Identification of differentially expressed proteins in SH-SY5Y cells treated with resveratrol☆

Ying Wang, Zhong Dong, Hongyan Fan, Ming Chang, Guoyi Li, Linsen Hu

1Department of Neurology, First Affiliated Hospital, Jilin University, Changchun 130021, Jilin Province, China

2Department of Pharmacology, Jilin Medical College of Basic Medicine, Jilin 132011, Jilin Province, China

INTRODUCTION

Multiple factors contribute to the pathogenesis of neurodegenerative diseases[1-2]. Antioxidants and antiapoptotic factors alone are unlikely to be sufficient for the treatment of these conditions. Recently,natural bioactive compounds have caught the attention of researchers for the prophylaxis and treatment of neurodegenerative diseases. One promising candidate is the polyphenol compound resveratrol[3].

Resveratrol (3, 4′, 5-trans-trihydroxystilbene,C14H12O3), is a naturally occurring polyphenol phytoalexin[4]. This compound has been shown to have a wide range of biochemical and physiological properties,including the ability to inhibit platelet aggregation, anti-inflammatory actions,antitumor activity, and antioxidant functions[5-7]. Several reports have demonstrated that resveratrol protects SH-SY5Y cells against damage induced by neurotoxins such as 6-hydroxydopamine,which is used to generate experimental Parkinson's disease[8]. Although there is a great deal of evidence indicating that resveratrol is neuroprotective, the mechanisms are not completely understood.

In this study, a proteomic strategy was used to examine changes in protein expression induced by resveratrol in SH-SY5Y cells.

RESULTS

Two-dimensional electrophoresis protein maps and grouping

Four batches of cells from different generations (the fifth, sixth, seventh, and eighth generations) were randomly divided into resveratrol-treated groups (R1-4) and control groups (C1-4). Resveratrol-treated groups were treated with the compound at 25 μmol/L, and control groups were treated with the dimethyl sulfoxide vehicle (Table 1).

Table 1 Sample groupings and fluorescence labeling

50 μg of protein from vehicle and resveratrol-treated SH-SY5Y cells, labeled with Cy3 and Cy5 dyes, respectively, and the internal standard labeled with Cy2, were mixed and resolved on 24 cm pH 4-7 immobilized pH gradient strips. This was followed by 12.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis(SDS-PAGE). Images were captured using a Typhoon 9400 scanner at three different wavelengths(520 nm, 580 nm, and 670 nm). Figure 1A and B shows representative fluorescence images of two-dimensional gels of protein extracts from vehicle-treated SH-SY5Y cells (Cy3 image) and resveratrol-treated SH-SY5Y cells(Cy5 image). Up to 1 660 protein spots were detected in each image. Spot detection, quantification, and image matching were performed using Decyder software. The pixel volume of each spot was calculated, normalized,and compared between the two groups with the t-test,performed with Decyder software. Differentially expressed proteins were selected for further analysis when P ＜ 0.05 and the average fold change was ＜ -1.3 or ＞ 1.3. A total of 34 protein spots were ultimately selected (Figure 1C); of these, 11 protein spots were upregulated in the resveratrol-treated group, and 23 were downregulated. Among these, 4 spots were finally identified (Figure 1D). Figure 1E–H shows detailed information on spot 3, which was identified as endoplasmic reticulum oxidoreductin 1-like protein alpha (Ero1-Lα),which was downregulated in resveratrol-treated cells.

Meanwhile, three-dimensional simulation of the protein spot was performed using DeCyder (Figure 1 G and H).

Identification of the differentially expressed proteins

Following the manufacturer’s recommended protocols,the protein spots of interest were cut with the Ettan Spot Picker from the preparative gels. After digestion using the Ettan Digester and analysis by matrix-assisted laser desorption ionization time-of-flight mass spectrometry(MALDI-TOF-MS), four protein spots were unambiguously identified. three of these were downregulated (P ＜ 0.05), and one was upregulated (P ＜0.05). The spot numbers marked in Figure 1D designate the identified proteins. Table 2 provides detailed information, including spots numbers, National Center for Biotechnology Information accession numbers, the names of the identified proteins, expectation (the probability of misidentification based on mass spectral data alone),coverage (the ratio of the protein sequences covered by the matched peptides), and treatment-to-control fluorescence ratios of the four spots. A representative MALDI-TOF-MS peptide mass fingerprint spectrum of trypsin-digested spot 3 is shown in Figure 2. This spot,identified as Ero1-Lα, was decreased by resveratrol.

Western blot analysis of Ero1-Lα

Western blot analysis was performed to validate the identity of the Ero1-Lα spot. Protein from resveratroltreated cells, labeled with Cy5, was subjected to 13 cm pH 4-7 immobilized pH gradient electrophoresis followed by 12.5% SDS-PAGE. Fluorescent images were obtained with a Typhoon 9400 scanner. Figure 3A shows a two-dimensional difference gel electrophoresis(2D-DIGE) image. The area of the gel containing the spot of interest was marked and excised. Figure 3B shows the excised gel slice subjected to imaging with the Typhoon 9400 scanner. The target protein (Ero1-Lα) was successfully detected after incubation with the primary antibody and the Cy3-labeled secondary antibody.

Figure 3C shows the image obtained with the Typhoon 9400 scanner.

Figure 1 Two-dimensional difference gel electrophoresis gel (2D-DIGE) images of proteins from vehicle and resveratrol-treated cells. (A) Image of proteins from vehicle-treated cells labeled with Cy3. (B) Image of proteins from resveratrol-treated cells labeled with Cy5. (C) Differentially expressed proteins are marked with lines. (D) Proteins whose expressions were significantly changed by resveratrol and identified with the help of matrix-assisted laser desorption ionization time-of-flight mass spectrometry.(E–H) Representative. DeCyder images of part of a 2D-DIGE gel showing detail on the endoplasmic reticulum oxidoreductin 1-like protein alpha (Ero1-Lα) spot downregulated by resveratrol. The Ero1-Lα protein spot was selected out of the yellow coil. E and G: vehicle-treated; F and H: resveratrol-treated. E–F are zoomed images of the protein spots in DeCyder-BVA software, in which the Ero1-Lα protein spot was selected out of the yellow coil, and red and green for other protein spots (not identified). D indicates the corresponding positions for four spots (1-4) of interest identified with mass spectrometry in gels A and B. The same molecular weight marker was used for gels A–D.

Table 2 Protein changes in SH-SY5Y cells treated with resveratrol and identified by mass spectrometry

Figure 2 Typical matrix-assisted laser desorption ionization time-of-flight mass spectrometry peptide mass fingerprint spectrum of trypsin-digested spot 3 in Table 1.The mass fingerprint spectrum data matched to Ero1-Lα,whose expectation value was 0.019 and protein sequence coverage was 23.4%. The X-axis represents the mass-to-charge ratio (m/z), whereas the Y-axis represents relative abundance. The mass fingerprint spectrum data for the other three proteins are shown in supplementary Figure 1 online.

Figure 3 Western blot analysis of Ero1-like protein alpha(Ero1-Lα) validated the identification by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. (A) Image of gel of proteins from resveratrol-treated cells. The area of the gel containing the protein spot of interest is outlined and the protein of interest is indicated. (B) Image of the excised gel for scanning with a Typhoon 9400 scanner. (C) Ero1-Lα was successfully detected.

DISCUSSION

Resveratrol is a naturally occurring phytoalexin with multiple biological activities. Although the neuroprotective effects of resveratrol are well known, the underlying molecular mechanisms are not completely understood. In this study, with the help of 2D-DIGE and mass spectrometry, we detected and identified the proteomic changes in human SH-SY5Y cells treated with resveratrol. The human SH-SY5Y cell line, derived from a childhood neuroblastoma, is frequently used to study neurotoxicity and neuroprotection in experimental models of Parkinson’s disease[9]. From these experiments, four proteins whose expression levels were significantly changed by resveratrol treatment, namely Ero1-Lα, p21-activated kinase 1 (PAK1), Archain1(ARCN1) protein, and T cell receptor beta chain, were identified with confidence. The former three were downregulated, and the latter was upregulated. The effects of resveratrol on the expression of these proteins may be involved in mediating its neuroprotective actions.

Ero1 is a glycosylated flavoenzyme tightly associated with the lumenal face of the endoplasmic reticulum[10]and interacts with the thioredoxin-like protein disulfide isomerase in eukaryotes, initiating the transfer of oxidizing equivalents to folding proteins for protein disulfide bond formation[11]. There is only one Ero1 in yeast, meanwhile human cells contain two Ero1 paralogs,Ero1-Lα and Ero1-Lβ[12]. The expression of these isoforms appears to be regulated in different ways;Ero1-Lα expression is mainly controlled by cellular oxygen tension, and Ero1-Lβ expression is triggered mainly by the unfolded protein response[13].

Particularly in cells with heavy secretory loads, Ero1 may contribute to the generation of reactive oxygen species(ROS). Because ROS has been implicated in numerous neurodegenerative disorders[10], increased levels of Ero1 may contribute to neuronal injury. It has been shown that resveratrol reduces the production of ROS in various cell types[14].

In our study, the reduced expression of Ero1-Lα induced by resveratrol suggests that the antioxidant action of the compound may be partly due to its ability to reduce levels of Ero1-Lα. Furthermore, resveratrol may activate SIRT1,which has become a focus of neurodegenerative diseases because of its role in delaying senescence[15].

Ero1-Lα and associated proteins also modulate SIRT1 activities. Ero1-Lα expression decreased and was associated with increased SIRT1 activity when adipocytes secreted adiponectin in response to resveratrol treatment[16]. The links between resveratrol,Ero1-Lα and SIRT1 are intriguing, but require further investigation to clarify.

ARCN1 is a subunit of the coat protein I complex. It is an intracellular protein which may be involved in vesicle structure and trafficking and plays a fundamental role in delivering molecules and organelles to their proper destination for normal cellular functions[17]. ARCN1, a protein involved in intracellular trafficking, appears to play an important role in the central nervous system, as a mutation in the gene leads to Purkinje cell degeneration in mice[18]. Our results demonstrating that ARCN1 protein levels are reduced in SH-SY5Y cells treated with resveratrol provides new clues for further studies on the molecular mechanisms of action of this phytoalexin.

The p21-activated kinases (PAKs) are a family of serine/threonine kinases that are essential for crucial cellular functions, including cell morphogenesis, motility,survival, angiogenesis, gene transcription, and hormone signaling[19]. In humans, six known PAK isoforms are classified into two subfamilies based on domain structure and regulatory mechanisms[20]. One member, PAK1, is an important signal integrator in cytoskeletal regeneration, cell motility, adhesion, and tumorigenesis.

It has been reported that higher PAK1 mRNA and protein levels were detected in various human malignancies,indicating the involvement of PAK1 in the pathogenesis and clinical progression of human tumors[21-22].

Downregulation of PAK1 in liver cancer cells treated by resveratrol may reduce rampant cell proliferation[23]. In addition, suppression of PAK2 by resveratrol was involved in methylglyoxal-triggered apoptosis in osteoblasts and ethanol-induced apoptosis of ESC-B5 cells[24-25]. In this study, the level of PAK1 was decreased in SH-SY5Y cells treated with resveratrol. Thus, we speculate that resveratrol may protect neurons from apoptosis by decreasing levels of PAK1.

T cell receptors are composed of either α/β or γ/δ chain heterodimers which recognize antigen-derived peptides bound to major histocompatibility complex molecules on antigen presenting cells and initiate the immune response[26]. Previous studies have demonstrated that immune reactions and inflammation mediated by T cells are involved in neurological diseases such as multiple sclerosis and amyotrophic lateral sclerosis[27-28]. Recently,Fujii and his colleagues characterized infiltrating T cells in Japanese encephalitis virus infected brains, by analyzing the sequences of complementarity determining region 3 of the T cell receptor[29], in an effort to understand the role of T cells in inflammatory diseases of the nervous system. Although some authors suggest that inflammation in the nervous system may be neuroprotective by releasing neurotrophic factors[30], it has been reported that T cell receptor mediated inflammation may aggravate axonal loss in multiple sclerosis[27]. This suggests that inflammation in the nervous system is multifaceted, with both positive and negative consequences.

Resveratrol can inhibit T cell function[31], but in this study,resveratrol increased expression of the T cell receptor.Therefore, the possible immunomodulatory roles of resveratrol in the nervous system need further investigation.

In summary, in this study, we examined proteomic changes in SH-SY5Y cells induced by resveratrol treatment. We report for the first time that the expression of Ero1-Lα, ARCN1, and the T cell receptor beta chain are affected by resveratrol treatment in SH-SY5Y cells.

These results should provide valuable clues for further studies exploring the neuroprotective mechanisms of resveratrol.

MATERIALS AND METHODS

Design

Proteomic analysis of an in vitro experiment.

Time and setting

The experiments were conducted at the Neurology Proteomics Laboratory of First Affiliated Hospital, Jilin University from October 2009 to May 2010.

Materials

Resveratrol (3, 4’, 5-trans-trihydroxystilbene), molecular formula C14H12O3, molecular weight 228.25, was purchased from Sigma (St. Louis, MO, USA). SH-SY5Y cells were obtained from American Type Culture Collection (ATCC, Rockville, MD, USA).

Methods

SH-SY5Y cell culture and treatments

Human SH-SY5Y cells were cultured in Dulbecco’s modified Eagle’s medium (Gibco, Grand Island, NY, USA)containing 5% heat-inactivated fetal bovine serum(Si-ji-qing, Hangzhou, China) and penicillin(100 units/mL)/streptomycin (100 units/mL). The cells were grown in 25 cm2cell culture flasks at 37°C in a humidified atmosphere containing 5% CO2. Cell density for the experiment was 2×105/mL. Resveratrol (Sigma)was prepared as an 80 mmol/L stock solution in dimethyl sulfoxide (Sigma) and stored at -20°C. Just before use,the resveratrol stock was diluted to the appropriate concentration in medium without serum. The final concentration of dimethyl sulfoxide in the experiments was 0.03%. SH-SY5Y cells for resveratrol-treated groups were incubated for 48 hours prior to treatment and for 24 hours with resveratrol (25 μmol/L), and then harvested.

As a control, SH-SY5Y cells for vehicle-treated groups were exposed to the same concentration of dimethyl sulfoxide.

Analysis of differentially expressed proteins by 2D-DIGE

After harvesting cells, proteomic sample preparation,2D-DIGE image analysis, and protein identification were carried out as described by our group previously[32-33]. In brief, cells were collected and protein samples were extracted. According to the manufacturer’s instructions,the protein samples were cleaned and quantified.

Following the manufacturer’s recommended protocols,50 μg protein from resveratrol and vehicle-treated groups was labeled with CyDye fluors (Amersham Biosciences-GE Healthcare, Uppsala, Sweden), Cy3 and Cy5, respectively. Cy3 and Cy5-labeled samples, and the internal standard labeled with CyDye fluor Cy2, were mixed in rehydration solution and loaded on an immobilized pH gradient strip (24 cm, pH 4-7)(Amersham Biosciences-GE Healthcare) for isoelectric focusing on an IPGphor. After equilibration, the pH gradients strips were placed on 12.5% sodium dodecyl sulfate polyacrylamide gel in Ettan DALT Six system(Amersham Biosciences-GE Healthcare) for electrophoresis at 2 W/gel overnight. Fluorescence images were collected with a Typhoon 9400 scanner(Amersham Biosciences-GE Healthcare). Matching,quantification, and statistical analyses were carried out with DeCyder Differential in-Gel Analysis software. We used the Ettan Picker (Amersham Biosciences-GE Healthcare) for excising the interested spots from the preparative gels. After a series of operations including digestion and dehydration, equal volumes of sample and α-HCCA (Amersham Biosciences-GE Healthcare) were spotted and mixed on the MALDI-TOF-MS target slides by Ettan Spotter (Amersham Biosciences-GE Healthcare). Peptide extracts were analyzed on a MALDI-TOF Pro MS (Identification parameters are the same as those described by Zhang[32-33]). Peptide mass fingerprints were queried against the National Center for Biotechnology Information database with the search engine ProFound with Homo sapiens as the species searched. The basic requirement for identification was that the expectation value (chance of misidentification)was less than 0.05 and the coverage (the ratio of the protein sequence covered by the matched peptides) was more than 20%. The final results were further confirmed with the Swiss-Prot protein database with the search engine Mascot (us.expasy.org/sprot).

Western blot

We made western blot gels similar to gels for DIGE analysis. A 50 μg protein sample of resveratrol-treated SH-SY5Y cells was labeled with Cy5, resolved on a 13 cm pH 4–7 immobilized pH gradient strip and then subjected to 12.5% SDS-PAGE. After electrophoresis,the protein spot of interest was excised. The proteins were transferred to a polyvinylidene fluoride membrane(Amersham Biosciences-GE Healthcare). After blocking non-specific binding sites for 1 hour with 5% bovine serum albumin in Tris-buffered saline (TBS) containing 0.1% Tween-20 (Amersham Biosciences-GE Healthcare),the membrane was incubated overnight at 4°C with goat anti-Ero1-Lα antibody (1: 500; Santa-Cruz Biotechnology,Santa Cruz, CA, USA). Then the membrane was washed 3× with TBS containing 0.1% Tween-20, and incubated with rabbit anti-goat IgG-Cy3 secondary antibody(1: 2 000; Sigma) at room temperature for 20 minutes.

Acquisition of the fluorescence images was performed with a Typhoon 9400 scanner after washing the membrane 3× with TBS containing 0.1% Tween-20.

Statistical analysis

Data were statistically analyzed using SPSS 11.0 software (SPSS, Chicago, IL, USA) and were expressed as mean ± SD. Student’s t-tests in DeCyder Differential Analysis software (Amersham Biosciences-GE Healthcare) were used to analyze differences between vehicle and resveratrol-treated cells. A P value of less than 0.05 was considered statistically significant.

Author contributions:Ying Wang designed and performed all research experiments, performed data analyses, and drafted the manuscript. Zhong Dong developed the methods for processing of two-dimensional gel-based protein separation.Hongyan Fan coordinated the data analyses. Ming Chang and Guoyi Li coordinated the in vitro study. Linsen Hu initiated the study and drafted the manuscript. All authors read and approved the final manuscript

Conflicts of interest:None declared.

Supplementary information:Supplementary data associated with this article can be found in the online version, by visiting www.nrronline.org, and entering Vol. 6 No. 21, 2011 after selecting the “NRR Current Issue” button on the page.

[1]Gilgun-Sherki Y, Melamed E, Offen D. Oxidative stress induced-neurodegenerative diseases: the need antioxidants that penetrate the blood brain barrier. Neuropharmacology.2001;40(8):959-975.

[2]Sheldon AL, Robinson MB. The role of glutamate transporters in neurodegenerative diseases and potential opportunities for intervention. Neurochem Int. 2007;51(6-7):333-355.

[3]Oomen CA, Farkas E, Roman V, et al. Resveratrol preserves cerebrovascular density and cognitive function in aging mice.Front Aging Neurosci. 2009;1:4.

[4]Frémont L. Biological effects of resveratrol. Life Sci.2000;66(8):663-673.

[5]Pace-Asciak CR, Hahn S, Diamandis EP, et al. The red wine phenolics trans-resveratrol and quercetin block human platelet aggregation and eicosanoid synthesis: implications for protection against coronary heart disease. Clin Chim Acta.1995;235(2):207-219.

[6]Birrell MA, McCluskie K, Wong S, et al. Resveratrol, an extract of red wine, inhibits lipopolysaccharide induced airway neutrophilia and inflammatory mediators through an NF-kappaB-independent mechanism. FASEB J. 2005;19(7):840-841.

[7]Mehta RG, Pezzuto JM. Discovery of cancer preventive agents from natural products: from plants to prevention. Curr Oncol Rep.2002;4(6):478-486.

[8]Chao J, Li H, Cheng KW, et al. Protective effects of pinostilbene, a resveratrol methylated derivative, against 6-hydroxydopamine-induced neurotoxicity in SH-SY5Y cells. J Nutr Biochem. 2010;21(6):482-489.

[9]Xie HR, Hu LS, Li GY. SH-SY5Y human neuroblastoma cell line:in vitro cell model of dopaminergic neurons in Parkinson's disease.Chin Med J (Engl). 2010;123(8):1086-1092.

[10]Sevier CS, Kaiser CA. Ero1 and redox homeostasis in the endoplasmic reticulum. Biochim Biophys Acta.2008;1783(4):549-556.

[11]Papp E, Nardai G, Mandl J, et al. FAD oxidizes the ERO1-PDI electron transfer chain: The role of membrane integrity. Biochem Biophys Res Commun. 2005;338(2):938-945.

[12]Pollard MG, Travers KJ, Weissman JS. Ero1p: a novel and ubiquitous protein with an essential role in oxidative protein folding in the endoplasmic reticulum. Mol Cell. 1998;1(2):171-182.

[13]Gess B, Hofbauer KH, Wenqer RH, et al. The cellular oxygen tension regulates expression of the endoplasmic oxidoreductase ERO1-Lalpha. Eur J Biochem. 2003;270(10):2228-2235.

[14]Zheng Y, Liu Y, Ge J, et al. Resveratrol protects human lens epithelial cells against H2O2-induced oxidative stress by increasing catalase, SOD-1, and HO-1 expression. Mol Vis.2010;16:1467-1474.

[15]Tang BL, Chua CE. SIRT1 and neuronal diseases. Mol Aspects Med. 2008;29(3):187-200.

[16]Qiang L, Wang H, Farmer SR. Adiponectin secretion is regulated by SIRT1 and the endoplasmic reticulum oxidoreductase Ero1-L alpha. Mol Cell Biol. 2007;27(13):4698-4707.

[17]Faulstich D, Auerbach S, Orci L, et al. Architecture of coatomer:molecular characterization of delta-COP and protein interactions within the complex. J Cell Biol. 1996;135(1):53-61.

[18]Xu X, Kedlaya R, Higuchi H, et al. Mutation in Archain 1, a Subunit of COPI coatomer complex, causes diluted coat color and Purkinje cell degeneration. PLoS Genet. 2010;6(5):1-13.

[19]Bokoch GM. Biology of the p21-activated kinases. Annu Rev Biochem. 2003;72:743-781.

[20]Jaffer ZM, Chernoff J. p21-activated kinases: three more join the Pak. Int J Biochem Cell Biol. 2002;34(7):713-717.

[21]Adam L, Vadlamudi R, Mandal M, et al. Regulation of microfilament reorganization and invasiveness of breast cancer cells by kinase dead p21-activated kinase-1. J Biol Chem.2000;275(16):12041-12050.

[22]Ito M, Nishiyama H, Kawanishi H, et al. P21-activated kinase 1: A new molecular marker for intravesical recurrence after transurethral resection of bladder cancer. J Urol. 2007;178(3 Pt 1):1073-1079.

[23]Parekh P, Motiwale L, Naik N, et al. Downregulation of cyclin D1 is associated with decreased levels of p38 MAP kinases, Akt/PKB and Pak1 during chemopreventive effects of resveratrol in liver cancer cells. Exp Toxicol Pathol. 2011;63(1-2):167-173.

[24]Chan WH, Wu HJ, Shiao NH. Apoptotic signaling in methylglyoxal-treated human osteoblasts involves oxidative stress, c-Jun N-terminal kinase, caspase-3, and p21-activated kinase 2. J Cell Biochem. 2007;100(4):1056-1069.

[25]Huang LH, Shiao NH, Hsuuw YD, et al. Protective effects of resveratrol on ethanol-induced apoptosis in embryonic stem cells and disruption of embryonic development in mouse blastocysts.Toxicology. 2007;242(1-3):109-122.

[26]Davis MM, Boniface JJ, Reich Z, et al. Ligand recognition by alpha beta T cell receptors. Annu Rev Immunol. 1998;16:523-544.

[27]Wang C, Gold BG, Kaler LJ, et al. Antigen-specific therapy promotes repair of myelin and axonal damage in established EAE.J Neurochem. 2006;98(6):1817-1827.

[28]Holmøy T. T cells in amyotrophic lateral sclerosis. Eur J Neurol.2008;15(4):360-366.

[29]Fujii Y, Kitaura K, Nakamichi K, et al. Accumulation of T-cells with selected T-cell receptors in the brains of Japanese encephalitis virus-infected mice. Jpn J Infect Dis. 2008;61(1):40-48.

[30]Kerschensteiner M, Stadelmann C, Dechant G, et al.Neurotrophic cross-talk between the nervous and immune systems: implications for neurological diseases. Ann Neurol.2003;53(3):292-304.

[31]Hushmendy S, Jayakumar L, Hahn AB, et al. Select phytochemicals suppress human T-lymphocytes and mouse splenocytes suggesting their use in autoimmunity and transplantation. Nutr Res. 2009;29(8):568-578.

[32]Fernández EA, Girotti MR, López del Olmo JA, et al. Improving 2D-DIGE protein expression analysis by two-stage linear mixed models: assessing experimental effects in a melanoma cell study.Bioinformatics. 2008;24(23):2706-2712.

[33]Zhang L, Chang M, Li H, et al. Proteomic changes of PC12 cells treated with proteasomal inhibitor PSI. Brain Res.2007;1153:196-203.