The unfolded protein response: mechanisms and therapy of neurodegeneration

Abstract

Activation of the unfolded protein response is emerging as a common theme in protein-misfolding neurodegenerative diseases, with relevant markers observed in patient tissue and mouse models. Genetic and pharmacological manipulation of the pathway in several mouse models has shown that this is not a passive consequence of the neurodegeneration process. Rather, overactivation of the protein kinase RNA-like ER kinase (PERK, encoded by EIF2AK3) branch of the unfolded protein response directly contributes to disease pathogenesis through the critical reduction in neuronal protein synthesis rates, essential for learning and memory and for neuronal survival. The pharmacological inhibition of this process in these models is strikingly neuroprotective, resulting in the discovery of the first small molecule preventing neurodegeneration and clinical disease in vivo This now represents a potential generic approach for boosting memory and preventing neurodegeneration across the spectrum of these disorders, albeit with some exceptions, independent of disease-specific proteins. Targeting the unfolded protein response, and particularly PERK-branch mediated translational failure is thus an increasingly compelling strategy for new treatments for dementia and neurodegenerative disease.

Keywords: mouse models of neurodegeneration; neurodegeneration; neuroprotection; treatment of dementia; unfolded protein response.

Dysregulation of the unfolded protein response is an emerging theme in neurodegenerative disease. Smith and Mallucci present mechanistic insights into this phenomenon from animal models that explain human neuropathological findings. They report that modulation of the pathway can have profound neuroprotective effects, with therapeutic implications for the treatment of dementia.

Figure 1

The unfolded protein response. Cells have adopted an intrinsic network of protein quality control that promotes efficient protein folding and either the re-folding or degradation of misfolded and aggregated protein. The UPR is one such protective mechanism that acts to restore protein homeostasis (proteostasis) upon the accumulation of misfolded and aggregated protein in the endoplasmic reticulum. The UPR signals through three transmembrane proteins, PERK, IRE1 and ATF6. PERK phosphorylates eIF2α, leading to a rapid attenuation of protein synthesis as well as facilitating the non-canonical translation of ATF4 mRNA. ATF4 upregulates proteins involved in folding, amino acid metabolism, autophagy and redox balance. eIF2α is also a hub for signalling through the related integrated stress response (ISR), where other kinases phosphorylate eIF2α with resultant translational attenuation and ATF4 upregulation. The IRE1 and ATF6 branches of the UPR activate the transcription factors XBP1 and ATF6 (N), respectively. Target genes encode proteins involved in protein folding, endoplasmic reticulum-associated protein degradation (ERAD) and lipid biosynthesis.

Figure 2

Genetic and pharmacological inhibition of PERK signalling is neuroprotective in several models of neurodegenerative disease. The accumulation of misfolded proteins such as PrP and tau results in chronic activation of the PERK branch of the UPR, leading to elevated levels of eIF2α-P. Consequently, there is a sustained reduction in global protein synthesis rates, causing a decrease in the levels of key synaptic proteins, driving synaptic failure, memory impairment and ultimately neuronal loss. In prion-diseased mice, inhibiting the dephosphorylation of eIF2α using salubrinal exacerbates disease, whereas reducing the levels of eIF2α-P through the overexpression of GADD34 is neuroprotective. Pharmacological inhibition of PERK signalling using GSK2606414 reduces PERK-P and eIF2α-P levels in prion-diseased mice, tauopathy mice and in Drosophila expressing TDP-43. As a result of this, in the two mouse models, protein synthesis rates are restored, enabling the restoration of synaptic plasticity, memory formation and promoting neuronal survival. Genetic reduction of eIF2alpha-P levels due to PERK or ISR kinase haploinsufficiency restores memory in Alzheimer's disease mouse models. ISRIB restores protein synthesis rates downstream of eIF2α-P and enhances memory formation in wild-type (WT) mice and is neuroprotective in prion-diseased mice. [1] Moreno et al., 2013; [2] Radford et al., 2015; [3] Kim et al., 2014; [4] Moreno et al., 2012; [5] Costa-Mattioli et al., 2007; [6] Ma et al., 2013; [7] Sidrauski et al., 2013; [8] Halliday et al., 2015. Aβ = amyloid-β; AD = Alzheimer’s disease.

Figure 3

PERK inhibition by GSK2606414 prevents clinical disease and neuronal loss in prion-diseased and frontotemporal dementia mice. (A) Prion-diseased mice were treated with vehicle or GSK2606414 from 7 weeks post-infection (wpi). At 12 weeks post-infection, GSK2606414-treated mice lacked the clinical signs of prion disease, with normal posture and movement of hind legs. (B) At 12 weeks post-infection, GSK2606414-treated prion-diseased mice showed marked neuroprotection in the hippocampus. (C) GSK2606414-treated frontotemporal dementia mice show normal grooming, posture and movement compared to vehicle-treated mice. (D) GSK2606414 treatment resulted in marked neuroprotection in the frontotemporal dementia mice, with the preservation of hippocampal neurons (i–ii). Immunostaining showed reduced levels of phosphorylated tau in the GSK2606414-treated mice (iii–iv).