Introduction

AMPA-type glutamate receptors (AMPARs) play an important role in fast excitatory synaptic transmission in the central nervous system. The functional properties of AMPAR depend on the precise composition of the core subunit (GluA1–4)1 and the associated auxiliary subunits, including a family of transmembrane AMPAR regulatory proteins (TARPs)2, such as Type I TARPγ2 (stargazin), γ3, γ4 and γ8, and Type II γ5 and γ73,4. Type I TARPs play pivotal roles in the neuronal trafficking of AMPARs and channel properties3,5,6, but Type II TARPs mainly modulate AMPAR channel function7.

As an AMPAR auxiliary subunit8, stargazin delivers AMPARs through the secretory pathway to the plasma membrane, and then shuttles AMPARs to postsynaptic densities (PSDs) by binding to PSD-95 and related PDZ-containing proteins5,9. Thus, the density of AMPARs at the neuronal membrane and synapse can be controlled by stargazin and PSD-95. Following the discovery of stargazin as an AMPAR auxiliary subunit, three subunits of TARP family (γ3, γ4, and γ8) were identified3. Due to the high level of expression of TARPγ8 in the hippocampus, many studies have been focused on the modulation of TARPγ8 on AMPAR4,10,11. Transgenic mice with a mutation of TARPγ8 show a decrease in the expression of the AMPAR protein level12. Knockout of γ8 reduces the level of the AMPAR protein and impairs AMPAR-mediated synaptic function4,13. It also prevents against kainate-induced neuronal death12, suggesting that the TARPγ8 subunit is critical for the expression and trafficking of AMPAR, which is essential for long-term potentiation (LTP)10,14.

The most efficient way to regulate the level of cellular proteins is through selective proteolysis15. Calpain, a calcium-activated non-lysosomal protease16, cleaves many target proteins, such as AMPAR17,18, contributing to glutamate excitotoxicity under disease conditions, including stroke and epilepsy. Caspase, a cysteine-dependent and calcium-regulating protease19,20, was also reported to modulate AMPAR function via activity-dependent AMPAR cleavage21,22,23. Whether there is a similar regulatory mechanism of the AMPAR-associated protein TARP is unclear.

In this study, we investigated the effects and the cellular mechanism of AMPA exposure on TARPγ8 expression and AMPA current in different culture stages. We found that pre-exposure of hippocampal neurons to non-toxic concentrations of AMPA24,25 caused a development-dependent decrease in TARPγ8, which could be prevented by calpain inhibitor but not caspase inhibitor, suggesting that the calpain-mediated cleavage of TARP was dependent on development.

Materials and methods

Animals

Sprague-Dawley rats were obtained from the Experimental Animal Center of Zhengzhou University (Zhengzhou, China). All experimental procedures were approved by the Ethics Committee for Animal Experimentation of Xinxiang Medical University, Xinxiang, China. In addition, all efforts were made to minimize animal suffering and the number of animals used.

Primary hippocampal neuronal culture

Primary cultures of hippocampal neurons were obtained from Sprague-Dawley rat embryos (E18.5) as previously described26. The hippocampus was isolated under sterile conditions and was enzymatically treated with 0.25% papain in HBSS (5.4 mmol/L KCl, 137 mmol/L NaCl, 0.4 mmol/L KH2PO4, 0.34 mmol/L Na2HPO4·7H2O, 10 mmol/L glucose, and 10 mmol/L HEPES) for 10 min. Then, the cells were mechanically dissociated by trituration and seeded at a density of 1×106 cells/cm2 onto plates coated with poly-D-lysine (Sigma-Aldrich, St Louis, MO, USA). The neurons were cultured in Neurobasal medium (Gibco, Invitrogen Corp, Grand Island, NY, USA) supplemented with 2% B27 (Gibco), 1% glutamine (Sigma-Aldrich), and 1% penicillin/streptomycin (Sigma-Aldrich Corp) at 37 °C in a 95% air/5% CO2 humidified incubator (Thermo Scientific, Waltham, MA, USA). Half of the medium was changed twice a week. Hippocampal neurons were incubated for at least 7 d in vitro (DIV) before being used for experiments.

Western blotting analysis

The proteins extracted from the hippocampal tissues of Sprague-Dawley rats and primary cultured hippocampal neurons were denatured at 95 °C for 5 min and were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred onto PVDF membranes (Millipore, Boston, MA, USA) using a Mini Trans-Blot electrophoretic transfer cell (Bio-Rad, Hercules, CA, USA) following the manufacturer's instructions. Transferred membranes were blocked in 5% skim milk dissolved in Tris-buffered-saline pH 7.5/0.1% Tween-20 (TBST) for 2 h at room temperature and incubated with one of the following primary antibodies: TARPγ8 (1:1000, 50 kDa, Millipore) at room temperature overnight, or GluA1 (1:1000, 110 kDa, Abcam, Shanghai, China), GluA2 (1:2000, 106 kDa, Abcam), and β-actin (loading control, 1:20 000, 42 kDa) at 4 °C overnight. The membranes were washed three times with TBST and then incubated with appropriate conjugated hydrogen peroxidase HRP secondary antibody (1:2000, Abcam) for 1 h at room temperature. Proteins were visualized using the enhanced chemiluminescence (ECL) reagent (Beyotime, Shanghai, China), and band optical density (OD) analysis was performed using Quantity One software (Bio-Rad).

The relative expression of TARPγ8 and AMPARs was determined using β-actin as a loading control by calculating the ratio of ODs for each protein of interest versus β-actin OD in each individual subject. To avoid interassay variations, the values obtained were normalized with the values measured for the control cultures in each experiment. The data are presented as the mean±SEM from 5 independent experiments.

Whole-cell recordings

AMPA-induced currents were recorded using the whole-cell patch clamp technique under voltage clamp mode from cultured hippocampal neurons 7–8 and 17–18 d after plating. Recordings were performed using borosilicate glass patch pipettes (2–5 MΩ) connected to an Axonpatch 700B amplifier (Axon Instruments, Molecular Devices, Saint Clara, CA, USA). Data acquisition was performed using a digitizer (DigiData 1440, Axon Ins) and off-line analysis was carried out with pClamp10.0 (Axon Ins). The whole-cell currents were filtered at 2 kHz and sampled at 10 kHz. The cells were focused on the stage of a Nikon FN1 (Tokyo, Japan) inverted microscope and was locally superfused with a Tyrode solution containing the following (in mmol/L): 137 NaCl, 2 CaCl2, 1 MgCl2, 10 glucose, 5 KCl, and 10 HEPES/NaOH (pH 7.4), 1 μmol/L tetrodotoxin (TTX), 500 nmol/L MK-801. The whole-cell recording pipette was filled with a solution containing the following (in mmol/L): potassium gluconate 140, KCl 5, HEPES 10, EGTA 0.2, MgCl2 2, MgATP 4, Na2GTP 0.3, and Na2-phosphocreatine 10 at pH 7.2. Following seal rupture, series resistance (<20 MΩ) or membrane resistance (300–500 MΩ) was monitored throughout the whole-cell recording and data were discarded if the resistance changed by more than 20%. All recordings were obtained at 31±1 °C using an automatic temperature controller (Warner Ins).

AMPAR-mediated currents were recorded with AMPA (100 μmol/L) application in cells (held at −60 mV). AMPA was puffed by pressure from a micropipette using a Picospritzer III (World Precision Instruments, Stevenage, Hertfordshire, UK) every 30 s at 10 psi for a duration of 50 ms from patch pipettes of the same dimensions as those used for recording (tip 2–4 μmol/L), with the pipette tip located approximately 50 μm away from the soma surface, so that the cell recorded could be totally immersed in the drug solution during perfusion.

Antibodies and reagents

We purchased the following primary antibodies: polyclonal rabbit anti-TARP antibody (Millipore), rabbit anti-GluA1 (Abcam), mouse anti-GluA1 (Abcam), and β-actin (Abcam). The following secondary antibodies were used: HRP-conjugated goat anti-rabbit and anti-mouse IgG (Abcam). α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]quinoxaline-2,3-dione (NBQX), (2R)-amino-5-phosphonovaleric acid (AP5) and tetrodotoxin (TTX) were purchased from Tocris Bioscience (Bristol, UK). All other reagents were obtained from Sigma-Aldrich.

Statistical analysis

All statistical tests were performed using SigmaStat software (SPSS Inc, Chicago, IL, USA). The results are expressed as the mean±SEM. Differences between mean values were evaluated using Student's t-test or one-way analysis of variance (ANOVA) followed by Tukey's HSD test. Differences were accepted as significant when P<0.05.

Results

Pretreatment with AMPA decreased TARPγ8 protein expression level

To study the potential impact of AMPAR activation on TARPγ8 expression, we measured the TARPγ8 protein level in neuronal culture at different stages. As shown in Figure 1, Western blotting analysis indicated that TARPγ8 expression did not differ between neurons cultured for 7–8 DIV (100% as the control) and 17–18 DIV (87.8%±9.4% of control, n=4) in the control conditions. AMPA (100 μmol/L) pre-treatment induced a significant decrease of TARPγ8 expression in neurons cultured for 17–18 DIV (47.4%±8.8% of control, n=4) compared with the neurons cultured for 7–8 DIV (83.2%±11.1% of control, P<0.05, n=4, ANOVA). In the 7–8 DIV cultures, TARPγ8 expression was not significantly different between the control (100%) and AMPA group (83.2%±11.8% of control, n=4). However, in the 17–18 DIV cultures, TARP expression was significantly decreased in the AMPA group (47.8%±8.3%, P<0.05 vs 87.5%±9.6% for non-agonist group, n=4, ANOVA). These results indicate that prolonged AMPAR stimulation causes a down-regulation of TARPγ8 levels in rat hippocampal neurons.

Figure 1
figure 1

AMPA-induced decrease in TARPγ8 protein expression in prolonged cultures of rat hippocampal neurons. (A) Western blots show the levels of TARPγ8 (50 kDa) and β-actin (42 kDa) of hippocampal neurons cultured for 7–8 DIV and 17–18 DIV in the absence and presence of AMPA (100 μmol/L, 4 h). (B) Bar graph shows the quantification of TARPγ8 band densities. The normalization ratio in this case was obtained using β-actin. The data are presented as the mean±SEM, bP<0.05, n=4, one-way ANOVA.

PowerPoint slide

Full size image

AMPA-induced decrease of TARPγ8 levels can be blocked by AMPA and NMDA receptor antagonists

To confirm the activity-dependent decrease of TARPγ8 levels in the prolonged cultures, the antagonist of AMPA and/or NMDA receptors was applied. As shown in Figure 2, the AMPA-mediated decrease of TARPγ8 expression can be prevented by the AMPA/KA receptor antagonist NBQX (10 μmol/L) and NMDA receptor antagonist D-AP5 (20 μmol/L). Compared with controls, the TARPγ8 expression in the presence of AMPA was significantly decreased (64.4%±5.6% of control, P<0.05, n=5, ANOVA), and the TARPγ8 level was not significantly altered in the presence of AMPA+NBQX (94.3%±5.2% of control, n=5). However, compared with control, TARPγ8 expression was increased in the presence of AMPA+D-AP5 (151.7%±5.5% of control, P<0.05, n=5, ANOVA) or in the presence of AMPA+NBQX+AP5 (141.4%±4.7% of control, P<0.05, n=5, ANOVA). These results indicated that the AMPA-mediated decrease of TARPγ8 can be blocked by NBQX treatment and reversed by AP5 treatment, which suggested that both AMPAR and NMDA receptor activation is involved in the AMPA-induced decrease of TARPγ8 expression.

Figure 2
figure 2

AMPA-induced decrease in TARPγ8 expression was blocked by AMPAR and NMDAR antagonists. (A) Western blots show levels of TARPγ8 (50 kDa) and β-actin (42 kDa) of hippocampal neurons cultured for 17–18 DIV following AMPA treatment (100 μmol/L, 4 h) in the presence of the AMPAR and/or NMDAR antagonists NBQX and/or AP5 (added simultaneously with AMPA treatment). (B) The bar graph shows the quantification of TARPγ8 band densities. The normalization ratio in this case was obtained using β-actin. The data are presented as the mean±SEM, bP<0.05, n=5, one-way ANOVA.

PowerPoint slide

Full size image

The protein levels of TARPγ8 and GluA1 but not GluA2 are mediated by calpain and lysosomes

Previous studies have demonstrated that calpain mediates the GluA1 expression level18. The AMPA-induced decrease of the TARPγ8 expression level is involved in NMDA receptor activation and calcium influx, and calpain activity is strongly calcium-dependent; therefore, we further tested whether calpain contributes to the AMPA-induced decrease of TARPγ8 expression. As shown in Figure 3, the AMPA-induced decrease of TARPγ8 expression was blocked by calpain inhibitor, calpeptin (50 μmol/L). Compared with the control, the TARPγ8 expression level in the presence of AMPA was significantly decreased (65.4%±3.7% of control, P<0.05, n=5, ANOVA), but TARPγ8 expression was not significantly different between cultures in the presence of calpeptin (96.1%±6.3% of control, n=5) and in the presence of AMPA+calpeptin treatment (98.3%±5.4% of control, n=5).

Figure 3
figure 3

Calpain and lysosome inhibitors blocked the AMPA-induced decrease of TARPγ8. (A) Western blots show levels of TARPγ8 (50 kDa) and β-actin (42 kDa) of hippocampal neurons cultured for 17–18 DIV in the absence and presence of AMPA (100 μmol/L, 4 h) following the administration of the calpain inhibitor calpeptin, the lysosome enzyme inhibitor leupeptin, or the caspase inhibitor zVAD (added simultaneously with AMPA treatment, n=5). (B) The bar graph shows the quantification of the TARPγ8 band densities. The normalization ratio in this case was obtained using β-actin. The data are presented as the mean±SEM, bP<0.05, n=5, one-way ANOVA. Cal, calpeptin; Leu, leupeptin.

PowerPoint slide

Full size image

We also tested whether lysosomes' enzymatic activity is involved in the AMPA-induced decrease of TARPγ8 expression. The AMPA-induced decrease of TARPγ8 expression was blocked by the lysosomal enzyme inhibitor leupeptin (50 μmol/L). Compared with the control, the TARPγ8 expression in the cultures in the presence of leupeptin (96.2%±8.5% of control, n=5) and in the presence of AMPA+leupeptin (96.7%±5.4% of control, n=5) was not significantly different (Figure 3).

Previous studies have demonstrated that caspase mediates the GluA1 protein expression level27. Therefore, we tested whether caspase is involved in activity-dependent decrease in TARPγ8 expression level. As shown in Figure 3, the TARPγ8 protein level was not different between control and 50 μmol/L zVAD, a caspase inhibitor, treatment (94.4%±2.2% of control, n=5). In the activity-dependent condition, the AMPA-induced decrease in the TARPγ8 expression level was 68.1%±2.3% of control (P<0.05, n=5, ANOVA), and there was no significant difference in TARPγ8 expression between zVAD+AMPA and AMPA treatment alone. These results indicated that calpain and lysosomes, but not caspase, mediated the activity-dependent decrease of the TARPγ8 expression level.

In another set of experiments, we tested whether activity-dependent modulation also occurs in the expression levels of GluA1 and GluA2. As shown in Figure 4A and 4C, AMPA induced a decrease in GluA1 (83.3%±2.2% of control, P<0.05, n=5, ANOVA) in prolonged cultured hippocampal neurons (17–18 DIV), but the GluA1 level did not change significantly following the application of calpeptin (102.1%±3.4% of control, n=5) or calpeptin+AMPA (109.7%±3.6% of control, n=5). Similarly, the GluA1 level did not change significantly following the application of leupeptin (98.5%±4.1% of control, n=5) or leupeptin+AMPA (101.8%±4.3% of control, n=5). zVAD (50 μmol/L) slightly reduced the GluA1 level without statistical significance (96.9%±1.3% of control, n=5). AMPA induced no change in the GluA2 level (Figure 4B and 4D). On average, the GluA2 level was 98.4%±4.7% for AMPA alone, 104.6%±3.5% for calpeptin alone, 101.4%±2.3% for calpeptin+AMPA, 99.2%±5.1% for zVAD alone, 92.3%±4.8% for zVAD+AMPA, 98.4%±4.4% for leupeptin alone and 98.6%±9.2% for leupeptin+AMPA (n=5 for each group).

Figure 4
figure 4

Calpain and lysosome inhibitors blocked the AMPA-induced decrease of GluA1. Western blots show levels of GluA1 (110 kDa) (A) and GluA2 (106 kDa) (B) of hippocampal neurons cultured for 17–18 DIV in the absence and presence of AMPA (100 μmol/L, 4 h) following the administration of the calpain inhibitor calpeptin, the lysosome enzyme inhibitor leupeptin, or the caspase inhibitor zVAD (added simultaneously with AMPA treatment). Bar graph shows the quantification of GluA1 (C) and GluA2 (D) band densities. The normalization ratio in this case was obtained using β-actin. The data are presented as the mean±SEM, bP<0.05, n=5, one-way ANOVA. Cal, calpeptin; Leu, leupeptin.

PowerPoint slide

Full size image

Taken together, these data indicate that AMPA reduces the GluA1 expression in cultured hippocampal neurons, which is mediated by calpain and/or lysosome enzyme.

AMPA-induced decrease in AMPA current

To confirm whether there was an activity-dependent functional change in AMPAR under the same conditions of AMPA-induced regulation of TARPγ8 and GluA1, we examined AMPA-induced whole-cell currents in cultured hippocampal neurons pretreated with AMPA (100 μmol/L). The application of AMPA induced an inward current, which was completely blocked by the AMPA/kainate receptor antagonist (10 μmol/L NBQX), indicating that it is mediated primarily by AMPARs. As shown in Figure 5, pretreatment with AMPA did not significantly alter the AMPA-induced current in cultured neurons (7–8 DIV) (14.55±1.92 pA/pF) compared to those from the sham treatment group (13.81±1.39 pA/pF, n=12, Figure 5A, B). However, pretreatment with AMPA markedly suppressed the subsequently recorded AMPA-induced current in cultured neurons (17–18 DIV) (8.38±0.82 pA/pF, n=12, P<0.05, ANOVA) compared to those from the sham treatment group (14.13±1.34 pA/pF, n=12). The activity-dependent decrease in AMPA-induced current was blocked by pretreatment with calpeptin (13.88±1.25 pA/pF, n=12) or leupeptin (14.13±1.04 pA/pF, n=12), but not by zVAD (7.02±0.98 pA/pF, n=12, P<0.05, ANOVA), compared to the sham treatment group. The AMPA-induced current in the sham treatment group was not altered by pretreatment with calpeptin (14.38±13.88 pA/pF, n=12), leupeptin (13.0±0.76 pA/pF, n=12), or zVAD (13.35±0.93 pA/pF, n=12, Figure 5C and 5D). These data suggest that the activity-dependent down-regulation of AMPA-induced current is also mediated by calpain.

Figure 5
figure 5

AMPA-induced changes in AMPA-induced inward current in prolonged culture of rat hippocampal neurons. (A) Typical whole-cell currents induced by AMPA (100 μmol/L, 50 ms) recorded in hippocampal neurons cultured for 7–8 DIV in the absence and presence of AMPA (100 μmol/L, 4 h). (B) Average current density in cultured hippocampal neurons following different treatments. (C) Typical whole cell currents induced by AMPA (100 μmol/L, 50 ms) recorded in hippocampal neurons cultured for 17–18 DIV in the absence and presence of AMPA (100 μmol/L, 4 h) following the administration of the calpain inhibitor calpeptin, the lysosome enzyme inhibitor leupeptin, or the caspase inhibitor zVAD (added simultaneously with AMPA treatment). The holding potential was −60 mV. (D) Average current density in cultured hippocampal neurons following different treatments (n=12 for all groups). The data are presented as the mean±SEM, bP<0.05, one-way ANOVA.

PowerPoint slide

Full size image

Discussion

In this study, we found that (1) the pre-treatment of rat hippocampal neurons with the receptor agonist AMPA resulted in the down-regulation of TARPγ8 and AMPAR subunit GluA1 expression and diminished the AMPA current; (2) the decrease of TARPγ8, GluA1 expression and AMPA current can be prevented by calpain and lysosome protease inhibitors.

AMPA induced a decrease in TARPγ8 expression

A major finding in this study was the significant decrease of TARPγ8 expression in the hippocampus in an activity-dependent manner. To our knowledge, this is the first study to report the activity-dependent change of the TARP subunit, which suggests that, similar to AMPAR, TARP may also undergo changes in expression level during glutamate-mediated excitation and excitotoxicity and that this change may indirectly affect AMPAR function. The change in the TARP subunit is also development-dependent. The functional consequence of such a change is unknown, but it may be related to the age-related changes in AMPAR function in vivo. The amplitudes of spontaneous or miniature EPSCs are reduced in the anterior piriform cortex of aged mice, indicative of an aging-related decrease in AMPAR function28. The decreased AMPAR function in aging cannot be explained by the unaltered expression of GluA1 or by the reduction of GluA2/3 expression in the aged human brain29, but it may be associated with the altered trafficking or subcellular location of AMPAR. As TARP contributes to AMAPR trafficking, the decreased expression of the TARP subunit revealed in this study may provide a mechanism for aging-related changes in AMPAR subunits and function.

The AMPA-induced decrease of the TARPγ8 level can be blocked by AMPA and NMDA receptor antagonists

Although TARPγ8 expression showed a decreasing tendency in prolonged culture, there was no significant difference in TARPγ8 expression in cultures between 7–8 DIV and 17–18 DIV. AMPA did not change the TARPγ8 expression level in the cultures for 7–8 DIV, but significantly decreased it in the 17–18 DIV culture. Such a down-regulation of the TARPγ8 expression level can be prevented by AMPAR or NMDAR antagonist, suggesting that both AMPAR and NMDAR activation were involved. NMDAR's involvement in increased calcium transience in TARPγ8 over-expressing cells has been reported30.

AMPA-induced decreases of TARPγ8 and GluA1 expression are calpain-dependent.

It is assumed that the AMPA-induced tonic activation of AMPAR and membrane depolarization cause NMDAR activation and Ca2+ influx, which may activate a series of calcium-dependent events, such as the activation of calpain and other proteolytic enzymes. Calpain inhibitor III blocks the decrease of stargazin under seizure conditions induced by kainic acid31, which demonstrates that calpain mediates TARP expression levels under pathological conditions. In our results, we found that the AMPA-induced TARPγ8 decrease can be blocked by a pan-calpain inhibitor. One explanation is that the activation of AMPAR and NMDAR may increase cytoplasmic Ca2+ concentrations, resulting in the activation of calpain and the subsequent cleavage of TARPγ8. The calpain-mediated down-regulation of TARPγ8 in our results was supported by aging-related increases in calpain activity32,33.

Activity-dependent down-regulation of GluA1 observed in this study was in agreement with the results that prolonged glutamate or NMDA treatment reduced the level of GluA1, but not GluA2, subunits in a calpain-dependent manner in cortical cultures18,34. Previous studies have shown that glutamate or kainate treatment decreased the GluA1 expression level, which can be blocked by calpain inhibitor18,24. These results indicate that the Ca2+-activated protease calpain was involved in GluA1 cleavage. Interestingly, other proteolytic enzymes, such as caspase, cleave GluA1 but not GluA2 in cortical culture27,35, indicating that the activation of these proteolytic enzymes modulates AMPAR function through cleavage of GluA1, but not GluA2.

Previous studies have demonstrated that certain caspases, such as caspase-3, an apoptotic enzyme, cleave GluA1 under apoptotic stimulation27,35. Our results show that caspase inhibitor did not prevent the AMPA-induced down-regulation of TARPγ8 and GluA1, suggesting that the activity-dependent conditions do not cause caspase activation, even in prolonged culture (Lu and Mattson, unpublished data). We also found that the lysosome-mediated proteolysis inhibitor leupeptin can block the AMPA-induced down-regulation of TARPγ8 and GluA1, and this may be related to its non-specific inhibitory role in many other proteolytic enzymes, including calpain36. Thus, the AMPA-induced down-regulation of TARPγ8 and GluA1 occurs mainly through calpain activation.

AMPA-induced down-regulation of AMPA-induced currents

The whole-cell patch clamp recording in hippocampal neurons demonstrated that the down-regulation of AMPA-induced current under activity-dependent conditions was blocked by calpain and lysosome enzyme inhibitors, suggesting that there was an activity-dependent functional change in AMPAR under the same conditions as caused the down-regulation of TARPγ8 and GluA1. Calpain-mediated suppression of AMPAR-mediated currents in cortical pyramidal neurons was also observed in the conditions of prolonged glutamate or NMDA treatment after ischemic insult18.

Given the close association between TARPs and AMPAR subunits9,37, the calpain-mediated decrease of TARPγ8 may contribute to the altered AMPAR function10. Rouach et al10 found that both GluA1 and GluA2/3 were reduced in TARPγ8 knockout mice, which was different from the results of this study. We found that the AMPA-induced down-regulation of GluA1 and TARPγ8, but not GluA2/3, likely occurs through a mechanism of calpain activation in the presence of AMPA. Our results were similar to the reported selective down-regulation of GluA1 but not GluA2/3 under trophic factor withdrawal35,38. Because the activity-dependent down-regulation of the GluA1 level occurs, it is difficult to define the contribution of the down-regulated TARPγ8 to the simultaneous down-regulation of GluA1, but not GluA2/3.

Limited studies have adopted prolonged neuronal culture as an aging model in vitro39,40. In this study, we observed the decreased expression of TARP in the prolonged culture of hippocampal neurons, which was consistent with the results found in aged hippocampal tissue in vivo. The decreased expression of GluA1 in vitro did not agree with the results found in the older human brain. Thus, the in vitro model of aging cannot fully mimic the normal aging process. However, the results observed in this study may provide some important clues to better understand the mechanism of normal aging and neurodegeneration and better strategies for preventing and treating aging-related diseases, such as Alzheimer's disease and stroke24,41,42.

In conclusion, this study demonstrates the activity-dependent down-regulation of TARPγ8 and GluA1 in prolonged neuronal culture, emphasizing a role of calpain. The down-regulation of TARPγ8 likely contributed to the impaired AMPAR trafficking and thus impaired fast neuronal transmission, regardless of the changes in AMPAR expression.

Author contribution

Jian-gang WANG performed the experiments, analyzed the data and wrote the paper; Ya-li WANG performed the experiments and analyzed the data; Fang XU and Jing-xi ZHAO performed the experiments; Si-yuan ZHOU, Yi YU and Xiao-fang WANG analyzed the data; Paul L CHAZOT designed the research; Cheng-biao LU designed the experiments, performed the experiments and wrote the paper.