NMDA receptor mediates tau-induced neurotoxicity by calpain and ERK/MAPK activation

Abstract

The altered function and/or structure of tau protein is postulated to cause cell death in tauopathies and Alzheimer's disease. However, the mechanisms by which tau induces neuronal death remain unclear. Here we show that overexpression of human tau and of some of its N-terminal fragments in primary neuronal cultures leads to an N-methyl-D-aspartate receptor (NMDAR)-mediated and caspase-independent cell death. Death signaling likely originates from stimulation of extrasynaptic NR2B-subunit-containing NMDARs because it is accompanied by dephosphorylation of cAMP-response-element-binding protein (CREB) and it is inhibited by ifenprodil. Interestingly, activation of NMDAR leads to a crucial, sustained, and delayed phosphorylation of extracellular-regulated kinases 1 and 2, whose inhibition largely prevents tau-induced neuronal death. Moreover, NMDAR involvement causes the fatal activation of calpain, which, in turn, degrades tau protein into a 17-kDa peptide and possibly other highly toxic N-terminal peptides. Some of these peptides are hypothesized, on the basis of our in vitro experiments, to initiate a negative loop, ultimately leading to cell death. Thus, inhibition of calpain largely prevents tau degradation and cell death. Our findings unravel a cellular mechanism linking tau toxicity to NMDAR activation and might be relevant to Alzheimer's disease and tauopathies where NMDAR-mediated toxicity is postulated to play a pivotal role.

Conflict of interest statement

Conflict of interest statement: No conflicts declared.

Figures

Fig. 1.

Reduced viability of tau-(1–441)-infected neurons. (A) CGCs were infected at 4 days in vitro with either Lac-Z- or tau-expressing adenovirus vectors at the MOIs indicated. Survival was assessed 24 and 48 h later by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay. Each data point is the mean ± SE of triplicate determinations of three independent experiments and is expressed as percentage of Lac-Z-infected cells, considering the value obtained in these cells as 100% (∗, P < 0.05; ∗∗, P < 0.01 compared with Lac-Z-infected neurons). (B) Western blot analysis of lysates from tau-(1–441)-infected CGCs performed with mAb 9E10 and normalized with β-actin (Upper) and with mAb TAU-1 (Lower) and normalized with a nonspecific (N.S.) band. (C) Immunofluorescence analysis of tau-(1–441)-infected cortical neurons at MOIs of 30 (a) and 120 (b), immunostained with mAb 9E10 (green). Forty-eight hours after infection, nuclei were stained with Hoechst 33258 (blue). A cortical neuron expressing tau with a sign of degeneration (varicosity along a neurite) is shown (c). (Scale bar, 20 μm.) (D) Micrographs of Hoechst 33258-stained cortical neurons after 48 h of infection. Values corresponding to condensed nuclei and to MTT assay are reported below. (E) Western blot analysis of the NR1 subunit expression in CGCs treated with the antisense and scrambled ODNs. (F) CGCs were incubated with NR1 antisense and scrambled ODNs and then infected with Lac-Z and tau-(1–441) vectors for 48 h, when survival was determined. Data are expressed as in A.

Fig. 2.

Impact of various tau fragments on CGC viability. (A) Diagram of tau vectors used in this study. All constructs are derived from the longest htau isoform and expressed with 6 Myc epitope tags fused to the N terminus. (B) CGCs were infected at various MOIs with the vectors indicated. Viability was determined 24 and 48 h after infection and is expressed as reported for Fig. 1A. ∗∗, P < 0.01; ∗, P < 0.05 compared with Lac-Z-infected CGCs at an MOI of 30.

Fig. 3.

Tau toxicity involves extrasynaptic NMDAR, CREB dephosphorylation, and ERK1/2 activation. (A) Survival of CGCs infected with vectors indicated in the absence or in the presence of ifenprodil (10 μM), CNQX (40 μM), or ω-conotoxin MCVII (1 μM), a blocker of N-, P-, and Q-type calcium channels, reported to be inhibited by ifenprodil. Data are reported as in Fig. 1A. (B) Lysates from Lac-Z-, tau-(1–441)-, and tau-(1–44)-infected CGCs treated with MK-801 (10 μM), ifenprodil (10 μM), or CNQX (40 μM) for 24 h were probed for phospho-CREB (P-CREB) and total CREB. Representative blots of two independent experiments with similar results are shown. (C and D) Time course analysis of phospho-ERK1/2 (P-ERK1/2) analyzed by Western blotting from lysates of Lac-Z-, tau-(1–441)-, and tau-(1–44)-infected CGCs in the absence (C) or in the presence (D) of MK-801. Representative blots of three independent experiments with similar results are shown. (E) Survival at 24 and 48 h of tau-infected neurons in the absence or in the presence of UO126 (5 μM) or PD98059 (15 μM). Data are reported as in Fig. 1A. ∗, P < 0.05; ∗∗, P < 0.01 compared with tau-untreated cells). (F) Both wild-type (ERK1+/+) and ERK1-deficient (ERK1−/−) CGCs were prepared from 6-day-old mice and infected with the vectors indicated. Data are reported as in Fig. 1A. ∗, P < 0.05 compared with Lac-Z-infected CGCs. (Upper Right) Survival was determined 48 h later by the MTT assay. (Lower Right) Western blot detecting ERK1/2 in CGC lysates from ERK1+/+ and ERK1−/− mice.

Fig. 4.

Tau toxicity is accompanied by NMDAR-dependent activation of calpain. (A) Western blot analysis, with antibody against calpain, of lysates from CGCs infected with the vectors indicated, in the absence or in the presence of MK-801 or calpeptin (10 μM) for 24 h (arrowhead: mature calpain; arrow, active calpain). (B) Viability of CGCs 24 h after infection with LacZ or tau vectors, in the absence or in the presence of calpeptin (10 μM) or Ad-calpastatin. Data are reported as in Fig. 1A. ∗∗, P < 0.01 by unpaired t test compared with untreated tau-infected cells.