Chronic alcohol-induced long-lasting working memory deficits are associated with altered histone H3K9 dimethylation in the prefrontal cortex

Introduction

Epigenetic modifications have emerged as key contributors to the enduring behavioral, molecular and epigenetic neuroadaptations during withdrawal from chronic alcohol exposure. The present study investigated the long-term consequences of chronic alcohol exposure on spatial working memory (WM) and associated changes of transcriptionally repressive histone H3 lysine 9 dimethylation (H3K9me2) in the prefrontal cortex (PFC).

Methods

Male C57BL/6 mice were allowed free access to either 12% (v/v) ethanol for 5 months followed by a 3-week abstinence period or water. Spatial WM was assessed through the spontaneous alternation T-maze test. Alcoholic and water mice received daily injections of GABAB agonist baclofen or saline during alcohol fading and early withdrawal. Global levels of histone modifications were determined by immunohistochemistry.

Results

Withdrawal mice displayed WM impairments along with reduced prefrontal H3K9me2 levels, compared to water-drinking mice. The withdrawal-induced decrease of H3K9me2 occurred concomitantly with increased level of permissive H3K9 acetylation (H3K9ac) in the PFC. Baclofen treatment rescued withdrawal-related WM deficits and fully restored prefrontal H3K9me2 and H3K9ac. Alcohol withdrawal induced brain region-specific changes of H3K9me2 and H3K9ac after testing, with significant decreases of both histone marks in the dorsal hippocampus and no changes in the amygdala and dorsal striatum. Furthermore, the magnitude of H3K9me2 in the PFC, but not the hippocampus, significantly and positively correlated with individual WM performances. No correlation was observed between H3K9ac and behavioral performance. Results also indicate that pre-testing intraperitoneal injection of UNC0642, a selective inhibitor of histone methyltransferase G9a responsible for H3K9me2, led to WM impairments in water-drinking and withdrawal-baclofen mice. Collectively, our results demonstrate that alcohol withdrawal induced brain-region specific alterations of H3K9me2 and H3K9ac, an effect that persisted for at least three weeks after cessation of chronic alcohol intake.

Conclusion

The findings suggest a role for long-lasting decreased H3K9me2 specifically in the PFC in the persistent WM impairments related to alcohol withdrawal.

Acute pre-learning stress selectively impairs hippocampus-dependent fear memory consolidation: Behavioral and molecular evidence

Despite compelling evidence that stress or stress-related hormones influence fear memory consolidation processes, the understanding of molecular mechanisms underlying the effects of stress is still fragmentary. The release of corticosterone in response to pre-learning stress exposure has been demonstrated to modulate positively or negatively memory encoding and/or consolidation according to many variables such as stress intensity, the emotional valence of the learned material or the interval between stressful episode and learning experience. Here, we report that contextual but not cued fear memory consolidation was selectively impaired in male mice exposed to a 50 min-period of restraint stress just before the unpaired fear conditioning session. In addition to behavioral impairment, acute stress down-regulated activated/phosphorylated ERK1/2 (pERK1/2) in dorsal hippocampal area CA1 in mice sacrificed 60 min and 9 h after unpaired conditioning. In lateral amygdala, although acute stress by itself increased the level of pERK1/2 it nevertheless blocked the peak of pERK1/2 that was normally observed 15 min after unpaired conditioning. To examine whether stress-induced corticosterone overflow was responsible of these detrimental effects, the corticosterone synthesis inhibitor, metyrapone, was administered 30 min before stress exposure. Metyrapone abrogated the stress-induced contextual fear memory deficits but did not alleviate the effects of stress on pERK1/2 and its downstream target phosphorylated CREB (pCREB) in hippocampus CA1 and lateral amygdala. Collectively, our observations suggest that consolidation of hippocampus-dependent memory and the associated signaling pathway are particularly sensitive to stress. However, behavioral normalization by preventive metyrapone treatment was not accompanied by renormalization of the canonical signaling pathway. A new avenue would be to consider surrogate mechanisms involving proper metyrapone influence on both nongenomic and genomic actions of glucocorticoid receptors.

The histone H3 lysine 9 methyltransferase G9a/GLP complex activity is required for long-term consolidation of spatial memory in mice

The G9a/G9a-like protein (GLP) histone lysine dimethyltransferase complex and downstream histone H3 lysine 9 dimethylation (H3K9me2) repressive mark have recently emerged as key transcriptional regulators of gene expression programs necessary for long-term memory (LTM) formation in the dorsal hippocampus. However, the role for hippocampal G9a/GLP complex in mediating the consolidation of spatial LTM remains largely unknown. Using a water maze competition task in which both dorsal hippocampus-dependent spatial and striatum-dependent cue navigation strategies are effective to solve the maze, we found that pharmacological inhibition of G9a/GLP activity immediately after learning disrupts long-term consolidation of previously learned spatial information in male mice, hence producing cue bias on the competition test performed 24 h later. Importantly, the inhibition of hippocampal G9a/GLP did not disrupt short-term memory retention. Immunohistochemical analyses revealed increases in global levels of permissive histone H3K9 acetylation in the dorsal hippocampus and dorsal striatum at 1 h post-training, which persisted up to 24 h in the hippocampus. Conversely, H3K9me2 levels were either unchanged in the dorsal hippocampus or transiently decreased at 15 min post-training in the dorsal striatum. Finally, the inhibition of G9a/GLP activity further increased global levels of H3K9 acetylation while decreasing H3K9me2 in the hippocampus at 1 h post-training. However, both marks returned to vehicle control levels at 24 h. Together, these findings support the possibility that G9a/GLP in the dorsal hippocampus is required for the transcriptional switch from short-term to long-term spatial memory formation.

Sustained corticosterone rise in the prefrontal cortex is a key factor for chronic stress-induced working memory deficits in mice

Exposure to prolonged, unpredictable stress leads to glucocorticoids-mediated long-lasting neuroendocrine abnormalities associated with emotional and cognitive impairments. Excessive levels of serum glucocorticoids (cortisol in humans, corticosterone in rodents) contribute notably to deficits in working memory (WM), a task which heavily relies on functional interactions between the medial prefrontal cortex (PFC) and the dorsal hippocampus (dHPC). However, it is unknown whether stress-induced increases in plasma corticosterone mirror corticosterone levels in specific brain regions critical for WM. After a 6 week-UCMS exposure, C57BL/6 J male mice exhibited increased anxiety- and depressive-like behaviors when measured one week later and displayed WM impairments timely associated with increased plasma corticosterone response. In chronically stressed mice, basal phosphorylated/activated CREB (pCREB) was markedly increased in the PFC and the CA1 area of the dHPC and WM testing did not elicit any further increase in pCREB in the two regions. Using microdialysis samples from freely-moving mice, we found that WM testing co-occurred with a rapid and sustained increase in corticosterone response in the PFC while there was a late, non-significant rise of corticosterone in the dHPC. The results also show that non-stressed mice injected with corticosterone (2 mg/kg i.p.) before WM testing displayed behavioral and molecular alterations similar to those observed in stressed animals while a pre-WM testing metyrapone injection (35 mg/kg i.p.), a corticosterone synthesis inhibitor, prevented the effects of UCMS exposure. Overall, the abnormal regional increase of corticosterone concentrations mainly in the PFC emerges as a key factor of enduring WM dysfunctions in UCMS-treated animals.

Targeting the Glucocorticoid Receptors During Alcohol Withdrawal to Reduce Protracted Neurocognitive Disorders

Persistent regional glucocorticoid (GC) dysregulation in alcohol-withdrawn subjects emerges as a key factor responsible for protracted molecular and neural alterations associated with long-term cognitive dysfunction. Regional brain concentrations of corticosterone vary independently from plasma concentrations in alcohol-withdrawn subjects, which may account for the treatment of alcohol withdrawal-induced persistent pathology. Thus, from a pharmacological point of view, a main issue remains to determine the relative efficacy of compounds targeting the GC receptors to attenuate or suppress the long-lasting persistence of brain regional GC dysfunctions in abstinent alcoholics, as well as persistent changes of neural plasticity. Data from animal research show that acting directly on GC receptors during the withdrawal period, via selective antagonists, can significantly counteract the development and persistence of cognitive and neural plasticity disorders during protracted abstinence. A critical remaining issue is to better assess the relative long-term efficacy of GC antagonists and other compounds targeting the corticotropic axis activity such as gamma-aminobutyric acid A (GABA_A) and GABA_B agonists. Indeed, benzodiazepines (acting indirectly on GABA_A receptors) and baclofen (agonist of the GABA_B receptor) are the compounds most widely used to reduce alcohol dependence. Clinical and preclinical data suggest that baclofen exerts an effective and more powerful counteracting action on such persistent cognitive and endocrine dysfunctions as compared to diazepam, even though its potential negative effects on memory processes, particularly at high doses, should be better taken into account.

Protocols to Study Declarative Memory Formation in Mice and Humans:Optogenetics and Translational Behavioral Approaches

Declarative memory formation depends on the hippocampus and declines in aging. Two functions of the hippocampus, temporal binding and relational organization (Rawlins and Tsaltas, 1983; Eichenbaum et al., 1992 ; Cohen et al., 1997 ), are known to decline in aging (Leal and Yassa, 2015). However, in the literature distinct procedures have been used to study these two functions. Here, we describe the experimental procedures used to investigate how these two processes are related in the formation of declarative memory and how they are compromised in aging ( Sellami et al., 2017 ). First, we studied temporal binding using a one-trial learning procedure: trace fear conditioning. It is classical Pavlovian conditioning requiring temporal binding since a brief temporal gap separates the conditioned stimulus (CS) and unconditioned stimulus (US) presentations. We combined the trace fear condition procedure with an optogenetic approach, and we showed that the temporal binding relies on dorsal (d)CA1 activity over temporal gaps. Then, we studied the interaction between temporal binding and relational organization in declarative memory formation using a two-phase radial-maze task in mice and its virtual analog in humans. The behavioral procedure comprises an initial learning phase where subjects learned the constant rewarding /no rewarding valence of each arm, followed by a test phase where the reward contingencies among the arms remained unchanged but where the arms were recombined to assess flexibility, a cardinal property of declarative memory. We demonstrated that dCA1-dependent temporal binding is necessary for the development of a relational organization of memories that allows flexible declarative memory expression. Furthermore, in aging, the degradation of declarative memory is due to a reduction of temporal binding capacity that prevents relation organization.

Changes in striatal activity and functional connectivity in a mouse model of Huntington's disease

Hereditary Huntington’s disease (HD) is associated with progressive motor, cognitive and psychiatric symptoms. A primary consequence of the HD mutation is the preferential loss of medium spiny projection cells with relative sparing of local interneurons in the striatum. In addition, among GABAergic striatal projection cells, indirect pathway cells expressing D2 dopamine receptors are lost earlier than direct pathway cells expressing D1 receptors. To test in vivo the functional integrity of direct and indirect pathways as well as interneurons in the striatum of male R6/1 transgenic mice, we assessed their c-Fos expression levels induced by a striatal-dependent cognitive task and compared them with age-matched wild-type littermates. We found a significant increase of c-Fos+ nuclei in the dorsomedial striatum, and this only at 2 months, when our HD mouse model is still pre-motor symptomatic, the increase disappearing with symptom manifestation. Contrary to our expectation, the indirect pathway projection neurons did not undergo any severer changes of c-Fos expression regardless of age in R6/1 mice. We also found a decreased activation of interneurons that express parvalbumin in the dorsomedial striatum at both presymptomatic and symptomatic ages. Finally, analysis of c-Fos expression in extended brain regions involved in the cognitive learning used in our study, demonstrates, throughout ages studied, changes in the functional connectivity between regions in the transgenic mice. Further analysis of the cellular and molecular changes underlying the transient striatal hyperactivity in the HD mice may help to understand the mechanisms involved in the disease onset.

Alcohol withdrawal induces long-lasting spatial working memory impairments: relationship with changes in corticosterone response in the prefrontal cortex

This study intends to determine whether long-lasting glucocorticoids (GCs) dysregulation in the prefrontal cortex (PFC) or the dorsal hippocampus (dHPC) play a causal role in the maintenance of working memory (WM) deficits observed after alcohol withdrawal. Here, we report that C57/BL6 male mice submitted to 6 months alcohol consumption (12 percent v/v) followed by 1 (1W) or 6 weeks (6W) withdrawal periods exhibit WM deficits in a spatial alternation task and an exaggerated corticosterone rise during and after memory testing in the PFC but not the dHPC. In contrast, emotional reactivity evaluated in a plus-maze is altered only in the 1W group. No behavioral alterations are observed in mice still drinking alcohol. To determine the causal role of corticosterone in the withdrawal-associated long-lasting WM deficits, we further show that a single intraperitoneal injection injection of metyrapone (an inhibitor of corticosterone synthesis) 30 minutes before testing, prevents withdrawal-associated WM deficits and reestablishes PFC activity, as assessed by increased phosphorylated C-AMP Response Element-binding protein (CREB) immunoreactivity in withdrawn mice. Finally, we show that intra-PFC blockade of mineralocorticoid receptors by infusion of spironolactone and, to a lesser extent, of GCs receptors by injection of mifepristone reverses the WM deficits induced by withdrawal whereas the same injections into the dHPC do not. Overall, our study evidences that long-lasting GCs dysfunction selectively in the PFC is responsible for the emergence and maintenance of WM impairments after withdrawal and that blocking prefrontal mineralocorticoid receptors receptors restores WM in withdrawn animals.

Different implications of the dorsal and ventral hippocampus on contextual memory retrieval after stress

This study assessed the relative contributions of dorsal (dHPC) and ventral (vHPC) hippocampus regions in mediating the rapid effects of an acute stress on contextual memory retrieval. Indeed, we previously showed that an acute stress (3 electric footschocks; 0.9 mA each) delivered 15 min before the 24 h-test inversed the memory retrieval pattern in a contextual discrimination task. Specifically, mice learned in a four-hole board two successive discriminations (D1 and D2) varying by the color and texture of the floor. Twenty-four hours later, nonstressed animals remembered accurately D1 but not D2 whereas stressed mice showed an opposite memory retrieval pattern, D2 being more accurately remembered than D1. We showed here that, at the time of memory testing in that task, stressed animals exhibited no significant changes neither in pCREB activity nor in the time-course evolution of corticosterone into the vHPC; in contrast, a significant decrease in pCREB activity and a significant increase in corticosterone were observed in the dHPC as compared to nonstressed mice. Moreover, local infusion of the anesthetic lidocaine into the vHPC 15 min before the onset of the stressor did not modify the memory retrieval pattern in nonstress and stress conditions whereas lidocaine infusion into the dHPC induced in nonstressed mice an memory retrieval pattern similar to that observed in stressed animals. The overall set of data shows that memory retrieval in nonstress condition involved primarily the dHPC and that the inversion of memory retrieval pattern after stress is linked to a dHPC but not vHPC dysfunction.

Hippocampal Injections of Oligomeric Amyloid β-peptide (1-42) Induce Selective Working Memory Deficits and Long-lasting Alterations of ERK Signaling Pathway

Increasing evidence suggests that abnormal brain accumulation of soluble rather than aggregated amyloid-β1-42 oligomers (Aβo(1-42)) plays a causal role in Alzheimer's disease (AD). However, as yet, animal's models of AD based on oligomeric amyloid-β1-42 injections in the brain have not investigated their long-lasting impacts on molecular and cognitive functions. In addition, the injections have been most often performed in ventricles, but not in the hippocampus, in spite of the fact that the hippocampus is importantly involved in memory processes and is strongly and precociously affected during the early stages of AD. Thus, in the present study, we investigated the long-lasting impacts of intra-hippocampal injections of oligomeric forms of Aβo(1-42) on working and spatial memory and on the related activation of ERK1/2. Indeed, the extracellular signal-regulated kinase (ERK) which is involved in memory function had been found to be activated by amyloid peptides. We found that repeated bilateral injections (1injection/day over 4 successive days) of oligomeric forms of Aβo(1-42) into the dorsal hippocampus lead to long-lasting impairments in two working memory tasks, these deficits being observed 7 days after the last injection, while spatial memory remained unaffected. Moreover, the working memory deficits were correlated with sustained impairments of ERK1/2 activation in the medial prefrontal cortex (mPFC) and the septum, two brain areas tightly connected with the hippocampus and involved in working memory. Thus, our study is first to evidence that sub-chronic injections of oligomeric forms of Aβo(1-42) into the dorsal hippocampus produces the main sign of cognitive impairments corresponding to the early stages of AD, via long-lasting alterations of an ERK/MAPK pathway in an interconnected brain networks.

Behavioral Neuroadaptation to Alcohol: From Glucocorticoids to Histone Acetylation

A prime mechanism that contributes to the development and maintenance of alcoholism is the dysregulation of the hypothalamic-pituitary-adrenal axis activity and the release of glucocorticoids (cortisol in humans and primates, corticosterone in rodents) from the adrenal glands. In the brain, sustained, local elevation of glucocorticoid concentration even long after cessation of chronic alcohol consumption compromises functional integrity of a circuit, including the prefrontal cortex (PFC), the hippocampus (HPC), and the amygdala (AMG). These structures are implicated in learning and memory processes as well as in orchestrating neuroadaptive responses to stress and anxiety responses. Thus, potentiation of anxiety-related neuroadaptation by alcohol is characterized by an abnormally AMG hyperactivity coupled with a hypofunction of the PFC and the HPC. This review describes research on molecular and epigenetic mechanisms by which alcohol causes distinct region-specific adaptive changes in gene expression patterns and ultimately leads to a variety of cognitive and behavioral impairments on prefrontal- and hippocampal-based tasks. Alcohol-induced neuroadaptations involve the dysregulation of numerous signaling cascades, leading to long-term changes in transcriptional profiles of genes, through the actions of transcription factors such as [cAMP response element-binding protein (CREB)] and chromatin remodeling due to posttranslational modifications of histone proteins. We describe the role of prefrontal-HPC-AMG circuit in mediating the effects of acute and chronic alcohol on learning and memory, and region-specific molecular and epigenetic mechanisms involved in this process. This review first discusses the importance of brain region-specific dysregulation of glucocorticoid concentration in the development of alcohol dependence and describes how persistently increased glucocorticoid levels in PFC may be involved in mediating working memory impairments and neuroadaptive changes during withdrawal from chronic alcohol intake. It then highlights the role of cAMP-PKA-CREB signaling cascade and histone acetylation within the PFC and limbic structures in alcohol-induced anxiety and behavioral impairments, and how an understanding of functional alterations of these pathways might lead to better treatments for neuropsychiatric disorders.

Altered hippocampal information coding and network synchrony in APP-PS1 mice

β-amyloid is hypothesized to harm neural function and cognitive abilities by perturbing synaptic transmission and plasticity in Alzheimer's disease (AD). To assess the impact of this pathology on hippocampal neurons' ability to encode flexibly environmental information across learning, we performed electrophysiological recordings of CA1 hippocampal unit activity in AD transgenic mice as they acquired an action-reward association in a spatially defined environment; the behavioral task enabled the precise timing of discrete and intentional behaviors of the animal. We found that the proportion of behavioral task-sensitive cells in wild-type (WT) mice typically increased, whereas the proportion of place cells decreased with learning. In AD mice, this learning-dependent change of cell-discharge patterns was absent, and cells exhibited similar firings from the beginning to firings attained at the late learning stage in wild-type cells. These inflexible hippocampal representations of task and space throughout learning are accompanied by remarkable alterations of local oscillatory activity in the theta and ultra-fast ripple frequencies as well as learning abilities. The present data offer new insights into the in vivo cellular and network processes by which β-amyloid and other AD mutations may exert its harmful effects to produce cognitive and behavioral impairments in early stage of AD.

Rescuing prefrontal cAMP-CREB pathway reverses working memory deficits during withdrawal from prolonged alcohol exposure

Both human and animal studies indicate that alcohol withdrawal following chronic alcohol consumption (CAC) impairs many of the cognitive functions which rely on the prefrontal cortex (PFC). A candidate signaling cascade contributing to memory deficits during alcohol withdrawal is the protein kinase A (PKA)/cAMP-responsive element binding (CREB) cascade, although the role of PKA/CREB cascade in behavioral and molecular changes during sustained withdrawal period remains largely unknown. We demonstrated that 1 week (1W) or 6 weeks (6W) withdrawal after 6-month CAC impairs working memory (WM) in a T-maze spontaneous alternation task and reduces phosphorylated CREB (pCREB) in the PFC but not the dorsal CA1 region (dCA1) of the hippocampus compared with CAC and water conditions. In contrast, both CAC-unimpaired and withdrawn-impaired mice exhibited decreased pCREB in dCA1 as well as reduced histone H4 acetylation in PFC and dCA1, compared with water controls. Next, we showed that enhancing CREB activity through rolipram administration prior to testing improved WM performance in withdrawn mice but impaired WM function in water mice. In addition, WM improvement correlates positively with increased pCREB level selectively in the PFC of withdrawn mice. Results further indicate that direct infusion of the PKA activator (Sp-cAMPS) into the PFC significantly improves or impairs, respectively, WM performance in withdrawn and water animals. In contrast, Sp-cAMPS had no effect on WM when infused into the dCA1. Collectively, these results provide strong support that dysregulation of PKA/CREB-dependent processes in prefrontal neurons is a critical molecular signature underlying cognitive decline during alcohol withdrawal.

Region- and age-specific patterns of histone acetylation related to spatial and cued learning in the water maze

Epigenetic processes, such as histone acetylation, are critical regulators of learning and memory processes. In the present study, we investigated whether training in either a spatial or a cued water maze task undergoes selective changes of histone H3 and H4 acetylation within the hippocampus and the dorsal striatum of C57BL/6 mice. We also attempted to provide new insights into the relationships between deregulation in histone acetylation and age-associated memory deficits. In young mice, spatial training increased acetylation of histones H3 and H4 selectively in the dorsal hippocampal CA1 region and the dentate gyrus (DG) whereas cued training significantly enhanced acetylation of both histones selectively in the dorsal striatum. Our data also revealed age-related differences in histone acetylation within the hippocampus and striatum according to task demands. Specifically, age-related spatial memory deficits were associated with opposite changes of H3 (increase) and H4 (decrease) acetylation in CA1 and DG. After cued learning, both histone acetylation levels were reduced in the striatum of aged mice compared with corresponding young-adults but remained well above those of cage-controls. Collectively, our findings suggest an important role for histone acetylation in regulating the relative contributions of the hippocampus and striatum to learning spatial and cued memory tasks.

Partial loss in septo-hippocampal cholinergic neurons alters memory-dependent measures of brain connectivity without overt memory deficits

The functional relevance of septo-hippocampal cholinergic (SHC) degeneration to the degradation of hippocampus-dependent declarative memory (DM) in aging and Alzheimer's disease (AD) remains ill-defined. Specifically, selective SHC lesions often fail to induce overt memory impairments in animal models. In spite of apparent normal performance, however, neuronal activity within relevant brain structures might be altered by SHC disruption. We hypothesized that partial SHC degeneration may contribute to functional alterations within memory circuits occurring in aging before DM decline. In young adult mice, we studied the effects of behaviorally ineffective (saporin-induced) SHC lesions - similar in extent to that seen in aged animals - on activity patterns and functional connectivity between three main neural memory systems: the septo-hippocampal system, the striatum and the amygdala that sustain declarative, procedural and emotional memory, respectively. Animals were trained in a radial maze procedure dissociating the human equivalents of relational/DM and non-R/DM expressions in animals. Test-induced Fos activation pattern revealed that the partial SHC lesion significantly altered the brain's functional activities and connectivity (co-activation pattern) despite the absence of overt behavioral deficit. Specifically, hippocampal CA3 hyperactivity and abnormal septo-hippocampo-amygdalar inter-connectivity resemble those observed in aging and prodromal AD. Hence, SHC neurons critically coordinate hippocampal function in concert with extra-hippocampal structures in accordance with specific mnemonic demand. Although partial SHC degeneration is not sufficient to impact DM performance by itself, the connectivity change might predispose the emergence of subsequent DM loss when, due to additional age-related insults, the brain can no longer compensate the holistic imbalance caused by cholinergic loss.

Activating a memory system focuses connectivity toward its central structure

This report investigates in what way functional connectivity may explain how two memory systems that share almost all their structures, can function as separate systems. The first series of experiments was aimed at demonstrating the reliability of our experimental design by showing that acquisition of the spatial version of a water cross-maze task (stimulus-stimulus associations) was impaired by dorsal hippocampal lesions whereas the cue version (stimulus-reinforcement association) was altered by amygdala lesion. Then, we evaluated how these two tasks induce different patterns of connectivity. The connectivity was evaluated by calculating the correlations between the zif-268 immunoreactivity of 22 structures composing the hippocampus and the amygdala systems. We designed a new statistical procedure to demonstrate double dissociations on the basis of brain regional intercorrelations. Our data show that the correlations between the hippocampus and the other structures of the memory system are higher in the place-learning group compared to the cue-learning group, whereas they are enhanced with the amygdala in the latter group compared to the former. This demonstrates that the activation of a memory system consists in the focusing of functional connectivity toward the central structure of the system. This may explain how several memory systems can share the same structures while remaining independent.

Alteration of CREB phosphorylation and spatial memory deficits in aged 129T2/Sv mice

Phosphorylation of cAMP-response element binding protein (CREB) is required for hippocampus-dependent long-term memory formation. The present study was designed to determine whether spatial memory deficits in aged mice were associated with alteration of hippocampal CREB phosphorylation. We examined the temporal pattern of CREB activation in 5-6 and 23-24-month-old 129T2/Sv mice trained on a spatial reference memory task in the water maze. Phosphorylated CREB (pCREB), total CREB (t-CREB) and c-Fos immunoreactivity (ir) were measured at four time points after the end of training. In young mice, pCREB-ir was significantly increased 15 and 60 min after training in the CA1 region and dentate gyrus. In aged mice sacrificed 15 min after training, pCREB-ir in these structures was reduced whereas t-CREB-ir remained unchanged compared to respective young-adults. An age-related reduction of c-Fos-ir also occurred selectively in hippocampal CA1 region. Since reduced pCREB-ir in CA1 from the 15 min-aged group strongly correlated with individual learning performance, we suggest that altered CREB phosphorylation in CA1 may account for spatial memory impairments during normal aging.

Modafinil restores memory performance and neural activity impaired by sleep deprivation in mice

The original aims of our study have been to investigate in sleep-deprived mice, the effects of modafinil administration on spatial working memory, in parallel with the evaluation of neural activity level, as compared to non-sleep-deprived animals. For this purpose, an original sleep deprivation apparatus was developed and validated with continuous electroencephalography recording. Memory performance was evaluated using spontaneous alternation in a T-maze, whereas the neural activity level was estimated by the quantification of the c-Fos protein in various cerebral zones. This study allowed altogether: First, to evidence that a diurnal 10-h sleep deprivation period induced an impairment of spatial working memory. Second, to observe a decrease in c-Fos expression after sleep deprivation followed by a behavioural test, as compared to non-sleep-deprived mice. This impairment in neural activity was evidenced in areas involved in wake-sleep cycle regulation (anterior hypothalamus and supraoptic nucleus), but also in memory (frontal cortex and hippocampus) and emotions (amygdala). Finally, to demonstrate that modafinil 64 mg/kg is able to restore on the one hand memory performance after a 10-h sleep deprivation period, and on the other hand, the neural activity level in the very same brain areas where it was previously impaired by sleep deprivation and cognitive task.

Biphasic ERK1/2 activation in both the hippocampus and amygdala may reveal a system consolidation of contextual fear memory

There is accumulating evidences to suggest that memory consolidation in some conditions involves two waves of neuronal plastic change. Using two fear conditioning procedures in male C57BL/6J mice, we have recently shown that consolidation of the foreground contextual fear memory required two waves of ERK1/2 activation in hippocampal CA1, while consolidation of cue conditioning was only associated with the early phase of activation. The present experiment further showed that this bi-phasic pattern of ERK1/2 activation was not restricted to hippocampal CA1, but could also be observed in other fear memory-related brain areas. The unpaired conditioning procedure (context in foreground) induced two waves of ERK1/2 activation in hippocampal CA1 and CA3, as well as in the LA and BLA nuclei of the amygdala. In contrast, the paired conditioning procedure (context in background) led to a transient early phase only in hippocampal CA1 and LA. In addition, ERK1/2 phosphorylation in the hippocampus was found to correlate with that in the amygdala nuclei specifically after the unpaired procedure. Taken together, our data suggest that the observed biphasic pattern of neuronal plastic events may reflect the interplay between hippocampal and amygdala activity-dependent plasticity critical for the system consolidation of contextual fear memory.

Extracellular hippocampal acetylcholine level controls amygdala function and promotes adaptive conditioned emotional response

Ample data indicate that tone and contextual fear conditioning differentially require the amygdala and the hippocampus. However, mechanisms subserving the adaptive selection among environmental stimuli (discrete tone vs context) of those that best predict an aversive event are still elusive. Because the hippocampal cholinergic neurotransmission is thought to play a critical role in the coordination between different memory systems leading to the selection of appropriate behavioral strategies, we hypothesized that this cholinergic signal may control the competing acquisition of amygdala-mediated tone and contextual conditioning. Using pavlovian fear conditioning in mice, we first show a higher level of hippocampal acetylcholine release and a specific pattern of extracellular signal-regulated kinase 1/2 (ERK1/2) activation within the lateral (LA) and basolateral (BLA) amygdala under conditions in which the context is a better predictor than a discrete tone stimulus. Second, we demonstrate that levels of hippocampal cholinergic neurotransmission are causally related to the patterns of ERK1/2 activation in amygdala nuclei and actually determine the selection among the context or the simple tone the stimulus that best predicts the aversive event. Specifically, decreasing the hippocampal cholinergic signal not only impaired contextual conditioning but also mimicked conditioning to the discrete tone, both in terms of the behavioral outcome and the LA/BLA ERK1/2 activation pattern. Conversely, increasing this cholinergic signal not only disrupted tone conditioning but also promoted contextual fear conditioning. Hence, these findings highlight that hippocampal cholinergic neurotransmission controls amygdala function, thereby leading to the selection of relevant emotional information.

CREB antisense oligodeoxynucleotide administration into the dorsal hippocampal CA3 region impairs long- but not short-term spatial memory in mice

The transcription factor cAMP response-element binding protein (CREB) has a pivotal role in hippocampal synaptic plasticity and hippocampus-dependent long-term memory. We recently demonstrated that the dorsal hippocampal CA3 region is involved in memory consolidation of spatial information tested on a Morris water maze in mice. To test whether activation of CREB in the CA3 region is required for memory consolidation of spatial information, bilaterally cannulated mice were infused 18 h before the beginning of the behavioral training with antisense or control sense CREB oligodeoxynucleotides (ODNs) or buffer. Mice were then subjected to massed training in a spatial version of the water maze and tested for retention 0 or 24 h after the last training session. We showed that CREB antisense ODN-infusion in the CA3 region impaired long-term memory when tested 24 h later but had no effect on spatial acquisition or short-term memory tested immediately after behavioral training. These findings provide evidence that the regionally restricted activation of CREB in the dorsal hippocampal CA3 region is critical for the long-term memory consolidation phase of spatial learning but not for short-term memory.

Extinction of auditory fear conditioning requires MAPK/ERK activation in the basolateral amygdala

Whereas the neuronal substrates underlying the acquisition of auditory fear conditioning have been widely studied, the substrates and mechanisms mediating the acquisition of fear extinction remain largely elusive. Previous reports indicate that consolidation of fear extinction depends on the mitogen-activated protein kinase/extracellular-signal regulated kinase (MAPK/ERK) signalling pathway and on protein synthesis in the medial prefrontal cortex (mPFC). Based on experiments using the fear-potentiated startle paradigm suggesting a role for neuronal plasticity in the basolateral amygdala (BLA) during fear extinction, we directly addressed whether MAPK/ERK signalling in the basolateral amygdala is necessary for the acquisition of fear extinction using conditioned freezing as a read-out. First, we investigated the regional and temporal pattern of MAPK/ERK activation in the BLA following extinction learning in C57Bl/6J mice. Our results indicate that acquisition of extinction is associated with an increase of phosphorylated MAPK/ERK in the BLA. Moreover, we found that inhibition of the MAPK/ERK signalling pathway by intrabasolateral amygdala infusion of the MEK inhibitor, U0126, completely blocks acquisition of extinction. Thus, our results indicate that the MAPK/ERK signalling pathway is required for extinction of auditory fear conditioning in the BLA, and support a role for neuronal plasticity in the BLA during the acquisition of fear extinction.

Foreground contextual fear memory consolidation requires two independent phases of hippocampal ERK/CREB activation

Fear conditioning is a popular model for investigating physiological and cellular mechanisms of memory formation. In this paradigm, a footshock is either systematically associated to a tone (paired conditioning) or is pseudorandomly distributed (unpaired conditioning). In the former procedure, the tone/shock association is acquired, whereas in the latter procedure, the context/shock association will prevail. Animals with chronically implanted recording electrodes show enhanced amplitude of the extracellularly recorded field EPSP in CA1 pyramidal cells for up to 24 h after unpaired, but not paired, fear conditioning. This is paralleled by a differential activation of the ERK/CREB pathway in CA1, which is monophasic in paired conditioning (0-15 min post-conditioning), but biphasic (0-1 h and 9-12 h post-conditioning) in unpaired conditioning as revealed by immunocytochemistry and Western blotting. Intrahippocampal injection of the MEK inhibitor U0126 prior to each phase prevents the activation of both ERK1/2 and CREB after unpaired conditioning. Block of any activation phase leads to memory impairment. We finally reveal that the biphasic activation of ERK/CREB activity is independently regulated, yet both phases are critically required for the consolidation of long-term memories following unpaired fear conditioning. These data provide compelling evidence that CA1 serves different forms of memory by expressing differential cellular mechanisms that are dependent on the training regime.

Effects of age and spatial learning on adenylyl cyclase mRNA expression in the mouse hippocampus

Adenylyl cyclase (AC) subtypes have been implicated in memory processes and synaptic plasticity. In the present study, the effects of aging and learning on Ca2+/calmodulin-stimulable AC1, Ca2+-insensitive AC2 and Ca2+/calcineurin-inhibited AC9 mRNA level were compared in the dorsal hippocampus of young-adult and aged C57BL/6 mice using in situ hybridization. Both AC1 and AC9 mRNA expression were downregulated in aged hippocampus, whereas AC2 mRNA remained unchanged, suggesting differential sensitivities to the aging process. We next examined AC mRNA expression in the hippocampus after spatial learning in the Morris water maze. Acquisition of the spatial task was associated with an increase of AC1 and AC9 mRNA levels in both young-adult and aged groups, suggesting that Ca2+-sensitive ACs are oppositely regulated by aging and learning. However, aged-trained mice had reduced AC1 and AC9, but greater AC2, mRNA levels relative to young-trained mice and age-related learning impairments were correlated with reduced AC1 expression in area CA1. We suggest that reduced levels of hippocampal AC1 mRNA may greatly contribute to age-related defects in spatial memory.

Resistance to extinction is associated with impaired immediate early gene induction in medial prefrontal cortex and amygdala

Extinction of classical fear conditioning is thought to involve activity-dependent potentiation of synaptic transmission in the medial prefrontal cortex (mPFC), resulting in the inhibition of amygdala-dependent fear responses. While many studies have addressed the mechanisms underlying extinction learning, it is unclear what determines whether extinction memory is consolidated or whether spontaneous recovery of the fear response occurs. Here we show, using a combined electrophysiological and immunocytochemical approach, that spontaneous recovery of conditioned fear in mice is associated with a prolonged expression of long-term depression of synaptic transmission in the mPFC and the failure of induction of the immediate-early genesc-Fos and zif268 in the mPFC and the basolateral nucleus of the amygdala. This suggests that coordinated activity-dependent changes in gene expression in the mPFC and the amygdala may underlie the formation of long-term fear extinction memory.

Chronic ethanol consumption restores the age-related decrease in neurogranin mRNA level in the hippocampus of mice

Neurogranin (Ng) is a Ca(2+)-sensitive calmodulin-binding neuron-specific protein that has been implicated in the regulation of numerous post-synaptic signalling pathways. Here, we investigate the effects of 5 months low level ethanol consumption (approximately 20% of total calories intake) on Ng mRNA expression in the brain of adult (approximately 7-8 months) and aged (approximately 21-22 months) mice using in situ hybridization histochemistry. Results showed that ageing was accompanied by a decrease in amounts of mRNA coding for Ng, especially in the hippocampus (approximately 25% of adults) known to play a critical role in higher cognitive functions. Chronic ethanol consumption restored this decline up to pre-senescent (adult) levels without altering Ng mRNA levels in adult mice. On the basis of recent data indicating a central role for Ng in the regulation of hippocampal synaptic plasticity and spatial learning, our results suggest that moderate ethanol consumption might have a beneficial influence on cognitive deterioration during senescence. Such a possibility is in fact congruent with recent follow-up studies conducted in elderly people.

Selective age-related changes in the PKC-sensitive, calmodulin-binding protein, neurogranin, in the mouse brain

Brain ageing is associated with a dysregulation of intracellular calcium (Ca(2+)) homeostasis which leads to deficits in Ca(2+)-dependent signalling pathways and altered neuronal functions. Given the crucial role of neurogranin/RC3 (Ng) in the post-synaptic regulation of Ca(2+) and calmodulin levels, age-dependent changes in the levels of Ng mRNA and protein expression were analysed in 3, 12, 24 and 31-month-old mouse brains. Ageing produced significant decreases in Ng mRNA expression in the dorsal hippocampal subfields, retrosplenial and primary motor cortices, whereas no reliable changes were seen in any other cortical regions examined. Western blot indicated that Ng protein expression was also down-regulated in the ageing mouse brain. Analysis of Ng immunoreactivity in both hippocampal CA1 and retrosplenial areas indicated that Ng protein in aged mice decreased predominantly in the dendritic segments of pyramidal neurones. These data suggest that age-related changes of post-synaptic Ng in selected brain areas, and particularly in hippocampus, may contribute to altered Ca(2+)/calmodulin-signalling pathways and to region-specific impairments of synaptic plasticity and cognitive decline.

Differential regulation of Ca(2+)-calmodulin stimulated and Ca(2+)-insensitive adenylyl cyclase messenger RNA in intact and denervated mouse hippocampus

The Ca(2+)-calmodulin stimulated AC1 and Ca(2+)-insensitive AC2 are major isoforms of adenylyl cyclase, playing an important role in synaptic plasticity in the mammalian brain. We studied the pattern of expression of AC1 and AC2 genes in the hippocampus of C57BL/6 mice. We found that there were differences in their patterns of distribution in the dentate gyrus. AC1 messenger RNA was detected both in the dentate granule cell bodies and the corresponding molecular field whereas AC2 messenger RNA was preferentially distributed in the dentate granule cell layer, suggesting that AC1 and AC2 messenger RNA are differentially regulated in the dentate gyrus. In order to examine the regulation of AC1 and AC2 expression in response to synaptic deafferentation and reinnervation, the distribution patterns of the two AC messenger RNA in the hippocampal fields and the parietal cortex were analysed 2, 5, 9 and 30 days following an unilateral entorhinal cortex lesion. Interestingly, we found significantly reduced levels of AC1 hybridization signal following the lesion whereas the level of AC2 messenger RNA remained unaffected in all lesioned groups. The changes in AC1 messenger RNA were transient, with a maximal reduction at five days postlesion, and were restricted to the granule cell bodies and stratum moleculare of the deafferented dentate gyrus. No significant change in AC1 messenger RNA levels was detected in other hippocampal fields nor for any other postlesion times studied. These findings suggest that, at least in the dentate gyrus, messenger RNA for AC1 and AC2 might be differentially compartmentalized in cell bodies and dendritic fields. The activity-dependent regulation of AC1 messenger RNA levels by afferent synapses may provide an elegant mechanism for achieving a selective local regulation of AC1 protein, close to its site of action.

The role of Ca2+/calmodulin-stimulable adenylyl cyclases as molecular coincidence detectors in memory formation

Evidence from systems as diverse as mollusks, insects and mammals has revealed that adenylyl cyclase, cyclic adenosine 3',5'-monophosphate (cAMP) cascade, cAMP-dependent protein kinases and their substrates are required for the cellular events underlying the short-term and long-term forms of memory. In Aplysia and Drosophila models, the coincident activation of independent paths converge to produce a synergistic activation of Ca2+/calmodulin-stimulable adenylyl cyclase, thereby enhancing the cAMP level that appears as the primary mediator of downstream events that strengthen enduring memory. In mammals, in which long-term memories require hippocampal function, our understanding of the role of adenylyl cyclases is still fragmentary. Of the differently regulated isoforms present in the hippocampus, the susceptibility of type 1 and type 8 to stimulation by the complex Ca2+/calmodulin and their expression in the hippocampus suggest a role for these two isoforms as a molecular coincidence device for hippocampus-related memory function. Here, we review the key features of Ca2+/calmodulin stimulable adenylyl cyclases, as well as the involvement of cAMP-regulated signaling pathway in the processes of learning and memory.

Ca2+-sensitive adenylyl cyclases, key integrators of cellular signalling

The concept of second messenger signalling originated from the discovery of the role of cyclic AMP, although it is now known that cytosolic calcium [Ca2+]i mediates numerous signalling pathways and plays an equally vital role in many cellular events. In the last few years there has been a great deal of interest in the substantial molecular and functional diversity of mammalian adenylyl cyclases (ACs). Although AC was viewed as a generic activity, which was either stimulated or inhibited by stimulatory or inhibitory receptors, respectively, acting via alpha-subunits of trimeric GTP-regulatory proteins, the recent cloning of nine full-length isoforms, which significantly differ in their regulatory properties and tissue distributions, has revealed an unexpected level of complex regulation. In fact, each AC may integrate convergent inputs from many distinct signal-generating pathways. The nine isoforms can be divided into four distinct families, which reflect their distinct patterns of regulation by betagamma subunits of G-proteins, protein kinase C (PKC) and Ca2+. The mechanisms of regulation are often highly synergistic or conditional, suggesting a function of ACs as coincident detectors. Since all nine isoforms can be regulated either directly or indirectly by Ca2+ or PKC, a complex range of responses is possible. The Ca2+ concentration that stimulates the major ACs in brain has been found to inhibit AC activity in a number of peripheral tissues and cell lines. The purpose of this article is to review many of the important aspects about the distinct regulatory properties and cellular distribution of Ca2+-regulated ACs. Indeed, the notion that Ca2+ and cAMP are "synarchic" messengers acting in concert to regulate cellular activity was formally proposed some time ago. Here, we will focus on acute interactions between Ca2+ and cAMP and attempt to understand how AC activities can be regulated by discrete, physiological [Ca2+]i rises in intact cells. All Ca2+-regulated isoforms have characteristic distribution patterns in the brain. Also discussed are emerging insights on the temporal and spatial regulation of Ca2+- and cAMP-regulated pathways which may enable cell stimuli to elicit specific responses.

Dependence of the Ca2+-inhibitable adenylyl cyclase of C6-2B glioma cells on capacitative Ca2+ entry

The ability of adenylyl cyclases to be regulated by physiological transitions in Ca2+ provides a key point for integration of cytosolic Ca2+ concentration ([Ca2+]i) and cAMP signaling. Ca2+-sensitive adenylyl cyclases, whether endogenously or heterologously expressed, require Ca2+ entry for their regulation, rather than Ca2+ release from intracellular stores (Chiono, M., Mahey, R., Tate, G., and Cooper, D. M. F. (1995) J. Biol. Chem. 270, 1149-1155; Fagan, K., Mahey, R., and Cooper, D. M. F. (1996) J. Biol. Chem. 271, 12438-12444). The present study compared the regulation by capacitative Ca2+ entry versus ionophore-mediated Ca2+ entry of an endogenously expressed Ca2+-inhibitable adenylyl cyclase in C6-2B cells. Even in the face of a dramatic [Ca2+]i rise generated by ionophore, Ca2+ entry via capacitative Ca2+ entry channels was solely responsible for the regulation of the adenylyl cyclase. Selective efficacy of BAPTA over equal concentrations of EGTA in blunting the regulation of the cyclase by capacitative Ca2+ entry defined the intimacy between the adenylyl cyclase and the capacitative Ca2+ entry sites. This association could not be impaired by disruption of the cytoskeleton by a variety of strategies. These results not only establish an intimate spatial relationship between an endogenously expressed Ca2+-inhibitable adenylyl cyclase with capacitative Ca2+ entry sites but also provide a physiological role for capacitative Ca2+ entry other than store refilling.

Immunological assessment of the distribution of type VII adenylyl cyclase in brain

The localization of the nine identified isoforms of adenylyl cyclase in brain has been largely based on determination of patterns of mRNA expression. A polyclonal antibody has now been developed that specifically recognizes Type VII adenylyl cyclase. This antibody was used for immunocytochemical analysis of the distribution of Type VII adenylyl cyclase in rat brain. Labeling of Type VII adenylyl cyclase was observed in several areas, including cerebellum, caudate-putamen, nucleus accumbens, hippocampus and cerebral cortex. In some of these areas, the staining of the adenylyl cyclase protein suggested the possibility of presynaptic localization. For example, in situ hybridization showed Type VII adenylyl cyclase mRNA concentrated in cerebellar granule neurons. The cerebellar granule cell layer, however, showed little immunostaining, while punctate immunostaining was observed in the molecular layer. These results suggested that protein synthesized in the granule neurons may be targeted to the neuron terminals. Punctate staining in the caudate-putamen, globus pallidus and nucleus accumbens also suggested the possibility of axonal and/or dendritic localization of Type VII adenylyl cyclase in these regions. Labeling of the soma of cerebellar Purkinje cells, cortical pyramidal and non-pyramidal cells and interneurons in the cerebellum and hippocampus was also observed. Type VII adenylyl cyclase, like the other adenylyl cyclase isoforms, has distinct regulatory characteristics, including sensitivity to stimulation by Gsalpha and G protein betagamma subunits, modulation by protein kinase C, and high sensitivity to stimulation by ethanol. These characteristics, and the discrete localization of this enzyme, may contribute to its ability to provide signal integration and/or control of neurotransmitter release in particular neurons or brain areas.

Ca(2+)-inhibitable adenylyl cyclase and pulmonary microvascular permeability

Intracellular mechanisms responsible for endothelial cell disruption are unknown, although either elevated cytosolic Ca2+ ([Ca2+]i) or decreased adenosine 3',5'-cyclic monophosphate (cAMP) promotes permeability. Recent identification that Ca(2+)-inhibitable adenylyl cyclase establishes an inverse relationship between [Ca2+]i and cAMP in macrovascular endothelial cells provided a possible mechanism of development of permeability. However, these data utilized an in vitro model; lacking was evidence supporting 1) expression of Ca(2+)-inhibitable adenylyl cyclase in pulmonary microvascular endothelium and 2) Ca2+ inhibition of adenylyl cyclase and cAMP content as a paradigm for inflammatory mediator-induced permeability in the intact circulation. We therefore addressed these issues in microvascular endothelial cells derived from rat lung and in an isolated perfused rat lung preparation. Results demonstrate expression of a Ca(2+)-inhibitable adenylyl cyclase in microvascular endothelial cells. Furthermore, data suggest that Ca2+ inhibition of adenylyl cyclase is necessary for development of microvascular permeability in the intact circulation. We conclude Ca2+ inhibition of cAMP represents a critical step in genesis of microvascular permeability in the intact pulmonary circulation.

Regulation of growth hormone secretion by amino acid neurotransmitters in the goldfish (I): Inhibition by N-methyl-D, L-aspartic acid

High levels of the amino acid neurotransmitter glutamate were found in the goldfish hypothalamus and pituitary using high performance liquid chromatography with fluorometric detection. A specific polyclonal antibody to glutamate was generated in the rabbit for immunocytochemistry. Localization studies demonstrated that glutamatergic neurons of undetermined origin innervate the particular part of the goldfish adenohypophysis where somatotrophs and gonadotrophs are located. Intraperitoneal and brain third ventricle injection of the glutamate agonist N-methyl-D,L-aspartic acid (NMA) inhibited GH release in vivo. The gonadal steroid estradiol plays an important role in regulating GH secretion by stimulating basal serum GH levels and enhancing the inhibitory effects of NMA on GH secretion. Taken together, these results demonstrate that glutamate is an important regulator of GH secretion in goldfish.

Adenylate cyclases: critical foci in neuronal signaling

Current findings show that adenylate cyclases comprise a heterogeneous multigene family, members of which are variously regulated by the alpha and beta gamma subunits of G proteins, by Ca2+ and by protein kinases. In the CNS, individual isoforms of adenylate cyclase are expressed discretely in select regions of the brain. At the subcellular level, adenylate cyclases can be concentrated into dendritic spines, thereby increasing their susceptibility to multiple regulatory influences. Altogether, such findings greatly expand knowledge of the potential role of this archetypical signaling system in the modulation of neuronal function.

Immunohistochemical localization of adenylyl cyclase in rat brain indicates a highly selective concentration at synapses

Only three isoforms of adenylyl cyclase (EC 4.6.1.1) mRNAs (AC1, -2, and -5) are expressed at high levels in rat brain. AC1 occurs predominantly in hippocampus and cerebellum, AC5 is restricted to the basal ganglia, whereas AC2 is more widely expressed, but at much lower levels. The distribution and abundance of adenylyl cyclase protein were examined by immunohistochemistry with an antiserum that recognizes a peptide sequence shared by all known mammalian adenylyl cyclase isoforms. The immunoreactivity in striatum and hippocampus could be readily interpreted within the context of previous in situ hybridization studies. However, extending the information that could be gathered by comparisons with in situ hybridization analysis, it was apparent that staining was confined to the neuropil--corresponding to immunoreactive dendrites and axon terminals. Electron microscopy indicated a remarkably selective subcellular distribution of adenylyl cyclase protein. In the CA1 area of the hippocampus, the densest immunoreactivity was seen in postsynaptic densities in dendritic spine heads. Labeled presynaptic axon terminals were also observed, indicating the participation of adenylyl cyclase in the regulation of neurotransmitter release. The selective concentration of adenylyl cyclases at synaptic sites provides morphological data for understanding the pre- and postsynaptic roles of adenylyl cyclase in discrete neuronal circuits in rat brain. The apparent clustering of adenylyl cyclases, coupled with other data that suggest higher-order associations of regulatory elements including G proteins, N-methyl-D-aspartate receptors, and cAMP-dependent protein kinases, suggests not only that the primary structural information has been encoded to render the cAMP system responsive to the Ca(2+)-signaling system but also that higher-order strictures are in place to ensure that Ca2+ signals are economically delivered and propagated.

The characterization of a novel human adenylyl cyclase which is present in brain and other tissues

We characterized a human cDNA clone which encodes a novel adenylyl cyclase. Data from Southern and Northern blot analysis, and analysis of sequence similarity with a recently cloned mouse adenylyl cyclase (10), indicated that the human adenylyl cyclase was a species variant of type VII adenylyl cyclase. The sequence of the novel human adenylyl cyclase indicated it was a member of the type II adenylyl cyclase family, and we compared the regulatory characteristics of the novel human enzyme with those of type II adenylyl cyclase. The human type VII and rat type II adenylyl cyclases, expressed in human embryonic kidney 293 cells, were activated by prostaglandin E1 (PGE1), but only type VII was activated by isoproterenol. The stimulation of type VII adenylyl cyclase by PGE1 and isoproterenol was attenuated by pretreatment of the cells with staurosporine. Phorbol 12,13-dibutyrate synergistically enhanced the stimulation of both type VII and type II enzyme activity by PGE1 and by the constitutively active Gs mutant Gs (Q227L). The human type VII adenylyl cyclase activity was unresponsive to capacitatively induced changes in intracellular Ca2+. The functional characteristics of human type VII adenylyl cyclase resemble those of the rat type II enzyme, but the enzymes may respond differently to in vivo phosphorylation conditions. While the mRNA for adenylyl cyclase type II was found in several brain areas, the message for type VII adenylyl cyclase was localized primarily to the cerebellar granule cell layer.

Adenylyl cyclases and the interaction between calcium and cAMP signalling

Adenylyl cyclase is the prototypical second messenger generator. Nearly all of the eight cloned adenylyl cyclases are regulated by one or other arm of the phospholipase C pathway. Functional and ultrastructural investigations have shown that adenylyl cyclases are intimately associated with sites of calcium ion entry into the cell. Oscillations in cellular cyclic AMP levels are predicted to arise because of feedback inhibition of adenylyl cyclase by Ca2+. Such findings inextricably intertwine cellular signalling by cAMP and internal Ca2+ and extend the known regulatory modes available to cAMP.

Adenylyl cyclase mRNA expression does not reflect the predominant Ca2+/calmodulin-stimulated activity in the hypothalamus

Only three (Types I, II, V) of the six currently-described subtypes of adenylyl cyclase are prominently expressed in the rat brain. These species are differently sensitive to Ca2+, beta gamma subunits of G-proteins and protein kinase C. A knowledge of the susceptibility of the cAMP-signalling system in particular brain regions to these diverse modes of regulation can shed light on the mechanism of action of the neurotransmitters that modify neuronal activity in such regions. Cyclic AMP is extensively involved in the physiological functions of the hypothalamus. We have used in situ hybridization histochemistry with synthetic oligonucleotides to examine the expression in the rat hypothalamus of the three major brain subtypes of adenylyl cyclase-Ca2+/calmodulin-stimulable (Type I), Ca(2+)-insensitive (Type II) and Ca(2+)-inhibitable (Type V). The hypothalamus expresses high levels only of Type II mRNA, particularly in the supraoptic and paraventricular nuclei. Curiously, the strong expression of the Ca(2+)-insensitive Type II mRNA and the lack of expression of the major brain specific Type I mRNA does not correlate with the adenylyl cyclase activity, which is largely Ca2+/calmodulin stimulable in plasma membranes prepared from the hypothalamus.

Type VIII adenylyl cyclase. A Ca2+/calmodulin-stimulated enzyme expressed in discrete regions of rat brain

A cDNA that encodes type VIII adenylyl cyclase has been isolated from two rat brain libraries. The open reading frame encodes a 1248-amino acid protein predicted to have two sets of six transmembrane spans and two putative nucleotide binding domains as is characteristic of other mammalian adenylyl cyclases. Two type VIII messages are detected in rat brain with estimated sizes of 5.5 and 4.4 kilobases. In situ hybridization indicates that the type VIII messages are most abundantly expressed in the granule cells of the dentate gyrus, the pyramidal cells of hippocampal fields CA1-CA3, the entorhinal cortex, and the piriform cortex. Hybridization is also detected in the neocortex, the amygdaloid complex, and regions of the thalamus and hypothalamus. Stable expression of the type VIII cDNA in human embryonal kidney cells leads to the appearance of a novel 165-kDa glycoprotein in the membrane fraction. Stimulation of these cells with agents that increase intracellular Ca2+ results in up to 43-fold increases in cAMP accumulation over that of control cells transfected with the expression vector. Addition of isoproterenol alone does not lead to type VIII-specific effects in intact cells. Adenylyl cyclase activity in membranes prepared from type VIII-transformed cells is stimulated up to 40-fold by the addition of Ca2+/calmodulin (EC50 = 53 nM calmodulin). The addition of activated recombinant alpha subunit of Gs synergistically increases the Ca2+/calmodulin-stimulated activity. A possible role for type VIII adenylyl cyclase in long-term potentiation is discussed.

Selective expression of one Ca(2+)-inhibitable adenylyl cyclase in dopaminergically innervated rat brain regions

Type I adenylyl cyclase, which can be stimulated by elevated cellular levels of Ca2+, has been proposed to provide a positive coincidence signal detection system, which can integrate signals arising via Gs- and Ca(2+)-mediated pathways. The occurrence of this adenylyl cyclase in brain regions implicated with associative learning in invertebrates and with the mammalian model of plasticity--hippocampal long-term potentiation, supports the notion that the ability of this species of adenylyl cyclase to detect two signals simultaneously may play a role in this neuronal function. In the present study, two recently cloned, closely-related adenylyl cyclases (Types V and VI), are shown to be inhibited by physiological elevation in [Ca2+]i. As a first step towards probing the neuronal significance of Ca(2+)-inhibitable adenylyl cyclases, their distribution was evaluated by in situ hybridization analysis of the rat brain. Strikingly distinct patterns of gene expression were found, ranging from a highly selective distribution of Type V mRNA within the striatum, nucleus accumbens and olfactory tubercle, to a weak and ubiquitous distribution of Type VI mRNA. Type V AC mRNA is expressed exclusively in medium-sized striatal neurons, which also express D1-dopaminergic (Gs-linked) and M1-muscarinic cholinergic (Ca(2+)-linked) receptors. Thus the adenylyl cyclase is primed for simultaneous detection of opposing regulatory influences. The utility of this novel mode of signal detection to dopaminergic function remains to be established.

Discrete expression of Ca2+/calmodulin-sensitive and Ca(2+)-insensitive adenylyl cyclases in the rat brain

Physiological studies and inferences from invertebrate models implicate Ca2+/calmodulin-sensitive adenylyl cyclase with memory processes. However, Ca2+/calmodulin-insensitive adenylyl cyclase also occurs in brain, and its neuronal functions are less clear. Two oligonucleotide probes, based on rat cDNAs for Types I and II adenylyl cyclase, which appear to correspond to these functional forms, were used to conduct in situ hybridization analysis of the relative abundance and localization of these two species in the rat brain as a first step in evaluating their neuronal role. Quite discrete patterns of expression were encountered; in some areas, both species were co-expressed, but in others, little overlap was observed. The differential expression of the two mRNAs suggests that discrete roles may be fulfilled by the two adenylyl cyclases in neural tissues.

Preparation and characterization of a specific antibody for the immunohistochemical detection of L-dopa in paraformaldehyde-fixed rodent brains

A rat polyclonal antiserum has been obtained after coupling of L-3,4-dihydroxyphenylalanine (L-DOPA) to larger proteins using a low concentration of glutaraldehyde. The antiserum was tested for its affinity and specificity using an enzyme-linked-immunosorbent-assay (ELISA). From competition experiments, the most immunoreactive compound was found to be the non-reduced L-DOPA conjugate. Our specific L-DOPA antiserum enables us to visualize L-DOPA molecule on brain of guinea pigs and rats. We examined the immunohistochemical distribution of the polyclonal L-DOPA antiserum after the fixation of brains with a mixture of paraformaldehyde and picric acid. The presence of L-DOPA-immunoreactive (IR) neurons and fibers was described in the posterior, dorsal and periventricular hypothalamic areas and in the arcuate nucleus. Finally, the distribution of L-DOPA-IR cells was compared to that of tyrosine hydroxylase (TH)-IR cells, by means of a double staining procedure. The presence of two populations of TH-IR cells (TH-positive/L-DOPA-negative and TH-positive/L-DOPA-positive cells) was described in the dorsal part of the hypothalamus.

Simultaneous monitoring of cytosolic free calcium and exocytosis at the single cell level

Abstract Quinacrine, a fluorescent basic molecule, accumulates in secretory granules of pituitary cells, as was revealed by its colocalization with immunoreactive prolactin. Thus quinacrine fluorescence may be used to monitor secretory activity at the single cell level. Rat pituitary cells in primary culture were loaded with quinacrine and stimulated with physiological secretagogues, such as thyrotrophin-releasing hormone or bradykinin, which induced a multiphasic lowering of fluorescence, corresponding to the loss of quinacrine contained in exocytosed granules. Quinacrine was further used in combination with the fluorescent calcium probe fura-2, in order to monitor simultaneously exocytosis and variations in the cytosolic free calcium concentration, [Ca(2+)](i). With an appropriate selection of the excitation wavelengths, in dual excitation microfluorimetry experiments, it was possible to distinguish between fluorescence changes due to altered [Ca(2+)](i) versus quinacrine exocytosis. Transient elevations of [Ca(2+)](i) were provoked in individual pituitary cells by enhancing calcium influx through voltage gated channels. In part of the cells an initial increase in [Ca(2+)](i) coincided with stimulated quinacrine release. The approach was also applied to cells of the neuroblastoma line NCB20, where stimulation with bradykinin caused a transient rise in [Ca(2+)](i), concomitantly with enhanced exocytosis. No increase in exocytosis was ever detected without an elevation of [Ca(2+)](i), suggesting that in both cellular systems, an increase in [Ca(2+)](i), is absolutely necessary, but not sufficient to induce secretion.

Quantitative image analysis with densitometry for immunohistochemistry and autoradiography of receptor binding sites--methodological considerations

Major technical progress in the development of computer-based image analysis has made possible the entry of autoradiography and immunohistochemistry into a new era where quantification by densitometry has become easily accessible. Autoradiography could become quantitative and displayed adequate reproducibility with the help of emulsion-coated films and the use of scales of standards of known radioactivity exposed and analyzed in parallel to the tissue sections. Immunohistochemistry after revelation by a color-based enzymatic technique can also become quantitative, providing that standardization of the crucial steps of the procedure and calibration through a parallel treatment of a scale of antigen standards can be ensured. Such an approach is described here in the rat with reference to tyrosine hydroxylase (TH), the main synthesizing enzyme for catecholamines, and with dopamine (DA) itself, a catecholaminergic neurotransmitter. The different parts of the procedure, which can influence the results, such as the fixation of the animals by perfusion and the evaluation of the fluctuations via the calibration curve, are discussed in detail. Biological validation of the proposed procedure is described by reference to experiments already well documented biochemically, such as the induction effect of reserpine on TH in the rat locus coeruleus and the depleting effect of alpha-methyltyrosine (AMPT), a well-known blocker of TH activity, on rat striatal DA content. Finally the importance of restricting the measurements to the (pseudo)linear portion of the calibration curve is illustrated by the autoradiographic identification of the differential intrastriatal repartition of the dopaminergic D1 and D2 receptor sites, particularly the dual patch-matrix compartments.

The metabolism of exogenous L-dopa in the brain: an immunohistochemical study of its conversion to dopamine in non-catecholaminergic cells of the rat brain

The characterization and localization of non-catecholaminergic cells producing dopamine after L-Dopa load have been investigated in the normal rat brain by a direct immunohistochemical labelling of amines using specific antibodies. The detection of dopamine-containing non-catecholaminergic cells has been achieved in rats given a commonly used mixture of L-Dopa plus peripheral decarboxylase inhibitor, and compared to controls. Results indicate that serotoninergic neurons tend toward a switch of their metabolism into dopamine production after L-Dopa load in a dose-dependent manner. In addition small non-aminergic cells, identified as aromatic amino-acid decarboxylase-containing cells, were observed to produce dopamine after exogenous L-Dopa load. Possible implications of such results concerning the mode of action of L-Dopa in the brain are discussed.