Biochemical Journal

Research article

Quantitative proteomic analysis of induced pluripotent stem cells derived from a human Huntington's disease patient

Jung-Il Chae , Dong-Wook Kim , Nayeon Lee , Young-Joo Jeon , Iksoo Jeon , Jihye Kwon , Jumi Kim , Yunjo Soh , Dong-Seok Lee , Kang Seok Seo , Nag-Jin Choi , Byoung Chul Park , Sung Hyun Kang , Joohyun Ryu , Seung-Hun Oh , Dong Ah Shin , Dong Ryul Lee , Jeong Tae Do , In-Hyun Park , George Q. Daley , Jihwan Song

Abstract

HD (Huntington's disease) is a devastating neurodegenerative genetic disorder caused by abnormal expansion of CAG repeats in the HTT (huntingtin) gene. We have recently established two iPSC (induced pluripotent stem cell) lines derived from a HD patient carrying 72 CAG repeats (HD-iPSC). In order to understand the proteomic profiles of HD-iPSCs, we have performed comparative proteomic analysis among normal hESCs (human embryonic stem cells; H9), iPSCs (551-8) and HD-iPSCs at undifferentiated stages, and identified 26 up- and down-regulated proteins. Interestingly, these differentially expressed proteins are known to be involved in different biological processes, such as oxidative stress, programmed cell death and cellular oxygen-associated proteins. Among them, we found that oxidative stress-related proteins, such as SOD1 (superoxide dismutase 1) and Prx (peroxiredoxin) families are particularly affected in HD-iPSCs, implying that HD-iPSCs are highly susceptible to oxidative stress. We also found that BTF3 (basic transcription factor 3) is up-regulated in HD-iPSCs, which leads to the induction of ATM (ataxia telangiectasia mutated), followed by activation of the p53-mediated apoptotic pathway. In addition, we observed that the expression of cytoskeleton-associated proteins was significantly reduced in HD-iPSCs, implying that neuronal differentiation was also affected. Taken together, these results demonstrate that HD-iPSCs can provide a unique cellular disease model system to understand the pathogenesis and neurodegeneration mechanisms in HD, and the identified proteins from the present study may serve as potential targets for developing future HD therapeutics.

  • apoptosis
  • cytoskeleton-associated proteins
  • Huntington's disease
  • induced pluripotent stem cell (iPSC)
  • oxidative stress
  • proteomic analysis

INTRODUCTION

HD (Huntington's disease) is a devastating autosomal-dominant neurodegenerative disorder, caused by abnormal expansion of CAG (cytosine-adenine-guanine) repeats in exon 1 of the Htt (huntingtin) protein-encoding HTT gene [1]. Eventually, these abnormal expansions will be translated into poly(Q)s (polyglutamines) in the N-terminus of the Htt protein, which form aggregates in the cytoplasm and nucleus. Notably, in the brain of a HD patient, the aggregated N-terminal fragments of elongated Htt protein are located as neuronal intranuclear inclusions and dystrophic neuritis in cortex and striatum [2]. It is known that several factors are involved in pathogenesis of HD, including excitotoxicity, impaired energy metabolism and oxidative stress [3]. People carrying the HD mutation gradually develop personality changes, involuntary movements, weight loss and eventually dementia. There are currently no cures for HD.

To understand the pathogenesis and to develop therapeutics in HD, various model systems have been developed that include transgenic mice carrying the normal or mutant htt gene, fibroblasts from HD patients or immortalized neurons expressing a mutant N-terminal fragment of human HTT gene [2]. Previously, we have isolated two HD-iPSC (induced pluripotent stem cell; HD and HD2) lines derived from the skin fibroblast of a juvenile HD patient carrying 72 CAG repeats [4], which was generated by retroviral infection of four pluripotency factors {Oct4 (Octamer-binding protein 4), Sox2 [SRY (sex-determining region Y)-box 2], Klf4 (Krüppel-like factor 4) and c-Myc}. Compared with transgenic animal models or immortalized cell lines, it is believed that the iPSC model system can represent the pathology of HD better. Moreover, the pluripotent nature of HD-iPSCs can provide an unlimited supply of cells for biochemical studies, drug screening, cell therapy and so on. In order to be used in cell therapy, it will be essential that the mutation of the HTT gene in HD-iPSCs should be corrected prior to grafting. For clinical application, it will be also important to generate transgene-free reprogrammed iPSC under GMP (Good Manufacturing Practice) conditions.

In the present study, we employed a proteomic approach using two previously established HD patient-derived iPSC lines (HD-iPSCs, HD and HD2) [4] to identify and address the differences of HD-iPSCs in protein expression profiles, compared with normal hESCs (human embryonic stem cells; H9) and iPSCs (551-8) at undifferentiated stages. The proteome techniques used are very powerful tools to understand the mechanisms of biological process. However, there are only a few studies reporting proteomic analyses on iPSCs, and no study has been reported regarding HD-iPSCs. To identify and characterize the changes of proteome profiles in HD-iPSCs, when compared with those of H9 and 551-8 cells, we carried out high-throughput image analysis, followed by LC-MS/MS (liquid chromatography tandem MS) on the differentially expressed proteins in each sample. We found that the differentially expressed proteins are key regulators in oxidative stress, DNA damage and expression of cytoskeleton-associated proteins. As a result, it is likely that HD-iPSCs exhibit increased apoptotic cell death and reduced neuronal differentiation, compared with H9 and 551-8 cells. Our results indicate that HD-iPSCs are highly susceptible to oxidative stress, followed by apoptotic cell death. Therefore HD-iPSCs can serve as a unique cellular disease model system to understand the pathogenesis and neurodegeneration mechanisms in HD. Moreover, the identified proteins from the present study can provide useful clues on the potential targets for developing future HD therapeutics.

MATERIALS AND METHODS

Culture and neuronal differentiation of H9, 551-8 and HD cells

hESCs (H9), normal iPSCs (551-8) and HD-iPSCs (HD and HD2 cells) were cultured and maintained according to the methods described previously [46]. Neuronal differentiation was induced by co-culturing the cells with PA6 stromal cells [6]. To do this, undifferentiated H9, 551-8 and HD colonies were mechanically dissected and transferred on to freshly prepared PA6 cells in DM (differentiation medium)-PA6, and 4 days later, KO-SR (knock-out serum replacement) in DM-PA6 was replaced by N2 supplements. In the following ~11–13 days, definitive neural rosette-like structures containing neuroepithelial cells were formed, which were mechanically detached and transferred on to a non-sticky Petri dish for suspension culture for 6 days to form neurospheres. To differentiate into mature neurons, neurospheres were directly plated on to PLO/FN (polyornithine and fibronectin)-coated dishes and were cultured for one week in DM supplemented with 20 ng/ml BDNF (brain-derived neurotrophic factor; R&D Systems) in the absence of bFGF (basic fibroblast growth factor). To quantify the efficiency of neuronal differentiation, the number of colonies forming neural rosette-like structures out of the total number of colonies was counted. The areas containing neural rosette-like structures out of the entire areas in each colony were measured using ImageJ (http://rsbweb.nih.gov/ij/).

Immunocytochemistry

To analyse the marker expression of H9, 551-8, HD and HD2 cells, the following primary antibodies were used: anti-(human-specific nuclei) (1:200 dilution, Chemicon), anti-(human-specific mitochondria) (1:200 dilution, Chemicon), anti-(type III β-tubulin) (Tuj1) (1:500 dilution, Chemicon), anti-MAP2 (microtubule-associated protein 2; 1:200 dilution, Chemicon), anti-Prx (peroxiredoxin) 1 (1:100 dilution, Santa Cruz Biotechnology), anti-Prx2 (1:100 dilution, Santa Cruz Biotechnology), anti-Prx6 (1:100 dilution, Santa Cruz Biotechnology), anti-Cfl-1 (Cofilin-1; 1:500 dilution, Abcam), anti-Stmn-1 (Stathmin-1; 1:100 dilution, Santa Cruz Biotechnology), anti-Facn-1 (Fascin-1; 1:100 dilution, Santa Cruz Biotechnology) and anti-Sept (septin)-2 (1:100 dilution, Santa Cruz Biotechnology). Secondary antibodies used were goat anti-(mouse IgG)-conjugated Alexa Fluor® 555 (1:200 dilution, Molecular Probes), goat anti-(rabbit IgG)-conjugated Alexa Fluor® 488 (1:200 dilution, Molecular Probes) and goat anti-(mouse IgM)-conjugated Alexa Fluor® 555 (1:200 dilution, Molecular Probes). The staining patterns were examined and photographed using a confocal laser-scanning microscope imaging system (LSM510, Carl Zeiss).

Western blot analysis

Aliquots (30 μg) of protein extracts from H9 hESCs, 551-8 hiPSCs (human iPSCs), and HD- and HD2-iPSCs at undifferentiated or differentiated stages were loaded and separated by SDS/PAGE (12 and 15% gels). The proteins were then transferred on to a nitrocellulose membrane, which was blocked for 2 h at 25°C with 3% BSA in TBST [10 mM Tris/HCl (pH 7.4), 140 mM NaCl and 0.1% Tween-20] and incubated with polyclonal antibodies against SOD1 (superoxide dismutase 1), Prx1, Prx2, Prx6 (Ab Frontier), BTF3 (basic transcription factor 3; Santa Cruz Biotechnology), phospho-ATM (ataxia telangiectasia mutated), phospho-H2A.x (histone H2A.x; Millipore), phospho-p53, Bid (BH3-interacting domain death agonist; cleaved form), caspase-9, caspase-3 (cleaved form), caspase-7 (cleaved form), PARP [poly(ADP-ribose) polymerase; cleaved form; Cell Signaling Technology] and β-actin (Santa Cruz Biotechnology) at 4°C overnight. After washing with TBST, the membranes were incubated with secondary antibodies for 1 h at 37°C and visualized by enhanced chemiluminescence (Amersham Biosciences). The membranes were then scanned and the signal intensity of each band was determined using LAS 3000 (Fuji). The relative protein levels in each sample were normalized to the level of β-actin.

RNA extraction and qRT-PCR (quantitative real-time PCR)

Total RNA was isolated from H9 hESCs, 551-8 hiPSCs, HD and HD2 cells using TRIzol® reagent (Invitrogen). RNA (1 μg) was reverse-transcribed into cDNA. qRT-PCR primers were targeted against BTF and ATM. Quantification of genes was performed using SYBR Green gene expression assays (Eppendorf Realplex 2). PCR amplification was generated using gene-specific primers (Table 1). The level of target gene expression was determined by the comparative Ct method, whereby the target is normalized to endogenous β-actin. The Ct value is the cycle number at which the fluorescence level reaches threshold. The ΔCt value is determined by subtracting the Ct value of the β-actin control from the Ct value of the target gene [ΔCt=Ct(target)−Ct(β-actin)]. This relative value of target to endogenous reference is described as the fold of β-actin=2−ΔCt.

View this table:
Table 1 Primer sets used in qRT-PCR analysis

TUNEL (terminal deoxynucleotidyltransferase-mediated dUTP nick-end labelling) assay

H9, 551-8, HD and HD2 cells at undifferentiated and differentiated neuronal stages were fixed in 4% paraformaldehyde in PBS for 20 min at room temperature (25°C). After three washes in PBS for 10 min, they were stained using the In Situ Cell Death Detection Kit (Roche) according to the manufacturer's protocol. All of the samples were counterstained with DAPI (4′,6-diamidino-2-phenylindole). The number of TUNEL-positive cells were counted and processed for statistical analysis.

Computational pathway analysis

The dataset was uploaded to the MetaCore™ software (GeneGo) to elucidate regulatory pathway in HD-iPSCs. The differentially expressed proteins were mapped in the pathway using the transcription regulation algorithm in the MetaCore™ software, and the molecular relationships among genes were graphically represented.

RESULTS

Morphological characteristics and neuronal differentiation of H9, 551-8 and HD cells

In order to study the proteomic profiles of HD-iPSCs, we initially carried out comparative morphological analyses among HD-iPSCs, H9 hESCs (control) and 551-8 hiPSCs (normal iPSC control), in which different stages of neuronal differentiation were compared in each cell line simultaneously. We also included another HD-iPSC line, called HD2, which was derived from the same HD patient simultaneously when HD-iPSC was established. Since HD cells exhibited very similar properties with HD2 cells, considering the limitation of sample numbers, our initial proteomic analysis was carried out using HD cells alone, although neuronal differentiation studies, as well as validation studies using differentially expressed marker proteins were performed on both HD-iPSC lines (see below).

Figure 1(A) outlines our five-stage differentiation protocol, which involves undifferentiated cells maintained on mouse embryonic fibroblasts (Stage 1), co-culture of undifferentiated cells with PA6 stromal cells (Stage 2), and isolation of neural rosettes (Stage 3), followed by neurosphere formation in suspension culture (Stage 4). Afterwards, we further differentiated the neurospheres into mature neurons (Stage 5). Figure 1(B) shows morphological characteristics of each cell line at various different stages of neuronal differentiation, in which, although the undifferentiated stages exhibit the highest similarity, some variable morphological differences were detected in each cell line at later stages.

Figure 1 Morphological characteristics and neuronal differentiation of H9, 551-8, HD and HD2 cells

(A) Experimental scheme showing the culture conditions for maintaining undifferentiated cells, as well as the step-wise differentiation procedures into neuronal lineages. (B) Representative morphology at undifferentiated (Stage 1), neural rosette (Stage 3), neurosphere (Stage 4) and mature neuron (Stage 5) stages in H9, 551-8, HD and HD2 cells. AA, ascorbic acid; BDNF, brain-derived neurotrophic factor; bFGF, basic fibroblast growth factor; DMEM, Dulbecco's modified Eagle's medium; GMEM, Glasgow minimum essential medium; KO-SR, knock-out serum replacement; MEF, mouse embryonic fibroblasts; PLO/FN, polyornithine and fibronectin. Scale bar=100 μm.

We have recently shown that the initial neural-forming efficiency of HD and HD2 at Stage 3 was lower, compared with normal hESC (H9) and iPSC (F5) lines [7]. In the present study, when the number of colonies forming neural rosette-like structures was counted out of total number of colonies at Stage 3, we found that H9 forms the highest percentage of rosette-forming efficiency (86.60±1.83%), and that F5, HD and HD2 cells exhibit significantly reduced rosette-forming efficiency (18.28±0.81%, 32.97±1.90% and 26.82±1.57% respectively) [7]. In the present study, we observed that the percentage of rosette-forming colonies of 551-8 cells was 29.32±1.56%. Similarly, we also showed that the total areas of rosettes to be formed in each colony at Stage 3 were highest in H9 (71.37±1.67%) and were reduced in F5, HD and HD2 cells (27.32±9.29%, 42.09±8.53% and 31.61±3.56% respectively) [7]. We also observed that the percentage of rosette-forming areas in 551-8 cells was 32.61±3.322%, indicating the efficiency of initial neuronal differentiation in HD-iPSCs at Stage 3 is significantly lower than in H9 cells and slightly higher than in F5, 551-8 and HD2 cells. As for the reduced efficiency of neuronal differentiation in 551-8 or F5 cells compared with HD cells, we speculate that this might be due to the residual expression of the KLF4 transgene that was used in making the iPSCs, which was shown to control the expression of miR-371-3 [8]. It is also possible that HD-iPSC lines can be influenced by the expression of the KLF4 transgene, but their proteomic expression profiles and the subsequent validation results clearly indicate that HD-iPSC lines are intrinsically different from 551-8 or F5 cells.

In the present study, we extensively examined the efficiency of neuronal differentiation at Stage 5, in which neurospheres were attached to form neurite outgrowth. In this case, since only the selected population of neural rosette-like structures was allowed to form neurospheres, we speculated that the efficiency of mature neuron formation would be comparable unless there are significant intrinsic differences among each cell line. To compare the extent of neuronal differentiation and neurite outgrowth in each cell line, we immunostained the Stage 5 samples using an antibody against MAP2 (Figure 2A) and counted the total number of MAP2-positive cells (Figure 2B). As expected, we found that H9 forms the highest percentage of MAP2-positive neurons (83.50±1.62%). However, unlike Stage 3, 551-8 forms more MAP2-positive neurons than HD and HD2 cells (50.95±1.81%, 43.33±2.19% and 36.70±2.60% respectively) at Stage 5. No measurements were made for F5 cells. We also measured the average length of neurites in MAP2-positive neurons (Figure 2C), and found that the extent of neuronal differentiation and neurite outgrowth were significantly reduced in both HD and HD2 lines (352.88±7.97 μm and 336.99±18.06 μm respectively), compared with H9 or 551-8 cells (506.86±13.29 μm and 417.31±16.59 μm respectively). Interestingly, we observed that the extent of neuronal differentiation and neurite outgrowth in 551-8 cells was lower than in H9 cells, but higher than in HD and HD2 cells (Figures 2B and 2C).

Figure 2 Efficiency of neuronal differentiation in H9, 551-8 and HD cells at Stage 5

(A) Immunocytochemical staining using an antibody against MAP2 showing the morphological features of neuronal cells at Stage 5 in each cell. (B) Histogram showing the percentage of MAP2positive neurons, which have been counterstained with DAPI. (C) Sigma plot diagram showing the distribution and average length of neurites in MAP2-positive neurons. The significance of differences was evaluated by one-way ANOVA (SAS version 8.0). *P<0.05, **P<0.001. Scale bar=50 μm.

Proteome analysis of differentially expressed proteins among H9, 551-8 and HD cells

To determine the proteome differences of each cell line, we conducted high-resolution 2-DE (two-dimensional electrophoresis) mapping using whole proteins extracted from H9, 551-8 and HD cells at undifferentiated stages. Total proteins were separated on 2-DE gels and visualized through silver staining. Approximately, more than 2500 protein spots were mapped individually from the H9, 551-8, and HD 2-DE gels (Supplementary Figure S1 at http://www.BiochemJ.org/bj/446/bj4460359add.htm). We primarily focused on spots that showed more than 2-fold changes in H9 compared with 551-8, 551-8 compared with HD and H9 compared with HD cells, and then selected differentially expressed protein spots in each group. The selected protein spots were excised from the 2-DE gels for identification process. A total of 26 spots were analysed by LC-MS/MS, and then the identified peptides were compared with known proteins using the IPI human protein database (http://www.ebi.ac.uk/IPI/IPIhuman.html). Among the identified peptides, spots which showed a statistically significant difference (P<0.05) were selected and summarized in Table 2.

View this table:
Table 2 Identification of differentially expressed proteins in H9, 551-8 and HD cells

NEDD, neural-precursor-cell-expressed developmentally down-regulated.

As shown in Table 2, when H9 and 551-8 cells were compared, we observed that a total of 14 proteins showed different expression patterns, in which eight proteins were up-regulated and six proteins were down-regulated in H9 and 551-8 cells respectively. By contrast, when 551-8 and HD cells were compared, we found that 17 proteins were differentially expressed (Figure 3, Table 2 and Supplementary Figure S1). Figure 3(A) shows some representative protein spots that were differentially expressed, whereas Figure 3(B) shows fold changes of each candidate protein expression in each cell. In the case of HD-iPSCs, seven proteins were down-regulated and ten proteins were up-regulated when compared with 551-8 cells. When compared with H9 cells, HD cells showed 20 differentially expressed protein patterns, in which ten proteins were down-regulated and ten proteins were up-regulated. These results suggest that there are some intrinsic differences at protein levels in HD patient-derived iPSC (HD), when compared with normal hESCs (H9) and iPSCs (551-8), which will, in turn, also alter the cellular and biochemical properties of HD-iPSCs.

Figure 3 Analysis of differentially expressed protein spots among H9, 551-8 and HD cells

(A) Enlarged images showing the differentially expressed protein spots. Note the up-regulated spots (Prx1, Prx2, Prx6 and BTF3) and down-regulated spots (SOD1, GST, Gpx1, Cfl-1 and Stmn-1) in HD-iPSCs. Arrows indicate the differentially expressed protein spots in each sample. (B) Quantification of differentially expressed protein spots. Intensities [optical density (OD)/background)] of each spot were analysed and presented in a histogram. Results are means±S.E.M for at least three independent experiments. The significance of differences was evaluated by paired two-tailed Student's t test (SAS version 8.0). *P<0.05, ‡P<0.01 and †P<0.001 compared with H9 cells.

Classification and biological network analysis of the identified proteins from H9, 551-8 and HD cells

In order to characterize the differentially expressed proteins, a total of 26 proteins were classified into functional categories according to biological processes using information from Gene Ontology (http://www.geneontology.org) and UniProt (http://www.expasy.uniprot.org). As a result, they were grouped into several different categories as follows: cell death regulation (23%), oxidative and cellular stress (10%), reactive oxygen species metabolic process, catabolic process, and redox homoeostasis (20%) and others (47%) (Supplementary Figure S2 at http://www.BiochemJ.org/bj/446/bj4460359add.htm). Importantly, we found that programmed cell death, oxidative stress- and cellular oxygen-associated proteins, including SOD1 and Prx families, constitute the major portion of differentially expressed proteins in HD-iPSCs (53% of classified biological processes). These results strongly suggest that HD patient-derived iPSCs may be highly susceptible to intracellular or extracellular stresses, such as oxidative stress and apoptosis signals.

To deduce the possible relationships among the identified proteins by 2-DE analysis, we carried out biological network analysis using MetaCore™ software. After the networks were built by the shortest paths algorithm, our 26 proteins were uploaded and were mapped on the transcriptional regulation pathway (Supplementary Figure S3 at http://www.BiochemJ.org/bj/446/bj4460359add.htm). This network shows that our identified proteins may play a role in the regulation of various transcription factors, such as p53, c-Myc, E2F1, YY1 (Yin and Yang 1) and NF-κB (nuclear factor κB) in HD-iPSCs, which are known as common transcriptional factors and play important roles in the regulation of various cellular processes, including cell proliferation, differentiation and development. Although we did not detect these transcription factors directly in the 2-DE analysis, it will be likely that alterations of transcription factors can happen through the gene regulation network by identified proteins in HD-iPSCs.

Up-regulation of oxidative stress-related proteins in HD-iPSCs

To verify the 2-DE-based proteome data, we performed Western blotting using the same protein samples used in the 2-DE analysis (Figure 4). In this case, we also included protein samples from HD2 cells to confirm whether both HD and HD2 cells give rise to similar results. Among the identified proteins from the 2-DE analysis, we selected SOD1, GST (glutathione transferase), Gpx1 (glutathione peroxidase 1) and Prx families (Prx1, Prx2 and Prx 6), which are all known as representative antioxidant molecules, for further verification. This was because oxidative stress and the role of antioxidants for defence are known to be important in several degenerative diseases, including HD [9,10].

Figure 4 Expression of oxidative stress-related proteins in H9, 551-8, HD and HD2 cells

(A) Total proteins were extracted from H9, 551-8, HD and HD2 cells, and were used for Western blot analysis. Oxidative stress-related antibodies including SOD1, GST, Gpx1, Prx1, Prx2 and Prx6 were used with β-actin as a loading control. (B) Quantification of protein expression levels after normalization using β-actin. Results are means±S.E.M for three independent experiments. The significance of differences was evaluated by paired two-tailed Student's t test (SAS version 8.0). §P<0.01 and *P<0.001 compared with H9 cells; +P<0.05, ‡P<0.01 and †P<0.001 compared with 551-8 cells. (C) Immunocytochemical staining showing the expression of Prx1, Prx2 and Prx6 in each cell. (D) Histograms showing the relative expression of Prx1 in cytoplasm and nucleus. Scale bar=50 μm.

In the Western blotting analysis, we observed that SOD1, GST and Gpx1 were strongly expressed in both H9 and 551-8 cells, compared with HD and HD2 cells, and their expression levels were the lowest in both HD-iPSC lines (Figure 4A). It is known that SOD, GST and Gpx1 have roles in reducing intracellular superoxide levels, detoxifying endogenous compounds, such as peroxidized lipids, and protecting the organism from oxidative damage. Therefore reduced levels of SOD1, GST and Gpx1 in the HD-iPSC lines suggest that they are highly susceptible to oxidative stress, compared with the H9 and 551-8 cells.

On the other hand, we found that expression of Prx family members, including Prx1, Prx2 and Prx6, were up-regulated in the two HD-iPSC lines, compared with H9 and 551-8 cells. Prxs have been implicated as important indicators for cellular ROS (reactive oxygen species) signals, since they mediate antioxidant processes and control cytokine-induced peroxide levels. Therefore elevation of the cellular levels of Prxs under an oxidative stress environment can protect oxidative stress-mediated toxicity through their antioxidant activities. In our 2-DE results, we found that Prx1, Prx2 and Prx6 were highly up-regulated in HD-iPSCs compared with H9 and 551-8 cells. Consistent with these observations, the Western blotting results also showed the elevation of Prx1, Prx2 and Prx6 expression in both the HD- and HD2-iPSC lines (Figure 4A). Therefore we speculated that elevation of Prx1, Prx2 and Prx6 levels in HD-iPSC lines is a defence mechanism against cellular oxidative stress, which may have been transmitted from the specific pathological conditions of the HD patient.

Figure 4(B) shows the fold changes of each candidate protein expression, in which SOD1, GST and Gpx1 expression decreases significantly in HD-iPSC lines, whereas Prx family expression increases conversely in both the HD and HD2 lines. When compared with H9 cells, the fold changes of expression levels of each marker in HD and HD2 cells were as follows: SOD1 (0.48±0.03 and 0.34±0.02), GST (0.6±0.03 and 0.45±0.05), Gpx1 (0.27±0.02 and 0.19±0.02), Prx1 (1.93±0.12 and 1.95±0.12), Prx2 (1.76±0.12 and 1.91±0.19) and Prx6 (1.95±0.16 and 1.97±0.18).

We also performed immunocytochemical staining using antibodies against Prx1, Prx2 and Prx6, in order to examine the localization patterns of Prx proteins under specific conditions. Prx1, Prx2 and Prx6 are normally expressed in the cytoplasm, although Prx1 is also known to be expressed in the nucleus, especially when the cells are attacked by ROS [11]. From our immunocytochemical staining results, both Prx2 and Prx6 were localized in the cytoplasm, regardless of the cell type. However, in the case of Prx1, which was mainly expressed in the cytoplasm of H9 and 551-8 cells, was predominantly detected in the nucleus of HD and HD2 cells (Figures 4C and 4D). This result raises the possibility that oxidative stress in HD-iPSCs may give rise to the up-regulation of Prx1 protein, which resulted in the translocation to the nucleus [11].

In a parallel experiment, we further differentiated H9, 551-8, HD and HD2 cells into mature neurons (Figure 1B, Stage 5) and carried out Western blot and immunocytochemical analyses using markers specific for oxidative stress (SOD1, GST, Gpx1, Prx1, Prx2 and Prx6). The Western blot analysis showed their expression patterns are similar to those from undifferentiated stages (Supplementary Figure S4A at http://www.BiochemJ.org/bj/446/bj4460359add.htm). Fold change analysis further revealed that two HD-iPSC lines persistently respond to oxidative stress after neuronal differentiation, and their expression levels were adjusted to those of H9 cells. Fold changes of expression levels of marker proteins in HD and HD2 cells of differentiated stage (Stage 5) were compared with those of undifferentiated stage (Stage 1) as follows: SOD1 (HD,−0.37±0.02 compared with −0.52±0.04 and HD2, −0.51±0.02 compared with −0.65±0.02), GST (HD, −0.43±0.02 compared with −0.4±0.03 and HD2, −0.49±0.02 compared with −0.55±0.05), Gpx1 (HD, −0.53±0.01 compared with −0.73±0.02 and HD2, −0.61±0.01 compared with −0.81±0.02), Prx1 (HD, 4.37±0.3 compared with 1.92±0.12 and HD2, 4.71±0.31 compared with 1.95±0.2), Prx2 (HD, 3.52±0.17 compared with 1.76±0.12 and HD2, 4.58±0.33) and Prx6 (HD, 3.46±0.3 compared with 1.95±0.16 and HD2, 3.63±0.3 compared with 1.97±0.18) (Supplementary Figure S4B). We also carried out immunocytochemical analysis on the differentiated neurons and found that the expression levels of Prx1, Prx2 and Prx6 proteins were highly increased in HD and HD2 cells compared with H9 cells (Supplementary Figure S4C). Interestingly, we also observed that these proteins are slightly increased in 551-8 cells, suggesting oxidative stress might also affect normal iPSCs after neuronal differentiation.

Taken together from our Western blotting and immunocytochemical staining results, up-regulation of Prx1, Prx2 and Prx6 proteins and down-regulation of SOD1, GST and Gpx1 proteins in HD and HD2 cells indicate a strong induction or recruitment of factors that have protective roles against oxidative stress that may come from HD pathological conditions.

Induction of DNA damage-mediated apoptosis in HD-iPSCs

Since it is well known that apoptotic cell death can be caused by stressful conditions, including oxidative stress, we next investigated whether there are more apoptotic cells in HD-iPSC lines, compared with H9 and 551-8 cells. To address this question, we used a TUNEL assay and found that TUNEL-positive cells were significantly higher in HD (23.51±2.57%) and HD2 cells (26.57±0.80%) than in H9 (1.37±0.73%) or 551-8 (7.22±2.85%) cells at undifferentiated stages (Figure 5A). Similarly, we also found that TUNEL-positive cells were significantly higher in HD (36.14±1.61%) and HD2 cells (40.47±2.85%) than in H9 (9.49±2.45%) or 551-8 (14.96±2.33%) cells at differentiated neuronal stages (Figure 5B).

Figure 5 TUNEL assay showing the proportions of apoptotic cells in H9, 551-8, HD and HD2 cells

(A) Immunocytochemical staining of TUNEL-positive cells at undifferentiated stages. DAPI was used for counter-staining the cells. Histograms showing the relative expression of TUNEL-positive cells at undifferentiated stages. Scale bar=20 μm. (B) Immunocytochemical staining of TUNEL-positive cells at differentiated stages and their quantifications. Scale bar=50 μm. **P<0.001.

Next we investigated whether the apoptotic cell death in HD-iPSCs was caused by DSBs (double-strand breaks) or DNA fragmentation that take place under stressful conditions, including oxidative stress, metabolic disorder or genetic mutation. Among the specifically up-regulated proteins in HD-iPSCs, BTF3 [also known as NACB (nascent-polypeptide-associated complex β polypeptide)] is a well-known transcription factor, which is involved in the transcription initiation through direct binding to TATA and CAAT box sequences in the proximal promoter [12,13]. In addition, ATM was recently proposed as a target gene of BTF3. As it is well known that ATM is activated under oxidative stress and plays an important role in DNA-damage-mediated signalling [14], we speculated that, when the cells are attacked by ROS and oxidative stress, BTF3 might up-regulate ATM and then the apoptotic signal is transmitted by ATM activation to induce apoptotic changes in the cells. Having this in mind, we examined the expression levels of BTF3 and ATM transcripts in H9, 551-8, HD and HD2 cells using qRT-PCR. Interestingly, both BTF3 and ATM expression were significantly up-regulated in HD and HD2 cells compared with H9 and 551-8 cells (Figure 6A).

Figure 6 Analysis of apoptotic cell death in H9, 551-8, HD and HD2 cells

(A) qRT-PCR analyses showing relative expression levels of BTF3 (left-hand panel) and ATM (right-hand panel) transcripts. After normalization with β-actin, results are means±S.E.M for three independent experiments. (B) Western blots showing the expression of key components involved in apoptosis, such as BTF3, ATM, p53 and H2A.x. (C) Histograms showing the fold changes of their relative expression levels. (D) Western blots showing the expression of mitochondria-dependent caspase signalling components. (E) Histograms showing the fold changes of their relative expression levels. The significance of differences was evaluated by paired two-tailed Student's t test (SAS version 8.0). §P<0.01 and *P<0.001 compared with H9 cells; +P<0.05, ‡P<0.01 and †P<0.001 compared with 551-8 cells. p-, phospho-.

We also examined the translational control of key components involved in apoptosis, such as BTF3, ATM and p53 expression by Western blot analysis. According to the Western blotting results, BTF3 was significantly up-regulated both in HD and HD2 cells and its expression level was 3.29±0.15- and 3.57±0.13-fold higher than those of H9 and 551-8 cells, of which results are similar to the qRT-PCR results (i.e. 3.45±0.46 and 3.7±0.32). Since ATM is known as a downstream target of BTF3 under oxidative stress, we examined the status of phosphorylation of ATM on Ser1981 that normally occurs in response to oxidative stress-inducers such as H2O2, which is critical for sustained occupancy of ATM on DNA DSB sites [14,15]. p-ATM (S1981) [ATM phosphorylated at Ser1981] was increased in HD-iPSCs when the total proteins were reacted to an antibody against p-ATM (S1981) (Figures 6B and 6C). Auto-phosphorylation of ATM makes subsequent modification of downstream regulators, such as p53 (Ser15), MDM2 (murine double minute 2) (Ser395), Chk2 (checkpoint kinase 2; Thr68) and H2A.X (Ser139) through its activated kinase in response to DSBs [16]. For this reason, we then examined the expression levels of p53 (Ser15) and H2A.X (Ser139) and observed an increased level of phosphorylation at each specific site in the HD-iPSC lines. These results suggest that HD-iPSCs undergo oxidative-stress-mediated cellular apoptosis, which is involved with DNA damage through oxidation of DNA.

To further investigate the apoptotic processes in HD-iPSCs, we examined the expression of central regulators and effectors involved in apoptosis. As shown in Figure 6(D), cleaved Bid, which is an active form of Bid, was significantly increased. In addition, we detected high levels of active forms of caspases (cleaved forms of caspase-9, -3, and -7) from HD and HD2 cells. Finally, a DNA repair-involved protein, PARP, which is known as a marker for undergoing apoptosis, was more extensively cleaved in HD and HD2 cells (Figures 6D and 6E).

Taken together, these results strongly suggest that cellular oxidative stress in HD-iPSCs can cause DNA damage, followed by activation of consecutive ATM-mediated signalling (i.e. phosphorylation of p53 and H2A.X) via its kinase activity. Activation of substrates of ATM in HD-iPSCs can lead to DNA-damage-induced apoptotic cell death through the mitochondrial pathway.

Down-regulation of cytoskeleton-associated proteins in HD-iPSCs

We then investigated whether increased apoptotic cell death in HD-iPSCs in response to oxidative stress could affect the expression of cytoskeleton-associated proteins, including Cfl-1, Stmn-1, Facn-1 and Sept-2 at differentiated neuronal stages (Stage 5) among each cell line (H9, 551-8, HD and HD2 cells). These proteins are previously known to be strongly expressed during neuronal differentiation [1719]. To compare the expression levels, we performed a Western blot analysis and found that after the differentiation of H9, 551-8, HD and HD2 cells into neuronal lineages, although high levels of Cfl-1, Stmn-1, Facn-1 and Sept-2 proteins were detected in H9 and 551-8 cells, their expression was significantly decreased in both HD and HD2 cells (Figures 7A and 7B). Immunocytochemical staining further confirmed that these proteins are significantly down-regulated in HD and HD2 cells (Figure 7C). Taken together, these results suggest that expression of cytoskeleton-associated proteins was highly reduced in HD-iPSCs. Although it is currently unknown whether these proteins are directly involved in neuronal differentiation, it appears that neuronal differentiation was highly affected in HD-iPSCs.

Figure 7 Expression of cytoskeleton-associated proteins at the differentiated neuronal stage

(A) Western blots showing the expression levels of Cfl-1, Stmn-1, Fscn-1 and Sept-2. (B) Histograms showing the fold changes of their relative expression levels. The significance of differences was evaluated by paired two-tailed Student's t test (SAS version 8.0). §P<0.01 and *P<0.00 compared with H9 hESCs; +P<0.05, ‡P<0.01 and †P<0.001 compared with 551-8 cells. (C) Immunocytochemical staining showing the expression patterns of Cfl-1, Stmn-1, Facn-1 and Sept-2 at Stage 5 in each cell line. Scale bars=50 μm in upper images and 20 μm in lower images.

DISCUSSION

HD is a progressive and autosomal dominantly inherited neurodegenerative disorder, characterized by selective degeneration of medium spiny projection neurons in the striatum, which leads to severe impairments of motor functions, such as chorea. In addition to genetic causes, in which CAG repeats abnormally expand in HD patients, several factors are also known to be involved in the pathogenesis of HD, which include excitotoxicity, impaired energy metabolism and oxidative stress [20]. In the present study, we performed a comparative proteomic analysis using protein samples from undifferentiated H9, 551-8 and HD cells, in an attempt to identify and characterize the differentially expressed proteins in HD patient-derived iPSCs (Supplementary Figure S1). In our comparative studies, we mainly focused on the differentially expressed proteins in HD-iPSCs compared with H9 or 551-8 cells. Although more iPSC samples are needed in order to generalize our results, the present study nevertheless represents the first proteomic analysis on HD-iPSCs and will provide useful insights on the pathogenesis and neurodegeneration of HD.

According to our results, compared with normal hESCs (H9) or hiPSCs (551-8), HD-iPSCs (HD and HD2) are shown to be highly susceptible to oxidative stress, which acts as one of the major factors involved in HD pathogenesis [2123]. More recently, Parkinson's disease-related LRRK2 (leucine-rich repeat serine/threonine-protein kinase 2) mutant iPSC-derived dopaminergic neurons are also shown to have increased susceptibility to oxidative stress [24], reiterating the significance of oxidative stress in neurodegeneration. In our proteomic analysis, several oxidative stress response proteins, such as Prxs (Prx1, Prx2 and Prx6) were strongly up-regulated in HD-iPSC lines (Figures 3 and 4). Prxs reduce H2O2 and peroxynitrite through their catalytic cysteine residues, so they can protect the cells from oxidative cellular damage via their self–oxidization upon reaction with peroxide under oxidative stress conditions [25]. In keeping with our findings, it has been shown previously that Prx1, Prx2 and Prx6 were significantly induced in the stratum of HD patients [23]. Prx1, Prx2 and Prx6 are normally expressed in the cytoplasm, but we observed that Prx1 is predominantly expressed in the nucleus (Figures 4C and 4D). It is known that, whereas cytoplasmic Prx1 regulates H2O2-dependent NF-κB activation, nuclear Prx1 regulates NF-κB–DNA binding through elimination of H2O2 as a p50 subunit oxidant. Therefore, it is likely that translocated Prx1 in the nucleus under HD pathological environment can facilitate the degeneration of cells. Prx1 also acts as a H2O2 scavenger when the cells are attacked by ROS. It will be important to identify the signals which make Prx1 translocate to the nucleus in HD-iPSCs.

In contrast with Prxs, antioxidant enzymes, such as SOD1, GST and Gpx1, were shown to be down-regulated in HD-iPSCs. SOD can convert superoxide to H2O2, thereby reducing intracellular superoxide level, and the decreased SOD activity accelerates the production of ROS. In the nervous system, SOD is mainly detected in neurons at high levels, especially in the cortical layer. Therefore when the cells are damaged, SOD activity is down-regulated rapidly. In the present study, we found that SOD activity is highly decreased in HD-iPSCs, which may lead to an increase of intracellular ROS. GST is another antioxidant enzyme, which catalyses the conjugation of reduced glutathione and detoxifies the endogenous compounds, such as peroxidized lipids. Gpx catalyses the reduction of hydroperoxides to the corresponding alcohol at the expense of GSH. Gpx1 is the most abundant enzyme that is largely restricted to the cytosol, but is also present in mitochondria. Its antioxidant activity is very important in the brain, as demonstrated in Gpx1-knockout mice. We speculate that down-regulation of these antioxidant enzymes is due to the increased level of ROS, which may be related to the mutant Htt proteins in HD-iPSCs. It is likely that mutant Htt proteins can directly affect the production of ROS.

In our proteomic analysis, we also found that BTF3 is predominantly up-regulated in HD-iPSCs (Figures 3 and 6). BTF3 is a transcription factor that can activate transcription of RNA polymerase II through physiological binding to promoter regions such as the TATA and CAAT boxes [12,26]. BTF3 is also known to be involved in cell-cycle regulation and apoptosis [2729]. More recently, it is known that BTF3 can activate ATM [30], which is involved in the DNA-damage-related apoptosis pathway [16]. When DNA is damaged by genotoxic stresses such as UV, γ-irradiation or ROS, ATM can phosphorylate the tumour suppressors p53 and Chk2 [16]. Subsequently, the activated p53 via phosphorylation induces its target gene, Bax, which is involved in mitochondrial apoptotic pathway by release of Cytc (cytochrome c), followed by activation of caspases [31]. In the present study, BTF3 is strongly up-regulated in HD and HD2 cells, in response to oxidative stress/damage, and/or metabolic dysfunction, which, in turn, will trigger the induction of the DNA damage-repair protein ATM, followed by activation of the p53-mediated apoptosis pathway.

DNA damage by oxidative stress, such as ROS, leads to apoptosis via activation of ATM in mitochondria. ROS activates MAPK (mitogen-activated protein kinase) that activates both p38 and JNK (c-Jun N-terminal kinase) through phosphorylation. These subsequent activations cleave Bid, which is translocated to mitochondria by Bax and subsequently facilitates Cytc release from mitochondria [32]. Bax, a Bcl-2 protein family member, is required for translocation of tBid (truncated Bid), which can be induced through phosphorylation of p53 at Ser15 [31,33]. The interacted form as a dimer of tBid and Bax in mitochondria membrane can release Cytc. The released Cytc, coupled with a key initiator caspase, caspase-9, which acts as a proteosome, in turn further activates the effector caspases, such as caspase-3 and -7 [34]. Activation of these effector caspases results in the cleavage of a DNA-repair-involved protein, PARP, which gives rise to apoptosis via stimulation of AIF (apoptosis-inducing factor) released from mitochondria [35]. In the present study, we observed the elevated levels of cleaved Bid, cleaved caspases (caspase-9, -3 and -7) and cleaved PARP in HD and HD2 cells. Moreover, TUNEL-positive cells were significantly increased. Therefore these results demonstrate that ROS is generated in HD-iPSCs, and then the mitochondria-dependent apoptotic pathway is triggered through sequential activation of caspases.

From our analysis, we also found that the expression of cytoskeleton-associated proteins, including Cfl-1, Stmn-1, Facn-1 and Sept-2, are highly down-regulated in HD-iPSCs. Cytoskeleton is composed of actin and microtubule filaments. Actin is a major cytoskeletal protein in neurons, and the dynamics of its assembly are involved in many aspects of cell motility, vesicle transport and membrane turnover [36]. Thus disorganization of actin results in failure of neuronal differentiation. Cfl-1 plays as a key regulator, together with the ADF (actin depolymerizing factor) in actin dynamics, which is implicated in neuronal function [37]. Abnormal expression of Cfl-1 or neurodegenerative stimulation could form rod-like inclusions in neurons [38]. Stmn-1 is associated with the assembly of microtubule filaments and is highly expressed in the brain [39,40]. It is also known that Stmn-1 can destabilize microtubules through a direct interaction with tubulin [41]. Furthermore, we have previously reported that Stmn-1 is highly expressed in hESC-derived NESs (neuroectodermal spheres), which can differentiate into three neural lineages: astrocytes, oligodendrocytes and mature neurons [18]. Facn-1 has important regulatory roles in cell motility and invasion via the regulation of cytoskeletal structures [42,43]. Previous reports suggest that Facn-1 is expressed in the developing nervous system at high levels and its expression is mainly detected in medullary epithelium of human fetal and adult brains [19,44]. Sept-2 is required for the organization of actin cytoskeleton such as cellular membranes, actin filaments and microtubules [45,46]. Septins can oligomerize with other Septin family members such as Sept-4–Sept-5–Sept-8, Sept-7–Sept-9b–Sept-11 and Sept-2–Sept-6–Sept-7. Among these hetero-oligomerized complexes, only the Sept-2–Sept-6–Sept-7 complex can assemble into filaments [47,48]. It was also reported that Sept-2 is expressed in NESs at high levels [18]. According to our results, expression of cytoskeleton-associated proteins is highly affected in HD and HD2 cells (Figures 7A and 7B). Moreover, expression patterns of these proteins at differentiated neuronal stage (Stage 5) suggest that neuronal differentiation is also impaired in the HD-iPSC lines (Figure 7C). Although it is currently unknown whether these proteins are directly involved in neuronal differentiation, it is likely that oxidative-stress-induced cell death in HD-iPSCs affects the formation of cytoskeleton-associated proteins, which will, in turn, influence neuronal differentiation either directly or indirectly. Furthermore, we observed that the differentially expressed antioxidant enzymes in HD-iPSC lines at undifferentiated stages were similarly altered during neuronal differentiation (Supplementary Figure S4). It is also possible that antioxidant enzymes, such as SOD1, GST, Gpx1 and the Prx family (Prx1, Prx2 and Prx6) might cause a delay in neuronal differentiation in the HD-iPSC lines.

In summary, we have carried out a comparative proteomic analysis, and isolated and characterized the differentially expressed proteins in HD-iPSCs at undifferentiated stage. Since pathogenic features of HD only become apparent at certain ages and only in specific brain regions, mostly medium spiny projection neurons in the striatum, it is striking that significant changes are detected in undifferentiated HD-iPSCs at proteomic levels. Interestingly, we found that HD-iPSCs are highly susceptible to oxidative stress, which leads to increased apoptotic cell death. HD-iPSCs also exhibit dysregulation of cytoskeleton-associated proteins, which affects neuronal differentiation. Considering the significance of iPSC research and the increasing interests in proteome profiles of iPSCs, in particular from disease-specific iPSC lines, the present study provides the first report on the quantitative proteomic analysis of an iPSC line derived from a juvenile HD patient carrying 72 CAG repeats, which contains some useful information or insights on the pathogenesis or neurodegeneration mechanisms of HD using HD-iPSCs. As for the early onset of HD phenotypes at the proteomic level, despite no obvious cellular phenotype judged by the absence of mutant Htt protein expression (results not shown), it will be mainly due to the high numbers of CAG repeats (i.e. 72 CAG repeats). Therefore in order to draw generalized conclusions on HD-iPSC, we obviously need to extend the present study using more HD-iPSC lines, in particular those carrying different lengths of CAG repeats. In addition, since HD-iPSC carries genetic mutations, it will be essential to correct the mutant gene in order to be used for cell therapy. In this case, it will be interesting to examine whether the increased expression of oxidative stress-related proteins, etc. discovered from the present study will diminish and/or return to normal levels in proteomic analysis after gene correction of HD mutant genes in HD-iPSCs.

AUTHOR CONTRIBUTION

Jung-Il Chae contributed to conception and design of the study, and wrote the paper. Dong-Wook Kim wrote the paper and performed the Western blot and protein network analyses. Nayeon Lee performed stem cell culture and differentiation experiments, including immunocytochemical analysis. Young-Joo Jeon, Joohyun Ryu and Byoung Chul Park carried out the proteomic analysis and quantitative evaluation. Iksoo Jeon, Jihye Kwon and Jumi Kim performed the TUNEL assay and immunocytochemical analysis. Yunjo Soh, Dong-Seok Lee, Kang Seok Seo, Nag-Jin Choi and Sung Hyun Kang provided theoretical input and critical advice on the biochemical experiments. Seung-Hun Oh, Dong Ah Shin, Dong Ryul Lee and Jeong Tae Do participated in the data analysis. In-Hyun Park and George Daley originally established the 551-8, F5 and HD-iPSC lines and also provided theoretical and technical advice for stem cell experiments. Jihwan Song designed and supervised the entire research, and wrote the paper.

FUNDING

This work was supported by the Next-Generation BioGreen 21 Program [grant number PJ008116062011], the Rural Development Administration, Republic of Korea (to J.-I.C.), the Korea Food and Drug Administration [grant number S-11-04-2-SJV-993-0-H] and the Korea Health Technology R&D Project of the Ministry of Health and Welfare of the Republic of Korea [grant number A111016 (to J.S.)], the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology [grant numbers 2012R1A1A2006827 (to J.S.), 2010-0021532 (to J.-I.C.) and 2011-0026986 (to D.-W.K.)], and research funds of Chonbuk National University in 2012 (to D.-W.K.).

Acknowledgments

We thank Professor Patrik Brundin and Professor Jiayi Li (Lund University, Lund, Sweden), Dr Manho Kim, Dr Wooseok Im and Dr Hoon Ryu (Seoul National University, Seoul, Korea) and Dr Seung-Jae Lee (Konkuk University, Seoul, Korea) for useful discussions on Huntington's disease and its pathogenesis. We also thank members of the Song laboratory for useful discussion and support throughout the present study.

Abbreviations: ATM, ataxia telangiectasia mutated; Bid, BH3-interacting domain death agonist; BTF3, basic transcription factor 3; Chk2, checkpoint kinase 2; Cfl-1, Cofilin-1; Cytc, cytochrome c; DAPI, 4′,6-diamidino-2-phenylindole; 2-DE, two-dimensional electrophoresis; DM, differentiation medium; DSB, double-strand break; Facn-1, Fascin-1; Gpx1, glutathione peroxidase 1; GST, glutathione transferase; H2A.x, histone H2A.x; HD, Huntington's disease; hESC, human embryonic stem cell; hiPSC, human iPSC; Htt, huntingtin; iPSC, induced pluripotent stem cell; Klf4, Krüppel-like factor 4; LC-MS/MS, liquid chromatography tandem MS; MAP2, microtubule-associated protein 2; NES, neuroectodermal sphere; NF-κB, nuclear factor κB; PARP, poly(ADP-ribose) polymerase; p-ATM, (S1981), ATM phosphorylated at Ser1981; Prx, peroxiredoxin; qRT-PCR, quantitative real-time PCR; ROS, reactive oxygen species; Sept, septin; SOD1, superoxide dismutase 1; Stmn-1, Stathmin-1; tBid, truncated Bid; TBST, 10 mM Tris/HCl (pH 7.4), 140 mM NaCl and 0.1% Tween-20; TUNEL, terminal deoxynucleotidyltransferase-mediated dUTP nick-end labelling

References

View Abstract