Neurobiology of Disease 45 (2012) 508–518

Contents lists available at SciVerse ScienceDirect

Neurobiology of Disease

j ourna l homepage: www.e lsev ie r .com/ locate /ynbd i

In vitro drug treatments reduce the deleterious effects of aggregates containingpolyAla expanded PHOX2B proteins

Eleonora Di Zanni a,1, Tiziana Bachetti a,1, Sara Parodi a,2, Paola Bocca b, Ignazia Prigione b, Simona Di Lascio c,Diego Fornasari c, Roberto Ravazzolo a,d, Isabella Ceccherini a,⁎a Laboratorio di Genetica Molecolare, Istituto Giannina Gaslini, Largo G. Gaslini 5, 16148 Genova, Italyb Laboratorio di Oncologia, Istituto Giannina Gaslini, Largo G. Gaslini 5, 16148 Genova, Italyc Department of Pharmacology, School of Medicine, Università degli Studi di Milano and CNR-Institute of Neuroscience, Milan, Italyd Dipartimento di Pediatria e CEBR, Università degli Studi di Genova, 16148 Genova, Italy

⁎ Corresponding author. Fax: +39 010 3779797.E-mail address: isa.c@unige.it (I. Ceccherini).

1 These authors equally contributed to the work.2 Current address: Department of Neuroscience and B

tute of Technology, Genoa, Italy.Available online on ScienceDirect (www.scienced

0969-9961/$ – see front matter © 2011 Elsevier Inc. Alldoi:10.1016/j.nbd.2011.09.007

a b s t r a c t

a r t i c l e i n f o

Article history:Received 3 June 2011Revised 9 August 2011Accepted 13 September 2011Available online 21 September 2011

Keywords:Congenital Central HypoventilationSyndromePHOX2BPolyalanine expansionsIntracytoplasmic aggregatesDrug treatment17-AAGCurcumin

Heterozygous in frame duplications of the PHOX2B gene, leading to polyalanine (polyAla) expansions rangingfrom +5 to +13 residues of a 20-alanine stretch, have been identified in the vast majority of patients affectedwith Congenital Central Hypoventilation Syndrome (CCHS), a rare neurocristopathy characterized by absenceof adequate autonomic control of respiration with decreased sensitivity to hypoxia and hypercapnia. Ventilatorysupports such as tracheostomy, nasal mask or diaphragm pacing represent the only options available for affected.Wehave already shown that the severity of the CCHS phenotype correlateswith the length of polyAla expansions,ultimately leading to formation of toxic intracytoplasmic aggregates and impaired PHOX2Bmediated transactiva-tion of target gene promoters, such as DBH. At present, there is no specific treatment to reduce cell aggregatesand to ameliorate patients' respiration. In this work, we have undertaken in vitro analyses aimed at assessingthe effects of molecules on the cellular response to polyAla PHOX2B aggregates. In particular, we tested 17-AAG,ibuprofen, 4-PBA, curcumin, trehalose, congo red and chrysamine G for their ability to i) recover the nuclear loca-lisation of polyAla expanded PHOX2B, ii) rescue of PHOX2B mediated transactivation of the DBH promoter, andiii) clearance of PHOX2B (+13 Ala) aggregates. Our data have suggested that 17-AAG and curcumin are effectivein vitro in both rescuing the nuclear localization and transactivation activity of PHOX2B carrying the largest expan-sion of polyAla and promoting the clearance of aggregates of these mutant proteins inducing molecular mecha-nisms such as ubiquitin–proteasome (UPS), autophagy and heat shock protein (HSP) systems.

rain Technologies, Italian Insti-

irect.com).

rights reserved.

© 2011 Elsevier Inc. All rights reserved.

Introduction

Congenital central hypoventilation syndrome (CCHS) is a rare neu-rocristopathy characterized by decreased sensitivity to hypoxia and hy-percapnia with absence of adequate autonomic control of breathing,especially during sleep (Coleman et al., 1980; Patwari et al., 2010). Inparticular, patients with CCHS show an adequate ventilation whileawake but hypoventilate during sleep (Patwari et al., 2010). CCHS is achronic disorder, with no effective treatment available, except ventila-tory supports such as tracheostomy, nasal mask or diaphragm pacing.CCHS is inherited as an autosomal dominant trait with reduced pene-trance, associated with mutations of the PHOX2B gene. This gene en-codes for a transcription factor expressed in several different districts

of the autonomic nervous system, in particular in the developing hind-brain and peripheral nervous system aswell as in all noradrenergic cen-ters and in specific neuronal groups, such as those involved in themedullary control reflexes of autonomic functions (Brunet and Pattyn,2002; Pattyn et al., 1997, 1999).

The most frequent mutations found in CCHS patients are heterozy-gous in-frame duplications within a 20 alanine stretch, leading to expan-sions from+5 to+13 alanine residues, though frameshift, nonsense andmissense mutations, frequently reported in association with Hirsch-sprung disease and tumors of neural crest origin, are also identified insmall subsets of patients (Amiel et al., 2003; Matera et al., 2004; Sasakiet al., 2003; Weese-Mayer and Berry-Kravis, 2004). A correlation be-tween length of alanine expanded stretches and phenotype severity hasalready been reported in CCHS patients (Matera et al., 2004; Patwariet al., 2010); moreover, functional studies have correlated polyalanine(polyAla) expansions with decreasing PHOX2B-mediated activation ofthe DBH and PHOX2A promoters, due to polyAla length-dependent for-mation of cytoplasmic aggregates which prevent the protein from enter-ing the nucleus, binding DNA and activating transcription (Bachetti et al.,2005, 2007; Trochet et al., 2005).

509E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

PolyAla, as well as polyglutamine (polyQ) expansions belong to aclass of defective trinucleotide repeats known to associate with manyneurodegenerative diseases (Albrecht and Mundlos, 2005) and toshare typical features such as formation of amyloid-likefibrils character-ized by a β-sheet conformation thus leading to insoluble protein aggre-gates (Blondelle et al., 1997; Forood et al., 1995; Scheuermann et al.,2003; Shinchuk et al., 2005). To reduce aggregate formation and celldeath, therapeutic strategies, tested so far in in vitromodels of neurode-generative disorders, have targeted protein misfolding and chaperonesexpression. Indeed, several chaperones, members of the heat shock pro-tein (HSP) family, have been shown to interact with huntingtin proteinscontaining different polyQ lengths, and to co-localize with intracellularaggregates. Furthermore, transient overexpression of these HSPs doesreduce aggregate formation as well as cellular toxicity induced by ex-panded polyQ tracts (Abu-Baker et al., 2003; Bachetti et al., 2007; Baoet al., 2004; Cummings et al., 1998; Hay et al., 2004; Jana et al., 2000;Wang et al., 2005). Therefore, compounds that either induce HSPs ex-pression, interact with the secondary configuration of mutated proteins,or reduce aggregates formation may be beneficial to rescue the cellularphenotype in association with polyAla expansions. To this end, wehave investigated the effects of the following molecules on the cellularresponse to polyAla PHOX2B aggregates: 1) the antibiotic geldanamycin(GA), already shown to activate a heat shock response and to inhibithuntingtin aggregation in a cell culture model of Huntington's disease(Sittler et al., 2001), in addition to promoting nuclear localisation andclearance of PHOX2B misfolded proteins (Bachetti et al., 2007) and, forthis reason, used as positive control in our cellular model; 2) the GA an-alog 17-AAG (17-(Allylamino)-17-demethoxygeldanamycin), which iscurrently in clinical trials as an anticancer drug, and specifically bindsto and inhibits HSP90 (Neckers and Neckers, 2002); 3) ibuprofen,known to induce HSP70 expression and to reduce poly(A) binding pro-tein, nuclear 1 (PABPN1) aggregation (Wang et al., 2005;Wang and Bag,2008); 4) curcumin (diferuloylmethane), amajor component of turmer-ic (Curcuma Longa) at present under phase II clinical investigation forthe treatment of various tumors, and already investigated in differentneurological disorders (Cheng et al., 2001; Sharma et al., 2004) for itspotent anti-inflammatory, antioxidant, and anti-protein-aggregation ac-tivities (Cole et al., 2007); 5) sodium 4-phenylbutyrate (4-PBA), a his-tone deacetylase inhibitor, FDA approved drug for management ofurea cycle disorders, has been shown to improve survival and attenuatestriatal atrophy in the R6/2 transgenicmousemodel of Huntington's dis-ease (Gardian et al., 2005); 6) trehalose, already shown to reduce aggre-gation and toxicity of mutant PABPN1 in vitro and in vivo, and to delaythe onset of muscle weakness in OPMD transgenic mice, that developa progressive neuro-muscular phenotype accompanied by the forma-tion of aggregates in skeletal muscle nuclei (Davies et al., 2006);7) congo red, and its analog chrysamine G, acting by inhibiting aggrega-tion of amyloid β-peptide and its neurotoxic effect in cultures of rat hip-pocampal neurons (Burgevin et al., 1994; Fraser et al., 1992; Lorenzo andYankner, 1994) and also huntingtin aggregation in vitro (Heiser et al.,2000). Following treatments with these molecules, we have observedthat 17-AAG and curcumin exerted the major effects on both rescue ofthe PHOX2B (+13Ala) nuclear localization and transactivation activity,thus disclosing some of the mechanisms underlying polyAla expansionpathogenesis and drug mediated effects.

Material and methods

Cell line and cell cultures

COS-7 cells were cultured in Dulbecco's modified essential medium(DMEM, Sigma) supplemented with 10% fetal bovine serum (FBS)(Gibco, New Zealand), 1% L-glutamine, 100 U/ml penicillin and 100 g/mlstreptomycin in an atmosphere of 95% air and 5% CO2 at 37 °C. HeLacells were cultured in minimal essential medium (MEM, Euroclone) sup-plemented with 10% fetal bovine serum (FBS) (Gibco, New Zealand), 1%

L-glutamine, 1% non-essential aminoacids, 100 U/ml penicillin and100 g/ml streptomycin in an atmosphere of 95% air and 5% CO2 at 37 °C.

Fluorescence microscopy analysis

105 COS-7 cells were plated in 35 mm dishes 24 h prior to transfec-tion. Cells were transiently transfected with 700 ng of the GFP reporterplasmids pcDNA3.1/CT-GFP-TOPO [PHOX2B wt] and pcDNA3.1/CT-GFP-TOPO [dup39], encoding the wild type and the mutant PHOX2B(+13 Ala), this latter carrying 13 extra alanine residues in the 20 ala-nine stretch, respectively. Transfection was performed with FugeneHD transfection reagent (Roche). At the time of transfection, cellswere also addedwith different concentrations of compounds for further48 h. 48 h after transfection, cellswerewashed oncewith PBS 1x (Phos-phate Buffered Saline Tablets, Dulbecco's Formula, 1X; pH=7.4, MPBiomedical) and then fixed with MeOH/Acetone 1:1 for 3 min at roomtemperature. Nuclei were stainedwith DAPI (Roche). For quantificationof PHOX2B aggregates at least 100 cells were examined with a ZeissAxiophot fluorescence microscope and the GFP fluorescence allowedto assess the subcellular PHOX2B localization thus classified as “nucle-ar” or “nuclear and cytoplasmic”. The most effective concentrationwas assessed for each compound by at least three independentexperiments.

Analysis of LC3B localisation

105 HeLa cells were plated in 35 mm dishes 24 h prior to transfec-tion. 48 h after co-transfection with the pcDNA3.1/CT-GFP-TOPO[dup39] expression construct and the pEX-HcRed-hLC3WT (Addgeneplasmid 24991; Tanida et al., 2008), cells were washed with PBS 1xand then fixed with MeOH/Acetone 1:1 for 3 min at room temperature.Nuclei were stained with DAPI (Roche) and analyzed with a Zeiss Axio-phot fluorescence microscope.

Evaluation of proteasome activity

HeLa cells were transiently transfected with the ZsProSensor-1plasmid (Clontech), expressing the fluorescent ZsGreen protein, andadded with MG132 for 24 h. Expression of the ZsGreen protein, a C-terminal fusion of ZsGreen, a naturally occurring green fluorescentprotein, with the mouse ornithine decarboxylase degradation do-main, normally targeted for rapid degradation by the proteasome,was evaluated by direct microscope fluorescence.

Transcriptional activity assay

5×104 HeLa cells were seeded in 24-wells plate 24 h prior totransfection and transiently co-transfected, using Lipofectamine2000,with 370 ng of pcDNA3.1TOPO-PHOX2B wild type and pcDNA3.1-TOPO-dup39mutant PHOX2B expression constructs (this latter carry-ing 13 extra alanine residues in the 20 alanine stretch) (Bachetti et al.,2005) together with 70 ng of a construct containing four copies of do-main II of the DBH promoter, already known to be a PHOX2B targetsequence, cloned upstream of the Luciferase reporter gene (Adachiet al., 2000; Benfante et al., 2007). The pRL-CMV plasmid, expressingthe Renilla Luciferase gene, was co-trasfected and used as an internalcontrol. Luciferase activity (Dual-Luciferase Reporter Assay System,Promega) was performed 48 h after transfection with a TD-20/20Luminometer following manufacturer's instructions.

Compounds and treatment

17-(Allyloamino)-17-demethoxygeldanamycin (Sigma) (17-AAG)was prepared at 1 mM stock solution in DMSO and diluted in freshmedium at the final concentration of 25 nM, 50 nM, 100 nM and300 nM at the time of transfection. Curcumin (Sigma) was prepared

510 E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

at 11 mg/ml stock solution in DMSO and diluted in fresh medium atthe final concentration of 0.25 μM, 0.5 μM, 1 μM, 5 μM, 8 μM, 10 μMand 15 μM. Congo Red (Calbiochem) was prepared at 10 mg/mlstock solution in H2O and diluted in fresh medium at the final concen-tration of 0.5 μM, 1 μM and 2 μM. Chrysamine G (Calbiochem) wasprepared at 1 mg/ml stock solution in DMSO and diluted in fresh me-dium at the final concentration of 0.5 μM, 1 μM, 5 μM and 10 μM. 4-PBA (Calbiochem) was prepared at 10 mg/ml stock solution in H2Oand diluted in fresh medium at the final concentration of 1 mM and5 mM. Ibuprofen (Sigma) was prepared at 100 mg/ml stock solutionin H20 and diluted in fresh medium at the final concentration of8 μM, 16 μM, 32 μM, 50 μM, 100 μM and 1 mM. Trehalose (Sigma)was prepared at 1 M stock solution and diluted in fresh medium atthe final concentration of 10 μM, 100 μM, 1 mM, 10 mM and100 mM. 3-methyladenine (3-MA, Sigma) at 1 M stock in DMEMwas added to cells at the time of transfection to give a final concentra-tion of 10 μM. MG132 (Sigma) at 10 mM stock in water was added tocells at the time of transfection to give a final concentration of 10 μM.

Western blotting assay

4×105 HeLa cells were plated in 60 mm dishes and were treatedwith 17-AAG, GA and curcumin. After 48 h, cells were washed withPBS 1×, centrifuged and lysed with RIPA buffer (Tris–HCl 50 mM pH7.5, NaCl 150 mM, Triton-X 1%, SDS-20 0.1%, Na deoxycholate 1%, Pro-tease Inhibitor mix 1×). Total cell lysates were quantified and equalamounts were electrophoresed on 10% SDS-PAGE and transferred toa polyvinylidene difluoride membrane (Millipore). Proteins wereidentified by probing the membrane with the rabbit anti-LC3B anti-body (Sigma #L7543, dilution 1:1000), specifically addressed againstthe LC3B, with the mouse monoclonal HSP70 specific antibody (SantaCruz Biotechnology #SC24, dilution 1:1000) and then with a rabbitanti-mouse IgG-HRP(Dako #PO161, dilution 1:20,000) and a goatanti-rabbit IgG HRP (Dako #PO448, dilution 1:15,000), respectively.Signals were detected using the chemiluminescence reagent ECL ad-vance (Amersham) and protein levels in each sample were evaluatedby comparison with corresponding amounts of the housekeeping β-actin with the goat anti actin antibody (Santa Cruz Biotechnology#SC1616, dilution 1:2500).

Analysis of compound-mediated cell death and cytotoxicity

HeLa cells were cultured in 60 mm dishes and treated with allcompounds. Analysis of apoptosis was performed with the V-PE Apo-ptosis Detection kit (Pharmingen, BD Biosciences, San Jose, CA) fol-lowed, after 48 h and 96 h, by flow cytometry (FACSCalibur, BDBiosciences, Cell Quest software). Results were expressed as percentof GFP positive cells. Compound-mediated cytotoxicity was moni-tored using Real-Time Cell Analyzer single-plate (RTCA SP) Instru-ment, xCELLigence System (Roche Applied Science, Mannheim,Germany). The system monitors cellular events in real time measur-ing electrical impedance across interdigitated micro-electrode inte-grated on the bottom of tissue culture E-plates. The impedancemeasurement provides quantitative information about the biologicalstatus of the cells, including cell number, viability, and morphology.The Cell Index (CI) is derived as a relative change in measured electri-cal impedance to reflect the integrated cellular status in the culture.Background of the E-plates was determined in 100 μl medium andsubsequently 100 μl of the HeLa cell suspension was added (13,000cells/well). 2 h after seeding, different concentrations of various com-pounds were added in quadruplicate. Cells were grown for 48 h, withimpedance measured every 30 min. CI values were calculated andplotted on a graph. Standard deviations of well replicates for thetwo cell types subjected to different treatments were analyzed withthe RTCA Software.

Analysis of PHOX2B protein levels

Transient transfections were performed plating 4×105 HeLa cellsin 60 mm diameter dishes using Fugene HD Transfection Reagentwith 1.5 μg of the GFP reporter plasmids pcDNA3.1/CT-GFP-TOPO[PHOX2B wt] and pcDNA3.1/CT-GFP-TOPO [dup39]. Cells were trea-ted at the time of transfection and 48 h later, the mean fluorescenceintensity (MFI) of an identical amount of cells expressing the GFPtagged mutant protein was measured by FACS analysis (FACSCalibur,BD Biosciences, Cell Quest software).

Statistical analysis

ANOVA was performed to compare the treatment effect on nucle-ar localization and transcriptional activity among compounds. Analy-sis was run using the GraphPad Prism (version 5.04) software (http://www.graphpad.com/prism/), starting from raw data in triplicate foreach compounds under analysis, including untreated and DMSO-trea-ted cells. Pairwise t-tests and the Dunnett's test, performed by usingthe above statistical package, have then allowed to identify the bestperforming compounds (those showing a statistically significant dif-ference between treated and untreated cells) to continue the study.A correlation analysis was finally used (GraphPad Prism software)to test statistical association between amount of plasmid DNA trans-fected and nuclear localization of the exogenous protein.

Results

Effects of compounds on PHOX2B aggregates and nuclear localisation

Previous results of ours showed that polyAla expansions inducePHOX2B mis-localisation with presence of nuclear and cytoplasmicPHOX2B aggregates (Bachetti et al., 2005) and that the correct func-tion of mutant PHOX2B proteins could be rescued by administrationof the antibiotic geldanamycin (GA) (Bachetti et al., 2007).

Starting from this latter observation, we wondered whether addi-tional drugs could reduce formation of mutant protein aggregates andrestore the nuclear function of the transcription factor. COS-7 cells,being characterized by a large and well readable microscopic mor-phology, were used to evaluate aggregate formation as alreadyreported (Bachetti et al., 2005). To this end, COS-7 cells transientlytransfected with pcDNA3.1/CT-TOPO-GFP[dup39], encoding PHOX2Bcarrying +13Ala, were added with increasing amounts of 4-PBA, cur-cumin, ibuprofen, chrysamine G (CG), trehalose, congo red (CR) and17-(Allylamino)-17-demethoxygeldanamycin (17-AAG). Knowing al-ready its effect, GA 360 nM was used as positive control, and increas-ing amounts of each compound were tested to first assess the mosteffective concentrations, able to rescue a correct nuclear localisationof PHOX2B proteins. After 48 h, a fluorescence microscope analysisallowed to assess the drug mediated nuclear localization rescue(Fig. 1A). As expected (Bachetti et al., 2005, 2007), the wild type pro-tein was localized only in the nuclear compartment while dup39transfected cells showed aggregates in the cytoplasmatic compart-ment, in the absence of any treatment; on the other hand, we ob-served drug-dependent translocation of dup39 PHOX2B proteins inthe correct nuclear compartment. Fig. 1B reports the effect of treat-ments with the different compounds, with DMSO only and withoutany treatment on the nuclear localization of the PHOX2B mutant pro-tein. Although the nuclear rescue of the PHOX2B (+13Ala) proteinwas not complete with any drug, compared to cells expressing theWT protein, the Analysis of Variance (ANOVA) showed that overalltreated and untreated cells significantly differed (F=10.6, df 10;22,Pb0.0001). Particularly curcumin, congo red, trehalose, chrysamineG and 17-AAG showed the highest effect on the sub-cellular localisa-tion and on preventing the formation of protein aggregates. Indeed,17-AAG induced more than 60% rescue of the nuclear localisation of

Fig. 1. Nuclear localization of mutant PHOX2B following drug treatments. (A) Microscope fluorescence analysis of COS-7 cells transiently transfected with constructs encoding thewild type (wt), and the +13Ala (dup39). Nuclei were stained with DAPI and images at the right end represent the merge of the two adjacent pictures. (B) The effect of treatmentson the rescue of a correct sub-cellular localization of the PHOX2B (+13Ala) protein is shown as percentage of COS-7 cells transfected with pcDNA3.1/CT-TOPO-GFP[dup39] andcharacterized by an exclusively nuclear localization. The effect of either none or DMSO only treatments is reported in the left end of the diagram. Significant differences, reflectingthe effect of the treatment compared to the negative DMSO control, are indicated by asterisks (Student's t-test, *pb0.05, **pb0.01).

511E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

512 E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

dup39-PHOX2B already at 25 nM concentration, and with dose de-pendent effects. Similarly, also the effect of curcumin was dose de-pendent, but showing an increase of aggregates at concentrationshigher than 10 μM. This observation is consistent with the biphasic ef-fect of curcumin, already suggested to stimulate proteasome activityat low doses and to impair its function at high concentrations (Aliand Rattan, 2006).

Effect of treatments on rescue of PHOX2B-mediated DBH promotertransactivation

We already demonstrated a strict correlation between progressivedecrease of DBH promoter transcriptional activation and length of thePHOX2B alanine expanded tract in HeLa cells, with the dup39PHOX2B construct showing the lowest transactivation activity (Bachettiet al., 2005; Trochet et al., 2005). We wondered whether the above de-scribed treatments, in addition to allow a correct nuclear localization ofthe PHOX2B (+13Ala) protein, could also rescue its transcriptional ac-tivity. To this aim, HeLa cells were co-transfected with either the wildtype or dup39-PHOX2B expression constructs, along with a constructcontaining the DBH promoter cloned upstream of the Luciferase report-er gene. Luciferase assays revealed that incubation for 48 h with either300 nM 17-AAG/GA or curcumin restored most of the transcriptionalfunction of mutant PHOX2B, inducing more than 80% and 70% of theLuciferase's activity in mutant PHOX2B compared to wild typePHOX2B, respectively (Fig. 2). This observation was confirmed byANOVA: the statistically significant difference demonstrated amongtreated and untreated cells (F=6.03, df 9;20, Pb0.0004) disappearedwhen data obtained after treatments with curcumin, 17-AAG and GAwere removed (PN0.05). Pairwise t-tests andDunnett's test, successive-ly performed, have further confirmed the 17-AAG, GA and curcumin asthe best performing compounds.

Analysis of apoptosis and cytotoxic effect of treatments

To evaluate the cytotoxic potential of the compounds under anal-ysis, we tested 17-AAG, GA and curcumin with the xCELLigence sys-tem (Roche). This system monitors cellular events in real time,without incorporating any label. The measurement of the impedanceprovides a quantitative information about the biological status of thecells, including cell viability. Untreated (NT) HeLa cells were analyzed

Fig. 2. Transactivation of the DBH promoter by polyalanine expanded PHOX2B followingdrug treatments. Transcriptional activity obtained after co-transfection of PHOX2B con-structs carrying polyAla expanded tracts with a Luciferase reporter gene construct contain-ing the DBH regulatory region. The white and gray bars represent transactivation induced,in the absence of any treatment, by PHOX2B-WT and PHOX2B (+13Ala) respectively,while the dark bars refer to the effects of treatments during the co-transfection step. Valuesare themean±SDof three independent experiments performed in duplicate. Significant dif-ferences, reflecting the effect of the treatment compared to the negative+13Ala control, areindicated by asterisks (Student's t-test, *pb0.05, **pb0.01).

in the same experiment as a control to evaluate cell growth, adher-ence to plastic, time to reach confluence and plateau value.

The graph of the impedance demonstrates that GA becomes toxicafter 20 h of treatment, but cells that survive do restore their viabilityafter 48 h. Under the same conditions, 17-AAG 100 nM did not resultas toxic, while curcumin turned out to induce deleterious effects sincethe beginning of treatment, with very slow cell growth (Fig. 3A). Noneof the other molecules showed any toxic effect (data not shown).

To deepen into the toxicity of these compounds, we investigatedcell apoptosis by staining cells with annexin V and 7-amino-actinomycin D (7-AAD), followed by FACS analysis. Viable cells withintact membranes exclude 7-AAD, a vital dye, whereas membranesof dead and damaged cells are permeable to 7-AAD. Annexin V and7-AAD negative cells were therefore considered to be viable. AnnexinV positive and 7-AAD negative cells were considered to be in early ap-optosis and annexin V positive and 7-AAD positive cells were consid-ered to be apoptotic. Finally, annexin V negative and 7-AAD positivecells were considered as necrotic cells and represented only a lowproportion of the whole population (see Fig. 1S). No statistically sig-nificant difference could be demonstrated between treated anduntreated cells (Student's t-test PN0.05). HeLa cells treated withcompounds under study were analyzed by FACS after 48 h and5 days. The percentage of viable cells after 48 h decreased only incells treated with curcumin, 17-AAG and GA compared to untreatedcells, while survival at 5 days was not affected. This is confirmed bythe decrease of annexin V and 7-AAD negative cells, and by the in-crease of apoptotic cells, though not statistically significant, in bothtreated and untreated cells (Figs. 3B and C).

Clearance of PHOX2B polyalanine aggregates induced by treatments

The accumulation of mutant proteins is revealed by cytoplasmicaggregates formation. We already reported that clearance of these ag-gregates involves the ubiquitin–proteasome system (UPS) and autop-hagy (Bachetti et al., 2007). We then hypothesized that keeping theamount of mutant protein as low as possible may facilitate its refold-ing and consequently its nuclear localisation. To verify this hypothe-sis, COS-7 cells transiently transfected with increasing amounts ofpcDNA3.1/CT-GFP-TOPO [dup39] (300 ng, 500 ng, 750 ng, 1 μg)were subjected to immunofluorescence analysis. When we trans-fected 1 μg of pcDNA3.1/CT-GFP-TOPO [dup39], we observed high re-tention of cytoplasmic PHOX2B with aggregates formation comparedto cells transfected with 300 ng of the same plasmid (Fig. 4A). Thesedata demonstrate that there is an inverse statistically significant cor-relation between the amount of plasmid transfected and the nuclearlocalisation of the mutated protein, provided the two highestamounts of plasmid are considered as a single point, as both amountssaturate the system giving similar results (r=−0.998; P=0.0421).

Since immunofluorescence analysis showed a recovery of the nu-clear localization and the Luciferase assay revealed a rescue of thePHOX2B function, we investigated whether the mutated PHOX2Bcould be eliminated by drug treatments, and if the recovery of locali-zation and function was due to protein refolding. To this end, cellswere transiently transfected with pcDNA3.1/CT-GFP-TOPO [dup39]and GFP fluorescence was assessed by quantitative FACS analysis. Adecrease of the mean fluorescence intensity (MFI) of GFP for cellstreated with GA, 17-AAG and curcumin did confirm elimination ofmutant proteins by treatment with these agents (Fig. 4B). This sug-gests that clearance of dup39-PHOX2B results in dissolving aggre-gates and refolding mutant proteins. In particular, to investigate therole of HSP70, known to co-localize with PHOX2B aggregates and tobe induced by many different compounds, western blot analysiswas performed in duplicate on treated HeLa cells. As expected, wefound HSP70 induction by GA and 17-AAG (Fig. 4C) while curcumindid not show any effect on expression of HSP70 (not shown), thussuggesting it may promote refolding through a different path. The

Fig. 3. Viability of cells treatedwith different concentration of the compounds under analysis. (A) Output from the xCELLigence system (Roche). The xCELLigence curves, recorded across 48 htime, show the impedance of untreated (red) and treated cells (17-AAG green curve, curcumin orange curve, GA blue curve). Cytofluorimetric detection of apoptosis in HeLa cells, eitheruntreated or treated with different compounds, was assessed after 48 h of treatment (B) and after 5 days of treatment(C), by staining with annexin V and 7AAD followed by flow cytometry.Results are expressed as percentage of cells in each of the four states: alive (annexin V−, 7AAD−), early apoptosis (annexin V+, 7AAD−), late apoptosis (annexin V+, 7AAD+) and dead(annexin V−, 7AAD+).

513E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

same experiment, performed on cells transfected with PHOX2B-WTshowed that it is more successfully eliminated, likely due to itsmore soluble conformation with respect to the mutant protein, thusdemonstrating that these drug treatments can eliminate also thewild type protein, when overexpressed (data not shown).

Aggregate clearance induced by drug treatments involves the proteasomeand autophagy pathways

As suggested by analysis of protein levels, treatments with 17-AAG,GA and curcumin act by inducing elimination of themutated protein. Inorder to understand the molecular mechanisms involved in the clear-ance of aggregates in our in vitro system, we used 3-methyladenine(3-MA) and MG132 to inhibit the autophagy pathway and ubiquitin–proteasome system (UPS) respectively, as involvement of these twoprocesses has already been reported to explain the effects of the drugsunder analysis (Ali and Rattan, 2006; Bachetti et al., 2007; Riedelet al., 2010). The ability of 3-MA andMG132 to inhibit autophagy path-way and ubiquitin–proteasome system (UPS) in cell lines has been con-firmed by checking these two drugs in our cell system, as reported inFig. 2S. COS-7 cells transfected with pcDNA3.1/CT-TOPO-GFP [dup39]were then tested by immunofluorescence. Co-treatment of cells,expressing dup39, with either 17-AAG, curcumin or GA and the autop-hagy inhibitor 3-MA led to an increase in PHOX2B aggregates localizedin the nucleus and in the cytoplasm, compared to 3-MA untreated cells.Similarly, we could confirm the UPS involvement in the drug mediated

elimination, after observing formation of protein aggregates in the cyto-plasmic and perinuclear region in thepresence of the proteasome inhib-itor MG132 (Fig. 5). These findings are consistent with previousobservations of ours showing involvement of both proteasome andautophagy in the clearance of misfolded PHOX2B (Bachetti et al., 2007).

We already know that cells carrying the largest expanded alaninetract (dup39) form not only cytoplasmic and nuclear aggregates, butalso aggresome, a very large cytosolic perinuclear aggregate consid-ered as a marker of the proteasome impairment, a circumstance in-ducing the autophagic process (Bachetti et al., 2007). To confirminvolvement of the autophagy in PHOX2B elimination, HeLa cellswere co-transfected with either pcDNA3.1/CT-GFP-TOPO [PHOX2Bwt] or [dup39] expression constructs, along with pEX-HcRed-hLC3WT, a construct expressing LC3B wild type tagged to HcRed.LC3B is processed during the formation of autophagosomes in a cyto-solic isoform of 18 kDa (LC3B-I), that is converted to a 16 kDa isoformassociated with the autophagosomal membranes (LC3B-II), a markerof autophagosomes formation. The intracellular distribution of LC3Bis wide, being present both in the nucleus and in the cytoplasm,with formation of dots reflecting the autophagic activity of the cell.Here we show a different pattern of LC3B localisation between wtand PHOX2B (+13Ala): a quite exclusively diffuse localisation inthe presence of wt PHOX2B and, conversely, formation of LC3B dotsco-localizing with the peri-nuclear aggregates of PHOX2B (+13Ala).This result suggests that autophagy takes part in the elimination ofmutant PHOX2B aggregates (Fig. 6A).

Fig. 4. Amount of mutant PHOX2B protein in treated and untreated transfected cells and drug mediated HSP70 induction. (A) Proportion of cells transfected with 1 μg, 750 ng,500 ng or 300 ng of the pcDNA3.1/CT-GFP-TOPO [dup39] plasmid and showing a nuclear localization of fluorescence. Values are the mean±SD of three independent experiments.(B) PHOX2B (+13Ala) protein level, expressed as mean fluorescence intensity (MFI), in HeLa cells transfected with the pcDNA3.1/CT-GFP-TOPO [dup39] construct and treated withdifferent compounds. Values are mean±SD of three independent experiments. (C) Western blot analysis of lysates from HeLa cells treated with 17-AAG and GA and probed with aHSP70 antibody. The amount of the housekeeping actin level is represented in the bottom line.

514 E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

To further characterize this response, we wondered whether com-pounds under study were able to induce autophagy in our in vitro sys-tem. To this end, HeLa native cells were incubated for 48 h with either17-AAG (100 nM and 300 nM), GA and curcumin. Cell lysates weresubjected to immunoblot analysis and the amount of LC3B was com-pared to β-actin. Evaluation of the 16 kDa band revealed that GA and17-AAG induced activation of autophagy pathway (Fig. 6B). This hasbeen confirmed after quantitative analysis of three independent ex-periments suggesting that the LC3B-II protein is induced more incells treated with 17-AAG and GA than in untreated cells (Fig. 6C).

Discussion

Though the genetic basis of congenital central hypoventilation syn-drome (CCHS) is known, and some pathogenetic mechanisms associatedwith PHOX2Bmutations have recently beendisclosed (Amiel et al., 2003;Bachetti et al., 2005;Matera et al., 2004), this lifelong disorder, with ven-tilator supports representing the only options available at present for pa-tients, is still lacking an ameliorative pharmacological strategy. Onlyrecently, the fortuitous observation of a case of CO2-chemosensitivity re-covery in two CCHS women who took a progestin contraceptive, deso-gestrel, has demonstrated that response to hypoxia and hypercapniacan be pharmacologically restored in CCHS (Straus et al., 2010) andprompted us to look formolecules able to counteract the pathological ef-fects of polyAla PHOX2B aggregates.

We have recently shown that in vitro treatment of cells expressingpolyAla PHOX2B expansions with geldanamycin could prevent aggre-gate formation, induce clearance of intracellular inclusions and rescuethe transcriptional activity of PHOX2B (+13Ala) (Bachetti et al.,2007). As a straightforward continuation, here we have investigated

the effects of candidate therapeutic agents in an in vitro model ofCCHS, particularly focusing on molecules already used in clinical trialsfor other pathologies, namely 17-AAG and curcumin (Cheng et al.,2001; Heath et al., 2008; Lao et al., 2006; Ramanathan et al., 2007;Ronnen et al., 2006).

In order to test the effects of these drugs, we evaluated the ability ofthe selected compounds to recover the correct nuclear localisation ofPHOX2B (+13Ala), that is limited to 30% of analyzed cells. All the mol-ecules under analysis succeeded in reducing aggregation and enhancingtranslocation of the transcription factor to the nucleus. Particularly,after 48 h from the addition of any of these drugs, we could observean increment, although variable, of the percentage of cells showing anuclear localisation of PHOX2B (+13Ala). Indeed, while 17-AAG recov-ered a nuclear localisation of PHOX2B in more than 60% of cells, othercompounds showed a rescue in about 50% of cells, with 4-PBA beingthe least effective in term of presence of mutant PHOX2B in the nucleusin only 40% of the cells. In addition, to find out the most suitable drugconcentrations, immunofluorescence experiments were performedalso with different drug amounts. While 17-AAG as GA showed adose-dependent effect, curcumin displayed both a recovery at lowdoses and an increment of aggregates at high doses. Such a biphasicdose response to curcumin is already known to lead, at low concentra-tions, to increased activity of proteasome and increased expression ofheat-shock proteins, while 10 μM curcumin can induce inhibition ofproteasome activity (Ali and Rattan, 2006).

After assessing the most effective dose for each of the drugs understudy, we showed that treatments of PHOX2B (+13Ala) containingcells with curcumin and 17-AAG could lead to a significant incrementin transactivation of the DBH target promoter. Though 17-AAG can re-cover the correct localization of PHOX2B (+13Ala) at both 100 nM

Fig. 5. Effect of proteasome and autophagy inhibition on PHOX2B(+13Ala) cellular localization and aggregate formation. Immunofluorescence analysis of COS-7 cells expressingPHOX2B (+13Ala), treated with 17-AAG, curcumin and GA, either alone or with 3-methyladenine (3-MA) and MG132 to inhibit the autophagy pathway and ubiquitin–proteasomesystem (UPS), respectively. Arrows indicate peri-nuclear and cytoplasmic PHOX2B aggregates. Nuclei were stained with DAPI and images in the right box in each row represent themerge of the two adjacent pictures.

515E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

Fig. 6. Intracellular distribution and induction of LC3B following drug treatments. (A) Immunofluorescence detection in HeLa cells, co-transfected with pEX-HcRed-hLC3WT andpcDNA3.1/CT-GFP-TOPO [wt]/[dup39] constructs (green), of LC3B expression (red). Arrows indicate the co-localisation of LC3B dots with PHOX2B aggregates. Cell nuclei werestained with DAPI. The figure on the right represents the merge of adjacent images (100× magnification). (B) Western blot analysis of the two LC3B forms (I and II) involved inthe autophagy pathway was performed in total lysates from HeLa cells either untreated or treated with 17-AAG, GA and curcumin. The level of the reference actin gene isshown below. (C) Diagram reporting a quantitative estimation of the LC3B induction by treatments, as assessed by the above WB.

516 E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

and 300 nM, DBH promoter activity was rescued only by a treatmentwith 300 nM of 17-AAG, thus suggesting that different doses can havevariable effects for any given drug.

Although 4-PBA has anti-aggregation properties on polyQ inclu-sions, and is also known as a histone deacetylase inhibitor believedto promote gene expression (Gardian et al., 2005), expectations onpossible effects of 4PBA on polyAla PHOX2B expansions were not sup-ported by our in vitro experiments.

To assess whether and atwhich extent treatments could induce tox-icity, we investigated cellular viability and proliferation in real time,during the first 48 h after treatment, by using the xCELLigence system(Roche). Curcumin and GA were the only compounds that resulted tobe cytotoxic, as also demonstrated by FACS analysis with apoptosismarkers. Actually, curcumin, 17-AAG and GA showed cytotoxic effectsafter 48 h but not across a longer period of 5 days. Further studiesmay help to reconcile our observations with what reported in differentin vivomodels, suggesting no toxic effect inmice after administration of17-AAG for 20 weeks (Waza et al., 2005). Curcumin is a natural com-pound present in turmeric, a commonly used spice. In our in vitro sys-tem curcumin revealed cytotoxic effect at a concentration of 8 μM. Incontrast, in K562 cells, curcumin showed cytotoxic effect at concentra-tions above 20 μM and several phase I clinical trials demonstrated thatdietary curcumin administered at doses as high as 12 g per day waswell tolerated (Cheng et al., 2001; Lao et al., 2006; Teiten et al., 2009).Recent studies have examined the cytotoxic effects of curcumin at theblastocyst stage of mouse embryos and reported that 24 μM curcuminsignificantly increased apoptosis, thus decreasing total cell number,while treatments with 6 and 12 μM curcumin had the same effects asuntreated controls (Chen et al., 2010).

We have further observed that elimination of PHOX2B (+13Ala),due to 17-AAG, curcumin and GA, does enhance the refolding whichprompts the correct localisation of the remaining PHOX2B proteinsto the nucleus. Indeed, refolding mediated by molecular chaperonsis believed to be more effective in the absence of insoluble aggregates,namely once part of the mutant protein has been degraded. A criticalissue, however, still concerns the question that also the wild typePHOX2B protein could be partially eliminated. Similarly, studies ontherapeutic effect of 17-AAG, and its ability to degrade polyQ-expandedmutant Androgen Receptor in Spinal Bulbar Muscular Atrophy, sug-gested a more efficient degradation of mutant expanded proteinsboth in cells and transgenic mice, as compared to the wild-type protein(Waza et al., 2005). However, in our in vitro system, while degradationof thewild type PHOX2Bprotein seems to affect neither its nuclear loca-lisation nor its function, even under treatment conditions, degradationof dup39-PHOX2B is a critical step as the heat shock protein system be-comesmore effective and successfully aids in refoldingmutant proteins,thus allowing their correct localisation and function. Moreover, mutantproteins have been shown to form cytoplasmic and insoluble aggre-gates, whereaswild type proteins are predominantly expressed in a sol-uble form, and localize in the nucleus. The clearance of mutatedproteins in the presence of insoluble aggregates could result quite diffi-cult for the cells, and for this reason we may hypothesize that refoldingmediated by molecular chaperons is one of the major processes coun-teracting the detrimental effects of polyAla expanded PHOX2B proteins.

Reasoning that drug mediated elimination of mutant proteinsmight be advantageous to achieve correct cell functions, here wehave demonstrated an inverse correlation between amount of trans-fected plasmid and nuclear localisation of the corresponding mutant

517E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

protein. This can be explained considering that, if cell degradation sys-tems are not saturated by accumulation ofmutant proteins and forma-tion of aggregates, then partial elimination of PHOX2B (+13Ala), andimprovement in refolding andUPS, can bemore easily achieved. In ad-dition, recent studies have demonstrated that the transcription factorPHOX2B carrying the expanded polyAla region can sequestrate thewild type protein into nuclear aggregates, an observation which ac-counts for the already suggested dominant negative effect ofPHOX2Bmutations (Bachetti et al., 2005; Trochet et al., 2005). Consis-tent with what previously reported (Bachetti et al., 2007; Heat et al.,2008; Ronnen et al., 2006), we have demonstrated that curcuminand 17-AAG also act by improving the proteasome and autophagyme-diated degradation of mutant PHOX2B, thus confirming that PHOX2Bis eliminated through these two pathways. In particular, the use ofMG132 and 3-methyladenine allowed to confirm that blocking pro-teasome and autophagy leads to mis-localization of PHOX2B(+13Ala) with increased proportions of cells characterized by aggre-gates in the cytosolic compartment.

Clearance of PHOX2B (+13Ala) by autophagy has already beenreported, after demonstrating the presence of aggresomes in cells trans-fected with the dup39 PHOX2B construct and treated with lactacystin(Bachetti et al., 2007). Consistent with such activation of autophagy incells transfectedwith PHOX2B (+13Ala), aggresome formation becomesevident, and stands as a marker for proteasome activity impairment dueto excessive amount of aggregation-prone misfolded proteins in the cell(Johnston et al., 2002). In addition, here we have shown the co-localisa-tion ofmutant protein aggregateswith LC3, amarker of autophagosomesand a quantitative analysis has revealed a 1.8-fold and 2.8-fold inducedautophagy in 17-AAG and GA treated compared to untreated cells, re-spectively. We also report that, under the same conditions, the markerLC3 co-localizes with aggregates in some of the cells. This is consistentwith a recent study showing that 17-AAG is an effective inducer of theautophagic pathway (Ronnen et al., 2006). Therefore, aggregates degra-dation mediated by 17-AAG and GA involves both proteasome andautophagy, this latter system playing a role particularly when protea-some activity is impaired (Di Zanni et al., 2011).

It is well-known that 17-AAG, GA and curcumin treatments causeupregulation of heat shock proteins, especially HSP70 (Ali and Rattan,2006; Bachetti et al., 2007; Cheng et al., 2001; Riedel et al., 2010; Teitenet al., 2009; Waza et al., 2005). Therefore, the success of these treat-ments in our in vitro CCHSmodel is probably due to the combined effectof i) refolding of mutant proteins, promoted by HSP70 induction, ii) re-moving aggregates and iii) degrading mutant proteins, made possibleby the UPS and autophagy systems (Di Zanni et al., 2011).

In conclusion, our drug evaluation study has revealed that 17-AAGand curcumin are effective in rescuing in vitro the nuclear localizationand transactivation activity of PHOX2B carrying the largest expansionof polyAla and in promoting the clearance of aggregates of these mu-tant proteins. We have also obtained evidences that such effects arelikely mediated by the UPS, autophagy and HSP systems, which remainmajor druggable targets to ameliorate the cellular consequences of thepresence of proteins containing expanded homopolymers. 17-AAGand curcumin are presently in clinical trials for other diseases (Heathet al., 2008; Teiten et al., 2010), and we have confirmed that in ourin vitro system their cytotoxicity is quite limited. All these observa-tions make us believe that developing a therapeutic strategy for CCHSpatients is possible and a pharmacological approach may represent aconcrete option to be offered to these patients in a hopefully nearfuture.

Supplementary materials related to this article can be found on-line at doi:10.1016/j.nbd.2011.09.007.

Conflict of interest statement

The authors have no actual or potential conflict of interest to disclose.

Acknowledgments

We are grateful to Dr Luis Galietta (Gaslini Institute, Genoa, I) for thegift of the pZProSensor plasmid. The financial support of FondazioneMariani (grant# 08/69 to I. C.),Ministery of Health (Progetto Finalizzato2006 and Progetto Strategico 2009 to I. C.) and Fondazione Cariplo(grant no. 2010–0688 to D.F.) are gratefully acknowledged. We arealso thankful to the “Associazione italiana per la Sindrome da Ipoventi-lazione Centrale Congenita” for their generous support to our work.

References

Abu-Baker, A., Messaed, C., Laganiere, J., Gaspar, C., Brais, B., Rouleau, G.A., 2003. In-volvement of the ubiquitin–proteasome pathway and molecular chaperones inoculopharyngeal muscular dystrophy. Hum. Mol. Genet. 12, 2609–2623.

Adachi, M., Browne, D., Lewis, E.J., 2000. Paired-like homeodomain proteins Phox2a/Arix and Phox2b/NBPhox have similar genetic organization and independentlyregulate dopamine beta-hydroxylase gene transcription. DNA Cell Biol. 19,539–554.

Albrecht, A., Mundlos, S., 2005. The other trinucleotide repeat: polyalanine expansiondisorders. Curr. Opin. Genet. Dev. 15, 285–293.

Ali, R.E., Rattan, S.I., 2006. Curcumin's biphasic hormetic response on proteasome activ-ity and heat-shock protein synthesis in human keratinocytes. Ann. N. Y. Acad. Sci.1067, 394–399.

Amiel, J., Laudier, B., Attie-Bitach, T., Trang, H., de Pontual, L., Gener, B., et al., 2003.Polyalanine expansion and frameshift mutations of the paired-like homeoboxgene PHOX2B in congenital central hypoventilation syndrome. Nat. Genet. 33,459–461.

Bachetti, T., Matera, I., Borghini, S., Di Duca, M., Ravazzolo, R., Ceccherini, I., 2005. Dis-tinct pathogenetic mechanisms for PHOX2B associated polyalanine expansionsand frameshift mutations in congenital central hypoventilation syndrome. Hum.Mol. Genet. 14, 1815–1824.

Bachetti, T., Bocca, P., Borghini, S., Matera, I., Prigione, I., Ravazzolo, R., et al., 2007. Gel-danamycin promotes nuclear localisation and clearance of PHOX2B misfolded pro-teins containing polyalanine expansions. Int. J. Biochem. Cell Biol. 39, 327–339.

Bao, Y.P., Sarkar, S., Uyama, E., Rubinsztein, D.C., 2004. Congo red, doxycycline, andHSP70 overexpression reduce aggregate formation and cell death in cell modelsof oculopharyngeal muscular dystrophy. J. Med. Genet. 41, 47–51.

Benfante, R., Flora, A., Di Lascio, S., Cargnin, F., Longhi, R., Colombo, S., et al., 2007. Tran-scription factor PHOX2A regulates the human alpha3 nicotinic receptor subunitgene promoter. J. Biol. Chem. 282, 13290–13302.

Blondelle, S.E., Forood, B., Houghten, R.A., Pérez-Payá, E., 1997. Polyalanine-based pep-tides as models for self-associated beta-pleated-sheet complexes. Biochemistry 36,8393–8400.

Brunet, J.F., Pattyn, A., 2002. Phox2 genes— from patterning to connectivity. Curr. Opin.Genet. Dev. 12, 435–440.

Burgevin, M.C., Passat, M., Daniel, N., Capet, M., Doble, A., 1994. Congo red protectsagainst toxicity of beta-amyloid peptides on rat hippocampal neurones. Neurore-port 5, 2429–2432.

Chen, C.C., Hsieh, M.S., Hsuuw, Y.D., Huang, F.J., Chan, W.H., 2010. Hazardous effects ofcurcumin on mouse embryonic development through a mitochondria-dependentapoptotic signaling pathway. Int. J. Mol. Sci. 11, 2839–2855.

Cheng, A.L., Hsu, C.H., Lin, J.K., Hsu, M.M., Ho, Y.F., Shen, T.S., et al., 2001. Phase I clinicaltrial of curcumin, a chemopreventive agent, in patients with high-risk or pre-ma-lignant lesions. Anticancer Res. 21, 2895–2900.

Cole, G.M., Teter, B., Frautschy, S.A., 2007. Neuroprotective effects of curcumin. Adv.Exp. Med. Biol. 595, 197–212.

Coleman, M., Boros, S.J., Huseby, T.L., Brennom,W.S., 1980. Congenital central hypoven-tilation syndrome. A report of successful experience with bilateral diaphragmaticpacing. Arch. Dis. Child. 55, 901–903.

Cummings, C.J., Mancini, M.A., Antalffy, B., DeFranco, D.B., Orr, H.T., Zoghbi, H.Y., 1998.Chaperone suppression of aggregation and altered subcellular proteasome locali-zation imply protein misfolding in SCA1. Nat. Genet. 19, 148–154.

Davies, J.E., Sarkar, S., Rubinsztein, D.C., 2006. Trehalose reduces aggregate formationand delays pathology in a transgenic mouse model of oculopharyngeal musculardystrophy. Hum. Mol. Genet. 15, 23–31.

Di Zanni, E., Ceccherini, I., Bachetti, T., 2011. Toward a therapeutic strategy for polyala-nine expansions disorders: In vivo and in vitro models for drugs analysis. Eur. J.Paediatr. Neurol. 15, 449–452.

Forood, B., Pérez-Payá, E., Houghten, R.A., Blondelle, S.E., 1995. Formation of an ex-tremely stable polyalanine beta-sheet macromolecule. Biochem. Biophys. Res.Commun. 211, 7–13.

Fraser, P.E., Nguyen, J.T., Chin, D.T., Kirschner, D.A., 1992. Effects of sulfate ions on Alz-heimer beta/A4 peptide assemblies: implications for amyloid fibril-proteoglycaninteractions. J. Neurochem. 59, 1531–1540.

Gardian, G., Browne, S.E., Choi, D.K., Klivenyi, P., Gregorio, J., Kubilus, J.K., et al., 2005.Neuroprotective effects of phenylbutyrate in the N171-82Q transgenic mousemodel of Huntington's disease. J. Biol. Chem. 280, 556–563.

Hay, D.G., Sathasivam, K., Tobaben, S., Stahl, B., Marber, M., Mestril, R., et al., 2004. Pro-gressive decrease in chaperone protein levels in a mouse model of Huntington'sdisease and induction of stress proteins as a therapeutic approach. Hum. Mol.Genet. 13, 1389–1405.

518 E. Di Zanni et al. / Neurobiology of Disease 45 (2012) 508–518

Heath, E.I., Hillman, D.W., Vaishampayan, U., Sheng, S., Sarkar, F., Harper, F., et al., 2008.A phase II trial of 17-allylamino-17-demethoxygeldanamycin in patients with hor-mone-refractory metastatic prostate cancer. Clin. Cancer Res. 14, 7940–7946.

Heiser, V., Scherzinger, E., Boeddrich, A., Nordhoff, E., Lurz, R., Schugardt, N., et al., 2000.Inhibition of huntingtin fibrillogenesis by specific antibodies and small molecules:implications for Huntington's disease therapy. Proc. Natl. Acad. Sci. U.S.A. 97,6739–6744.

Jana, N.R., Tanaka, M., Wang, G., Nukina, N., 2000. Polyglutamine length-dependent in-teraction of Hsp40 and Hsp70 family chaperones with truncated N-terminal hun-tingtin: their role in suppression of aggregation and cellular toxicity. Hum. Mol.Genet. 9, 2009–2018.

Johnston, J.A., Illing, M.E., Kopito, R.R., 2002. Cytoplasmic dynein/dynactin mediates theassembly of aggresomes. Cell Motil. Cytoskeleton 53, 26–38.

Lao, C.D., Demierre,M.F., Sondak, V.K., 2006. Targeting events inmelanoma carcinogenesisfor the prevention of melanoma. Expert Rev. Anticancer Ther. 6, 1559–1568.

Lorenzo, A., Yankner, B.A., 1994. Beta-amyloid neurotoxicity requires fibril formationand is inhibited by congo red. Proc. Natl. Acad. Sci. U.S.A. 91, 12243–12247.

Matera, I., Bachetti, T., Puppo, F., Di Duca, M., Morandi, F., Casiraghi, G.M., et al., 2004.PHOX2B mutations and polyalanine expansions correlate with the severity of therespiratory phenotype and associated symptoms in both congenital and lateonset Central Hypoventilation syndrome. J. Med. Genet. 41, 373–380.

Neckers, L., Neckers, K., 2002. Heat-shock protein 90 inhibitors as novel cancer chemo-therapeutic agents. Expert Opin. Emerg. Drugs 7, 277–288.

Pattyn, A., Morin, X., Cremer, H., Goridis, C., Brunet, J.F., 1997. Expression and interac-tions of the two closely related homeobox genes Phox2a and Phox2b during neu-rogenesis. Development 24, 4065–4075.

Pattyn, A., Morin, X., Cremer, H., Goridis, C., Brunet, J.F., 1999. The homeobox genePhox2b is essential for the development of autonomic neural crest derivatives. Na-ture 399, 366–370.

Patwari, P.P., Carroll, M.S., Rand, C.M., Kumar, R., Harper, R., Weese-Mayer, D.E., 2010.Congenital central hypoventilation syndrome and the PHOX2B gene: a model ofrespiratory and autonomic dysregulation. Respir. Physiol. Neurobiol. 173, 322–335.

Ramanathan, R.K., Egorin, M.J., Eiseman, J.L., Ramalingam, S., Friedland, D., Agarwala, S.S.,et al., 2007. Phase I and pharmacodynamic study of 17-(allylamino)-17-demethoxy-geldanamycin in adult patients with refractory advanced cancers. Clin. Cancer Res.13, 1769–1774.

Riedel, M., Goldbaum, O., Schwarz, L., Schmitt, S., Richter-Landsberg, C., 2010. 17-AAGinduces cytoplasmic alpha-synuclein aggregate clearance by induction of autop-hagy. PLoS One 5, e8753.

Ronnen, E.A., Kondagunta, G.V., Ishill, N., Sweeney, S.M., Deluca, J.K., Schwartz, L., et al.,2006. A phase II trial of 17-(Allylamino)-17-demethoxygeldanamycin in patientswith papillary and clear cell renal cell carcinoma. Invest. New Drugs 24, 543–546.

Sasaki, A., Kanai, M., Kijima, K., Akaba, K., Hashimoto, M., Hasegawa, H., et al., 2003. Mo-lecular analysis of congenital central hypoventilation syndrome. Hum. Genet. 114,22–26.

Scheuermann, T., Schulz, B., Blume, A., Wahle, E., Rudolph, R., Schwarz, E., 2003. Trinu-cleotide expansions leading to an extended poly-L-alanine segment in the poly (A)binding protein PABPN1 cause fibril formation. Protein Sci. 12, 2685–2692.

Sharma, R.A., Euden, S.A., Platton, S.L., Cooke, D.N., Shafayat, A., Hewitt, H.R., et al., 2004.Phase I clinical trial of oral curcumin: biomarkers of systemic activity and compli-ance. Clin. Cancer Res. 10, 6847–6854.

Shinchuk, L.M., Sharma, D., Inouye, H., Wetzel, R., Kirschner, D.A., 2005. Poly-(L-ala-782 nine) expansions form core beta-sheets that nucleate amyloid assembly. Pro-teins 783 (61), 579–589.

Sittler, A., Lurz, R., Lueder, G., Priller, J., Lehrach, H., Hayer-Hartl, M.K., et al., 2001.Geldanamycin activates a heat shock response and inhibits huntingtin aggrega-tion in a cell culture model of Huntington's disease. Hum. Mol. Genet. 10,1307–1315.

Straus, C., Trang, H., Becquemin, M.H., Touraine, P., Similowski, T., 2010. Chemosensi-tivity recovery in Ondine's curse syndrome under treatment with desogestrel.Respir. Physiol. Neurobiol. 171, 171–174.

Tanida, I., Yamaji, T., Ueno, T., Ishiura, S., Kominami, E., Hanada, K., 2008. Considerationabout negative controls for LC3 and expression vectors for four colored fluorescentprotein-LC3 negative controls. Autophagy 4, 131–134.

Teiten, M.H., Reuter, S., Schmucker, S., Dicato, M., Diederich, M., 2009. Induction ofheat shock response by curcumin in human leukemia cells. Cancer Lett. 279,145–154.

Teiten, M.H., Gaascht, F., Eifes, S., Dicato, M., Diederich, M., 2010. Chemopreventive po-tential of curcumin in prostate cancer. Genes Nutr. 5, 61–74.

Trochet, D., Hong, S.J., Lim, J.K., Brunet, J.F., Munnich, A., Kim, K.S., et al., 2005. Molecularconsequences of PHOX2B missense, frameshift and alanine expansion mutationsleading to autonomic dysfunction. Hum. Mol. Genet. 14, 3697–3708.

Wang, Q., Mosser, D.D., Bag, J., 2005. Induction of HSP70 expression and recruitment ofHSC70 and HSP70 in the nucleus reduce aggregation of a polyalanine expansionmutant of PABPN1 in HeLa cells. Hum. Mol. Genet. 14, 3673–3684.

Wang, Q., Bag, J., 2008. Induction of expression and co-localization of heat shock poly-peptides with the polyalanine expansion mutant of poly(A)-binding protein N1after chemical stress. Biochem. Biophys. Res. Commun. 370, 11–15.

Waza, M., Adachi, H., Katsuno, M., Minamiyama, M., Sang, C., Tanaka, F., et al., 2005. 17-AAG, an Hsp90 inhibitor, ameliorates polyglutamine-mediated motor neuron de-generation. Nat. Med. 11, 1088–1095.

Weese-Mayer, D.E., Berry-Kravis, E.M., 2004. Genetics of congenital central hypoventi-lation syndrome: lessons from a seemingly orphan disease. Am. J. Respir. Crit. CareMed. 170, 16–21.