Partial restoration of protein synthesis rates by the small molecule ISRIB prevents neurodegeneration without pancreatic toxicity

Abstract

Activation of the PERK branch of the unfolded protein response (UPR) in response to protein misfolding within the endoplasmic reticulum (ER) results in the transient repression of protein synthesis, mediated by the phosphorylation of the alpha subunit of eukaryotic initiation factor 2 (eIF2α). This is part of a wider integrated physiological response to maintain proteostasis in the face of ER stress, the dysregulation of which is increasingly associated with a wide range of diseases, particularly neurodegenerative disorders. In prion-diseased mice, persistently high levels of eIF2α cause sustained translational repression leading to catastrophic reduction of critical proteins, resulting in synaptic failure and neuronal loss. We previously showed that restoration of global protein synthesis using the PERK inhibitor GSK2606414 was profoundly neuroprotective, preventing clinical disease in prion-infected mice. However, this occured at the cost of toxicity to secretory tissue, where UPR activation is essential to healthy functioning. Here we show that pharmacological modulation of eIF2α-P-mediated translational inhibition can be achieved to produce neuroprotection without pancreatic toxicity. We found that treatment with the small molecule ISRIB, which restores translation downstream of eIF2α, conferred neuroprotection in prion-diseased mice without adverse effects on the pancreas. Critically, ISRIB treatment resulted in only partial restoration of global translation rates, as compared with the complete restoration of protein synthesis seen with GSK2606414. ISRIB likely provides sufficient rates of protein synthesis for neuronal survival, while allowing some residual protective UPR function in secretory tissue. Thus, fine-tuning the extent of UPR inhibition and subsequent translational de-repression uncouples neuroprotective effects from pancreatic toxicity. The data support the pursuit of this approach to develop new treatments for a range of neurodegenerative disorders that are currently incurable.

Figure 1

Scheme of the PERK branch of the UPR. Unfolded proteins activate PERK, which phosphorylates eIF2α. This represses translation at the level of initiation. Some proteins, however, escape this repression and are preferentially translated after eIF2α phosphorylation, such as ATF4, which leads to the translation of the pro-apoptotic CHOP. Chronic translational repression leads to neurodegeneration in prion disease. Signaling through the integrated stress response (ISR) can also lead to eIF2α-P and translational repression. The sites of action of GSK2606414 and ISRIB are shown

Figure 2

ISRIB restores translation in prion-diseased mice, downstream of eIF2α phosphorylation. (a) ISRIB treatment (blue bars) lowers ATF4 levels while leaving eIF2α-P unchanged in prion-diseased mice when compared with vehicle treated (gray bars) animals, confirming its site of action downstream of eIF2α-P. Representative immunoblots of hippocampal lysates and bar chart quantitating relative levels of proteins in three independent samples are shown. (b) ISRIB treatment partially restores translation as measured by [³⁵S]-methionine incorporation into hippocampal slices, compared with vehicle-treated animals (n=3-4 for each group). ***P<0.001, Student's t-test, two-tailed. Bar graphs show mean values±S.E.M.

Figure 3

ISRIB confers neuroprotection in prion-diseased mice, via a mechanism independent of prion replication. (a) Representative images of hematoxylin and eosin-stained hippocampal sections from uninfected control (left hand panels) and prion-infected mice treated with vehicle (central panels) or ISRIB (right hand panels). Vehicle-treated mice show extensive neuronal loss in the CA1-3 region, with associated spongiosis, while ISRIB treatment prevents neurodegeneration and reduces spongiosis. Scale bar, top row 400 μm, bottom row 50 μm). (b) ISRIB treatment does not affect the levels of total PrP and PrP^Sc. Total PrP and PrP^Sc levels, detected with and without proteinase K (PK) digestion, were equivalent in prion-infected mice treated with vehicle or ISRIB. Representative immunoblots of three independent hippocampal lysate samples for total PrP and PrP^Sc after PK (50 μg/ml) digestion. Control samples are from mice inoculated with normal brain homogenate

Figure 4

ISRIB treatment significantly extends survival in prion-infected animals compared with vehicle-treated mice. Kaplan–Meier plot, controls n=9 (black bar), vehicle n=9 (gray bar), ISRIB treated n=12 (blue bar). **P<0.005, Mantel–Cox test

Figure 5

ISRIB is not toxic to the pancreas, unlike GSK2606414. (a) GSK2606414 treatment (green bar) mildly raises blood glucose levels compared with control (black bar), vehicle-treated (gray bar) and ISRIB-treated (blue bar) mice (n=6–9 for each group). (b) GSK2606414 treatment leads to a significant reduction in pancreas weight, while ISRIB treatment has no effect (n=3–6 for each group). (c) Representative images of hematoxylin and eosin-stained pancreas sections. GSK2606414 treatment leads to extensive destruction of exocrine acinar pancreatic tissue, while ISRIB-treated tissue is histologically normal. Scale bar=200 μm. *P<0.05, ***P<0.001, Student's t-test, two-tailed. Bar graphs show mean values±S.E.M.

Figure 6

GSK2606414 fully inhibits UPR/ISR-induced translational repression, whereas ISRIB only partially inhibits translational attenuation. (a) Immunoblot of puromycinylated (nascent) proteins, phosphorylated eIF2α and total eIF2α recovered from lysates of AR42j rat pancreatic acinar cells and resolved by SDS–PAGE. Where indicated the cells were exposed to the ER stress-inducing agent thapsigargin (Tg, 0.3 μM) in the presence or absence of ISRIB (100 nM) or the PERK kinase inhibitor GSK2606414 (1 μM) for 30 min before lysis. Puromycin (10 μg/ml) was added during the last 10 min of the incubation. Where indicated (‘Washout') the cells were exposed to the aforementioned concentration of ISRIB or GSK2606414 for 30 min, washed thoroughly with media and incubated further for 1.5 or 7.5 h, before thapsigargin challenge and puromycin labeling. The extent of puromycin labeling (‘Translation') integrated across the surface of each lane is stated below the immunoblot (this value is set arbitrarily to 100 in the reference lane 1, and scaled linearly in the other lanes). (b) Reversal of ISR-mediated inhibition of protein synthesis in the experimental conditions described in a (Lane 1, 2, 3 and 6) was quantified by integrating the puromycinylated protein signal across the surface of each lane and applying the algebraic operator: Rⁿ=[(Pⁿ−P^Tg)÷(P^UT−P^Tg)] × 100. Rⁿ is the reversal in lane n. Pⁿ, P^UT and P^Tg are the puromycinylated protein signals from the sample of interest (lane n), the untreated sample (lane 1) and the sample exposed to thapsigargin alone (lane 2), respectively. Shown are mean values±S.E.M. (n=4, *P<0.05, Student's t-test). (c) Blot: immunoblot of puromycinylated (nascent) proteins, phosphorylated eIF2α and total eIF2α recovered from lysates of AR42j rat pancreatic acinar cells and resolved by SDS–PAGE. Where indicated the cells were exposed to the ER stress-inducing agent thapsigargin (TG, 0.3 μM) in the presence or absence of the indicated concentrations of ISRIB for 30 min before lysis. Puromycin (10 μg/ml) was added during the last 10 min of the incubation. Bar chart: reversal of ISR-mediated inhibition of protein synthesis in the experiment described in ‘a' was quantified by integrating the puromycinylated protein signal across the surface of each lane and applying the algebraic operator: Rⁿ=[(Pⁿ−P^Tg)÷(P^UT−P^Tg)] × 100. Rⁿ is the reversal in lane n. Pⁿ, P^UT and P^Tg are the puromycinylated protein signals from the sample of interest (lane n), the untreated sample (lane 2) and the sample exposed to thapsigargin alone (lane 10), respectively. Shown are mean values±variation in an experiment reproduced twice. (d) Blot: as in ‘a' above, but applying the indicated concentrations of GSK2606414. Bar chart: as in (c) above, but pertaining to GSK2606414. (e) Luciferase expression in CHO cells expressing a luciferase reporter construct under control of the CHOP promoter. Tunicamycin treatment leads to robust luciferase expression, which is completely inhibited by GSK2606414 treatment. ISRIB treatment only partially reduces expression (n=9 for each group). Shown are mean values±S.E.M, *P<0.05, ***P<0.001 Student's t-test, two-tailed