Dithiol-based Compounds Maintain Expression of Antioxidant Protein Peroxiredoxin 1 That Counteracts Toxicity of Mutant Huntingtin*

Background: Decreased antioxidant activity is implicated in neurodegenerative diseases, including Huntington disease (HD). Results: Mutant Huntingtin (mHtt) triggered loss of peroxiredoxin 1 (Prx1) expression in two nerve cell models. Dimercaptopropanol treatment attenuated both mHtt-induced toxicity and loss of Prx1 expression. Conclusion: Chemical or genetic means of maintaining Prx1 expression counters mHtt toxicity. Significance: Dithiol compounds may offer new treatment options for neurodegenerative diseases. Mitochondrial dysfunction and elevated reactive oxygen species are strongly implicated in both aging and various neurodegenerative disorders, including Huntington disease (HD). Because reactive oxygen species can promote the selective oxidation of protein cysteine sulfhydryl groups to disulfide bonds we examined the spectrum of disulfide-bonded proteins that were specifically altered in a HD context. Protein extracts from PC12 cells overexpressing the amino-terminal fragment of the Huntingtin (Htt) protein with either a nonpathogenic or pathogenic polyglutamine repeat (Htt-103Q) were resolved by redox two-dimensional PAGE followed by mass spectrometry analysis. Several antioxidant proteins were identified that exhibited changes in disulfide bonding unique to Htt-103Q expressing cells. In particular, the antioxidant protein peroxiredoxin 1 (Prx1) exhibited both decreased expression and hyperoxidation in response to mutant Htt expressed in either PC12 cells or immortalized striatal cells exposed to 3-nitropropionic acid. Ectopic expression of Prx1 in PC12 cells attenuated mutant Htt-induced toxicity. In contrast, short hairpin RNA-mediated knockdown of Prx1 potentiated mHtt toxicity. Furthermore, treatment with the dithiol-based compounds dimercaptopropanol and dimercaptosuccinic acid suppressed toxicity in both HD cell models, whereas monothiol compounds were relatively ineffective. Dimercaptopropanol treatment also prevented mutant Htt-induced loss of Prx1 expression in both cell models. Our studies reveal for the first time that pathogenic Htt can affect the expression and redox state of antioxidant proteins; an event countered by specific dithiol-based compounds. These findings should provide a catalyst to explore the use of dithiol-based drugs for the treatment of neurodegenerative diseases.


Mitochondrial dysfunction and elevated reactive oxygen species are strongly implicated in both aging and various neurodegenerative disorders, including Huntington disease (HD).
Because reactive oxygen species can promote the selective oxidation of protein cysteine sulfhydryl groups to disulfide bonds we examined the spectrum of disulfide-bonded proteins that were specifically altered in a HD context. Protein extracts from PC12 cells overexpressing the amino-terminal fragment of the Huntingtin (Htt) protein with either a nonpathogenic or pathogenic polyglutamine repeat (Htt-103Q) were resolved by redox two-dimensional PAGE followed by mass spectrometry analysis. Several antioxidant proteins were identified that exhibited changes in disulfide bonding unique to Htt-103Q expressing cells. In particular, the antioxidant protein peroxiredoxin 1 (Prx1) exhibited both decreased expression and hyperoxidation in response to mutant Htt expressed in either PC12 cells or immortalized striatal cells exposed to 3-nitropropionic acid. Ectopic expression of Prx1 in PC12 cells attenuated mutant Httinduced toxicity. In contrast, short hairpin RNA-mediated knockdown of Prx1 potentiated mHtt toxicity. Furthermore, treatment with the dithiol-based compounds dimercaptopropanol and dimercaptosuccinic acid suppressed toxicity in both HD cell models, whereas monothiol compounds were relatively ineffective. Dimercaptopropanol treatment also prevented mutant Htt-induced loss of Prx1 expression in both cell models. Our studies reveal for the first time that pathogenic Htt can affect the expression and redox state of antioxidant proteins; an event countered by specific dithiol-based compounds. These findings should provide a catalyst to explore the use of dithiolbased drugs for the treatment of neurodegenerative diseases.
Huntington disease (HD) 3 is an inherited adult onset neurodegenerative disorder characterized clinically by chorea, psychiatric disturbances and dementia, and pathologically by the loss of striatal and cortical cell neurons. The disease is caused by an expansion of a CAG trinucleotide repeat region in exon 1 of the HTT gene, which encodes Huntingtin (Htt), a ubiquitously expressed protein in the brain and peripheral tissues with an uncertain molecular function (1). Individuals with HD have a CAG expansion that results in enlargement of the polyglutamine (poly(Q)) tract within the N terminus of Htt to greater than 36 residues. Longer poly(Q) stretches are associated with earlier onset of HD and more severe disease symptoms (2). The precise mechanism of HD pathophysiology is poorly defined but evidence exists that multiple neurodegenerative pathways are involved including mitochondrial impairment, oxidative stress, transcriptional dysregulation, elevated apoptosis, changes in intracellular transport, signaling dysfunction, and altered protein interactions and activity (1).
Mutant Htt (mHtt) containing a poly(Q) repeat greater than 36 has a high predisposition to misfold and disrupt normal processes essential for cellular homeostasis (3). Among these, mitochondrial dysfunction and elevated reactive oxygen species (ROS) production are strongly involved in HD progression (4). Although mitochondria produce most of the cellular ATP, they are also a major source of ROS production via electron leakage from the respiratory chain (especially complexes I and III). Several studies have shown that mHtt is found in association with the outer mitochondrial membrane in brain tissue from HD transgenic mice and in isolated mitochondria from both lymphoblasts and postmortem brain tissue from HD patients (5)(6)(7). In addition, isolated mitochondria from HD mice exhibit decreased membrane potential, increased propensity to depolarize at lower calcium loads, and elevated sensitivity to calcium-induced cytochrome c release compared with controls (5,6). Transcription of peroxisome proliferator-activated receptor, a coactivator 1␣ (PGC1␣), a key transcriptional co-activator that induces expression of genes that regulate mitochondrial respiration and oxidative stress, is repressed in mHtt-expressing neurons (8). Impaired mitochondrial respiration and ATP synthesis have been detected in postmortem brain samples from HD patients and in various HD cell and animal models (9). Collectively these findings strongly indicate that perturbed mitochondrial function contributes to HD pathogenesis.
Expression of mHtt in cultured non-neuronal or neuronal cells has been shown to increase both ROS production and toxicity, which can be rescued by treatment with the thiolbased antioxidants N-acetyl-L-cysteine (NAC) and glutathione (GSH) (10,11). Antioxidant treatment has also been shown to reduce cytotoxicity associated with the expression of mHtt in transgenic nematode and mouse models of HD (12,13). However, the exact mechanism by which antioxidants counter mHtt-induced toxicity is poorly defined.
Although ROS have traditionally been viewed as agents that cause nonspecific damage to DNA, lipids, and proteins, recent evidence has shown that ROS can act as second messengers and selectively target proteins leading to a change in their activity or function (14). The specificity of ROS-mediated signaling has become an extremely active area of research over the last decade. Cysteine represents one of the major amino acids modified by ROS in a highly selective and reversible manner. Oxidation of protein cysteine sulfhydryl groups (Cys-SH) can lead to the formation of covalent disulfide bonds (Cys-S-S-Cys) or higher oxidized species such as sulfenic (Cys-SOH), sulfinic (Cys-SO 2 H), or sulfonic (Cys-SO 3 H) acids. Functional consequences of Cys-SH oxidation include altered protein interactions, misfolding, catalytic inactivation, and decreased antioxidant capacity (15). The specificity of ROS-mediated oxidation of cysteine residues within proteins that participate in a wide array of cellular processes has made these proteins attractive targets for therapeutic manipulation in diseases strongly linked to oxidative stress (16).
One of the best examples of redox regulation is found in the peroxiredoxin family of proteins. Peroxiredoxins (Prxs) were originally characterized as abundant antioxidant proteins that contain a catalytic cysteine residue and detoxify hydrogen peroxide (H 2 O 2 ), peroxynitrate, and a range of organic hydroperoxides using reducing equivalents supplied by the thioredoxin system (17). However, recent studies have shown that Prxs participate in a wide array of cellular processes including neuronal differentiation, cell signaling, molecular chaperoning, and mitochondrial function in both a catalytic dependent and independent manner (18 -20). For example, Prxs are able to interact with and inactivate protein kinases such as JNK, c-Abl, and ASK1 in a redox-regulated manner (21).
In an effort to determine the spectrum of disulfide-bonded proteins (DSBP) that are altered in an HD context, and hence potential targets of antioxidants, we resolved protein extracts from PC12 cells that inducibly express exon 1 of the HTT gene with either a 25 (nonpathogenic) or 103 (pathogenic) poly(Q) repeat using a novel two-dimensional polyacrylamide gel electrophoresis (PAGE) technique to separate DSBP. Following mass spectrometry analysis, a number of antioxidant proteins were identified that displayed alterations in disulfide bonding only in Htt-103Q expressing cells. In particular, Prx1 was shown to exhibit a progressive decrease in expression and a concomitant increase in protein sulfonylation following induction of mHtt expression. Testing of various thiol-based antioxidants revealed that dimercaptopropanol (DMP) and the structurally related compound dimercaptosuccinic acid (DMSA) were specifically able to rescue mHtt-induced toxicity in PC12 cells, whereas monothiol reducing agents were relatively ineffective. In addition, DMP was able to protect against 3-nitropropionic acid-induced toxicity in a rodent HD striatal cell line. DMP-mediated protection correlated with the maintenance of Prx expression and suppression of Prx1 sulfonylation. These novel findings suggest that dithiol-based compounds can selectively protect against mHtt-induced toxicity.
Cell Lines-The PC12 HD nerve cell model was obtained from Dr. Erik Schweitzer (UCLA) and maintained in DMEM supplemented with 10% horse serum, 5% FBS, and 1% penicillin/streptomycin. These cells are stably transfected with a plasmid encoding the entire exon 1 of the HD gene with either a 25 (nonpathogenic) or 103 (pathogenic) poly(Q) repeat fused inframe to the coding sequence for the enhanced GFP at the carboxyl terminus (22) under control of an ecdysone-regulated promoter. Transcription of the transgene was activated by the addition of 2.5 M tebufenicide (an ecdysone analog) to the culture medium. These cell lines exhibit extensive cell death (Ͼ50%) 48 h after induction of Htt-103Q expression (22). Immortalized striatal cell lines derived from a knock-in transgenic mouse containing homozygous htt loci with a humanized exon 1 with either 7 (STHdhQ7) or 111 (STHdhQ11) poly(Q) repeats (23) were obtained from the Corriel Institute (Camden, NJ) and maintained in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin at 33°C in a humidified CO 2 incu-bator. Striatal cells were treated with 10 mM 3-nitropriopionic acid for up to 48 h.
Determination of Mitochondrial Produced ROS-Mitochondrial ROS production was visualized using the fluorescent dye Mitotracker Red CM-H2XRos (Invitrogen). PC12 Htt-25Q-GFP and Htt0103Q-GFP cells were seeded at 1 ϫ 10 5 in polylysine-treated 6-well glass bottom dishes and induced the following day with 2.5 M tebufenocide. At 24 and 48 h postinduction, medium was aspirated and cells were incubated in DMEM containing 200 nM Mitotracker Red for 20 min at 37.5°C in a humidified CO 2 incubator. The medium was removed and the cells were incubated for 5 min with PBS containing 10 g/ml of Hoechst stain to stain nuclei, followed by two washes with PBS. Phenol Red-free DMEM was added to the culture prior to visualization using a Zeiss AxioObserver A1 epifluorescent inverted microscope with a ϫ100 oil objective and equipped with appropriate filters. Pictures were taken of three random fields of view for each experiment using a Q Imaging (Retiga 1300 monochrome 10-bit) camera and Q Capture software. Mitotracker Red fluorescence was quantified with ImageJ software.
Protein Extract Preparation-For two-dimensional PAGE and one-dimensional immunoblot experiments, PC12 cells were seeded at 2.0 ϫ 10 6 cells in 10-cm dishes and transgene expression was induced 24 h after seeding with 2.5 M tebufenocide. At various time points after induction, cells were washed with cold phosphate-buffered saline (PBS), incubated for 3 min with PBS containing 40 mM iodoacetamide to prevent postlysis thiol-disulfide exchange, and scraped into 100 l of Triton X-100 extraction buffer (50 mM Tris, pH 7.5, 1% Triton X-100, 1 mM phenylmethanesulfonyl fluoride, 40 mM iodoacetamide). The cell suspension was rocked on ice for 20 min and centrifuged at 15,000 ϫ g for 10 min. Subsequently the supernatant (containing soluble proteins) was aliquoted and frozen at Ϫ80°C. Striatal cell extracts were prepared by seeding striatal cells at 2.5 ϫ 10 5 cells in 60-mm dishes and, following various treatments over the course of 48 h, were washed twice with ice-cold PBS and harvested in 100 l of Triton X-100 extraction buffer as described above. Protein content of cell extracts was determined using the colorimetric Bio-Rad DC Protein Assay Kit.
Redox Two-dimensional PAGE-Separation of DSBP by redox two-dimensional PAGE was performed as described previously (16,24). In brief, protein extracts (300 g) from Htt-25Q-GFP or Htt-103Q-GFP expressing PC12 cells were resolved for 5 h at constant current, under nonreducing conditions through a 1.5-mm thick, 12% polyacrylamide gel using a Hoefer SE600 electrophoresis apparatus (San Francisco, CA). Following electrophoresis in the first dimension, gel strips containing resolved proteins were excised and immersed in SDS sample buffer (0.05 M Tris-HCl, pH 6.8, 5% glycerol, and 2% SDS) containing 100 mM dithiothreitol (DTT) for 20 min at room temperature to reduce disulfide bonds. Gel strips were then washed with SDS-running buffer and immersed for 10 min in SDS sample buffer containing 100 mM iodoacetamide. Each gel strip was then applied horizontally to a second gel (1.5 mm thick, 12% polyacrylamide) and run for 14 h at constant current (10 mA/gel). After electrophoresis in the second dimension, gels were fixed in 50% methanol overnight and silver stained as described previously (24).
MS Identification of Disulfide-linked Proteins-Gel spots from redox two-dimensional gels were excised and in-gel digested with trypsin using a MassPREP automated digester station (PerkinElmer Life Sciences). Gel pieces were de-stained of silver using 50 mM sodium thiosulfate 5-hydrate and 15 mM potassium ferricyanide, followed by protein reduction using 10 mM DTT, alkylation using 55 mM iodoacetamide, and finally digested with trypsin. Peptides were extracted using a solution of 1% formic acid and 2% acetonitrile and lyophilized. Prior to mass spectrometry analysis, dried peptide samples were re-dissolved in 10% acetonitrile and 0.1% TFA (trifluoroacetic acid). Aliquots of digested samples were mixed with an equal volume of MALDI matrix (␣-cyano-4-hydroxycinnamic acid, prepared as 5 mg/ml in 6 mM ammonium phosphate monobasic, 50% acetonitrile, 0.1% trifluoroacetic acid) at a 1:1 ratio. Mass spectrometry data were obtained using a 4700 Proteomics Analyzer, MALDI TOF-TOF (Applied Biosystems, Foster City, CA). Data acquisition and data processing were, respectively, done using 4000 Series Explorer and Data Explorer (both from Applied Biosystems). The instrument is equipped with a 355-nm Nd:YAG laser and run in the positive reflector mode. Monoisotopic masses were determined after internal calibration and searched against the NCBInr rat protein data base using the Mascot program and a mass tolerance of 25 ppm. Each mass spectrum was collected as a sum of 1000 shots.
RNA Extraction and RT-PCR-RNA was extracted from PC12 cells using TRIzol reagent (Invitrogen) following the manufacturer's protocol. Briefly, 1 g of total RNA was reverse transcribed into cDNA using a 1st Strand cDNA Synthesis kit for RT-PCR (Quanta BioSciences Inc). PCR was performed on a Mastercycler thermocycler (Eppendorf) for 30 cycles for Prx1 and 27 cycles for GAPDH as a control. Each cycle included denaturation at 94°C for 20 s, annealing at 57°C for 45 s, and elongation at 72°C for 45 s. The primers used for Prx1 were forward, 5Ј-CCTGTAGCTCGACTCTGCTGATAG3Ј, reverse, 5Ј-GCGGCCAACAGGAAGATC-3Ј and for GAPDH were forward, 5Ј-CCTTCATTGACCTCAACTAC-3Ј, reverse, 5Ј-GGAAGGCCATGCCAGTGAGC-3Ј. The size of the PCR products was determined by 1% agarose gel electrophoresis with visualization by ethidium bromide staining.
Immunoblot Analysis-Triton X-100-soluble protein extracts (30 g) were resolved by 12% nonreducing and reducing PAGE using a Hoefer SE600 electrophoresis apparatus. Extracts resolved under reducing conditions were treated with 100 mM DTT and 2% ␤-ME and boiled for 5 min. Gels were electroblotted onto Immobilon-P membrane (Millipore, Bedford, MA), blocked with 1% milk and 3% BSA (Sigma) in Trisbuffered saline with 0.5% Tween 20 (TBST). Membranes were incubated with polyclonal rabbit antibodies against Prx1, sulfonylated Prx (Ab Frontier, South Korea), GFP (Sigma), and actin (Cell Signaling) overnight. Following incubation, membranes were washed with TBST and further hybridized with horseradish peroxidase-conjugated secondary antibodies (Bio-Rad). Detection was performed using Pierce ECL Western blotting detection reagents (ThermoScientific) and visualized using a ChemiDoc digital imaging system (Bio-Rad). Densitometric analysis was performed using Quantity One software (Bio-Rad). Data presented are the average of three separate experiments. Statistical analysis of densitometric averages was performed using a one-way ANOVA followed by a post-hoc Tukey test.
Prx1 Expression Plasmids and Transfection Procedure-Expression constructs containing FLAG-tagged versions of both wild type and cysteine mutant (C52S) Prx cDNAs were generously supplied by Dr. Hyunjung Ha (Chungbuk National University, Cheonju Korea). PC12-Htt-103Q-GFP cells were seeded in 35-mm dishes at 3 ϫ 10 5 cells/dish and co-transfected with either pcDNA or FLAG-tagged Prx expression constructs along with a mCherry Fluorescent Protein reporter plasmid at a 3:1 molar ratio for a total of 4 g of DNA per dish. Plasmid DNA was mixed with 4 l of Lipofectamine 2000 (Invitrogen) in Opti-MEM, added to PC12 cell cultures, and 6 h later the transfection medium was replaced with DMEM containing 10% horse serum and 5% FBS. The following day transfected cells were treated with 2.5 M tebufenocide to induce mHtt expression and at 48 and 72 h post-induction, cells were treated Hoechst stain (10 g/ml) in PBS for 10 min, washed twice with PBS, and trypsinized. Cells were counted using a hemocytometer by both light and epifluorescent microscopy. Cell viability was assessed by comparing the average number of Hoeschtstained, mCherry-positive cells following tebufenocide induction versus cells transfected with the same plasmid but without tebufenocide treatment. Mean cell viability ratios from three separate experiments were analyzed using a one-way ANOVA and post hoc Tukey test.
MTT and Trypan Blue Cell Viability Assays-For MTT cell viability assays PC12 Htt-103Q-GFP cells were seeded in quadruplicate at 7,000 cells/well in 96-well plates (100 l total culture volume). Thiol-based compounds were added at various concentrations (1, 5, 10, 50, 100, 500, and 1000 M) to 96-well plates 24 h after seeding. One hour after compound exposure, transgene expression was induced with 2.5 M tebufenocide. 48 h after induction of Htt-103Q-GFP expression 10 l of 2 mg/ml of MTT solution prepared in PBS was added to the culture media and incubated at 37°C for 2 h. Media was thoroughly aspirated from the wells and the purple formazan product of MTT was released from cells by the addition of 100 l of DMSO. Absorbance was measured at a test wavelength of 595 nm and reference wavelength of 655 nm using a Bio-Rad 3550 microplate reader.
Due to interference of 3-nitropropionic acid with the MTT assay, cell viability in striatal cells was measured by trypan blue dye exclusion. In brief, STHdhQ7 and STHdhQ111 cells were seeded in triplicate in 12-well plates at 2.5 ϫ 10 4 cells/well and treated with 10 mM 3-NP the following day. 24 h after 3-NP treatment, cells were detached with 100 l of trypsin solution and an equal volume of trypan blue was added. Cells were counted using a hemocytometer and cell viability was determined by comparing the number of trypan blue excluding cells in treated versus nontreated cells. Mean cell viability ratios from three separate experiments were analyzed using a one-way ANOVA and post hoc Tukey test.

Increased Mitochondrial ROS in mHtt Expressing PC12
Cells-We utilized a PC12 nerve cell model that has been stably transfected with a plasmid encoding the entire exon 1 of the HTT gene with either a 25 (nonpathogenic) or 103 (pathogenic) poly(Q) repeat fused in-frame to the coding sequence for the enhanced GFP at the carboxyl terminus (22). Transcription of the transgene is driven by an ecdysone-regulated promoter that can be turned on by the addition of ecdysone or tebufenocide (an ecdysone analog). Although several studies have shown ele-vated ROS formation in mHtt expressing mammalian cells (10,11,25), the exact site of ROS production is poorly defined. We therefore examined mitochondrial derived ROS levels in Htt expressing PC12 cells using MitoTracker Red CM-H 2 XROS, a dye that is specifically targeted to mitochondria and only fluoresces in the presence of ROS (26). Induction of Htt-25Q-GFP FIGURE 2. Redox two-dimensional PAGE analysis of protein extracts from PC12 cells expressing nonpathogenic and pathogenic Htt. Triton X-100soluble proteins were isolated from PC12 cells following tebufenocide-induced expression of an amino-terminal Huntingtin (Htt) protein containing either a nonpathogenic (25Q) or pathogenic (103Q) polyglutamine repeat. Following 2 days of induction, extracts were resolved by sequential nonreducing/reducing SDS-PAGE (redox two-dimensional PAGE) and DSBP, spots in off-diagonal zones, were visualized by silver staining. PC12 cells expressing Htt-25Q (A) showed little or no increase in DSBP following induced expression. In contrast, PC12 cells expressing Htt-103Q (B) displayed a pronounced increase in some DSBP (spots 1 and 4 -6) and reduction in others (spots 2 and 3). The numbered spots were subsequently identified by mass spectrometery (see Table 1). The gels presented are representative of three separate experiments.

TABLE 1 MS identification of DSBP altered in PC12 cells expressing mutant Huntingtin
Silver-stained DSBP that were altered in a mHtt-dependent manner in Fig. 2 were picked, in-gel digested with trypsin, and identified by MALDI-TOF mass spectrometry. The relative change in disulfide bonding (1 increased, 2 decreased) of induced versus uninduced cells is indicated. A probability based MOWSE score greater than 50 was considered significant. expression in PC12 cells resulted in diffuse green fluorescence, whereas Htt-103Q-GFP expressing cells exhibited large perinuclear protein aggregates (Fig. 1A). Elevated mitochondrial ROS were detected at 2 days post-induction only in Htt-103Q-GFP expressing cells, whereas mitochondrial ROS levels remained constant in either uninduced or induced Htt-25Q-GFP expressing cells (Fig. 1B). Therefore, expression of Htt containing a pathogenic poly(Q) repeat results in elevated mitochondrial ROS production. mHtt Expression Promotes Changes in Protein Disulfide Bonding-To determine the spectrum of proteins that specifically undergo changes in disulfide bonding in response to mHttinduced ROS production, cell lysates were analyzed by redox two-dimensional PAGE. This technique, also known as diagonal electrophoresis, allows for the separation of disulfide-linked proteins that appear as distinct spots in silver-stained gels above and below a strong diagonal line (16). Lysates were processed from cells cultured for 48 h in either the presence of tebufenocide or a vehicle control (DMSO). Redox two-dimensional PAGE analysis of protein extracts revealed a number of proteins that exhibited both increased and decreased disulfide bonding unique to Htt-103Q expressing cells (Fig. 2). Specifically, spots 2 and 3 exhibited a diminishment in disulfide bonding, whereas spots 1 and 4 -8 displayed increased disulfide bonding. In contrast, minimal alterations were observed in DSBP following induction of nonpathogenic Htt-25Q expression. To ensure that the redox two-dimensional PAGE technique did not artificially induce disulfide bonding, protein extracts were treated in both dimensions with the reducing agent DTT. As expected, a clear diagonal line was revealed with no spots in the off-diagonal zones (supplemental Fig. 1).

Identification of Proteins That Undergo Changes in Disulfide
Bonding in Response to mHtt Expression-Protein spots that exhibited reproducible migration following redox two-dimensional PAGE were excised, in-gel digested with trypsin, and identified by matrix-assisted laser desorption/ionization timeof-flight (MALDI-TOF) mass spectrometry. Analysis of the protein spots revealed that half of the proteins identified (Table  1) were categorized as antioxidant proteins. Specifically the peroxiredoxin (Prx) family of antioxidant proteins Prx1, -2, and -4 were identified as proteins that exhibited changes in disulfide bonding following mHtt expression. In addition, the antioxidant Cu/Zn-superoxide dismutase was shown to have increased disulfide bonding. The four other proteins that were identified (Table 1 and Fig. 2, spots 5-7) have various intracellular roles including molecular chaperoning, cytoskeletal remodeling, and vesicular trafficking.
mHtt Promotes Decrease in Prx1 Expression-Prx1 is an abundant intracellular protein within mammalian cells and plays an essential role in protecting against oxidative stress (17). In this study, Prx1 also exhibited the most robust decrease in disulfide bonding following redox two-dimensional PAGE analysis. Thus, we chose to examine Prx1 expression in the PC12 cell model using one-dimensional immunoblot analysis under both nonreducing and reducing conditions. Under normal cellular conditions, Prx1 predominately exists as a disulfide-linked homodimer. During its normal catalytic cycle, the homodimeric form of Prx1 can be recycled back to its mono-meric form via thioredoxin (17). However, exposure to high intracellular ROS can overoxidize the catalytic cysteine residue of Prx1 to a sulfonic acid thereby rendering the protein unable to form a disulfide bond (27). Immunoblot analysis of nonreduced protein extracts revealed a decrease in the homodimeric disulfide-linked form of Prx1 at the 2-and 3-day time points following induction of mHtt expression (Fig. 3A). Also apparent under nonreduced conditions was a monomeric form of Prx1 (Fig. 3A). Blotting using an antibody that specifically recognizes sulfinic and sulfonic acid derivatives of Prxs (Prx-SO 2-3 ) revealed that the monomeric form of Prx1 was indeed a hyperoxidized form of the protein (Fig. 3A). In contrast, Prx1 expression levels and the redox state remained unaltered in cells expressing Htt-25Q. RT-PCR analysis revealed that levels of the Prx1 mRNA transcript decreased dramatically 48 h after induction of mHtt expression (Fig. 3B). To rule out the possibility that decreased Prx1 expression following mHtt induction was simply due to nonspecific degradation of protein or mRNA in dying cells, a live-dead cell assay was performed on the adherent cells used in expression analysis studies. As expected, an overall decrease in the total number of cells was observed following induction of mHtt expression compared with non- treated cells. However, under induced conditions the remaining adherent cells showed greater that 90% viability for up to 3 days (supplemental Fig. 2). These findings indicate that mHtt specifically promotes the loss of Prx1 expression in PC12 cells.
Overexpression of Wild-type Prx1 Suppresses mHtt-induced Toxicity-Because mHtt expression promotes the loss of endogenous Prx1, it was asked if ectopic expression of Prx1 can attenuate mHtt-induced toxicity. PC12 cells were transfected with cDNA expression constructs that allow high level expression of FLAG-tagged versions of either wild type or a catalytically inactive version (C52S) of Prx1. Unlike endogenous Prx1, levels of ectopically expressed wild type or mutant Prx1 were maintained following induction of mHtt expression over the course of 3 days (Fig. 4A). Moreover, cell viability analysis revealed that elevated expression of wild type Prx1 attenuated mHtt-induced toxicity at 48 and 72 h (Fig. 4, B and C). In contrast, mutant Prx1 exhibited only modest protection against mHtt-induced toxicity at 48 h (Fig. 4B) and had no protective effect at 72 h (Fig. 4C). These findings indicate that maintenance of Prx1 levels can protect against toxicity initiated by mHtt expression.
Dimercaptopropanol Suppresses mHtt-induced Toxicity, Prx1 Hyperoxidation, and Loss of Prx1 Expression-Because mHtt expression was shown to induce both loss of Prx1 expression and cell death in PC12 cells it was asked if thiol-based antioxidants could counter these effects. Various thiol compounds were examined for their ability to counter mHtt toxicity including ␤-ME, DTT, cysteamine (Cys), NAC, reduced eeGSH, DMP, DMSA, and penicillimine. DMP, and the related compound DMSA, exhibited the most robust protection (greater than 60% viability) against mHtt-induced toxicity in PC12 cells (Fig. 5). In contrast, all the other thiol-based compounds either exhibited very modest protection (i.e. DTT) or had no effect on cell viability at the concentrations tested. Because DMP may have antioxidant effects, we examined mitochondrial ROS production in Htt-103Q expressing PC12 cells. DMP treatment failed to attenuate mitochondrial ROS production in mHtt expressing cells (supplemental Fig. 3). Therefore, DMP is unlikely to confer neuroprotection against mHtt toxicity by nonspecific antioxidant effects.
We next sought to determine whether the neuroprotective properties of DMP also correlated with suppression of altered disulfide bonding initiated by mHtt expression using redox two-dimensional PAGE. Exposure of Htt-103Q expressing cells to 100 M DMP, a concentration that is strongly neuroprotective, resulted in minimal changes in disulfide bonding following induction of mHtt expression (Fig. 6A). In particular, the mHttinduced loss of disulfide-linked homodimers of peroxiredoxin 1 and 2 (spots 3 and 2, respectively, in Fig. 2) was completely prevented by pretreatment with DMP. RT-PCR and immunoblot analysis revealed that DMP pretreatment maintained both Prx1 mRNA and protein levels following induction of mHtt expression (Fig. 6, B and C).
Recent studies have indicated that hyperoxidation (sulfonylation) of peroxiredoxins not only leads to their catalytic inactivation but also renders the proteins more susceptible to degradation (28,29). Immunoblot analysis using an anti-Prx-SO 2-3 antibody revealed that induction of mHtt expression resulted in increased sulfonylation of Prx1 relative to total Prx1 levels (Fig.  6C). In contrast, DMP pretreatment prevented both mHtt-induced loss of Prx1 expression and sulfonylation.
Knockdown of Endogenous Prx1 Potentiates mHtt Toxicity-To further determine the ability of Prx1 to attenuate mHtt toxicity, endogenous Prx1 expression was knocked down using shRNAs specific to Prx1. Western blot analysis revealed an ϳ65% knockdown in Prx1 protein levels in PC12-Htt-103Q cells stably transfected with two shRNA expressing constructs targeting different regions of the Prx1 mRNA compared with control shRNA-transfected cells (Fig. 7A). A significant decrease in viability was observed in Prx1-shRNA expressing cells compared with the control cells following induction of mHtt expression at both 24 and 48 h (Fig. 7, B and C). Furthermore, DMP treatment was unable to fully restore viability in Prx1-shRNA expressing cells when compared with control cells. These findings suggest that DMP-mediated suppression of mHtt toxicity works, in part, by maintaining expression of Prx1.

DMP Prevents 3-NP-induced Cell Death and Prx1
Hyperoxidation in a Striatal Cell Model of HD-To determine whether DMP exerted similar protective effects in another cell model of HD, we examined immortalized striatal cell lines derived from a knock-in transgenic mouse in which either 7 glutamine repeats (STHdHQ7) or 111 glutamine repeats (STHdHQ111) were introduced into exon 1 of the mouse Htt gene (23). Treatment with 3-NP, a mitochondrial complex II toxin, induces mitochondrial depolarization and cell death at significantly higher levels in STHdhQ111 cells compared with STHdhQ7 cells (30). Pre-treatment of STHdhQ111 cells with 10 M DMP before 3-NP exposure restored cell viability to a level comparable with STHdhQ7 cells treated with 3-NP alone (Fig. 8A). Immunoblot analysis revealed that 3-NP treatment promoted a dramatic loss of Prx1 expression in STHdHQ111 cells but had little effect on Prx1 levels in 3-NP-treated STHdHQ7 cells. In addition, increased sulfonylation of Prx1 and another Prx isoform of ϳ22 kDa was observed in both STHdhQ7 and STHdhQ111 cells following 3-NP treatment (Fig. 8B). However, the degree or Prx1 oxidation, relative to total Prx1 levels, was higher in 3-NPtreated STHdhQ111 cells compared with STHdhQ7 cells (Fig.  8C). Moreover, DMP pre-treatment prevented Prx1 sulfonylation and the overall diminishment of Prx1 protein levels in STHdhQ111 cells following 3-NP exposure. Therefore, DMP treatment can prevent loss of Prx1 expression, Prx1 hyperoxidation, and cell death associated with either elevated mHtt expression or exposure to a mitochondrial toxin that exacerbates HD pathology.

DISCUSSION
ROS have traditionally been viewed as deleterious molecules that nonspecifically damage cellular macromolecules and contribute to various forms of neurodegeneration and aging. Elevated oxidation of DNA, protein, and lipids has frequently been detected in postmortem brain tissue from patients afflicted with various neurodegenerative diseases. Whether the elevation in oxidative damage associated with disease states is a causal or correlative event has long been the subject of debate. However, over the last decade numerous studies have shown that ROS-mediated oxidation of cysteine residues acts as a key signaling mechanism controlling multiple cellular processes including protein phosphorylation, gene expression, chromatin remodeling, protein turnover, molecular chaperoning, metabolism, autophagy, antioxidant defense, and mitochondrial function in a specific and reversible manner (31). In fact, it is now recognized that under physiological conditions transient low level generation of ROS appears to be essential to maintain cellular homeostasis (32). However, relatively few studies have attempted to detect oxidative modifications to cysteine residues in a disease context. One study revealed that glucose-6phosphate dehydrogenase, a key enzyme that generates NADPH reducing equivalents in cells, is elevated in Alzheimer disease brain along with a concomitant increase in reactive sulfhydryls in the neuronal cytoplasm (33). However, the identity of proteins with reactive sulfhydryl groups that are altered in neurodegenerative diseases is poorly defined.
Here we show for the first time that increased expression of pathogenic Htt promotes aberrant disulfide bonding in nerve cell lines. In particular, mHtt expression promotes increased disulfide bonding of the antioxidant proteins Cu/Zn-superox- ide dismutase 1 (SOD1) and Prx4 while decreasing disulfide bonding of Prx1 and Prx2. SOD1 is a cytosolic protein responsible for the enzymatic conversion of the superoxide anion to hydrogen peroxide. Prx1 and Prx2 are predominately cytosolic, whereas Prx4 is a secreted peroxiredoxin isoform that is also found in the endoplasmic reticulum. All three Prx isoforms belong to the typical 2-cysteine family of peroxiredoxins that undergo a catalytic cycle in which the N-terminal cysteine is oxidized by H 2 O 2 to a sulfenic acid that then reacts with the C-terminal cysteine of another subunit to produce an intermolecular disulfide (17). This disulfide is then reduced by the Trx/ TrxR system, completing the catalytic cycle. During oxidative stress the sulfenic intermediate is hyperoxidized to a sulfinic acid or a sulfonic acid, which is incapable of forming a disulfide bond leading to inactivation of Prx peroxidase activity (27). Oxidation of the catalytic cysteine to a sulfinic acid in Prxs 1-4 is reversed by a reaction catalyzed by sulfiredoxin (34). However, overoxidation of all Prx isoforms to a sulfonic state is irreversible.
In this study, we observed that mHtt expression promoted a decrease in Prx1 expression and a concurrent increase in Prx1 sulfonylation. Hyperoxidation of Prxs has been show to render these proteins susceptible to ubiquitin-mediated proteolysis (28,29). Therefore, mHtt may affect Prx1 expression at both transcriptional and post-translational levels. In contrast, DMP treatment prevented both mHtt-induced loss of Prx1 expression and increased sulfonylation. Prx1 expression is controlled by the transcription factor NF-E2 related factor-2 (Nrf2) (35). Elevated ROS can promote oxidative modifications to Kelchlike ECH-associated protein 1 (Keap1), a negative regulator of  Ͻ 0.05, n ϭ 3). B, Western blot analysis of protein extracts from striatal cells revealed that 3-NP treatment promoted both loss of Prx1 expression and increased sulfonylation in STHdHQ111 cells but not in STHdHQ7 cells. DMP pre-treatment prevented 3-NP-mediated loss of Prx1 expression and elevated sulfonylation. C, densitometric analysis of Western blots revealed a significantly higher ratio (*, p Ͻ 0.05, n ϭ 3) of sulfonylated Prx1 to total Prx1 following a 48-h 3-NP treatment in STHdHQ111 cells. In contrast, the sulfonylated to total Prx1 ratio in DMP-treated STHdHQ111 cells was markedly lower following 3-NP exposure.
Nrf2, leading to the stabilization and nuclear translocalization of Nrf2 (36). Once in the nucleus, Nrf2 dimerizes with a small Maf protein and the heterodimer then binds antioxidant response element sequences within the promoters of antioxidant and phase II detoxifying genes. Besides activating transcription of Prx1, Nrf2 can also activate expression of sulfiredoxin, an enzyme that can reverse oxidation of Prx sulfinic acids to sulfhydryl groups and various other antioxidant enzymes (37). The observation that DMP exposure can partially rescue mHtt-induced toxicity even when Prx1 expression is knocked down (Fig. 7) suggests that this compound is likely to alter expression of other neuroprotective genes. It is possible that DMP may affect Nrf2 activation via a redox-dependent mechanism leading to increased expression of Prx1, sulfiredoxin, and other antioxidant enzymes even in the presence of mHtt. Future studies will evaluate this hypothesis.
We have previously shown that Prx1 levels are elevated and overoxidized Prx2 isoforms predominate in Alzheimer diseased brain tissue (38). Overexpression of Prx1 alone can protect against amyloid-␤ toxicity in nerve cells. Furthermore, amyloid-␤-resistant nerve cells up-regulate multiple Prx isoforms, in addition to the reductive enzymes required to maintain Prxs in an active state, thereby countering toxicity initiated by amyloid-␤ exposure (38). A recent proteomic study showed that the Prx isoforms 1, 2, and 6 are elevated in the striatum and cortex of HD patients compared with unaffected individuals (39). However, this study also detected the accumulation of acidic (i.e. sulfonylated) forms of Prx1 and Prx6 in HD patients as well. Collectively these findings indicate that maintenance of Prxs at a high level and in an active state is key to ensuring optimal neuronal cell survival in the presence of pathogenic proteins associated with various neurodegenerative diseases.
The observation that oxidative damage to macromolecules is increased in HD has spurred interest in evaluating the ability of various antioxidants to attenuate mHtt-induced toxicity. A number of studies have focused on the use of the molecule Coenzyme Q10 (CoQ10) for the treatment of HD (40). CoQ10 functions in the inner membrane of mitochondria as an electron carrier from enzyme complex I and complex II to complex III and can also act as a lipid soluble antioxidant, particularly in its reduced form (ubiquinol, CoQ10H2). CoQ10 has been shown to significantly increase survival and delay motor symptoms in transgenic HD mouse models (13). However, clinical trials testing various formulations of CoQ10 in HD patients have, so far, failed to fully recapitulate the positive effects of this compound observed in animal models (41,42). Whether CoQ10 is able to completely target affected brain regions in HD patients and either increase mitochondrial function or detoxify elevated ROS is uncertain. The modest effect of this antioxidant in HD patients may be related to its relative lack of specificity in targeting oxidized intracellular targets. Although, CoQ10 can detoxify lipid peroxides, it does so to varying extents in different tissues and in a relatively nonselective manner. Furthermore, it is now recognized that under physiological conditions, the transient generation of ROS, within certain limits, appears to be essential to maintain cellular homeostasis, whereas countering low level ROS with broad-acting antioxidants can actually be detrimental (32). For example, administration of antioxidants has been shown to inhibit autophagy and increase toxicity of poly(Q) proteins (43). Therefore, identification of compounds that possess good bioavailability and display some degree of specificity in targeting oxidized proteins may be key factors for the treatment of neurodegenerative diseases.
In this study we demonstrate for the first time that specific dithiol-based compounds confer selective resistance to mHttinduced neurotoxicity. Although many commonly used monothiol antioxidants such as ␤-ME, NAC, and eeGSH were tested in the mHtt-inducible PC12 cell model, only the dithiols DMP, DMSA, and to a lesser extent DTT, showed neuroprotection. This observation suggests that the protective effect of DMP cannot fully be attributed to the nonspecific and widespread reduction in intracellular disulfide bonding. DMP was previously shown in an unbiased blind screen of over 1000 FDA approved compounds to be neuroprotective against mHtt toxicity in the same PC12 model used in this study (22). In addition, DMP, also known as British anti-Lewisite (BAL), was shown to reduce pathophysiological symptoms of HD and halt disease progression in a long term study of two patients conducted in 1955 (44). At the time DMP was believed to work by chelating metals but subsequent studies using the clinically more tolerable metal chelator penicillamine failed to alleviate HD clinical symptoms (45). It is intriguing to note that penacillamine did not confer protection against mHtt expression in PC12 cells (Fig. 5) suggesting that the protective effect of DMP is unlikely to be mediated by metal chelation. Although DMP exhibited the highest level of protection against mHtt toxicity, it has poor bioavailability properties and requires intramuscular injections. DMSA also exhibited protection in this study, albeit to a lesser extent than DMP. DMSA is orally active and can cross the blood-brain barrier (46,47) making it an excellent candidate drug for further testing as a therapy for HD. Moreover, the structural relationship between the dithiols DMP and DMSA and their strong neuroprotective effects in HD cell culture models suggest that these molecules may be a good starting point for further development of dithiol-based compounds to treat HD and other neurodegenerative disorders.