Contents lists available at ScienceDirect

Neurobiology of Disease

journal homepage: www.elsevier.com/locate/ynbdi

Targeting PERK signaling with the small molecule GSK2606414 preventsneurodegeneration in a model of Parkinson's disease

Gabriela Mercadoa,b,c,⁎, Valentina Castilloa,b,c, Paulina Sotoa,b,c, Nélida Lópeza,b,c,Jeffrey M. Axtend, Sergio P. Sardie, Jeroen J.M. Hoozemansf, Claudio Hetza,b,c,g,h,⁎,1

a Biomedical Neuroscience Institute, Faculty of Medicine, University of Chile, Santiago, Chileb Program of Cellular and Molecular Biology, Institute of Biomedical Sciences, University of Chile, Santiago, Chilec Center for Geroscience, Brain Health and Metabolism, Santiago, Chiled Virtual Proof of Concept Discovery Performance Unit, GlaxoSmithKline, PA, USAeNeuroscience Therapeutic Area, Sanofi, 49 New York Avenue, Framingham, MA 01701, USAfDepartment of Pathology, VU University Medical Center, Amsterdam Neuroscience, Amsterdam, The Netherlandsg Buck Institute for Research on Aging, Novato, CA 94945, USAhDepartment of Immunology and Infectious Diseases, Harvard School of Public Health, Boston, MA 02115, USA

A R T I C L E I N F O

Keywords:Parkinson's diseaseUPRISRER stressProteostasisPERKGSK2606414

A B S T R A C T

Parkinson's disease (PD) is the second most common neurodegenerative disorder, leading to the progressivedecline of motor control due to the loss of dopaminergic neurons in the substantia nigra pars compacta (SNpc).Accumulating evidence suggest that altered proteostasis is a salient feature of PD, highlighting perturbations tothe endoplasmic reticulum (ER), the main compartment involved in protein folding and secretion. PERK is acentral ER stress sensor that enforces adaptive programs to recover homeostasis through a block of proteintranslation and the induction of the transcription factor ATF4. In addition, chronic PERK signaling results inapoptosis induction and neuronal dysfunction due to the repression in the translation of synaptic proteins. Herewe confirmed the activation of PERK signaling in postmortem brain tissue derived from PD patients and threedifferent rodent models of the disease. Pharmacological targeting of PERK by the oral administration ofGSK2606414 demonstrated efficient inhibition of the pathway in the SNpc after experimental ER stress stimu-lation. GSK2606414 protected nigral-dopaminergic neurons against a PD-inducing neurotoxin, improving motorperformance. The neuroprotective effects of PERK inhibition were accompanied by an increase in dopaminelevels and the expression of synaptic proteins. However, GSK2606414 treated animals developed secondaryeffects possibly related to pancreatic toxicity. This study suggests that strategies to attenuate ER stress levels maybe effective to reduce neurodegeneration in PD.

1. Introduction

Parkinson's disease (PD) is the second most common neurodegen-erative disease and the predominant cause of movement problems, af-fecting 1–2% of the population over 60 years of age (Forman et al.,2004; Vila and Przedborski, 2004) with increasing incidence worldwidein the last years (Dorsey and Bloem, 2017). PD is a slowly evolvingdisorder, recognized mostly by its motor symptoms such as bradyki-nesia, resting tremor and postural rigidity (Postuma et al., 2015). Theimpairment of motor control in PD is the result of a progressive

degeneration of dopaminergic neurons in the substantia nigra parscompacta (SNpc) (Damier et al., 1999), resulting in the gradual de-pletion of dopamine content in the striatum. PD is histopathologicallycharacterized by the accumulation of intracellular protein inclusionscalled Lewy Bodies (LBs), where aggregates of misfolded α-Synucleinare the major components (Spillantini et al., 1997). There is no cure forPD and available palliative treatments only aim to restore dopaminedeficits. PD is part of a group of human diseases characterized by theaccumulation of misfolded proteins in the brain collectively defined asprotein misfolding disorders (PMDs). These group of diseases include

https://doi.org/10.1016/j.nbd.2018.01.004Received 27 September 2017; Received in revised form 21 December 2017; Accepted 8 January 2018

⁎ Corresponding authors at: Institute of Biomedical Sciences, University of Chile, Independencia 1027, Sector B, Santiago, Chile.

1 www.hetzlab.cl.E-mail addresses: mariamercado@med.uchile.cl (G. Mercado), clahetz@med.uchile.cl, chetz@hsph.harvard.edu (C. Hetz).

Abbreviations: PD, Parkinson's disease; SNpc, substantia nigra pars compacta; PrDs, prion-related disorders; UPR, unfolded protein response; ISR, integrated stress response; PERK, RNA-like ER kinase; ATF4, activating of transcription-4; ATF3, activating of transcription-3; CHOP, C/EBP-homologous protein; TRB3, tribbles-related protein 3; TH, tyrosine hydroxylase; 6-OHDA, 6-hydroxydopamine; AAV, adeno-associated virus

Neurobiology of Disease 112 (2018) 136–148

0969-9961/ © 2018 Elsevier Inc. All rights reserved.

T

amyotrophic lateral sclerosis (ALS), Alzheimer's disease (AD), Prion-related disorders (PrDs), Huntington's disease among others (Bertramand Tanzi, 2005; Soto, 2003). Although the neuronal populations andthe clinical manifestation of specific PMDs are diverse, most of theseconditions share the occurrence of proteostasis alterations which mayunderlay the occurrence of protein misfolding and neuronal dysfunc-tion (Ciechanover and Kwon, 2017; Hetz and Saxena, 2017; Kaushikand Cuervo, 2015).

Emerging evidence suggests that a reduction in the buffering ca-pacity of the proteostasis network during aging operates as risk factor toundergo neurodegeneration (Kaushik and Cuervo, 2015). One of thenodes of the proteostasis network more affected in PD involves dys-function of the folding capacity of the endoplasmic reticulum (ER)(Michel et al., 2016). In fact, ER stress is a common feature of PD asreported in patient derived post-mortem brain tissue (Conn et al., 2004;Hoozemans et al., 2007; Selvaraj et al., 2012; Slodzinski et al., 2009)and various animal and cellular models of the disease (reviewed inMercado et al., 2016). ER stress was also shown to be a major patho-logical signature in stem cell derived-dopaminergic neurons from PDpatients (Chung et al., 2013; Heman-Ackah et al., 2017). At the mole-cular level, the induction of chronic ER stress by α-Synuclein has beenrelated to a direct inhibition of the vesicular trafficking between ER andGolgi (Cooper et al., 2006; Gitler et al., 2008; Su et al., 2010;Thayanidhi et al., 2010). In addition, its accumulation inside the ERresults in abnormal interactions with BiP (Bellucci et al., 2011; Collaet al., 2012a; Colla et al., 2012b), which may alter protein folding andmaturation. ER stress has been also linked to other PD genes, includingLRRK2, Parkin, Pael-R, DJ-1, and ATP13A2 (Mercado et al., 2016;Tsujii et al., 2015). Altogether, several studies delineate a paradigmwhere proteostasis defects and abnormal protein accumulation duringaging results in neuronal loss and synaptic dysfunction in PD. Im-portantly, ER stress is also emerging as a common and transversal pa-thological mechanism of most frequent neurodegenerative diseases(Hetz and Saxena, 2017; Scheper and Hoozemans, 2015; Smith andMallucci, 2016).

To alleviate ER stress, cells activate the unfolded protein response(UPR), an integrated signaling pathway that aims to re-establish pro-teostasis (Hetz, 2012; Walter and Ron, 2011). However, chronic or ir-reversible ER stress can lead to cell death by apoptosis (Tabas and Ron,2011; Urra et al., 2013). The UPR is mediated by three main stresssensors known as Inositol requiring enzyme alpha (IRE1α), RNA-like ERkinase (PERK) and activating of transcription-6 (ATF6). PERK is a type-Itransmembrane kinase located at the ER membrane controlling cell fateunder ER stress (Hetz and Papa, 2017). Activation of PERK leads to thephosphorylation of the eukaryotic translation initiator factor eIF2α,inhibiting general protein synthesis (Harding et al., 1999). This eventquickly reduces the load of proteins entering the ER, having an im-portant pro-survival role (Harding et al., 2000). Phosphorylation ofeIF2α allows the selective translation of certain mRNAs containingshort open reading frames in their 5′-untranslated regions, highlightingthe mRNA encoding for activating of transcription-4 (ATF4) as a majorsubstrate (Harding et al., 2000). ATF4 controls the expression of acluster of genes involved in redox control, amino acid metabolism,macroautophagy, protein synthesis and folding (Han et al., 2013;Harding et al., 2003). Under prolonged or chronic stress, ATF4 alsoregulates the expression of pro-apoptotic genes such as the C/EBP-homologous protein (CHOP, also known as GADD153), tribbles-relatedprotein 3 (TRB3) and members of the BCL-2 protein family (Ma et al.,2002; Ohoka et al., 2005; Zinszner et al., 1998). The levels of eIF2αphosphorylation are negatively controlled by two phosphatase com-plexes, the ER stress inducible form composed by PPP1R15A/GADD34(a target of ATF4/CHOP) and the constitutive phosphatase PPP1R15B/CreP (Tsaytler and Bertolotti, 2013). Dephosphorylation of eIF2α mayburst protein synthesis on stressed cells, inducing oxidative stress andproteotoxicity (Marciniak et al., 2004; Zinszner et al., 1998). Im-portantly, in addition to PERK, eIF2α phosphorylation is a convergent

point of several stress pathways mediated by the kinases GCN2, PKRand HRI, known as the integrated stress response (ISR) (Pakos-Zebruckaet al., 2016).

The involvement of ER stress in neurodegenerative diseases iscomplex and may depend on the specific outputs controlled by the UPR.Extensive studies have linked PERK signaling with PMDs and otherpathological conditions affecting the nervous system. For example,CHOP expression was reported as a factor mediating dopaminergicneuron loss in toxicological models of PD in mice (Silva et al., 2005).ATF4 overexpression at the SNpc using adeno-associated viruses (AAV)also reduced dopaminergic neuron survival (Gully et al., 2016). Simi-larly, PERK inhibition was shown to protect against mutant Pink1 andParkin in fly models (Celardo et al., 2016). In contrast, the ATF6 andIRE1α pathways have been proposed to be mostly neuroprotective inthe context of experimental PD (Credle et al., 2015; Egawa et al., 2011;Sado et al., 2009; Valdes et al., 2014). Studies in ALS models showedthat PERK haploinsufficiency exacerbates the progression of the disease(Wang et al., 2011). Consistent with this, treatment of animals withinhibitors of eIF2α phosphatases, or the genetic disruption of GADD34,delay ALS progression (Hetz and Saxena, 2017). In contrast, Atf4 defi-ciency protected against ALS possible due to a reduction in the levels ofapoptosis (Matus et al., 2013), suggesting a dual role of the pathway inthis specific disease. In models of multiple sclerosis (Clayton and Popko,2016) and spinal cord injury (Valenzuela et al., 2012; Valenzuela et al.,2016b), PERK signaling operates as an essential survival factor of oli-godendrocytes, potentiating locomotor recovery (Clayton and Popko,2016). Similar results were reported in models of Charcot–Marie–Toothdisease, where Chop deficiency or inhibition of eIF2α phosphatasesimproves motor function involving protection of Schwan cells (Daset al., 2015; Pennuto et al., 2008; Sidoli et al., 2016). All together, thesestudies indicate that eIF2α phosphorylation is protective in models ofneurodegeneration, possibly by reducing the levels of ER stress andabnormal protein aggregation, whereas ATF4/CHOP expression trig-gers neuronal apoptosis during late disease stages.

Recent findings revealed a new pathological role of PERK signalingin neurodegenerative diseases, where sustained translational repressionhas adverse consequences to synaptic function (Mercado and Hetz,2017). Studies in models of PrDs indicated that chronic eIF2α phos-phorylation represses the synthesis of essential synaptic proteins,leading to behavioral impairment (Moreno et al., 2012). A small mo-lecule that inhibits PERK known as GSK2606414 was identified throughstructure-guided lead optimization (Axten et al., 2012). GSK2606414 isorally active (Moreno et al., 2013) and effective in cancer models(Atkins et al., 2013; Axten et al., 2012). Administration of GSK2606414prevented translational repression on mouse models of PrDs andTauopathies, providing strong neuroprotection (Moreno et al., 2013;Radford et al., 2015). However, another study suggested that PERKactivation protects against Tau pathogenesis (Bruch et al., 2017). In thecontext of experimental AD, genetic ablation of Perk expression in thebrain improved neuronal physiology and reduced cognitive decline (Maet al., 2013). In addition, local expression of ATF4 in the axonal com-partment was shown to trigger neurodegenerative cascade that propa-gates in a cell-nonautonomous manner (Baleriola et al., 2014). Theseobservations are consistent with the observation that the control ofprotein synthesis at the level of eIF2α phosphorylation by the ISR isessential to fine-tune neuronal plasticity (reviewed in (Buffington et al.,2014)). In summary, all these reports suggest a complex scenariowhere, depending on the disease context, PERK signaling may havecontrasting and even opposite effects, impacting neuronal survival/apoptosis, stress mitigation, and synaptic function.

In the current study, we investigated the possible contribution ofPERK to dopaminergic neuron loss. Analysis of human PD postmortembrain tissue, in addition to the brain of three different rodent models ofthe disease, indicated a clear molecular signature associated withPERK/ATF4 activation. We then explored the possible neuroprotectiveeffects of the oral administration of GSK2606414 in a mouse model of

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

137

PD. We confirmed that this route of drug delivery inhibits the target atthe SNpc when animals were exposed to the classical ER stress-inducingagent tunicamycin. Using a pharmacological model of PD, we demon-strated that blocking PERK signaling significantly protects nigral-do-paminergic neurons, improving motor performance. Importantly,GSK2606414 administration elevated the levels of dopamine and itsmetabolite DOPAC, in addition to restoring the levels of the synapticproteins SNAP25 and VAMP2 in the striatum of animals exposed to thePD-inducing neurotoxin. Our results suggest that strategies to reducechronic ER stress signals emerging from PERK may have importanteffects in improving dopaminergic neuron function and survival in PDpatients.

2. Materials and methods

2.1. PD human tissue

Human brain specimens were obtained from the Biobank Pathologyunit, VU University Medical Center under approval of the ethics com-mittee of the VU University Medical Center. Human autopsy materialwas obtained with informed consent for research. Neuropathologicaldiagnosis was performed using immunohistochemical stainings forAmyloidβ, Tau, α-Synuclein, TDP-43 and p62/SQSTM1. Analysis offormalin-fixed and paraffin-embedded tissue from different parts of thebrain was performed, including the frontal cortex (F2), temporal polecortex, parietal cortex (superior and inferior lobule), occipital polecortex including BA17 and 18, amygdala and the hippocampus, es-sentially CA1 and entorhinal area of the parahippocampal gyrus.Staging of pathology was evaluated according to a Braak and Braak forneurofibrillary tangles (Braak and Braak, 1991) and Lewy body ac-cording to Braak et al. (Braak et al., 2003). Formalin fixed paraffinsections (5 μm thick) were mounted on superfrost plus tissue slides(Menzel-Gläser, Germany) and deparaffinized. Subsequently, sectionswere immersed in 0.3% H2O2 in methanol for 30min to quench en-dogenous peroxidase activity. Sections were treated in 10mM pH 6.0citrate buffer heated by microwave during 10min for antigen retrieval.Primary antibodies rabbit anti-pPERK (Thr980, 1:800 dilution, SantaCruz Biotechnology, USA), rabbit anti-peIF2α (Ser52, 1:500 dilution,Sigma, USA) and mouse anti-α-Synuclein (1:10.000 dilution, cloneLB509, Zymed Laboratories, USA) were diluted in normal antibodydiluent (Immunologic, The Netherlands) and applied overnight at 4 °C.Between incubation steps, sections were rinsed with phosphate bufferedsaline (PBS, pH 7.4). As secondary step, sections were incubated withEnVision detection system (goat anti-mouse/rabbit horseradish perox-idase, DAKO) for 30min at room temperature (RT). Signals were de-veloped using 3-amino-9-ethyl-carbazole (AEC, Zymed, USA) and nu-clei were stained with hematoxylin. The immunohistochemistry (IHC)score ranged from 0 to 2 (0: no signal; 1: moderate signal; 2 strongimmunoreactivity). The IHC score for each case was an average of atleast four microscopic fields (magnification ×10).

2.2. Animal treatments

Three months-old male wild-type C57BL/6 mice and female adultSprague–Dawley rats weighing approximately 200 g were used. In ad-dition to transgenic mice expressing mutant α-Synuclein mutant A53T(αSynA53TTg) previously described in (Giasson et al., 2002) and ob-tained from the Jackson's laboratory, USA. All animals were housed in a12 h light/dark cycle, with ad libitum access to food and water. Allexperiments were performed in accordance with the guidelines set bythe animal care and use committee of the Faculty of Medicine at theUniversity of Chile, with approved animal experimentation protocolsCBA#0717 and CBA#0658 FMUCH. GSK2606414 (GlaxoSmithKline,USA) was formulated in the vehicle (0.5% HPMC and 0.1% Tween-80 inwater, pH 4.0) as a suspension and administrated by oral gavage at finalconcentration of 100mg/kg/day. In the case of mice treatments, oral

gavage was applied twice a day [50mg/kg]. And in the case of rats, oralgavage was applied once a day [100mg/kg].

2.3. Pharmacokinetics

GSK2606414 concentrations were quantified in serum (1:1–1:4 di-lution with water) and brain tissue homogenates (frontal cortex, 1:4dilution with 20:80 methanol:water) at Alliance Pharma (Malvern, PA,USA) by protein precipitation followed by HPLC-MS/MS. The lowerlimit of quantification of GSK2606414 was 1.0 ng/mL.

2.4. Pharmacological stimulation of ER stress

To determine the efficacy of GSK2606414 in vivo injections with theER stress agent tunicamycin were performed directly into the SNpcusing brain stereotaxis after a pretreatment with GSK2606414.Tunicamycin injections [5mg/mL] were performed by injecting 2 μL inthe SN using a 5 μL Hamilton syringe (Hamilton, USA) at the followingcoordinates for mice: antero-posterior (AP): +0.29 cm, medio-lateral(ML): −0.13 cm relative to bregma and dorso-ventral (DV): −0.42 cmrelative to skull surface (according to the atlas of Franklin and PaxinosSecond Edition, 2013). For rats, injections were performed followingcoordinates: AP: −0.52 cm, MD: −0.2 cm relative to bregma DV:−0.78 cm relative to skull surface (according to the atlas of Paxinos andWatson Seventh Edition, 2014). Injections were conducted at a rate of0.5 μL/min and the needle was left in place for 5min before retraction.Animals were euthanized 24 h after tunicamycin injections for bio-chemical analysis.

2.5. Genetic model of PD in rats

Recombinant AAV vectors were produced using the plasmid kindlyprovided by Dr. Aebischer team (Coune et al., 2011). Purification ofAAV particles and titration were performed as described before(Valenzuela et al., 2012). Two μl of viral preparation (1.5×109 VGs/mL) were injected in the right brain hemisphere of female adult Spra-gue–Dawley rats using a 5 μL Hamilton syringe at a speed of 0.2 μL/min. The needle was left in place for an additional 5min before beingslowly withdrawn. The SNpc was targeted at the following coordinates:AP: −0.52 cm, MD: −0.2 cm relative to bregma DV: −0.78 cm relativeto skull surface (according to the atlas of Paxinos and Watson SeventhEdition, 2014). Rats were euthanized 3months after surgical procedurefor biochemical analysis.

2.6. Toxicological model of PD in mice

6-OHDA (Sigma-Aldrich, USA) was dissolved fresh at a concentra-tion of 4 μg/μL in 0.2% ascorbic acid (Sigma-Aldrich, USA). Injectionsof 6-OHDA were performed in a single point, injecting 2 μL in the rightstriatum using a 5 μL Hamilton syringe (Hamilton, USA) at the fol-lowing coordinates: AP: +0.07 cm, ML: −0.17 cm relative to bregmaand DV: −0.31 cm relative to skull surface (according to the atlas ofFranklin and Paxinos Second Edition, 2013). The injection was con-ducted at a rate of 0.5 μL/min and the needle was left in place for 5minbefore needle retraction. Mice were euthanized 21 days after the sur-gical procedure for histological and biochemical analysis.

2.7. Motor behavioral studies in mice

All motor tests were performed 7 days before the injection of 6-OHDA (base line) and then 7, 14 and 21 days after injections untilanimal's euthanasia. The cylinder test was used to evaluate spontaneousforepaw akinesia associated to 6-OHDA hemi-lesions as we reported(Castillo et al., 2015). Animals were placed in in a cylinder of trans-parent acrylic of 10 cm in diameter and 20 cm high and left to explorefreely. The camera was placed above the cylinder and movements were

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

138

recorded for 5min. The analysis was performed blind. The results wereplotted as a percentage of touches with left forepaw (contralateral toinjection side) respect to total of number touches with both forepaws.For the rotarod assay, mice were positioned on a rotating wheel (ModelLE8500, Panlab SL) oriented horizontally using acceleration mode,where the speed of rotating wheel start with 4 rpm and increases to40 rpm over 2min. The test was performed 5 times in one day and thetime of latency to fall recorded. The results were plotted as percentageof base line performance. Hanging test was performed as we recentlyreported (Woehlbier et al., 2016). In brief, individual mice were placedhanging with their forepaws on a 39 cm length horizontal wire and35 cm from the base. The reaction of the mouse and the body positionwas observed for 30 s and recorded with a video-camera. The test wasperformed 3 times in one day. A score was derived from each trial. Ascore of 0 was given when the mouse could not hold onto the barfor> 10 s, a score of 1 when the mouse maintain itself on the wire withthe forepaws, a score of 2 if the mouse maintain itself with the forepawsand tried to use its hind paws but without success, a score of 3 if themouse used forepaws and one or two of the hind paws, a score of 4 ifthe mouse used all four paws and the tail and 5 if the mouse activelyescapes of the horizontal wire up to the vertical support and 6 if themouse is able to down the vertical support to the base in<30 s.

2.8. Quantitation of biogenic amines

Mice injected with 6-OHDA and treated with vehicle orGSK2606414 [100mg/kg/day] were euthanized after 21 days. Brainswere rapidly removed and the striatum dissected. Striatal tissues werestored at −80 °C. The tissue extracts were then analyzed using UltraPerformance Liquid Chromatography–Tandem Mass Spectrometry(UPLC-MSMS) in the Biomarkers Core Lab at Columbia UniversityMedical Center. The tissue homogenate was spiked with the internalstandard cocktail, and the protein was precipitated with 0.3 M per-chloric acid and centrifuged at 12,000 g for 5min. The supernatant wasmixed with 1M sodium bicarbonate and 1% dansyl chloride, vortexedand heated at 60 °C for 10min. After chilled on ice for 2min, salt wasprecipitated with 400 μL of acetonitrile, centrifuged at 12,000 g for5min and the upper phase was mixed with 400 μL of ethyl acetate. Thesupernatant after centrifugation was evaporated to dryness under anitrogen stream and suspended in 30 μL acetonitrile for LCMS analysis.LCMS analysis was carried out on a Waters Xevo TQ MS ACQUITY UPLCsystem (Waters, Milford, MA, USA). The system was controlled byMassLynx Software 4.1. Five microliters of the sample was loaded ontoa Waters ACQUITY UPLC BEH Phenyl column (3× 100mm, 1.7 μm),maintained at 40 °C. The flow rate was 300 μL/min. The initial condi-tions were 50% phase A (water containing 0.1% formic acid) and 50%mobile phase B (acetonitrile containing 0.1% formic acid). Solvent Bwas increased linearly to 99% over 5min and maintained till 7.5 minwith a total run time of 10min. Positive ESI-MS/MS with multiple re-action monitoring (MRM) mode was performed using the followingparameters: capillary voltage 4 kV, source temperature 150 °C, deso-lvation temperature 500 °C, desolvation gas flow 1000 L/h.Concentrations of biogenic amines were quantitated by comparing in-tegrated peak areas for those of each species against those of knownamounts of purified standards.

2.9. Tissue preparation for biochemical analysis

Animals were sacrificed by CO2 asphyxiation, brains were removedand the ventral midbrain (containing entire SNpc), striatum, and cortexfrom both hemispheres were rapidly dissected on the surfaced of an ice-cold plastic dish. The tissue was homogenized on 100 μL of ice-cold0.1 M phosphate buffer saline (PBS) (pH 7.4) supplemented proteaseinhibitor and phosphatase inhibitor cocktails (Roche applied science,USA). The homogenate was divided in two fractions for preparing totalmRNA and protein extracts followed by standard methods and

quantification protocols. Protein extraction was performed in TENbuffer (10mM Tris pH 8.0; 1 mM EDTA; 0.1 M NaCl) containing pro-tease inhibitor cocktail and phosphatase inhibitor cocktail (Roche ap-plied science, USA) followed by sonication. Sample quantification wasperformed with micro-BCA assay kit (Pierce, USA). Antibodies and di-lutions used are the following: β-actin (1:20,000, MP Biomedicals,USA), ATF4 (1:2,000, Santa Cruz, USA), α-Synuclein (1:2,000, BDBiosciences, USA), peIF2α (1:500; Abcam, USA), VAMP2 (1:20,000;Abcam, USA), SNAP25 (1:10,000, Abcam, USA), TH (1:1000; Merck-Millipore, USA), p-RIP1 (1:500; Cell Signaling, USA) and p-RIP3(1:1000; Abcam, USA).

2.10. RNA extraction and real time PCR

Total RNA was isolated from ventral midbrain (containing entireSN), striatum and cortex. After homogenization in PBS, we purifiedtotal mRNA using Trizol reagent (Invitrogen, USA) and then cDNA wassynthesized from 1 μg of RNA using Reverse Transcription Kit (AppliedBiosystems, USA). Quantitative real time PCR was performed in aStratagene light-cycler system employing the EVA green fluorescentreagent (BioDyne, USA) and the primers listed in SupplementaryTable 1. XBP1 mRNA splicing assays in mice (Rodriguez et al., 2012)and rats (Toda et al., 2006) were performed as we previously described.

2.11. Histological analysis

Animals were anesthetized with the mixture of ketamine/xylazineanesthesia (Ketamine: 100mg/kg, Xylazine: 10mg/kg) and perfusedthrough the ascending aorta with isotonic saline followed by ice-cold4% paraformaldehyde in 0.1M PBS (pH 7.4). Brains were removed,post-fixed overnight at 4 °C in the same solution and subsequentlyplaced on 30% sucrose (Merck, USA) at 4 °C for 48 h. Brains were frozenin Optimal Cutting Temperature compound (Tissue Tek, USA), andcoronal sections of 25 μm containing the rostral striatum and midbrainwere cut on a microtome (Leica, Germany) with a temperature con-trolled freezing stage (Physitemp Instruments, USA). For im-munohistochemical analysis, nigral and striatal tissue sections werequenched with 3% H2O2 at 37 °C for 30min, blocked at room tem-perature for 2 h with 0.5% bovine serum albumin and 0.1% triton X-100 and incubated overnight at 4 °C with the primary antibody anti-TH(rabbit 1:1,000; Millipore-Sigma, Germany), then incubated for 2 h atroom temperature in a biotinylated goat anti-rabbit antibody (1:500;Vector Laboratories, USA). Sections were then incubated with theavidin-biotin-peroxidase complex (Vector Laboratories, USA) for30min at room temperature and developed with 3,3′-diaminobezidine(Sigma-Aldrich, USA) before mounting on glass slides with Entellanmedium (Merck, Germany). Formalin fixed midbrain paraffin sections(5 μm thick) were mounted on superfrost plus tissue slides (ThermoScientific, USA) and deparaffinized. Subsequently, sections were im-mersed in 0.3% H2O2 in methanol for 30min to quench endogenousperoxidase activity. Sections were treated in 10mM pH 6.0 citratebuffer heated in a steamer during 30min for antigen retrieval andblocked at room temperature for 2 h with 1% goat normal serum and0.1% triton X-100. Primary antibodies anti-peIF2α (rabbit 1:100;Invitrogen, USA) and anti-pPERK (rabbit 1:500; Santa CruzBiotechnology, USA) were diluted in blocking solution and appliedovernight at 4 °C. Between incubation steps, sections were rinsed withTris Buffered Saline (TBS, pH 7.4). As secondary step, sections wereincubated for 2 h at room temperature in a biotinylated goat anti-rabbitantibody (1:500; Vector Laboratories, USA). Sections were then in-cubated with the avidin-biotin-peroxidase complex (VectorLaboratories, USA) for 1 h at room temperature and developed with3,3′-diaminobezidine (Sigma-Aldrich, USA). Images were acquiredusing an Olympus IX71 microscope.

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

139

2.12. Cell counting

To determine the percentage of TH-positive cells loss in the SNpc ofinjected animals, the number of nigral-dopaminergic cells in the in-jected and non-injected side was obtained counting in a blinded mannerthe total number of TH-positive cells obtained by im-munohistochemistry in midbrain correlative tissue sections every100 μm containing the entire SNpc as we reported before (Castillo et al.,2015; Valdes et al., 2014). Results were expressed as the percentage ofTH-positive neuron loss in the injected side compared to the non-in-jected side.

2.13. Densitometry analysis of striatal dopaminergic innervation

Striatal dopaminergic innervation was analyzed by measuring theoptical density (OD), integrated density of gray pixel values, correctedfor non-specific background, of TH immunoreactivity on tissue sectionsevery 100 μm covering the entire striatum. Sections were scanned usingan Epson Perfection V600 Photo scanner, and OD of the striatum wasanalyzed using the ImageJ software (free NIH software: http://rsbweb.nih.gov/ij/). Results are expressed as the percentage of TH im-munoreactivity loss with respect to the contralateral non-injectedhemisphere.

2.14. Statistical analyses

Data are expressed as mean ± SEM. After confirming normal dis-tribution with Skewness/Kurtosis statistic test, Student's t-test was usedto analyze differences in histological, behavioral and biochemicalanalyses. Two-way ANOVA followed by Bonferroni post-test was usedin behavioral experiments and body weight. Statistical analyses wereperformed using GraphPad Prism 5.0 software. Statistical differenceswere considered significant for values of p < .05.

3. Results

3.1. Activation of PERK/ATF4 signaling in the brain of PD patients androdent models of the disease

Previous studies have reported the induction of ER stress in PDbrain, showing a correlation between the accumulation of Lewy bodiesand the activation of PERK at the SNpc (Hoozemans et al., 2007;Selvaraj et al., 2012). To further study the contribution of PERK to PDpathogenesis, we performed a global analysis of the brain of PD patientsusing postmortem tissue, in addition to incidental cases that display α-Synuclein histopathology with no clinical manifestations, and com-pared them with healthy control subjects (Table 1). We analyzed sev-eral brain areas corresponding to the medulla oblongata, pons, mid-brain, hippocampus and neocortex. Formalin-fixed paraffin-embeddedtissue was stained for phosphorylated PERK (pPERK) and eIF2α(peIF2α), followed by quantification of the IHC score (see Methods). Aclear increase in the levels of phosphorylated PERK and eIF2α wasobserved in both PD and incidental cases in all brain areas tested(Fig. 1A, B). In addition, we confirmed the presence of α-Synucleininclusions in the samples analyzed (Fig. 1C). These results suggest thatthe activation of PERK signaling in PD cases may occur very earlyduring the disease process based on the positive signals observed inincidental cases.

We next monitored the levels of ER stress in different animal modelsof PD. Analysis of α-Synuclein transgenic mice expressing the A53Tmutation (Giasson et al., 2002) at the symptomatic (9–10months ofage) and pre-symptomatic (4 months of age) stages indicated a stronginduction of ATF4 using western blot analysis of brain cortex (Fig. 1D).We then analyzed a different model of PD in rats, based on the targetedoverexpression of α-Synuclein in the SNpc induced by the unilateralinjection of AAV coding for human α-Synuclein (Coune et al., 2011).

Three months after AAV delivery, the ventral midbrain, containing theSNpc, of control and injected sides were dissected and the levels ofATF4 target genes were determined using real time PCR. Increasedexpression of Atf3, Trb3 and Chop mRNA levels was observed in α-Sy-nuclein overexpressing midbrain when compared with the non-injectedcontrol sides (Fig. 1E).

We then moved forward and analyzed markers of ATF4 activation ina classical toxicological model of PD. Three months old mice were in-jected with 8 μg of the neurotoxin 6-hydroxydopamine (6-OHDA) intothe right striatum. Animals were euthanized at different time pointsafter injection and the levels of Atf3, Trb3 and Chop mRNA levelsmeasured by real time PCR. A progressive increase in the levels ofATF4-target genes was observed in midbrain tissue starting from 1weekafter the injection of 6-OHDA that was sustained for three weeks(Fig. 1F). Consistent with these results, increased levels of phosphory-lated PERK (Fig. 1G, Supplementary Fig. S1) and eIF2α (Fig. 1H, Sup-plementary Fig. S2) were observed in the SNpc of animals injected with6-OHDA as revealed using IHC. Overall, these experiments indicate thatactivation of PERK is a salient feature of PD.

3.2. Oral administration of GSK2606414 inhibits PERK signaling at theSNpc

To test the therapeutic potential of GSK2606414 in experimentalPD, we first evaluated the bioavailability and the activity of this smallmolecule in vivo in the target tissue. Mice were treated withGSK2606414 [100mg/kg/day] through oral administration for3 weeks. Then, 12 h after the last treatments animals were sacrificedand compound levels detected in serum and brain cortex (see methods).Using this regimen of administration, significant levels of GSK2606414were detected in both samples, reaching mean concentrations of4150 ng/mL and 5200 ng/g in serum and tissue, respectively (Fig. 2A).These data confirmed adequate exposure of GSK2606414 in the brain toengage and inhibit PERK, in line with previous studies (Moreno et al.,2013; Radford et al., 2015). We also measured the basal levels of eIF2αphosphorylation and observed a clear reduction in the striatum of micetreated with GSK2606414 using western blot analysis (Fig. 2B).

We then monitored the efficacy of GSK2606414 in inhibiting PERKsignaling at the SNpc. To this aim, we performed a pharmacologicalstimulation of the UPR by the intra-nigral injection of 10 μg of tunica-mycin in mice using brain stereotaxis. Animals were pre-treated during1 week with GSK2606414 [100mg/kg/day] and then exposed to tuni-camycin. Twenty-four hours later, ventral midbrain containing theSNpc was dissected and mRNA levels of Atf3, Trb3 and Chop mRNAwere measured by real time PCR. A significant inhibition in the upre-gulation of Trb3 was observed in GSK2606414 treated animals, whereasAtf3 and Chop only showed a trend of reduction (Fig. 2C). We validatedthese experiments also in rats by pre-treating animals for 3 days withvehicle or GSK2606414 [100mg/kg/day] followed by tunicamycininjection. A significant inhibition in the upregulation of Trb3 and Chopwas observed after induction of ER stress in this additional animalmodel (Fig. 2D). To monitor the specificity of the small molecule used,we then determined the levels of XBP1 mRNA splicing, a parallel UPRsignaling pathway activated by the stress sensor IRE1α (Walter andRon, 2011). No effects of GSK2606414 administration were observedon XBP1 mRNA splicing in the same samples (Fig. 3E, F). In addition, nochanges in the induction of ATF6 target gene by ER stress were ob-served when animals were treated with GSK2606414 (SupplementaryFig. S3A, B), indicating that GSK2606414 specifically blocked the ac-tivity of PERK without altering other UPR branches.

3.3. PERK inhibition protects dopaminergic neurons against experimentalPD

To explore the protective potential of GSK2606414 on dopami-nergic neuron survival, we performed unilateral stereotaxic injections

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

140

of 8 μg of 6-OHDA into the striatum of mice, followed by the oral ad-ministration of GSK2606414 [100mg/kg/day] or vehicle starting thesame day of the neurotoxin injections and continued for a total of threeweeks. We analyzed the neurodegenerative process at the SNpc bymonitoring nigral-dopaminergic neuron loss using tyrosine hydroxylase(TH) staining by IHC in midbrain tissue as we previously reported(Castillo et al., 2015; Valdes et al., 2014) (Fig. 3A). Importantly, thedecrease in TH-positive neurons observed in this toxicological model isdue to dopaminergic neuron loss and not to the downregulation in theexpression of the TH marker as monitored after NeuN or DAT staining(previously described in (Valdes et al., 2014)). We quantified TH-po-sitive cells in serial midbrain tissue sections every 100 μm covering theentire SNpc (Supplementary Fig. S4). Using this method, we observed a49% of neuronal loss in vehicle treated animals, whereas GSK2606414provided strong neuroprotection reflected on a 29% reduction in thecontent of TH-positive cells (Fig. 3B, left panel), indicating an efficacyof protection close to the 45% (Fig. 3B, right panel). We also detectedthe protection induced by GSK2606414 by measuring TH protein levelsin the SNpc using western blot analysis (Supplementary Fig. S5). Asexpected, a similar degree of striatal denervation induced by 6-OHDA(loss of TH-positive staining) was observed in vehicle and GSK2606414treated animals, indicating that the neurotoxin was equally effective ininducing damage at the striatum in both groups (Fig. 3C, Fig. 4H,Supplementary Fig. S6). Taken together, these results indicate that thepharmacological targeting of PERK protects dopaminergic neuronsagainst a parkinsonian-inducing neurotoxin.

3.4. GSK2606414 administration improves motor performance, andrecovers dopamine content and synaptic protein expression on a model of PD

To determine if the neuroprotective effect of GSK2606414 treat-ment translate into improved motor control, we performed severalbehavioral tests in animals injected with 6-OHDA, including cylindertest, hanging wire test and the rotarod. Mice were monitored in thecylinder test to measure forepaw akinesia overtime after 6-OHDA in-jection, and baseline recorded one week prior to the experiment.Forepaw akinesia was strongly induced by the 6-OHDA hemilaterallesion only in vehicle treated animals, while GSK2606414 treated ani-mals show a slight increase in forepaw akinesia that was not significantat any time point analyzed (Fig. 4A). Then, we performed the hanging

wire test, which measures coordination and muscle strength of animals.Remarkably, the drop observed in the performance of animals in thistest 14 days post 6-OHDA injection was completely inhibited in theGSK2606414 treated group (Fig. 4B, Supplementary Movies 1 and 2).Similarly, using rotarod assay, GSK2606414 administration reduced thedecay in performance observed 21 days after 6-OHDA injection(Fig. 4C). Taken together, these results indicate that the protection ofdopaminergic neurons observed at the histological level induced byGSK2606414 treatment translated into a significant attenuation in themotor and coordination defects.

To obtain insights about the physiological and functional state ofdopaminergic neurons after GSK2606414 administration, we monitoredthe levels of dopamine and its metabolite DOPAC using UltraPerformance Liquid Chromatography–Tandem Mass Spectrometry(UPLC-MSMS). 6-OHDA triggered a reduction near to 60% in the levelsof DOPAC and dopamine after 21 days of injection (Fig. 4D). Re-markably, a complete recovery in the levels of DOPAC and dopaminewere observed in in GSK2606414 treated animals (Fig. 4D).

One of the pathological effects of chronic PERK signaling is therepression in the synthesis of certain synaptic proteins (Smith andMallucci, 2016). Importantly, changes in synaptic protein content havebeen reported in PD models including the administration of 6-OHDA,associated with alterations in the expression of SNAP25 and VAMP2(Xiong et al., 2014). We measured protein levels of these two synapticmarkers in dissected striatum using western blot analys (Fig. 4E).Quantification of VAMP2 and SNAP25 protein levels confirmed a re-duction after 3 weeks of 6-OHDA injections, which was partially re-verted by GSK2606414 (Fig. 4F, G). As control, we determined TH le-vels in the striatum, which were comparable between groups (Fig. 4H)similar to the histological staining presented in Fig. 3C. Overall, thesefindings indicate that GSK2606414 administration provides globalneuroprotection at the level of neuronal survival, synaptic function anddopamine production, associated with improved motor performanceand coordination.

3.5. GSK2606414 induces slight body weight loss and pancreatic toxicity ina model of PD

Previous studies indicated adverse effects of GSK2606414 admin-istration due to pancreas toxicity (Atkins et al., 2013; Halliday et al.,

Table 1Summary of post mortem human tissue for control and Parkinson's disease patients.

Case Neuropathological diagnosis Gender Age PMI Cause of death Braak score for LB Braak score for NFT

1 PD M 82 19 h PD/pneumonia 6 42 PD M 77 1 day PD/pneumonia 6 33 PD M 77 2 days PD/pneumonia 6 14 PD M 74 1 day PD 6 15 PD M 84 1 day PD 6 26 PD M 57 2 days PD 6 17 PD M 68 1 day PD/septic shock 6 18 Incidental M 77 1 day Heart failure/ischaemia N.D. N.D.9 Incidental M 67 1 day Myocardial infarction N.D. N.D.10 Incidental M 73 1 day Heart failure N.D. N.D.11 Incidental F 62 1 day Myocardial infarction N.D. N.D.12 Incidental M 88 1 day Myocardial infarction N.D. N.D.13 Control F 61 6 h Euthanasia N.D. N.D.14 Control M 66 1 day Myocardial infarction N.D. N.D.15 Control F 73 1 day Hodgkin's disease N.D. N.D.16 Control F 32 1 day Lung carcinoma N.D. N.D.17 Control M 69 1 day Septic shock N.D. N.D.18 Control M 49 1 day Myocardial infarction N.D. N.D.19 Control F 63 1 day Pulmonal hypotensia N.D. N.D.20 Control F 68 1 day Lung fibrosis N.D. N.D.21 Control M 65 1 day Lung carcinoma N.D. N.D.22 Control F 66 1 day Hypotensia N.D. N.D.

Clinical data of cases investigated by histological analysis of postmortem brain tissue.(PMI, postmortem interval; LB, Lewy body; NFT, neurofibrillary tangles; N.D., none determined.)

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

141

2015; Moreno et al., 2013). However, animals were able to undergotreatments for months leading to important neuroprotective effects inmodels of PrDs and frontotemporal dementia (Moreno et al., 2013;Radford et al., 2015). In our experimental setting, oral administrationof GSK2606414 [100mg/kg/day] for a period of for 21 days after 6-OHDA injection led to a decrease in body weight of around 19%(Fig. 5A). This drop in body weight was accompanied by transient andsmall fluctuations in glucose levels that were not significant at 21 dayspost 6-OHDA injection (Fig. 5B). Histological analysis of the pancreasusing hematoxylin eosin staining indicated that administration of

GSK2606414 for 21 days led to signs of tissue damage with disin-tegrated pancreatic acini and lower content of islets when comparedwith vehicle-treated animals (Fig. 5C). These experiments indicate thatalthough GSK2606414 treatment has strong neuroprotective effects onour PD model, it results in slight undesired side effects possibly relatedto pancreatic dysfunction.

4. Discussion

Parkinson´s disease is an invalidating condition with no available

Fig. 1. Activation of PERK/ATF4 pathway in the brain of PD patients and animal models of the disease. Left panels: representative images from brain tissue from PD patients stained for(A) phosphorylated PERK (pPERK), (B) phosphorylated eIF2α (peIF2α) and (C) αSynuclein (αSyn). Right panels: immunoreactivity score (see Methods) was quantified in different brainregions including the medulla oblongata, pons, midbrain, hippocampus and neocortex of control cases (n=10), incidental cases (n= 5) and PD patients (n= 7). (D) Protein levels ofATF4, α-Synuclein (αSyn) and actin as loading control were measured in the brain cortex of mutant α-Synuclein transgenic (αSynA53T Tg) or non-transgenic (Non-Tg) littermate controlmice by Western blot analysis in pre-symptomatic (4 months) and symptomatic (9 and 10months) animals. Bottom panel: protein levels of ATF4 and αSyn quantified and normalized byactin levels in the brain cortex of Non-Tg and symptomatic (9 and 10months) αSynA53T Tg mice. (E) Adeno-associated vectors (AAV) coding for the human wild-type form of α-Synuclein(AAV αSyn) were injected unilaterally into the SNpc of rats and 3months later mRNA levels of Aft3, Trb3 and Chop relative to beta actin were determined by real time PCR (n= 4 pergroup). (F) Wild type mice were injected unilaterally in the striatum with the neurotoxin 6-OHDA and then euthanized at indicated time points after injections to measured mRNA levelsof Aft3, Trb3 and Chop relative to beta actin by real time PCR (n=3 for 24 h, 3 days and 1week groups; n=4 for the 2 and 3 weeks groups). Wild-type mice were injected unilaterally inthe striatum with the neurotoxin 6-OHDA and after 2 and 3weeks the levels of phosphorylated PERK (pPERK) (G) and phosphorylated eIF2α (peIF2α) (H) were determined using immunestaining midbrain tissue sections containing the SNpc. Representative images of the injected and non-injected (control) sides SNpc of animals 2 (upper panels) and 3 (bottom panels)weeks after injections are shown. Scale bar: 250 μm. For statistical analysis, mean and SEM are presented followed by Student's t-test. *: p < .05 and **: p < .01.

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

142

cure. Although several pathological mechanisms have been proposed toexplain the loss of dopaminergic neurons in PD, several studies indicatethat proteostasis disruption at the level of the ER is a salient andtransversal feature of the disease. Chronic ER stress may result in de-leterious events that promote dopaminergic neuron dysfunction anddeath. The current view of the involvement of the UPR in neurode-generative diseases is highly complex due to the divergent roles ofspecific components of UPR in distinct diseases (Hetz and Saxena, 2017;Hoozemans et al., 2007; Smith and Mallucci, 2016). Studies in mousemodels of PD suggested that activation of the ATF6 and IRE1α/XBP1signaling branches are neuroprotective, involving an increased buf-fering capacity, reducing ER stress levels (Credle et al., 2015; Egawaet al., 2011; Sado et al., 2009; Valdes et al., 2014). Moreover, underresting conditions these two signaling pathways have been proposed tobalance proteostasis to sustain dopaminergic neuron physiology (Egawaet al., 2011; Valdes et al., 2014). Although ATF4/CHOP expression hasa relevant role in inducing neuronal loss in toxin-based models of PD(Gully et al., 2016; Silva et al., 2005), the therapeutic potential of

targeting PERK signaling as an intervention strategy has not been ex-plored. Recent advances in the field led to the identification of severalsmall molecules to inhibit or activate specific UPR signaling branches.In the current study, we investigated the potential of inhibiting theenzymatic activity of PERK to improve the function and survival ofdopaminergic neurons. We validated the presence of ER stress and signsof PERK/ATF4 activation in human brain tissue derived from PD pa-tients, in addition to incidental cases of PD. Moreover, we confirmedthat the pathway is engaged in various rodent models of PD as an earlyresponse even before behavioral alterations are observed. Motivated bythese observations we then determined the therapeutic potential ofinhibiting PERK signaling with the small molecule GSK2606414 in ananimal model of PD.

The oral administration of GSK2606414 is effective in preventingneurodegeneration and synaptic dysfunction in models of PrPs andtauopathies (Moreno et al., 2013; Radford et al., 2015). GSK2606414administration also reduces tissue damage and cognitive impairment ina model of brain injury induced by hemorrhage (Yan et al., 2017).

Fig. 2. Effects of GSK2606414 administration on PERK signaling at the SNpc. (A) Wild-type mice were treated with GSK2606414 [100mg/kg/day] or vehicle for 3 weeks and 12 h afterthe last administration, serum and brain tissue were collected to determine GSK2606414 levels by HPLC-MS/MS (n=11 for GSK2606414). (B) Wild-type mice were treated withGSK2606414 [100mg/kg/day] or vehicle for 3 weeks and 12 h after the last administration, brain striatal tissue was dissected and phosphorylation of eIF2α determined using westernblot (n= 3 per group). (C) Mice or (D) rats were pre-treated with GSK2606414 [100mg/kg/day] or vehicle and then injected with tunicamycin (Tm) or its vehicle DMSO through brainstereotaxis directly into the SNpc (n= 3 per group). 24 h after Tm injections relative mRNA levels of Aft3, Atf4 and Chop were determined by real time PCR in dissected midbrain usingbeta actin as control. (E) Mice or rats (F) were pre-treated with GSK2606414 [100mg/kg/day] or vehicle and then injected with Tm or its vehicle DMSO (n=3 per group). 24 h after Tminjections XBP1 mRNA splicing levels were determined by semi quantitative RT-PCR in mRNA extracted from dissected midbrain. For statistical analysis, mean and SEM are presentedfollowed by Student's t-test. *: p < .05.

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

143

However, the ability of GSK2606414 to target the SNpc has not beentested before. Our results demonstrate that GSK2606414 is active in thebrain and specifically engages the target in a model of chronic ER stress.For proof-of-concept, we used a rapidly evolving model of PD in micebased on the injection of a neurotoxin that selectively damages dopa-minergic neurons. We found that the oral administration ofGSK2606414 augmented the survival of dopaminergic neurons at theSNpc, associated with increased dopamine levels and motor functions.At the molecular level, PERK inhibition may not only reduce deleteriousproapoptotic signals downstream of ATF4, but may also improve do-paminergic neuron function by restoring the levels of synaptic proteins.These effects contrast with the known role of eIF2α phosphorylation asa general stress mitigation output of ISR.

Our current study is the first demonstrating that pharmacologicalinhibition of PERK has neuroprotective potential at the SNpc. A pre-vious report used a generic approach to target the ISR in two models ofα-Synuclein-mediated using salubrinal (Colla et al., 2012a), a smallmolecule that inhibits the two eIF2α phosphatase complexes (Boyceet al., 2005). That study demonstrated neuroprotection at the beha-vioral level, improving motor performance, however, no protection bysalubrinal administration was observed at the level of nigral-dopami-nergic neurons (Colla et al., 2012a). Importantly, alternative substratesfor PERK kinase have been identified including NRF2 (Cullinan et al.,2003), which controls the antioxidant response (Cullinan and Diehl,2004) and is implicated in neurodegenerative diseases (Johnson andJohnson, 2015). Due to the side effects of GSK2606414 to pancreaticfunction, the use of this drug for the treatment of PD patients has lowtranslational potential (Mercado and Hetz, 2017). However, as proof-of-concept the use of the PERK inhibitor allowed us to propose that newsafe molecules to target the pathway may have the ability to attenuateneurodegeneration in PD. One recent study questioned the specificity ofGSK2606414 and suggested a higher affinity for RIP1 kinase (Rojas-Rivera et al., 2017), a central component of the necroptosis machinery.However, that study only used artificial in vitro systems to characterizethis activity and our data confirmed the direct effects of GSK2606414on PERK signaling. In addition, we did not detect signs of RIP1 or RIP3kinase activation in our experimental model (Supplementary Fig. S7).Another small molecule known as ISRIB inhibits the consequences ofeIF2α phosphorylation by binding to the eIF2B complex and preservingglobal translational capacity (Sekine et al., 2015; Sidrauski et al.,2015). ISRIB improves synaptic function at basal levels (Sidrauski et al.,2013), and provided neuroprotection in models of PrDs without signs ofpancreatic toxicity (Halliday et al., 2015). However, ISRIB has poorsolubility and lower relative activity compared with PERK in inhibitingprotein translation (Sekine et al., 2015; Sidrauski et al., 2015). Inter-estingly, a recent report screened for molecules that mimic the cellulareffects of ISRIB using a library of FDA-approved drugs (Halliday et al.,2017). This study identified dibenzoylmethane and trazodone as strongcandidates to target the ISR, having significant therapeutic effects inmodels of neurodegeneration (Halliday et al., 2017). We are currentlytesting the efficacy of these drugs in PD models.

Overall, it is predicted that the prolonged administration of com-pounds that target PERK will have deleterious effects to the function ofhighly secretory cells like pancreatic beta cells, exocrine pancreas andother tissues that require PERK signaling to sustain their normal func-tion (Dufey et al., 2014). Thus, in the scenario of treating a PD patient(10–15 years of treatment), inhibiting the PERK pathway using systemicadministration of small molecules carries a high risk of adverse sideeffects. In addition, studies using genetic ablation of PERK in the brainhave shown important defects in the basal activity of cortical neurons,leading to behavioral alterations (Costa-Mattioli et al., 2009; Ounallah-Saad et al., 2014; Trinh et al., 2012a; Trinh and Klann, 2013). Fur-thermore, reduction of eIF2α phosphorylation enhances long-termmemory formation (Costa-Mattioli et al., 2005; Costa-Mattioli et al.,2007; Ounallah-Saad et al., 2014; Zhu et al., 2011). The role of the UPRdownstream target ATF4 in neuronal physiology is controversial givenboth negative and positive results on behavioral tasks (Bartsch et al.,1995; Costa-Mattioli et al., 2005; Costa-Mattioli et al., 2007; Chenet al., 2003; Pasini et al., 2015; Trinh et al., 2012b). GSK2606414 wasshown to alter the basal function of certain neuronal circuits, where itcan enhance taste-dependent learning and plasticity (Ounallah-Saadet al., 2014), whereas it impairs working memory (Zhu et al., 2016).Thus, strategies for the local manipulation of PERK in the SNpc shouldbe considered in the future using gene therapy to downregulate PERKand thus avoid the systemic effects generated by PERK inhibition. Insupport of this idea, delivery of AAVs to express ATF4 into the SNpctriggers spontaneous dopaminergic neuron apoptosis (Gully et al.,2016). Our study suggests that drugs to alleviate ER stress levels mayhave important beneficial effects in PD. Two major molecular targets of

Fig. 3. GSK2606414 treatment protects dopaminergic neurons against 6-OHDA-inducedneurotoxicity. (A) Wild-type mice were hemilaterally injected with 6-OHDA into the rightstriatum, and from the same day were treated with vehicle or GSK2606414 [100mg/kg/day]. After 3 weeks, dopaminergic neurons were visualized in midbrain tissue sections byanti-TH immunostaining. Scale bar: 500 μm. (B) Anti-TH immunohistochemistry wasperformed in midbrain tissue sections every 100 μm and nigral dopaminergic neuronswere quantified and expressed as a percentage of neuron loss compared to the non-in-jected (control) side (left panel). In addition to calculate normalized neuronal damageusing vehicle treated animals as 100% (right panel). (C) Anti-TH immunohistochemistryanalysis was performed in striatal sections to quantify 6-OHDA-induced denervation inboth injected and non-injected sides. Scale bar: 1mm. (D) The integrated density of pixelintensity was calculated from images of anti-TH immunohistochemistry covering theentire striatum and expressed as the percentage of loss compared with the non-injected(control) sides. For quantifications, mean and SEM are presented (n= 12 for vehicle;n= 14 for GSK2606414). For statistical analysis, mean and SEM are presented followedby Student's t-test. **: p < .01.

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

144

α-Synuclein that perturbed ER function involve the direct inhibition ofER to Golgi trafficking (Cooper et al., 2006; Gitler et al., 2008) and alsoATF6 inhibition (Credle et al., 2015), resulting in deleterious levels ofstress. Gene therapy is emerging as a feasible strategy to specificallytarget the SNpc and deliver therapeutic genes (Castillo, Mercado, andHetz, 2015; Coune et al., 2012). We speculate that a more direct way toreduce ER stress and improve the functionality and survival of dopa-minergic neurons will be the delivery of active XBP1, ATF6, or ERchaperones to the SNpc. For example, the enforced expression of XBP1sinto the SNpc using recombinant viruses has been shown to protectdopaminergic neurons against PD-inducing neurotoxins (Sado et al.,2009; Valdes et al., 2014). Similarly, the local delivery of BiP into the

SNpc has a strong neuroprotective effect on a model of α-Synuclein-induced degeneration (Gorbatyuk et al., 2012).

The therapeutic potential of the UPR to treat PD may impact otheraspects of the disease that are beyond synaptic protein expression andER stress buffering. In this line, we have described new biologicalfunctions of XBP1, where it can enhance synaptic plasticity by con-trolling the expression of BDNF (Martinez et al., 2016), in addition toimproving axonal regeneration (Onate et al., 2016). AAV-XBP1s injec-tions can also improve motor function after axonal damage (Valenzuelaet al., 2012), in addition to reduce abnormal protein aggregation inmodels of Huntington's disease (Zuleta et al., 2012). Additional studieshave also demonstrated that strategies to alleviate ER stress using gene

Fig. 4. GSK2606414 treatment improves motor performance and dopaminergic neuron function. Animals were injected with 6-OHDA in the right striatum and from the same day treatedwith GSK2606414 [100mg/kg/day] or vehicle. (A) Spontaneous forepaw use during explorative behavior was assessed over time by the cylinder test. Touches on a cylinder wall werescored for each forepaws. Results were plotted as a percentage of touches with left forepaw (contralateral to injection side) in relation to the total of number touches with both forepaws.(B) Motor performance was quantified using a hanging test as indicated in material and methods to calculate a score of the performance. (C) Motor coordination was assessed by rotarodtest. The rotarod was set with a start speed of 4 rpm, acceleration rate 20 rpm/min, and the time of latency before fall was recorded. The mean on five trials for each animal is presented asa percentage from the base line performance. In all tests mean and SEM are presented (n=9 for vehicle; n=10 for GSK2606414). Statistical analysis was performed by two ways ANOVAfollowed by Bonferroni post-test: *: p < .05 and ***: p < .001. (D) Biogenic monoamines including DOPAC and dopamine were measured by UPLC-MSMS in the 6-OHDA injected andnon-injected (control) striatum of vehicle and GSK2606414 treated animals. Results are presented as percentage of the corresponding non-injected sides in means and SEM. Statisticalanalysis were performed by two ways ANOVA followed by Bonferroni post-test: **: p < .01 and ***: p < .001. (n= 4 per vehicle and n=7 per GSK2606414). (E) Levels of the synapticproteins VAMP2, SNAP25 and TH were determined by Western blot in striatal tissue dissected from vehicle and GSK2606414 [100mg/kg/day] treated animals from the 6-OHDA lesionedand the non-injected (control) sides. β-actin was measured as loading control. (F) Levels VAMP2 protein were quantified relative to β-actin (n= 4 per group). (G) Levels SNAP25 werequantified relative to β-actin (n= 4 per group). (H) Levels TH were quantified relative to β-actin (n= 4 per group). For statistical analysis, mean and SEM are presented followed byStudent's t-test. *: p < .05.

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

145

therapy are neuroprotective in various neurodegenerative diseases(Valenzuela et al., 2016a). GSK2606414 was also shown to attenuateER stress-induced brain inflammation, impacting cytokine productionand glial activation (Guthrie et al., 2016), suggesting additional bene-ficial impacts of targeting the UPR in brain diseases. More efforts areneeded to define the significance of the UPR as a target to treat PD. Insummary, our results demonstrate the significance of PERK signaling todopaminergic neuron function and survival on a model of PD. Thisstudy may also contribute to understanding the differential neuronal

vulnerability observed in PD patients.

5. Conclusions

In this report we provide evidence indicating that the pharmaco-logical inhibition of PERK using the oral delivery of GSK2606414 re-sults in strong neuroprotection in a model of PD. PERK inhibition alsoincreases dopamine levels and motor performance possibly associatedwith an augmented levels of key synaptic proteins. Additional studiesusing alternative approaches to locally target this pathway at the SNpcare needed to develop potential therapies for PD.

Supplementary data to this article can be found online at https://doi.org/10.1016/j.nbd.2018.01.004.

Conflict of interest

G.M., V.C., P.S., N.L., J.J.M.H and C.H. declare that they have noconflicts of interest. J.M.A. is an employee of GlaxoSmithKlineCompany. S.P.S. is an employee of Sanofi Company.

Author contributions

G.M. and C.H. designed the study. G.M. and V.C. participated inexperimental design, performed experiments and analyzed the data.P.S. and N.L. performed experiments. J.M.A. provided GSK2606414compound expertise and contributed to critical discussions of experi-mental design and data analyses. S.P.S. contribute with the AAV vectorsto over-express α-Synuclein. J.J.M.H. performed human tissue samplesstudies. G.M., V.C. and C.H. wrote or contributed to writing themanuscript. All authors read and approved the final version of themanuscript.

Acknowledgements

We thank Alexis Martinez for experimental support and JavieraPonce for technical support in animal care supervision. We also thanksto Dr. Patrick Aebischer and Dr. Bernard Schneider for supplying theplasmid encoding human WT α-Synuclein used in AAVs production.This work was directly funded by FONDECYT 11140738 (GM), MichaelJ Fox Foundation for Parkinson's Research – Target Validation grant9277 and 12473, Millennium Institute P09-015-F and FONDAP pro-gram 15150012 (CH). We also thank the support from ALS TherapyAlliance 2014-F-059, Muscular Dystrophy Association 382453,FONDEF ID16I10223, FONDEF D11E1007, FONDECYT 1140549,ALSRP Therapeutic Idea Award AL150111, US Office of NavalResearch-Global (ONR-G) N62909-16-1-2003, U.S. Air Force Office ofScientific Research FA9550-16-1-0384 and CONICYT-Brazil 441921/2016-7 (CH). V.C. was supported by a CONICYT Master's fellowship.

References

Atkins, C., et al., 2013. Characterization of a novel PERK kinase inhibitor with antitumorand antiangiogenic activity. Cancer Res. 73, 1993–2002.

Axten, J.M., et al., 2012. Discovery of 7-methyl-5-(1-{[3-(trifluoromethyl)phenyl]acetyl}-2,3-dihydro-1H-indol-5-yl)-7H-p yrrolo[2,3-d]pyrimidin-4-amine (GSK2606414), apotent and selective first-in-class inhibitor of protein kinase R (PKR)-like endoplasmicreticulum kinase (PERK). J. Med. Chem. 55, 7193–7207.

Baleriola, J., et al., 2014. Axonally synthesized ATF4 transmits a neurodegenerativesignal across brain regions. Cell 158, 1159–1172.

Bartsch, D., et al., 1995. Aplysia CREB2 represses long-term facilitation: relief of re-pression converts transient facilitation into long-term functional and structuralchange. Cell 83, 979–992.

Bellucci, A., et al., 2011. Induction of the unfolded protein response by alpha-synuclein inexperimental models of Parkinson's disease. J. Neurochem. 116, 588–605.

Bertram, L., Tanzi, R.E., 2005. The genetic epidemiology of neurodegenerative disease. J.Clin. Invest. 115, 1449–1457.

Boyce, M., et al., 2005. A selective inhibitor of eIF2alpha dephosphorylation protects cellsfrom ER stress. Science 307, 935–939.

Braak, H., Braak, E., 1991. Neuropathological stageing of Alzheimer-related changes.Acta Neuropathol. 82, 239–259.

Fig. 5. GSK2606414 treatment induces body weight loss and pancreatic dysfunction.Wild-type mice were injected with 8 μg of 6-OHDA in the right striatum and treated withvehicle or GSK2606414 [100mg/kg/day] for 21 days. (A) Body weight was measuredover time during treatment and presented as mean and SEM per group. Statistical analysiswas performed by two ways ANOVA followed by Bonferroni post-test: ***: p < .001. (B)Glucose in blood was determinate and expressed as [mg/dL]. For statistical analysis,mean and SEM are presented followed by Student's t-test. *: p < .05. (C) Upper panels:representative images of hematoxylin and eosin-stained pancreas sections of vehicle andGSK2606414 [100mg/kg/day] treated animals. Arrowheads indicate pancreatic islets.Scale bar: 100 μm. Lower panels: show magnifications of the insets indicated by thediscontinue lines rectangles.

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

146

Braak, H., et al., 2003. Staging of brain pathology related to sporadic Parkinson's disease.Neurobiol. Aging 24, 197–211.

Bruch, J., et al., 2017. PERK activation mitigates tau pathology in vitro and in vivo.EMBO Mol Med. 9, 371–384.

Buffington, S.A., Huang, W., Costa-Mattioli, M., 2014. Translational control in synapticplasticity and cognitive dysfunction. Annu. Rev. Neurosci. 37, 17–38.

Castillo, V., Mercado, G., Hetz, C., 2015. Gene therapy in Parkinson's disease: targetingthe endoplasmic reticulum proteostasis network. Neural Regen. Res. 10, 1053–1054.

Castillo, V., et al., 2015. Functional role of the disulfide isomerase ERp57 in axonal re-generation. PLoS One 10, e0136620.

Celardo, I., et al., 2016. Mitofusin-mediated ER stress triggers neurodegeneration inpink1/parkin models of Parkinson's disease. Cell Death Dis. 7, e2271.

Chen, A., et al., 2003. Inducible enhancement of memory storage and synaptic plasticityin transgenic mice expressing an inhibitor of ATF4 (CREB-2) and C/EBP proteins.Neuron 39, 655–669.

Chung, C.Y., et al., 2013. Identification and rescue of alpha-synuclein toxicity inParkinson patient-derived neurons. Science 342, 983–987.

Ciechanover, A., Kwon, Y.T., 2017. Protein quality control by molecular chaperones inneurodegeneration. Front. Neurosci. 11, 185.

Clayton, B.L., Popko, B., 2016. Endoplasmic reticulum stress and the unfolded proteinresponse in disorders of myelinating glia. Brain Res. 1648, 594–602.

Colla, E., et al., 2012a. Endoplasmic reticulum stress is important for the manifestationsof alpha-synucleinopathy in vivo. J. Neurosci. 32, 3306–3320.

Colla, E., et al., 2012b. Accumulation of toxic alpha-synuclein oligomer within en-doplasmic reticulum occurs in alpha-synucleinopathy in vivo. J. Neurosci. 32,3301–3305.

Conn, K.J., et al., 2004. Identification of the protein disulfide isomerase family memberPDIp in experimental Parkinson's disease and Lewy body pathology. Brain Res. 1022,164–172.

Cooper, A.A., et al., 2006. Alpha-synuclein blocks ER-Golgi traffic and Rab1 rescuesneuron loss in Parkinson's models. Science 313, 324–328.

Costa-Mattioli, M., et al., 2005. Translational control of hippocampal synaptic plasticityand memory by the eIF2alpha kinase GCN2. Nature 436, 1166–1173.

Costa-Mattioli, M., et al., 2007. eIF2alpha phosphorylation bidirectionally regulates theswitch from short- to long-term synaptic plasticity and memory. Cell 129, 195–206.

Costa-Mattioli, M., et al., 2009. Translational control of long-lasting synaptic plasticityand memory. Neuron 61, 10–26.

Coune, P.G., Schneider, B.L., Aebischer, P., 2012. Parkinson's disease: gene therapies.Cold Spring Harb. Perspect. Med. 2, a009431.

Coune, P.G., et al., 2011. Rab1A over-expression prevents Golgi apparatus fragmentationand partially corrects motor deficits in an alpha-synuclein based rat model ofParkinson's disease. J. Parkinsons Dis. 1, 373–387.

Credle, J.J., et al., 2015. Alpha-synuclein-mediated inhibition of ATF6 processing intoCOPII vesicles disrupts UPR signaling in Parkinson's disease. Neurobiol. Dis. 76,112–125.

Cullinan, S.B., Diehl, J.A., 2004. PERK-dependent activation of Nrf2 contributes to redoxhomeostasis and cell survival following endoplasmic reticulum stress. J. Biol. Chem.279, 20108–20117.

Cullinan, S.B., et al., 2003. Nrf2 is a direct PERK substrate and effector of PERK-depen-dent cell survival. Mol. Cell. Biol. 23, 7198–7209.

Damier, P., et al., 1999. The substantia nigra of the human brain. II. Patterns of loss ofdopamine-containing neurons in Parkinson's disease. Brain 122 (Pt 8), 1437–1448.

Das, I., et al., 2015. Preventing proteostasis diseases by selective inhibition of a phos-phatase regulatory subunit. Science 348, 239–242.

Dorsey, E.R., Bloem, B.R., 2017. The Parkinson pandemic—a call to action. JAMA Neurol.75 (1), 9–10.

Dufey, E., et al., 2014. Cellular mechanisms of endoplasmic reticulum stress signaling inhealth and disease. 1. An overview. Am. J. Phys. Cell Physiol. 307, C582–94.

Egawa, N., et al., 2011. The endoplasmic reticulum stress sensor, ATF6alpha, protectsagainst neurotoxin-induced dopaminergic neuronal death. J. Biol. Chem. 286,7947–7957.

Forman, M.S., Trojanowski, J.Q., Lee, V.M., 2004. Neurodegenerative diseases: a decadeof discoveries paves the way for therapeutic breakthroughs. Nat. Med. 10,1055–1063.

Giasson, B.I., et al., 2002. Neuronal alpha-synucleinopathy with severe movement dis-order in mice expressing A53T human alpha-synuclein. Neuron 34, 521–533.

Gitler, A.D., et al., 2008. The Parkinson's disease protein alpha-synuclein disrupts cellularRab homeostasis. Proc. Natl. Acad. Sci. U. S. A. 105, 145–150.

Gorbatyuk, M.S., et al., 2012. Glucose regulated protein 78 diminishes alpha-synucleinneurotoxicity in a rat model of Parkinson disease. Mol. Ther. 20, 1327–1337.

Gully, J.C., et al., 2016. Up-regulation of activating transcription factor 4 induces severeloss of dopamine nigral neurons in a rat model of Parkinson's disease. Neurosci. Lett.627, 36–41.

Guthrie, L.N., et al., 2016. Attenuation of PKR-like ER kinase (PERK) signaling selectivelycontrols endoplasmic reticulum stress-induced inflammation without compromisingimmunological responses. J. Biol. Chem. 291, 15830–15840.

Halliday, M., et al., 2015. Partial restoration of protein synthesis rates by the small mo-lecule ISRIB prevents neurodegeneration without pancreatic toxicity. Cell Death Dis.6, e1672.

Halliday, M., et al., 2017. Repurposed drugs targeting eIF2α-P-mediated translationalrepression prevent neurodegeneration in mice. Brain 140, 1768–1783.

Han, J., et al., 2013. ER-stress-induced transcriptional regulation increases proteinsynthesis leading to cell death. Nat. Cell Biol. 15, 481–490.

Harding, H.P., Zhang, Y., Ron, D., 1999. Protein translation and folding are coupled by anendoplasmic-reticulum-resident kinase. Nature 397, 271–274.

Harding, H.P., et al., 2000. Regulated translation initiation controls stress-induced gene

expression in mammalian cells. Mol. Cell 6, 1099–1108.Harding, H.P., et al., 2003. An integrated stress response regulates amino acid metabolism

and resistance to oxidative stress. Mol. Cell 11, 619–633.Heman-Ackah, S.M., et al., 2017. Alpha-synuclein induces the unfolded protein response

in Parkinson's disease SNCA triplication iPSC-derived neurons. Hum. Mol. Genet. 26,4441–4450.

Hetz, C., 2012. The unfolded protein response: controlling cell fate decisions under ERstress and beyond. Nat. Rev. Mol. Cell Biol. 13, 89–102.

Hetz, C., Papa, F.R., 2017. The unfolded protein response and cell fate control. Mol. Cell69 (2), 169–181.

Hetz, C., Saxena, S., 2017. ER stress and the unfolded protein response in neurodegen-eration. Nat. Rev. Neurol. 13, 477–491.

Hoozemans, J.J., et al., 2007. Activation of the unfolded protein response in Parkinson'sdisease. Biochem. Biophys. Res. Commun. 354, 707–711.

Johnson, D.A., Johnson, J.A., 2015. Nrf2—a therapeutic target for the treatment ofneurodegenerative diseases. Free Radic. Biol. Med. 88, 253–267.

Kaushik, S., Cuervo, A.M., 2015. Proteostasis and aging. Nat. Med. 21, 1406–1415.Ma, Y., et al., 2002. Two distinct stress signaling pathways converge upon the CHOP

promoter during the mammalian unfolded protein response. J. Mol. Biol. 318,1351–1365.

Ma, T., et al., 2013. Suppression of eIF2alpha kinases alleviates Alzheimer's disease-re-lated plasticity and memory deficits. Nat. Neurosci. 16, 1299–1305.

Marciniak, S.J., et al., 2004. CHOP induces death by promoting protein synthesis andoxidation in the stressed endoplasmic reticulum. Genes Dev. 18, 3066–3077.

Martinez, G., et al., 2016. Regulation of memory formation by the transcription factorXBP1. Cell Rep. 14, 1382–1394.

Matus, S., et al., 2013. Functional contribution of the transcription factor ATF4 to thepathogenesis of amyotrophic lateral sclerosis. PLoS One 8, e66672.

Mercado, G., Hetz, C., 2017. Drug repurposing to target proteostasis and prevent neu-rodegeneration: accelerating translational efforts. Brain 140, 1544–1547.

Mercado, G., et al., 2016. ER stress and Parkinson's disease: pathological inputs thatconverge into the secretory pathway. Brain Res. 1648 (Pt B), 626–632.

Michel, P.P., Hirsch, E.C., Hunot, S., 2016. Understanding dopaminergic cell deathpathways in Parkinson disease. Neuron 90, 675–691.

Moreno, J.A., et al., 2012. Sustained translational repression by eIF2alpha-P mediatesprion neurodegeneration. Nature 485, 507–511.

Moreno, J.A., et al., 2013. Oral treatment targeting the unfolded protein response pre-vents neurodegeneration and clinical disease in prion-infected mice. Sci. Transl. Med.5, 206ra138.

Ohoka, N., et al., 2005. TRB3, a novel ER stress-inducible gene, is induced via ATF4-CHOP pathway and is involved in cell death. EMBO J. 24, 1243–1255.

Onate, M., et al., 2016. Activation of the unfolded protein response promotes axonalregeneration after peripheral nerve injury. Sci. Rep. 6 (21709).

Ounallah-Saad, H., et al., 2014. Genetic or pharmacological reduction of PERK enhancescortical-dependent taste learning. J. Neurosci. 34, 14624–14632.

Pakos-Zebrucka, K., et al., 2016. The integrated stress response. EMBO Rep. 17,1374–1395.

Pasini, S., et al., 2015. Specific downregulation of hippocampal ATF4 reveals a necessaryrole in synaptic plasticity and memory. Cell Rep. 11, 183–191.

Pennuto, M., et al., 2008. Ablation of the UPR-mediator CHOP restores motor functionand reduces demyelination in Charcot-Marie-Tooth 1B mice. Neuron 57, 393–405.

Postuma, R.B., et al., 2015. MDS clinical diagnostic criteria for Parkinson's disease. Mov.Disord. 30, 1591–1601.

Radford, H., et al., 2015. PERK inhibition prevents tau-mediated neurodegeneration in amouse model of frontotemporal dementia. Acta Neuropathol. 130, 633–642.

Rodriguez, D.A., et al., 2012. BH3-only proteins are part of a regulatory network thatcontrol the sustained signalling of the unfolded protein response sensor IRE1alpha.EMBO J. 31, 2322–2335.

Rojas-Rivera, D., et al., 2017. When PERK inhibitors turn out to be new potent RIPK1inhibitors: critical issues on the specificity and use of GSK2606414 and GSK2656157.Cell Death Differ. 24, 1100–1110.

Sado, M., et al., 2009. Protective effect against Parkinson's disease-related insults throughthe activation of XBP1. Brain Res. 1257, 16–24.

Scheper, W., Hoozemans, J.J., 2015. The unfolded protein response in neurodegenerativediseases: a neuropathological perspective. Acta Neuropathol. 130, 315–331.

Sekine, Y., et al., 2015. Stress responses. Mutations in a translation initiation factoridentify the target of a memory-enhancing compound. Science 348, 1027–1030.

Selvaraj, S., et al., 2012. Neurotoxin-induced ER stress in mouse dopaminergic neuronsinvolves downregulation of TRPC1 and inhibition of AKT/mTOR signaling. J. Clin.Invest. 122, 1354–1367.

Sidoli, M., et al., 2016. Ablation of Perk in Schwann cells improves myelination in theS63del Charcot-Marie-Tooth 1B mouse. J. Neurosci. 36, 11350–11361.

Sidrauski, C., et al., 2013. Pharmacological brake-release of mRNA translation enhancescognitive memory. elife 2, e00498.

Sidrauski, C., et al., 2015. Pharmacological dimerization and activation of the exchangefactor eIF2B antagonizes the integrated stress response. elife 4, e07314.

Silva, R.M., et al., 2005. CHOP/GADD153 is a mediator of apoptotic death in substantianigra dopamine neurons in an in vivo neurotoxin model of parkinsonism. J.Neurochem. 95, 974–986.

Slodzinski, H., et al., 2009. Homocysteine-induced endoplasmic reticulum protein (herp)is up-regulated in parkinsonian substantia nigra and present in the core of Lewybodies. Clin. Neuropathol. 28, 333–343.

Smith, H.L., Mallucci, G.R., 2016. The unfolded protein response: mechanisms andtherapy of neurodegeneration. Brain 139, 2113–2121.

Soto, C., 2003. Unfolding the role of protein misfolding in neurodegenerative diseases.Nat. Rev. Neurosci. 4, 49–60.

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

147

Spillantini, M.G., et al., 1997. Alpha-synuclein in Lewy bodies. Nature 388, 839–840.Su, L.J., et al., 2010. Compounds from an unbiased chemical screen reverse both ER-to-

Golgi trafficking defects and mitochondrial dysfunction in Parkinson's diseasemodels. Dis. Model. Mech. 3, 194–208.

Tabas, I., Ron, D., 2011. Integrating the mechanisms of apoptosis induced by endoplasmicreticulum stress. Nat. Cell Biol. 13, 184–190.

Thayanidhi, N., et al., 2010. Alpha-synuclein delays endoplasmic reticulum (ER)-to-Golgitransport in mammalian cells by antagonizing ER/Golgi SNAREs. Mol. Biol. Cell 21,1850–1863.

Toda, H., et al., 2006. Behavioral stress and activated serotonergic neurotransmissioninduce XBP-1 splicing in the rat brain. Brain Res. 1112, 26–32.

Trinh, M.A., Klann, E., 2013. Translational control by eIF2alpha kinases in long-lastingsynaptic plasticity and long-term memory. Neurobiol. Learn. Mem. 105, 93–99.

Trinh, M.A., et al., 2012a. Brain-specific disruption of the eIF2alpha kinase PERK de-creases ATF4 expression and impairs behavioral flexibility. Cell Rep. 1, 676–688.

Trinh, M.A., et al., 2012b. Brain-specific disruption of the eIF2alpha kinase PERK de-creases ATF4 expression and impairs behavioral flexibility. Cell Rep. 1, 676–688.

Tsaytler, P., Bertolotti, A., 2013. Exploiting the selectivity of protein phosphatase 1 forpharmacological intervention. FEBS J. 280, 766–770.

Tsujii, S., Ishisaka, M., Hara, H., 2015. Modulation of endoplasmic reticulum stress inParkinson's disease. Eur. J. Pharmacol. 765, 154–156.

Urra, H., et al., 2013. When ER stress reaches a dead end. Biochim. Biophys. Acta 1833,3507–3517.

Valdes, P., et al., 2014. Control of dopaminergic neuron survival by the unfolded proteinresponse transcription factor XBP1. Proc. Natl. Acad. Sci. U. S. A. 111, 6804–6809.

Valenzuela, V., et al., 2012. Activation of the unfolded protein response enhances motor

recovery after spinal cord injury. Cell Death Dis. 3, e272.Valenzuela, V., et al., 2016a. Gene therapy to target ER stress in brain diseases. Brain Res.

1648 (Pt B), 561–570.Valenzuela, V., et al., 2016b. Injury to the nervous system: a look into the ER. Brain Res.

1648, 617–625.Vila, M., Przedborski, S., 2004. Genetic clues to the pathogenesis of Parkinson's disease.

Nat. Med. 10 (Suppl), S58–62.Walter, P., Ron, D., 2011. The unfolded protein response: from stress pathway to

homeostatic regulation. Science 334, 1081–1086.Woehlbier, U., et al., 2016. ALS-linked protein disulfide isomerase variants cause motor

dysfunction. EMBO J. 35 (8), 845–865.Xiong, Y., et al., 2014. Differential expression of synaptic proteins in unilateral 6-OHDA

lesioned rat model—a comparative proteomics approach. Proteomics 14, 1808–1819.Yan, F., et al., 2017. Pharmacological inhibition of PERK attenuates early brain injury

after subarachnoid hemorrhage in rats through the activation of Akt. Mol. Neurobiol.54, 1808–1817.

Zhu, P.J., et al., 2011. Suppression of PKR promotes network excitability and enhancedcognition by interferon-gamma-mediated disinhibition. Cell 147, 1384–1396.

Zhu, S., et al., 2016. PERK regulates working memory and protein synthesis-dependentmemory flexibility. PLoS One 11, e0162766.

Zinszner, H., et al., 1998. CHOP is implicated in programmed cell death in response toimpaired function of the endoplasmic reticulum. Genes Dev. 12, 982–995.

Zuleta, A., et al., 2012. AAV-mediated delivery of the transcription factor XBP1s into thestriatum reduces mutant huntingtin aggregation in a mouse model of Huntington'sdisease. Biochem. Biophys. Res. Commun. 420, 558–563.

G. Mercado et al. Neurobiology of Disease 112 (2018) 136–148

148

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具