What is the Function of AMPA receptor subunit GluR3 in
Memory Formation and Consolidation?
Eva H.H. Albers University of Amsterdam
Eva H. H. Albers
Student number: 10137688
MSc Brain and Cognitive Sciences, Cognitive Neuroscience track University of Amsterdam
Supervisor: Dhr. Dr. H. W. Kessels Synaptic Plasticity and Behavior, Netherlands Institute for Neuroscience Co-assessor: Dhr. Dr. H.J. Krugers
Structural and Functional Plasticity of the Nervous System, Swammerdam Institute for Life Sciences
July 2013 – July 2014 50 ECS
Abstract
Introduction: Synaptic plasticity is thought to be one of the mechanisms underlying learning
and memory and is largely driven by AMPA-R trafficking. AMPA-R complexes in the adult hippocampus exist mainly in GluR1/2 and GluR2/3 heteromers. Synaptic insertion and removal of GluR1-containing AMPA-Rs is activity dependent and known to promote LTP and LTD. The function of AMPA subunit GluR3 is largely unknown. The aim of this study is to examine the role of GluR3 in memory formation and consolidation. In the first experiment it was hypothesized that replacement of GluR1/2s by GluR2/3s could underlie memory consolidation and is facilitated by sleep. The second experiment explored how GluR3 affects memory formation mediated by the physiological stress response.
Methods & Results: In the first experiment wild type c57bl/6J mice were contextually fear
conditioned, followed by 6 hours sleep deprivation, lastly memory performance was tested. The results showed no effect of sleep deprivation on memory performance. In the second experiment GluR3 knockout and wild type mice were injected with epinephrine, contextually fear conditioned with a low shock intensity and memory performance was tested two hours after conditioning. GluR3 knockout mice showed better memory performance upon increased levels of epinephrine than wild type littermates did. Also, GluR3 knockout mice injected with epinephrine showed increased memory compared to GluR3 knockout mice with vehicle. Together, this suggests that the physiological stress response enhances memory formation and that GluR3 has a suppressing effect on the formation of these memories.
Conclusions: It is unlikely that replacement of GluR1/2s by GluR2/3s is dependent on sleep
and thereby underlies memory consolidation. Theoretically it is more conceivable that GluR1 and GluR3 represent two independent memory systems, or that GluR3 plays a memory suppressing role in stress mediated memory formation, rather than a stabilizing role. However, the exact mechanisms remain to be studied.
Survival in a constantly changing environment requires the use of learning and memory. Synaptic plasticity is thought to be one of the neuronal mechanisms underlying learning and memory (Kandel, Schwartz & Jessell, 2000). Synaptic plasticity refers to the strengthening or weaking of synapses depending on its co-activity with other synapses (Hebb, 1949).
Glutamate is the major excitatory neurotransmitter in vertebrates (Byrne, 2008). The two major glutamatergic receptors are N-methyl-D-aspartate (NMDA) and α-amino-3-hydroxy-5-methyl-4-isoxazole propionic-acid (AMPA).
Synaptic plasticity depends upon activity-dependent insertion or removal of AMPA receptors in the synapse (Kandel, Sschwartz & Jessell, 2000). Presynaptic release of
glutamate acts on postsynaptic AMPA receptors (AMPA-Rs), which leads to depolarization. The depolarization triggers the release of magnesium blocks from NMDA receptors, resulting in a calcium influx (Byrne, 2008). Calcium in flux in the post-synaptic neuron activates Ca2+/calmodulin-dependent protein kinase II (CamKII), which phosphorylates AMPA
receptors and increases channel conductance (Benke, Lüthi, Isaac & Collingridge, 1998) and synaptic insertion of AMPA-Rs (Plant et al., 2006). Term Potentiation (LTP) and Long-Term Depression (LTD) refer to a change in synaptic plasticity induced by prior activation
and indicate, respectively, an increased or decreased receptor density and, hence,
susceptibility for synaptic transmission (Lømo, 1966). LTP and LTD are associated with changes in synaptic efficacy and are mediated by phosphorylation and synaptic delivery of AMPARs (Malenka & Nicoll, 1999).
AMPA-Rs exist in four subunits: GluR1, GluR2, GluR3 and GluR4 (Dingledine et al., 1999). These subunits have different functions and different distribution over the course of ontogenetic development (Lohman & Kessels, 2013). In the adult hippocampus the
composition of the AMPA subunits are mainly the GluR1/2 heteromers and GluR2/3 heteromers (Wenthold et al., 1996; Lohman & Kessels, 2013). The role of GluR1 is widely studied (e.g. Hayashi et al., 2000; Hu et al., 2007; Lu et al., 2013; Malino & Malenka, 2002; Rumpel et al., 2005; Vyazovkiy et al., 2008). The trafficking of GluR1 containing AMPA-Rs in and out of the synapse is thought to be the major mechanism driving synaptic potentiation and depression (Kessels & Malinow, 2009). It has been shown, in vitro, that GluR1/2’s are inserted in the synapse during LTP (Hayashi et al., 2000) and during new experiences, in vivo (Rumpel et al., 2005). On the contrary, the role of GluR3 is largely unknown.
There are two proposed pathways of AMPA trafficking (Malinow, Mainen & Hayashi, 2000). The constructive pathway controls modification of synaptic strength via synaptic insertion or removal of GluR1-containing AMPA-Rs. This pathway is activity dependent. The second is the maintenance pathway, which provides a constant turnover of GluR2/3s in and out the synapse. The synaptic insertion of this pathway is thought to be much slower than the immediate activity dependent constructive pathway.
It is hypothesized that the constitutive trafficking of GluR2/3s mainly replaces AMPA-Rs that are already present in the synapse, thereby not modifying synaptic strength (Shi et al., 2001). The replacement of GluR1-containing AMPA-Rs by GluR2-containing AMPA-Rs could be a mechanism to stabilize synapses (Shi et al., 2001) and may reflect memory consolidation (Plant et al., 2006). Since GluR1/2s and GluR2/3s are the most common AMPA-Rs in the adult hippocampus (Wenthold et al., 1996) the replacement of GluR1-containing subunits by GluR2-GluR1-containing subunits (Shi et al., 2001) suggests a role of GluR3 in stabilization of synapses and possibly in memory consolidation.
GluR1 and GluR3 appear to have different effects on learning and memory formation. Mice lacking GluR1 showed a reduced memory 24-hours after cued (Humeau et al., 2007) and contextual (Feyder et al., 2007) fear conditioning, whereas mice lacking GluR3 showed no memory deficits compared to wild type littermates (Humeau et al., 2007). Furthermore,
preliminary data from our laboratory showed that GluR1 knock-out (KO) mice have a complete lack of memory ten minutes after contextual fear conditioning. However, the fear memory occurs slowly over the course of 24 hours. Taken together these results suggest that GluR1/2 containing AMPA-Rs are associated with immediate memory, whereas GluR2/3 seem to underlie slowly increasing memory consolidation over a time span of 24 hours. Memory consolidation happens over a time course of hours to days (Abel & Lattal, 2001), which seems to coincide with the time course of the gradually occurring fear memory that is observed in GluR1 KO mice.
Consolidation refers to the process of strengthening the newly formed memory trace in order to stabilize the memory (Abel & Lattal, 2001). In the classical view, consolidation reflects the transition of information from immediate and instable short term memory (STM) to the more stable long term memory (LTM; Byrne, 2008). Memory consolidation appears to be dependent on sleep. Lack of sleep is known to impair memory consolidation in mice (Cai et al., 2009; Graves et al., 2003; Silva et al., 2004; Takatsu-Coleman et al., 2013) and rats (Hagewoud et al., 2011; Hagewoud et al., 2010; Leenaars et al., 2013).
There is evidence that AMPA subunit composition is related to sleep. The levels of GluR1 are higher after wakefulness than after sleep, whereas the levels of GluR2 remain the same (Vyazovskiy et al., 2008). This suggests that the levels of GluR3 are higher after sleep than after wakefulness. Replacement of GluR1/2s by GluR2/3s during sleep could potentially explain this shift in AMPA subunit composition.
This raises the question whether sleep is promotes the synaptic insertion of GluR2/3s. Sleep is defined as a reversible disconnection from the environment (Tononi & Cirelli, 2014). It is plausible that synaptic insertion of GluR3-containing AMPA-Rs is dependent on sleep to drive synaptic stabilization and memory consolidation because new experiences drive GluR1-containing AMPA-Rs into the synapse (Rumpel et al., 2005). In order to stabilize existing synapses, the system must be disconnected from the environment to prevent new information to enter the system. This happens during sleep.
In summary, GluR2/3s are propsed to stabilize synapses (Shi et al., 2001) and facilitate memory consolidation (Plant et al., 2006). Lack of sleep impairs memory consolidation (Cai et al., 2009; Graves et al., 2003; Hagewoud et al., 2011; Hagewoud et al., 2010; Leenaars et al., 2013; Silva et al., 2004; Takatsu-Coleman et al., 2013). The composition of AMPA subunits differs between periods of wakefulness and sleep (Vyazovskiy et al.,
2008).Therefore, it is hypothesized that the replacement of GluR1/2s by GluR2/3 thought to underlie memory consolidation is dependent on sleep.
In order to test the role of GluR3 in sleep dependent consolidation, the first aim of the first experiment was to develop complimentary contextual fear conditioning and sleep deprivation protocols that detect the sleep deprived memory deficits in wild type mice. This aim is necessary before proceeding to the actual aim of this study. The second aim was to examine whether GluR3 mediates sleep dependent memory consolidation. It was hypothesized that GluR3 KO mice are resistant to sleep deprivation.
Stress response mediated memory encoding
In addition to the proposed role of GluR3 in consolidation of memories, GluR3 might also be involved in stress- mediated memory formation. Stress can both impair (Trammel & Clore, 2014) and enhance (Cahill, Gorski & Le, 2003) learning and memory encoding in both humans and animals. Two of the most important stress related neuromodulators, epinephrine and corticosterone, show bidirectional effects on learning (McGaugh, 2000). Acute stress has been shown to increase synaptic GluR1 levels in the hippocampus, which lead to the
facilitation of LTP (Whitehead et al., 2013). This is mediated by norepinephrine, which indirectly phosphorylates the c-tail of GluR1 and, hence, facilitates synaptic delivery of GluR1 (Hu et al., 2007). Furthermore, glucocorticoids induce in- or decreased LTP and facilitate consolidation of contextual fear conditioning via AMPA-R trafficking (Krugers, Hoogenraad & Groc, 2010). In contrast, norepinephrine (NE) can also decrease AMPA mediated EPSCs (Dihn et al., 2009). Contrary to the immediate effects of glucocorticoids, the slower HPA-axis mediated effects of glucocorticoids reduce LTP and thus suppress synaptic transmission, also via AMPA-Rs (Krugers, Hoogenraad & Groc, 2010).
Preliminary data from our laboratory show an increased synaptic efficacy in wild type mice and in mice lacking GluR1 ten minutes after fear conditioning (M. Renner, unpublished data). Two hours after fear conditioning the increase in mEPSC frequency is gone. Mice lacking GluR3 subunits do not show this transient increase synaptic efficacy. These findings indicate that the increase in synaptic efficacy depends on GluR3. Furthermore, the increase in mEPSC frequency is present in all recorded cells, which is in contrast to the selective synaptic strengthening that occurs with learning. Also, the same effects on synaptic efficacy are seen upon epinephrine injections (M. Renner, unpublished data). These results point towards the involvement of GluR3 in the stress-related component of fear conditioning. Therefore, the
aim of the second experiment is to examine the role of GluR3 under influence of stress and how this affects memory formation.
Methods
EXPERIMENT 1: THE ROLE OF GLUR3 IN SLEEP DEPENDENT MEMORY Animals and housing conditions
Forty, eight weeks old, C57Bl/6J male mice (Harlan Laboratories, Horst, The
Netherlands) were used to test different fear conditioning and sleep deprivation protocols. Mice were housed in groups of four and kept on a 12-hours day-night cycle, with onset of the light period at 7 a.m., onset of the dark period at 7 p.m. Animals had ad libitum access to water and food. All experiments were conducted with approval of the experimental animal committee (DEC) of the Royal Netherlands Academy of Sciences (KNAW) and in
accordance with Dutch legislation (Wet op Dierproeven, 1996). Procedure
Upon arrival, the mice were given two weeks to habituate to the new environment and reduce stress originating from the change in housing situation. After two weeks littermate mice were randomly assigned to either sleep (S) or sleep deprivation (SD). To test the effectiveness of different sleep deprivation and fear conditioning protocols, three protocols were designed (Cai et al., 2009; Graves et al., 2003).
MILD PROTOCOL. In the first protocol mice were fear conditioned with 3 shocks of 0.40 mA (1 second per shock). Directly after conditioning the SD mice (N = 4) were put in the sleep deprivation device (SDD) and kept awake for 6 hours. The control littermates (N = 4) were put back in their home cage and the home cage was put in an isolated room, to ensure undisturbed rest. After 6 hours either SD or undisturbed sleep (S) the mice were put back in their normal housing room. 24 Hours after fear conditioning, the mice were tested in the context where they received the shock.
NATURAL PROTOCOL. In the second protocol the natural sleep and wake cycle, rather than artificially induced wakefulness, was used (as described in: Cai et al., 2009). Mice were fear conditioned with 3 shocks of 0.80 mA (1 second per shock) either at the onset of the light period (between 7 a.m. and 8 a.m.), after their awake period, or at the onset of the dark period
(between 7 p.m. and 8 p.m.), after their asleep period. After conditioning the mice were put back in their normal housing room. The mice were tested 12 hours after conditioning, that is, after their natural awake period or after their natural sleep period.
IMMEDIATE PROTOCOL. The third protocol consisted of fear conditioning with 3 shocks of 0.4 mA (1 second per shock) at the onset the of light period, followed by either 6 hours of sleep deprivation or 6 hours of undisturbed rest. To refrain the mice from any substantial sleep between conditioning and testing, the mice in the third protocol are tested immediately after the 6 hours of sleep deprivation or undisturbed sleep.
Figure 1. Time line of the three different sleep deprivation paradigms: in the mild protocol (top) animals are
sleep deprived for six hours after fear conditioning and tested 24 hours after conditioning, in the natural protocol (middle) animals are either conditioned at the onset of the light period (natural sleep) or at the onset of the dark period (natural wakefulness) and tested 12 hours later, in the immediate protocol (bottom) animals were sleep deprived for six hours after fear conditioning and tested immediately after the sleep deprivation.
Contextual Fear Conditioning
Mice were conditioned in a chamber (29 cm high, 31.5 cm wide, 23 cm deep) with two metal walls, two transparent Plexiglas walls and grid floor with stainless steel bars through which the foot shock was delivered (Med Associations Inc., Georgia, VT). A camera, which
was mounted on the ceiling of the cage, provided a live view and was used to make
recordings of the conditioning and test trials (PCTV). The cage was cleaned with 70% ethanol between every trial.
On the day of conditioning the mice were put in the fear conditioning cage. The first two minutes were used to measure baseline activity. After two minutes three shocks (duration: 1 second) were delivered with one minute in between (see Procedure for the different shock intensities that were used per protocol). After the three shocks mice were kept in the fear conditioning cage for two minutes to measure post-shock freezing and motion.
On the day of testing the mice were put back in the fear conditioning cage for two minutes, not receiving a shock.
Sleep deprivation
Sleep deprivation was performed by means of an automated sleep deprivation box with minimal confounding locomotor activity, as described and tested by Leenaars et al. (2011). The device was adjusted for mice, by reducing the space of the individual compartments of the cage from 180° to 90°. Additionally, a second food tray and drink bottle were added in order to sleep deprive two mice at the same time. Mice had unrestricted access to water and food and the floor was covered with a thin layer of bedding. The apparatus was controlled by customized software (DEPRI, Netherlands Institute for Neuroscience). The speed of the rotating floor was kept constant at 2 cycles per minute and the direction altered every 2 minutes from clockwise (CW) to counter clockwise (CCW) and vice versa. These parameters forced the mice to move at least every 7.5 seconds ((2 cycles/ minute)/4 compartments = 7.5 second/ compartment). The mice were monitored via a live view camera on a remote desktop access. Contrary to the description in Leenaars et al. (2011), in this study there was no
increase in the rotation velocity in order to compensate for increasing sleep pressure. Visual inspection showed that increased rotation velocity led to severe exhaustion of the mice, thereby not reducing the confounding locomotor activity and violating the welfare criteria of the mice.
Data Analyses
The freezing values for the first and second sleep deprivation protocol were scored manually. Freezing was quantified with a stopwatch and calculated as the time that mice showed no movement except movement necessary for respiration, relative to the total time of
the video. The videos of the third protocol were analyzed with the MATLAB based software “mousemove”, developed by Kopec et al. (2007). Videos were converted from ~.mpg to ~.avi extension, had a resolution of 720 x 576 pixels and a frame rate of 11 Hz. Motion was
calculated as the amount of pixels per frame that differed significantly from the variation in a manually defined background area. Basal motion was calculated as the average number of motion pixels in the first two minutes (pre-shock). Test motion was calculated as the average number motion pixels during the entire test video (2 minutes). Freezing values were
calculated as the percentage of frames that did not contain more than 5 pixels movement for at least one second, compared to the total number of frames. Freezing values were calculated for the time of testing only, because under normal circumstances mice do not show freezing behavior in an unconditioned context. Additionally, a correlation between manual scoring and software was calculated based on test movies (N = 8) of the second protocol. The correlation was high (R = .978, p = .000) so freezing values could be compared between different protocols.
To statistically test the effect of sleep deprivation on memory consolidation a multivariate analysis of variance (MANOVA) was performed for each protocol with post-shock freezing and test freezing as dependent variables and sleep condition (undisturbed sleep or natural sleep vs. sleep deprivation or natural wakefulness) as between subject variable. For the third protocol the same analysis was beside for freezing also performed for motion, with post-shock motion and test motion as dependent variables.A p-value of .05 was considered significant. Descriptive statistics are presented as (M ± SEM) and error bars in the figures represent standard error of the mean (SEM).
EXPERIMENT 2: THE ROLE OF GLUR3 IN STRESS RESPONSE MEDIATED MEMORY Animals and housing conditions
Forty 8-12 week old male GluR3 knock-out (KO) and wild type littermates were used, obtained through inbred breeding at the Netherlands Institute for Neuroscience with mutant ancestors (B6x129P2-Gria3tm1Dgen/Mmnc) and originally ordered from Mutant Mouse Regional
Resource Centers (Davis, CA). Mice were, depending on the litter size, housed in groups of two to six animals and kept on a 12-hours day-night cycle, with onset of the light period at 7 a.m. and onset of the dark period at 7 p.m. Animals had ad libitum access to water and food. All experiments were conducted with approval of the experimental animal committee (DEC)
of the Royal Netherlands Academy of Sciences (KNAW) and in accordance with Dutch legislation (Wet op Dierproeven, 1996).
Procedure
Upon arrival in the new housing room, the mice were given a week to habituate, since the GluR3 mice were bred in the institute and internal transfer is thought to causes less stress than transport from an external company. One day before the experiment the mice were weighted to determine the injection amount per mice (5 ml solution/ kg). Epinephrine hydrochloride (Sigma-Aldrich Chemie BV, Zwijndrecht, The Netherlands) was diluted in saline (0.1
mg/ml). Ten minutes before onset of the conditioning the mice were injected intraperitoneally with either epinephrine (0.5mg/kg) or saline (control condition). Genotypes and injection solution were counterbalanced. Two minutes after the mice were put in the conditioning cage one mild shock (0.20 mA, 1 second duration) was delivered. Mice were kept in the
conditioning cage for two more minutes and then put back in the home cage.
Figure 2. Time line of the experiment. The mice were injected with either epinephrine or saline ten minutes
before onset of the conditioning session. The conditioning session consisted of two minutes of exploration after which one foot shock of 0.2 mA was delivered through the grid floor. Two minutes after the shock mice were put back their home cage. The memory test (duration: two minutes) takes place two hours after the conditioning session.
Contextual Fear Conditioning
For the fear contextual fear conditioning the same chamber was used as described before (“Experiment 1, Contextual Fear Conditioning”), but the parameters were slightly different (see “Experiment 2, Procedure”).
The contextual fear conditioning and test videos were analyzed with Matlab software, as described before (“Experiment 1, Data Analyses”), except the conditioning and test videos of four mice that contained noise from reflections. Those videos were analyzed manually. A correlation (R = .897, p = .000) between manual freezing and Matlab software freezing was calculated based on test videos (N = 13) of animals of the second protocol. Since the
correlation was high, the freezing data of the manually scored videos was added to the freezing scores that were analyzed with software. Additionally, for this experiment also pre-shock freezing was calculated, since the epinephrine injections only could lead to freezing behavior even before delivery of the shock.
To examine whether epinephrine modulated the formation and consolidation fear memory and whether this effect is mediated by GluR3, a multivariate analysis of variance was
performed (MANOVA) with genotype (knock-out vs. wild type) and injection solution (epinephrine vs. saline) as between subject factors. The dependent variables were ‘pre-shock freezing’ and ‘test freezing’. The same analysis was performed for the dependent variable motion (SMPs). When significant interactions were found, additional t-tests for independent samples were be done to reveal the significantly different contrast. A p-value of .05 was considered significant. Descriptive statistics are presented as (M ± SEM) and error bars in the figures represent standard error of the mean (SEM).
Results
EXPERIMENT 1: THE ROLE OF GLUR3 IN SLEEP DEPENDENT MEMORY In order to test the role of GluR3 in in sleep dependent memory formation and consolidation the first aim of this study was to develop a fear conditioning and sleep deprivation protocol in wild type (C57Bl/6J) mice that shows sleep deprivation related deficits in memory consolidation.
Firstly, no difference in freezing immediately after the shock is found (F(1, 6) = 0.96, p = .365), so the animals did not differ in their freezing behavior before they received either undisturbed sleep (N = 4) or sleep deprivation (N = 4). However, at the test time of this protocol (24 hours after contextual fear conditioning), also no difference in freezing was seen between mice that were sleep deprived for six hours after fear conditioning (47.44 ± 13.75)
and mice that received undisturbed sleep (42.65 ± 8.23). Figure 3 shows the mean percentage of freezing 24 hours after conditioning.
Figure 3. The effects of 6 hours of sleep (N = 4) and sleep deprivation (N = 4) on memory formation and
consolidation 24 hours after contextual fear conditioning (3x0.40mA, 1-sec) in wild type C57Bl/6J male mice. Analyses of variance yielded no significant effects.
Secondly, testing the mice twelve hours after conditioning, either after their natural awake period (dark) or after their natural sleep period (light), did not show any significant difference in freezing behavior (F(1, 14) = 0.558, p = .467). Interestingly, a significant difference was found in post-shock freezing (F(1,14) = 10.08, p = .0067). The mice that were conditioned in the morning, that is, at the onset of the light period before their natural sleep, showed slightly less freezing immediately after the shock (49.44 ± 2.79) than the mice that were conditioned in the evening, that is, at the onset of the dark period before their natural wakefulness (63.277 ± 3.35). Figure 4 shows the percentage of freezing immediately after fear conditioning (left graph: ‘conditioning’) and 12 hours after conditioning (right graph: ‘test session’).
Figure 4. The effects of sleep (N = 8) and natural wakefulness (N = 8) on freezing immediately after (left) and 12
hours after (right) contextual fear conditioning (3x0.80mA, 1-sec) in wild type C57Bl/6J male mice. Multivariate analysis of variance yielded a significant (F(1,14) = 10.08, p = .0067) effect on post-shock freezing. Mice that were conditioned at the onset of the light period showed more freezing (49.44 ± 2.79) than mice that were conditioned at the onset of the dark period (63.277 ±3.35). No significant effects were found after 12 hours.
Thirdly, when the mice were tested six hours after contextual fear conditioning, which was immediately after the end of the six hours of sleep deprivation, also no significant difference in freezing (F(1,14) = 0.40, p = .538) was found between mice that were sleep deprived (45.09 ± 6.41) and mice that were given undisturbed sleep (39.52 ± 6.05). These results are shown in figure 5 (left graph: freezing). Additionally, since the data of this protocol was analyzed with Matlab software rather than manually, the motion of the mice could be quantified. Similar to the freezing values, no significant differences in motion (F(1,14) = 0.001, p = .981) were found between mice that were sleep deprived (88.06 ± 17.47) and mice that had undisturbed sleep (88.57 ± 17.47). Figure 5 (right graph: motion) depicts the average motion pixels that mice moved during the test session, six hours after conditioning. Finally, the mice did not differ in basal motion (F(1,14) = 1.44, p = .25), nor in their amount of motion (F(1,14) = 0.24, p = .634) and freezing (F(1,14) = 0.59, p = .457) immediately after the shock, so the mice did not behavior differently before the sleep intervention.
Figure 5. The effects of undisturbed sleep (N = 8) and sleep deprivation (N = 8) on a) freezing levels and b)
motion 6 hours after contextual fear conditioning (3x0.40mA, 1-sec). No significant effects were found.
EXPERIMENT 2: THE ROLE OF GLUR3 IN STRESS RESPONSE MEDIATED MEMORY To examine how epinephrine modulated freezing behavior in response to a mild shock and whether this modulation is mediated by GluR3 a MANOVA was done with genotype (WT vs. KO) and injection solution (saline vs. epinephrine) as between-subject factors and pre-shock freezing and freezing as dependent measures.
A main effect of injection solution on pre-shock freezing was found (F(1,33) = 7.159, p = .012). Animals that were injected with epinephrine showed a significant higher (3.896 ± 1.16) percentage immobility the two minutes before the first shock was delivered than the animals that were injected with saline (0.750 ± 0.17). These results are shown in figure 6. This suggests that the injection with epinephrine can elicit immobility, which is possibly freezing behavior. However, the effect size is relatively small and therefore the effects is less relevant.
Figure 6. The effect of epinephrine on immobility before delivery of the foot-shocks. Mice that are injected with
epinephrine (N = 18) show significantly (p = .012) more immobility than mice injected with saline (N = 20). * p < .05
Epinephrine might influence the sensitivity to shock, as a result of which the memory formation is affected. Therefore, the behavioral response to the shock was examined. The Matlab based software mousemove (Kopec et al., 2007) calculates the number of pixels that change per frame compared to the previous frame. Figure 7b shows an example of the significant motion pixels (SMPs) over time. Two minutes after the mice are put in the fear conditioning box the shock is delivered. This is reflected by a peak in the number of pixels that change significantly. This is considered as the response to the shock (‘shock motion’). Shock motion is calculated by extracting the maximum value in SMP per mouse within a time window of five seconds before until five seconds after delivery of the shock. The averages per condition are depicted in figure 7a. A two-way-ANOVA with injection solution (epinephrine vs. saline) and genotype (GluR3 KO vs. GluR3 WT) as between subject factors and motion (SMP) as dependent variable revealed no significant (F(3,34) = 1.631, p = .200) difference. This indicates that the response to the mild foot shock is the same in mice of both genotypes and that epinephrine does not modulate the sensitivity to the shock. The results are shown in figure 7a.
Figure 7b. An example of the amount of motion over time during the contextual fear conditioning session. The
total duration of the conditioning session was 4 minutes. After two minutes the shock was delivered. The response to the shock is reflected by increased motion at 120 seconds.
Figure 7a. The response to the foot shock in SMPs. There was no significant difference in amount of movement
at the moment of the shock between mice with epinephrine or saline injections, nor between GluR3 KO mice and wild type littermates.
A significant interaction effect of genotype (WT vs. KO) x injection solution (saline vs. epinephrine) was found on test freezing (F(1,34) = 5.763, p = .022). This indicates that epinephrine had a different effect on freezing behavior in GluR3 KOsthan it had in wild type littermates. Additional independent t-tests revealed two significant contrasts. First, GluR3 KOmice showed significantly more (24.79 ± 4.23) freezing than their wild type littermates (10.79 ± 2.72), but only when injected with epinephrine ( t (17) = 2.893, p = .010) and not with saline (p = .721). This contrast represent the effects of GluR3 under influence of epinephrine and indicates that GluR3 has a suppressing effect on freezing. Secondly, GluR3 KOmice showed significantly ( t (16) = 3.651, p = .002) more freezing when injected with epinephrine (24.79 ± 4.23) compared to saline injections (7.77 ± 1.867). This difference in
freezing between epinephrine and saline is not present in wild type littermates (p = .692). This second effect shows that the increased freezing of the GluR3 KO mice is specific for
epinephrine, since the GluR3 KOdo not show this increase in freezing behavior when injected with saline. The results are depicted in figure 8.
Figure 8. The effect of epinephrine on freezing levels 2 hours after contextual fear conditioning in GluR3 KO
mice and wild type littermates. GluR3 KO mice freeze significantly more with epinephrine than with saline (p = .002) and significantly more than wild type littermates with epinephrine (p = .010). * p < .05, ** p < .01
No significant effects of epinephrine or genotype on motion are found. Motion and
freezing are both measures of the amount of the behavioral activity of mice. Often high levels of freezing are associated with low levels of motion. However, it is possible that differences are found in freezing, but not in motion. This indicates that mice that freeze more, also move more during the time that they do not freeze.
Conclusions and discussion
EXPERIMENT 1: THE ROLE OF GLUR3 IN SLEEP DEPENDENT MEMORY
The goal of this experiment was to investigate whether the replacement of GluR1/2s by GluR2/3 depends on sleep. However, preceding that goal it needed to be replicated that sleep
deprivation impairs contextual fear conditioning in wild type mice. Unfortunately, the different sleep deprivation and fear conditioning protocols did not yield any results in wild type mice. The absence of expected results could be due to a couple of factors. Firstly, the number of animals that were used in previous studies are higher, varying from 12 (Silva et al., 2004) to 23 (Chang et al., 2013) per group. However, regarding the small variance in the data of this study it not plausible that adding more animals would yield effects.
Secondly, stress can impair learning and memory consolidation (Trammel & Clore, 2014). The papers (Chang et al., 2013; Silva et al., 2004) that found effects of sleep deprivation on memory consolidation after contextual fear conditioning in mice used sleep deprivation protocols that are known to induce high levels of stress (for references, see Leenaars et al., 2011). It is possible that the previously found impairments on contextual memory
consolidation were partly caused by stress induced by the method of sleep deprivation. On the contrary, the study that used the natural sleep/wake cycle of mice found memory enhancing effects after natural sleep as opposed to wakefulness (Cai et al., 2009). This study shows that effects on memory are also found in sleep deprivation protocols that elicit no to little stress, namely the natural sleep-wake cycle of mice. Therefore, it is not plausible that stressful sleep deprivation protocols only, can explain impaired memory consolidation.
Thirdly, although sleep deprivation with the semi-automated apparatus has proven effective with rats (Leenaars et al., 2011) it might be the case that frequent interruption of sleep is not sufficient to interfere with memory consolidation in mice. Short episodes of sleep in between sleep fragmentation have shown to be sufficient to attenuate learning impairments (Lahl et al., 2008; Mednick et al., 2003). Micro sleeps occur after sustained wakefulness and refer to short periods (3 to 15 seconds) in which the animal is unresponsive to external stimuli together with an EEG pattern that resembles the slow waves seen in sleep (Tirunahari et al., 2003). The rotation velocity used in this study allowed the mice to maximally sleep for 7.5 seconds. This might have been sufficient to rescue memory consolidation. To control for the possible confound of micro sleeps, EEG measurement should be recorded in future research, together with possible alterations in sleep deprivation method that reduce the ability for micro sleeps. One alternative method for sleep deprivation is enriched environment (Vyazovskiy, Borbéli & Tobler, 2000). This method draws upon the natural tendency of mice to explore novel objects and has shown to effectively prolong wakefulness in mice for six hours (Vyazovskiy et al., 2004). The active interaction with the environment is thought to recruit more neuronal capacity, thereby diminishing the prevalence of local and micro sleeps.
Independent memory mechanisms
Alternatively, it is possible memory of wild type animals in contextual fear conditioning is not susceptible to sleep deprivation. In a recent review Bannerman and colleagues (2014) proposed that short and long term memory are two independent mechanisms, rather than two sequential stages of one memory system. This claim is based on evidence that short term spatial memory (spatial working memory) is not a prerequisite for long term spatial memory (spatial reference memory). If short and long term memory were two sequential stages, the latter could not emerge in absence of the first. Mice lacking AMPA-R subunit GluR1 do not show short term memory in a novelty preferences task (Sanderson et al., 2009), nor ten minutes after contextual fear conditioning (Supplementary Figure S1.). However, GluR1 KO mice do show memory 24 hours after the last exposure in a novelty preference task
(Sanderson et al., 2009) and 24 hours after contextual fear conditioning (Supplementary Figure S1.). Also, performance on the Morris water maze, which is a spatial memory task that is acquired over consecutive days, is unimpaired in GluR1 KO mice (Zamanillo et al., 1999). This indicates that GluR1 is crucial for short term memory, whereas long term memory is independent of GluR1.
The neuronal mechanism underlying long term memory is yet unknown, although
independent of GluR1. It is shown that LTP induction in the hippocampus of GluR1 KO mice is severely impaired (Zamanillo et al., 1999), which is consistent with the absence of short term memory in these mice. Interestingly, a more recent paper shows that hippocampal LTP can be induced independent of GluR1 (Romberg et al., 2009). The GluR1 dependent LTP can be induced with a single tetanus stimulation, whereas GluR1 independent LTP requires multiple tetanus stimulations (Romberg et al., 2009). The GluR1 independent form of LTP is mediated by GluN2B, via which PKC phosphorylates GluR2 (Romberg et al., 2009). In the absence of GluR1, GluR2 mainly forms GluR2/3 heteromers. This suggests that the GluR1 independent LTP involves GluR3. In the light of the distinction between GluR1 dependent short and GluR1 independent long term memory, extending the hypothesized involvement of GluR3 in LTP suggests a possible role of GluR3 in long term memory.
Besides two independent memory systems Bannerman and colleagues (2014) propose that these mechanisms compete against each other. GluR1 KO mice outperform wild type
littermates on long term memory tasks (Sanderson et al., 2009). This indicates that mice lacking GluR1 have an advantage on tasks that require long term memory. Similarly, data
from our laboratory show that GluR3 KO mice have enhanced short term memory compared to wild types after contextual fear conditioning (M. Renner, unpublished data, see:
Supplementary Figure S2.). Thus, mice lacking GluR3 appear to have an advantage on tasks that require immediate memory. Together, these data suggest that mice lacking a specific AMPA subunit have an advantage on a memory task that recruits another AMPA subunit. The immediate short term memory is dependent on GluR1 (Bannerman et al., 2014; Sanderson et al., 2009; Supplementary Figure S1.), whereas slowly emerging long term memory could rely on synaptic insertion of GluR2/3s. However, this is very indirect evidence and the hypothesis is fairly speculative, since other forms of plasticity could also account for the
GluR1-independent form of learning. Memory specific role of sleep
It is conceivable that GluR1 dependent immediate memory is not susceptible to sleep deprivation, whereas slow, hypothetically GluR3 dependent, memory would be affected by sleep deprivation. Namely, the slow GluR1 independent memory occurs gradually over the course of 24 hours. This time course coincides with the presence of sleep. Also, the levels of GluR3 are increased after sleep. The slowly occurring memory resembles the slow and constitutive trafficking of GluR2/3s in and out of the synapse. Together, this suggests that sleep facilitates slow memory and that GluR3 might be involved. If so, sleep deprivation would affect slow memory, not immediate memory. Contextual fear conditioning is an immediate memory task, which would not be affected by sleep deprivation in mice that possess both memory systems, that is, wild type mice.
However, assuming that the two memory systems compete with each other, mice lacking one memory systems would be able to acquire memories with the other memory system. Mice lacking immediate memory, that is, GluR1 KO mice, depend on slow long term system for memory acquisition. This can explain the slow learning curve of GluR1 KO mice after contextual fear conditioning, even though fear conditioning primarily recruits immediate memory. Next, we will test whether slow memory is facilitated by sleep. Also, we will test whether a memory task that draws upon slow memory is more susceptible to sleep
deprivation. Contextual fear conditioning is a single trial, immediate memory dependent task. Memory tasks that require multiple trials, for example, maze learning, might be driven by the mechanisms underlying slow learning. Therefore it is plausible that if slow memory is
Replacement vs. competition
The proposed hypothesis regarding independent memory systems (Bannerman et al., 2014) sheds a new light on hypothesis that presumes replacement of GluR1/2s by GluR2/3s. The data from the current study show that sleep deprivation in wild type mice does not affect memory consolidation after contextual fear conditioning. This can be explained by the
presence of two independent memory systems, of which the slow GluR1 independent memory theoretically would be susceptible to sleep deprivation, whereas the immediate GluR1
dependent memory would be resistant to sleep deprivation.
Furthermore, the replacement hypothesis postulates a stabilizing role of GluR3 in memory consolidation. From this perspective, it is predicted that the memory of GluR3 KO mice would become less stable over time. However, data from our laboratory contradicts this prediction. The memory of mice lacking GluR3 remains stable up to at least one month after contextual fear conditioning (M. Renner, unpublished data, see: Supplementary Figure S2).
Also, the gradual replacement of GluR1-containing AMPA-Rs by the slow, constant cycling of GluR3-containing AMPA-Rs (Shi et al., 2001), presumes a conditional role of GluR1 and short term memory in the sequential transition GluR2/3 replacement to long term memory. However, mice lacking GluR1-containing AMPA-Rs are still capable of synaptic insertion of GluR2/3s and memory acquisition. On the contrary, the suggestion of
placeholders (Shi et al., 2001) could forestall this argument.
Moreover, replacement of GluR1/2s by GluR2/3s does not provide an explanation for the outperforming effects of the GluR3 and GluR1 KO mice compared to their wild type
littermates. The model of competing memory systems can explain the enhanced performance of GluR3 KO mice on immediate memory tasks and the performance of GluR1 KO mice on long term memory tasks.
Lastly, both replacement as well as competing memory systems can provide a theoretical view that makes it conceivable that synaptic insertion of GluR3 is facilitated by sleep. However, the replacement model provides a more direct reason for GluR3 to benefit from sleep. Namely, the absence of new experiences results in a decreased synaptic insertion of GluR1, as a consequence of which the relatively slow turnover of GluR2/3 can replace the quickly inserted GluR1-containing subunits. The role of GluR3 in the dual-memory model is still speculative, since there is only scarce and indirect evidence for the involvement of GluR3 in slow memory. Also, the role of sleep in slow memory remains elusive. On the contrary,
both models argue that the absence of GluR1 due to sleep prevents GluR1-containing AMPA-R to occupy the synapse. In the replacement model this leads to a ‘catch up’ opportunity for GluR2/3 replacement, whereas in the competition model this ‘allows’ GluR3 to form memories. The model of independent dual-memory systems can account for all of the discussed findings and is therefore considered more conceivable.
In summary, it is plausible that long and short term memory represent independent mechanisms. AMPA-R subunits GluR1 and GluR3 could underlie these competing memory systems. Theoretically, sleep can selectively facilitate long, but not short term memory, possibly via synaptic insertion of GluR3. However, the coupling of GluR1 and GluR3 to respectively immediate and slow memory is still highly speculative. Also, the mechanisms for synaptic insertion of GluR3 and the proposed role of sleep therein remain elusive and are subject to further investigation.
EXPERIMENT 2: THE ROLE OF GLUR3 IN STRESS RESPONSE MEDIATED MEMORY The goal of this experiment was to examine the role of GluR3 in the stress related memory formation. Preliminary data from our laboratory has shown a transient increase in synaptic efficacy upon fear conditioning in GluR1 KO mice. These mice do not show memory ten minutes after fear conditioning, which suggests that the transient increase in synaptic efficacy does not contribute to memory formation. We hypothesized that the
emotional component of fear conditioning, the stress response, could contribute to this effect. The transient increase in synaptic efficacy is absent in GluR3 KO mice. This points towards a role of GluR3 in the emotional component of fear conditioning.
The data of this experiment show that epinephrine in the absence of GluR3 enhances memory. However, in the presence of GluR3 this memory enhancement is attenuated. This indicates that GluR3 suppresses memory. The difference in memory performance between genotypes is absent in the absence of epinephrine. The modulatory effect of GluR3 are specific for epinephrine.
Downstream of epinephrine multiple different hormonal and neuromodulatory effects are affected. Increased levels of epinephrine release norepinephrine (NE) from the locus
coeruleus, activate the autonomous nervous system, which via the HPA-axis releases
has been shown that norepinephrine, via a downstream pathway involving cAMP and PKA phosphorylates GluR1 (Hu et al., 2007). This explains the memory enhancing effects of epinephrine in mice lacking GluR3.
However, regarding the versatile effects of epinephrine on different neurochemical components, epinephrine injections are not specific enough to determine based on the current data which neuromodulatory component is responsible for activation of the GluR3 mediated memory suppression. The candidate pathways are norepinephrine and corticosteroids. Firstly, norepinephrine acts on β-adrenergic receptors downstream of which cAMP and PKA activate GluR1 (Hu et al., 2009). Another downstream pathway of norepinephrine includes CaMKII and subsequent PKC activation (Hu et al., 2007). PKC is known to interact with AMPA-R subunit GluR2 thereby regulating surface expression of AMPA-Rs (Romberg et al., 2009). Thus, norephinephrine release could indirectly affect GluR2/3 complexes.
Secondly, corticosteroids can regulate AMPA trafficking and memory (Krugers,
Hoogenraad & Groc, 2010). Upon release of corticosterone a rapid non-genomic effect arises, which is caused by mineralocorticoid receptors. These mineralocorticoids activate synaptic insertion of AMPA-Rs, enhance LTP and are thought to underlie increased memory encoding. Also, slow genomic effects are induced and involve glucocorticoid receptors. The
glucocorticoids regulate lateral diffusion of AMPA-Rs, decrease LTP and facilitate LTD. These processes are associated with reduced memory encoding for event following a high salience stressful event.
Interestingly, recent in vitro data from our laboratory has shown that norepinephrine only is not sufficient to cause the massive increase in mEPSC frequency that is seen shortly after fear conditioning. If the GluR3 dependent memory suppression is related to the GluR3 mediated massive and transient increase in mEPSC frequency, norepinephrine might not be the only neuromodulatory component via which increased levels of epinephrine affect the GluR3 mediated memory suppression. This will be examined further in future studies.
The suppressed memory performance could be related to the genomic effects of glucocorticoids on the synaptic incorporation of GluR2-containing AMPA-Rs that is associated with decreased LTP induction and facilitation of LTD (Krugers, Hoogenraad & Groc, 2010). However, the time course of the slow genomic effects (i.e. app. 30 minutes and on) of glucocorticoids on AMPA trafficking and synaptic plasticity is not consistent with the timing of the relatively short interval (i.e. ten minutes) between the epinephrine injections and the fear conditioning, that is, the moment of memory encoding. Regarding this interval it
would be expected that the rapid non-genomic effects of mineralocorticoids, associated with an increased LTP, would cause enhanced memory encoding. In contrast, the time course of the rapid non-genomic effects of mineralocorticoids on synaptic insertion of AMPA-Rs is consistent with the transient increased synaptic efficacy ten minutes after fear conditioning. However, there is no specific evidence that GluR3-containing AMPA-Rs are involved in the rapid effects that mineralocorticoids induce. The rapid activity-dependent synaptic insertion of AMPA-Rs is merely associated with GluR1-containing AMPA-Rs (Kessels & Malinow, 2009). In conclusion, there are some indirect indications for both norepinephrine as well as corticosteroids to be involved in GluR3 mediated memory suppression. Nevertheless, the exact signaling pathway remains to be determined.
Notably, caution should be taken since the association between the increase in synaptic efficacy and memory formation is not yet established. Although both effects are mediated by GluR3 and both the attenuated memory performance and the increase in mEPSC frequency are been shown to depend on epinephrine, it remains that the increase in synaptic efficacy is only hypothetically consequence of the stress response. Thus, even though both findings can be related to GluR3 and are due to increased levels of epinephrine, they might be due to different mechanisms downstream of epinephrine. Therefore, in the process of investigating which pathway, downstream of epinephrine, affects GluR3 functioning and, hence, memory formation, it should be noted that in vitro results are not directly translations to in vivo processes and vice versa.
In summary, the data from this experiment shows that GluR3 is involved in modulating memory formation during stress response. More specifically, the data suggest that GluR3 during a stressful experience suppresses memory formation or consolidation. Together with the GluR3 mediated massive, transient increased synaptic efficacy that occurs after contextual fear conditioning, this indicates that during stressful situation GluR3 mediates an increase in synaptic efficacy and attenuates memory formation.
GENERAL DISCUSSION
The function of AMPA subunit GluR3 in synaptic plasticity is scarcely studied. The aim of this study was to investigate the function of GluR3 in learning and memory consolidation. The results show that sleep does not play a critical role in consolidation of contextual fear memories in wild type mice. These data, together with contextual fear conditioning data of
mice lacking either GluR1 or GluR3, are inconclusive with respect to the hypothesized replacement of GluR1/2s by GluR2/3s. In addition, the data can also point towards the presence of two independent memory systems: one reflecting immediate, short term memory and the other representing slowly emerging long term memory. These memory systems can be driven by respectively GluR1 and GluR3. Also, the role of GluR3 in memory formation during stress was examined. These data show that GluR3 attenuates memory formation under influence of epinephrine. The effect of GluR3 on memory formation can be mediated by norepinephrine or corticosteroids, which are released upon increased levels of epinephrine.
If GluR1 and GluR3 represent respectively immediate and slow memory, and if these memory systems compete against each other for the encoding of memories, the attenuated memory performance of wild type littermate mice might be explained by the supposed competition between both systems. GluR3 KO mice lack this competition, since these mice mainly possess GluR1/2 heteromers and GluR1 is essential for the acquisition of quick, immediate memories. The increased levels of epinephrine and downstream release of norepinephrine cause the phosphorylation of GluR1 (Hu et al., 2007). Therefore GluR3 KO mice exhibit increased memory performance compared to the memory performance of GluR3 KO mice in the vehicle condition. Thus, in GluR3 KO mice the epinephrine mediated
enhanced GluR1 is, in the absence of competing GluR3, able to encode large part of the fear memory.
Wild type mice have both memory systems. In these mice the immediate GluR1 dependent memory system is also enhanced through increased levels of epinephrine. However, if the gradually occurring, hypothetically GluR3 dependent, memory system competes with the immediate memory system and has occupied part of the fear memory two hours after memory encoding, yet without expression of the memory, this can explain the attenuated memory performance of wild type mice. Following this line of thought, the wild type mice should have equal memory performance to the GluR3 KO mice when the slow memory has fully emerged, that is, 24 hours after conditioning.
Thus, either GluR3 can have an intrinsically suppressing effect on memory formation independent of the time of memory retrieval, or the hypothetically GluR3 dependent memory is gradually emerging over the course of 24 hours and is therefore not fully present on the short term (2 hours after conditioning). These possibilities can be examined with the use of a mouse model that lacks both GluR1 and GluR3. If GluR3 suppresses memory formation during a physiological stress response, the mice lacking both GluR1 and GluR3 should show
enhanced memory performance on long term memory (24 hours) compared to mice lacking solely GluR1, since no memory suppression takes place. Additionally, this also advocates that slow memory is caused by other mechanisms than GluR3, since long term memory would still be present in the absence of GluR3. On the contrary, if GluR3 reflects a memory system that slowly occurs after conditioning, mice lacking both GluR1 and GluR3 should show decreased long term memory (24 hours) compared to mice lacking solely GluR1.
Importantly, the caveat of working with global knockout mice models is the possibility that compensatory mechanisms might alter the functioning of the genetically modified nervous system, thereby not reflecting the same mechanisms as are present in wild type animals. To confirm future results, it would therefore be good to work with conditional knockout mice models.
In conclusion, there is little known about the function of AMPA-R subunit GluR3 in learning and memory. This study attempted to shed a light the role of GluR3 in memory consolidation and the role of sleep therein. Also, the influence of GluR3 on stress mediated memory formation was examined. Together these results suggested that GluR3 could represent a slowly emerging memory system or that GluR3 has a suppressing effect on memory formation. However, a lot needs to be studied in order to draw conclusions on the proposed functions of GluR3. Among these topics for further investigation are the questions whether GluR3 plays a role in GluR1 independent slow memory. Also, whether and if so, how this long term memory is facilitated by sleep and whether sleep promoted synaptic insertion of GluR3. Finally, the mechanisms underlying the GluR3 mediated suppression of memory formation in stress related learning and whether this finding is compatible with the hypothesized role of GluR3 in long term remains subject to further investigation.
References
Abel, T., & Lattal, K. M. (2001). Molecular mechanisms of memory acquisition,
consolidation and retrieval. Current Opinion in Neurobiology, 11(2), 180–7. Retrieved from http://www.ncbi.nlm.nih.gov/pubmed/11301237
Bannerman, D. M., Sprengel, R., Sanderson, D. J., McHugh, S. B., Rawlins, J. N. P., Monyer, H., & Seeburg, P. H. (2014). Hippocampal synaptic plasticity, spatial memory and anxiety. Nature Reviews. Neuroscience, 15(3), 181–92. doi:10.1038/nrn3677
Benke, T.A., Lüthi, A., Isaac, J.T.R., & Collingridge, G.L. (1998). Modulation of AMPA receptor unitary conductance by synaptic activity. Nature, 393(6687), 793-7. doi: 10.1038/31709
Byrne, J.H. (Ed.). (2008). Learning and memory: a comprehensive reference. Oxford: Academic Press.
Cahill, L., Gorski, L., & Le, K. (2003). Enhanced Human Memory Consolidation With Post-Learning Stress: Interaction With the Degree of Arousal at Encoding. Post-Learning & Memory, 10(4), 270–274. doi:10.1101/lm.62403
Cai, D. J., Shuman, T., Gorman, M. R., Sage, J. R., & Anagnostaras, S. G. (2009). Sleep selectively enhances hippocampus-dependent memory in mice. Behavioral
Neuroscience, 123(4), 713–719. doi:10.1037/a0016415
Chang, H.-F., Su, C.-L., Chang, C.-H., Chen, Y.-W., & Gean, P.-W. (2013). The beneficial effects of leptin on REM sleep deprivation-induced cognitive deficits in mice. Learning & Memory (Cold Spring Harbor, N.Y.), 20(6), 328–35. doi:10.1101/lm.030775.113 Dinh, L., Nguyen, T., Salgado, H., & Atzori, M. (2009). Norepinephrine homogeneously
inhibits alpha-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate- (AMPAR-) mediated currents in all layers of the temporal cortex of the rat. Neurochemical Research, 34(11), 1896–1906. doi:10.1007/s11064-009-9966-z
Dingledine, R., Borges, K., Bowie, D., & Traynelis, S. F. (1999). The glutamate receptor ion channels. Pharmacological Reviews, 51(1), 7–61. Retrieved from
http://www.ncbi.nlm.nih.gov/pubmed/10049997
Feyder, M., Wiedholz, L., Sprengel, R., & Holmes, A. (2007). Impaired associative fear learning in mice with complete loss or haploinsufficiency of AMPA GluR1 receptors. Frontiers in Behavioral Neuroscience, 1(December), 4. doi:10.3389/neuro.08.004.2007 Graves, L. a., Heller, E. a., Pack, A. I., & Abel, T. (2003). Sleep Deprivation Selectively
Impairs Memory Consolidation for Contextual Fear Conditioning. Learning & Memory, 10(3), 168–176. doi:10.1101/lm.48803.LEARNING
Hagewoud, R., Bultsma, L. J., Barf, R. P., Koolhaas, J. M., & Meerlo, P. (2011). Sleep deprivation impairs contextual fear conditioning and attenuates subsequent behavioural, endocrine and neuronal responses. Journal of Sleep Research, 20(2), 259–266.
doi:10.1111/j.1365-2869.2010.00895.x
Hagewoud, R., Whitcomb, S. N., Heeringa, A. N., Havekes, R., Koolhaas, J. M., & Meerlo, P. (2010). A time for learning and a time for sleep: the effect of sleep deprivation on contextual fear conditioning at different times of the day. Sleep, 33(10), 1315–1322. Hayashi, Y., Shi, S.H., Esteban, J.A., Piccini, A., Poncer, J.C., & Malinow, R. (2000).
Driving AMPA receptors into synapses by LTP and CaMKII: requirement for GluR1 and PDZ domain interaction. Science, 287, 2265-2267.
Hebb, D.O. (1949). The organization of behavior: a neuropsychological theory. Oxford: Wiley.
Hu, H., Real, E., Takamiya, K., Kang, M. G., Ledoux, J., Huganir, R. L., & Malinow, R. (2007). Emotion Enhances Learning via Norepinephrine Regulation of AMPA-Receptor Trafficking. Cell, 131(1), 160–173. doi:10.1016/j.cell.2007.09.017
Humeau, Y., Reisel, D., Johnson, A. W., Borchardt, T., Jensen, V., Gebhardt, C., … Lüthi, A. (2007). A pathway-specific function for different AMPA receptor subunits in amygdala long-term potentiation and fear conditioning. The Journal of Neuroscience : The Official Journal of the Society for Neuroscience, 27(41), 10947–10956.
doi:10.1523/JNEUROSCI.2603-07.2007
Kandel, E.R., Schwartz, J.H., & Jessell, T.M. (2000). Principles of Neural Science (4th ed.). New York: McGraw-Hill.
Kessels, H. W., & Malinow, R. (2009). Synaptic AMPA receptor plasticity and behavior. Neuron, 61(3), 340–50. doi:10.1016/j.neuron.2009.01.015
Kopec, C. D., Kessels, H. W. H. G., Bush, D. E. a, Cain, C. K., LeDoux, J. E., & Malinow, R. (2007). A robust automated method to analyze rodent motion during fear conditioning. Neuropharmacology, 52(1), 228–233. doi:10.1016/j.neuropharm.2006.07.028
Krugers, H. J., Hoogenraad, C. C., & Groc, L. (2010). Stress hormones and AMPA plasticity and memory, 10. Retrieved from http://dx.doi.org/10.1038/nrn2913
Lahl, O., Wispel, C., Willigens, B., & Pietrowsky, R. (2008). An ultra short episode of sleep is sufficient to promote declarative memory performance. Journal of Sleep Research, 17(1), 3–10. doi:10.1111/j.1365-2869.2008.00622.x
Leenaars, C. H. C., Dematteis, M., Joosten, R. N. J. M. A, Eggels, L., Sandberg, H., Schirris, M., … Van Someren, E. J. W. (2011). A new automated method for rat sleep deprivation
with minimal confounding effects on corticosterone and locomotor activity. Journal of Neuroscience Methods, 196(1), 107–117. doi:10.1016/j.jneumeth.2011.01.014
Leenaars, C. H. C., Girardi, C. E. N., Joosten, R. N. J. M. A, Lako, I. M., Ruimschotel, E., Hanegraaf, M. a J., … Van Someren, E. J. W. (2013). Instrumental learning: An animal model for sleep dependent memory enhancement. Journal of Neuroscience Methods, 217(1-2), 44–53. doi:10.1016/j.jneumeth.2013.04.003
Lohmann, C., & Kessels, H. W. (2013). The developmental stages of synaptic plasticity. Journal of Physiology, 592(Pt 1), 13–31. doi:10.1113/jphysiol.2012.235119
Lømo, T. (1966). Frequency potentiation of excitatory synaptic activity in the dentate area of hippocampal formation [abstract]. Acta Physiologica Scandinavica, 68, 128.
Lu, W., Shi, Y., Jackson, A. C., Bjorgan, K., During, M. J., Seeburg, P. H., & Nicoll, R. A. (2013). NIH Public Access, 62(2), 254–268. doi:10.1016/j.neuron.2009.02.027.Subunit McGaugh, J.L. (2002). Memory—a century of consolidation. Science, 287, 248-251.
Malenka, R.C. & Nicoll, R.A. (1999). Long-term potentiation—a decade of progress? Science, 285(5435), 1870-4.
Malinow, R., Mainen, Z. F., & Hayashi, Y. (2000). LTP mechanisms: From silence to four-lane traffic. Current Opinion in Neurobiology, 10(3), 352–357.
Malinow, R., & Malenka, R. C. (2002). AMPA receptor trafficking and synaptic plasticity. Annual Review of Neuroscience, 25, 103–26.
doi:10.1146/annurev.neuro.25.112701.142758
Mednick, S., Nakayama, K., & Stickgold, R. (2003). Sleep-dependent learning: a nap is as good as a night. Nature Neuroscience, 6(7), 697–8. doi:10.1038/nn1078
Plant, K., Pelkey, K. a, Bortolotto, Z. a, Morita, D., Terashima, A., McBain, C. J., … Isaac, J. T. R. (2006). Transient incorporation of native GluR2-lacking AMPA receptors during hippocampal long-term potentiation. Nature Neuroscience, 9(5), 602–604.
doi:10.1038/nn1678
Romberg, C., Raffel, J., Martin, L., Sprengel, R., Seeburg, P. H., Rawlins, J. N. P., … Paulsen, O. (2009). Induction and expression of GluA1 (GluR-A)-independent LTP in the hippocampus. The European Journal of Neuroscience, 29(6), 1141–52.
doi:10.1111/j.1460-9568.2009.06677.x
Rumpel, S., Ledoux, J., Zador, A., & Malinow, R. (2005). Postsynaptic Receptor Trafficking Underlying a Form of Associative Learning, 308(April).
Sanderson, D. J., Good, M. a, Skelton, K., Sprengel, R., Seeburg, P. H., Rawlins, J. N. P., & Bannerman, D. M. (2009). Enhanced long-term and impaired short-term spatial memory in GluA1 AMPA receptor subunit knockout mice: evidence for a dual-process memory model. Learning & Memory (Cold Spring Harbor, N.Y.), 16(6), 379–386.
doi:10.1101/lm.1339109.Freely
Takatsu-Coleman, A. L., Zanin, K. a., Patti, C. L., Zager, A., Lopes-Silva, L. B., Longo, B. M., … Frussa-Filho, R. (2013). Short-term sleep deprivation reinstates memory retrieval in mice: The role of corticosterone secretion. Psychoneuroendocrinology, 38(10), 1967– 1978. doi:10.1016/j.psyneuen.2013.02.016
Tirunahari, V. L., Zaidi, S. A., Sharma, R., Skurnick, J., & Ashtyani, H. (2003). Microsleep and sleepiness : a comparison of multiple sleep latency test and scoring of microsleep as a diagnostic test for excessive daytime sleepiness, 4, 63–67. doi:10.1016/S1
Tononi, G., & Cirelli, C. (2014). Sleep and the Price of Plasticity: From Synaptic and Cellular Homeostasis to Memory Consolidation and Integration. Neuron, 81(1), 12–34.
doi:10.1016/j.neuron.2013.12.025
Trammell, J. P., & Clore, G. L. (2014). Does stress enhance or impair memory consolidation? Cognition & Emotion, 28(2), 361–74. doi:10.1080/02699931.2013.822346
Shi, S. H., Hayashi, Y., Esteban, J. a., & Malinow, R. (2001). Subunit-specific rules
governing AMPA receptor trafficking to synapses in hippocampal pyramidal neurons. Cell, 105(3), 331–343.
Silva, R. H., Chehin, a B., Kameda, S. R., Takatsu-Coleman, a L., Abilio, V. C., Tufik, S., & Frussa-Filho, R. (2004). Effects of pre- or post-training paradoxical sleep deprivation on two animal models of learning and memory in mice. Neurobiol.Learn.Mem., 82(1074-7427 (Print) LA - eng PT - Comparative Study PT - Journal Article PT - Research Support, Non-U.S. Gov’t SB - IM), 90–98. doi:10.1016/j.nlm.2004.04.005
Vyazovskiy, V., Borbély, a a, & Tobler, I. (2000). Unilateral vibrissae stimulation during waking induces interhemispheric EEG asymmetry during subsequent sleep in the rat. Journal of Sleep Research, 9(4), 367–71. Retrieved from
http://www.ncbi.nlm.nih.gov/pubmed/11123523
Vyazovskiy, V. V., Cirelli, C., Tononi, G., & Tobler, I. (2008). Cortical metabolic rates as measured by 2-deoxyglucose-uptake are increased after waking and decreased after sleep in mice. Brain Research Bulletin, 75(5), 591–597.
doi:10.1016/j.brainresbull.2007.10.040
Vyazovskiy, V. V, Welker, E., Fritschy, J.-M., & Tobler, I. (2004). Regional pattern of metabolic activation is reflected in the sleep EEG after sleep deprivation combined with
unilateral whisker stimulation in mice. The European Journal of Neuroscience, 20(5), 1363–1370. doi:10.1111/j.1460-9568.2004.03583.x
Wenthold, J., Petralia, S., & Niedzielski, S. (1996). Evidence for Multiple Neurons Complexes in Hippocampal, 76(6), 1982–1989.
Whitehead, G., Jo, J., Hogg, E. L., Piers, T., Kim, D.-H., Seaton, G., … Cho, K. (2013). Acute stress causes rapid synaptic insertion of Ca2+ -permeable AMPA receptors to facilitate long-term potentiation in the hippocampus. Brain : A Journal of Neurology, 136(Pt 12), 3753–65. doi:10.1093/brain/awt293
Zamanillo, D. (1999). Importance of AMPA Receptors for Hippocampal Synaptic Plasticity But Not for Spatial Learning. Science, 284(5421), 1805–1811.
Supplementary Figure S1.
Figure S1. Data from our laboratory showing that mice lacking GluR1 do not show memory ten minutes after
contextual fear conditioning (left). However, 24 hours after contextual fear conditioning mice lacking GluR1 do show memory (right). Together with the memory performance two hours after contextual fear conditioning (middle), these data indicate a slowly emerging memory over the course of 24 hours in the absence of GluR1. ** p < .01, **** p < .0001
Supplementary Figure S2.
Figure S2. Data from our laboratory showing that 1) mice lacking GluR3 (“GluA -/-”) show enhanced memory
performance immediately after contextual fear conditioning compared to wild type littermates (“GluA +/+”) and 2) the memory performance of mice lacking GluR3 remains stable up to at least one month after contextual fear conditioning (M. Renner, unpublished data).
