Leonardo Restivo

Hippocampal Neurogenesis Regulates Forgetting During Adulthood and Infancy

Forget It! When examining the relationship between the production of new neurons in the hippocampus and memory, studies have generally first manipulated hippocampal neurogenesis and afterward investigated memory formation and found that new neurons help to encode new memories. However, when investigating how similar manipulations of neurogenesis impact established hippocampus-dependent memories, Akers et al. (p. 598; see the Perspective by Mongiat and Schinder) uncovered a role for neurogenesis in memory clearance. Thus, the continuous addition of new neurons both degrades existing information stored in hippocampal circuits and simultaneously provides substrates for new learning. Addition of new neurons leads to remodeling of hippocampal circuitry and memory degradation. [Also see Perspective by Mongiat and Schinder] Throughout life, new neurons are continuously added to the dentate gyrus. As this continuous addition remodels hippocampal circuits, computational models predict that neurogenesis leads to degradation or forgetting of established memories. Consistent with this, increasing neurogenesis after the formation of a memory was sufficient to induce forgetting in adult mice. By contrast, during infancy, when hippocampal neurogenesis levels are high and freshly generated memories tend to be rapidly forgotten (infantile amnesia), decreasing neurogenesis after memory formation mitigated forgetting. In precocial species, including guinea pigs and degus, most granule cells are generated prenatally. Consistent with reduced levels of postnatal hippocampal neurogenesis, infant guinea pigs and degus did not exhibit forgetting. However, increasing neurogenesis after memory formation induced infantile amnesia in these species.

MEF2 negatively regulates learning-induced structural plasticity and memory formation

Memory formation is thought to be mediated by dendritic-spine growth and restructuring. Myocyte enhancer factor 2 (MEF2) restricts spine growth in vitro, suggesting that this transcription factor negatively regulates the spine remodeling necessary for memory formation. Here we show that memory formation in adult mice was associated with changes in endogenous MEF2 levels and function. Locally and acutely increasing MEF2 function in the dentate gyrus blocked both learning-induced increases in spine density and spatial-memory formation. Increasing MEF2 function in amygdala disrupted fear-memory formation. We rescued MEF2-induced memory disruption by interfering with AMPA receptor endocytosis, suggesting that AMPA receptor trafficking is a key mechanism underlying the effects of MEF2. In contrast, decreasing MEF2 function in dentate gyrus and amygdala facilitated the formation of spatial and fear memory, respectively. These bidirectional effects indicate that MEF2 is a key regulator of plasticity and that relieving the suppressive effects of MEF2-mediated transcription permits memory formation.

Viral-mediated expression of a constitutively active form of CREB in hippocampal neurons increases memory.

Synaptic activity‐dependent phosphorylation of the transcription factor cAMP response element binding protein (CREB) leads to CREB‐dependent gene transcription, a process thought to underlie long‐term hippocampal synaptic plasticity and memory formation. We previously reported that increasing CREB activity in glutamatergic neurons enhances synaptic plasticity and neuronal excitability. Whether these modifications are sufficient to promote hippocampal‐dependent memory formation was not determined. Here, we provide direct evidence that a brief increase in CREB‐dependent transcription in either CA1 or DG neurons, using in vivo viral vectors, is sufficient to boost memory for contextual representations, as tested in the contextual fear conditioning task, without affecting motor, pain, or anxiety behaviors.

The Promnesic Effect of G-protein-Coupled 5-HT4 Receptors Activation Is Mediated by a Potentiation of Learning-Induced Spine Growth in the Mouse Hippocampus

Pharmacological modulation of synaptic efficacy is a prominent target in the identification of promnesic compounds. Here, we report that pretraining administration of the serotonin 5-HT4 receptors (5-HT4Rs) partial agonist SL65.0155 enhances simultaneous olfactory discrimination performance and potentiates learning-induced dendritic spine growth in the mouse hippocampus. SL65.0155 does not affect spine density in the pseudo-trained mice and, by itself, does not promote spine growth. Injecting the 5-HT4 antagonist RS39604 prior to SL65.0155 prevents both the increase in performance and the additional formation of spines, thus confirming the 5-HT4Rs specificity of the observed effects. These findings provide evidence that 5-HT4Rs stimulation selectively increases experience-dependent structural plasticity in learning-activated hippocampal circuits.

Shifting to automatic.

How does the brain acquire and retain complex motor kills? In a well-known zen story, a centipede was asked how he could coordinate all of his numerous feet without stumbling. The centipede said that he had never given it a thought. From that time on, the centipede became unable to move. This story nicely illustrates the idea that many motor skills that we commonly use throughout our life are mostly automatic and not accessible to conscious recall. The benefit of this shift to automation (or habit) is that it frees us from continuously allocating attentional resources to motor sequences, making multi-tasking a possibility. However, habit formation represents only the last stage of the complex process of motor skill learning: the acquisition of motor skills is characterized by an initial phase of rapid improvement of the performance which is then followed by a later phase during which memory becomes more automatic as performance reaches an asymptotic level (Shiffrin and Schneider, 1977). Learning new motor sequences likely depends on a wide network of brain areas, necessarily involving the interaction between sensory and motor systems (see Pennartz et al., 2009 for recent progress). The fine tuning of new motor sequences is achieved through effortful evaluation of sensory information about the outcome of the movements, and with sufficient repetition only sensory feedback is needed to link movement sequences. This two-stage process of motor skill acquisition suggests that the brain networks underlying the expression of motor performance gradually reorganize over time. Indeed, recent evidence indicates that two distinct compartments of the monkey striatum contribute differentially to the early and late stages of procedural learning (Miyachi et al., 2002). The dorsomedial region of the striatum (DMS), which is primarily innervated by association cortices, seems to be preferentially recruited during initial stages of visuo-motor learning (Miyachi et al., 2002). In contrast, the dorsolateral portion (DLS), which is primarily innervated by sensorimotor areas, is more important for the gradual acquisition of automatic behavior (Miyachi et al., 1997; Tang et al., 2007). Building on these observations, Yin et al. (2009) now identify region specific changes in striatal neural activity that map onto different phases of skill learning. Yin and colleagues implanted mice with microelectrode arrays aimed at the DMS and the DLS and trained them in the accelerating rotarod task. The rotarod task requires the acquisition of complex movements that take both time and practice to learn. As with many new skills, performance in this task is characterized by rapid initial improvement on the first day of training, with performance asymptoting after 3 days of training. The authors used this task to finely dissect early and late phases of the skill learning process. In vivo recordings showed a functional dissociation between striatal regions: DMS neurons showed robust rate modulation during early training phase, while DLS showed increased rate modulation during the extended training period. Accordingly, DMS excitotoxic lesions impaired skill learning when performed early during training, but had no effect when performed after extended training. In contrast, DLS lesions affected both early and late phases of training suggesting that DLS and DMS activation is required to acquire the motor skill, while the DMS is disengaged as the skill becomes more automatic. Ex vivo recordings showed that DMS and DLS medium spiny neurons exhibited training phase-related changes in glutamatergic transmission. EPSP slope measured in response to increasing afferent stimulation in the two striatal regions revealed that the average synaptic strength was higher in DMS during early training, but increased in the DLS after extended training. Medium spiny striatal neurons can be segregated into two distinct populations on the basis of their projections: striatonigral neurons, projecting to substantia nigra, and striatopallidal neurons which send their projections to the external globus pallidus. These two populations exhibit distinct physiological properties and express different dopaminergic receptors, with striatonigral and striatopallidal neurons preferentially expressing D1 and D2 receptors, respectively. Using D2-eGFP mice, Yin and colleagues found that D2 expressing striatopallidal neurons located in DLS exhibited a significant increase in synaptic strength in comparison to D1 expressing neurons from the same region when mice underwent extended training. Accordingly, D1 blockade had no effect on performance when injected during the extended training phase, whereas D2 blockade impaired performance at both early and late training phases. Therefore, these data suggest that skill automatization likely involves an increase in synaptic activation of D2 expressing medium spiny neurons in the DLS. Yin and colleagues also noted that D2 expressing neurons have more and stronger inhibitory projections on D1 expressing neurons than the converse (Taverna et al., 2008), a difference which might provide a mechanistic basis for neural competition between striatopallidal and striatonigral circuits during skill learning and subsequent consolidation. The study by Yin and colleagues brings new, compelling evidence that functional reorganization of the internal circuitry of the striatum accompanies different stages of skill learning. This finding is further supported by Kimchi et al. (2009) data showing that progressive change of DLS activity is likely to occur also during learning of an instrumental task. Such dynamic reorganization of brain circuits seems to be a fundamental property of learning systems where initial information is subjected to rapid degradation and new memory traces have to be functionally integrated with previously acquired information (McClelland et al., 1995; Frankland and Bontempi, 2005). The advantage of functional circuit reorganization resides in the capability to integrate new information into several distinct neural contexts, making the new memory accessible in various situations. Moreover, as a consequence of functional embedding of new memories into distinct neuronal networks, memory traces may eventually become less prone to degradation and more durable. What happens to the skill memory once it has been acquired? Recent evidence shows that neural rewiring rapidly occurs in the motor cortex after training mice in the accelerating rotarod (Yang et al., 2009). These data might raise the question whether motor cortex rewiring precedes, co-occurs or is a consequence of synaptic strength changes taking place in the DLS. It is tempting to speculate that skill learning involves a three stage process where initial effortful learning is accomplished by DMS due to its connections with associative prefrontal cortices. Automatization might be subsequently achieved through a strengthening of synapses onto DLS neurons projecting to basal ganglia and thalamic structures which eventually convey information to the motor cortex where the new skill memory is embedded in a more ample, cortical-based, motor repertoire.

Synaptic Adaptations of CA1 Pyramidal Neurons Induced by a Highly Effective Combinational Antidepressant Therapy

BACKGROUND Antidepressants (AD) need to be chronically administered (weeks to months) to provide beneficial effects. Evidence suggests that combined administration of inhibitors of monoamine reuptake and phosphodiesterase type 4 allows a highly effective therapeutic action. Also, this coadministration more rapidly boosts the cyclic adenosine monophosphate (cAMP) pathway, which is normally activated during chronic treatment of single compounds. Little is known, however, about how this augmentation therapy affects the core mechanism of glutamatergic plasticity. We therefore investigated how in vivo combinational subchronic rolipram and imipramine (scRI) treatment affects depressive behavior, cAMP-dependent transcription, and glutamatergic transmission in the hippocampus, a region critically implicated in depression. METHODS Antidepressant properties of scRI were investigated through the forced swim test. Changes in cAMP-dependent transcription and synaptic transmission of CA1 pyramidal cells were explored with green fluorescent protein, enzyme-linked immunosorbent assay, electrophysiology recordings, and Golgi-Cox staining. RESULTS We demonstrate that scRI displays robust antidepressant properties compared with single-drug treatments and increases hippocampal c-Fos expression and brain-derived neurotrophic factor protein levels. These effects are accompanied by a specific increase in alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and N-methyl-D-aspartate receptors in already existing synapses. Finally, both acute and subchronic treatments led to enhancement of long-term potentiation but differently affected spine density and morphology, with scRI administration specifically resulting in a large increase in stubby spines. CONCLUSIONS We conclude that scRI is highly effective in providing antidepressive effects, including the hippocampal transcriptional alterations normally observed with longer single-drug treatments. Furthermore, we identified scRI-induced modifications in glutamatergic transmission that probably underlie the beneficial action of this combinational therapy.

Development of Adult-Generated Cell Connectivity with Excitatory and Inhibitory Cell Populations in the Hippocampus

New neurons are generated continuously in the subgranular zone of the hippocampus and integrate into existing hippocampal circuits throughout adulthood. Although the addition of these new neurons may facilitate the formation of new memories, as they integrate, they provide additional excitatory drive to CA3 pyramidal neurons. During development, to maintain homeostasis, new neurons form preferential contacts with local inhibitory circuits. Using retroviral and transgenic approaches to label adult-generated granule cells, we first asked whether a comparable process occurs in the adult hippocampus in mice. Similar to development, we found that, during adulthood, new neurons form connections with inhibitory cells in the dentate gyrus, hilus, and CA3 regions as they integrate into hippocampal circuits. In particular, en passant bouton and filopodia connections with CA3 interneurons peak when adult-generated dentate granule cells (DGCs) are ∼4 weeks of age, a time point when these cells are most excitable. Consistent with this, optical stimulation of 4-week-old (but not 6- or 8-week-old) adult-generated DGCs strongly activated CA3 interneurons. Finally, we found that CA3 interneurons were activated robustly during learning and that their activity was strongly coupled with activity of 4-week-old (but not older) adult-generated DGCs. These data indicate that, as adult-generated neurons integrate into hippocampal circuits, they transiently form strong anatomical, effective, and functional connections with local inhibitory circuits in CA3. SIGNIFICANCE STATEMENT New neurons are generated continuously in the subgranular zone of the hippocampus and integrate into existing hippocampal circuits throughout adulthood. Understanding how these cells integrate within well formed circuits will increase our knowledge about the basic principles governing circuit assembly in the adult hippocampus. This study uses a combined connectivity analysis (anatomical, functional, and effective) of the output connections of adult-born hippocampal cells to show that, as these cells integrate into hippocampal circuits, they transiently form strong connections with local inhibitory circuits. This transient increase of connectivity may represent an homeostatic process necessary to accommodate changes in the excitation/inhibition balance induced by the addition of these new excitatory cells to the preexisting excitatory hippocampal circuits.

The strain-specific involvement of nucleus accumbens in latent inhibition might depend on differences in processing configural- and cue-based information between C57BL/6 and DBA mice.

Latent inhibition (LI) consists of decreased associative strength between an elemental stimulus (CS: tone) paired with an unconditioned stimulus (US: footshock) following non-reinforced pre-exposure to the tone. In view of the differences shown by C57BL/6 (C57) and DBA/2 (DBA) mice in processing elemental vs. configural stimuli, the present experiments were designed (1) to assess whether these differences were likely to interfere with the capability of each strain to show LI, and (2) to verify the extent to which lesions of the nucleus accumbens, which have been reported to enhance attention towards contextual stimuli under certain conditions, might interfere with the development of LI. C57 and DBA mice with Nacc or sham lesions were given two periods (4 or 7 days) of pre-exposure to a CS (tone) then subjected to two CS-US pairings given on a single day. On the day after, freezing to the tone was examined in each group. Results show that, following the shorter period of pre-exposure, LI developed in sham-lesioned DBA but did not in sham-lesioned C57. Nacc lesions, however, were found (1) to block LI in DBA but (2) to promote LI in C57. After the longer period of pre-exposure LI was observed in both strain and lesion conditions. In general, these results confirm that strain differences in processing the tone as a single elemental cue (DBA) or, alternatively, as a part of a contextual configural stimulus (C57) can interfere with the development of LI. In addition, they indicate that Nacc lesions, that are susceptible to increase attention to the background, might modify the salience of the tone and produce opposite effect on LI according to the strain specialisation to show elemental or configural responding.

Defective Processing of Contextual Information May Be Involved in the Poor Performance of DBA/2 Mice in Spatial Tasks

These experiments examine the influence of context manipulations on radial maze performance in C57BL/6 (C57) and DBA/2 (DBA) mice. Animals from each strain were trained in two distinct contexts—poor cuing vs rich cuing—that were sucessively switched. The results first show that C57 performed better when trained under rich cuing conditions than under poor ones, whereas DBA performed poorly under both conditions. In addition, contextual manipulations were found to produce more drastic effects in C57 than in DBA mice. That is, C57 showed a strong performance decrement following each context shift, whereas DBA mice did not. In particular, the fact that DBA mice performed similarly under rich and poor cuing conditions and also reacted mildly—or did not react—to context shifts suggests a deficit in processing contextual information, which places important constraints on their capability to form spatial representations.

Enhanced procedural learning following beta-amyloid protein (1-42) infusion in the rat.

Wistar rats receiving intracerebroventricular infusion of the &bgr;-amyloid protein (A&bgr;1-42) or of the inactive fragment (A&bgr;1-42) were subjected to the cross-maze task. According to the standard protocol, rats were released from the south arm and trained to collect food at the end of the east arm. After a 5-day training period, they were given a probe trial during which they were released from the north arm and allowed to choose either the east arm (place learning) or the west arm (response learning). Control rats showed predominant place learning whereas all rats receiving (A&bgr;1-42) showed response learning. These data indicate that exposure to (A&bgr;1-42) does not only impair cognitive responding but elicits strong procedural (motor-based) responding.