Constitutive deficiency of the neurogenic hippocampal modulator AP2γ promotes anxiety-like behavior and cumulative memory deficits in mice from juvenile to adult periods

  1. Eduardo Loureiro-Campos
  2. António Mateus-Pinheiro
  3. Patrícia Patrício
  4. Carina Soares-Cunha
  5. Joana Silva
  6. Vanessa Morais Sardinha
  7. Bárbara Mendes-Pinheiro
  8. Tiago Silveira-Rosa
  9. Ana Verónica Domingues
  10. Ana João Rodrigues
  11. João Oliveira
  12. Nuno Sousa
  13. Nuno Dinis Alves  Is a corresponding author
  14. Luísa Pinto  Is a corresponding author
  1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Portugal
  2. ICVS/3B’s -PT Government Associate Laboratory, Portugal
  3. IPCA-EST-2Ai, Polytechnic Institute of Cávado and Ave, Applied Artificial Intelligence Laboratory, Campus of IPCA, Portugal

Abstract

The transcription factor activating protein two gamma (AP2γ) is an important regulator of neurogenesis both during embryonic development as well as in the postnatal brain, but its role for neurophysiology and behavior at distinct postnatal periods is still unclear. In this work, we explored the neurogenic, behavioral, and functional impact of a constitutive and heterozygous AP2γ deletion in mice from early postnatal development until adulthood. AP2γ deficiency promotes downregulation of hippocampal glutamatergic neurogenesis, altering the ontogeny of emotional and memory behaviors associated with hippocampus formation. The impairments induced by AP2γ constitutive deletion since early development leads to an anxious-like phenotype and memory impairments as early as the juvenile phase. These behavioral impairments either persist from the juvenile phase to adulthood or emerge in adult mice with deficits in behavioral flexibility and object location recognition. Collectively, we observed a progressive and cumulative impact of constitutive AP2γ deficiency on the hippocampal glutamatergic neurogenic process, as well as alterations on limbic-cortical connectivity, together with functional behavioral impairments. The results herein presented demonstrate the modulatory role exerted by the AP2γ transcription factor and the relevance of hippocampal neurogenesis in the development of emotional states and memory processes.

Editor's evaluation

The aim of this study was to examine the impact of AP2γ deficiency on the development of sensorimotor skills, cognitive, and emotional function. This paper will be of interest to scientists in the fields of developmental psychobiology and neurogenesis.

https://doi.org/10.7554/eLife.70685.sa0

Introduction

New cells are continuously generated, differentiated into neurons, and integrated into the preexisting neural networks in restricted regions of the postnatal mouse brain (Dennis et al., 2016; Kempermann et al., 2018; Moreno-Jiménez et al., 2019; Tobin et al., 2019). One of these so-called neurogenic niches is the subgranular zone (SGZ) of the hippocampal dentate gyrus (DG). Here, neural stem cells (NSC) give rise to mature neural cells including glutamatergic granular neurons in a finely tuned process with many developmental steps sensitive to different regulatory influences (Kempermann et al., 2004; Mateus-Pinheiro et al., 2017; Tobin et al., 2019; Toda et al., 2019). The last years have seen an increase in the number of studies on the functional specificity of hippocampal neurogenesis during development (DHN) and adult hippocampal neurogenesis (AHN). The ontogenetic interpretation of hippocampal neurogenesis assigns different functional relevance to DHN as the neurogenic process that establishes the basic repertoire of adaptable behaviors and AHN as the neurogenic process that underpins the adult brain’s ability to adapt functional behaviors (Abrous et al., 2021; Masachs et al., 2021; Tronel et al., 2015). This functional dissociation between these two stages of hippocampal neurogenesis allows us to understand how influences in the neurogenic process at different phases can lead to distinct impairments in emotional states and cognitive functions.

Postnatal hippocampal glutamatergic neurogenesis exhibits a regulatory transcriptional sequence (Sox2→ Pax6→ Ngn2→ AP2γ→ Tbr2→ NeuroD→ Tbr1) that recapitulates the hallmarks of the embryonic glutamatergic neurogenic process in the cerebral cortex (Hochgerner et al., 2018; Mateus-Pinheiro et al., 2017; Nacher et al., 2005). Transcriptional factors as Pax6, Ngn2, Tbr2, NeuroD, and Tbr1 have several and distinct roles in proliferation, cell kinetics, fate specification, and axonal growth (Englund et al., 2005; Götz et al., 1998; Hevner, 2019; Hevner et al., 2006; Hochgerner et al., 2018). Despite several efforts to understand the complex transcriptional network orchestration involved in the regulation of neurogenesis, both in early developmental stages and during adulthood, these are still to be fully understood (Bertrand et al., 2002; Brill et al., 2009; Englund et al., 2005; Hack et al., 2005; Hsieh, 2012; Mateus-Pinheiro et al., 2017; Waclaw et al., 2006). Recently, the transcription factor activating protein two gamma (AP2γ, also known as Tcfap2c or Tfap2c) was described to be an important regulator of glutamatergic neurogenesis in the adult hippocampus, being involved in the regulation of transient amplifying progenitors (TAPs) cells (Hochgerner et al., 2018; Mateus-Pinheiro et al., 2018; Mateus-Pinheiro et al., 2017). AP2γ belongs to the AP2 family of transcription factors that is highly involved in several systems and biological processes, such as cell proliferation, adhesion, developmental morphogenesis, tumor progression, and cell fate determination (Eckert et al., 2005; Hilger-Eversheim et al., 2000; Thewes et al., 2010). In addition, AP2γ is functionally relevant during embryonic neocortical development, playing detrimental roles in early mammalian extraembryonic development and organogenesis, being one of the molecular components regulating the number of upper layer neurons during ontogeny and phylogeny in the developing cortex (Pinto et al., 2009). AP2γ is critical for the specification of glutamatergic neocortical neurons and their progenitors, acting as a downstream target of Pax6 and being involved in the regulation of Tbr2 and NeuroD basal progenitors’ determinants (Mateus-Pinheiro et al., 2017; Pinto et al., 2009). Strikingly, in humans, defects in the AP2γ gene were reported in patients with severe pre- and post-natal growth retardation (Geneviève et al., 2005), and to be involved in the mammary, ovarian and testicular carcinogenesis (Hoei-Hansen et al., 2004; Li et al., 2002; Odegaard et al., 2006).

AP2γ deletion during embryonic development results in a specific reduction of upper layer neurons in the occipital cerebral cortex, while its overexpression increases region- and time-specific generation of neurons from cortical layers II/III (Pinto et al., 2009). AP2γ expression persists in the adult hippocampus, particularly in a sub-population of TAPs acting as a positive regulator of the cell fate modulators, Tbr2 and NeuroD, and therefore as a promoter of proliferation and neuronal differentiation (Mateus-Pinheiro et al., 2017). Conditional and specific downregulation of AP2γ in the adult brain NSCs decreases the generation of new neurons in the hippocampal DG and disrupts the electrophysiological synchronization between the hippocampus and the medial prefrontal cortex (mPFC). Furthermore, mice with conditional deletion of AP2γ exhibit behavioral impairments, particularly a deficient performance in cognitive-related tasks (Mateus-Pinheiro et al., 2018; Mateus-Pinheiro et al., 2017). These studies reveal the crucial modulatory role of AP2γ during embryonic cerebral cortex development as well as its influence on glutamatergic neurogenesis and hippocampal-dependent behaviors during adulthood. Still, it is relevant to assess the importance of AP2γ for postnatal neurophysiology and behavior since development, comprehending how defects in the neurogenic process since early development can impact on behavior at specific developmental stages.

Herein, we explored the neurogenic, behavioral, and functional effects of AP2γ constitutive and heterozygous deletion in mice from early postnatal development to adulthood. We revealed that AP2γ deficiency promotes downregulation of the hippocampal glutamatergic neurogenic process, compromises functional limbic-cortical connectivity, and leads to significant behavioral impairments on emotional and memory functional modalities. We showed that these impairments induced by AP2γ deficiency do not occur only during adulthood and that compromised hippocampal neurogenesis leads to emotional and memory deficits as early as the juvenile phase. These functional behavioral impairments are either maintained from the juvenile phase to adulthood or emerge as behavioral flexibility and object location recognition deficits in adult mice with AP2γ constitutive deletion.

Results

Constitutive AP2γ deficiency decreases proliferation and neurogenesis in the postnatal DG, reduces dendritic length and complexity of short-DCX+ cells, without affecting neuronal morphology of mature granular neurons

We sought to dissect the impact of constitutive and heterozygous deficiency of AP2γ in the modulation of postnatal neuronal plasticity in the hippocampus, including its effect in the hippocampal neurogenic niche on proliferation and morphology of doublecortin+ (DCX) cells and pre-existing DG granule neurons (Figure 1A). In juvenile (P31) and adult mice (P70 – P84), we assessed the expression of markers for different cell populations (Figure 1B) along the neurogenic process through western blot and immunofluorescence. We observed a significant decrease in the expression levels of AP2γ protein in the hippocampal DG of juvenile (Figure 1C) and adult (Figure 1D) AP2γ heterozygous knockout mice (Tfap2c+/-, henceforth referred as AP2γ KO), with concomitant reduction of Pax6 and Tbr2 protein levels, but not Sox2, an upstream regulator of AP2γ (Mateus-Pinheiro et al., 2017).

AP2γ constitutive and heterozygous deficiency reduces postnatal hippocampal neurogenesis both at juvenile and adult periods.

(A) Experimental timeline. (B) Transcriptional network of hippocampal neurogenesis under modulatory role of AP2γ. Western-blot analysis of AP2γ, Sox2, Pax6, and Tbr2 in juvenile (C) and adult (D) dentate gyrus (DG). (E) Hippocampal DG coronal sections stained for bromodeoxyuridine (BrdU) (green), doublecortin (DCX) (red), and DAPI (blue). BrdU+/DCX+ cells are indicated by white arrows and solely BrdU+ cell is identified with yellow arrows. (F–I) Cell counts of BrdU+ and BrdU+/DCX+ cells in the hippocampal DG of juvenile and adult mice. Data presented as mean ± SEM. Sample size: Western-blot analysis: nWT juvenile = 4; nAP2γ KO juvenile = 4; nWT adult = 4; nAP2γ KOadult = 4; Immunostainings: nWT juvenile = 6; nAP2γ KO juvenile = 6; nWT adult = 5; nAP2γ KOadult = 5 [Student’s t-test; ** p < 0.01; * p < 0.05; Statistical summary in Supplementary file 1]. Scale bars represent 50 μm. Abbreviations: WT, wild-type; AP2γ KO, AP2γ heterozygous knockout mice; O.D., optical density.

Figure 1—source data 1

Western-blot analysis at the juvenile and adult timepoints.

Western-blot analysis of AP2γ, Sox2, Pax6, and Tbr2 in juvenile (A) and adult (B) dentate gyrus (DG) protein extracts. (C and D) Representative Ponceau staining of membranes corresponding to the juvenile and adult timepoints. Surrounded by orange squares are demonstrated the quantified bands and denoted with a small orange star are identified the bands presented in the main Figure 1. Sample size: Western-blot analysis: nWT juvenile = 4; nAP2γ; KO juvenile = 4; nWT adult = 4; nAP2γ; KO adult = 4; Abbreviations: WT, wild-type; AP2γ KO, AP2γ heterozygous knockout mice.

https://cdn.elifesciences.org/articles/70685/elife-70685-fig1-data1-v2.pdf

Analysis of cell populations in the hippocampal neurogenic niche using BrdU and DCX labeling (Figure 1E) revealed that the number of BrdU+ and BrdU+/DCX+ cells is reduced in both juvenile and adult AP2γ KO mice, suggesting a decrease in the number of fast proliferating cells (TAPs) and neuroblasts, respectively (Figure 1F–I). Of note, we observed a decrease in these cell populations with age in both WT and AP2γ KO mice in agreement with previous reports (Kase et al., 2020; Katsimpardi and Lledo, 2018; Supplementary file 1).

To clarify the impact of AP2γ on the dendritic development and structural neuronal formation in the dorsal DG, we analyzed the dendritic structure and complexity of immature neurons based on DCX staining (Figure 2). DCX is a cytoskeletal protein expressed both in neuronal progenitors and immature neurons (Dioli et al., 2019; Francis et al., 1999; Horesh et al., 1999). We assessed the morphology of DCX+ cells only in adult animals as during juvenile period, the large amount of DCX+ cells does not allow a proper segmentation of dendrites through this methodology (Figure 2A). We divided the analysis of DCX+ immature neurons into those that exhibit dendrites that branch into the inner molecular layer (IML) – short-DCX+ cells – and DCX+ cells with dendritic trees that reach the medial/outer molecular layer (M/OML) – long-DCX+ cells, as previously described (Dioli et al., 2019; Figure 2B). Adult AP2γ KO mice presented decreased density of both short- and long-DCX+ cells in comparison to WT animals (Figure 2C and G). Moreover, deficiency of AP2γ led to a reduced dendritic length (Figure 2D and E) and neuronal arborization (Figure 2F) of short-DCX+ cells. Although the density of long-DCX+ cells is reduced in AP2γ KO mice, dendritic length and arborization complexity of the existing long cells are similar to the WT animals (Figure 2H-J). Interestingly, when we analyzed pre-existing granular hippocampal neurons, constitutive deficiency of AP2γ did not alter their dendritic length and the neuronal arborization complexity either during juvenile period (Figure 2—figure supplement 1A and C;) or at adulthood (Figure 2—figure supplement 1B and C), when compared to WT mice. These observations suggest that AP2γ deficiency in adult mice delays the maturation of granular neurons but has no impact on their definitive morphology.

Figure 2 with 1 supplement see all
AP2γ constitutive and heterozygous deficiency reduces DCX+ cell population, and the dendritic length of short-DCX+ cells and its arborization complexity in adult mice.

(A and B) Dorsal hippocampal DG coronal sections stained for doublecortin (DCX) (red) and DAPI (blue) with the corresponding color conversion images. These representative images include the DG subregions, the granular cell layer (GCL), the inner and medial/outer molecular layers (IML and M/OML, respectively). Short DCX+ cells are indicated by orange arrows and long DCX+ cells by yellow arrows. (B) Dendritic tree of short (left) and long (right) DCX+ cells are traced in green. Cell counts of short (C) and long (G) DCX+ cells. Dendritic length of reconstructed short (D and E) and long DCX+ (H and I) cells, and respective sholl analysis (F and J). Data presented as mean ± SEM. Sample size: nWT adult = 4; nAP2γ KOadult = 4. [Student’s t-test and Repeated measures ANOVA; ***p < 0.001, ** p < 0.01, * p < 0.05; Statistical summary in Supplementary file 1]. Scale bars represent 50 μm. Abbreviations: WT, wild-type; AP2γ KO, AP2γ heterozygous knockout mice; P, Postnatal day.

Taken together, the results herein revealed that AP2γ modulatory actions in the hippocampal neurogenic niche are similar and maintained in juvenile and adult mice. AP2γ transcription factor regulates NSCs proliferation and neuronal differentiation in the postnatal hippocampus by interacting with the different modulators involved in the transcriptional regulation of postnatal hippocampal neurogenesis. Furthermore, it suggests an important role of AP2γ transcription factor for a timely maturation of dendritic branching in newly born granular neurons in adulthood.

AP2γ KO mice display normal early postnatal development but anxiety-like behavior and memory impairments at juvenile period

In light of the negative impact of AP2γ heterozygous deficiency in the postnatal neurogenic process in the hippocampus, known as an important modulator of emotional and memory functions (Christian et al., 2014; Toda et al., 2019), we assessed early postnatal development and the behavioral performance of WT and AP2γ KO at juvenile and adult periods.

Evaluation of early postnatal development was performed through the assessment of somatic and neurobiological parameters during the first 21 postnatal days (Table 1; Figure 3—figure supplement 1). Despite a variation in the eye-opening day, responsiveness in sensory-motor functions, vestibular area-dependent tasks, and strength, as well as somatic parameters were similar in WT and AP2γ KO mice. Furthermore, all analyzed parameters were within the previously described range (Guerra-Gomes et al., 2021; Heyser, 2004). These observations suggest that constitutive heterozygous deletion of AP2γ has no impact on early postnatal development.

Table 1
Results from the milestones protocol tests included in the assessment of early postnatal neurodevelopment.

Sample size: nWT = 9; n AP2γ KO = 9. Abbreviations: WT, wild-type; AP2γ KO, AP2γ heterozygous mice; P, Postnatal day.

Milestone testWTDay (median)AP2γ KODay (median)Statistical test, significanceMann-Whitney testTypical range
Rooting77U = 31.50, p = 0.43P1 – P15
Ear twitch89U = 24.50, p = 0.16P6 – P14
Auditory startle87U = 37.50, p = 0.85P7 – P16
Open field88U = 39.50, p > 0.99P6 – P15
Walking109U = 38.00, p = 0.78P7 – P14
Surface righting77U = 38.50, p = 0.88P1 – P10
Negative geotaxis89U = 25.00, p = 0.17P3 – P15
Cliff aversion76U = 26.00; p = 0.22P1 – P14
Postural reflex98U = 26.00, p = 0.21P5 – P21
Air righting1211U = 29.00, p = 0.32P7 – P16
Wire suspension1311U = 21.50, p = 0.10P5 – P21
Grasping1616U = 30.00, p = 0.37P13 - P17
Eye-opening1211U = 9.00, p < 0.01P7 - P17
WT
Mean ±SEM
AP2γ KO
Mean ± SEM
Statistical test, significance
Repeated measures ANOVA
HomingTrial 16.89 s ± 0.446.89 s ± 1.32F (1,48) = 0.06, p = 0.80
Trial 26.33 s ± 1.126.89 s ± 1.12
Trial 35.22 s ± 1.575.78 s ± 1.57
Weight gain pattern8.99 g ± 0.539.22 g ± 0.16F (1,16) = 0.32, p = 0.58

We then performed an assessment of different behavioral domains in juvenile mice (P25 - P33; Figure 3A). The open-field (OF) and the novelty suppressed-feeding (NSF) tests were used to assess anxiety-like behavior. In the OF, juvenile AP2γ KO mice exhibited decreased distance traveled in the anxiogenic center of the arena, in comparison to WT animals, (Figure 3B) suggesting an anxiety-like phenotype promoted by AP2γ deficiency. Concomitantly, AP2γ KO mice exhibited an increased latency to reach and feed on the food pellet in the NSF test (Figure 3C and D). Of note, WT and AP2γ KO have similar locomotor activity (Figure 3—figure supplement 2A) and appetite drive (Figure 3—figure supplement 2B).

Figure 3 with 2 supplements see all
AP2γ constitutive and heterozygous deficiency increases anxiety-like behavior and promotes cognitive deficits in juvenile mice.

(A) Timeline of behavioral assessment. Anxiety-like behavior was assessed through the open-field (OF) (B) and the novelty suppressed feeding (NSF) (C and D) tests. Coping and anhedonic-like behavior by the tail-suspension (TST) (C) and the sucrose splash (SST) (D) tests. (E) To evaluate cognitive behavior, juvenile mice were subjected to the Morris water maze (MWM) (G), the object recognition (ORT) (H–J), the object-in-context recognition (OIC) (K and L) and the contextual fear conditioning (CFC) (M–Q). Data presented as mean ± SEM. Sample size: OF: nWT = 13; nAP2γ KO = 11; NSF: nWT = 13; nAP2γ KO = 15; TST: nWT = 16; nAP2γ KO = 13; SST: nWT = 15; nAP2γ KO = 10; MWM: nWT = 10; nAP2γ KO = 9; ORT: nWT = 11; nAP2γ KO = 8; OIC: nWT = 16; nAP2γ KO = 17; CFC: nWT = 16; nAP2γ KO = 17. [Student’s t-test, Repeated measures ANOVA and Two-way ANOVA; *** p < 0.001; ** p < 0.01; * p < 0.05; Statistical summary in Supplementary file 1]. Abbreviations: WT, wild-type; AP2γ KO, AP2γ heterozygous knockout mice.

The assessment of adaptive behavior to cope with an inescapable stressor and self-care were then evaluated throughout the tail suspension test (TST) and the sucrose splash test (SST), respectively. We observed no impact of AP2γ constitutive and heterozygous deletion in these emotional domains, as no alterations were detected in the immobility time in the TST (Figure 3E) and grooming time in the SST (Figure 3F).

Furthermore, we assessed memory in juvenile mice through the Morris water maze (MWM), the object recognition test (ORT), the object-in-context recognition test (OIC), and the contextual fear conditioning (CFC) test. In the MWM, WT and AP2γ KO juvenile mice present a high and steady escape latency to find the hidden platform throughout training, implying that, at this age, mice are unable to learn this spatial navigation task at this early age (Figure 3G). This phenotype is consistent with the postnatal maturation state of the DG (Hochgerner et al., 2018; Kozareva et al., 2019).

In the ORT (Figure 3H), despite no differences in the novel object location (Figure 3I), AP2γ KO mice displayed significant deficits in the novel object recognition, as denoted by a decreased preference to explore the novel object (Figure 3J). In the OIC behavioral paradigm (Figure 3K), WT mice spent a greater proportion of time exploring the out of context object, than the object in the familiar environment, whereas AP2γ KO mice showed no preference of object exploration (Figure 3L). Thus, AP2γ KO mice were not able to successfully discriminate between the two distinct contexts, demonstrating an impaired PFC-hippocampus-perirhinal cortex network (Barker and Warburton, 2020).

In the CFC, a behavior test described to be sensitive to changes in hippocampal neurogenesis (Gu et al., 2012), juvenile mice were subjected to two distinct context tests, aimed to test hippocampal-dependent memory, and a cue probe to assess the integrity of extrahippocampal memory circuits (Figure 3M; Gu et al., 2012; Mateus-Pinheiro et al., 2017). As expected, prior to conditioning, freezing behavior was low, and WT and AP2γ KO mice identically explored the acrylic cylinder (Figure 3N). In context A, juvenile AP2γ KO mice exhibited reduced freezing behavior when exposed to a familiar context (Figure 3O). No alterations in freezing behavior were observed neither in context B (Figure 3P) nor in the cue probe (Figure 3Q). These CFC results, along with the OIC findings, reveal that AP2γ KO juvenile mice exhibit deficits in contextual memory.

Despite no evident impact on early postnatal development, AP2γ constitutive heterozygous deficiency impairments in hippocampal neurogenesis since development leads to a hypersensitivity toward aversive stimulus, as revealed by the presence of an anxious-like phenotype and memory impairments. These results highlight the involvement of the AP2γ transcription factor in the proper acquisition of important behavioral functions.

Adult AP2γ KO mice maintain emotional and memory dysfunctions, while behavioral flexibility and object location recognition impairments emerge during adulthood

Adult WT and AP2γ KO mice were also tested for emotional and memory paradigms to determine if behavioral changes caused by the effects of AP2γ constitutive deletion on postnatal hippocampal neurogenesis were still present in adulthood (Figure 4A).

Figure 4 with 1 supplement see all
Behavioral assessment of adult mice.

(A) Experimental timeline. Anxiety-like behavior was assessed through open-field (OF) (B), elevated plus-maze (EPM) (C) and novelty suppressed feeding (NSF) (D and E) tests, while coping behavior was evaluated through forced-swimming test (FST) (G) and the tail-suspension test (TST) (G). Object recognition test (ORT) (H and I), object-in-context recognition (OIC) (J) and the contextual fear conditioning (CFC) (K–N) were performed to assess cognitive behavior. Data presented as mean ± SEM. Sample size: OF, EPM and FST: nWT = 12; nAP2γ KO = 14; NSF: nWT = 10; nAP2γ KO = 9;TST: nWT = 6; nAP2γ KO = 6; ORT: nWT = 12; nAP2γ KO = 9; OIC: nWT = 12; nAP2γ KO = 10; CFC: nWT = 7; nAP2γ KO = 6. [Student’s t-test and Two-way ANOVA; *** p < 0.001; ** p < 0.01; * p < 0.05; Statistical summary in Supplementary file 1]. Abbreviations: WT, wild-type; AP2γ KO, AP2γ heterozygous knockout mice.

The OF, the elevated plus-maze (EPM), and the NSF tests were performed to evaluate anxiety-like behavior. In the OF test, AP2γ KO mice showed a trend toward a decrease in the distance traveled in the anxiogenic center of the arena (P = 0.05, Figure 4B), with no changes in locomotor activity as denoted by similar average velocity (Figure 4—figure supplement 1A). In the EPM test, AP2γ KO mice spent significantly less time in the open arms than WT mice (Figure 4C). These results are in agreement with results from the NSF test, where AP2γ KO mice displayed increased latency to reach and feed from the food pellet (Figure 4D and E). No changes in appetite drive were observed between groups, since they fed similarly in the post-NSF food assessment (Figure 4—figure supplement 1B). The forced swimming test (FST) and TST were then performed to examine the adaptive behavior to cope with inescapable stressors. WT and AP2γ KO mice exhibit similar immobility time in FST (Figure 4F) and TST (Figure 4G). These observations suggest that anxiety-like behavior promoted by constitutive AP2γ deficiency persist in adult mice, with no alteration in coping behavior.

Additionally, memory performance was evaluated by the ORT, OIC, CFC, and MWM. ORT revealed a trend toward a decrease in the preference to explore the displaced object of AP2γ KO when compared to WT mice (p = 0.08, Figure 4H). However, no alterations were observed in preference toward the novel object (Figure 4I). In the OIC paradigm, results were similar to juvenile mice. AP2γ KO adult mice were not able to discriminate between the two contexts, since they showed no preference of object exploration, whereas WT animals spent a greater proportion of time exploring the out of context object (Figure 4J). Prior to conditioning, freezing behavior in the CFC was low, and WT and AP2γ KO performed similarly (Figure 4K). AP2γ KO mice exhibited reduced freezing behavior when exposed to a familiar conditioning context - context A (Figure 4L), whereas no alterations in freezing behavior were observed neither in context B (Figure 4M) nor in the cue probe (Figure 4N). These CFC observations, along with the results from the OIC paradigm, suggest that adult AP2γ KO mice have deficits in contextual memory and an intact associative non-hippocampal-dependent memory.

The experimental groups were also subjected to the MWM test for evaluation of spatial memory (Figure 5). In the reference memory task, that relies on hippocampal function integrity (Cerqueira et al., 2007), AP2γ KO and WT mice exhibit similar performance to reach the hidden platform along the training days (Figure 5A and B). When the platform was changed to the opposite quadrant to assess behavior flexibility, a task that relies not only in hippocampal neurogenesis (Anacker and Hen, 2017; Dupret et al., 2008) but also in prefrontal cortical areas (Hamilton and Brigman, 2015), adult AP2γ KO mice spent less time in the new quadrant than WT animals (Figure 5C) suggesting that constitutive AP2γ deficiency leads to impaired behavioral flexibility. Detailed analysis of the strategies adopted to reach the escape platform (Antunes et al., 2021; Garthe et al., 2009; Garthe and Kempermann, 2013; Mateus-Pinheiro et al., 2017; Ruediger et al., 2012) revealed that AP2γ KO mice delayed the switch from non-hippocampal dependent (‘Block 1’) to hippocampal-dependent (‘Block 2’) strategies (Figure 5D–H), suggesting an impairment of hippocampal function. No differences were found in the working memory task (Figure 5—figure supplement 1A and B).

Figure 5 with 1 supplement see all
Cognitive performance of adult mice in the Morris water maze test.

(A and B) Spatial reference memory was assessed as the average escape latency to find a hidden and fixed platform in each test day. (C) In the last testing day, animals were subjected to a reversal-learning task to test behavioral flexibility. (D) Schematic representation of typical strategies to find the platform during spatial memory evaluation grouped according to its dependence of the hippocampus (Block 1: Non-hippocampal dependent strategies; Block 2: Hippocampal dependent strategies). Average of each strategy used for WT (E) and AP2γ KO (F) mice, by trial number. The prevalence of each block along with trials, the distribution of strategies-block boundaries, and overall block length are shown for (G) WT and (H) AP2γ KO mice. Data presented as mean ± SEM. nWT = 10; nAP2γ KO = 10. [Repeated measures ANOVA and Student’s t-test; ***p < 0.001; Statistical summary in Supplementary file 1]. Abbreviations: WT, wild-type; AP2γ KO, AP2γ heterozygous knockout mice.

Overall, these results show that AP2γ constitutive deficiency negatively modulates not only early postnatal hippocampal neurogenesis but also AHN. These neurogenic dysfunctions result in emotional and memory deficits that persist in the adult mice, together with impairments in behavioral flexibility and object location recognition impairments that emerge during adulthood. Specifically, adult AP2γ KO mice exhibit anxiety-like behavior, contextual-memory impairments, and deficiencies in behavioral flexibility capacities. Moreover, due to the relevance of the hippocampus and mPFC to memory-related behavioral tests here assessed (OIC and MWM behavioral flexibility task) in adult mice, the functional integrity of these brain areas is suggested to be affected by constitutive and heterozygous deficiency of AP2γ.

Adult hippocampal-to-PFC functional connectivity is disrupted by AP2γ constitutive and heterozygous deficiency

Given the impact of AP2γ deficiency on emotional behavior and memory, we sought for a functional correlate by investigating related neurocircuits. In adult WT and AP2γ KO mice, we explored the integrity of the hippocampus-to-medial prefrontal cortex (mPFC) circuitry, assessing electrophysiological features of local field potentials (LFPs) simultaneously in these connected brain areas (Figure 6A and Figure 6—figure supplement 1A). In AP2γ KO mice, the temporal structure of LFPs recorded simultaneously from the dorsal hippocampus (dHip) and mPFC was affected. Specifically the spectral coherence between these regions (Adhikari et al., 2010; Oliveira et al., 2013; Sardinha et al., 2017) in AP2γ KO mice is significantly decreased in delta, theta and beta frequency bands when compared to WT littermates (Figure 6B), suggesting the importance of AP2γ for an intact dHip-mPFC functional connectivity. AP2γ constitutive deficiency had a subtle impact in PSD values in the dHip, specifically in the Theta and Beta frequency bands (Figure 6C). In the mPFC, PSD values in delta, theta and beta frequencies bands were significantly lower in AP2γ KO than in WT mice (Figure 6D). In contrast, deficiency of AP2γ did not exert an effect neither in the spectral coherence between the vHip and the mPFC (Figure 6—figure supplement 1B) nor in the PSD values in the vHip (Figure 6—figure supplement 1C).

Figure 6 with 1 supplement see all
In adult mice, AP2γ constitutive and heterozygous deficiency induces deficits in spectral coherence between the dorsal hippocampus (dHip) and the medial prefrontal cortex (mPFC), impacting on neuronal activity.

(A) Identification of the local field potential (LFP) recording sites, with a depiction of the electrode positions (upper panel), and representative Cresyl violet-stained sections, with arrows indicating electrolytic lesions at the recording sites (lower panel). (B) Spectral coherence between the dHip and mPFC (left panel). Group comparison of the coherence values for each frequency (right panel). (C) Power spectral density (PSD) was measured in the dHip (C) and mPFC (D). Heatmaps of PSD activity (upper panel) and group comparison for each frequency (lower panel). Each horizontal line in the Y-axis of the presented spectrograms represents an individual mouse. Frequency bands range: delta (1–4 Hz), theta (4–12 Hz), beta (12–20 Hz), low gamma (20–40 Hz), and High gamma (40–90 Hz). Data presented as mean ± SEM. nWT = 6; nAP2γ KO = 5. [Repeated measures ANOVA; *p < 0.05; Statistical summary in Supplementary file 1]. Abbreviations: WT, wild-type; AP2γ KO, AP2γ heterozygous knockout mice.

Figure 6—source code 1

Local field potentials analysis between the dorsal hippocampus (dHip) and medial prefontal cortex (mPFC).

https://cdn.elifesciences.org/articles/70685/elife-70685-fig6-code1-v2.zip

The electrophysiological studies revealed that AP2γ constitutive and heterozygous deficiency led to two outcomes: first, a significant decrease of coherence between the dHip and the mPFC indicating impairments in the ability of these regions to functionally interact; second, this decrease in interregional coherence was accompanied by a diminished neuronal activity in a range of frequencies in the mPFC, including in theta and beta frequencies, previously shown to be critically related with behavioral outputs dependent on cortico-limbic networks (Colgin, 2011; Fell and Axmacher, 2011; Oliveira et al., 2013).

Discussion

Herein, we show that despite a normal early postnatal acquisition of neurodevelopmental milestones, AP2γ constitutive and heterozygous deficiency causes emotional and memory deficits in juvenile mice that persist into adulthood, while behavioral flexibility and object location recognition impairments emerge only in adult mice. These findings suggest that AP2γ is required for the proper development and maturation of neural circuits involved in relevant behavioral functions from early development to adulthood.

Newly generated neurons are highly relevant to hippocampal functioning and hippocampal-associated behaviors (Anacker and Hen, 2017; Christian et al., 2014; Fang et al., 2018; Gonçalves et al., 2016). Impairments in hippocampal neurogenesis precipitate the emergence of depressive- and anxiety-like behaviors (Bessa et al., 2009; Hill et al., 2015; Mateus-Pinheiro et al., 2013a; Mateus-Pinheiro et al., 2013b; Revest et al., 2009; Sahay and Hen, 2007). Here, we assessed for the first time the longitudinal impact of AP2γ, a transcription factor that plays an important role on embryonic neuronal development (Pinto et al., 2009) and recently described as a novel regulator of adult hippocampal neurogenesis (Mateus-Pinheiro et al., 2018; Mateus-Pinheiro et al., 2017), on neural plasticity, function and behavior at different postnatal periods.

Characterization of the neurogenic process in the hippocampal DG in juvenile and adult constitutive AP2γ KO mice, revealed that in agreement with our previous reports (Mateus-Pinheiro et al., 2018; Mateus-Pinheiro et al., 2017), AP2γ regulates upstream neurogenic regulators as Pax6 and Tbr2. Other modulators of the TAP’s population, such as Ngn2 and Tbr2, have been shown to exert a similar control of hippocampal neurogenesis (Galichet et al., 2008; Hodge et al., 2012; Roybon et al., 2009; Tsai et al., 2015). Also, we observed, at juvenile period and during adulthood, that AP2γ plays an essential role in the regulation of pivotal neurogenic steps of NSCs proliferation and neuronal maturation. In addition to a decreased proliferation and neuroblasts population, AP2γ constitutive and heterozygous deficiency led to a reduced number of immature neurons and delayed morphological maturation of granular neurons in the adult dorsal hippocampal DG. Yet, these reduced neurogenesis and delayed maturation did not alter the definitive morphology of granular neurons, another form of hippocampal structural plasticity (Bessa et al., 2009; Mateus-Pinheiro et al., 2013a). Due to a large amount and density of DCX+ cells in the DG of juvenile mice (Kase et al., 2020; Katsimpardi and Lledo, 2018), we were unable to analyze their morphology and understand the impact of AP2γ deletion at this period.

Taking into consideration the embryonic and early postnatal developmental modulatory roles of AP2γ, and the severe and/or lethal malformations during development promoted by deficiencies in other members of the AP2 family (AP2α and AP2β) (Lim et al., 2005; Moser et al., 1997; Schorle et al., 1996), we sought to understand whether postnatal hippocampal neurogenesis deficits induced by constitutive and heterozygous deficiency of AP2γ impacts on behavior in specific developmental stages.

The developmental milestones protocol showed no impact of AP2γ constitutive and heterozygous deficiency in early postnatal neurodevelopment. Nevertheless, this deficiency in AP2γ promotes deficits in hippocampal neurogenesis altering the ontogeny of emotional and memory responses in juvenile and adult mice. At juvenile age and during adulthood, AP2γ deficiency led to the manifestation of anxiety-like behavior and significant impairments in different memory behavioral tasks dependent on hippocampal function (Garthe et al., 2009; Garthe and Kempermann, 2013; Gu et al., 2012; Jessberger et al., 2009; Ruediger et al., 2012; Treves et al., 2008). These functional behavioral impairments are either maintained from the juvenile phase until adulthood or emerge in adult mice with deficits in behavioral flexibility and object location recognition together with decreased AHN. Given the importance of AP2γ for the proliferation and expansion of a subpopulation of TAPs (Tbr2+), our results reveal that downregulating this specific cell population since early development promotes a significant long-term impact on different behavioral functions. Regulation of TAPs’ by AP2γ seems to be also relevant for the preservation of cognitive performance, as shown by its impact on hippocampal-dependent emotional and memory tasks. These observations are in agreement with previous publications where suppression of the TAP’s regulator Tbr2 exerted both an anxiety-like phenotype during the juvenile period and induced cognitive deficits during early adulthood (Veerasammy et al., 2020). Moreover, Ngn2, another regulator of TAP cells, is also reported as important for the modulation of cognitive behavior, namely in the rescue of cognitive function in the T-Maze task (Zhao et al., 2018).

AP2γ deficiency resulted in poor performance on the MWM’s behavioral flexibility task and compromised object context discrimination in the adult mice, both of which are memory tasks that rely on the interaction of the hippocampal and prefrontal cortical brain areas (Barker and Warburton, 2020; Hamilton and Brigman, 2015). Interestingly, adult mice with constitutive and heterozygous deletion of AP2γ present significant deficits of electrophysiological coherence between the dHip and the mPFC in a wide range of frequencies, previously associated to behavior outputs dependent on cortico-limbic networks (Colgin, 2011; Fell and Axmacher, 2011; Oliveira et al., 2013; Sardinha et al., 2017). Previously, we observed that conditional deletion of AP2γ during adulthood led to coherence impairments between the vHip-to-mPFC (Mateus-Pinheiro et al., 2017). This apparent discrepancy may be related to the specific time of AP2γ deletion on these different mice models. Yet, similar electrophysiological deficits in dHIP-mPFC coherence were observed in a rat model of hippocampal cytogenesis abrogation, which also denote long-term manifestation of emotional and cognitive deficits (Mateus-Pinheiro et al., 2021). Moreover, the integrity of the hippocampus-to-PFC circuitry was described to be relevant for the action of antidepressants, such as ketamine (Carreno et al., 2016), which promote neurogenesis, suggesting that AP2γ may be involved in conserving this neuronal circuit. Additionally, AP2γ plays an important role on cortical basal progenitors’ specification during embryonic development (Pinto et al., 2009) that might be affecting the mPFC activity. In fact, AP2γ KO mice presented impaired neuronal activity in the mPFC in a wide range of frequency bands, as detected by the general decrease of PSD signals recorded, and in concordance to previous studies (Mateus-Pinheiro et al., 2017). Thus, misspecification of upper cortical layers promoted by AP2γ deficiency since embryonic development may be contributing to the functional electrophysiological alterations, and eliciting cognitive defects herein observed.

Collectively, the findings of this study show that AP2γ transcription factor deficiency compromises the generation of hippocampal glutamatergic neurons, altering the ontogeny of behavior complexity associated with the hippocampus formation. Decreased hippocampal neurogenesis since early development in AP2γ KO mice leads to a compromised basic repertoire of adaptive behaviors. Impairments in DHN induced by deficiencies in AP2γ transcription factor leads to a hypersensitivity toward aversive stimulus as revealed by the presence of an anxious-like phenotype, and memory impairments shown in multiple cognitive tests. These functional behavioral impairments are either maintained from the juvenile phase until adulthood or emerge in adult mice as a result of the reduced AHN caused by constitutive deletion of AP2γ. These observations highlight the importance of an intact hippocampal neurogenic process for such functional outputs.

The current study adds a temporal window of analysis not yet studied and hints that future works should elucidate whether AP2γ can participate in the pathogenesis of neurodevelopmental and/or psychiatric disorders and whether positive modulation of AP2γ could be applied at younger ages as a therapeutic approach to revert deficits in hippocampal neurogenesis and associated behavioral impairments.

Materials and methods

Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Genetic reagent (Mus musculus, male)Tfap2c+/-Dr. Hubert Schorle (Bonn University Medical School)Male mice maintained in a 129/SV background
Chemical compound, drugBrdU, 50 mg/kgSigma-Aldrich# 9,285Intraperitoneal injection
Antibodyalpha-tubulin (Mouse monoclonal)Sigma#5,168WB (1:5000)
AntibodyAnti-AP2γ (Goat polyclonal)Abcam#31,288WB (1:500)
AntibodyAnti-Pax6 (Rabbit polyclonal)Millipore#2,237RRID:AB_1587367WB (1:1000)
AntibodyAnti-Sox2 (Mouse monoclonal)Abcam#7,935WB (1:500)
AntibodyAnti-mouse (Goat monoclonal)BioRad#1706516RRID:AB_11125547WB (1:10000)
AntibodyAnti-rabbit (Goat monoclonal)BioRad#1706515RRID:AB_11125142WB (1:10000)
AntibodyAnti-goat (Donkey polyclonal)Santa Cruz Biotechnologies#sc-2020RRID:AB_631728WB (1:7500)
Commercial assay or kitSuperSignal west Femto reagentThermoFisher#34,096
AntibodyAnti-BrdU (Rat monoclonal)Abcam#6,326RRID:AB_305426IF (1:100)
AntibodyAnti-Doublecortin (Rabbit polyclonal)Abcam#18,723RRID:AB_732011IF (1:100)
AntibodyAnti-Doublecortin (Goat monoclonal)Santa Cruz Biotechnologies#sc-8066RRID:AB_2088494IF (1:500)
AntibodyAlexa Fluor 488 (Goat Anti-rat polyclonal)Invitrogen#32,731IF (1:1000)
AntibodyAlexa Fluor 568 (Goat Anti-rabbit polyclonal)Invitrogen#11,011RRID:AB_143157IF (1:1000)
AntibodyAlexa Fluor 594 (Donkey Anti-Goat polyclonal)Invitrogen#A32758RRID:AB_2534105IF (1:1000)
other4',6-diamidino-2-phenylindole (DAPI stain)Sigma Aldrich#8,417IF (1:200)
Commercial assay or kitAnti-fade Fluorescence Mounting MediumAbcam#ab104135
Software, algorithmActivity Monitor softwareMedAssociates
Software, algorithmKinoscope softwareKokras et al., 2017
Software, algorithmFiji softwareSchindelin et al., 2012RRID: SCR_002285
Software, algorithmEthovision XT 11.5NoldusRRID:SCR_000441
Software, algorithmSignal SoftwareCED
Software, algorithmPrism v.8GraphPad Software IncRRID:SCR_002798
Software, algorithmMATLABMathWorks IncRRID:SCR_005547

Experimental model details

Request a detailed protocol

Wild-type (WT) and AP2γ heterozygous knock-out (Tfap2c+/-, henceforth referred as AP2γ KO) mice were kindly provided by Dr. Hubert Schorle. This previously generated mice line was obtained through a multistep vector- and animal-manipulation (Werling and Schorle, 2002). Briefly, a 6 kb genomic fragment spanning from exons 2–7 was used to insert a floxed neo/tk cassette 3’ of exon five as well as a loxP site 5’ of exon 5. Following electroporation into embryonic stem cells (ES) the correct clones were antibiotically selected using G418. Cre-mediated excision removed the region between the loxP sites, generating the AP2γ null allele. To generate the AP2γ KO mice, ESs with the null allele were injected into blastocysts resulting in chimeric animals. The offspring of germ line transmitting animals were tested for the presence of the AP2γ null allele. AP2γ KO transgenic mice line was maintained in a 129/SV background and identified by polymerase chain reaction (PCR) of genomic DNA (Werling and Schorle, 2002).

Along the study, distinctive cohorts (at least two per timepoint) of littermate WT and AP2γ KO male mice were submitted to molecular (n = 4–6 per group), behavioral (n = 6–17 per group) and electrophysiological (n = 5–6 per group) assessment at the different postnatal ages (early postnatal period: from postnatal day (P)one to P21; juvenile period, from P25 to P33; adulthood, from P70 to P84). None of the animals’ cohorts were used in multiple timepoints. Different cohorts of animals were used for each analyzed developmental stage to avoid any potential influence of habituation or learning from the repetition of behavior assessment and avoid any long-term impact of behavioral testing during development periods.

As a result of limiting the number of behavioral tests to which each animal was subjected, there is some variation in sample size across behavioral tests. All mice were housed and kept under standard laboratory conditions at 22°C ± 1°C, 55 % humidity, and ad libitum access to food and water on a 12 hr light/dark cycle (lights on 8 A.M. to 8 P.M.). Efforts were made to minimize the number of animals and their suffering. All experimental procedures performed in this work were conducted in accordance with European Regulation (European Union Directive 2010/63/EU) and approved by the Portuguese National Authority for animal experimentation, Direção-Geral de Alimentação e Veterinária (DGAV) with the project reference 0420/000/000/2011 (DGAV 4542).

Behavioral analysis

Early postnatal period

Request a detailed protocol
Developmental milestones protocol
Request a detailed protocol

Early postnatal neurodevelopment in mice was assessed according to previously validated protocols (Castelhano-Carlos et al., 2010; Guerra-Gomes et al., 2021; Hill et al., 2008; Santos et al., 2007). This consisted in a daily evaluation for the first 21 days of life. Mice subjected to early postnatal behavioral assessment did not perform any additional behavioral test later in life. From postnatal day P1 onward, newborn animals were evaluated in several parameters, including skin appearance, activity, and presence of milk spot in the stomach, indicator of correct maternal care and well-being. Pups were examined for the acquisition of developmental milestones until weaning (P21), every day at the same time, in the same experimental room, by the same experimenter. This daily scoring included tests to assess the acquisition of mature response regarding somatic parameters and neurobiological reflexes.

Somatic parameters
Request a detailed protocol

As a measure of morphological development, animals were daily weighed (weight ±0.01 g). The eye-opening day was also evaluated and considered when both eyes were opened. When both eyes opened on different days, score was set as one if only one of the eyes was open, and two when both eyes were open. The mature response was registered when both eyes were open.

Neurobiological reflexes
Request a detailed protocol

The assessment of the neurobiological reflexes included the daily performance of different tests. Of note, the scale of evaluation was distinct among tests. Tests including rooting, ear twitch, auditory startle, open field transversal, air righting, wire suspension, postural reflex were scored according to the absence (0) or presence (1) of a mature response. When it was possible to detect a gradual progression in performance as for walking, surface righting, grasping, negative geotaxis, cliff aversion, daily score was attributed between 0 and 3, with 0 representing absence, and three corresponding to the achievement of the mature response. The postnatal day in which animals achieved a mature response was registered. All tests were conducted in a smooth foam pad, and immediately after testing, the pups were returned to their home cage.

Labyrinthine reflex, body righting mechanism, coordination and strength

Request a detailed protocol

Surface righting reflex – P1 to P13 – This test consists of gently laying the animal on its back, and the mature response was considered when a pup was able to get right. If the animal did not respond within 30 s, the test was ended. Mature response was achieved when the pups were able to get right in less than 1 s for 3 consecutive days.

Negative geotaxis – P1 to P14 – Pups were placed head down in a horizontal grid, tilted 45° to the plane. The acquisition of a mature response was set when pups were able to head up in less than 30 s for 3 consecutive days.

Air righting – P8 to P21 – In this test, the pup was held upside down and released from a height of approximately 13 cm from a soft padded surface and released. A mature response was obtained when the animals landed on four paws for 3 consecutive days.

Cliff aversion – P1 to P14 – It evaluates the mouse pup’s ability to turn and crawl away when on the edge of a cliff. A mature response was achieved once the animal moved away in less than 30 s for 3 consecutive days.

Postural reflex – P5 to P21 – Pups were placed in a small plastic box and gently shaken up down and right. When the animals were able to maintain their original position in the box by extending four paws, mature response was acquired.

Wire suspension – P5 to P21 – This test evaluates forelimb grasp and strength. Pups were placed vertically to hold with their forepaws a 3 mm diameter metal wire suspended 5 cm above a soft foam pad. A mature response was achieved once the animal was able to grasp the bar, holding it with four paws.

Grasping – P5 to P21 – The mouse pup forelimb was touched with a thin wire to evaluate when the involuntary freeing reflex stopped. This reflex disappears with the development of the nervous system, as so, the mature response achieved when the animal grasped immediately and firmly the wire.

Tactile reflex:

Request a detailed protocol

Ear twitch – P7 to P15 – In this test, the mouse pup ear was gently stimulated with the tip of a cotton swab, three times. If the animal reacted, flattening the ear against the side of the head for three consecutive days, the mature response was reached.

Rooting – P7 to P12 – A fine filament of a cotton swab was used to gently and slowly rub the animal’s head, from the front to the back. It was considered a successful test if the pup moved its head toward the filament. The test was repeated on the other side of the head to evaluate the appearance of this neurobiological reflexes on both sides. If the animal did not react to the filament, the test was repeated. Mature response was obtained when animal reacted on both sides for three consecutive days.

Auditory reflex:

Request a detailed protocol

Auditory startle – P7 to P18 – We evaluated the reaction of pups to a handclap, at a distance of 10 cm. If pups quick and involuntary jumped for three consecutive days, a matured response was attributed.

Motor:

Request a detailed protocol

Open field transversal – P7 to P18 – To execute this test, animals were placed in a small and circle (13 cm diameter), and time to move was recorded. If the pup was not able to move, the test was ended. In case the mouse leaves the circle in less than 30 s, for three consecutive days, a mature response was reached.

Walking – P5 to P21 – In this paradigm, animals were able to freely move around for 60 s. The mature response was achieved when they showed a walking movement fully supported on their four limbs.

Juvenile and adult mice

During juvenile and adult periods and through different previously validated behavior tests, we assessed sequentially anxiety-like behavior, coping and anhedonic behavior, and the performance on multiple cognitive tasks. Individual cohorts of mice were exclusively subjected to behavioral characterization at one of these periods, either while juvenile or later at adulthood. To avoid any possible impact of habituation or learning from the repetition of behavior assessment, we exposed different cohorts of animals to a subset of behavioral paradigms. Thus, with this approach of limiting the number of behavioral tests each animal was exposed to, there is some variation in the sample number across behavioral paradigms. Behavioral tests were performed in the following order: Open-field, elevated-plus maze, novelty-suppressed feeding, sucrose splash, forced-swimming, tail suspension, object recognition, pattern separation, Morris water maze, and contextual fear conditioning. When mice were exposed to multiple behavioral paradigms, 1–2 days of resting were added between tests. Sucrose splash test was only performed in juvenile mice while elevated-plus maze and forced-swimming test were exclusively tested during the adult period.

Open field (OF) test

Request a detailed protocol

The OF test is a behavioral paradigm commonly used to assess anxiety-like and exploratory behavior, and general activity in rodents. This apparatus consists of a square arena of 43.2 cm x 43.2 cm, closed by a 30.5 cm high wall, with a lamp at the center (~850 lux). Mice were individually placed in the center of the OF arena, and their movement was tracked for 5 mins, using a 16-beam infrared system (MedAssociates, US). Average velocity of the animals was considered as a measure of locomotor capacity, and the percentage of distance traveled in the center of the arena (defined as the central 11 × 11 cm region) was considered as measurement of anxiety-like behavior. Data was analysed using the Activity Monitor software (MedAssociates, US). For juvenile behavioral assessment, mice were exposed to this paradigm at P25, and adult animals were between P70 and P84.

Elevated-plus maze (EPM) test

Request a detailed protocol

To assess anxiety-like behavior, the EPM test was also performed (Walf and Frye, 2007). This test is performed on a black propylene apparatus (ENV – 560; MedAssociates Inc, US) with two opposite open arms (50.8 cm x 10.2 cm) and two closed arms (50.8 cm x 10.2 cm x 40.6 cm) elevated 72.4 cm above the floor and dimly illuminated. The central area connecting both arms corresponds to an area of 10 × 10 cm. Animals were individually positioned in the center of the maze, facing an edge of a closed-arm, and were allowed to freely explore the maze for 5 min. All trials were recorded using an infrared photobeam system, and the percentage of time spent in the open arms was accessed through the EthoVision XT 11.5 tracking system (Ethovision, Noldus Information Technologies, Netherlands). For the adult behavioral assessment, animals were between P70 and P84.

Novelty suppressed feeding (NSF) test

Request a detailed protocol

Following ~18 h of food deprivation, animals were placed in an illuminated open-field arena with 43.2 cm x 43.2 cm, closed by a 30.5 cm high wall, with a lamp at the center (~850 lux). The arena floor was covered with 2 cm of the habitual bedding material, and a single food pellet was placed on a 5 cm square piece of filter paper at the center. Mice were placed facing a wall, and the latency to reach and feed on the food pellet was assessed, within a maximum limit of 10 min. Feeding behavior was defined as rearing with visible food consumption. Latency to reach and feed on the food pellet was used as a measure of anxiety-like behavior (Samuels and Hen, 2011). After this test trial, animals were single housed in homecages containing a pre-weighted food pellet, and allowed to feed during 10 min. The amount of food consumption in the home cage provided a measure of appetite drive (Bessa et al., 2009). For juvenile behavioral assessment, mice were exposed to this paradigm at P27, and adult animals were between P70 and P84.

Forced swimming test (FST)

Request a detailed protocol

For the assessment of adaptive behavior to cope with an inescapable stressor, we performed the forced swimming test (Porsolt et al., 1977). Briefly, each animal was individually placed in glass cylinders filled with water (23 °C; depth 30 cm) for 5 min. All sessions were video-recorded, and the immobility time, defined through a video tracking software Ethovision XT 11.5 (Noldus, Netherlands), was considered as a measure of coping behavior. Mice were considered immobile when all active behaviors (struggling, swimming, and jumping) were ceased. For immobility, the animals had to remain passively floating or making minimal movements required to maintain the nostrils above water. For the assessment of coping behavior, the first 3 min of the trial were considered as a habituation period and the last 2 min as the test period (Mateus-Pinheiro et al., 2017). For the adult behavioral assessment, animals were between P70 and P84.

Tail-suspension test (TST)

Request a detailed protocol

The TST is a commonly used behavioral test to assess adaptive behavior to cope with an inescapable stressor in rodents. Briefly, mice were suspended by the tail to the edge of a laboratory bench 80 cm above the floor (using adhesive tape) for 6 min. Trials were video-recorded, and the immobility and climbing times were automatically analyzed by the video tracking software Ethovision XT 11.5 (Noldus, Netherlands). In the TST, immobility time is normally defined as a readout of coping behavior. Animal models of depression typically exhibit longer periods of immobility during the TST (Krishnan and Nestler, 2011). For the juvenile behavioral assessment, mice were exposed to this paradigm at P26, and adult animals were between P70 and P84.

Splash-sucrose test (ST)

Request a detailed protocol

The splash-sucrose test consists of spraying a 10 % sucrose solution on the dorsal coat of mice in their home cage (Yalcin et al., 2008). Sucrose solution in the mice’s coat induces grooming behavior. After spraying the animals, animal’s behavior was video-recorded for 5 min and the time spent grooming was taken as an index of self-care and motivational behavior. Grooming was manually quantified using the behavioral scoring program Kinoscope (Kokras et al., 2017). For juvenile behavioral assessment, mice were exposed to this paradigm at P27.

Object recognition test (ORT)

Request a detailed protocol

The ORT was performed to assess short- and long-term memory (Leger et al., 2013). This test relies on rodents’ nature to explore and preference for novelty. First, mice were acclimatized to a testing arena (30 cm x 30 cm x 30 cm) under dim light (desk lamp pointing to a room wall ~150–200 lux at the arenas level) for 3 days during 20 min. After habituation, animals were presented with two equal objects for 10 min (training), positioned in the center of the arena. One hour later, one of the objects was moved toward an arena wall, and mice were allowed to freely explore the objects for 10 min. On the following day, animals returned to the arena for 10 min, with one of the objects being replaced by a novel object. The familiar and novel objects had different size, color, shape, and texture. These objects were made from building blocks as represented in Figure 3G. Between trials, the arena and objects were properly cleaned with 10 % ethanol. Sessions were recorded and manually scored through the behavioral scoring program Kinoscope (Kokras et al., 2017). The percentage of time exploring the moved- and the novel-object was used as a measure of short- and long-term memory, respectively. For juvenile behavioral assessment, mice were exposed to this paradigm between P27 – P31, and adult animals were between P70 – P84.

Object-in-Context Recognition Test (OIC)

Request a detailed protocol

The OIC test is used to measure animal’s ability to distinguish between similar events (Dere et al., 2007; Jain et al., 2012; Ramsaran et al., 2016). This behavioral paradigm was performed in white and grey arenas (30 cm x 30 cm x 30 cm) under dim white-light room illumination (desk lamp pointing to a room wall ~150–200 lux at the arenas levels). During the training phase, animals were placed in an open white chamber wall-patterned with laminated blue squares and allowed to explore two identical objects for 10 min. After 30 min inter-trial interval, mice were placed in a second open grey chamber wall-patterned with laminated white stripes. Animals were allowed to explore for 10 min two identical objects unique from the first trial. Both set of objects were made from building blocks and had distinct surface pattern (one set was smooth and the other softly wrinkled). The objects used in this test had a similar shape to the representation in Figure 3J. After 3 hr, mice were exposed for 10 min to the test trial. Mice from both experimental groups were randomly assigned to either context A or B. During the test trial, one of the objects presented in the sample trials A and B is also presented. The time that each mouse spent exploring the out of context object was compared with the time exploring the object in the familiar context (e.g. object from trial A in context from trial B). Between trials, the arena and objects were properly cleaned with 10 % ethanol. The test session was recorded and manually scored through the behavioral scoring program Kinoscope (Kokras et al., 2017). For juvenile behavioral assessment, mice were exposed to this paradigm at P29, and adult animals were between P70 and P84.

Contextual-fear conditioning (CFC)

Request a detailed protocol

The CFC test was performed in a white acrylic box with internal dimensions of 20 cm wide, 16 cm deep, and 20.5 cm high (MedAssociates). CFC apparatus had a fixed light bulb mounted directly above the chamber to provide a source of illumination. Each box contained a stainless-steel shock grid floor inside a clear acrylic cylinder (10 cm diameter with 15 cm depth), where the animals were placed. All animals were exposed to two probes: a context probe and a cue (light) probe, as previously described (Gu et al., 2012; Mateus-Pinheiro et al., 2017). All probes were recorded, and the freezing behavior was manually scored through Kinoscope (Kokras et al., 2017). For juvenile behavioral assessment, mice were exposed to this paradigm between P31 and P33, and adult animals were between P70 and P84. This behavioral paradigm took 3 days.

Day 1:
Request a detailed protocol

Animals were individually placed in the conditioning-white box (Context A) and received three pairings between a light (20 s) and a co-terminating shock (1 s, » 0.5 mA). The interval between pairings was 180 s, and the first light presentation started 180 s after the beginning of the trial. After the three pairings, mice remained in the acrylic box for 30 s, being after returned to their home cage. Between animals, the apparatus was properly cleaned with 10 % ethanol.

Day 2:
Request a detailed protocol

For the context probe, animals were placed into the same white acrylic chamber (context A), 24 hr after the light-shock pairings. The freezing behavior was monitored for 3 min. Two hours later, animals were introduced into a modified version of the chamber (Context B). This new box was sheeted with a black plasticized cover, sprayed with a vanilla scent. In this way, both contexts had distinct spatial and odor cues. Also in Context B, the ventilation was off, and the experimenter wore a different color of gloves and a lab coat. Freezing behavior was measured for 3 min. The freezing behavior state was defined as the total absence of motion, for a minimum of 1 s.

Day 3:
Request a detailed protocol

For the cue probe, the animals were set in Context B, and individually placed in this chamber 24 hr after the context probe. After 3 min, the light was turned on for 20 s, and the freezing behavior monitored for 1 min after light is turned off.

Morris water maze (MWM)

In the MWM test, several cognitive domains were assessed: working- and spatial-reference memory and behavioral flexibility. Additionally, the strategies used to reach the platform were also analyzed. MWM was performed in a circular white pool (170 cm diameter) filled with water at 22 °C to a depth of 31 cm in a room with and dim light and extrinsic clues (triangle, square, cross, and horizontal stripes). The pool was divided into four quadrants by imaginary lines, and a clear-acrylic cylinder platform (12 cm diameter; 30 cm high), placed in one of the quadrants. All trials were video recorded by a tracking system (Viewpoint, France). For juvenile behavioral assessment, mice were exposed to this paradigm between P29 and P32 and adult animals were between P70 and P84.

Working memory task

Request a detailed protocol

The working memory task (Alves et al., 2017; Cerqueira et al., 2007) evaluates the cognitive domain that relies on the interplay between the hippocampal and prefrontal cortex (PFC) functions. In this task, animals had to learn the position of the hidden platform and to retain this information for four consecutive daily trials. The task was performed for 4 days and in a clockwise manner the platform was repositioned in a new quadrant each day. During the daily trials, animals had different starting positions (north, east, west, and south). Trials ended when mice reached the platform within the time limit of 120 s. If the animals did not reach the platform during the trial time, they were guided to the platform and allowed to stay for 30 s. The time and path to reach the platform were recorded.

Reference memory task

Request a detailed protocol

After working memory evaluation (days 1-4), spatial-reference memory, a hippocampal dependent function, was assessed by keeping the platform in the same quadrant during three consecutive days (days 4-6) (Morris, 1984). During the daily trials, animals had different starting positions (north, east, west, and south). The time and path to reach the platform were recorded for each trial.

Reversal learning task:

Request a detailed protocol

On the last day of MWM testing, reversal-learning performance, a PFC dependent function, was assessed. This was conducted by positioning the platform in a new (opposite) quadrant. Animals were tested in 4 trials. The percentage of time spent in the new and old quadrant containing the platform was used as readout of behavioral flexibility.

Search strategies analysis:

Request a detailed protocol

Throughout the Morris water maze, the mice adopted strategies to reach the hidden platform were evaluated as previously described (Garthe and Kempermann, 2013; Mateus-Pinheiro et al., 2017; Ruediger et al., 2012). Quantitative analyses and strategy classification were completed by assessing different parameters collected through the Viewpoint software: (1) thigmotaxis (Tt): most of the swim distance (>70%) happened within the outer ring area (8 cm from the pool border; (2) random swim (RS): most of the swim distance (>80%) occurred within the inner circular area, and all quadrants were explored with a percentage of swim distance not below 50% for none of the quadrants; non-circular trajectories; (3) scanning (Sc): most of the swim pattern and distance (>80%) happened within the inner circular area, with balanced exploration in all quadrants of the pool; non-circular trajectories, with a percentage (80%), with a balanced exploration of all pool quadrants; swim distance in the platform corridor area 80%); swim distance in the platform corridor area >60%, with shifts in the trajectories directions; (6) focal search (FS): directed trajectories to the platform zone, with swim exploration within the perimeter of the escape platform (30cm); (7) directed swim (DSw): directed trajectories to the hidden platform, without much exploration of the pool. For simplification, we defined two blocks of strategies: Block 1, that comprises the ‘non-hippocampal dependent strategies’ (Tt, RS, and Sc), and Block 2, comprising the defined ‘hippocampal dependent strategies’ (DS, FS, and DSw). These blocks were defined when a sequence of at least three trials within the same block were reached.

Electrophysiological studies

Request a detailed protocol

Electrophysiological recordings were obtained from anesthetized mice (sevoflurane 2.5%; 800 mL/min). Surgical procedure was performed to insert platinum/iridium concentric electrodes (Science Products) in the target positions following the mouse brain atlas (from Paxinos): prelimbic region of the medial prefrontal cortex (mPFC): 1.94 mm anterior to bregma, 0.4 mm lateral to the midline, 2.5 mm below bregma; dorsal hippocampus (dHIP): 1.94 mm posterior to bregma, 1.2 mm lateral to the midline, 1.35 mm below bregma; ventral hippocampus (vHIP): 3.8 mm posterior to bregma, 3.3 mm lateral to the midline, 3.4 mm below bregma. LFP signals obtained from mPFC, dHIP, and vHIP were amplified, filtered (0.1–300 Hz, LP511 Grass Amplifier, Astro-Med), acquired (Micro 1401 mkII, CED) and recorded through the Signal Software (CED). Local field activity was recorded at the sampling rate of 1000 Hz during 100s. After electrophysiological recordings, a biphasic 0.7 mA stimulus was delivered to mark the recording sites. Then, mice were deeply anesthetized with sodium pentobarbital, brains removed, immersed in paraformaldehyde (PFA) 4% for 48 hr and sectioned (50 µm) in a vibratome (Leica VT 1000S, Germany). Coronal slices containing the mPFC, dHip vHip were stained for Cresyl Violet to check for recording sites. Animals with recording positions outside at least in one of the two regions under study (mPFC and dHip or vHip) were excluded from the analysis. Coherence analysis was based on multi-taper Fourier analysis.

Coherence was calculated by custom-written MATLAB scripts, using the MATLAB toolbox Chronux (http://www.chronux.org; Figure 6—source code 1: Local field potentials analysis between the dHip and mPFC; and Figure 6—figure supplement 1—source code 1: Local field potentials analysis between the vHip and mPFC) (Mitra and Pesaran, 1999 ). Coherence was calculated for each 1 s long segments and their mean was evaluated for all frequencies from 1 to 90 Hz. The power spectral density (PSD) of each channel was calculated through the 10 x log of the multiplication between the complex Fourier Transform of each 1s long data segment and its complex conjugate. The mean PSD of each channel was evaluated for all frequencies from 1 to 90 Hz (Oliveira et al., 2013). Both coherence and PSD measurements were assessed in the following frequencies: delta (1–4 Hz), theta (4–12 Hz), beta (12–20 Hz); low (20–40 Hz), and high gamma (40-90 Hz). BrdU labeling: to assess proliferation of fast-dividing progenitor cells, and its impact on the generation of adult-born neurons, animals from all groups were injected intraperitoneally with the thymidine analogous 5-bromo-2’-deoxyuridine or bromodeoxyuridine (BrdU, 50 mg/kg; Sigma-Aldrich, US) that is incorporated in the DNA during the S-phase. BrdU injections were performed once, at the end of the behavioral assessment, 24 hr prior to occision.

Western blot analysis

Request a detailed protocol

Hippocampal DG of juvenile and adult AP2γ KO mice and WT littermates were carefully macrodissected out after occision. The tissue was weighted and homogenized in RIPA buffer [containing 50 mM Tris HCl, 2 mM EDTA, 250 mM NaCl, 10 % glycerol, 1 mM PMSF protease inhibitors (Roche, Switzerland)] and then sonicated (Sonics & Materials, US) for 2 min. Samples were centrifuged for 25 min at 10,000 rpm and 4 °C. The protein concentration of the supernatant was determined using Bradford assay. Samples with equal amounts of protein, 30 μg, were analyzed using the following primary antibodies: alpha-tubulin (#5168; Sigma, mouse, 1:5000), AP2γ (#31288; goat, 1:500; Abcam, UK), Pax6 (#2237; rabbit, 1:1000; Millipore, US), Sox2 (#7935; mouse, 1:500; Abcam, UK) and Tbr2 (#2283; rabbit, 1:500; Millipore, US). Secondary antibodies were used from BioRad (Anti-mouse, 1:10.000; #1706516; Anti-rabbit, 1:10.000; #1706515, US) and Santa-Cruz Biotechnologies (Anti-goat, 1:7500; #A2216, US). Membranes were developed using SuperSignal west Femto reagent (#34096; ThermoFisher, US) and developed in Sapphire Biomolecular Imager from Azure Biosystems (US). After developing, images were quantified using AzureSpot analysis software (Azure Biosystems, US).

Immunostaining procedures

Request a detailed protocol

All mice were deeply anesthetized and then transcardially perfused with cold 0.9 % NaCl, followed by 4 % paraformaldehyde (PFA). Brains were carefully removed from the skull, postfixed in 4 % PFA, and then cryoprotected in 30 % sucrose solution. The brains were coronally sectioned in the vibratome (Leica VT 1000 S, Germany) with a thickness of 50 mm, extending over the entire length of the hippocampal formation. Coronal sections containing the dorsal hippocampal dentate gyrus (DG) were further stained to assess cell proliferation, quantify the population of neuroblasts, immature neurons and its morphology. Analyses were focused in the dorsal hippocampal DG due to the results obtained in the electrophysiological studies.

For the cell proliferation and neuroblasts population assessment, brain sections containing the dorsal hippocampal DG were double stained for BrdU (#6326; rat, 1:100; Abcam, UK) and doublecortin (DCX; #18723; rabbit, 1:100; Abcam, UK). Appropriate secondary fluorescent antibodies were used (Alexa Fluor 488 Goat Anti-rat, #32731; 1:1000; Invitrogen, US; and Alexa Fluor 568 Goat Anti-rabbit, #11011; 1:1000; Invitrogen, US). For cell nuclei labeling, 4',6-diamidino-2-phenylindole (DAPI, 1:200; Sigma Aldrich) was used. The density of each cell population in the DG was determined by normalizing the number of positive cells by the corresponding area of the region.

For the morphology analysis of immature neurons, brain sections containing the dorsal hippocampal DG were stained for DCX (#sc-8066; goat, 1:500; Santa Cruz Biotechnology, US) and the corresponding secondary antibody was used (Alexa Fluor 594 donkey anti-goat, #A32758, 1:1000; Invitrogen, US), as previously described (Dioli et al., 2019). For cell nuclei labeling, 4',6-diamidino-2-phenylindole (DAPI, 1:200; Sigma Aldrich) was used.

All sections were mounted with anti-fade fluorescence mounting medium (Fluorshield Mounting Medium, #ab104135, Abcam, UK). Analysis and cell counting were performed using a confocal microscope (Olympus FluoViewTM FV1000, Germany) and an optical microscope (Olympus BX51, Germany). The observer was blind to the experimental groups.

Morphological analysis

Request a detailed protocol

Dendritic reconstruction of DCX+ cells was performed by the analysis of confocal stack images (Olympus FluoViewTM FV1000, Germany) by the simple neurite tracer plugin (Dioli et al., 2019; Longair et al., 2011; Schindelin et al., 2012) using the open-access Fiji software, as previously described (Dioli et al., 2019; Valero et al., 2014). For a better dendrite segmentation, the colored confocal images were converted to black and grey. In this study, only DCX+ cells that branch into the granular cell layer (GCL) and reach the molecular layer (ML) were studied. Sholl analysis was also assessed using for that the ‘Sholl analysis plugin (http://fiji.sc/Sholl_Analysis)’ and based on the quantification of the number of intersections between dendrites and the surface of circles with a radius increment of 20 μm. The dendritic length and the neuronal complexity were analyzed with the Skeletonize (https://imagej.net/Skeletonize3D) and AnalyzeSkeleton (https://imagej.net/AnalyzeSkeleton). The experimenter was blind to the animal’s genotype.

3D dendritic morphology of pre-existing granule neurons in the DG of juvenile and adult mice was performed through the Golgi-Cox impregnation technique. Briefly, brains were immersed in Golgi-Cox solution for 14 days and then transferred to 30 % sucrose solution. Coronal sections (200 μm) were sectioned in a vibratome (Leica VT100S, Germany), collected and then blotted dry onto gelatine-coated microscope slides. Sections containing the dorsal hippocampus were then alkalinized in 18.8 % ammonia, developed in Dektol (Kodak, US), fixed in Kodak Rapid Fix, dehydrated and xylene cleared. Dendritic arborization was analyzed in the DG of WT and AP2γ KO animals (10 neurons per animal).

Data analysis and statistics

Request a detailed protocol

Statistical analysis was performed using Prism v.8 (GraphPad Software, US). Animals were randomly assigned to groups, balanced by genotypes. Sample sizes were determined by power analyses based on previously published studies (Mateus-Pinheiro et al., 2017) and normal distributions were assessed using the Shapiro-Wilk statistical test, taking into account the respective histograms and measures of skewness and kurtosis. For variables that followed the Gaussian distribution within groups, parametric tests were applied, while non-parametric tests were used for discrete variables. To compare the mean values for two groups, a two-tailed independent-sample t-test was applied. For comparisons between two time-points a two-way ANOVA was used. For longitudinal analyses (across days and different trials) a repeated measures ANOVA was used.

For the comparison of categorical variables (strength to grab, limb grasping, and clasping), crosstabulations were performed and the statistical test used was Fisher’s exact test (when Pearson Qui-Squared assumptions were not met).

Data is expressed either as mean ± SEM (standard error of the mean), as median, or as percentage, as stated in the figures’ legends. Statistical significance was set when p < 0.05. Statistical summary presented in Supplementary file 1.

Data availability

All data generated or analysed during this study are included in the manuscript and supporting file; Source Data file has been provided for Figure 1. Source Code files have been provided for Figure 6 and for Figure 6 - Figure supplement 1.

References

    1. Porsolt RD
    2. Bertin A
    3. Jalfre M
    (1977)
    Behavioral despair in mice: a primary screening test for antidepressants
    Archives Internationales de Pharmacodynamie et de Therapie 229:327–336.

Decision letter

  1. D Nora Abrous
    Reviewing Editor; Neurocentre Magendie, INSERM, France
  2. Catherine Dulac
    Senior Editor; Harvard University, United States
  3. Joram Mul
    Reviewer; University of Amsterdam, Netherlands

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Decision letter after peer review:

Thank you for sending your article entitled "Constitutive AP2γ deficiency reduces postnatal hippocampal neurogenesis and induces behavioral deficits in juvenile mice that persist during adulthood" for peer review at eLife. Your article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Catherine Dulac as the Senior Editor.

All reviewers found the paper interesting for scientists in the fields of developmental psychobiology and molecular mechanisms underlying adult hippocampal neurogenesis. The data analysis is rigorous, and the conclusions are justified by the data.

However, additional data from the different ontogenic stages would be helpful to strength the conclusions. In addition, given the authors previous publication on the impact of conditional knockdown of AP2γ (Mateus-Pinheiro et al., Mol Psy 2017), the authors need to emphasise what the novelty of this work is and how it advances their previous findings.

In addition, more information is needed to improve the manuscript. In particular, we recommend to better describe the animal model, the methods and use anova for the analysis of coherence or PSD at different frequencies.

Reviewer #1:

The aim of this study is to examine the impact of AP2γ deficiency on the emergence of behavioral impairments of emotional and cognitive functions. This longitudinal behavioral analysis show that AP2γ deficiency anxiety leads to memory impairments and an anxious-like phenotype during the juvenile period, a phenotype that persists into adulthood. In contrast, depression-like responses remain unchanged. These changes are associated with an alteration in hippocampal neurogenesis and an alteration in adult hippocampal-to-PFC functional connectivity.

Major strength:

Understanding when emotional and cognitive functions emerge during development is a fundamental question with regard to neurodevelopmental disorders. In this study, the authors addressed this question by determining the impact of AP2γ deficiency on the emergence of different behavioral responses. They found some neurobiological correlates for the deficits observed.

Major weakness:

It is unclear whether different batches of animals were used for the behavioral testing or alternatively whether the same batch of animals was tested during two different developmental periods. In this latter case, the impact of training during the juvenile period may have interfered with the behavior of animals tested at adulthood. In addition, it would have been interesting to test (1) anxiety before weaning, (2) spatial learning in a WM or contextual fear conditioning in juvenile/adolescent animals.

Given that the behavioral syndrome is a stable form from the juvenile period into adulthood, the role of adult hippocampal neurogenesis in this context is unclear.

The authors found no differences when assessing DG neuronal characteristics in juvenile and adult mice. It would be interesting to determine whether AP2γ deficiency impacts the dendritic arborisation of Dcx neurons in juvenile and adult mice.

Reviewer #2:

Loureiro-Campos, Dinis Alves et al. investigated impact of a constitutive AP2γ; heterozygous deletion in mice from early postnatal development until adulthood on hippocampal neurogenesis, emotional and cognitive behaviors, and limbic-cortical connectivity. The authors demonstrate that constitutive and heterozygous AP2γ deficiency does not alter early postnatal development but reduces BrdU-positive and BrdU/DCX double-positive cells, but not dendritic length, in the dentate gyrus of the hippocampus both at juvenile and adult periods. The authors also find that constitutive and heterozygous AP2γ deficiency in juvenile mice (P25-31) increases anxiety-like behavior (% distance center zone open-field test), without altering coping behavior in response to an acute stressor (immobility during tail suspension test) or depressive-like behavior (grooming during sucrose splash test), while promoting cognitive deficits (novel object recognition during object recognition test). Constitutive and heterozygous AP2γ deficiency in adult mice (P70-92) increases anxiety-like behavior (% distance center zone open-field test, % time in open arms elevated plus-maze), without altering coping behavior in response to an acute stressor (immobility during forced swim test or tail suspension test). Adult AP2γ+/- also showed altered behavior during the object recognition test (trend for decreased preference to explore novel object location) and deficits in contextual hippocampal-related memory (reduced freezing time during exposure to familiar context during the contextual fear conditioning test). Adult AP2γ+/- also showed impaired behavioral flexibility (reference memory test in Morris water maze) and a delayed switch from non-hippocampal-dependent to hippocampal-dependent strategies to reach the escape platform during consecutive Morris water maze trials. Finally, simultaneous assessment of electrophysiological features of local field potentials (LFPs) in the dorsal hippocampus and medial prefrontal cortex revealed an impaired functional connectivity between these two brain regions in adult AP2γ+/- mice compared to wild-type controls, and decreases in power spectral density, predominantly in the medial prefrontal cortex.

The conclusions of this paper are supported by the data. The manuscript is written clearly, and the figures are organized and informative. The Material and Methods section is very detailed.

Adding individual datapoints to appropriate graphs will increase clarity for the reader. The rationale behind the current study could be described in greater detail; it could, for example, be better explained why constitutive and heterozygous AP2;; deficiency was studied in this paper. A reader would also benefit from additional information how AP2γ+/- mice were exactly generated (e.g. compare with detailed information on experimental animals in Mateus-Pinheiro et al. 2017), and what the information form this paper specifically adds to previous findings form the same group with (mostly) constitutive and homozygous AP2γ deficiency in mice (Mateus-Pinheiro et al. 2017).

The data of this paper add to the knowledge on the role of AP2γ deficiency in mice on hippocampal neurogenesis, emotional and cognitive behaviors, and limbic-cortical connectivity.

Comments for the authors:

I have the following recommendations to improve the clarity of the manuscript.

– Adding individual datapoints to the graph would increase clarity for the reader.

– The rationale behind the current study could be described in greater detail; it could, for example, be better explained why constitutive and heterozygous AP2γ deficiency was studied in this paper.

– The reader would also benefit from detailed information how these mice were exactly generated (e.g. see information on Animals in Mateus-Pinheiro et al. 2017).

– The reader would also benefit what the information from this paper specifically adds to previous findings form the same group with (mostly) constitutive and homozygous AP2γdeficiency in mice (Mateus-Pinheiro et al. 2017)

– The repeated analysis of coherence or PSD at different frequencies within samples with simple Student t-tests feels incorrect. A repeated measures ANOVA might be more appropriate. At the minimum, a multiple comparison correction should be applied.

Reviewer #3:

The data presented build upon the authors previous findings reporting that conditional knockout of AP2γ in adulthood decreases hippocampal neurogenesis and interferes with cognitive performance. Here, the authors report that these impairments also occur following constitutive AP2γ knockdown and are apparent prior to adulthood, namely in the juvenile period. The authors also report that constitutive AP2γ knockdown disrupts dorsal hippocampus-medial prefrontal cortex (mPFC) coherence but does not affect ventral hippocampus-mPFC coherence which opposes the authors previous findings when investigating the impact of conditional AP2γ knockout in adulthood. The experiments are well designed and results are presented clearly. The authors conclusions are justified by the data. However, some weaknesses that require addressing include: correction of typos; provision of additional methodological details; provision of a more explicit statement as to the novelty of the findings of this work; a more detailed discussion including discussion of the lack of ventral hippocampus-mPFC impairments here versus their previous publication and a discussion on why some behavioural effects (novel object recognition) observed in the juvenile period were not sustained into adulthood.

Comments for the authors:

Given the authors previous publication on the impact of conditional knockdown of AP2γ on hippocampal neurogenesis, anxiety-like behaviour and cognition, the authors need to emphasise what the novelty of this work is and how it advances their previous findings. The translational value and opportunity for therapeutic development is not made explicitly clear and could be commented upon.

Introduction

There are some typos that should be fixed:

L49: "Postnatal mice brain" – please change to mouse brain.

L49 Boldrini paper is on human and not mouse hippocampal neurogenesis and thus should be removed from this sentence.

L53 "in a fined and tuned process" – change to a finely tuned process?

Methods

L318 "distinctive at cohorts"- delete "at"?

Why were only males assessed and not females also?

Open field test – please describe lighting conditions and the size of the area designated as the center of the Open Field.

The FST is normally 6 minutes long with immobility time of the last 4 minutes measured. Why do the authors only measure the last 2 minutes of a 5 minute test? Please provide supporting references for this modification.

Neither the tail suspension test nor the forced swim test are tests of "learned helplessness". Please revise the language used to describe these tests and the behaviour they induce. These tests were developed to measure antidepressant-like behaviour.

Since AP2γ regulates hippocampal neurogenesis , it is curious as to why the authors did not use behavioural tests known to be dependent on hippocampal neurogenesis such as pattern separation or antidepressant action in the novelty suppressed feeding test. This limitation and direction for future research should be acknowledged in the discussion.

Please name/describe the objects used in the novel object recognition test.

Why was morphological assessment of neurons restricted to the dorsal dentate gyrus? I am assuming this is based on the dHi/vHi-mPFC coherence findings but this is not made clear.

Results

Figure 1 – please add the time line and sequence of specific behavioural tests into the figure or describe the sequence and experimental days in which behavioural testing was conducted in the main text.

Fig1G – typo on y-axis – should be DCX and not DXC?

For consistency and to also show the declining rate of neurogenesis during increasing age. Please plot panels F and H, G and I using the same maximum y-axis length i.e. 20 and 15 respectively.

Panel E: Please show representative images from both WT and AP2γ+/- mice.

Please explain the large variation in n numbers across the behavioural tests e.g. in figure 2 and 3.

Fear conditioning – The authors report that AP2γ+/- mice freeze less in context A. This might suggest reduced anxiety which contradicts the OF and EPM findings. This should be discussed.

Figure 5 legend should clearly state whether these measurements were taken from adult or juvenile mice.

There seems to be duplication between Figure 1 and suppl Figure 1 whereby BrdU and BrdU/DCX cell counts and dendritic length are shown twice. Please correct.

Title of suppl Figure 5 should state whether these measurements are made in juvenile or adult mice.

Discussion

The authors should discuss why impairments in NOR in juveniles does not persist to adulthood.

The authors should discuss why in this experiment dHi-mPFC but not vHi-mPFC coherence was disrupted, but in their previous manuscript conditional knockout impaired vHi-mPFC coherence.

L276: "conditional deletion of AP2γ in adulthood lead to a less evident effect on emotional behavior, namely in anxiety-like behavior tested in the OF and EPM behavioral tests, when comparing with the constitutive mice model herein presented (Mateus-Pinheiro et al., 2017). This result indicates that constitutive deficiency of AP2γ may exert a longitudinal cumulative impact leading to more severe alterations in behavioral performance of mice". To support this conclusion, the authors should provide empirical data such as percentage difference in time spent in OA of EPM in AP2γ knockdown mice versus controls in both studies or else describe in more specific detail the findings in the previous paper.

[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Constitutive AP2γ deficiency reduces postnatal hippocampal neurogenesis and induces behavioral deficits in juvenile mice that persist during adulthood" for further consideration by eLife. Your revised article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluatino has been overseen by Catherine Dulac as the Senior Editor.

The comments below are intended to highlight the results you have achieved. It is to discuss them from a perspective more focused on the ontogeny of behaviors given the systematic comparison that you have undertaken. Moreover, (i) the new "object recognition task" is not a separation task, which does not take away from its interest, on the contrary, and (ii) the proposed formulation for "behavioral despair" is more adapted.

Reviewer #1:

In the revised version, the authors addressed most of my comments. However, they have added new behavioral data that deserve to be discussed more deeply in the context of ontogeny of emotion and memory.

1. The authors made several additional experiments. However, some of them have not been included in the revised version. In particular, results of the Figure 2 of the rebuttal letter should be added in figure 3 (in h). These results are really important because they show that animals are too immature to navigate through space using reference memory (Schenk et al. 1985). This is consistent with the delayed development of the hippocampus and indicates that deficits observed at adulthood in the WM are NOT linked to neurons born during the juvenile period or before.

2. The authors demonstrated that constitutive and heterozygous deletion of AP2γ has a complex influence on behaviors:

– Depression-like behaviors remain unchanged

– Some deficits induced in juveniles persist in adulthood : anxiety-like behavior, contextual fear memory (which by the way might be linked to enhanced anxiety levels )

– Some deficits observed in juveniles do not persist in adulthood: novel object recognition

– Some deficits are specific of adult: spatial reference memory, novel object location.

The authors should add value to their results by highlighting the complexity of these observations. This complexity should be discussed in term of ontogeny of behaviors, emergence of the different memory systems and different waves of neurogenesis.

3. There is too much emphasis in the discussion on adult neurogenesis. For example the observation that juveniles exhibit some deficits (anxiety like behavior, simple memory operation) is not relevant to adult neurogenesis! In contrast, the late emergence of spatial reference memory and behavioral flexibility might be! (Dupret et al., 2008; Garthe et al., 2009, see also review of Anacker and Hen R, 2017; Abrous et al., 2021).

4. The title do not reflect the results

5. Abstract/introduction : the concluding sentences do not highlight the novelty of the results.

6. Results

• Ln 109 the age of the animals should be indicated here and not later (ln126).

• Ln 126 The sentence is not clear enough: We assessed the morphology of DCX+ cells only in adult animals as during juvenile period (P31), the large amount of DCX+ cells does not allow a proper segmentation of dendrites through this methodology (Figure 2A).

• Ln 150-151: the subtitle does not reflect the results since memory deficits are not systematically observed. Moreover the word cognitive is rather vague (memory is more appropriate).

• The last paragraph of page 6 needs more breath.

• Ln 180: the task used in Figure 3J is not a "behavioral pattern separation" task. In this task, the ability to combine "what" and "where" information is measured (object in context) and mutant mice are impaired in their ability to contextualize the information. Among others, the work of Rasmaran et al., 2016 (Determinants of object-in-context and object-place-context recognition in the developing rat); Barker and Warburton 2011 (When Is the Hippocampus Involved in Recognition Memory). Barker and Warburton 2021 (Putting objects in context: A prefrontal-hippocampal-perirhinal cortex network) might be helpful for the discussion.

• A task measuring "behavioral pattern separation" is not really needed as behavioral flexibility also depends upon adult born neurons (Dupret et al., 2008; Garthe et al., 2009, see also review of Anacker and Hen R, 2017; Abrous et al., 2021).

• Fig3N (and also Fig4L): the authors wrote "No alterations in freezing behavior were observed in context B". These results have not been discussed. The % of freezing are high (40% and 60%) and are certainly higher than that measured before conditioning (please provide the results). Does that mean both groups of animals generalized and did not recognize the context?

• Ln 201 the subtitle does not reflect the results since not all deficits persist (and some appear).

• Ln 222 the 2 contexts are not similar.

• The last paragraph of page 7 needs more breath.

• Ln 218: memory (and not cognitive) performances.

• Ln 227 the authors concluded that "These observations suggest that adult KO mice exhibit deficits in contextual hippocampal", a conclusion that does not fit with the lack of deficits observed in Figure 4H.

• Ln 245 This conclusion does not summarize the results appropriately

7. Methods

• Line 656. Reference memory: please indicate that animals were released for different departure points.

8. Figures

• Figure 2 reading could be improve by aligning:

– Juvenile (P31) hippocampal DG assessment" and "DAPI | DCX"

– Adult ( Age??) hippocampal DG assessment" and "DAPI | DCX"

– DCX+ cells counts and arborization in the adult hippocampal DG : Short dendrites

– DCX+ cells counts and arborization in the adult hippocampal DG : Long dendrites

• Figure 3 G-K : all these results are based on object recognition ("pattern separation "needs to be removed).

• Figure 4 H-J Same comment.

• Figure 3—figure supplement 1: AP2γ constitutive and heterozygous deficiency does not impact sensory motor development.

9. Conclusion

• The discussion focuses too much on adult neurogenesis which is not the critical and new point of this study. The behavioral data and the complex impact of the mutation deserve to be discussed more thoroughly.

Reviewer #2:

The authors have responded to all my comments and suggestions. This has resulted, in my opinion, in a manuscript with more clarity. I have one remaining comment:

Behavior during FST and TST

The authors now indicate behavior during the FST and TST as 'behavioral despair'. I disagree with this anthropomorphic interpretation and would like to refer to the following reviews and opinion pieces:

https://pubmed.ncbi.nlm.nih.gov/33955617/

https://pubmed.ncbi.nlm.nih.gov/33548153/

https://pubmed.ncbi.nlm.nih.gov/30738104/

https://pubmed.ncbi.nlm.nih.gov/27034848/

Therefore, I strongly suggest labeling immobile behavior during the FST and TST, in this manuscript and future manuscripts, as 'adaptive behavior to cope with an inescapable stressor'.

Reviewer #3:

The authors have addressed all of my previous concerns. The addition of the pattern separation test and novelty-suppressed feeding test has really enhanced the manuscript. The manuscript is very interesting, novel and an important contribution to the field. I very much enjoyed reading it.

https://doi.org/10.7554/eLife.70685.sa1

Author response

Reviewer #1:

[…] It is unclear whether different batches of animals were used for the behavioral testing or alternatively whether the same batch of animals was tested during two different developmental periods. In this latter case, the impact of training during the juvenile period may have interfered with the behavior of animals tested at adulthood.

We thank and totally agree with the reviewer’s suggestion to improve and clarify possible interferences related to a putative repeated behavioral assessment at different timepoints. In this study, individual cohorts of mice were used for behavioral characterization at early postnatal period, juvenile or later at adulthood. Our intention was to avoid any possible impact of habituation or learning from the repetition of behavior assessment, as evoked by the reviewer, and prevent a possible long-term impact of behavioral testing during development periods. In the ”Materials and Methods” section, of the revised manuscript, we added this important information to the reader.

In addition, it would have been interesting to test (1) anxiety before weaning, (2) spatial learning in a WM or contextual fear conditioning in juvenile/adolescent animals.

We acknowledge the reviewer for such pertinent suggestions.

1. We agree that the assessment of the anxiety-like state before weaning represents an additional time point of analysis between the developmental milestones and the behavioral assessment during the juvenile period. As such, in consequence of the reviewer’s suggestion, at postnatal day (P)21, wild-type (WT) and AP2γ KO mice were subjected to the open-field test. Results reveal that similarly to the ones observed during the juvenile period, before weaning, AP2γ KO mice show a significant decrease in distance traveled in the center of the open-field, in comparison to WT animals (Author response image 1A). These results reinforce the observation that constitutive and heterozygous deletion of AP2γ promotes an increase of anxiety-like behavior, manifested as early as P21, as shown in Author response image 1. Importantly, AP2γ deletion does not impact the locomotor activity during the juvenile period as denoted by the average velocity in the open-field test (Author response image 1).

Author response image 1
Before weaning, anxiety-like behavior was assessed through the open-field (OF) test.

AP2γ KO mice exhibited a decreased in distance traveled in center, when compared to WT , suggesting an anxious-like phenotype (A). No differences in locomotor activity was detected between groups as denoted by a similar average velocity during the test (B). Data presented as mean ± SEM. Sample Size: OF: nWT = 10; nAP2γ KO = 9. [Student’s t-test].

2. Furthermore, we totally agree with the reviewer’s suggestion to evaluate the impact of constitutive and heterozygous deletion of AP2γ on cognitive performance during the juvenile period, as performed during adulthood. Thus, we performed the Morris water maze (MWM), as also suggested by the editor, and the contextual fear conditioning (CFC) test on WT and AP2γ KO juvenile mice. Results from the CFC reveal that juvenile AP2γ KO mice exhibit a significant decrease in freezing behavior on context probe A when compared to WT animals. No significant alterations in freezing behavior were observed either at context probe B or cue probe. These results are in line with those observed during adulthood, suggesting an early and specific impact of constitutive and heterozygous AP2γ deletion on context-dependent fear memory, observed during the juvenile period and still detected in adult mice. Results of CFC in juvenile mice are now included as figure 3 L-O in the revised manuscript. On the other hand, results from the MWM in juvenile mice were unable to provide further evidence of the impact of AP2γ deletion on cognitive performance, and more particularly on spatial reference memory. Along the days of training, both WT and AP2γ KO mice did not reveal a significant decrease in escape latency to find the hidden platform, suggesting that, at this early age, mice were not able to successfully learn this complex task (Author response image 2A). The results of this cognitive task are presented in (Author response image 2) .

Author response image 2
Cognitive performance of juvenile mice in the MWM test.

(A) Spatial reference memory was assessed as the average escape latency to find a hidden and fixed platform in each test day. Data presented as mean ± SEM. Sample size: MWM: nWT = 10; nAP2γKO = 9. [Repeated measures ANOVA].

Additionally, to assess the performance in a cognitive task particularly dependent on hippocampal neurogenesis (Treves et al., 2008), we subjected both experimental groups (WT and AP2γ KO mice) to the pattern separation paradigm during the juvenile period and adulthood. We observed that both juvenile and adult AP2γ KO mice were unable to distinguish changes of objects between distinct contexts as evidenced by a similar exploration of objects in the familiar and novel context. These observations constitute further evidence of a strong impact of the constitutive and heterozygous deletion of AP2γ on multiple cognitive modalities. Results from the pattern separation task are included as figures 3J-K (juvenile) and 4J (adult) in the revised manuscript.

Given that the behavioral syndrome is a stable form from the juvenile period into adulthood, the role of adult hippocampal neurogenesis in this context is unclear.

We thank the reviewer for this comment. In fact, the results of this study reveal that constitutive and heterozygous deletion of AP2γ promotes a significant impact on hippocampal neurogenesis and on behavior which is already evidenced during the juvenile period. In this context, we agree with the reviewer that AP2γ deletion does not impact solely on adult hippocampal neurogenesis but impacts this neurogenic process early on development. As so, throughout the manuscript namely in the “Results” and “Discussion” sections, we clearly state this perspective.

The authors found no differences when assessing DG neuronal characteristics in juvenile and adult mice. It would be interesting to determine whether AP2γ deficiency impacts the dendritic arborisation of Dcx neurons in juvenile and adult mice.

We acknowledge the pertinent reviewer’s suggestion and we agree that assessing the impact of AP2γ deletion on the morphology of immature DCX+ neurons would represent a more refine measurement than the morphology of pre-existing mature neurons in the dentate gyrus. In this context, we analyzed the dendritic arborization of DCX+ neurons in juvenile and adult mice. We assessed the morphology of short (dendrites extend to the inner molecular layer) and long (dendritic tree reaches the outer molecular layer) DCX+ cells. In adult mice, we observed a significant decrease in the arborization of short DCX+ cells in AP2γ KO mice when compared to WT. In addition to a decrease in the total dendritic length, we found a decreased complexity of short DCX+ cells in AP2γ KO mice denoted by a reduced number of intersections along their dendritic tree. Interestingly, we detected no differences between the dendritic arborization of long DCX+ cells in the DG of adult WT and AP2γ KO mice. These observations suggest that in addition to a decrease in the number of DCX+ cells (Figure 2C and G), in adult mice, AP2γ deficiency delays the maturation of granular neurons but has no impact on their definitive morphology. In juvenile mice, as expected (Kase et al., 2020; Katsimpardi and Lledo, 2018), we found in both groups a larger amount and density of DCX+ cells than at adulthood. In consequence of large overlap of these cells, it was impossible to distinguish individual dendritic trees of DCX+ cells which prevent us to analyze their morphology and understand the impact of AP2γ deletion at this period. The analysis of dendritic arborization of DCX+ cells in the hippocampal DG of adult mice is now presented in figure 2 of the revised version of the manuscript.

Reviewer #2:

[…] I have the following recommendations to improve the clarity of the manuscript.

– Adding individual datapoints to the graph would increase clarity for the reader.

We thank the reviewer for this relevant suggestion to improve the clarity of the manuscript. In the revised version of the figures, we now included individual data points to every graph included in this study.

– The rationale behind the current study could be described in greater detail; it could, for example, be better explained why constitutive and heterozygous AP2γ deficiency was studied in this paper.

– The reader would also benefit what the information from this paper specifically adds to previous findings form the same group with (mostly) constitutive and homozygous AP2γ deficiency in mice (Mateus-Pinheiro et al. 2017)

We acknowledged the reviewer’s comment and suggestion for a better understanding of the study’s aims. In previous studies, we observed that AP2γ transcription factor was critically required for neuronal specialization and development for specific and defined periods, for proper embryonic development of the cerebral cortex (Pinto et al., 2009), and a continuous generation of the neurons in the adult hippocampal dentate gyrus (Mateus-Pinheiro et al., 2017). In the latter, we used conditional AP2γ knockout mice and showed that conditional deletion of AP2γ in the adult mice also triggered hippocampal-dependent cognitive deficits. In this context, it is important to investigate the cumulative impact of AP2γ deficiency since development. For that we used in this study a constitutive and heterozygous mice model of AP2γ deletion (homozygous mice are not viable) unabling us, for the first time, to assess the importance of AP2γ for postnatal neurodevelopment, particularly on the process of hippocampal neurogenesis, and the impact of its deletion to important hippocampal-dependent behaviors. Furthermore, in light of the importance of AP2γ for the proliferation and expansion of a subpopulation (Tbr2+) of transient amplifying progenitors (TAPs), we also intend to reveal the long-term impact of downregulating this cell population since early development. As such, in this study, we applied a longitudinal assessment of postnatal hippocampal neurogenesis and neuronal morphology, regional electrophysiological coherence, and behavioral performance. This extensive and multimodal evaluation of the impact of AP2γ deficiency was performed during the juvenile period and adulthood. Additionally, in the same animal model, we assessed the acquisition of neurodevelopmental milestones early after birth. To understand the cumulative effects of AP2γ deletion, we also intended to determine at which stages of postnatal neurodevelopment AP2γ deficiency was impacting, either at the neurophysiological or/and behavioral level.

Recognizing the importance of frame this study among previous and related publications, in the revised version of the manuscript we made it more clear to the reader.

– The reader would also benefit from detailed information how these mice were exactly generated (e.g. see information on Animals in Mateus-Pinheiro et al. 2017).

We appreciate the suggestion to improve the description of how mice included in this study were generated. In the revised version, in the “Materials and methods” section of the manuscript, it is now described how this mice were generated, obtained and maintained.

– The repeated analysis of coherence or PSD at different frequencies within samples with simple Student t-tests feels incorrect. A repeated measures ANOVA might be more appropriate. At the minimum, a multiple comparison correction should be applied.

We appreciate and agree with the reviewer’s suggestion to change the statistical analysis of the electrophysiological recordings. For the analysis of spectral coherence or power spectral density, we applied, in the revised version of the manuscript, ANOVA Repeated Measures for a more appropriate comparison between groups and the different frequencies.

Reviewer #3:

[…] Given the authors previous publication on the impact of conditional knockdown of AP2γ on hippocampal neurogenesis, anxiety-like behaviour and cognition, the authors need to emphasise what the novelty of this work is and how it advances their previous findings. The translational value and opportunity for therapeutic development is not made explicitly clear and could be commented upon.

We thank the reviewer’s comment and suggestion to frame and contextualize this study and enumerate what it adds to previous publications. In previous studies, we observed that AP2γ transcription factor was critically required for neuronal specialization and development for specific and defined periods, for proper embryonic development of the cerebral cortex (Pinto et al., 2009), and a continuous generation of the neurons in the adult hippocampal dentate gyrus (Mateus-Pinheiro et al., 2017). In the latter, conditional deletion of AP2γ in the adult mice also triggered hippocampal-dependent cognitive deficits. In this context, it is important to investigate the cumulative impact of AP2γ deficiency since development. The constitutive and heterozygous mice model of AP2γ deletion (homozygous mice are not viable) unable us to, for the first time, assess the importance of AP2γ for postnatal neurodevelopment, particularly on the process of hippocampal neurogenesis, and the impact of its deletion to important hippocampal-dependent behaviors. Furthermore, in light of the importance of AP2γ for the proliferation and expansion of a subpopulation (Tbr2+) of transient amplifying progenitors (TAPs), we also intended to reveal the long-term impact of downregulating this population since early development. As such, in this study, we applied a longitudinal assessment of postnatal hippocampal neurogenesis and neuronal morphology, regional electrophysiological coherence, and behavioral performance. This extensive and multimodal evaluation of the impact of AP2γ deficiency was performed during the juvenile period and at adulthood. Additionally, in the same animal model, we assessed the acquisition of neurodevelopmental milestones early after birth. To understand the cumulative effects of AP2γ deletion, we also intended to determine at which stages of postnatal neurodevelopment AP2γ deficiency was impacting, either at the neurophysiological or/and behavioral level. Thus, through this work, we concluded that constitutive deficiency of AP2γ leads to neurogenic and behavioral impairments since the juvenile period until adulthood, and that these deficits do not occur only in adult mice. With the conditional deletion of AP2γ during adulthood from our study in Molecular Psychiatry 2017 (Mateus-Pinheiro et al., 2017), we were not able to understand how deficiencies of AP2γ transcription factor could modulate postnatal neurodevelopment, particularly hippocampal neurogenesis, and hippocampal-dependent behaviors. Hence, the current study adds a temporal window of analysis not yet studied, and hints that in future works some positive modulation of AP2γ could be applied at younger ages, as a therapeutic approach, to revert deficits in hippocampal neurogenesis in the context of neurological disorders.

Recognizing the importance of frame this study among previous and related publications, in the revised version of the manuscript we made it more clear to the reader.

Introduction

There are some typos that should be fixed:

L49: "Postnatal mice brain" – please change to mouse brain.

As suggested by the reviewer, we increased the accuracy of the first sentence of the introduction section changing the terminology to “mouse brain”.

L49 Boldrini paper is on human and not mouse hippocampal neurogenesis and thus should be removed from this sentence.

We acknowledge the reviewer’s suggestion, and in the revised manuscript we removed the reference Boldrini et., 2018 when referring to the neurogenic process in mice.

L53 "in a fined and tuned process" – change to a finely tuned process?

We appreciate the suggestion and we corrected the sentence in the revised version of the manuscript.

Methods

L318 "distinctive at cohorts" – delete "at"?

Following the reviewer’s suggestion, we deleted ‘at’ from the referred sentence.

Why were only males assessed and not females also?

We thank the reviewer’s comment. As referred to in the subsection ‘animals‘ from the ‘Materials and Materials’, all the animals used in this study were male mice. We acknowledge that it would be interesting to further study the impact of AP2γ constitutive and heterozygous deletion in female mice. In fact, we are currently extending our studies using also female animals. However, taking into consideration the extensive timeline and the multiple cohorts of animals already used for this study, it would not be feasible to double the number of mice. Thus, we decided to restrict it to male mice Moreover, aware of the observed results in emotional-related behaviors, and the impact of estrus cycle in anxiety-like phenotype (Chari et al., 2020; Mora et al., 1996), we would need to either confine the study to female mice with synchronized cycles or add a stressful vaginal smears for cytology.

Open field test – please describe lighting conditions and the size of the area designated as the center of the Open Field.

We appreciate the suggestion to improve the description of the experimental condition of this behavioral test. In the revised version of the manuscript we described the lighting conditions and the size of the defined center of the open-field.

The FST is normally 6 minutes long with immobility time of the last 4 minutes measured. Why do the authors only measure the last 2 minutes of a 5 minute test? Please provide supporting references for this modification.

The reviewer is right. In fact the FST in mice is typically performed in a 6 minutes-long protocol. However, we adapted it to a 5 minutes-long protocol, being the first 3 mins of habituation period and the last 2 mins of test trial as previously described (Mateus-Pinheiro et al., Molecular Psychiatry, 2017). Importantly, the results obtained in this FST protocol are in accordance with those observations from TST, where we found no differences in behavioral despair between WT and AP2γ KO mice.

Neither the tail suspension test nor the forced swim test are tests of "learned helplessness". Please revise the language used to describe these tests and the behaviour they induce. These tests were developed to measure antidepressant-like behaviour.

We thank the reviewer’s comment. In fact, we agree that “learned helplessness” is not the most appropriate term to describe the behavior assessed by tail suspension test and forced swimming test. Thus, throughout the revised manuscript, we changed the term “learned helplessness” to “behavioral despair”.

Since AP2γ regulates hippocampal neurogenesis , it is curious as to why the authors did not use behavioural tests known to be dependent on hippocampal neurogenesis such as pattern separation or antidepressant action in the novelty suppressed feeding test. This limitation and direction for future research should be acknowledged in the discussion.

We appreciate and agree with this pertinent question raised by the reviewer. It is crucial to assess the behavioral impact of the constitutive and heterozygous deletion of AP2γ, a key regulator of adult hippocampal neurogenesis, through behavioral tests particularly dependent on this neurogenic process. As such, we performed the novelty suppressed feeding (NSF) test in both juvenile and adult wild-type and AP2γ KO mice. We observed both in juvenile and adult mice an anxious-like phenotype in AP2γ KO mice, denoted by increased latency to the food pellet in comparison to WT animals. Results from NSF are included in the revised version of the manuscript as figures 3C-D (juvenile) and 4D-E (adult). Furthermore, we assessed the performance of juvenile and adult WT and AP2γ KO mice in the pattern separation task. Results from both periods reveal an impaired capacity of AP2γ KO mice to distinguish changes of objects between distinct contexts as evidenced by a similar exploration of objects in the familiar and novel context. Together, these observations constitute further evidence of the impact of constitutive and heterozygous AP2γ deletion on hippocampal neurogenesis and consequent impairment on dependent behaviors. Results from the pattern separation task are now included as figures 3J-K (juvenile) and 4J (adult) in the revised manuscript.

Please name/describe the objects used in the novel object recognition test.

As suggested by the reviewer, in the revised manuscript we added a detailed description of the objects used in the novel object recognition test.

Why was morphological assessment of neurons restricted to the dorsal dentate gyrus? I am assuming this is based on the dHi/vHi-mPFC coherence findings but this is not made clear.

We acknowledge the reviewer to raise this important point. In our study, the assessment of the neuromorphological impact of constitutive and heterozygous deletion of AP2γ was performed in the dorsal hippocampus. As pointed out by the reviewer, a reason for this option is related to the fact that constitutive and heterozygous deletion of AP2γ solely impacts on the mPFC-dorsal hippocampus coherence. Nevertheless, we cannot rule out a possible impact on the morphology of mature granular neurons in the ventral hippocampus. Indeed, it would be of interest in next studies the assessment of putative changes in number and morphology, particularly of DG neurons in the ventral hippocampus.

Results

Figure 1 – please add the time line and sequence of specific behavioural tests into the figure or describe the sequence and experimental days in which behavioural testing was conducted in the main text.

We thank the reviewer’s suggestion for clarity on the sequence of behavioral testing used in this study. In the ”Materials and Methods” section of the revised manuscript, we included the order and day of when the behavioral tests were performed, during juvenile and adult periods.

Fig1G – typo on y-axis – should be DCX and not DXC?

As suggested by the reviewer, we corrected this and other typos along the revised version of the manuscript.

For consistency and to also show the declining rate of neurogenesis during increasing age. Please plot panels F and H, G and I using the same maximum y-axis length i.e. 20 and 15 respectively.

We appreciate the reviewer’s suggestion to increase consistency and enable a comparative analysis of neurogenesis in DG with age. As suggested, the quantification of BrdU+ and BrdU+/DCX+ cells for juvenile and adult periods in the revised figures, is now presented with the same y-axis length.

Panel E: Please show representative images from both WT and AP2γ+/- mice.

We followed the suggestion of the reviewer, and we included in panel E of figure 1, images with representative BrdU and DCX stainings in the hippocampal DG of both experimental groups, WT and AP2γ KO mice.

Please explain the large variation in n numbers across the behavioural tests e.g. in figure 2 and 3.

We acknowledge the reviewer for this observation. We agree with the reviewer that there is a variation in the number of animals that performed the different behavioral tests within each period, juvenile or adulthood. This variation results from our intention to prevent overexposure mice with all behavior tests (a total of 13 different tests). We believed this would impact the normal performance of the animals, leading to habituation and training to behavioral testing, and would represent a cofounding effect for the interpretation of the results. Thus, different cohorts of mice containing both experimental groups WT and AP2γ KO were exposed to a representative portion of behavior tests. Each individual cohort was slightly different in their size. Furthermore, some cohorts of mice had commonly performed one or more behavioral tests, which resulted in the combination of data from these cohorts.

Fear conditioning – The authors report that AP2γ+/- mice freeze less in context A. This might suggest reduced anxiety which contradicts the OF and EPM findings. This should be discussed.

We thank the reviewer’s comment. The assessment of cognitive performance in the contextual fear conditioning in both juvenile (now included in the revised version of the manuscript) and adult mice revealed that after one day of training and an acquired light-shock association, AP2γ KO mice display decreased freezing behavior when exposed to the previous context (A), when compared to WT mice. As the context test is performed after the acquisition of context-dependent fear memory, these observations suggest a specific deficit in contextual hippocampal-related memory. In our interpretation, similar observation from exposure to context A that preceded light-shock association would eventually suggest reduced anxiety-like behavior, which would be in line with observations from OF, EPM and NSF.

Figure 5 legend should clearly state whether these measurements were taken from adult or juvenile mice.

Title of suppl Figure 5 should state whether these measurements are made in juvenile or adult mice.

We thank the reviewer for these relevant observations. In the legend from the electrophysiological assessment (now Figure 6), we added that these observations were obtained from adult mice. Likewise, the same note was included in the title of Figure 6—figure supplement 1.

There seems to be duplication between Figure 1 and suppl Figure 1 whereby BrdU and BrdU/DCX cell counts and dendritic length are shown twice. Please correct.

We agree that in the previous version of the manuscript, there was a duplication of data in figure 1 and supplementary figure 1, which was removed. In the revised version, the quantification of BrdU+ cells and BrdU+/DCX+ cells in the dentate gyrus either from juvenile and adult mice are included in Figure 1, while neuromorphological analysis of DG granular neurons is included in Figure 2—figure supplement 1.

Discussion

The authors should discuss why impairments in NOR in juveniles does not persist to adulthood.

We appreciate the reviewer’s suggestion. In fact, juvenile AP2γ KO mice exhibit a decreased performance in the novel object recognition task, when compared to WT mice, which is not observed when assessed at adulthood. Yet, we believe it is worthy to note that adult AP2γ KO mice present a trend towards significance (p = 0.08) for cognitive deficits in a similar task when challenged to recognize a change in the location of the object. Furthermore, evidence from other cognitive tests assessed both in juvenile and adult mice, pattern separation and contextual fear conditioning, revealed significant differences between experimental groups at both periods, suggesting that constitutive and heterozygous deletion of AP2γ promotes an early and strong impact on cognitive performance.

The authors should discuss why in this experiment dHi-mPFC but not vHi-mPFC coherence was disrupted, but in their previous manuscript conditional knockout impaired vHi-mPFC coherence.

We thank the reviewer for this pertinent observation. In fact, in this study we sought to understand how constitutive and heterozygous deletion of AP2γ affects the functional connectivity between the hippocampus and the medial prefrontal cortex (mPFC). We observed that the spectral coherence between dorsal hippocampus (dHIP) and mPFC is significantly decreased in delta, theta, and beta frequency bands. This impaired interregional coherence was accompanied by a reduced neuronal activity in the mPFC, including in delta, theta, beta, and low gamma frequencies. We found no differences in the connectivity between the ventral-hippocampus (vHip) and the mPFC. Taking into consideration our previously published work (Mateus-Pinheiro et al., 2017), this was an unexpected result, since conditional deletion of AP2γ in adult mice led to coherence impairments between the vHip-to-mPFC. However, we must take into consideration the time specificity on these mice models of AP2γ deletion. The first and previously published study, deletion of AP2γ was specifically performed at adulthood, while in the current study, mice have a constitutive deficiency of AP2γ since embryonic development. The time in which deficiency of the transcription factor AP2γ is induced might lead to different brain circuitry changes/remodelling, altering for example the connectivity between distinct hippocampal regions and the mPFC, and in this way, contributing to distinct functional readouts as shown in our studies. Although we are in both modulating the hippocampal neurogenic process, we have to bear in mind that AP2γ deficiency since embryonic development might induce not accounted differences contributing to these functional readouts, and the other way around. Yet, it is important to note that similar electrophysiological deficits in dHIP-mPFC coherence were observed in a rat model of hippocampal cytogenesis abrogation, which also denote long-term manifestation of emotional and cognitive deficits (Mateus-Pinheiro et al., 2021).

L276: "conditional deletion of AP2γ in adulthood lead to a less evident effect on emotional behavior, namely in anxiety-like behavior tested in the OF and EPM behavioral tests, when comparing with the constitutive mice model herein presented (Mateus-Pinheiro et al., 2017). This result indicates that constitutive deficiency of AP2γ may exert a longitudinal cumulative impact leading to more severe alterations in behavioral performance of mice". To support this conclusion, the authors should provide empirical data such as percentage difference in time spent in OA of EPM in AP2γ knockdown mice versus controls in both studies or else describe in more specific detail the findings in the previous paper.

We thank the reviewer for this pertinent observation. We agree that the referred statement is not supported by the data of the present study. To fully compare with precision the impact of the conditional and the constitutive heterozygous deletion of AP2γ, a more refine experiment would be needed. In this experiment would be necessary to provide the same experimental conditions for behavioral testing and a similar degree of deletion. In this context, in the revised version of the manuscript, we reformulated this inaccurate statement.

References:

Chari T, Griswold S, Andrews NA, Fagiolini M. 2020. The Stage of the Estrus Cycle Is Critical for Interpretation of Female Mouse Social Interaction Behavior. Front Behav Neurosci 14:1–9. doi:10.3389/fnbeh.2020.00113Kase Y, Kase Y, Shimazaki T, Okano H. 2020. Current understanding of adult neurogenesis in the mammalian brain: How does adult neurogenesis decrease with age? Inflamm Regen. doi:10.1186/s41232-020-00122-xKatsimpardi L, Lledo PM. 2018. Regulation of neurogenesis in the adult and aging brain. Curr Opin Neurobiol. doi:10.1016/j.conb.2018.07.006Mateus-Pinheiro A, Alves ND, Patrício P, Machado-Santos AR, Loureiro-Campos E, Silva JM, Sardinha VM, Reis J, Schorle H, Oliveira JF, Ninkovic J, Sousa N, Pinto L. 2017. AP2γ controls adult hippocampal neurogenesis and modulates cognitive, but not anxiety or depressive-like behavior. Mol Psychiatry 22:1725–1734. doi:10.1038/mp.2016.169Mateus-Pinheiro A, Patrício P, Alves ND, Martins-Macedo J, Caetano I, Silveira-Rosa T, Araújo B, Mateus-Pinheiro M, Silva-Correia J, Sardinha VM, Loureiro-Campos E, Rodrigues AJ, Oliveira JF, Bessa JM, Sousa N, Pinto L. 2021. Hippocampal cytogenesis abrogation impairs inter-regional communication between the hippocampus and prefrontal cortex and promotes the time-dependent manifestation of emotional and cognitive deficits. Mol Psychiatry. doi:10.1038/s41380-021-01287-8Mora S, Dussaubat N, Díaz-Véliz G. 1996. Effects of the estrous cycle and ovarian hormones on behavioral indices of anxiety in female rats. Psychoneuroendocrinology 21:609–620. doi:10.1016/S0306-4530(96)00015-7Pinto L, Drechsel D, Schmid M-T, Ninkovic J, Irmler M, Brill MS, Restani L, Gianfranceschi L, Cerri C, Weber SN, Tarabykin V, Baer K, Guillemot F, Beckers J, Zecevic N, Dehay C, Caleo M, Schorle H, Götz M. 2009. AP2γ regulates basal progenitor fate in a region- and layer-specific manner in the developing cortex. Nat Neurosci 12:1229–1237. doi:10.1038/nn.2399Treves A, Tashiro A, Witter MP, Moser EI. 2008. What is the mammalian dentate gyrus good for? Neuroscience 154:1155–1172. doi:10.1016/j.neuroscience.2008.04.073[Editors' note: further revisions were suggested prior to acceptance, as described below.]

Reviewer #1:

In the revised version, the authors addressed most of my comments. However, they have added new behavioral data that deserve to be discussed more deeply in the context of ontogeny of emotion and memory.

1. The authors made several additional experiments. However, some of them have not been included in the revised version. In particular, results of the Figure 2 of the rebuttal letter should be added in figure 3 (in h). These results are really important because they show that animals are too immature to navigate through space using reference memory (Schenk et al. 1985). This is consistent with the delayed development of the hippocampus and indicates that deficits observed at adulthood in the WM are NOT linked to neurons born during the juvenile period or before.

Following the reviewer’s suggestion, we included these results in Figure 3 of the revised manuscript.

2. The authors demonstrated that constitutive and heterozygous deletion of AP2γ has a complex influence on behaviors:

– Depression-like behaviors remain unchanged

– Some deficits induced in juveniles persist in adulthood : anxiety-like behavior, contextual fear memory (which by the way might be linked to enhanced anxiety levels )

– Some deficits observed in juveniles do not persist in adulthood: novel object recognition

– Some deficits are specific of adult: spatial reference memory, novel object location.

The authors should add value to their results by highlighting the complexity of these observations. This complexity should be discussed in term of ontogeny of behaviors, emergence of the different memory systems and different waves of neurogenesis.

We thank the reviewer for raising such a pertinent point. Indeed, constitutive and heterozygous deletion of AP2γ transcription factor leads to complex behavior alterations in juvenile and adult animals. We acknowledge this complexity in the discussion of the revised version of the manuscript.

3. There is too much emphasis in the discussion on adult neurogenesis. For example the observation that juveniles exhibit some deficits (anxiety like behavior, simple memory operation) is not relevant to adult neurogenesis! In contrast, the late emergence of spatial reference memory and behavioral flexibility might be! (Dupret et al., 2008; Garthe et al., 2009, see also review of Anacker and Hen R, 2017; Abrous et al., 2021).

We believe these two pertinent points (2 and 3) raised by the reviewer are interconnected. As so, we decided to address the arguments together.

In the last years, we have seen an increased number of studies focusing on the functional specificity of hippocampal neurogenesis during development (DHN) and adult hippocampal neurogenesis (AHN). The ontogenetic interpretation of hippocampal neurogenesis assigns different functional relevance to DHN (as the neurogenic process that establishes the basic repertoire of adaptable behaviors) and AHN (as the neurogenic process that underpins the adult brain's ability to adapt functional behaviors) (Abrous et al., 2021). This functional dissociation between these two types of hippocampal neurogenesis is very interesting because it allows us to understand how influences in the neurogenic process at different stages, whether during development or adulthood, can lead to distinct impairments in emotional states and cognitive functions.

AP2γ transcription factor plays detrimental roles in early mammalian extraembryonic development and organogenesis, being one of the molecular components regulating the number of upper layer neurons during ontogeny and phylogeny in the developing cortex (Pinto et al., 2009). Also, AP2γ expression persists in the adult hippocampus, acting as a promoter of proliferation and neuronal differentiation (Mateus-Pinheiro et al., 2017). In this context, we sought to comprehend how defects in the neurogenic process since the early development through constitutive and heterozygous deletion of AP2γ impact function and behavior in specific developmental stages: early postnatal, juvenile phase, and adulthood. Our findings revealed that constitutive and heterozygous AP2γ deletion has a significant impact on hippocampal neurogenesis and behavioral dimensions in juvenile and adult mice, without affecting sensory-motor development. We agree with the reviewer that AP2γ deletion impacts neurogenesis earlier in development rather than just adult hippocampal neurogenesis. As so, throughout the manuscript, namely in the Results and Discussion sections, we clearly state this perspective.

Memory and emotional responses are not distinct ontogenetic processes, but they progress from simple to complex capacities (Abrous et al., 2021). We can hypothesize that AP2γ transcription factor deficiencies can alter the ontogeny of the hippocampal neurogenic process, altering the development of behavior complexity associated with the hippocampus. To address this possibility, we would need to trace the early postnatally born hippocampal neurons, as Marie Lods and colleagues previously performed (Lods et al., 2021), and infer their activity throughout the juvenile phase and adulthood. Moreover, we would need to discriminate how AP2γ transcription factor deficiencies distinctly impair DHN and AHN and their functional outputs. In this study, we observed that at both the juvenile phase and during adulthood hippocampal neurogenesis is reduced, being the transcriptional network underlying this neurogenic process impaired. Furthermore, we noticed that adult AP2γ KO mice showed not only reduced hippocampal neurogenesis but also presented delayed maturation of granular neurons. However, we did not follow the activation/inactivation of these granular neurons throughout different behavioral tests, nor did we analyze how AP2γ transcription factor deficiencies could impact dentate gyrus engrams differently during the juvenile phase and adulthood.

Nonetheless, as AP2γ KO mice have decreased hippocampal neurogenesis since early development, the basic repertoire of adaptive behaviors might be already compromised, leading to behavioral deficits as early as the juvenile phase. Impairments in early postnatal neurogenesis induced by deficiencies in AP2γ transcription factor contributes to a hypersensitivity towards aversive stimulus, as uncovered by the presence of an anxious-like phenotype, and memory impairments as revealed by the AP2γ KO mice performance in multiple cognitive tests. Moreover, the adult emergence of behavioral flexibility by AHN (Abrous et al., 2021) and the impaired performance of adult AP2γ KO mice in anxious behavior testing, contextual memory, and cognitive behavior flexibility demonstrate the importance of an intact adult hippocampal neurogenic process for such functional outputs.

Recognizing the reviewer's comments and the fact that some concepts had not yet been fully explored in our work, we further detailed them in the revised manuscript.

4. The title do not reflect the results

We understand the reviewer's concerns about the title, and we revised it to reflect the findings of our study.

5. Abstract/introduction : the concluding sentences do not highlight the novelty of the results.

As implied by the reviewer, we updated both the concluding sentences of the Abstract and Introduction.

6. Results

• Ln 109 the age of the animals should be indicated here and not later (ln126).

Following the reviewer’s suggestion, we added the animals’ age at this point (ln125).

• Ln 126 The sentence is not clear enough: We assessed the morphology of DCX+ cells only in adult animals as during juvenile period (P31), the large amount of DCX+ cells does not allow a proper segmentation of dendrites through this methodology (Figure 2A).

We appreciate the advice to make the results presentation clearer. We added the word "only" in response to the reviewer's suggestion (ln142).

• Ln 150-151: the subtitle does not reflect the results since memory deficits are not systematically observed. Moreover the word cognitive is rather vague (memory is more appropriate).

We understand the reviewer's concern about the subtitle, and we revised it to reflect the findings specifically to these results.

• The last paragraph of page 6 needs more breath.

Following the reviewer’s suggestion, in the revised manuscript, we updated this paragraph.

• Ln 180: the task used in Figure 3J is not a "behavioral pattern separation" task. In this task, the ability to combine "what" and "where" information is measured (object in context) and mutant mice are impaired in their ability to contextualize the information. Among others, the work of Rasmaran et al., 2016 (Determinants of object-in-context and object-place-context recognition in the developing rat); Barker and Warburton 2011 (When Is the Hippocampus Involved in Recognition Memory). Barker and Warburton 2021 (Putting objects in context: A prefrontal-hippocampal-perirhinal cortex network) might be helpful for the discussion.

• A task measuring "behavioral pattern separation" is not really needed as behavioral flexibility also depends upon adult born neurons (Dupret et al., 2008; Garthe et al., 2009, see also review of Anacker and Hen R, 2017; Abrous et al., 2021).

We thank the reviewer for such pertinent correction. In the revised manuscript, we updated this behavioral paradigm nomenclature and reference. Moreover, we further discussed the relevance of these behavioral findings in the context of our work and for the field.

• Fig3N (and also Fig4L): the authors wrote "No alterations in freezing behavior were observed in context B". These results have not been discussed. The % of freezing are high (40% and 60%) and are certainly higher than that measured before conditioning (please provide the results). Does that mean both groups of animals generalized and did not recognize the context?

To address the reviewer's concern, we included the results of the freezing behavior analysis before conditioning (Figure 3N and Figure 4K). As expected, freezing behavior prior to conditioning was low, both during the juvenile phase and adulthood. WT and AP2γ KO mice identically explored the acrylic cylinder in which they were placed, denoted by the low levels of freezing behavior.

Concerning the freezing behavior in Context B, the percentages were lower than those observed in the conditioning context and light-cued probes. However, as referred by the reviewer, the generalization and the misplaced recognition of the new context after the conditioning could be a concerning observation. Nonetheless, the difference in the percentage of freezing of WT mice between Context A and B is of relevancy. WT mice demonstrated a lower average percentage of freezing in Context B during the juvenile and adulthood evaluations. These findings show that WT mice do not generalize and instead recognize Context B as new. Both in the juvenile and adult phases, the freezing percentages in context B were still lower than the ones observed in the conditioning contexts and light-cued probes.

• Ln 201 the subtitle does not reflect the results since not all deficits persist (and some appear).

We understand the reviewer’s concern about the subtitle, and we revised it to reflect the observed findings.

• Ln 222 the 2 contexts are not similar.

The reviewer is correct. In fact, the two contexts are not similar. The different contexts were identical solely in terms of behavioral paradigm (object disposition and assessment). In the revised manuscript, we corrected this suggestion.

• The last paragraph of page 7 needs more breath.

Following the reviewer’s suggestion, in the revised manuscript, we updated this paragraph.

• Ln 218: memory (and not cognitive) performances.

We updated this term in the revised manuscript in response to the reviewer's suggestion.

• Ln 227 the authors concluded that "These observations suggest that adult KO mice exhibit deficits in contextual hippocampal", a conclusion that does not fit with the lack of deficits observed in Figure 4H.

In fact, despite a trend towards a decreased preference to explore the displaced object (Figure 4H), adult AP2γ KO animals present no differences in the preference for the novel object in the object recognition test (ORT) (Figure 4I) when compared to the WT group. This was quite surprising since, at the juvenile window of analysis, this was already an established phenotype.

Memory and emotional responses are not singular ontogenetic processes, but they progress from simple to complex capacities (Abrous et al., 2021). During rodents’ adulthood, it is already well described that animals in normal/healthy conditions can recognize novel objects, locations and discriminate objects in different contexts (Abrous et al., 2021; Ainge and Langston, 2012). Moreover, these abilities emerge early in the juvenile phase (around P17) and reach maturity over adolescence (Ramsaran et al., 2016).

In the behavior assessment at juvenile period performed in this work, we confirmed that the WT mice were able to recognize the novel object and its displacement. In these ORT tasks, WT animals displayed an average exploration of the displaced and novel object as high as 80%, with AP2γ KO mice showing impaired performance in the novel object recognition. During adulthood, in the novel location task, WT mice showed similar percentages of exploration of the displaced object (~80%). However, in the recognition task, the exploration of the novel object dropped under 60%. Nevertheless, results from the Object-in-context (OIC) tasks and the Contextual Fear Conditioning (CFC) test point out that AP2γ constitutive deletion induced impairments in information contextualization. In OIC, both during the juvenile phase and adulthood, WT mice spent a greater proportion of time exploring the out of context object than the object in the familiar environment, whereas AP2γ KO mice showed no preference for object exploration (Figure 3K and Figure 4J). Moreover, in the CFC behavioral paradigm, juvenile and adult AP2γ KO mice exhibited reduced freezing behavior when exposed to the conditioning context where they have previously received foot shocks (Figure 3O and Figure 4L). These results reinforce the idea that deletion of the AP2γ transcription factor induces contextual memory impairments.

We understand the reviewers’ point of view, regarding the lack of deficits in this specific novel object exploration task. However, we believe that with the results obtained in the previous round of revisions in the OIC paradigm, as well as with the addition of the CFC evaluation at the juvenile phase, we were able to corroborate our conclusion. Taking into consideration this pertinent point raised by the reviewer, we will further detail our perspective in the Results section of the manuscript (ln215-217 and ln251-253).

• Ln 245 This conclusion does not summarize the results appropriately

Following the reviewer’s suggestion, in the revised manuscript, we changed the referred paragraph.

7. Methods

• Line 656. Reference memory: please indicate that animals were released for different departure points.

As suggested by the reviewer, we added the information about the release departure points to the reference memory task description.

8. Figures

• Figure 2 reading could be improve by aligning:

– Juvenile (P31) hippocampal DG assessment" and "DAPI | DCX"

– Adult ( Age??) hippocampal DG assessment" and "DAPI | DCX"

– DCX+ cells counts and arborization in the adult hippocampal DG : Short dendrites

– DCX+ cells counts and arborization in the adult hippocampal DG : Long dendrites

• Figure 3 G-K : all these results are based on object recognition ("pattern separation "needs to be removed).

• Figure 4 H-J Same comment.

• Figure 3—figure supplement 1: AP2 constitutive and heterozygous deficiency does not impact sensory motor development.

We thank the reviewer for these suggestions to improve the clarity of the Main and Supplementary Figures. Thus, in the revised version of the manuscript, we followed the indications herein presented.

9. Conclusion

• The discussion focuses too much on adult neurogenesis which is not the critical and new point of this study. The behavioral data and the complex impact of the mutation deserve to be discussed more thoroughly.

We thank the reviewer for the suggestions given throughout the reviewing process. In the revised manuscript, we updated the Discussion and Conclusion considering all indications herein presented.

Reviewer #2:

The authors have responded to all my comments and suggestions. This has resulted, in my opinion, in a manuscript with more clarity. I have one remaining comment:

Behavior during FST and TST

The authors now indicate behavior during the FST and TST as 'behavioral despair'. I disagree with this anthropomorphic interpretation and would like to refer to the following reviews and opinion pieces:

https://pubmed.ncbi.nlm.nih.gov/33955617/

https://pubmed.ncbi.nlm.nih.gov/33548153/

https://pubmed.ncbi.nlm.nih.gov/30738104/

https://pubmed.ncbi.nlm.nih.gov/27034848/

Therefore, I strongly suggest labeling immobile behavior during the FST and TST, in this manuscript and future manuscripts, as 'adaptive behavior to cope with an inescapable stressor'.

We acknowledge the reviewer for such pertinent suggestions, as well as the references shared with us. We understand this point of view regarding FST and TST and how the analysis of these paradigms has been shifting in the last years (Gorman‐Sandler and Hollis, 2021; Molendijk and de Kloet, 2019; Molendijk and Kloet, 2021). Thus, throughout the revised manuscript, we changed the term “behavioral despair” to either “adaptive behavior to cope with an inescapable stressor” or “coping behavior”. In some contexts, we had to summarize the suggested terminology for a simpler version. Nevertheless, we always referred that this coping mechanism is related to an inescapable stressor (either the water in the FST or the suspension in the TST).

References:

Abrous DN, Koehl M, Lemoine M. 2021. A Baldwin interpretation of adult hippocampal neurogenesis: from functional relevance to physiopathology. Mol Psychiatry 1–20. doi:10.1038/s41380-021-01172-4

Ainge JA, Langston RF. 2012. Ontogeny of neural circuits underlying spatial memory in the rat. Front Neural Circuits 6:1–10. doi:10.3389/fncir.2012.00008

Gorman‐Sandler E, Hollis F. 2021. The forced swim test: Giving up on behavioral despair. Eur J Neurosci ejn.15270. doi:10.1111/ejn.15270

Lods M, Pacary E, Mazier W, Farrugia F, Mortessagne P, Masachs N, Charrier V, Massa F, Cota D, Ferreira G, Abrous DN, Tronel S. 2021. Adult-born neurons immature during learning are necessary for remote memory reconsolidation in rats. Nat Commun 12:1778. doi:10.1038/s41467-021-22069-4

Mateus-Pinheiro A, Alves ND, Patrício P, Machado-Santos AR, Loureiro-Campos E, Silva JM, Sardinha VM, Reis J, Schorle H, Oliveira JF, Ninkovic J, Sousa N, Pinto L. 2017. AP2γ controls adult hippocampal neurogenesis and modulates cognitive, but not anxiety or depressive-like behavior. Mol Psychiatry 22:1725–1734. doi:10.1038/mp.2016.169

Molendijk ML, de Kloet ER. 2019. Coping with the forced swim stressor: Current state-of-the-art. Behav Brain Res 364:1–10. doi:10.1016/j.bbr.2019.02.005

Molendijk ML, Kloet ER. 2021. Forced swim stressor: Trends in usage and mechanistic consideration. Eur J Neurosci ejn.15139. doi:10.1111/ejn.15139

Pinto L, Drechsel D, Schmid M-T, Ninkovic J, Irmler M, Brill MS, Restani L, Gianfranceschi L, Cerri C, Weber SN, Tarabykin V, Baer K, Guillemot F, Beckers J, Zecevic N, Dehay C, Caleo M, Schorle H, Götz M. 2009. AP2γ regulates basal progenitor fate in a region- and layer-specific manner in the developing cortex. Nat Neurosci 12:1229–1237. doi:10.1038/nn.2399

Ramsaran AI, Westbrook SR, Stanton ME. 2016. Ontogeny of object-in-context recognition in the rat. Behav Brain Res 298:37–47. doi:10.1016/j.bbr.2015.04.011

https://doi.org/10.7554/eLife.70685.sa2

Article and author information

Author details

  1. Eduardo Loureiro-Campos

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Validation, Visualization, Writing – original draft, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5834-5851
  2. António Mateus-Pinheiro

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Formal analysis, Methodology
    Competing interests
    No competing interests declared
  3. Patrícia Patrício

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Formal analysis, Methodology, Writing – review and editing
    Competing interests
    No competing interests declared
  4. Carina Soares-Cunha

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Investigation, Methodology, Writing – review and editing
    Competing interests
    No competing interests declared
  5. Joana Silva

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
  6. Vanessa Morais Sardinha

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  7. Bárbara Mendes-Pinheiro

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Formal analysis, Investigation, Methodology
    Competing interests
    No competing interests declared
  8. Tiago Silveira-Rosa

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-6159-1623
  9. Ana Verónica Domingues

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Methodology
    Competing interests
    No competing interests declared
  10. Ana João Rodrigues

    Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    Contribution
    Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-1968-7968
  11. João Oliveira

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    3. IPCA-EST-2Ai, Polytechnic Institute of Cávado and Ave, Applied Artificial Intelligence Laboratory, Campus of IPCA, Barcelos, Portugal
    Contribution
    Formal analysis, Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-1005-2328
  12. Nuno Sousa

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Formal analysis, Funding acquisition, Methodology, Resources, Writing – review and editing
    Competing interests
    No competing interests declared
  13. Nuno Dinis Alves

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Present address
    Department of Psychiatry, Columbia University, NY 10032, New York, United States
    Contribution
    Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Resources, Validation, Visualization, Writing – original draft, Writing – review and editing
    Contributed equally with
    Luísa Pinto
    For correspondence
    nda2114@cumc.columbia.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8712-3710
  14. Luísa Pinto

    1. Life and Health Sciences Research Institute (ICVS), School of Medicine, University of Minho, Braga, Portugal
    2. ICVS/3B’s -PT Government Associate Laboratory, Guimarães, Portugal
    Contribution
    Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review and editing
    Contributed equally with
    Nuno Dinis Alves
    For correspondence
    luisapinto@med.uminho.pt
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-7724-0446

Funding

Fundação para a Ciência e a Tecnologia (SFRH/BD/131278/2017)

  • Eduardo Loureiro-Campos

Fundação para a Ciência e a Tecnologia (SRFH/BD/120124/2016)

  • Bárbara Mendes-Pinheiro

Fundação para a Ciência e a Tecnologia (SFRH/BD/135273/2017)

  • Tiago Silveira-Rosa

Fundação para a Ciência e a Tecnologia (SFRH/BD/147066/2019)

  • Ana Verónica Domingues

Fundação para a Ciência e a Tecnologia (CEECIND/03887/2017)

  • Carina Soares-Cunha

Fundação para a Ciência e a Tecnologia (IF/00328/2015)

  • João Oliveira

Fundação para a Ciência e a Tecnologia (2020.02855.CEECIND)

  • Luísa Pinto

Fundação Bial (037/18)

  • João Oliveira

Fundação Bial (427/14)

  • Luísa Pinto

Fundação para a Ciência e a Tecnologia (UIDB/50026/2020 and UIDP/50026/2020)

  • Eduardo Loureiro-Campos
  • António Mateus-Pinheiro
  • Joana Silva
  • Vanessa Morais Sardinha
  • Bárbara Mendes-Pinheiro
  • Tiago Silveira-Rosa
  • Ana Verónica Domingues
  • Ana João Rodrigues
  • João Oliveira
  • Nuno Sousa
  • Luísa Pinto

Fundação para a Ciência e a Tecnologia (PPBI-POCI-01-0145-FEDER-022122)

  • Eduardo Loureiro-Campos
  • António Mateus-Pinheiro
  • Joana Silva
  • Vanessa Morais Sardinha
  • Bárbara Mendes-Pinheiro
  • Tiago Silveira-Rosa
  • Ana Verónica Domingues
  • Ana João Rodrigues
  • João Oliveira
  • Nuno Sousa
  • Luísa Pinto

Fundação para a Ciência e a Tecnologia (IF/01079/2014)

  • João Oliveira

Fundação para a Ciência e a Tecnologia (PTDC/MED-NEU/31417/2017)

  • João Oliveira

La Caixa Foundation (LCF/PR/HR20/52400020) (100010434)

  • Ana João Rodrigues

European Research Council (10100 3187)

  • Ana João Rodrigues

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would like to acknowledge Dr. Hubert Schorle for providing the AP2γ mice line. ELC, AMP, PP, CSC, JS, TSR, BMP, AVD, JFO, NDA, and LP received fellowships from the Portuguese Foundation for Science and Technology (FCT) (IF/00328/2015 to JFO; 2020.02855.CEECIND to LP). This work was funded by FCT (IF/01079/2014, PTDC/MED-NEU/31417/2017 Grant to JFO), BIAL Foundation Grants (037/18 to JFO and 427/14 to LP), “la Caixa” Foundation (ID 100010434 to AJR), under the agreement LCF/PR/HR20/52400020), the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement No 101003187 to AJR), and Nature Research Award for Driving Global Impact - 2019 Brain Sciences (to LP). This was also co-funded by the Life and Health Sciences Research Institute (ICVS), and by FEDER, through the Competitiveness Internationalization Operational Program (POCI), and by National funds, through the Foundation for Science and Technology (FCT) - project UIDB/50026/2020 and UIDP/50026/2020. Moreover, this work has been funded by ICVS Scientific Microscopy Platform, member of the national infrastructure PPBI - Portuguese Platform of Bioimaging PPBI-POCI-01–0145-FEDER-022122; by National funds, through the Foundation for Science and Technology (FCT) - project UIDB/50026/2020 and UIDP/50026/2020.

Ethics

Efforts were made to minimize the number of animals and their suffering. All experimental procedures performed in this work were conducted in accordance with the EU Directive 2010/63/EU and approved by the Portuguese National Authority for animal experimentation, Direção-Geral de Alimentação e Veterinária (DGAV) with the project reference 0420/000/000/2011 (DGAV 4542).

Senior Editor

  1. Catherine Dulac, Harvard University, United States

Reviewing Editor

  1. D Nora Abrous, Neurocentre Magendie, INSERM, France

Reviewer

  1. Joram Mul, University of Amsterdam, Netherlands

Version history

  1. Received: May 26, 2021
  2. Preprint posted: June 8, 2021 (view preprint)
  3. Accepted: December 2, 2021
  4. Accepted Manuscript published: December 3, 2021 (version 1)
  5. Version of Record published: December 24, 2021 (version 2)

Copyright

© 2021, Loureiro-Campos et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,149
    Page views
  • 136
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Eduardo Loureiro-Campos
  2. António Mateus-Pinheiro
  3. Patrícia Patrício
  4. Carina Soares-Cunha
  5. Joana Silva
  6. Vanessa Morais Sardinha
  7. Bárbara Mendes-Pinheiro
  8. Tiago Silveira-Rosa
  9. Ana Verónica Domingues
  10. Ana João Rodrigues
  11. João Oliveira
  12. Nuno Sousa
  13. Nuno Dinis Alves
  14. Luísa Pinto
(2021)
Constitutive deficiency of the neurogenic hippocampal modulator AP2γ promotes anxiety-like behavior and cumulative memory deficits in mice from juvenile to adult periods
eLife 10:e70685.
https://doi.org/10.7554/eLife.70685

Further reading

    1. Chromosomes and Gene Expression
    2. Developmental Biology
    Virginia L Pimmett, Mounia Lagha
    Insight

    Imaging experiments reveal the complex and dynamic nature of the transcriptional hubs associated with Notch signaling.

    1. Cell Biology
    2. Developmental Biology
    Simon Schneider, Andjela Kovacevic ... Hubert Schorle
    Research Article

    Cylicins are testis-specific proteins, which are exclusively expressed during spermiogenesis. In mice and humans, two Cylicins, the gonosomal X-linked Cylicin 1 (Cylc1/CYLC1) and the autosomal Cylicin 2 (Cylc2/CYLC2) genes, have been identified. Cylicins are cytoskeletal proteins with an overall positive charge due to lysine-rich repeats. While Cylicins have been localized in the acrosomal region of round spermatids, they resemble a major component of the calyx within the perinuclear theca at the posterior part of mature sperm nuclei. However, the role of Cylicins during spermiogenesis has not yet been investigated. Here, we applied CRISPR/Cas9-mediated gene editing in zygotes to establish Cylc1- and Cylc2-deficient mouse lines as a model to study the function of these proteins. Cylc1 deficiency resulted in male subfertility, whereas Cylc2-/-, Cylc1-/yCylc2+/-, and Cylc1-/yCylc2-/- males were infertile. Phenotypical characterization revealed that loss of Cylicins prevents proper calyx assembly during spermiogenesis. This results in decreased epididymal sperm counts, impaired shedding of excess cytoplasm, and severe structural malformations, ultimately resulting in impaired sperm motility. Furthermore, exome sequencing identified an infertile man with a hemizygous variant in CYLC1 and a heterozygous variant in CYLC2, displaying morphological abnormalities of the sperm including the absence of the acrosome. Thus, our study highlights the relevance and importance of Cylicins for spermiogenic remodeling and male fertility in human and mouse, and provides the basis for further studies on unraveling the complex molecular interactions between perinuclear theca proteins required during spermiogenesis.