plasticity onto the odor-specific microcircuit dynamics.

5.2. Oxytocin (OT)

Another neuropeptide that has recently garnered attention for its role in abGC development is oxytocin, which has historically been studied as a hormone controlling lactation and parturition (Uvnäs-Moberg, 2023). However, it has been increasingly appreciated that oxytocinergic neurons project to a diverse set of brain areas (Liao et al., 2020; Sharma et al., 2019; Warfvinge et al., 2020). Interestingly, developing abGCs dynamically express oxytocin receptors (Oxtr) throughout the activity-dependent critical period and respond to oxytocinergic input (Pekarek et al., 2022). The origin of oxytocinergic input is not definitively known, but current evidence suggests the paraventricular hypothalamic nucleus as a prominent source. This is because paraventricular oxytocinergic fibers innervate the RMS, and released oxytocin may diffuse throughout the migratory pathway and into the OB (Pekarek et al., 2022).

Knockout (KO) of oxytocin receptor (Oxtr) from abGCs causes morphologic abnormalities, characterized by neurite overgrowth and higher dendritic spine density. These changes are accompanied by a pronounced reduction in functional synaptic input. Loss of oxytocin signaling leads to disrupted post-synaptic and synaptogenic machinery as measured by single-cell transcriptomic analysis and translating-RNA profiling (Pekarek et al., 2022). Furthermore, whole cell electrophysiological recordings show decreased AMPA/NMDA current ratios in Oxtr KO abGCs. Taken together, these loss-of-function studies demonstrate that oxytocinergic signaling regulates abGC morphogenesis and promotes maturation of MTC inputs onto abGCs (Pekarek et al., 2022).

5.3. Other Neuropeptides

A number of other neuropeptides are present in the OB from both local and distal sources which warrant further study of their function in abGC plasticity and integration. For example, vasopressin is a neuropeptide that structurally resembles oxytocin and has strong cross-reactivity with oxytocin receptors (Baribeau & Anagnostou, 2015; Z. Song & Albers, 2018). It is mostly secreted by hypothalamic nuclei, and it has typically been described for its roles in social behavior (Churchland & Winkielman, 2012). However, vasopressin is also expressed by a large population of interneurons in the rat OB, where its action has been shown to be important for social recognition (Tobin et al., 2010). In recordings from acute brain slices, it has been shown that activation of the vasopressin receptor on granule cells leads to dose-dependent long-term potentiation (Namba et al., 2016).

Interestingly, it has also been shown that the OB is innervated by a subpopulation of orexin-A neurons from the lateral hypothalamus, and olfactory sensory processing is directly modulated by orexin A (Qi et al., 2023). For example, in vivo application (both local and intracerebroventricular) of orexin significantly increases mitral cell firing (Apelbaum et al., 2005). Considering the reciprocal connectivity between mitral cells and granule cells, abGC activity and plasticity are likely to be influenced by orexinergic projections.

Somatostatin (SST) is also present within the bulb and is expressed by local interneurons residing within the inner part of EPL. Interestingly, SST has been shown to affect gamma oscillations, as recorded through local field potentials. Briefly, olfactory gamma oscillations represent the electric interplay between the MTC / GC reciprocal dendrodendritic synapses (Rojas-Líbano & Kay, 2008). More specifically, in vivo intra-bulbar injection of SST inhibitors significantly decreases gamma oscillations (Lepousez et al., 2010). In contrast, similar applications of SST agonists significantly increased gamma oscillations. Lastly, application of SST inhibitors reduces olfactory performance, while SST agonists improve olfactory performance and discrimination (Lepousez et al., 2010; Lepousez et al., 2010). Together, it is likely that such physiological and behavioral changes occurred as a consequence of alterations in the reciprocal connections that MTCs have with granule cells.

Of course there are several other molecular mediators of activity-dependent abGC plasticity, whose underlying mechanisms are worthy of further investigation. For example, extracellular matrix glycoprotein tenascin-R (TNR) is expressed by pre-existing GCs in an activity-dependent way, and regulates the radial migration and spine development of abGCs (David et al., 2013; Saghatelyan et al., 2004). Trophoblast glycoprotein 5T4 is expressed by a subtype of GCs located within the MCL and superficial GCL (Imamura et al., 2006), whose expression is sensory-input-dependent and contributes to activity-dependent dendritic development of abGCs (Yoshihara et al., 2012). However, these are just a handful of potential players and pathways that warrant further investigation.

Discussion and future directions

A central question underlying SVZ-OB adult-neurogenesis is how activity input shapes the cellular and synaptic connectivity of integrating abGCs. Namely, what molecular programs impact the survival of abGCs to guide the morphological and functional development of abGCs. Ultimately, these programs that govern the ongoing integration of new granule cells into the mature olfactory bulb circuitry define the functionality and integrity of olfactory processing.

How does OB microcircuitry coordinate with either the global olfactory circuit function or other behaviorally-related circuits to influence abGC integration? abGCs receive local glutamatergic input from OB principal cells (i.e., MTCs) and diverse modulatory input from both local interneurons and centrifugal projections, as well as other state-dependent brain centers. These inputs, while conveying behavioral context, also inherently influence abGC activity and plasticity. Combinatorial activity cues in the form of excitatory and inhibitory neurotransmitters, neuromodulators, neuropeptides, and neurotrophic factors, take place at different input sites along the basal and apical dendrites of abGCs. These activity-dependent signaling cues perform their function not only on their own, but also in spatiotemporally-coordinated patterns, which is consistent with a model that various activity inputs cooperate to modulate firing properties and the dynamics of abGCs and OB circuitry. Accordingly, any of these signals not only have the capacity to contribute to abGC survival and activity, but to also facilitate abGC long-term circuit plasticity.

Notably, Ca-dependent signaling through NMDARs mediates crosstalk between synaptic activity and nuclear transcriptional programs. Activity-activated TFs can integrate multiple signaling pathways and induce versatile effectors and cellular responses. Neuropeptide signaling mediated by GPCRs also plays an important role in abGC plasticity by correlating transcriptional changes to ever-changing physiological and brain states. Adding to the dimensionality of such mechanisms are the local signaling pathways that regulate targeted RNA translation, protein modification, degradation, and mobilization, all of which may affect cytoskeleton dynamics and synaptic composition.

The past two decades has witnessed the emergence and improvement of many techniques that facilitate the precise and elegant studies of adult neurogenesis. Two-photon in vivo imaging of genetically labeled adult-born neurons with fluorescent proteins and Ca²⁺ indicators provides the capacity to track their development longitudinally and correlate their activity with morphological features and spine dynamics. Also, optogenetic and chemogenetic based experimentation has enabled activation or inactivation of targeted cell types or projections to dissect input-specific roles of different types of activity in abGC plasticity. Combining these approaches with behavioral analysis (e.g., olfactory discrimination learning) and electrophysiological assays, the field promises to generate more insight into how different sources and patterns of activity differentially contribute to the structural and functional integration of abGCs. Glutamatergic feedback from the olfactory cortex has also been studied for its influence on behaviorally-relevant abGC plasticity. Research into the roles of GABAergic projections from the olfactory cortex and other neuro-modulatory projections onto abGCs, which are known to modulate GC firing and circuit dynamics, may provide additional details into circuit mechanisms underlying context-dependent abGC plasticity.

Finally, investigation of the molecular mechanisms that underlie abGC integration has been facilitated by molecular-based RNA-sequencing (including bulk RNA-seq, single-cell RNA-seq, and translating-RNA profiling) and CRISPR-Cas9 based KO techniques. These approaches have revealed many important players, including activity-regulated TFs, effector genes, and pathways related to cytoskeletal dynamics, ionotropic receptor signaling, and proteasomal equilibrium. Undoubtedly there are many more details underlying such activity-dependent signaling pathways yet to be revealed, especially at the levels of the epigenome, local translatome, and protein interactome. For example, in both MCs and GCs of the OB, increased in histone acetylation (an epigenetic phenomenon) is associated with aversive olfactory learning in young rats (Wang et al., 2013). Given that adult-born neurons can translate transient activity input into long-term plasticity, and that Calcium-dependent TFs may recruit priming complexes and chromatin modifiers to induce epigenetic changes, identifying the underlying epigenetic mechanisms will certainly further our understanding of abGC integration programs. Moreover, regulated local translation of pre-existing RNA within dendritic compartments and spines has been known as a way of toggling input-specific plasticity mechanisms. Coupling terminal-specific stimulation, MERFISH (K. Chen et al., 2015), and immunohistochemistry with super-resolution imaging techniques such as expansion microscopy (F. Chen et al., 2015) and/or STORM (Rust et al., 2006) also promises to provide a deeper insight into the input-specific plasticity contributed by such mechanisms. Lastly, in vivo proximity-based proteomics through methods such as BioID and TurboID can reveal dynamic protein interactomes with spatiotemporal specificity (Guo et al., 2023; Xu et al., 2021), which may also provide needed clues to novel protein interactions within spines or synapses.

In summary, although much has been learned over the last couple of decades regarding how different forms of activity influence rewiring and maintenance of adult brain tissue through ongoing adult neurogenesis, a deeper understanding of the activity-dependent mechanisms and the molecular underpinnings of adult-born neuron circuit integration will provide needed insight into the genetic programs and therapeutic targets that may be harnessed for future neuronal repair or replacement therapies.