11723 research outputs found
Sort by
Neuronal activity in dorsal CA1 drives memory consolidation and depends on Arc/Arg3.1
No full textLearning and memory are fundamental aspects of human cognition, encompassing complex processes that facilitate the encoding, consolidation, and retrieval of information and experiences that are essential for survival, adaptation, and abstract thinking. Despite decades of rigorous research, understanding the mechanisms behind memory encoding and consolidation remains a significant challenge. On the molecular level, novel gene transcription initiated during memory acquisition starts a cascade of molecular and cellular events concluding with structural and functional modulations of neurons, synapses, and memory-relevant neural networks. On a systemic level, a prevalent and re-emerging theory suggests that memories are integrated into engrams, which consist of groups of neurons (engram neurons) that are active during learning and are reactivated during memory retrieval. At the heart of this hypothesis lies the assumption that the collective activity pattern of engram neurons constitutes the very code of the stored information and contains distinctive elements setting it apart it from other acquired memories. The replay of these activity patterns can revoke the specific memory. Recent development of miniaturized head-mounted microscopes, so called miniscopes, combined with genetically-encoded calcium indicators (GECIs), provides a platform for recording the simultaneous activity of hundreds of individual neurons in the brains of freely behaving animal models. A surge of recent studies utilizing these techniques, provided invaluable information on neural activity supporting memory acquisition and retrieval. Nonetheless, many fundamental questions remain unanswered, in particular, what activity patterns and possibly engrams, support the persistence of memories over long-periods of time while maintaining their specificity? How does the process of molecular consolidation affect the emergence and time-dependent evolution of memory engrams and their activity?
The aim of this thesis is to address these questions by employing miniscopes and GECI technologies to perform longitudinal calcium imaging from mice performing a contextual fear memory task over several weeks. To unveil the contribution of molecular consolidation to formation and time-dependent evolution of memory engrams, I chose to investigate Arc/Arg3.1 KO mice, which fully lack the ability to consolidate long-term memory.
The activity-regulated cytoskeleton-associated protein Arc/Arg3.1 is coded by the immediate-early gene (IEG) ARC/ARG3.1, known for its central role in memory. Arc/Arg3.1 expression in the hippocampus is low under baseline activity, but is rapidly and strongly upregulated by plasticity-inducing stimuli, novel experience, learning and memory retrieval. Previous work from our lab and others, demonstrated that genetic deletion or inhibition of ARC/ARG3.1, resulted in failure to stabilize synaptic plasticity and a concurrent loss of long-term memories. Given the importance of the hippocampus to episodic memory and its dynamic expression of Arc/Arg3.1, we chose to investigate neural activity and memory engrams in the dorsal CA1 (dCA1) region.
We injected jGCaMP8m, a fast and sensitive GECI into the dCA1 of adult mice and implanted GRIN lenses for chronic imaging. The Inscopix miniscope and acquisition system were used to acquire calcium signals in vivo. Calcium imaging data were pre-processes to remove noise and motion artifacts. Calcium imaging data were motion-corrected within and between all imaging sessions of the longitudinal study. We conducted source extraction to obtain footprints of individual neurons and extracted their calcium traces, which were further refined with the aid of a Random Forest model. Finally, calcium traces were deconvolved into Calcium spikes using the OASIS algorithm. Calcium spikes were then used to identify place cells and neurons responsive to shock, tone, freezing, and mobility. Additionally, we conducted multi-electrode in vivo recordings from urethane-anesthetized mice in dorsal CA1, to examine action potential firing in WT and KO mice.
In part 1, we evaluated exploratory behavior in an open field arena, to examine possible influence of the surgical procedure and the tethering of the miniaturized microscopes on mice behavior. We found no significant differences in exploratory behavior between non-implanted and implanted mice, confirming preservation of native behavior under the imaging conditions. In all mice, exploratory activity in the same arena, declined on subsequent testing days, a sign of natural habituation. We noted stronger habituation in the KO mice.
Interestingly, calcium activity rates were lower in dCA1 of KO mice compared with WT littermates. Reduced calcium activity was associated with smaller amplitudes and shorter durations of complex calcium events but not with their numbers. Deconvolution of calcium events indicated a decrease in the number of spikes contributing to each event. Since calcium events result primarily from action-potential mediated influx through voltage dependent calcium channels, both reduction in action potential firing rates and alterations in calcium influx, might be responsible for the reduced calcium activity in the KO mice. We tested this possibility by recording extracellularly, action potentials (spikes) in dCA1 of urethane-anesthetized mice. We observed similar firing rates in KO and WT neurons, pointing towards the intriguing possibility of altered calcium influx regulation in the KO mice, as the source of reduced calcium spike activity.
A central function of CA1, is to encode location through place activity of its neurons. We thus analyzed calcium spikes for signs of place cell activity. Our findings reveal that KO mice contain slightly lower percentage of place cells, with similar place field size but weaker tuning, compared to WT. Across the three testing days, Place cells displayed higher and more stable activity rates, compared to other neurons for both WT and KO. Remapping of place cells between testing days was predominant in both genotypes, with non-significantly increased stability of WT neurons. In total, these findings show rather subtle differences between WT and KO place cell properties, indicating that despite the difference in calcium spike activity, Arc/Arg3.1 KO mice, maintained basic ability for functional space encoding in dCA1.
The second and major part of the thesis, investigates the link between neural activity and memory consolidation over long time periods. Implanted WT and KO mice underwent contextual fear conditioning and were tested in the conditioning context and in a novel context twice: 7 days (recent memory) and 21 (remote memory) days after conditioning. Firstly, we asked how WT and KO mice encoded contextual, sensory and behavior-associated information during the conditioning session. Behavioral analysis showed prolonged freezing in all mice following conditioning, reaffirming successful fear conditioning in WT and KO mice, as previously reported. Analysis of the simultaneously recorded calcium traces revealed a widespread reduction in calcium activity rates after the conditioning, that was particularly pronounced in the WT neurons. In both genotypes, activity rates decreased in two-thirds of the neurons but increased in the remaining third. We identified neurons with heightened responses to either tone or shock stimuli or to both. In both WT and KO mice, tone and shock neurons constituted, each, roughly 9% of all neurons and dual-responding neurons less than 1%. No significant differences were detected in the distribution of stimulus-responsive neurons between KO and WT neurons, or in their averaged population activity. Examination of the post-conditioning activity of these neurons, revealed that contrary to the overall inhibition of CA1 neurons, most tone-responsive neurons were excited in WT and KO mice alike. Divergent modulations were observed for the shock-responsive neurons, that were largely inhibited in the WT, but excited in the KO mice. During pre- and post-shock periods, mice engaged in two types of behavior: movement as an exploratory activity and freezing, signifying risk assessment, anxiety and fear. The proportion of time mice engaged with each behavior, changed dramatically after conditioning. We classified neurons, as either mobility- or freezing-responsive when their activity rates were significantly increased during concurrent behavior. Neurons without significantly preferred responses were classified as “others”. Prior to conditioning (Pre-shock), the distribution of neuronal classes was overall similar among WT and KO mice, constituting of a minority of freeze-neurons followed by a larger fraction of mobility neurons and a majority of neurons classified as others. In the post-conditioning episode, freezing bouts increased in number and duration and concordantly the number of freezing-responsive neurons. An even stronger modulation was observed for the mobility-neurons, most of which lost their preferred response altogether. One interpretation of these findings is that, fear-induced freezing is largely supported by removal of mobility-related activity, in addition to recruitment of fear-related activity. The similarity of these activity modulations among WT and KO, reflects their similar behavioral responses. Analysis of place cell activity was performed also for the conditioning context, showing similar fractions of place cells in WT and KO, and for both genotypes, only a small subset of neurons maintained their classification as place cells across both pre- and post-shock episodes (less the 2%).
In summary, analysis of calcium-spike activity, revealed similar fractions of tone, shock, mobility, freezing and place cell specific neurons, in WT and KO mice, suggesting that despite reduced activity rates in KO neurons, encoding of these functional properties, emerged independently of Arc/Arg3.1 presence and remained sufficiently efficient to support the rapid aversive learning observed in the Arc/Arg3.1 KO mice. To our surprise, conditioning silenced the majority of neurons in dCA1 of both genotypes.
We next investigated neural activity patterns during recent and remote memory recall. As shown in the first chapter, WT mice showed significantly higher percentage freezing in the conditioning context during recent and remote tests, but much weaker freezing in the novel contexts, indicating successful formation of strong and specific context memory. In contrast, the KO mice displayed weak and indiscriminate freezing in the conditioning and novel contexts, indicating their inability to consolidate the memory. Analysis of the collective population calcium activity, revealed a distinct pattern associated with the detection of familiar and novel environments. When WT mice were placed in the familiar conditioning-environment, the hippocampus calcium activity was reduced but significantly increased when the mice were placed shortly afterwards in a novel environment; this pattern was entirely absent in KO mice. We investigated the degree of correlation between neurons in the same mice, classifying significant ones as correlation partners. Analysis of WT mice, revealed a stark increase in the number of correlation-partners in the familiar conditioning-environment but not in the novel ones. Again, this coactivation signature was absent in KO neurons. Place cell activity displayed a similar modulation in WT mice, where their numbers and activity rates dropped in the familiar- but increased in the novel contexts. A similar but weaker pattern of place cell recruitment and activity was observed in KO mice. A striking difference between WT and KO became apparent upon analysis of the place field size: whereas in WT the field size increased after shock and was retained in the familiar context but reduced in the novel, the exact opposite modulation was observed in KO place cells. These findings confirm that place cells activity, contributes to distinction between familiar and novel environment up to three weeks post conditioning. Moreover, these findings suggest that while formation of place cells was not strongly affected by the loss of Arc/Arg3.1 in the KO mice, their modulation along the process of memory consolidation, was strongly altered.
Finally, we attempted to detect neural engrams that encode the conditioning context and are reliably reactivated during recent and remote memory tests. We also expected such engrams to differ significantly from those active in the unfamiliar contexts. We explored two engram definitions based on their heightened activity, either during the pre-shock episode or the during recent memory test in the familiar context. The number of reactivated neurons, based on either engram-definition was similar among genotypes and environments. However, activity rates of neurons allocated to the engram and their correlation values, displayed a significant environment-dependent modulation in WT mice but not in KO neurons. In total, our exploration of putative memory engrams supports the notion that highly active neurons are more likely to be allocated into the engram. However, it also suggests that rates and correlations of activity contribute most significantly to a memory signature rather than the size or identity of specific neurons.
In total, this thesis presents new findings on activity and functional encoding in Arc/Arg3.1 KO mice and sheds new light on the neural activity code underlying long-term memory consolidation. These include the following:
1. Arc/Arg3.1 regulates activity-dependent calcium influx in the hippocampus.
2. The emergence and formation of neurons encoding sensory, spatial, and behavioral information in dCA1 are independent of Arc/Arg3.1 plasticity.
3. Enhanced activity rates and neural synchronization underlie successful memory retrieval and distinguish familiar from novel information.
4. Modulation of activity rates and neural synchronization along the process of memory consolidation is driven by Arc/Arg3.1 expression in the brain.Lernen und Gedächtnis sind grundlegende Bestandteile menschlicher Kognition. Sie umfassen komplexe Prozesse welche für das Überleben, Adaptionen und abstraktes Denken essenziell sind. Hierzu zählen die Kodierung, Konsolidierung und der Abruf von Informationen und Er-fahrungen. Trotz jahrzehntelanger intensiver Forschung sind die zugrundeliegenden Mecha-nismen der Kodierung und Konsolidierung des Gedächtnisses bis heute unbekannt. Auf mole-kularer Ebene setzt eine neue Gentranskription während des Gedächtniserwerbs eine Kaskade molekularer und zellulärer Ereignisse in Gang welche in einer strukturellen und funktionellen Modulation von Neuronen, Synapsen und gedächtnisrelevanten neuronalen Netzwerken resul-tiert. Auf der systemischen Ebene schlägt eine weit verbreitete Theorie vor, dass Erinnerungen in Engramme codiertwerden. Diese Engramme bestehen aus Gruppen von Neuronen (En-grammneuronen), welche während des Lernens aktiv sind und beim Abrufen von Erinnerungen reaktiviert werden. Im Mittelpunkt der Theorie steht die Annahme, dass das kollektive Aktivi-tätsmuster der Engrammneuronen den eigentlichen Code der gespeicherten Information darstellt und charakteristische Elemente enthält, welche sie von anderen erworbenen Erinnerungen un-terscheiden. Die Wiederholung dieser Aktivitätsmuster kann die spezifische Erinnerung wie-derherstellen. Die jüngste Entwicklung miniaturisierter kopfgetragener Mikroskope, so genann-ter Miniskope, in Verbindung mit genetisch kodierten Kalziumindikatoren (GECIs) bietet eine Plattform für die gleichzeitige Aufzeichnung der Aktivität von Hunderten einzelner Neuronen im Gehirn frei agierender Tiermodelle. Eine Reihe neuerer Studien, bei denen diese Techniken zum Einsatz kamen, lieferten wichtige Informationen über die neuronale Aktivität, die den Erwerb und das Abrufen von Gedächtnisinhalten unterstützt. Dennoch bleiben viele grundlegen-de Fragen unbeantwortet. Insbesondere die Frage, welche Aktivitätsmuster und potenzielle Engramme die Persistenz von Erinnerungen über lange Zeiträume hinweg unterstützen und dabei ihre Spezifität erhalten? Eine weitere bislang unbeantwortete Frage ist inwiefern sich der Prozess der molekularen Konsolidierung auf die Entstehung und zeitabhängige Entwicklung von Gedächtnis-Engrammen und deren Aktivität auswirkt?
Ziel dieser Arbeit ist es, diese Fragen zu beantworten, indem Miniskope und GECI-Technologien eingesetzt werden. Diese ermöglichen eine longitudinale Kalziumbildgebung anhand dieser die neuronale Aktivität von Mausmodellen, während kontextuelle Furchtge-dächtnisaufgaben über mehrere Wochen verfolgt werden kann. Um den Beitrag der molekula-ren Konsolidierung zur Bildung und zeitabhängigen Entwicklung von Gedächtnis-Engrammen aufzudecken, habe ich mich für die Untersuchung von Arc/Arg3.1 KO-Mäusen entschieden, denen die Fähigkeit zur Konsolidierung des Langzeitgedächtnisses vollständig fehlt.
Das aktivitätsregulierte Zytoskelett-assoziierte Protein Arc/Arg3.1 wird durch das Immediate-Early-Gen (IEG) ARC/ARG3.1 kodiert, das für seine zentrale Rolle im Gedächtnis bekannt ist. Die Expression von Arc/Arg3.1 im Hippocampus ist bei Grundaktivität gering, wird aber durch plastizitätsfördernde Reize, neue Erfahrungen, Lernen und Gedächtnisabruf schnell und stark hochreguliert. Frühere Arbeiten unserer Arbeitsgruppe und anderer haben gezeigt, dass die ge-netische Deletion oder Inhibition von ARC/ARG3.1, zu einer fehlenden Stabilisierung der sy-naptischen Plastizität und zu einem gleichzeitigen Verlust des Langzeitgedächtnisses führt. An-gesichts der Bedeutung des Hippocampus für das episodische Gedächtnis und seiner dynami-schen Expression von Arc/Arg3.1 haben wir uns entschieden, die neuronale Aktivität und die Gedächtnis-Engramme in der dorsalen CA1-Region (dCA1) zu untersuchen.
Wir injizierten jGCaMP8m, ein schnelles und empfindliches GECI, in die dCA1 von adulten-Mäusen und implantierten GRIN-Linsen für die chronische Bildgebung. Das Inscopix-Miniskop und das Aufnahmesystem wurden zur Erfassung von Kalziumsignalen in vivo ver-wendet. Die Kalzium-Bilddaten wurden vorverarbeitet, um Rauschen und Bewegungsartefakte zu entfernen. Die Kalzium-Bilddaten wurden innerhalb und zwischen allen Bildgebungssitzun-gen der Längsschnittstudie bewegungskorrigiert. Wir führten eine Quellenextraktion durch, um Fußabdrücke einzelner Neuronen zu erhalten, und extrahierten deren Kalziumspuren, die mit Hilfe eines Random Forest-Modells weiter verbessert wurden. Schließlich wurden die Kalzi-umspuren mit Hilfe des OASIS-Algorithmus zu Kalziumspikes dekonvolviert. Die Kalzium-spikes wurden dann zur Identifizierung von Ortszellen und Neuronen verwendet, die auf Schock, Tonus, Einfrieren und Mobilität reagieren. Zusätzlich führten wir in vivo Multi-Elektroden-Aufnahmen von Mäusen im dorsalen CA1 durch, um das Feuern von Aktionspo-tentialen bei WT- und KO-Mäusen zu untersuchen.
Im ersten Teil des Projektes untersuchten wir das Erkundungsverhalten in einer offenen Felda-rena, um mögliche Einflüsse des chirurgischen Eingriffs und der Anbringung der Miniaturmik-roskope auf das Verhalten der Mäuse zu untersuchen. Wir fanden keine signifikanten Unter-schiede im Erkundungsverhalten zwischen nicht-implantierten und implantierten Mäusen, wel-ches die Bewahrung des natürlichen Verhaltens unter den Bildgebungsbedingungen bestätigt. Bei allen Mäusen nahm die Erkundungsaktivität in derselben Arena an den folgenden Testtagen ab, welches ein Zeichen für die natürliche Gewöhnung ist. Wir stellten eine stärkere Gewöh-nung bei den KO-Mäusen fest.
Interessanterweise waren die Kalziumaktivitätsraten in dCA1 von KO-Mäusen im Vergleich zu WT-Wurfgeschwistern geringer. Die geringere Kalziumaktivität ging mit kleineren Amplituden und kürzeren Dauern komplexer Kalziumereignisse einher, nicht aber mit deren Anzahl. Die Dekonvolution der Kalziumereignisse zeigte eine Abnahme der Anzahl von Spikes, die zu jedem Ereignis beitragen. Da Kalziumereignisse in erster Linie aus dem durch Aktionspotenziale vermittelten Einstrom von Kalziumionen durch spannungsabhängige Kalziumkanäle resultie-ren, könnten sowohl eine Verringerung der Feuerfrequenz von Aktionspotenzialen als auch Veränderungen des Kalziumeinstroms für die verringerte Kalziumaktivität in den KO-Mäusen verantwortlich sein. Wir testeten diese Möglichkeit durch extrazelluläre Aufzeichnungen von Aktionspotenzialen (Spikes) in dCA1 von mit Urethan betäubten Mäusen. Wir beobachteten ähnliche Feuerfrequenzen in KO- und WT-Neuronen, welches auf eine veränderte Regulierung des Kalziumeinstroms in den KO-Mäusen als Ursache für die verringerte Kalziumspike-Aktivität hindeutet.
Eine zentrale Funktion der CA1-Region besteht darin, durch die Ortsaktivität seiner Neuronen den Ort zu kodieren. Daher haben wir die Kalziumspikes auf Anzeichen von Ortszellenaktivität untersucht. Unsere Ergebnisse zeigen, dass KO-Mäuse einen geringfügig niedrigeren Prozent-satz an Ortszellen mit einer ähnlichen Größe des Ortsfeldes, aber einer schwächeren Abstim-mung im Vergleich zu WT-Mäusen aufweisen. Über die drei Testtage hinweg zeigten die Orts-zellen höhere und stabilere Aktivitätsraten im Vergleich zu anderen Neuronen, sowohl bei WT als auch bei KO-Mäusen. Das Neuzuordnung von Ortszellen zwischen den Testtagen war bei beiden Genotypen vorherrschend, wobei die Stabilität der WT-Neuronen nicht signifikant er-höht war. Insgesamt zeigen diese Ergebnisse eher subtile Unterschiede zwischen der WT- und KO-Ortszellen, was darauf hindeutet, dass in Arc/Arg3.1 KO-Mäuse trotz des Unterschieds in der Kalziumspike-Aktivität die grundlegende Fähigkeit zur funktionellen Raumkodierung in dCA1 erhalten blieb.
Im zweiten Tei
Analysis of meiosis-specific gene expression regulation and establishment of a cytological framework of female meiosis by live-cell imaging in Arabidopsis thaliana
No full text
