Transgenic mouse and cell culture models demonstrate a lack of mechanistic connection between endoplasmic reticulum stress and tau dysfunction

Abstract

In vivo aggregation of tau protein is a hallmark of many neurodegenerative disorders, including Alzheimer's disease (AD). Recent evidence has also demonstrated activation of the unfolded protein response (UPR), a cellular response to endoplasmic reticulum (ER) stress, in AD, although the role of the UPR in disease pathogenesis is not known. Here, three model systems were used to determine whether a direct mechanistic link could be demonstrated between tau aggregation and the UPR. The first model system used was SH-SY5Y cells, a neuronal cultured cell line that endogenously expresses tau. In this system, the UPR was activated using chemical stressors, tunicamycin and thapsigargin, but no changes in tau expression levels, solubility, or phosphorylation were observed. In the second model system, wild-type 4R tau and P301L tau, a variant with increased aggregation propensity, were heterologously overexpressed in HEK 293 cells. This overexpression did not activate the UPR. The last model system examined here was the PS19 transgenic mouse model. Although PS19 mice, which express the P301S variant of tau, display severe neurodegeneration and formation of tau aggregates, brain tissue samples did not show any activation of the UPR. Taken together, the results from these three model systems suggest that a direct mechanistic link does not exist between tau aggregation and the UPR.

Fig. 1

SH-SY5Y cells, 8 days following differentiation by retinoic acid, were untreated (DMSO) or treated with tunicamycin (TM), thapsigargin (TG), okadaic acid (OA), or a combination of OA and TM or TG as described in Materials and Methods. A: A segment of the XBP1 gene containing the region spliced by IRE1 was amplified from cDNA of cell lysates by PCR. PCR products were separated on a 2.5% agarose gel to resolve unspliced (U) and spliced (S) XBP1. B: Quantification of XBP1 splicing from the agarose gels was performed as described in Materials and Methods. Error bars represent the standard deviation from the mean of two biological replicates. Asterisks indicate a statistically significant difference from DMSO control with a 95% confidence level.

Fig. 2

A: Levels of total eIF2α and phosphorylated eIF2α in lysates from SH-SY5Y cells, which were treated as described in Figure 1, were measured by Western blotting. Glyceraldehyde-3-phosphate de-hydrogenase (GAPDH) was used as a protein loading control. B: Quantification from Western analysis of phosphorylated eIF2α normalized to GAPDH levels. Error bars represent the standard deviation from the mean of two biological replicates. Asterisks indicate a statistically significant difference from DMSO control with a 95% confidence level.

Fig. 3

A: Levels of total GSK-3β and phosphorylated GSK-3β in lysates from SH-SY5Y cells, which were treated as described in Figure 1, were measured by Western blotting. GAPDH was used as a protein loading control. B: Quantification from Western analysis of phosphorylated GSK-3β levels normalized to GAPDH levels. Error bars represent the standard deviation from the mean of two biological replicates. Asterisks indicate a statistically significant difference from DMSO control with 95% confidence level.

Fig. 4

A: Levels of total tau and tau phosphorylated at serine 199 or 202 in lysates from SH-SY5Y cells, which were treated as described in Figure 1, were measured by Western blotting. GAPDH was used as a loading control. Quantification from Western analysis of phosphorylated tau levels normalized to GAPDH levels for phosphorylation at serine 199/202 (B), serine 262 (C), and serine 396 (D). Error bars represent the standard deviation from the average of two (B,C) or three (D) biological replicates. Asterisk indicates statistically significant difference from DMSO control with a 95% confidence level.

Fig. 5

T-REx 293 cells, which were either untransfected or stably transfected with a plasmid containing the gene for either WT tau, P301L tau, or β-galactosidase (β-gal), were treated with tetracycline to induce protein expression as described in Materials and Methods. Expression of recombinant tau was confirmed by Western blotting for tau and phosphorylated tau (serines 199/202).

Fig. 6

Protein expression was induced in T-REx 293 cells as described for Figure 5. A segment of the XBP1 gene containing the region spliced by Ire1 was amplified via PCR from cDNA collected from T-REx 293 cell lysates. PCR products were separated on a 2.5% agarose gel to resolve unspliced (U) and spliced (S) XBP1. SH-SY5Y cells treated with TG for 1 hr were used as a positive control. Results shown are representative of two biological replicates.

Fig. 7

Protein expression was induced in T-REx 293 cells as described in Figure 5. A: Levels of total eIF2α and phosphorylated eIF2α in T-REx 293 cell lysates were measured by Western blotting. GAPDH was used as a protein loading control. B: Quantification from Western analysis of phosphorylated eIF2α levels normalized to GAPDH levels. Error bars represent the standard deviation from the average of two biological replicates.

Fig. 8

A segment of the XBP1 gene containing the region spliced by Ire1 was amplified via PCR from cDNA of WT and PS19 mouse tissue samples. PCR products were separated on a 2.5% agarose gel to resolve unspliced (U) and spliced (S) XBP1. SH-SY5Y cells treated with TM for 2.5 hr (6 months and 11 months PS19) or TG for 1 hr (11 months WT) were used as a positive control. Results shown for PS19 mice are representative of three replicate specimens. Ctx, cerebral cortex; HP, hippocampus; Cereb, cerebellum; B. Stem, brainstem; S.C., spinal cord; Olf. Bulb, olfactory bulb.

Fig. 9

Levels of BiP and calreticulin mRNA in tissue samples from the cerebral cortex, hippocampus, and cerebellum were measured by qPCR and normalized to levels of Rpl19, a ribosomal protein not regulated by ER stress. The relative -fold changes in mRNA levels for PS19 mice compared with a nontransgenic control are plotted. Error bars represent standard deviation from the mean of samples from three transgenic animals.