|Home | About | Journals | Submit | Contact Us | Français|
The contribution of the N-methyl D-aspartate receptors (NMDARs) to synaptic plasticity declines during aging and the decline is thought to contribute to memory deficits. Here, we demonstrate that an age-related shift in intracellular redox state contributes to the decline in NMDAR responses through Ca2+/calmodulin-dependent protein kinase II (CaMKII). The oxidizing agent xanthine/xanthine oxidase (X/XO) decreased the NMDAR mediated synaptic responses at hippocampal CA3-CA1 synapses in slices from young (3–8 mo), but not aged (20–25 mo) rats. Conversely, the reducing agent dithiothreitol (DTT) selectively enhanced NMDAR response to a greater extent in aged hippocampal slices. The enhancement of NMDAR responses facilitated induction of long-term potentiation (LTP) in aged but not young animals. The DTT-mediated growth in the NMDAR response was not observed for the alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor (AMPAR) mediated synaptic responses. A similar increase was observed by intracellular application of the membrane impermeable reducing agent, L-glutathione (L-GSH), through the intracellular recording pipette, indicating the increased NMDAR response was dependent on intracellular redox state. DTT enhancement of the NMDAR response was dependent on CaMKII activity and was blocked by the CaMKII inhibitor – myristoylated autocamtide-2 related inhibitory peptide (myr-AIP), but not by inhibition of the activity of protein phosphatases - PP1 and calcineurin (CaN/PP2B) or protein kinase C. CaMKII activity assays established that DTT increased CaMKII activity in CA1 cytosolic extracts in aged but not in young animals. These findings indicate a link between oxidation of CaMKII during aging, a decline in NMDAR responses, and altered synaptic plasticity.
The NMDAR is a major source of Ca2+ influx into the post synaptic neuron during the induction of LTP at hippocampal CA3-CA1 synapses (Bliss and Collingridge, 1993). CA1 region specific knockout of the NR1 subunit of the NMDAR abolishes LTP, in addition to impairing spatial memory (Tsien et al., 1996). Aged, memory impaired animals exhibit deficits in LTP-induction. Importantly, the NMDAR component of the synaptic response is decreased in aged animals (Barnes et al., 1997; Billard and Rouaud, 2007) suggesting that age-related LTP and memory deficits are due to a decrease in the NMDAR mediated component of synaptic transmission (Foster, 1999; Rosenzweig and Barnes, 2003; Foster, 2007), also referred to here as “NMDAR hypo-function”. Numerous studies indicate that NMDARs contribute less Ca2+ to the induction of LTP in the area CA1 of aged hippocampus, when compared to the young hippocampus (Norris et al., 1998a; Shankar et al., 1998; Boric et al., 2008). It is however unclear what age-related mechanism underlies the NMDAR hypo-function.
Age-related alterations that may contribute to the NMDAR deficits include altered subunit expression, composition, and splice forms (Magnusson et al., 2005; Magnusson et al., 2006). However, there is a debate concerning whether NMDAR subunit expression decreases at hippocampal CA3-CA1 synapses (Foster, 2002). In addition, it is possible that functional differences are related to post-translational modifications associated with oxidation or phosphorylation state rather than number and/or type of receptor subunits (Foster, 2007). Previous research examining the ability of reducing and oxidizing (redox) agents to modulate NMDAR activity in cell cultures and in tissue from neonates suggests that redox state is an important determinant of NMDAR function (Aizenman et al., 1989; Aizenman et al., 1990; Bernard et al., 1997; Choi and Lipton, 2000; Choi et al., 2001), possibly through oxidation of extracellular cysteine residues on the NMDAR (Lipton et al., 2002). Intracellular signaling molecules which affect NMDAR function are also sensitive to redox state. The aged brain exhibits an increase in oxidative damage (Foster, 2006; Poon et al., 2006), and a decrease in redox buffering capacity (Parihar et al., 2008); however, the effect of redox state on NMDAR function during aging has not been examined.
The current studies confirm that the NMDAR mediated synaptic potentials are decreased at CA3-CA1 synapses of the aged hippocampus. We now provide evidence that the age-related decline in NMDAR mediated synaptic responses and impaired induction of LTP in area CA1 of the hippocampus are related to the intracellular redox state. NMDAR responses were modified by redox agents in an age-dependent manner. However, using a combination of extracellular and intracellular recordings with the relatively membrane impermeable L-GSH we found that intracellular redox state mediates that age-dependent shift in NMDAR responses. Furthermore, we demonstrate that the mechanism for the age-dependent redox modulation of NMDARs involves CaMKII activity. These findings establish age-related changes in the redox state as a mechanism for the decrease in NMDAR response.
Procedures involving animals have been reviewed and approved by the Institutional Animal Care and Use Committee and were in accordance with guidelines established by the U.S. Public Health Service Policy on Human Care and Use of Laboratory Animals. Male Fischer 344 rats, young (3–8 mo) and aged (20–25 mo), were obtained from National Institute on Aging colony at Harlan Sprague Dawley Inc (Indianapolis, IA). All animals were group housed (2 per cage), and maintained on a 12:12 hr light schedule, and provided ad lib access to food and water.
The animals were deeply anaesthetized using isoflurane (Webster, Sterling, MA) and decapitated with a guillotine (MyNeurolab, St Louis, MO). The brains were rapidly removed and hippocampi were dissected. Hippocampal slices (~ 400µm) were cut parallel to the alvear fibers using a tissue chopper (Mickle Laboratory Engineering Co, Surrey, UK). The slices were incubated in a holding chamber (at room temperature) with artificial cerebrospinal fluid (ACSF). The ACSF consisted of (in mM): NaCl 124, KCl 2, KH2PO4 1.25, MgSO4 2, CaCl2 2, NaHCO3 26, and D-glucose 1. The ACSF was bubbled with 95% O2 and 5% CO2 for 20 min prior to measurement of pH. The pH was adjusted to 7.4 using 1–2 drops of 10M NaOH in 4L of ACSF. At least 30 min before recording, slices were transferred to a standard interface recording chamber (Warner Instrument, Hamden, CT). The chamber was continuously perfused with oxygenated ACSF at the rate of 2 mL/min and the temperature was maintained at 30 ± 0.5°C.
Extracellular field excitatory postsynaptic potentials (fEPSP) from stratum radiatum of the CA1 region of the hippocampus were recorded using glass micropipettes (4–6 MΩ) filled with recording medium (ACSF). Two concentric bipolar stimulating electrodes (outer pole: stainless steel, 200 µm diameter; inner pole: platinum/iridium, 25 µm diameter) (FHC, Bowdoinham, ME) were localized to the middle of the stratum radiatum to stimulate CA3 inputs onto CA1. Diphasic stimulus pulses of 100 µs duration were delivered by a stimulator (SD9 Stimulator; Grass Instrument, West Warwick, RI) and alternated between the two pathways such that each pathway was activated at 0.033 Hz. The signals were amplified, filtered between 1 Hz and 1 kHz, and stored on a computer disk for off-line analysis. Two cursors were placed to cover the initial descending phase of the waveform and the maximum negative slope (mV/ms) of the fEPSP and NMDAR-fEPSP was determined by a computer algorithm which determined the maximum change across a set of 20 consecutively recorded points (20-kHz sampling frequency) between the two cursors. In a few cases, two cursors were placed to cover the entire waveform; and the maximum amplitude (mV) of the NMDAR-fEPSP and NMDAR mediated intracellular EPSP was determined by a computer algorithm which calculated the valley voltage from zero (for NMDAR-fEPSP) or the maximum voltage from zero (for NMDAR mediated intracellular EPSP) between the cursors.
For measuring the paired-pulse ratio, paired pulses were delivered through a single stimulating electrode at varying inter pulse intervals. The first pulse was set to elicit 50% of the maximal fEPSP. The various inter pulse intervals between successive pulses were 50 ms, 100 ms, 150 ms, and 200 ms. The ratio of the maximum negative slope of the second pulse to the slope of the first pulse was computed as the paired-pulse ratio. To measure changes in the presynaptic fiber volley (PFV) upon application of DTT and X/XO, the average PFV amplitudes from the last 5 minutes of drug application were normalized to the average PFV amplitude recorded during the baseline.
For induction of LTP, the stimulation intensity was set to elicit 50% of the maximal fEPSP. After stable baseline recording at 0.033 Hz for at least 20 min, high frequency stimulation (HFS) was delivered to the pathway at 100 Hz for 1 sec (100 pulses) at the baseline stimulus intensity, and recorded for 60 min post-HFS. A simultaneously recorded control (non-HFS) pathway received the test stimulus but not the HFS. In some cases, the slices were pre incubated with the reducing agent DTT for at least 45 min before delivering the HFS. The average fEPSP slope corresponding to the last 5 minutes from each pathway was used to compare changes in synaptic strength relative to the baseline.
To obtain the NMDAR-mediated field excitatory postsynaptic potential (NMDAR-fEPSP) of CA3-CA1 synaptic transmission, slices were incubated in ACSF containing low extracellular Mg2+ (0.5 mM), 6, 7-dinitroquinoxaline-2, 3-dione (DNQX, 30 µM), and picrotoxin (PTX, 10 µM). In each case the baseline response was collected for at least 10 min before experimental manipulations (drug application). Changes in the transmission properties of the synapses, induced by drug application, were calculated as percentage change from the averaged baseline responses.
To obtain the NMDAR-mediated intracellular synaptic potentials from CA1 pyramidal neurons, slices were incubated in ACSF containing low extracellular Mg2+, DNQX and PTX as described above. Sharp microelectrodes were pulled from thin walled (1 mm) borosilicate capillary glass using a Flaming/Brown horizontal micropipette puller (Sutter Instruments, San Rafael, CA). Microelectrode tips were filled with 3M potassium acetate and in some cases included reduced L-GSH (0.7–1.4 mM). The microelectrode resistances ranged from 42–55 MΩ and 39–50 MΩ for 3M potassium acetate and 3M potassium acetate + L-GSH containing microelectrodes respectively. Microelectrodes were visually positioned in the CA1 pyramidal cell layer using a dissecting microscope (SZH10, Optical Elements Corp, Washington, DC) and a bipolar stimulating electrode was positioned to stimulate the CA3 afferents onto CA1 pyramidal neurons. On cell entry, positive or negative current was applied to clamp the neuronal membrane potential at −65 mV. Only neurons with a resting membrane potential less than −59 mV, and an input resistance >23 MΩ, and an action potential amplitude rising ≥78 mV from the point of spike initiation were included in the analysis. The mean resting membrane potential and holding current was −63 ± 2 mV, −0.07 ± 0.15 nA for the control cells and −63 ± 2 mV, 0.07 ± 0.18 nA for the L-GSH injected neurons. Inclusion of L-GSH in the pipette had no significant effect on the resting membrane potential, holding current, input resistance, action potential amplitude or the microelectrode resistance. Diphasic stimulus pulses of 100 µs duration were delivered at 0.033 Hz and the stimulation intensity was adjusted to elicit an intracellular NMDAR synaptic response, which was below action potential threshold. Baseline response recording began within 3 minutes after cell entry.
All the drugs were prepared according to the manufacturer’s specifications and ultimately dissolved in ACSF prior to bath application on the slices. 6,7-Dinitroquinoxaline-2,3(1H,4H)-dione (DNQX) (Sigma, St. Louis, MO); 4-[(2S)-2-[(5-isoquinolinylsulfonyl) methylamino]-3-oxo- 3-(4-phenyl-1-piperazinyl)propyl] phenyl isoquinolinesulfonic acid ester (KN-62), okadaic acid (OA) (Tocris Bioscience, Ellisville, MO) and FK-506 (LC Laboratories, Woburn, MA) were initially dissolved in a small amount of dimethyl sulfoxide (DMSO; Sigma, St. Louis, MO) and diluted in ACSF to a final DMSO concentration of less than 0.01% and to a final DNQX, KN-62, OA and FK-506 concentration of 30, 10, 1 and 10 µM respectively. Picrotoxin (PTX; Tocris Bioscience, Ellisville, MO) and 5,5′-Dithiobis(2-nitrobenzoic acid) (DTNB; Sigma, St. Louis, MO) were initially dissolved in a small amount of ethanol and diluted in ACSF to a final ethanol concentration of 0.0001% and to a final PTX and DTNB concentration of 10 µM and 0.5 mM respectively. Xanthine (Calbiochem, San Diego, CA) was initially dissolved in a small amount of 0.1N NaOH and finally dissolved in ACSF whose pH was readjusted to 7.4 using 10N hydrochloric acid. DTT, DL-2-Amino-5-phosphonovaleric acid (AP-5), reduced L-GSH (Sigma, St. Louis, MO), (±)-1-(5-Isoquinolinesulphonyl)-2-methylpiperazine dihydrochloride (H-7) (Tocris Bioscience, Ellisville, MO), bisindolylmaleimide I (Bis-I), myristoylated autocamtide-2 related inhibitory peptide (myr-AIP) (Calbiochem, San Diego, CA), and xanthine oxidase (Roche Diagnostics, Indianapolis, IN) were directly dissolved in ACSF. DMSO control studies were interleaved with experiments using DMSO as vehicle.
Hippocampal slices were incubated for 30 minutes in ACSF containing 10 µM of the ROS detection reagent 5-(and-6)-carboxy-2′,7′-dichlorodihydrofluorescein diacetate (c-H2DCFDA; Molecular Probes Inc, Eugene, OR). Slices incubated for 30 minutes in absence of c-H2DCFDA were used to detect background or auto fluorescence. Following incubation, fluorescent images were obtained with an Axiovert 40 CFL fluorescent microscope and Axiocam digital CCD camera (Carl Zeiss, Thornwood, NY). Fluorescence intensity was quantified as follows: fluorescent microscope was used to obtain images under uniform exposure time and intensity. The images were then converted to grayscale images in Adobe Photoshop 5.5; the resulting images were quantified by densitometry analysis using Image J software (http://rsbweb.nih.gov/ij). An area of about 225 µm along the medial-lateral axis and 187.5 µm along the anterior-posterior axis, centered on the CA1 pyramidal neurons was selected for analysis of fluorescence intensity. The mean gray value intensities obtained from the aged and young animals are represented as the mean fluorescence intensities. The mean fluorescence intensity from the dye-exposed slices is normalized to the mean fluorescence intensity obtained from dye-unexposed slices (harvested from the same animal) using the following relationship:
Where Fs and Fc are the mean fluorescence intensities obtained from dye-exposed and dye-unexposed slices respectively.
Hippocampi were isolated from aged F344 rats as described above. CA1 region was separated from the rest of the hippocampus, collected in an eppendorf tube, flash frozen in liquid nitrogen and stored at −80°C. The frozen CA1 tissue samples were placed in a dounce homogenizer containing 1 mL of the homogenization buffer (sucrose, 1M Tris pH 7.5, 1M KCl, protease inhibitor, protein phosphatase 1 inhibitor, protein phosphatase 2 inhibitor, 100 mg/mL sodium butyrate, 0.1 M phenylmethylsulfonyl fluoride; all prepared in distilled H2O) and homogenized using at least six strokes of the pestle. Homogenates were centrifuged at 7700 g for 10 minutes at 4°C. The supernatant (containing the cytosolic fraction) was carefully isolated and stored at −80°C. Protein concentrations of the cytosolic fractions were determined using the BCA assay method (Pierce, Rockford, IL). CaMKII activity in the cytosolic fraction was measured using the CaMKII assay kit (CycLex Co., Ltd, Nagano, Japan). Briefly, uniform amount of cytosolic extracts (protein concentration = 2.0 µg per well) were loaded onto micro titer wells coated with a specific peptide substrate for CaMKII – Syntide-2, along with kinase reaction buffer with or without Ca2+/calmodulin. Purified CaMKII (30 milli units per reaction; CycLex Co Ltd) was used as positive control and cytosolic extracts incubated with EGTA + myr-AIP (CaMKII specific peptide inhibitor) were used as negative control to obtain a measure of comparison for the CaMKII activity in the DTT treated and untreated cytosolic extracts. CaMKII activity is expressed as spectral absorbance units at 450 nm, normalized to the control.
All statistical analyses were performed using StatView 5.0 (SAS Institute Inc, NC). Student’s t-tests were used to examine for differences between treatments with significance set at p < 0.05. For tests involving more than one factor, analysis of variance (ANOVA) was employed, followed by Fisher’s protected least significant difference (PLSD) post hoc analysis to localize differences. In general only one or two slices per animal were employed for a given experimental condition; although, several conditions could be examined using slices from the same animal. Therefore, where stated, n represents the number of slices used in each experiment, unless stated as the number of animals in which case we used only one slice from each animal. In order to ensure the reliability of our main effect (increased NMDAR responses following DTT application in aged animals), experiments involving DTT application alone were interleaved with experiments involving kinase/phosphatase inhibitors, DTNB and AP-5, resulting in a relatively large number of slices (n=16) for this experiment.
The ROS detection reagent c-H2DCFDA was used to evaluate differential rates of ROS production in hippocampal slices from young and aged animals (Fig 1A and 1B). Dye-unexposed slices (aged control: 36.14 ± 4.42%, n = 3 animals; young control: 31.97 ± 3.16%, n = 3 animals) showed no significant difference in fluorescence across the age groups and were used to normalize the fluorescence intensity obtained from c-H2DCFDA treated slices. Incubation of the hippocampal slices from young and aged animals with c-H2DCFDA for 30 min resulted in significantly (unpaired t-test; p<0.05) enhanced fluorescence in aged animals (242.19 ± 20.96%, n = 3 animals) when compared to young animals (141.61 ± 11.78%, n = 3 animals). In order to exclude possible ROS outside the cells, aged hippocampal slices were incubated with superoxide dismutase (SOD; 121 units/mL) and catalase (260 units/mL) during the 30 min prior to imaging. There was no significant difference (p>0.05) in the c-H2DCFDA fluorescence observed between the SOD/catalase exposed (292.388± 28.04%, n = 3) and unexposed aged hippocampal slices.
The NMDAR mediated field excitatory postsynaptic potentials (NMDAR-fEPSPs) were studied to assess changes in the NMDAR response in hippocampal slices from young and aged animals. We confirmed that the NMDAR mediated synaptic transmission decreases in the area CA1 of the hippocampus during aging. The stimulation evoked presynaptic fiber volley (PFV) is an indicator of the level of axon activation that gives rise to the NMDAR-fEPSP (Fig 2A). To examine the relationship between PFV amplitude and the slope of the NMDAR-fEPSP across age groups, the PFV amplitude was separated into 0.4 mV bins and plotted against the corresponding NMDAR-fEPSP slope obtained from the aged and the young animals (Fig 2B). An ANOVA revealed that the slope of the NMDAR-fEPSP was reduced in the aged animals (n = 6 animals) when compared to young animals (n = 5 animals) [F (1, 47) = 27.47, p<0.0001]. The maximum slope of the NMDAR-fEPSP was 0.73 ± 0.14 mV and 2.87 ± 0.9 mV in aged and young animals, respectively.
For subsequent studies, the stimulation intensity was set to evoke a response which was 30% to 50% of the maximal NMDAR-fEPSP and the slope of the NMDAR-fEPSP was measured before and after pharmacological manipulations. To test the hypothesis that the decrease in the NMDAR response was related to oxidizing conditions X/XO, an enzyme substrate combination which produces two types of ROS - superoxide anion and hydrogen peroxide, was applied to hippocampal slices of young and aged animals. Paired t–test revealed that application of X/XO (20 µg/mL/ (0.25 units/mg of xanthine)) for 60 min, significantly decreased the slope of the NMDAR-fEPSP from the baseline levels in the young (p<0.01) but not in the aged animals. Furthermore, application of X/XO significantly [F(1,10) = 15.49, p<0.01] decreased the NMDAR-fEPSP slope to a greater extent in the young animals (66.37 ± 7.04%, n = 5), when compared to the aged animals (96.41 ± 6.14%, n = 7) (Fig 3A, 3B). Paired t-test on the percent change in the PFV indicated no effect (p > 0.05) of X/XO (young: 95.16 ± 8.29; aged: 108.23 ± 19.44). The effect of ROS on the NMDAR response of young animals was reversible, such that even in the presence of a higher concentration of xanthine oxidase (X/XO: (20 µg/mL)/(1 unit/mg of xanthine)), which decreased the NMDAR mediated synaptic response to 54.21 ± 5.79%, the response recovered to 88.11 ± 5.41% of the baseline (n = 5) following a 50 minute washout (Fig 3C).
The age-dependent sensitivity of NMDAR-fEPSP to oxidizing conditions suggests that components of the NMDAR signaling system are initially oxidized to a greater extent in aged animals. Furthermore, the recovery of the NMDAR-fEPSP following washout indicates that young animals possess sufficient antioxidant and/or redox buffering capacity to absorb the excess oxyradical production. To test whether the decline in the NMDAR response in aged animals might be due to the age-dependent increase in the formation of disulfide linkages on the cysteine residues, the reducing agent DTT was applied to hippocampal slices. DTT can reduce the disulfide bonds on proteins into free thiols, thus partially reversing the effect of oxidative stress. Paired t–test revealed that DTT significantly enhanced the slope of the NMDAR-fEPSP from the baseline levels in both the aged (p<0.0001) and young (p<0.05) animals. However, bath application of DTT (0.7 mM, 45 min) significantly [F(1,19) = 5.49, p<0.05] increased the slope of the NMDAR-fEPSP to a greater extent in the aged animals (171.38 ± 13.26%, n = 16) when compared to young animals (114.55 ± 4.41%, n = 5) (Fig 4A, 4B). Paired t-test on the PFV amplitude before and after application of DTT confirmed no change (p > 0.05) in the PFV amplitude for aged (102.81 ± 3.84) and young (97.08 ± 5.41) animals indicating that the effect of DTT on the NMDAR mediated response was not due to changes in excitability and/or the number of axons activated. The enhancement of the NMDAR-fEPSP in the aged animals was maintained (167.32 ± 20.91%, n = 5) following a 45 min washout of DTT (Fig 4C).
To test whether the NMDAR-fEPSP in aged animals was indeed affected by the redox environment, the oxidizing agent 5, 5’-dithiobis (2-nitrobenzoic acid) (DTNB) was applied subsequent to the DTT-mediated enhancement of the NMDAR response in aged animals. Bath application of DTNB (0.5 mM, 45 min) significantly decreased (p<0.05) the DTT-mediated increase in NMDAR-fEPSP in aged animals, such that the NMDAR-fEPSP slope was 86.49 ± 5.54% (n = 4) of the baseline (Fig 5A). A repeated measures ANOVA comparing the mean NMDAR-fEPSP slope corresponding to the last 5 min under each condition indicated a significant [F(2,9) = 6.24, p<0.05] difference. Post-hoc analysis revealed that the NMDAR-fEPSP slope was significantly (p<0.05) increased under DTT compared to baseline and DTNB (p<0.05), and no significant difference was observed between the NMDAR-fEPSP slopes in the baseline and under DTNB. Finally, the NMDAR-fEPSP was abolished by application of AP-5 (Fig 5B) indicating that the response was dependent on NMDARs.
To further examine the specificity of the DTT effects, slices were bathed in normal ACSF and the AMPAR component was recorded as the initial descending phase of the synaptic response. Application of DTT did not affect the AMPAR component of the synaptic response such that responses were 101.63 ± 2.89% (n = 10) of the baseline after application of DTT for 45 minutes (Fig. 5C & 5D). This confirms previous reports that DTT does not influence AMPAR function (Gozlan et al., 1995; Abele et al., 1998). Together, the results indicate that oxidation of sulfhydryl groups can rapidly regulate responsiveness of NMDARs, and that an age-related reduction in the NMDAR response is linked to the redox environment.
Previous research indicates that L-GSH is a cellular reducing agent which protects hippocampal neurons against oxidative stress (Shin et al., 2005; Shih et al., 2006; Yoneyama et al., 2008). However, L-GSH is relatively membrane impermeable such that exogenous application of L-GSH is not effective in increasing intracellular free thiols when compared to DTT (Mazor et al., 1996; Zou et al., 2001; Susankova et al., 2006). Extracellular application of the reduced form of L-GSH (0.7 mM, 45 min) did not alter NMDAR-fEPSP in the aged animals (104.83 ± 8.39%, n = 6) (Fig 6A). In contrast to the extracellular application, inclusion of the reduced form of L-GSH (0.7 mM to 1.4 mM) in the intracellular recording pipette significantly [F(1,6) = 6.87, p<0.05] enhanced the amplitude of the NMDAR mediated synaptic potentials (203.90 ± 31.38%, n = 5) in single hippocampal CA1 pyramidal neurons from aged hippocampus when compared to age matched control cells (91.55 ± 22.94%, n = 3), for which L-GSH was not included in the intracellular recording pipette (Fig 6B). Finally, intracellular delivery of L-GSH in the presence of nifedipine (10 µM) continued to enhance the amplitude of the NMDAR mediated synaptic potentials (223.54 ± 46.56%, n = 3) (Fig 6C) suggesting that the DTT effect is not due to differences in L-channel activity. The fact that intracellular delivery of glutathione enhanced the NMDAR response provides strong evidence for intracellular redox status as a mechanism for the age-associated modulation of NMDAR function.
For studies on LTP and paired-pulse facilitation, slices were bathed in normal ACSF in order to record the AMPAR and NMDAR component of the synaptic response. LTP was induced by a single episode of high frequency stimulation (HFS; 100Hz, 1second). The magnitude of LTP was significantly [F(1,17) = 12.14, p<0.01] greater in aged hippocampal slices pre-incubated with DTT for 45 min (136.53 ± 2.77%; n = 10), when compared to the aged controls not exposed to DTT (118.15 ± 4.63%; n = 9) (Fig 7A & 7C). In contrast there was no difference in the levels of HFS-induced LTP between young controls (130.17 ± 8.64%; n = 6) and the young slices exposed to DTT (117.09 ± 12.24%; n = 5) (Fig 7B & 7D). Thus, reducing conditions enhance NMDAR responses and the magnitude of LTP only in aged animals. Furthermore, examination of paired-pulses delivered at varying inter-pulse intervals (Δt = 50 ms, 100 ms, 150 ms, 200 ms), under control conditions and 45 min after the bath application of DTT indicated no effect of treatment across the four inter-pulse intervals (Fig 7E).
NMDAR function is controlled by several intracellular kinases (Ben-Ari et al., 1992; Westphal et al., 1999; Li et al., 2001). Furthermore, the enzymatic activity of these kinases can be regulated by the reduction and oxidation of the cysteine residues located in their structure (Raynaud et al., 1997; Griendling et al., 2000; Knapp and Klann, 2000). To test if serine/threonine (Ser/Thr) kinases were involved in the DTT mediated increase of the NMDAR response, the broad-spectrum and membrane-permeable Ser/Thr kinase inhibitor H-7 was bath applied prior to and during the application of DTT. In the presence of H-7 (10 µM, 45 min), DTT application failed to produce the robust increase (one-group t-test; p>0.05) in the NMDAR-fEPSP slope (111.86 ± 6.92%, n = 7) from the baseline levels (Fig 8A). Similarly, bath application of the CaMK inhibitor KN-62 (10 µM, 45 min) (Tokumitsu et al., 1990), prior to and during the application of DTT, blocked (97.9 ± 7.98%, n = 5) the DTT mediated increase in NMDAR-fEPSP in aged animals (Fig 8B). Finally, the specific inhibitor of CaMKII, myr-AIP (5 µM, 60 min), which was bath applied prior to and during the application of DTT, effectively blocked (104.48 ± 4.29%, n = 4) the DTT mediated increase in NMDAR-fEPSP in aged animals (Fig 8C). In contrast, the membrane permeable PKC inhibitor, Bis-I (500 nM, 45 min) (Knapp and Klann, 2002) failed to block the DTT mediated increase in the NMDAR-fEPSP (142.58 ± 13.06%, n = 6) (Fig 8D). An ANOVA comparison of the effect of DTT in the presence and absence of the kinase inhibitors, indicated a significant effect of kinase inhibitors on the DTT effect in aged animals [F(4,33) = 5.85, p<0.01]. Post hoc comparisons indicated that the DTT mediated increase in the NMDAR response was blocked by H-7, myr-AIP, and KN-62 but not Bis-I (Fig 8E). Finally, t-tests indicated no effect of 0.01% DMSO (96.38 ± 7.64, n = 5) or kinase inhibition per se (H-7: 103.01 ± 4.21%, n = 7; KN-62: 96.23 ± 8.3%, n = 5; myr-AIP: 100.52 ± 4.42%, n = 4; Bis-I: 110.72 ± 11.44%, n = 6) on the baseline NMDAR-fEPSP slope in aged animals.
The activity of Ser/Thr phosphatases like calcineurin (CaN) and protein phosphatase 1 (PP1), is thought to contribute to altered synaptic plasticity during aging (Foster et al., 2001). Furthermore, phosphatase activity decreases NMDAR function (Lieberman and Mody, 1994; Wang et al., 1994). To test if DTT effects were mediated by CaN, the CaN inhibitor FK-506 (10 µM) (Norris et al., 2008) was bath applied 45 minutes before and during the application of DTT on aged hippocampal slices. Application of FK-506 per se did not affect the NMDAR-fEPSP slope (109.61 ± 9.16%, n = 5). Importantly, FK-506 failed to block the DTT mediated increase in the NMDAR response in aged animals such that the NMDAR-fEPSP slope increased to 148.61 ± 16.42% (n = 5) (Fig 9A & 9C).
To examine the role of PP1, the PP1 inhibitor okadaic acid (OA) (1 µM, 30 min) (Schnabel et al., 2001) was applied prior to and during application of DTT to aged hippocampal slices. Application of OA significantly increased the NMDAR-fEPSP slope during the baseline recording period (121.46 ± 9.19%, n = 5, p<0.05). Therefore, a new stable baseline was recorded prior to the application of DTT. Relative to the new baseline, DTT increased the NMDAR-fEPSP slope (137.89 ± 3.99%, n = 5) in the presence of OA (Fig 9B & 9C). An ANOVA comparing the effect of DTT in the presence and absence of the phosphatase inhibitors, indicated no significant difference [F(2,15) = 1.37, p>0.05] (Fig 9D).
To determine whether DTT was directly influencing CaMKII activity, cytosolic extracts from CA1 region of the hippocampus from aged and young animals were assayed for CaMKII activity by examining the phosphorylation of the synthetic peptide, syntide-2, in the presence and absence of DTT. Relative to baseline control levels, cytosolic CaMKII activity was significantly enhanced (p<0.05) in the presence of 0.7 mM (113.11 ± 3.47%, n = 3) and 1.4 mM DTT (120.46 ± 3.14%, n=3) in the aged CA1 cytosolic extracts (Fig 10A). Higher levels of DTT (2.8 mM) resulted in a decrease in CaMKII activity, presumably due to denaturation of the enzyme. In contrast to the effect observed in aged animals, DTT had either no effect or decreased CaMKII activity in CA1 cytosolic extracts from young animals (Fig 10B). In both age groups, the CaMKII activity was inhibited (p<0.05) by the addition of calcium chelator EGTA (2 mM) and a CaMKII-specific peptide inhibitor myr-AIP (10 µM) (Aged: 5.36 ± 2.85%, n = 3; Young: 19.99 ± 9.01%, n = 3).
One possibility is that DTT was acting on the CaMKII activity regulator calmodulin (CaM). In this case the effect of DTT on CaMKII activity may have been reduced by the addition of exogenous and un-oxidized CaM. To test this idea the assay was repeated in the absence of exogenously added CaM. The CaMKII activity from aged CA1 cytosolic extract in the presence of 0.7 mM DTT, in the absence of exogenous CaM, (112.11 ± 3.91%, n = 3) was not enhanced beyond that observed following addition of exogenous CaM, suggesting that DTT effects were not mediated by reducing cysteines on CaM. Finally, addition of DTT to purified CaMKII (CycLex Co Ltd) decreased CaMKII activity (p<0.05) (0.7 mM DTT: 86.62 ± 6.04%, n = 3; 1.4 mM DTT: 70.72 ± 18.58%, n = 3) (Fig 10C), indicating that the DTT effects were specific for CaMKII present in the aged CA1 cytosolic extracts.
Previous work demonstrates that NMDAR responses are reduced in the area CA1 of the hippocampus of aged rat brain (Barnes et al., 1997; Billard and Rouaud, 2007), possibly due to altered receptor function (Cady et al., 2001). We confirmed an age-related decrease in the NMDAR response and demonstrate age-dependent effects of redox modulators on the NMDAR response. Our results reveal a role for intracellular redox state in contributing to the age-related decline in NMDAR mediated synaptic responses and impaired induction of LTP in area CA1 of the hippocampus. The effects of DTT application were specific for aged animals and not observed for young animals. Furthermore, the DTT mediated increase in the NMDAR response was not associated with a change in the PFV, was not mimicked for the AMPAR mediated response (Fig 5D & 5E) and was not blocked by nifedipine, suggesting the effect was not due to a general influence on cell excitability. Finally, the manner in which DTT affects the NMDAR response was specific to the intracellular redox state and the activity of CaMKII. Thus, we establish a mechanism that links general theories of aging (i.e. oxidative stress and redox state) with biological markers of age-related cognitive decline, reduced NMDAR responses, and impaired LTP.
The redox state of the extracellular cysteine residues of the NMDARs have been implicated in regulating the NMDAR function in cell cultures and neonatal animals (Aizenman et al., 1989; Aizenman et al., 1990; Bernard et al., 1997; Choi and Lipton, 2000; Choi et al., 2001). In contrast, our results are consistent with recent work in hippocampal cell cultures indicating a decrease in the intracellular redox ratio during aging, due in part to a deficit in intracellular GSH (Parihar et al., 2008). Specifically, we observed that intracellular, but not extracellular application of L-GSH enhanced the NMDAR response in the aged neurons, indicating that the age-related decrease in NMDAR function is due to a shift in redox state of the intracellular compartment.
The DTT mediated enhancement of NMDAR responses was specific to CaMKII activity. CaMKII inhibitors, myr-AIP and KN-62, blocked the DTT-mediated enhancement of the NMDAR response in aged animals. The DTT effects were not blocked by inhibition of PKC or phosphatases - CaN and PP1. The results point to CaMKII as a critical link between the intracellular redox state and the decrease in the NMDAR response. The role of CaMKII was confirmed by enzyme activity assays which established that DTT increased CaMKII activity only in CA1 cytosolic extracts of aged animals. DTT did not increase CaMKII activity in samples from young animals or in purified CaMKII.
Oxidation of methionine residues on CaM has been reported to decrease the ability of CaM to activate CaMKII (Robison et al., 2007). Our results are not likely due to CaM methionine oxidation since DTT selectively reduces cysteine residues (Ciorba et al., 1997; Cai and Sesti, 2009; Long et al., 2009). In addition, DTT had equivalent effects in activating CaMKII, regardless of whether exogenous CaM was added to the reaction. The results indicate that oxidation of CaMKII, rather than CaM, underlies the reduction in kinase activity and are consistent with a recent report demonstrating that oxidative stress induced by ischemia results in disulfide linkages on the cysteine residues of CaMKII which decrease kinase activity (Shetty et al., 2008). While the data provide a link between age-related changes in intracellular redox state, CaMKII activity, and NMDAR function, the exact mechanism through which CaMKII regulates the NMDAR response remains to be determined. In addition to regulating phosphorylation state of proteins, including glutamate receptors, synaptic CaMKII participates in protein-protein interactions with several proteins localized to the dendritic spine which could ultimately alter NMDAR location and function (Lisman et al., 2002; Robison et al., 2005). Indeed, reduced CaMKII activity is associated with a specific decrease in synaptic NMDARs and decreased LTP (Gardoni et al., 2009).
In addition to a role for CaMKII, we observed that PP1 inhibition resulted in a modest increase in the NMDAR-fEPSP in aged hippocampal neurons. Age differences in the NMDAR response, which depend on kinase/phosphatase activity, are reminiscent of the age-dependent effects of kinase and phosphatase inhibitors on the rapid component of synaptic transmission mediated by AMPARs (Norris et al., 1998b; Hsu et al., 2002; Foster, 2007) and suggest that PP1 activity contributes to a reduction in AMPAR and NMDAR components of synaptic transmission (Foster et al., 2001; Morishita et al., 2005), a characteristic specific to senescent CA1 synapses (Rosenzweig and Barnes, 2003). Our results indicate that a shift in the intracellular redox state during aging may cause or magnify the imbalance in the kinase/phosphatase activity, favoring phosphatases (Foster, 2007).
Finally, NMDAR function is critical to the induction of LTP and we observed that DTT improved LTP in hippocampal area CA1 in the aged animals. The interaction of NMDARs with CaMKII has been proposed as a model of memory (Lisman et al., 2002) and recent work indicates that disruption of the interaction between CaMKII and NMDAR impairs the induction of LTP and spatial learning (Zhou et al., 2007). We have provided evidence to indicate that an age-related shift in redox state is a biological mechanism that can progressively inhibit NMDAR function in the hippocampus during senescence. Together, the results suggest that age-related changes in the redox state contributes to a decline in CaMKII activity, which ultimately leads to a decline in the NMDAR response. The outcome of such senescent mechanisms is an alteration in the synaptic plasticity at the CA3-CA1 synapses which contributes to the age-related cognitive decline.
This work was supported by, the Evelyn F. McKnight Brain Research Foundation, National Institutes of Health Grant AG014979 and MH059891 to TCF, a UF Alumni fellowship to KB.