Ser46 phosphorylation and prolyl-isomerase Pin1-mediated isomerization of p53 are key events in p53-dependent apoptosis induced by mutant huntingtinAlice Grisona,1,2, Fiamma Mantovania,b,1, Anna Comela,b, Elena Agostonic, Stefano Gustincichc,d,e,Francesca Persichettic,f, and Giannino Del Sala,b,3

aLaboratorio Nazionale Consorzio Interuniversitario per le Biotecnologie, 34149 Trieste, Italy; bDepartment of Life Sciences, University of Trieste, 34100Trieste, Italy; cSector of Neurobiology, International School for Advanced Studies, 34136 Trieste, Italy; dInternational School for Advanced Studies Unit, ItalianInstitute of Technology, 34136 Trieste, Italy; eGiovanni Armenise–Harvard Foundation Laboratory, International School for Advanced Studies, 34136 Trieste,Italy; and fDepartment of Environmental and Life Sciences, University of Eastern Piedmont, 15121 Alessandria, Italy

Edited* by Carol Prives, Columbia University, New York, NY, and approved September 13, 2011 (received for review April 18, 2011)

Huntington disease (HD) is a neurodegenerative disorder caused bya CAG repeat expansion in the gene coding for huntingtin protein.Several mechanisms have been proposed bywhichmutant hunting-tin (mHtt) may trigger striatal neurodegeneration, including mito-chondrial dysfunction, oxidative stress, and apoptosis. Furthermore,mHtt induces DNA damage and activates a stress response. In thiscontext, p53 plays a crucial role inmediatingmHtt toxic effects. Herewe have dissected the pathway of p53 activation by mHtt in humanneuronal cells and in HD mice, with the aim of highlighting criticalnodes that may be pharmacologically manipulated for therapeuticintervention. We demonstrate that expression of mHtt causes in-creased phosphorylation of p53 on Ser46, leading to its interactionwith phosphorylation-dependent prolyl isomerase Pin1 and conse-quent dissociation from the apoptosis inhibitor iASPP, therebyinducing the expression of apoptotic target genes. Inhibition ofSer46 phosphorylation by targeting homeodomain-interacting pro-tein kinase 2 (HIPK2), PKCδ, or ataxia telangiectasia mutated kinase,as well as inhibition of the prolyl isomerase Pin1, prevents mHtt-dependent apoptosis of neuronal cells. These results provide a ratio-nale for the use of small-molecule inhibitors of stress-responsiveprotein kinases and Pin1 as a potential therapeutic strategy for HDtreatment.

Huntington disease is a dominantly inherited neurodegenera-tive disorder due to an expanded CAG repeat sequence in the

HD gene that elongates a segment of glutamine residues in theprotein huntingtin (Htt) (1). The most striking pathological man-ifestation of HD is a gradual loss of neurons, predominantly in thestriatum, causing motor abnormalities and cognitive decline (2).Toxic properties ofmutant huntingtin (mHtt) are believed to causeHD. Among them, mitochondrial dysfunction and generation ofreactive oxygen species (ROS) lead to DNA lesions (3, 4). In-terestingly, expression of full-length mHtt protein and N-terminalfragments containing the polyQ expansion elicit a DNA damageresponse, with activation of the ATM/ATR pathways (3, 4) andtheir downstream effectors, including the tumor suppressor p53 (3,5, 6). p53 mediates mitochondrial dysfunction and cytotoxicity inHD cells and in transgenic animal models, whereas its inhibitionprevents these phenotypes (5).p53 governs a wide array of pathways involved in genomic sta-

bility, antioxidant activities, and energy metabolism in addition topromoting either cytostatic or cytotoxic responses to intrinsic andexogenous sources of cellular stress (7). Given the complexity ofp53 functions within the cell, a better understanding of how sig-naling networks converge on this hub to modulate mHtt-de-pendent toxicity is required to shed light on the reduced ability ofthe brain neurons to survive. Regulation of p53 activities relies ona complex network of posttranslational modifications and proteininteractions (8, 9), which ultimately determine the outcome of thep53 response. This process entails site-specific phosphorylation byseveral DNA damage-activated protein kinases, including, amongothers, ataxia telangiectasia mutated (ATM), ATM and Rad3-re-lated (ATR), homeodomain-interacting protein kinase 2 (HIPK2),

and PKCδ. The subsequent transduction of stress-dependentphosphorylation into specific conformational changes of p53 thatfully unleash its apoptotic activity is performed by the prolylisomerase Pin1. This enzyme catalyzes cis/trans isomerization ofproline bonds preceded by phosphorylated serine or threonineresidues (pSer/Thr-Pro), thereby altering structure and functionsof its substrates (10, 11). Upon genotoxic insults, Pin1 binds mul-tiple sites on p53, promoting its accumulation in stressed cells, theactivation of its transcriptional functions, and the induction of itsapoptotic activity (12–14).Phosphorylation-dependent prolyl isomerization triggered by

Pin1 represents an essential mechanism in modulating severalsignaling pathways involved in DNA damage and apopotosis. Inthe CNS, Pin1 is highly expressed and regulates several substrates,including the hyperphosphorylated form of tau in Alzheimer’sdisease (15), whereas in Parkinson disease (PD) Pin1 facilitatesformation of α-synuclein inclusions (16).Based on these considerations, we reasoned that Pin1 could be

critical for mediating p53-dependent mHtt toxicity. Therefore,by investigating its role in this context, we may highlight crucialupstream events involved in HD pathogenesis and unveil at-tractive targets for development of novel therapeutic options tocounteract HD.

ResultsExpression of Mutant Huntingtin Promotes the Interaction of p53with Pin1. Analysis of postmortem brains of HD patients revealedhigh levels of p53 relative to healthy controls (Fig. 1A), in agree-ment with previous reports (5). To study the stress pathways re-sponsible for p53 activation in HD neurons, we then analyzed p53phosphorylation. Interestingly, in HD brains, p53 was phosphory-lated on Ser46 (Fig.1A), a modification that has been associatedwith activation of its apoptotic function upon stress (14, 17). Nu-clear accumulation of mHtt N-terminal fragments is observed inHD brains (18) and animal models (18). Expression of thesetruncated forms recapitulates many molecular and neurologicalHDphenotypes (19). TheN-terminal fragment (residues 1–171) ofeither wild-type or mutant Htt (bearing 21 and 150 polyQ, re-spectively) were thus expressed in SH-SY5Y human neuroblas-

Author contributions: F.M. and G.D.S. designed research; A.G., F.M., A.C., and E.A. per-formed research; E.A., S.G., and F.P. contributed new reagents/analytic tools; F.M. andG.D.S. analyzed data; and F.M., F.P., and G.D.S. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.

Freely available online through the PNAS open access option.1A.G and F.M. contributed equally to this work.2Present address: Sector of Neurobiology, International School for Advanced Studies,34136 Trieste, Italy.

3To whom correspondence should be addressed. E-mail: delsal@lncib.it.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1106198108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1106198108 PNAS | November 1, 2011 | vol. 108 | no. 44 | 17979–17984

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

15, 2

020

toma cells to verify whether p53 Ser46 phosphorylation was aconsequence of mHtt expression. Interestingly, mutant but notwild-type Htt induced the phosphorylation of endogenous p53 onSer46, in addition to the previously reported phosphorylation ofSer15 (6) (Fig. 1B). Because Ser46 phosphorylation generatesa target site for the prolyl isomerasePin1 (12, 14),weaskedwhetherPin1 might play a role in mediating activation of p53 upon mHttexpression in neuronal cells. Strikingly, expression of the mHtt-150Q fragment was sufficient to promote the interaction of en-dogenous p53 and Pin1 proteins in SH-SY5Y cells (Fig. 1C).Moreover, mHtt expression stimulated direct interaction of p53withPin1 as demonstrated byGST-Pin1 pull-down assays (Fig. 1D),and this effect was proportional to the amount of mHtt. Of note,Pin1 neither interacted withHtt (1–171) protein fragments (Fig. S1A and B) nor affected mHtt protein levels (Fig. S1 C and D).

Pin1 Mediates Activation of the p53 Pathway by mHtt. It has beenpreviously shown that expression of mHtt in SH-SY5Y cells trig-gers a p53-dependent response involving the activation of apo-ptotic genes, including Bax and PUMA (5). As shown in Fig. 2A,Pin1 potentiated the induction of p53 transcriptional activity bymHtt. Moreover, induction of endogenous PUMA in response tomHtt expression in these cells also required Pin1 (Fig. 2B).In agreement with previous reports (5), expression of mHtt in

SH-SY5Y cells provoked apoptosis, which was reduced by 50%upon silencing p53 expression (Fig. 2C). The same effect was ob-served upon silencing the expression of Pin1. Importantly, mHtt-induced apoptosis could be reestablished in Pin1-depleted cells byoverexpression of a siRNA-resistant Pin1 construct; this was noteffective in cells depleted of p53 (Fig. 2C), suggesting that the effectof Pin1 relies on p53. It is noteworthy that expression of a catalyt-ically inactive Pin1 mutant was unable to rescue knockdown ofendogenous Pin1, proving that the prolyl isomerase activity is es-sential for transducing mHtt-dependent stress into p53 activation.Similar results were also observed in another neuroblastoma cellline, SK-N-SH (Fig. S2). These data indicate that the Pin1/p53pathway plays a major role in neuronal apoptosis induced by mHtt.

We then analyzed activation of the p53 response in the brains ofHdhCAGknock-inmice inwhich the glutamine tract ofmouseHttis extended to 111 residues (HdhQ111) (20). Thesemice show striatalneurodegeneration, reactive gliosis, and gait abnormalities at olderage (after 24mo) (21). However, we observed stabilization of p53 inbrain extracts and the consequent transcriptional induction of thep53 target genep21WAF1 in the striatumof 12-mo-oldHdhQ111micecompared with theirWT littermates, HdhQ7 (Fig. 2D). This findingsuggests that activation of the p53 pathway by mHtt-associatedstress is an early event in HD pathogenesis and could precede neu-rological symptoms. HdhQ111 mice were then crossed with Pin1KO

mice (22) to verify whether Pin1 is required for p53 activation. Theinteraction between p53 and Pin1 was clearly detectable in proteinextracts obtained from HdhQ111/Pin1WT mouse brains (Fig. S3A),confirming what was observed in human cells. Importantly, incontrast to HdhQ111/Pin1WT mice, p53 transcriptional activity wasnot induced inHdhQ111/Pin1KOmice (Fig. 2D), and the expressionof the apoptotic p53 target PUMA in HdhQ111 mouse brains wasindeed dependent on Pin1 (Fig. 2E). We therefore analyzedstriatal neurodegeneration in 24-mo-old mice (Fig. S3 B and D).In agreement with published data (21), moderate neuronal losswas observed in HdhQ111 mice compared withWTHtt littermateson a Pin1 WT background (Fig. S3D). Pin1 KO mice also showeda similar reduction of striatal neurons number compared withWTlittermates, which could possibly be ascribed to the reported age-dependent neurodegeneration of mice lacking Pin1 (23). Impor-tantly, in Pin1 KO mice the numbers of striatal neurons did notfurther decrease on mHtt expression, suggesting that lack of p53activation (Fig. 2 D and E) might indeed prevent Htt-dependentneurodegeneration in mice devoid of Pin1.These results indicate that Pin1 plays a critical role for p53

activation in response to mHtt expression in striatal neurons ofa mouse model of HD pathogenesis.

Phosphorylation of p53 on Ser46 by HIPK2 and PKCδ Is an UpstreamEvent in the mHtt-Pin1-p53 Pathway. Our observations indicatedthat phosphorylation of p53 on Ser46 is triggered by mHtt

A B C

D

Fig. 1. Expression of mutant huntingtin induces the interaction of p53 with Pin1. (A) Phosphorylation of p53 on Ser46 in postmortem brains was comparedbetween HD patients (HD1 and HD2) and healthy controls (C) by Western blot. (Middle and Bottom) Levels of total p53 and actin as loading control. (B) SH-SY5Y cells were transfected with constructs expressing the N-terminal 1–171 Htt fragment with either 21Q or 150Q. p53 was immunoprecipitated from equalamounts of total cell lysates and analyzed by Western blot with antibodies specific for phosphorylated Ser46 and total p53. The levels of actin and Httproteins in input lysates are shown. (C) SH-SY5Y were transfected with indicated constructs. Cell lysates normalized for p53 protein levels were subjected tocoimmunoprecipitation to analyze interaction of endogenous p53 and Pin1. (D) H1299 cells were transfected with indicated constructs, and the interaction ofp53 with recombinant Pin1 protein was analyzed by GST pull-down of cell lysates normalized for p53 levels.

17980 | www.pnas.org/cgi/doi/10.1073/pnas.1106198108 Grison et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

15, 2

020

(Fig. 1 A and B). To define whether this or other phosphorylationsare responsible for activating p53 apoptotic response upon ex-pression ofmHtt, we used p53phosphorylationmutants with singleandmultiple substitutions of Ser/Thr with Ala residues within Pin1binding sites (14) (Fig. 3A). Expression of these proteins in p53-null H1299 cells demonstrated that Ser46 is required for mHtt-induced apoptosis, because a Ser46-Ala p53 mutant was unable tocause apoptosis in response to mHtt expression (Fig. 3B). Incontrast, a p53 mutant (p53 3M-S46wt) that lacked the remainingthree major Pin1 binding sites (i.e., Ser33, Thr81, and Ser315)could efficiently induce apoptosis downstream to mHtt expres-sion. This protein was phosphorylated on Ser46 in cells expressingmHtt (Fig. S4A), and its apoptotic activity was potentiated by Pin1

(Fig. 3B). Both p53 WT and p53 3M-S46WT were then able toinduce the proapoptotic p53 target PUMA upon transfection ofmHtt, whereas p53 S46A was almost inactive (Fig. 3B). Of note,similar experiments performed in mouse neuroblastoma cellsdemonstrated that phosphorylation of Ser58, the mouse homologof p53 Ser46, is essential for mHtt-induced apoptosis (Fig. S4B).Therefore, we concluded that phosphorylation of p53 on Ser46

is a crucial event in the pathway leading to neuronal death inducedby mHtt in human and mouse cells. Our data also indicate thatmodification at this site is sufficient for Pin1 to enhance p53’sapoptotic function in cells expressing mHtt.Among the protein kinases that catalyze phosphorylation of p53

on Ser46, HIPK2 plays a key role in unleashing p53’s apoptoticactivity upon DNA damage (24). HIPK2 is induced by cytotoxicstimuli through the ATM/ATR pathway (25), which becomesactivated upon mHtt-dependent stress (3). Interestingly, expres-sion of mHtt was sufficient to up-regulate HIPK2 protein levelsin SH-SY5Y cells, and this effect required ATM kinase activitybecause it was prevented by treatment with the ATM-specificinhibitor KU55933 (Fig. 3C).We thus inhibitedHIPK2 expression byRNAi, which dampened

mHtt-dependent phosphorylation of p53 on Ser46 with concomi-tant decrease of apoptosis (Fig. 3D). This result indicates a majorrole for HIPK2 in the activation of p53 by mHtt. Interestingly,depletion of HIPK2 did not fully prevent Ser46 phosphorylationtriggered by mHtt (Fig. 3D), implying the involvement of otherkinases in inducing apoptosis downstream of mHtt. We postulatedthat PKCδ might concur to this effect, because its activationhas been reported to regulate p53 during neuronal death (26).Knockdown of PKCδ by RNAi strongly decreased phospho-Ser46(Fig. 3D), consistent with the notion that this enzyme phosphor-ylates Ser46 upon DNA damage (27). Moreover, mHtt-dependentapoptosis was also reduced. Importantly, concomitant knockdownof both HIPK2 and PKCδ synergized to inhibit both Ser46 phos-phorylation and mHtt-induced apoptosis (Fig. 3D). The reductionof apoptosis exceeded that obtained by silencing p53, suggestingthat HIPK2 and PKCδ might also regulate p53-independent apo-ptotic pathways downstream to mHtt.

Pin1 Unlocks p53 from iASPP in Response to mHtt Expression. Wehave previously shown that Pin1-mediated isomerization of phos-phorylated Ser46-Pro47 site unleashes p53’s apoptotic potentialupon DNA damage, by leading to its dissociation from the apo-ptosis inhibitor iASPP (14). Because mHtt triggers both DNAdamage (3) and Ser46 phosphorylation, we investigated whether itcan induce p53 dissociation from iASPP. Coimmunoprecipitationexperiments highlighted that the interaction between p53 andiASPP was progressively lost upon expression of increasingamounts of mHtt in H1299 cells (Fig. 4A). This effect correlatedwith and required Ser46 phosphorylation (Fig. 4B): indeed, WTHtt caused neither p53 Ser46 phosphorylation (Fig. 4B) nor iASPPdetachment (Fig. 4C), and S46A mutation impaired the dissocia-tion of p53 from iASPP on mHtt expression (Fig. S4C). Impor-tantly, when Pin1 expression was silenced, p53 remained bound toiASPP regardless of mHtt expression (Fig. 4D). These data in-dicate that blocking Pin1 prevents p53 from inducing proapoptoticeffectors due to the sustained inhibition by iASPP even in presenceof mHtt-dependent stress.

Interfering with p53 Activation by Pin1 Prevents ApoptosisDownstream of mHtt. Given that p53 activation mediates the cy-totoxic effects of mHtt, we hypothesized that pharmacologic in-hibition of catalytic activity of either Pin1 or of the kinases thatphosphorylate p53 on the Pin1 target site Ser46 could prove ef-fective in preventing mHtt-induced apoptosis. In fact, treatmentwith the specific Pin1 inhibitor PiB (28) reduced mHtt-dependentapoptosis (Fig. 5A) with efficiency similar to knockdown of Pin1 orp53 (Fig. 2C). Intriguingly, inhibition of Pin1 also appeared toreduce Ser46 phosphorylation. Treatment of SH-SY5Y cells withrottlerin, a compound widely used to inhibit PKCδ activity, alsoreduced mHtt-dependent cellular toxicity and dampened Ser46

A B

C

D

E

Fig. 2. Pin1 activity is required for induction of p53-dependent apoptosis bymHtt. (A) The cooperative effect of mHtt and Pin1 on p53 transcriptionalactivity was evaluated by transfecting SH-SY5Y cells with pG13-Luc reporterand vectors expressing mHtt(1–171)150Q and increasing amounts (+, ++) ofPin1-HA. The graph shows means and SD of three independent experiments.(B) SH-SY5Y cells were transfected with indicated constructs and siRNA oli-gonucleotides, and the expression of the p53 target PUMA was evaluated byWestern blot after 48 h. (C) SH-SY5Y cells were transfected with the indicatedcombinations of constructs expressing mHtt(1–171)150Q-GFP, Pin1 siRNAoligonucleotides, and siRNA-resistant wild-type Pin1-HA (WT) or the catalyt-ically inactive mutant S67E (CI). Apoptosis of mHtt GFP-expressing cells wasevaluated by TUNEL assay after 48 h. The histograms show mean and SD ofthree independent experiments. (D) Expression of p21WAF1 mRNA was ana-lyzed by qRT-PCR from the striatum of 12-mo-old mice of the indicated gen-otypes (at least three mice for each genotype), normalizing for expression ofβ-actin. A t test was performed using homoschedastic variance throughgroups and one tail parameter. p53 was immunoprecipitated from equalamounts of brain lysates of the same mice and analyzed by Western blot.Actin protein levels in input lysates are shown. (E) The expression of the p53target PUMA in total brain lysates of 12-mo-old mice of the indicated geno-types (two mice for each genotype, 1 and 2) was analyzed by Western blot.

Grison et al. PNAS | November 1, 2011 | vol. 108 | no. 44 | 17981

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

15, 2

020

phosphorylation triggered by mHtt (Fig. 5B) similar to PKCδknockdown (Fig. 3D). Because specific inhibitors of HIPK2 arenot available, we attempted to pharmacologically interfere withthis pathway by inhibiting the upstream kinases. Treatment of SH-SY5Y cells with caffeine, a well-known inhibitor of ATM/ATRactivities, strongly reduced mHtt-dependent cellular toxicity andSer46 phosphorylation (Fig. 5C). We then focused on ATM,which, besides inducing HIPK2, directly phosphorylates p53 onSer46, in addition to Ser15 (29). Strikingly, the ATM-specific in-hibitor KU55933 was effective in preventing mHtt-induced apo-ptosis by reducing phosphorylation of p53 on both Ser46 andSer15 (Fig. 5D).Together, our experimental evidences support a model (Fig.

5E) where stress generated by mHtt triggers activation of ATM,HIPK2, and PKCδ kinases, which lead to phosphorylation of p53on Ser46. This process preludes Pin1-dependent prolyl isomeri-zation and consequent dissociation of p53 from iASPP as a pre-requisite for taking the apoptotic route.

DiscussionDespite the huge amount of data accumulated so far, a cure forHD is not yet available. Though attention has been especiallydevoted to the roles of downstream effectors, such as caspases, in

neurodegeneration, the identity of druggable upstreammediatorsof polyQ-dependent toxicity still remains elusive. In this work wehave focused on an emerging model of HD pathogenesis, wheremHtt evokes a canonical DNAdamage response in neuronal cells,with induction of ATM/ATRkinases and consequent activation ofp53 (3). We have observed that an important mark of p53 acti-vation, i.e., Ser46 phosphorylation, is induced downstream ofmHtt expression and is also evident in the brains of HD patients.Ser46 phosphorylation is specifically triggered by severe or per-sistent stress and represents a major event in shifting the p53 re-sponse from cell-cycle arrest to apoptosis (17, 30, 31). Here wehave shown that this modification is a prerequisite for execution ofapoptosis downstream ofmHtt.We have previously demonstratedthat recognition of phosphorylated Ser46 by the prolyl isomerasePin1 leads to dissociation of p53 from the apoptosis inhibitoriASPP (14). Our results indicate that in cells expressing mHtt,isomerization by Pin1 is an essential step for unleashing the ap-optotic potential of p53 from iASPP, thus allowing induction ofapoptotic effectors.HD neuropathology is characterized by massive loss of medium

spiny neurons in the striatum. The influence of p53 in HD patho-genesis in vivo has been clearly shown by using mHtt-transgenic flyand mouse models (5). Here we demonstrate that genetic ablation

A C

B D

Fig. 3. p53 Ser46 phosphorylation is promoted by mHtt through HIPK2 and PKCδ and is required for apoptosis induction by Pin1. (A) p53 scheme indicatingPin1 consensus sites (phospho–Ser/Thr-Pro). DBD, DNA binding domain; NLS, nuclear localization signal; TA, transactivation domain. p53 mutants have Ser/Thr-to-Ala substitutions in Pin1 consensus sites at residue 46 (p53 S46A) or at the three other major Pin1-binding sites at residues 33, 81, and 315 (p53 3M-S46wt). (B) p53-null H1299 cells were transfected with the indicated constructs (see also A). Apoptosis of mHtt GFP-expressing cells was evaluated by TUNELassay after 48 h. The histograms show mean and SD of three independent experiments. (C) SH-SY5Y cells were transfected with mHtt(1–171)150Q-GFP andtreated with the ATM inhibitor KU-55933 10 μM for 24 h. HIPK2 protein levels were evaluated by Western blot. (D) SH-SY5Y cells were transfected with theindicated combinations of mHtt(1–171)150Q-GFP expression construct and siRNA oligonucleotides for HIPK2 and PKCδ. Apoptosis of mHtt GFP-expressing cellswas evaluated by TUNEL assay after 48 h. The histograms show mean and SD of three independent experiments.

17982 | www.pnas.org/cgi/doi/10.1073/pnas.1106198108 Grison et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

15, 2

020

of Pin1 in HdhQ111 KI mice prevents precocious activation of p53,suggesting that Pin1 is required for induction of the p53 response,at least in early stages of HD. In fact, we observed that p53 tran-scriptional activity is induced in striatal neurons in vivo more than1 y before cell death, implying that these cells possess the ability todeal with chronic stress for long periods of time. Analysis of olderanimals also suggested that interference with p53 activation bytargeting Pin1 might reduce Htt-induced neurodegeneration.Our findings also imply that p53-independent pathways may

concur to HD-related toxicity. The p53 family member p73 hasbeen found relevant for mHtt-induced neuronal death (32), andthe ability of Pin1 to potentiate the apoptotic activity of both p53and p73 (33) might be critical in this respect.By describing how mHtt stimulates the activation of p53 by

triggering its phosphorylation-induced, Pin1-dependent isomeri-zation, we suggest that this polyQ-expanded protein causes neu-rotoxicity by acting as an upstream inducer rather than throughdirect interaction with p53 (5, 34); this implies that p53 activationmight represent a general pathogenic mechanism for polyglut-amine diseases sharing the occurrence of DNA lesions (4). In theCNS, p53 mediates neuronal death in response to excitotoxicityand oxidative stress (35), and its activity has been implicated inother neurodegenerative diseases, such as PD (36). Intriguingly,p53 has been shown to increase mHtt expression (37). Therefore,by activating p53, Pin1 might enforce a noxious loop triggered byHtt mutation, enhancing its toxic effects. In this respect, commongenetic polymorphisms (SNPs) affecting stress-induced p53 acti-vation (38) might impinge on HD pathogenesis (39). However,SNPs in the Pin1 promoter region have also been described asaffecting protein expression (40, 41), and it would be thus in-teresting to investigate whether Pin1 may act as a modifier ofHD pathogenesis.Identification of upstreampathogenic events inHD is crucial for

designing therapeutic interventions, and the vast knowledgeavailable on the p53 pathway provides an advantageous standpointto highlight regulatory nodes suitable formanipulation. Themodel

emerged from our data (Fig. 5E) details potential pharmacologicaltargets. First, we demonstrated that small-molecule inhibitors ofPin1 can protect neuronal cells from mHtt-induced apoptosisin vitro andmay therefore be effective as a therapeutic strategy fortreatment of HD. Although it is arguable that development ofclinically useful inhibitors of Pin1 awaits further improvement, ourresults may also indicate stress-induced p53 kinases as druggabletherapeutic targets. The dissociation of p53 from iASPP and theinduction of apoptotic effectors can indeed be prevented by in-terfering with phospho-Ser46 isomerization by Pin1 (14). Here wehave succeeded in reducing mHtt-dependent apoptosis of neuro-nal cells by inhibiting the activity of PKCδ andATMkinases, whichlead to Ser46 phosphorylation either directly or indirectly. Proteinkinases are a growing drug target class for diseases of peripheraltissues, and several candidate therapeutics targeting CNS kinasesare now in various stages of preclinical and clinical development.For instance, specific inhibitors of PKCδ have shown preclinicalin vivo efficacy in treatment of PD (42).Because clinical trials of molecules that may restore function-

ality to a single cellular pathway have failed, special attention hasbeen devoted to the identification of drugs that may interfere with

A B

C D

Fig. 4. mHtt induces dissociation of p53 from iASPP in a Ser46- and Pin1-dependent fashion. (A) H1299 cells were transfected with constructsexpressing p53 and increasing amounts (+, ++) of mHtt(1–171)150Q; theinteraction of p53 with endogenous iASPP protein was then analyzed bycoimmunoprecipitation. (B) H1299 cells were transfected with constructsexpressing p53 and Htt(1–171) fragments bearing either 21Q or 150Q. Totalcell lysates were analyzed by Western blot with antibodies specific for Ser46-phosphorylated and for total p53. (C) The interaction between p53 andiASPP proteins after transfection of increasing amounts of wt Htt(1–171)21Qin H1299 cells was determined as in A. (D) The effect of RNAi-mediatedknockdown of Pin1 on the dissociation of p53 from iASPP induced byoverexpressed mHtt was determined as in A.

A B C

E

D

Fig. 5. Inhibition of p53 phosphorylation on Ser46 or of Pin1-mediated isom-erization reduces mHtt-dependent apoptosis. (A–D) SH-SY5Y cells weretransfected with a construct expressing mHtt(1–171)150Q-GFP and treatedwith the Pin1 inhibitor PiB, 5 μM (A), the PKCδ inhibitor rottlerin, 5 μM (B), theATM/ATR inhibitor caffeine, 3 mM (C), the ATM-specific inhibitor KU-55933,10 μM (D), or with same amount of solvent as a control (–). Apoptosis of mHttGFP-expressing cells was evaluated by TUNEL assay after 24 h. Graphs showmeans and SD of three independent experiments. To detect p53 phosphory-lation, p53 was immunoprecipitated from equal amounts of total cell lysatesand analyzed by Western blot. The protein levels of Pin1, actin, and mHtt ininput lysates are shown. (E) Model for regulation of p53 by Pin1 upon cellularstress generatedbymutanthuntingtin. In cells expressingmHtt, the activities ofATM, HIPK2, and PKCδ lead to phosphorylation of p53 on Ser46. Subsequentprolyl isomerization of the phospho–Ser46-Pro47 site by Pin1 then unlocks p53from the apoptosis inhibitor iASPP, leading to induction of apoptotic genes.Pharmacologic interference with this pathway can be accomplished by use ofsmall-molecule inhibitors that target ATM, PKCδ, or Pin1, thereby preventingp53 cytotoxic activity.

Grison et al. PNAS | November 1, 2011 | vol. 108 | no. 44 | 17983

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

15, 2

020

different pathways at the same time. To this purpose, caffeineseems particularly interesting for its neuroprotective properties asa blocker of adenosine A2A receptors (ADORA2A) (43). Re-cently, a genetic variant of ADORA2A has been identified asa modifier of age at onset in HD (44). The combined actions asinhibitor of p53-mediated apoptotic pathway as well as neuro-protective molecule acting at adenosine receptors may increasecaffeine’s chances as a pharmacological intervention for HD.

MethodsCell Lines and HD Tissues. SH-SY5Y and SK-N-SH human neuroblastoma cellswere cultured in 1:1 Eagle’s minimal essential medium (MEM)/F12 Ham’smedium with 15% FBS, 0.5% GlutaMAX (Gibco), 1% nonessential aminoacids, and antibiotics. Neuro-2a mouse neuroblastoma cells were cultured inMEM with 10% FBS, 1% GlutaMAX, 1% nonessential amino acids, and anti-biotics. H1299 human lung carcinoma cells were cultured in Roswell ParkMemorial Institute medium with 10% FBS and antibiotics. PPIase-ParvulinInhibitor was from Calbiochem; Rottlerin and caffeine were from Sigma.Plasmids and siRNA oligonucleotides (Table S1) are described in SI Methods.

HD and control postmortem brain tissues were collected by the HarvardBrain Tissue Resource Center, McLean Hospital (Belmont, MA). HD brainswere assigned Vonsattel grade-3 pathology. Postmortem intervals were from12 to 32 h for controls and from 8 to 30 h for HD brains.

In Vitro Binding and Western Blot. Analysis of p53-Pin1 interaction by GST pull-down and coimmunoprecipitation was done as described (12). p53-iASPPcoimmunoprecipitation was performed as described (14). More details areprovided in SI Methods.

Antibodieswere anti-Pin1 polyclonal (12) andmonoclonal (G-8 Santa Cruz);anti-p53 Pab240, FL-393, and DO-1 (Santa Cruz); anti-iASPP pAbiASPPN1 andmAbiASPP49.3 (14); anti-Huntingtin MAB5490 (Millipore); anti–phospho-Ser15-p53 (Cell Signaling Technology); anti–phospho–Ser46-p53 (BD Phar-mingen); anti-PUMA ab-9643 (Abcam); anti-actin and anti-tubulin (Sigma),and anti-HSP90 F-8 (Santa Cruz). Anti-HIPK2 monoclonal antibody was a giftof L. Schmitz (University of Giessen, Giessen, Germany).

Mice Strains. HdhQ111 knock-in mice expressing the complete endogenous Httgene with 111 polyQ (20) and their wild-type littermates (HdhQ7) in C57BLbackgroundwereprovidedbyM.MacDonald (MassachusettsGeneral Hospital,Boston). Pin1 KO mice in C57BL background (22) were provided by A. Means(DukeUniversity, Durham,NC). HdhQ7/Q111:Pin1WT/KOmicewere intercrossed togive double-homozygous and heterozygous littermates. Genotyping for bothloci was performed by PCR on tail DNA as described (20, 22).

RT-PCR. Total RNA was extracted with QIAzol, and cDNA was transcribed usingQuantiTectReverseTranscriptionKit (Qiagen).RT-PCRwasperformedwithQuanti-Fast SYBR Green PCR Kit (Qiagen). Primer sequences are reported in Table S2.

Apoptosis Assays. TUNEL assays were performed with TMR Red in Situ CellDeath Detection Kit (Roche) following manufacturer’s instructions.

ACKNOWLEDGMENTS. We thank colleagues at Laboratorio NazionaleConsorzio Interuniversitario per le Biotecnologie for advice, G. Pastore andM. Maurutto for technical support, S. Soddu, X. Lu, and T. Hofmann forreagents, and G. Leanza for mouse immunohistochemistry. This work wassupported by Telethon Grant GGP07185 (to G.D.S. and F.P.) and AssociazioneItaliana per la Ricerca sul Cancro (G.D.S.).

1. The Huntington’s Disease Collaborative Research Group (1993) A novel gene con-taining a trinucleotide repeat that is expanded and unstable on Huntington’s diseasechromosomes. Cell 72:971–983.

2. Vonsattel JP, DiFiglia M (1998) Huntington disease. J Neuropathol Exp Neurol 57:369–384.

3. Illuzzi J, Yerkes S, Parekh-Olmedo H, Kmiec EB (2009) DNA breakage and induction ofDNA damage response proteins precede the appearance of visible mutant huntingtinaggregates. J Neurosci Res 87:733–747.

4. Bertoni A, et al. (2011) Early and late events induced by polyQ-expanded proteins:Identification of a common pathogenic property of polyQ-expanded proteins. J BiolChem 286:4727–4741.

5. Bae BI, et al. (2005) p53 mediates cellular dysfunction and behavioral abnormalities inHuntington’s disease. Neuron 47:29–41.

6. Illuzzi JL, Vickers CA, Kmiec EB (2011) Modifications of p53 and the DNA damageresponse in cells expressing mutant form of the protein huntingtin. J Mol Neurosci 45:256–268.

7. Vousden KH, Prives C (2009) Blinded by the light: The growing complexity of p53. Cell137:413–431.

8. Kruse JP, Gu W (2009) Modes of p53 regulation. Cell 137:609–622.9. Collavin L, Lunardi A, Del Sal G (2010) p53-family proteins and their regulators: Hubs

and spokes in tumor suppression. Cell Death Differ 17:901–911.10. Yeh ES, Means AR (2007) PIN1, the cell cycle and cancer. Nat Rev Cancer 7:381–388.11. Lu KP, Zhou XZ (2007) The prolyl isomerase PIN1: A pivotal new twist in phosphory-

lation signalling and disease. Nat Rev Mol Cell Biol 8:904–916.12. Zacchi P, et al. (2002) The prolyl isomerase Pin1 reveals a mechanism to control p53

functions after genotoxic insults. Nature 419:853–857.13. Zheng H, et al. (2002) The prolyl isomerase Pin1 is a regulator of p53 in genotoxic

response. Nature 419:849–853.14. Mantovani F, et al. (2007) The prolyl isomerase Pin1 orchestrates p53 acetylation and

dissociation from the apoptosis inhibitor iASPP. Nat Struct Mol Biol 14:912–920.15. Lu PJ, Wulf G, Zhou XZ, Davies P, Lu KP (1999) The prolyl isomerase Pin1 restores the

function of Alzheimer-associated phosphorylated tau protein. Nature 399:784–788.16. Ryo A, et al. (2006) Prolyl-isomerase Pin1 accumulates in Lewy bodies of Parkinson dis-

ease and facilitates formation of alpha-synuclein inclusions. J Biol Chem 281:4117–4125.17. Mayo LD, et al. (2005) Phosphorylation of human p53 at serine 46 determines pro-

moter selection and whether apoptosis is attenuated or amplified. J Biol Chem 280:25953–25959.

18. DiFiglia M, et al. (1997) Aggregation of huntingtin in neuronal intranuclear inclusionsand dystrophic neurites in brain. Science 277:1990–1993.

19. Schilling G, et al. (1999) Intranuclear inclusions and neuritic aggregates in transgenicmiceexpressing a mutant N-terminal fragment of huntingtin. Hum Mol Genet 8:397–407.

20. White JK, et al. (1997) Huntingtin is required for neurogenesis and is not impaired bythe Huntington’s disease CAG expansion. Nat Genet 17:404–410.

21. Wheeler VC, et al. (2002) Early phenotypes that presage late-onset neurodegenera-tive disease allow testing of modifiers in Hdh CAG knock-in mice. Hum Mol Genet 11:633–640.

22. Atchison FW, Capel B, Means AR (2003) Pin1 regulates the timing of mammalianprimordial germ cell proliferation. Development 130:3579–3586.

23. Liou YC, et al. (2003) Role of the prolyl isomerase Pin1 in protecting against age-dependent neurodegeneration. Nature 424:556–561.

24. Di Stefano V, Rinaldo C, Sacchi A, Soddu S, D’Orazi G (2004) Homeodomain-inter-acting protein kinase-2 activity and p53 phosphorylation are critical events for cis-platin-mediated apoptosis. Exp Cell Res 293:311–320.

25. Winter M, et al. (2008) Control of HIPK2 stability by ubiquitin ligase Siah-1 andcheckpoint kinases ATM and ATR. Nat Cell Biol 10:812–824.

26. Lee SJ, Kim DC, Choi BH, Ha H, Kim KT (2006) Regulation of p53 by activated proteinkinase C-delta during nitric oxide-induced dopaminergic cell death. J Biol Chem 281:2215–2224.

27. Yoshida K, Liu H, Miki Y (2006) Protein kinase C delta regulates Ser46 phosphorylationof p53 tumor suppressor in the apoptotic response to DNA damage. J Biol Chem 281:5734–5740.

28. Uchida T, et al. (2003) Pin1 and Par14 peptidyl prolyl isomerase inhibitors block cellproliferation. Chem Biol 10:15–24.

29. Kodama M, et al. (2010) Requirement of ATM for rapid p53 phosphorylation at Ser46without Ser/Thr-Gln sequences. Mol Cell Biol 30:1620–1633.

30. Oda K, et al. (2000) p53AIP1, a potential mediator of p53-dependent apoptosis, andits regulation by Ser-46-phosphorylated p53. Cell 102:849–862.

31. Smeenk L, et al. (2011) Role of p53 serine 46 in p53 target gene regulation. PLoS ONE6:e17574.

32. Hoshino M, et al. (2006) Transcriptional repression induces a slowly progressiveatypical neuronal death associated with changes of YAP isoforms and p73. J Cell Biol172:589–604.

33. Mantovani F, et al. (2004) Pin1 links the activities of c-Abl and p300 in regulating p73function. Mol Cell 14:625–636.

34. Steffan JS, et al. (2000) The Huntington’s disease protein interacts with p53 and CREB-binding protein and represses transcription. Proc Natl Acad Sci USA 97:6763–6768.

35. Mattson MP, Duan W, Pedersen WA, Culmsee C (2001) Neurodegenerative disordersand ischemic brain diseases. Apoptosis 6:69–81.

36. Alves da Costa C, Checler F (2011) Apoptosis in Parkinson’s disease: Is p53 the missinglink between genetic and sporadic Parkinsonism? Cell Signal 23:963–968.

37. Feng Z, et al. (2006) p53 tumor suppressor protein regulates the levels of huntingtingene expression. Oncogene 25:1–7.

38. Grochola LF, Zeron-Medina J, Mériaux S, Bond GL (2010) Single-nucleotide poly-morphisms in the p53 signaling pathway. Cold Spring Harb Perspect Biol 2:a001032.

39. Chattopadhyay B, Baksi K, Mukhopadhyay S, Bhattacharyya NP (2005) Modulation ofage at onset of Huntington disease patients by variations in TP53 and human caspaseactivated DNase (hCAD) genes. Neurosci Lett 374:81–86.

40. Segat L, et al. (2007) PIN1 promoter polymorphisms are associated with Alzheimer’sdisease. Neurobiol Aging 28:69–74.

41. Lu J, et al. (2009) A novel functional variant (−842G>C) in the PIN1 promoter con-tributes to decreased risk of squamous cell carcinoma of the head and neck by di-minishing the promoter activity. Carcinogenesis 30:1717–1721.

42. Chico LK, Van Eldik LJ, Watterson DM (2009) Targeting protein kinases in centralnervous system disorders. Nat Rev Drug Discov 8:892–909.

43. Prediger RD (2010) Effects of caffeine in Parkinson’s disease: From neuroprotection tothe management of motor and non-motor symptoms. J Alzheimers Dis 20(Suppl1):S205–S220.

44. Dhaenens CM, et al.; Huntington French Speaking Network (2009) A genetic variationin the ADORA2A gene modifies age at onset in Huntington’s disease. Neurobiol Dis35:474–476.

17984 | www.pnas.org/cgi/doi/10.1073/pnas.1106198108 Grison et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

15, 2

020