|Home | About | Journals | Submit | Contact Us | Français|
The type 2 ryanodine receptor/calcium release channel (RyR2), required for excitation-contraction coupling in the heart, is abundant in the brain. Chronic stress induces catecholamine biosynthesis and release, stimulating β-adrenergic receptors and activating cAMP signaling pathways in neurons. In a murine chronic restraint stress model neuronal RyR2 were phosphorylated by protein kinase A (PKA), oxidized and nitrosylated, resulting in depletion of the stabilizing subunit calstabin2 (FKBP12.6) from the channel complex and intracellular calcium leak. Stress-induced cognitive dysfunction, including deficits in learning and memory, and reduced long-term potentiation (LTP) at the hippocampal CA3-CA1 connection, were rescued by oral administration of S107, a novel compound developed in our laboratory that stabilizes RyR2-calstabin2 interaction, or by genetic ablation of the RyR2 PKA phosphorylation site at serine 2808. Thus, neuronal RyR2 remodeling contributes to stress-induced cognitive dysfunction. Leaky RyR2 could be a therapeutic target for treatment of stress-induced cognitive dysfunction.
Acute stress enhances cognitive function via activation of the sympathetic nervous system as part of the “fight-or-flight” response (McEwen and Sapolsky, 1995). In contrast, chronic stress causes cognitive dysfunction associated with glucocorticoid-induced hippocampal neuron damage (McEwen and Sapolsky, 1995). However, the precise mechanisms underlying stress-induced cognitive dysfunction remain poorly understood and there is currently no effective therapy.
Intracellular calcium (Ca2+) homeostasis plays a crucial role in neuron survival and function. Neuronal Ca2+ influx is mediated through voltage-gated channels and/or receptor-operated channels activated by neurotransmitters. Intracellular Ca2+ release occurs via ryanodine receptors (RyRs) and inositol (1,4,5)-trisphosphate receptors (IP3Rs) on the endoplasmic reticulum (ER) (Berridge et al., 2003). There are three isoforms of RyR in the brain. RyR1 and RyR2 are both expressed in the hippocampus. RyR2 expression is up-regulated in rat hippocampus following intensive training tasks (Cavallaro et al., 1997; Zhao et al., 2000). RyR2 knock-down impairs performance of mice in passive avoidance tests (Galeotti et al., 2008) whereas RyR1 knock-down did not (Galeotti et al., 2008).
Ca2+ released into the cytoplasm is pumped back into the ER by the ER Ca-ATPase (SERCA) or removed from the cell by the plasma membrane Ca-ATPase (PMCA) and Na+/Ca2+ exchanger. Altered Ca2+ homeostasis contributes to neuronal degeneration and cell death (Mattson, 2007) and defective postsynaptic signaling, presynaptic control of neurotransmitter release, and long-term synaptic plasticity (Berridge, 1998; Fitzjohn and Collingridge, 2002).
We have shown that binding of the subunit calstabin2 (FKBP12.6) to RyR2 channels in the heart is impaired under stress conditions by PKA phosphorylation at Serine 2808 (Wehrens et al., 2005) and oxidation/nitrosylation of the channel such that calstabin2 depleted RyR2 channels leak intracellular Ca2+ (Shan et al., 2010a; Shan et al., 2010b). A novel class of drugs (derivatives of 1,4-benzothiazepines) known as intracellular calcium release channel stabilizers, or Rycals, (e.g. S107 and JTV-519) (Bellinger et al., 2008b), prevent stress-induced depletion of calstabin from RyR and fix the pathological intracellular Ca2+ leak. Rycals inhibit heart failure progression in mice (Shan et al., 2010a), and prevent catecholamine-induced arrhythmias (Lehnart et al., 2008; Shan et al., 2010a; Wehrens et al., 2004; Yano et al., 2003).
We now show that stress-induced oxidation, nitrosylation and PKA hyperphosphorylation deplete calstabin2 from brain RyR2 resulting in leaky channels that contribute to stress-induced cognitive dysfunction in mice. The orally available small molecule S107 decreased ER Ca2+ leak via neuronal RyR2 and prevented stress-induced cognitive dysfunction. Genetic ablation of the RyR2 PKA phosphorylation site in RyR2-S2808A+/+ mice (Shan et al., 2010b; Wehrens et al., 2005) also protects against stress-induced cognitive dysfunction. Taken together, these results establish a crucial role for leaky neuronal RyR2 in stress-induced cognitive dysfunction and suggest a potential therapeutic approach to preventing stress-induced cognitive dysfunction and post-traumatic stress disorder (PTSD).
WT and RyR2-S2808A+/+ mice were subjected to chronic restraint stress (Copeland et al., 2005; Fukui et al., 1997; Patel et al., 2004) with or without treatment with the Rycal S107 in the drinking water (Figure. 1A). There was a >2-fold sustained increase in plasma corticosterone (CORT) in WT and RyR-S2808A+/+ mice following 3 weeks of chronic restraint stress (Aguilera, 1994; Carrasco and Van de Kar, 2003; McEwen, 2004, 2007) (Figure 1B, S1C). The immediate early gene c-Fos was significantly elevated after acute stress, and to a lesser degree after acute stress following chronic stress indicating some blunting of the HPA axis (Figure S1D-F) consistent with previous reports (Perrotti et al., 2004). The transcription factors CREB and ERK regulate tyrosine hydroxylase and catecholamine biosynthesis, and are phosphorylated under stress conditions (Brazelton et al., 2000; Impey et al., 1998; Liu et al., 2005; Liu et al., 2008; Sabban et al., 2004; Shaywitz and Greenberg, 1999). There was a significant increase in CREB (pCREB) and ERK (pERK) phosphorylation in hippocampal, cortical, and whole brain from stressed mice, consistent with a general stress response in the brain (Figure 1C-E, S2) and increased PKA activity (Figure S2G). PKA catalytic or regulatory subunits were unchanged (Figure 1E).
CRF, ACTH and CORT were significantly increased at 15 min after termination of chronic stress in all stressed groups (Figure S1A-C). CRF and ACTH fell to control levels one hr post-chronic stress (Figures S1A, B), however CORT levels remained significantly elevated at 24 hrs post-stress indicating a persistent stress response (Figure S1C). There was a significant but blunted elevation of CRF, ACTH and CORT after acute stress in chronically stressed WT and RyR2-S2808A+/+ mice (Figure S1A-C), indicating that the HPA-axis was intact but blunted (Figure S1A-C).
Neuronal RyR2 were PKA hyperphosphorylated, oxidized, S-nitrosylated and depleted of calstabin2 in hippocampal and whole brain (Figures 2A-E, S3A-E). Neuronal ER SERCA activity was unchanged in stressed mice (Figure S3F). S107 given in drinking water (75 mg/kg/d) prior to and during the stress protocol prevented calstabin2 depletion from the RyR2 macromolecular complex without altering PKA phosphorylation, oxidation or S-nitrosylation of RyR2 (Figure 2E). S107 improves cardiac and skeletal muscle function (Andersson et al., 2011; Bellinger et al., 2009; Shan et al., 2010a; Shan et al., 2010b). To show that the effects of S107 in the present study were due to its neural as opposed to extraneural actions, mice were treated with a related Rycal (S36, 75 mg/Kg/day) that also improves cardiac and skeletal muscle function but does not cross the blood-brain-barrier. In contrast to S107, S36 did not prevent calstabin2 depletion from neuronal RyR2 in hippocampal samples from stressed mice (Figure S4A-E).
PKA, H2O2 and Noc-12 individually caused a ~2-fold decrease in the binding affinity of calstabin2 to RyR2 (p<0.05) (Figures 2F-H, S5A,B), and in combination a ~ 4-fold decrease in the binding affinity of calstabin2 to RyR2 (p<0.01) (Figures 2H, S5A,B) as determined by 35S-calstabin2 binding, and ~80% depletion of calstabin2 from RyR2 assessed by co-immunoprecipitation (p<0.01) (Figure 2F,G). Anti-calstabin2 antibody co-immunoprecipitated RyR2 (Figure S5C). There was no stress-induced change in total neuronal RyR2 or calstabin2 (Figure S5D).
PKA phosphorylation, oxidation and nitrosylation of RyR2 cause calstabin2 depletion from cardiac RyR2 during heart failure (Lehnart et al., 2004; Shan et al., 2010a; Wehrens et al., 2004). Mice harboring RyR2 channels that cannot be phosphorylated by PKA (RyR2-S2808A+/+) are protected against heart failure progression (Shan et al., 2010b; Wehrens et al., 2006). Stress-induced depletion of calstabin2 from hippocampal or whole brain RyR2 was not observed in RyR2-S2808A+/+ mice (Figures 2A-E, S3A-E). Thus, S107 or genetic ablation of the RyR2 PKA phosphorylation site prevent calstabin2 depletion from neuronal RyR2 in stressed mice.
The stress-induced remodeling of the cardiac RyR2 macromolecular complex results in “leaky” channels (Lehnart et al., 2008; Shan et al., 2010a; Wehrens et al., 2004). We sought to determine whether stress-induced remodeling of hippocampal RyR2 channels also results in leaky channels with higher single-channel open probabilities (Po) under resting conditions (e.g. with low cytosolic activating [Ca2+]). In WT control mice the RyR Po was essentially zero at low cytosolic [Ca2+] (150 nM) (Figure 3A and F), whereas RyR channels from stressed mice exhibited a significantly increased Po (Figure 3B and F). RyR channels from S107 treated stressed mice exhibited low Po, similar to non-stressed mice (Figure 3C and F). RyR in non-stressed and stressed RyR2-S2808A+/+ mice showed no increased Po (Figure 3D-F). Whole brain RyR exhibited similar increases in single-channel Po and normalization when mice were treated with S107 compared to hippocampal RyR channels (Figure S3H).
Stress significantly increased ER Ca2+ leak (33.13 ± 4.2% of uptake in stress, n=7, versus 21.75 ± 1.82 % of uptake in non-stress, n=9, p<0.05 by one-way ANOVA) (Figure 3G, H). Treatment with S107 in the drinking water prior to and during the stress significantly reduced ER Ca2+ leak (13.51 ± 1.72% of uptake in S107 treated stress, n=7, versus stress, p<0.05 by one-way ANOVA) (Figure 3G and H). There was no significant difference in ER Ca2+ leak between stressed (10.10 ± 0.67 % of uptake, n=7) and non-stressed RyR2-S2808A+/+ (10.14 ± 1.74 % of uptake, n=7, p>0.05 by one-way ANOVA) or between WT and RyR2-S2808A+/+ mice (Figure 3G and H).
Thus, chronic stress-induced remodeling of neuronal RyR2 channels caused increased single-channel Po under resting conditions and ER Ca2+ leak that was prevented by treatment with S107 or genetic ablation of the RyR2-S2808 PKA phosphorylation site.
Mice were examined using a Morris Water Maze (MWM) to assess hippocampal-dependent learning and memory (Morris, 1984). Three month-old WT non-stressed and stressed mice exhibited no differences in swimming velocities indicating that there was no locomotor defect (Figure S3G). The time spent to reach the hidden platform (escape latency) was prolonged in WT-stressed mice (50.85 ± 3.09s, n=21) compared to non-stressed control mice (28.39 ± 3.79s, n=21, p<0.05) at day 3 of the 5-day trial (Figure 4A). Stressed WT mice exhibited prolonged escape latency on days 3-5, p<0.01, indicating impaired learning due to chronic restraint stress (Figure 4A).
S107 in the drinking water (75 mg/kg/day) significantly reduced escape latency at day 3 (26.61 ± 3.31s, n=22) compared to the non-treated stressed mice (50.85 ± 3.09s, n=21, p<0.05) (Figure 4A). A probe trial was performed on day 6 in which the mice attempted to locate the previously hidden platform that had been removed. The percentage of time spent in the target quadrant and the number of target crossings in control non-stressed WT mice (52.3 ± 4.8%; 4.3 ± 0.59, n=21) were significantly higher than those in stressed WT mice (32.50±2.86%, n=21, p<0.05; 1.59 ± 0.36, n=21, p<0.05) (Figure 4B and C), suggesting impaired learning in stressed mice. The stress-induced learning impairment was prevented by S107 treatment as both the time spent in the target quadrant and the number of target crossings were increased in the treated animals (54.58 ± 2.56%, n=22, p<0.05; 5.67 ± 0.34, n=22, p<0.05) (Figure 4B and C). A lower dose of S107 (50 mg/kg/day), or beginning treatment after the chronic stress did not prevent stress-induced cognitive dysfunction as assessed using the MWM (Figure S4I-K). The rycal S36 (75 mg/kg/day), which does not cross the BBB, did not improve either the time spent in the target quadrant or the number of target crossings (Figure S4F-H).
The novel object recognition (NOR) task was used to assess hippocampal-type memory (Bevins and Besheer, 2006; Denny et al., 2011). There were no significant differences in the total exploration duration among groups in phase-1, the habituation phase, when mice were exposed to two-identical objects (Figure S6). In phase-2, the testing phase, when one object was replaced by a novel object, WT stressed mice failed to discriminate between the novel object and the constant object. The discrimination index (DI) of WT stressed mice (49.22 ± 2.16, n=11) was significantly lower than controls (82.44 ± 1.50, n=12, p<0.05) (Figure S6). S107 treatment improved the DI of stressed mice to control level (82.59 ± 2.24, n=9, p<0.05 compared to stressed WT) (Figure S6). In agreement with the MWM results, the NOR test demonstrated that chronic stress impairs memory in WT mice, and pharmacologic (S107) rescues this impairment.
Elevated plus maze (EPM) (Pellow et al., 1985) and open-field (Hall, 1934) tests were used to assess anxiety and spontaneous activity in stressed WT mice. In the EPM, mice were placed at the center of the plus maze and allowed to move freely for 5 m. The time-spent in the closed-arm and open-arm, the total number of entries to the closed-arm and open-arm were recorded. WT stressed mice spent significantly longer time in the closed-arm (236.57 ± 9.18s, n=21) than non-stressed controls (197.86 ± 5.51s, n=21, p<0.05 by two-way ANOVA), and spent significantly shorter time in the open-arm (11.14 ± 3.24s, n=21) than controls (44.21 ± 5.0s, n=21, p<0.05 by two-way ANOVA) (Figure S7). The ratio of time-spent in the open-arm versus closed-arm was significantly reduced in WT stressed mice (0.06 ± 0.03, n=21) compared to non-stressed controls (0.22 ± 0.05, n=21, p<0.05 by two-way ANOVA) (Figure 4D). There were no significant differences in total number of entries to the closed-arm among experimental groups, but there were significantly fewer entries to the open-arm in WT stressed mice (1.43 ± 0.4, n=21) than controls (4.79 ± 0.46, n=21, p<0.05 by two-way ANOVA) (Figure S7). The ratio of the number of open-arm entries vs closed-arm entries was significantly reduced in WT stressed mice (0.23 ± 0.08, n=21) compared to non-stressed controls (0.47 ± 0.08, n=21, p<0.05 by two-way ANOVA) (Figure 4E). S107 treatment significantly reduced total time-spent in the closed-arm (180.94 ± 4.57s, n=22, p<0.05 by two-way ANOVA), but improved total time-spent in the open-arm (53.44 ± 4.79s, n=22, p<0.05 by two-way ANOVA) in WT stressed mice (Figure S7). S107 treatment also significantly improved the ratios of time-spent on the open-arm versus closed-arm and the number of open-arm entries versus closed-arm entries in stressed mice (0.29 ± 0.07, n=22, p<0.001; 0.48 ± 0.06, n=22, p<0.001 by two-way ANOVA) (Figure 4E). There were no differences in these two ratios when S107 treated non-stressed mice were compared to non-treated controls (0.27 ± 0.05, n=14, p=0.194; 0.51 ± 0.07, n=14, p=0.24 by two-way ANOVA) (Figure 4D and E) suggesting that the positive effects of S107 are not manifested in the absence of stress.
In the open-field test, mice were placed at the center of the experimental chamber and allowed to move for 6 m. The number of visits to the center area and peripheral area were recorded and analyzed. Within the first 3 m and the second 3 m, the ratios of the number of center visits versus peripheral visits for WT stressed mice were significantly lower (0.08 ± 0.02; 0.09 ± 0.03, n=22) than those of non-stressed WT controls (0.14 ± 0.03, n=21, p<0.01; 0.32 ± 0.06, p<0.01 by one-way ANOVA) (Figure 4F). These results are consistent with reduced spontaneous exploratory activity in the stressed mice. S107 significantly improved the exploratory activity and behavioral response to a novel environment in stressed mice as demonstrated by the increased ratios of the number of center visits versus periphery visits within both the first 3 m and second 3 m (0.20 ± 0.04, p<0.05; 0.22 ± 0.02, n=22, p<0.01 by one-way ANOVA) (Figure 4F). Taken together, the behavioral studies demonstrate that stressed mice have cognitive dysfunction manifested as impaired learning and memory, increased anxiety and reduced spontaneous exploratory activity. Pharmacologic (S107) therapy targeting the RyR channel prevented stress-induced cognitive dysfunction suggesting that neuronal ER Ca2+ leak causes cognitive dysfunction during stress.
Post-synaptic plasticity of hippocampal neurons was assessed using electrophysiological recordings of long-term potentiation (LTP) which reflects long-term hippocampal-type memory. Theta-burst stimulation (TBS) was used (12 bursts at 5 Hz, each consisting of four pulses delivered at 100 Hz) (Patterson et al., 2001) at the CA3-CA1 connection in hippocampal slices. LTP was reduced in stressed WT mice (205.9 ± 13.2%, n=7) compared to non-stressed mice (354.1 ± 26.1%, n=7, p<0.001 by two-way ANOVA) (Figure 4G). Treatment with S107 in the drinking water during stress prevented LTP reduction (330.4 ± 17.6%, n=7, p<0.001) (Figure 4G). The input-output curve was not altered by either stress or S107 treatment (Figure 4H). Thus, the stress-induced impairment of hippocampal post-synaptic plasticity was significantly improved by S107.
RyR2-S2808A+/+ mice that harbor RyR2 that cannot be PKA phosphorylated were subjected to behavioral testing after 3-weeks of chronic restraint stress. In MWM tests RyR2-S2808A+/+ mice showed the same escape latency (18.62 ± 3.05s, n=13) as WT mice (19.16 ± 3.0s, n=15, p=0.87 by one-way ANOVA) (Figure 5A) indicating normal cognitive function in the RyR2-S2808A+/+ mice. The time spent in the target quadrant and the number of target crossings also showed no differences between RyR2-S2808A+/+ (49.82 ± 4.87%; 3.73 ± 0.45 crossings, n=13) and WT mice (54.4 ± 3.62%, n=15, p=0.23; 4.13 ± 0.39 crossings, n=15, p=0.25) (Figure 5B and C). Stress failed to alter the escape latency in RyR2-S2808A+/+ mice, at day 5 of the training trial there were no significant differences in escape latency (17.43 ± 1.40s, n=16) compared to non-stressed RyR2-S2808A+/+ mice (18.62 ± 3.05s, n=13, p=0.85 by one-way ANOVA) (Figure 4A). Probe trials similarly showed no deficits in stressed RyR2-S2808A+/+ mice which exhibited the same time spent in the target quadrant and number of target crossings (50.44 ± 2.78%; 4.33 ± 0.31 crossings, n=16) as non-stressed RyR2-S2808A+/+ (49.82 ± 4.87%, n=13, p=0.40; 3.73 ± 0.45, n=13, p=0.14) (Figure 5B and C). There were no differences in swimming velocity between groups in the MWM tests (Figure S3G).
The NOR test indicated there was no effect of stress on memory in the RyR2-S2808A+/+ mice as the DI in phase-2 was unchanged (Figure S6).
EPM testing showed no significant differences in total time-spent in the closed-arm and open-arm between stressed RyR2-S2808A+/+ (161.31 ± 3.43 s in closed-arm, 44.77 ± 7.55 s in open-arm, n=16) and non-stressed RyR2-S2808A+/+ mice (166.21 ± 7.25 s in closed-arm, n=15, p=0.38; 47.64 ± 6.46 s in open-arm, n=15, p=0.65 by two-way ANOVA) (Figure S7). There were no significant differences in total number of entries to the closed-arm and open-arm between stressed RyR2-S2808A+/+ (9.15 ± 1.23 to the closed-arm, 4.62 ± 0.77 to the open-arm, n=16) and non-stressed RyR2-S2808A+/+ mice (9.43 ± 1.19 to the closed-arm, n=15, p=0.80; 4.86 ± 0.72 to the open-arm, n=15, p=0.95 by two-way ANOVA) (Figure S7). There were no differences in time-spent in the open-arm versus closed-arm between stressed RyR2-S2808A+/+ (0.29 ± 0.34, n=16) and non-stressed RyR2-S2808A+/+ (0.29 ± 0.31, n=15, p=0.92 by two-way ANOVA) (Figure 5D). There were also no differences in the number of open-arm entries versus closed-arm between stressed RyR2-S2808A+/+ (0.56 ± 0.06, n=16) and non-stressed RyR2-S2808A+/+ (0.59 ± 0.10, n=15, p=0.79 by two-way ANOVA) (Figure 5E), suggesting no significant changes in the level of anxiety in RyR2-S2808A+/+ mice under stress.
Open-field tests revealed that within the first 3 m and the second 3 m stress failed to alter the ratios of the number of center visits versus peripheral visits in RyR2-S2808A+/+ mice (0.17 ± 0.01, n=16 in stressed; 0.15 ± 0.02, n=15 in non-stressed, p=0.22 by one-way ANOVA) (Figure 5F), indicating no significant stress-induced changes in spontaneous activity in the RyR2-S2808A+/+ mice.
Electrophysiological recordings revealed no differences in LTP between the RyR2- S2808A+/+ (325.4 ± 11.6%, n=7) and WT mice (329.4 ± 18.5%, n=7, p=0.88 assessed by two-way ANOVA) (Figure 5G). Chronic stress failed to affect LTP in the RyR2-S2808A+/+ mice (313.9 ± 17.1%, n=7, p=0.56, assessed by 2-way ANOVA) (Figure 5G). Two-way ANOVA revealed no differences in the input-output curves between WT and RyR2-S2808A+/+ mice (Figure 5H), and there were no differences in the input-output relationships between stressed and non-stressed RyR2-S2808A+/+ mice (Figure 5H).
Thus, stress-induced defects in cognitive function and post-synaptic plasticity in hippocampal neurons were ameliorated by preserving RyR2-calstabin2 interaction either by administering S107 or by genetic ablation of the RyR2 PKA phosphorylation site.
In addition to RyR2, RyR1, which is required for skeletal muscle excitation-contraction coupling, is expressed in specific regions of the brain, including the hippocampus and cerebellum (Giannini et al., 1995). We have demonstrated a role for PKA hyperphosphorylation at Serine 2844 and oxidation and nitrosylation of RyR1 in promoting SR Ca2+ leak in skeletal muscle during extreme exercise, muscular dystrophy and aging (Andersson et al., 2011; Bellinger et al., 2008a; Bellinger et al., 2009; Bellinger et al., 2008b). PKA hyperphosphorylation at Serine 2844 and nitrosylation and oxidation of RyR1 cause depletion of calstabin1 from the channel complex destabilizing the closed state at low cytosolic activating [Ca2+] (~150 nM) (Marx et al., 2000; Reiken et al., 2003b; Ward et al., 2003).
We subjected RyR1-S2844A+/+ knock-in mice, which harbor RyR1 channels that cannot be PKA phosphorylated, to the stress protocol described in Figure 1A. Elevated plasma corticosterone levels (Figure 6A) and hippocampal pCREB and pERK phosphorylation confirmed the extent of stress in the RyR1-S2844A+/+ mice (Figure 6B-D).
Neuronal RyR1 were PKA hyperphosphorylated in hippocampus (Figure 6E and I) and whole brain (Figure S3H-K) following stress, and S-nitrosylated and oxidized (Figure 6E, G, and H). There was minimal depletion of calstabin1 from RyR1 (Figure 6E and I) in hippocampus compared to a significant reduction of calstabin1 from RyR1 in the whole brain (Figure S5E-G). Treatment with S107 prevented calstabin1 depletion from the RyR1 macromolecular complex in the whole brain (Figure S5E-G).
WT mice and RyR1-S2844A+/+ knock-in mice were examined using MWM, EPM, and open-field tasks. In the MWM task, ablation of the PKA phosphorylation site of RyR1 at Serine 2844 did not prevent learning impairment in stressed RyR1-S2844A+/+ mice characterized by prolonged escape latency after day 3 (47.3 ± 1.9s, n=12, p<0.05) compared to the non-stressed mice (37.5±2.8s, n=30) (Figure 7A) and less time spent in target quadrant (22.2 ± 0.2%, n=12, p<0.05) and reduced target crossings (1.3 ± 0.1, n=12, p<0.05) on the day 6 probe trial (Figure 7B and C). There were also no differences between RyR1-S2844A+/+ and WT in the escape latency observed during the 5 day trial (p=0.95 by one-way ANOVA) (Figure 7A). The time spent in the target quadrant and the number of target crossings were not significantly different between WT (45.3 ± 2.3%; 3.8 ± 0.4, n=30) and non-stressed RyR1-S2844A+/+ mice (44.3 ± 2.7%, n=30, p=0.38; 3.1 ± 0.5, n=30, p=0.15) (Figure 7B and C). No differences in swimming velocity among the groups were observed (Figure S3G).
EPM tests showed no significant differences in time-spent in the open-arm versus the closed-arm between RyR1-S2844A+/+ (0.29 ± 0.04, n=30) and WT mice (0.26 ± 0.04, n=30, p=0.66 by two-way ANOVA) (Figure 7D). There were also no differences in the number of open-arm versus closed-arm entries between RyR1-S2844A+/+ (0.47 ± 0.07, n=30) and WT (0.40 ± 0.07, n=30, p=0.55 by two-way ANOVA) (Figure 7D). Ablation of the RyR1 PKA phosphorylation site at serine 2844 failed to prevent the increase in anxiety in the stressed RyR1-S2844A+/+ mice. The time-spent on the open-arm versus closed-arm and the number of entries to the open-arm versus closed-arm in stressed RyR1-S2844A+/+ mice were significantly reduced (0.08 ± 0.02, n=12, p<0.05; 0.29 ± 0.05, n=12, p<0.05 by two-way ANOVA) compared to the non-stressed RyR1-S2844A+/+ mice (Figure 7E). Open-field tests from same group of animals as in the MWM and EPM revealed that, within the first 3 m and the second 3 m, the ratios of the number of center visits versus peripheral visits in stressed RyR1-S2844A+/+ mice were significantly lower (0.11 ± 0.02; 0.12 ± 0.02, n=12) than those of non-stressed RyR1-S2844A+/+ (0.17 ± 0.02, n= 30, p<0.05; 0.32 ± 0.03, p<0.05 by one-way ANOVA) (Figure 7F). There were no differences between the RyR1-S2844A+/+ and WT mice in the open-field tests (p=0.46, p=0.79 by one-way ANOVA) (Figure 7F). Thus, unlike RyR2, remodeling of the RyR1 macromolecular complex does not appear to play a role in stress-induced cognitive dysfunction.
There is a well established link between impaired learning, working and long-term memory and somatic factors, including mental stress and chronic diseases (Etgen et al., 2009). Stress affects the function and morphology of neurons (Kim and Diamond, 2002), and hippocampal neurogenesis (Mirescu and Gould, 2006). Moreover, it has been proposed that these effects could underlie neuropsychiatric disorders and PTSD (Airan et al., 2007; Mirescu and Gould, 2006). Long-term, chronic stress is a major contributor to the development of neuropsychiatric, cardiovascular, and autoimmune diseases, cancer and propensity to self-administer drugs of abuse (McEwen, 2008). The stress response is characterized by a rapid increase in norepinephrine (NE) and epinephrine (Epi) to maintain systemic “homeostasis” (Cannon, 1929; McMahon et al., 1992; Tsuda and Tanaka, 1985; Zigmond et al., 1989). Increased sympathetic activation and urinary NE and/or Epi concentrations are consistently observed in patients with PTSD (Southwick et al., 1999; Yehuda et al., 1998) suggesting long-term chronic stress (Miller and McEwen, 2006; Sherin and Nemeroff, 2011). Despite all of these parallels between chronic stress and PTSD, one potential caveat is that the chronic restraint stress model used in the present study likely does not recapitulate all of the parameters of PTSD.
Current therapy for PTSD remains largely supportive, and not mechanism based. Here we examined the potential role of stress-induced intracellular Ca2+ leak due to stress-induced remodeling of RyR/calcium release channels that have not previously been implicated in stress-induced cognitive dysfunction. PKA phosphorylation of RyR1 and RyR2 (Marx et al., 2000; Reiken et al., 2003a; Reiken et al., 2003b) at serine 2844 or 2808, respectively increases channel activity and intracellular Ca2+ release during acute stress (e.g. exercise) (Marx et al., 2000; Reiken et al., 2003b). RyR dysfunction and pathologic intracellular Ca2+ leak occur during intense chronic β-adrenergic stimulation due to sympathetic hyperactivity (Marx et al., 2000). We have developed a novel RyR-targeted small molecule Rycal (S107) that stabilizes the RyR1 and RyR2 closed states of PKA hyperphosphorylated, and oxidized/nitrosylated channels and prevents intracellular Ca2+ leak induced pathophysiology in relevant animal models of heart failure, exercise-induced arrhythmias, muscular dystrophies and sarcopenia (Andersson et al., 2011; Bellinger et al., 2009; Bellinger et al., 2008b; Lehnart et al., 2006; Wehrens et al., 2005; Wehrens et al., 2004).
Elevated intracellular Ca2+ has been associated with cognitive dysfunction due to brain injury (Deshpande et al., 2008), and neuro-degeneration including Alzheimer's and aging (Berridge, 2010; Keefe et al., 1989; Khachaturian, 1994). One of the downstream consequences of increased RyR2-mediated Ca2+ leak could be abnormal action potential conduction in hippocampal pyramidal neurons (Svoboda et al., 1997). Neuronal cell death could also be a consequence of Ca2+ leak since sustained increases in cytosolic Ca2+ resulting from excessive glutamate mediate excitotoxic cell death (Choi, 1988a, b; Hartley et al., 1993). Other downstream targets include Ca2+-dependent proteases and transcriptional activation through CaMKII, PKA and MAPK mediated activation of immediate early genes (Ghosh and Greenberg, 1995).
Intracellular Ca2+ homeostasis is also regulated by the IP3Rs and SERCA pumps on the ER (Berridge et al., 2003). Studies have shown improved cognitive function in mice due to blockade of IP3R in prefrontal cortex under stress condition (Brennan et al., 2008). Another study showed increased IP3R-mediated ER Ca2+ release in hippocampal neurons during hypoxia (Bickler et al., 2009). We did not observe stress-induced effects on IP3R1 (Figure S5H) or SERCA (Figure S3F) in contrast to the stress-induced remodeling of RyR2 that rendered the channels leaky. Although RyR1 and RyR2 are structurally and functionally homologous the fact that RyR1-S2844A+/+ mice did not protect against stress-induced cognitive dysfunction whereas RyR2-S2808A+/+ did suggests that RyR2 but RyR1 is involved in stressed-induced cognitive dysfunction.
Thus, neuronal RyR2 remodeling by PKA hyperphosphorylation, oxidation and nitrosylation plays an important role in chronic stress-induced cognitive dysfunction. Furthermore, leaky neuronal RyR2 channels could be a potential pharmacological target to ameliorate chronic stress-induced cognitive dysfunction and PTSD.
Wild type (WT) C57BL/6 mice and age-matched RyR2-S2808A+/+ and RyR1-S2844A+/+ knock-in mice (Wehrens et al., 2005) were subjected to restraint stress in 50 ml plastic restraint tubes (Plas-Labs Broome-Style Rodent Restrainer, from Fisher Scientific), without food and water from 6PM to 9AM for 3 weeks (Copeland et al., 2005). Schedules for applying restraint stress and performing cognition function tests are illustrated in Figure 1A. Investigators were blinded with respect to genotype and treatment.
MWM and open-field tests were analyzed using one-way ANOVA and comparison t-test. EPM and LTP measurement were analyzed using two-way ANOVA and comparison t-test. All other data were expressed as mean ± SEM. The results of Ca2+ leak assay were analyzed by one-way ANOVA with Bonferroni correction. Minimum statistically significant differences were established at P < 0.05.
The authors thank Drs. Steven Siegelbaum and Michael Shelanski for helpful comments and Shi-Xian Deng for synthesis of S107. This work was supported by RO1-HL56180 and RO1-NS049442.
Competing Financial Interests: Dr. Andrew R. Marks is a consultant for a start-up company, ARMGO Pharma Inc., which is targeting RyR channels for therapeutic purposes.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.