ARTICLE pubs.acs.org/jpr
Quantitative Proteomics Reveal up-regulated Protein Expression of the SET Complex Associated with Hepatocellular Carcinoma Chen Li,†,‡ Hong-Qiang Ruan,†,‡ Yan-Sheng Liu,†,‡ Meng-Jie Xu,† Jie Dai,† Quan-Hu Sheng,† Ye-Xiong Tan,§ Zhen-Zhen Yao,|| Hong-Yang Wang,§ Jia-Rui Wu,† and Rong Zeng*,† †
)
Key Laboratory of Systems Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China § Eastern Hepatobiliary Surgery Hospital, No. 225, Changhai Road, Shanghai 200438, China Department of Biochemistry & Molecular Biology, Second Military Medical University, Shanghai 200438, China
bS Supporting Information ABSTRACT: We combined culture-derived isotope tags (CDITs) with twodimensional liquid chromatographytandem mass spectrometry (2DLCMS/MS) to extensively survey abnormal protein expression associated with hepatocellular carcinoma (HCC) in clinical tissues. This approach yielded an in-depth quantitated proteome of 1360 proteins. Importantly, 267 proteins were significantly regulated with a fold-change of at least 1.5. The proteins upregulated in HCC tissues are involved in regulatory processes, such as the granzyme A-mediated apoptosis pathway (The GzmA pathway). The SET complex, a central component in the GzmA pathway, was significantly upregulated in HCC tissue. The elevated expressions of all of the SET complex components were validated by Western blotting. Among them, ANP32A and APEX1 were further investigated by immunohistochemistry staining using tissue microarrays (59 cases), confirming their overexpression in tumors. The up-regulation and nuclear accumulations of APEX1 was associated not only with HCC malignancy but also with HCC differentiation in 96 clinical HCC cases. Our work provided a systematic and quantitative analysis and demonstrated key changes in clinical HCC tissues. These proteomic signatures could help to unveil the underlying mechanisms of hepatocarcinogenesis and may be useful for the discovery of candidate biomarkers. KEYWORDS: hepatocellular carcinoma, clinical tissues, quantitative proteomics, biomarkers, the SET complex
’ INTRODUCTION Hepatocellular carcinoma (HCC) is a heterogeneous cancer that currently has no promising treatment; it affects people worldwide. According to epidemiological studies, hepatitis B virus (HBV), hepatitis C virus (HCV), aflatoxin and ethanol are well-established risk factors.13 Among these risk factors, HBV infection has an unequivocal causal relationship to the subsequent development of HCC.1 However, the natural course of HCC and its molecular pathogenesis are not clearly understood, partly due to its highly variable nature. Moreover, the diverse etiology, lack of markers for early diagnosis and high mortality/ morbidity associated with HCC present challenges for its diagnosis and treatment.3 Proteomics is currently considered to be the most promising technology in the global analysis of the abnormal regulatory processes involved in cancer and the search for clinically useful protein biomarkers for the early detection, diagnosis and prognosis of cancer, and for monitoring the response to therapy.4,5 Proteomic analysis has already been applied to tissue samples from HCC patients,611 including by our group.1215 Mass spectrometry (MS)-based labeling approaches can accurately measure relative proteome changes and have become r 2011 American Chemical Society
the favored quantitative proteomics technique.16 Stable isotope labeling by amino acids in cell culture (SILAC)17 is an accurate, robust, and very sensitive technique to simultaneously identify and quantify thousands of proteins in vivo.18 In 2005, SILAC technique was expanded into tissues by incorporation of culturederived isotope tags (CDITs).19 This precise proteomic method can identify and accurately compare thousands of proteins in different clinical samples, thus providing an opportunity for direct identification of novel biomarkers.18 Applying CDITs in the case of HCC, we employed a 2DLCMS/MS online system combining strong cation exchange chromatography (SCX) fractionation and reversed phase chromatography (RP) fractionation with high-accuracy mass spectrometry to accurately quantify the proteome in clinical HCC tissues. We used 13C-labeled hepatoma cell line HepG2- and normal liver cell line L02-derived protein mixtures as global internal standards to measure the differential proteomes between HCC/non-HCC tissues. The experimental pipeline is summarized in Figure 1. To our knowledge, this work provides the first Received: July 25, 2011 Published: November 14, 2011 871
dx.doi.org/10.1021/pr2006999 | J. Proteome Res. 2012, 11, 871–885
Journal of Proteome Research
ARTICLE
Figure 1. Quantitative proteome analysis in HCC/non-HCC tissues using 2D-LCMS/MS-CDIT. To accurately obtain proteome changes between clinical HCC and non-HCC tissues, HepG2 and L02 cells were cultured in heavy ([13C6]) leucine as the global internal standard. HCC or non-HCC tissue samples were equally mixed with stable isotope-labeled cultured cell samples early in the process to obviate the variation during sample preparation. The two mixed tissue-cell samples were subjected to trypsin in-solution digestion. The two peptide mixtures were separated with an online 2D-LCMS/MS system, followed by LTQ-Orbitrap analysis. The ratio between the two isotopic distributions, one from the tissue sample and one from the cultured cells labeled with isotopes, was then determined from full mass scans. Protein expression changes in HCC and non-HCC tissues were estimated by calculating the ratio of the two ratios (Ratio 1/Ratio 2), which canceled out the internal standards.
quantitative and comprehensive study of abnormal protein expression and proteomic phenotyping associated with clinical HCC using a culture-derived isotope tag strategy combined with highthroughput and accurate MS technology.
were rinsed three times with cold glutamine-free RPMI 1640 medium13,14 and homogenized in a liquid nitrogen-cooled mortar and pestle. The obtained powder was suspended in lysis buffer (8 mol/L urea, 4% CHAPS, 40 mmol/L Tris, 65 mmol/L DTT). Samples were sonicated on ice for approximately 2 min using an ultrasonic processor and centrifuged for 1 h at 25000 g. The concentrations were measured by a modified Bradford assay (Bio-Rad). All samples were stored at 80 °C until use.
’ EXPERIMENTAL PROCEDURES Materials
Access to human tissues complied with both Chinese laws and the guidelines of the institutional Ethics Committee; tissues from hepatocellular carcinoma patients were isolated from freshly partial hepatectomized tissues in the Shanghai Eastern Hepatobiliary Surgery Hospital. Urea, thiourea, tris, sodium dodecyl sulfate (SDS), glycine, acrylamide, bis-acrylamide, tetramethylethylenediamine (TEMED), bromophenole blue, 3-[(3-cholanidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), and dithiothreitol (DTT) were purchased from Bio-Bad (Hercules, CA). CLM-2262 L-leucine (U13C6, 98%) was obtained from Cambridge Isotope Laboratories (Andover, MA). Dialytic fetal bovine serum was purchased from Gibco Life Technologies (Gibco BRL, Grand Island, NY). D9785 cell culture powder, β-mercaptoethanol, L-glutamine, L-lysine, and L-methionine were from Sigma (St. Louis, MO). Sequencing grade TPCK-trypsin was purchased from Promega (Madison, WI). Acetone, ethanol and acetic acid were obtained from Shanghai Chemicals Corp. ECL Plus reagents were purchased from GE Healthcare (Uppsala, Sweden). All buffers were prepared with Milli-Q water.
Cell Culture and Sample Preparation
HepG2 and L02 were obtained from the Cell Bank, Institute of Biochemistry and Cell Biology, Chinese Academy of Sciences. HepG2 or L02 were cultured in D9785 medium supplemented with 13C-labeled L-leucine, 10% dialyzed FBS, L-glutamine, L-lysine, and L-methionine in 6-cm dishes. After at least six generations, HepG2 or L02 cells were transferred into 10-cm dishes. At 80% density, the cells were washed three times with cooled PBS, and lysed in lysis buffer. The two groups of 13 C-labeled cells were scraped from the dishes, sonicated on ice, and centrifuged for 1 h at 25000 g to remove DNA, RNA and any particulate materials. The concentrations were measured by a modified Bradford assay (Bio-Rad). All samples were stored at 80 °C until use. In-solution Digestion
For MS-based quantitation analysis, 13C-labeled HepG2 and13C-labeled L02 samples were mixed with equal amount of protein as internal standards. A pooled 500 μg protein sample from 5 HCC samples (P1P5, 100 μg/sample) or 5 non-HCC samples (P1P5, 100 μg/sample) was mixed with 500 μg of a 13 C-labeled leucine internal standard. The two tissue-cell sample mixtures were reduced with 10 mM DTT at 37 °C for 2 h and then carbamidomethylated with 50 mM IAA for 45 min at room temperature in the dark. The alkylated protein solutions were incubated with 4 volumes of cold acetone/ethanol (1:1, v/v) with 0.1% TFA at 20 °C for 12 h and centrifuged at 25000 g
Tissue Specimen and Sample Preparation by Nonenzymatic Method (NESP)
Nonenzymatic sample preparation was used to prepare the tissue samples.20 Eleven tumorous tissues and their adjacent paired nontumorous tissues (3 cm from the edge of HCC lesions, approximately 0.1 g) were isolated from freshly partially hepatectomized tissues. The pathologic data for 11 cases are summarized in Supplemental Table 1 (Supporting Information). The tissues 872
dx.doi.org/10.1021/pr2006999 |J. Proteome Res. 2012, 11, 871–885
Journal of Proteome Research
ARTICLE
for 1 h to remove the supernatant. The protein pellets were resuspended in a 50 mM ammonium chloride (pH 8.3) buffer with shaking, and then incubated with sequencing-grade modified trypsin (1:50) for 4 h at 37 °C. Next, trypsin was added to reach a final protease/protein ratio of 1:25. After 16 h, the digestion solutions were ultrafiltered using 10 kDa Microcon Centrifugal Filter Devices (Millipore) to remove trypsin, and then the two samples were lyophilized.
if all of the identified peptides assigned to them were also assigned to another protein group. Second, a single protein with the highest sequence coverage was selected from one protein group for further analysis. These two steps made all peptides distinct and only assigned to a single protein. For quantitative analysis, only the leucine-containing peptides were subjected to the program Census (version 1.33) as quantification candidates to determine the (12C6)-peptide/(13C6)peptide ratio.25 The results of quantification were filtered with the determinant scores (R2) g 0.5 and the correction factor (Ln) was set at 0.0 when exported. Peptides with negative R2 scores were removed and singleton peptides with less than two spectra counts were discarded. Protein quantification was performed as described in our previously published paper.26 Briefly, the SILAC ratio of all peptides was measured for protein quantification. Outliers were eliminated from all peptides that were assigned to the same protein using the biweight algorithm, while the weighted mean of the peptide ratio was determined as the protein expression ratio.
2D-LCMS/MS Analysis
An online 2D-LCMS/MS system was used as described previously21 with some modifications. Briefly, the sample was dissolved in 100 μL of a pH 2.0 buffer and then loaded onto the SCX column by a syringe pump at a flow rate of 3 μL/min. The SCX column was coupled with a Surveyor liquid chromatography system (Thermo Finnigan, San Jose, CA), consisting of a degasser, an MS pump, an autosampler, a C18 trap column (RP, 320 μm 20 mm, Column Technology, Inc., Fremont, CA) and an analytical C18 column (RP, 75 μm X 150 mm, Column Technology, Inc., Fremont, CA) online. The HPLC solvents used were 0.1% formic acid (v/v) aqueous (A) and 0.1% formic acid (v/v) acetonitrile (B). Instead of previously reported pH step gradient buffer, peptides were eluted from the SCX column with a continuous pH gradient from pH 2.0 to pH 8.5. Each eluted fractions was concentrated and desalted on a C18 trap column (RP, 320 μm 20 mm, Column Technology, Inc., Fremont, CA) at a flow rate of 3 μL/min after the split, and then subjected to an analytical C18 column (RP, 75 μm 150 mm, Column Technology, Inc., Fremont, CA) online. A reverse-phase gradient from 2 to 40% of the mobile phase B was used for 165 min at a flow rate of 100 μL/min before the split and 250 nL/min after the split. An LTQ-Orbitrap mass spectrometer (Thermo Electron Finnigan, San Jose, CA) equipped with a nanospray source from the same manufacture was used. The ion transfer capillary was 160 °C and the NSI voltage was 1.8 kV. A normalized CID collision energy of 35.0 was used. A full scan was obtained in the Orbitrap analyzer (R = 100000 at m/z 400) followed by MS/MS acquisition of the ten most intense ions in the LTQ spectrometer.
Data Normalization, Quantitation Accuracy Evaluation, and Differentially Expressed Protein Selection
To minimize system errors, we first normalized the two groups of protein ratios (HCC/13C-cells and non-HCC/13C-cells). The protein ratios were Loge (Ln) transformed and then normalized. For each group, the mean was zero and the standard deviation (σ) was one. Normalization of the raw ratios ensured that proteins with similar ratios of each group were close in Euclidean space. Therefore, quantitative protein results between HCC and non-HCC were obtained by comparing two groups of normalized ratios. To assess the quantification accuracy or to calculate the quantification error, the average relative standard deviation (average RSD) was determined. Each quantitative protein usually has several quantitative peptides, from which the RSD for most proteins can be obtained. The average RSD for the two groups of protein quantitative data indicated the accuracy of quantification. Accuracy ¼ ðaverage RSDHCC þ average RSDnon-HCC Þ=2 ¼ ð0:136 þ 0:144Þ=2 ¼ 0:14
Database Searching, Protein Identification, and Quantitative Analysis Using Census
ð1Þ
The accuracy indicates an acceptable quantification precision. Therefore, we considered 1.5-fold changes as the threshold for significant differences. In total, 267 significant differentially expressed proteins in HCC tissues were unambiguously quantified by 2D-LCMS/MS-CDIT.
The MS and MS/MS peak lists were searched against the database consisting of forward and reverse sequences22 of the human International Protein Index protein sequence database (Version 3.28, containing 68020 proteins, www.ebi.ac.uk/IPI) using the TurboSEQUEST program in the BioWorks 3.2 software package (University of Washington, licensed to Thermo Finnigan). Trypsin was designated as the protease and two missed cleavage site was allowed. Carbamidomethylation was searched as a fixed modification, and isotope labeled-leucine (+6.0201 Da) and oxidized methionine (+15.99492 Da) were allowed as variable modifications. The precursor and fragment ion mass tolerances were 500 ppm and 1 Da (Default), respectively. All of the out files were filtered by homemade Buildsummary software.23,24 A cutoff of 1% False discovery rate (FDR),22 precursor ion mass accuracy of 10 ppm, and deltaCN score of at least 0.1 regardless of the charge state were set to filter the identified peptides. In addition, all of the identified proteins with only one spectral count were deleted, and the final protein FDR of the identified 2383 proteome was as small as 1.55%. BuildSummary23,24 software was also used to eliminate redundancies for quantitation. First, protein groups were removed
Pathway Enrichment Analysis
Pathway enrichment analysis-based hierarchical clustering was performed according to previous work by Pan CP.27 Briefly, the quantified HCC/non-HCC proteome was divided into 5 quantiles corresponding to percentage cutoffs of 015%, 1525%, 2575%, 7585%, and 85100%. In contrast to the whole proteome background of the human database, the enrichment analyses for the Gene Ontology biological process and the BioCarta and KEGG pathways were performed separately for these quantiles using the bioinformatics resource DAVID.28 For hierarchical clustering, all the categories were first collated along with their p values (A modified Fisher’s exact test, EASE score)28 and then filtered for those categories that were enriched in at least one of the quantiles with a p value 0.70 (Figure 5C). The 95% CI range of AUC was 0.5990.848 (AUC = 0.724) in the wholecell or 0.6070.859 (AUC = 0.733) in the nucleus between the early and middle stages; by contrast, the 95% CI range of AUC was 0.6330.874 (AUC = 0.754) in the whole-cell or 0.684 0.905 (AUC = 0.795) in the nucleus between the early and late stages. Therefore, APEX1 could be associated with not only HCC malignancy, but also with HCC differentiation grading. Collectively, these results identified APEX1 and ANP32A as potential tissue biomarker candidates for HCC diagnosis. The combination of APEX1 and ANP32A might increase the sensitivity and specificity of the diagnosis. Importantly, both APEX1 expression and its nuclear accumulation seem to have the potential to serve as tumor-stage discriminators.
’ DISCUSSION First Application of CDITs to Accurately Quantify the Proteome of Clinical HCC/non-HCC Tissues
Here, we applied an online 2D-LC system and CDITs to HCC clinical tissues followed by quantitative proteomicsusing an LTQ-Orbitrap mass spectrometer. We successfully identified 2383 proteins and quantified 1360 proteins, displaying the high-throughput power of 2D-LCMS/MS-CDIT. Although several proteomic analyses have already been used on tissue samples from HCC patients611 including two of our previous works,1315 to the best of our knowledge, this report describes the first accurate, quantitative and comprehensive study on abnormal protein expression associated with clinical HCC that combines CDITs and high mass accuracy MS technology. Recently, Geiger T et al. reported the super-SILAC method to quantify breast cancer tissue with SILAC-labeled cell lines.18 When a single cell line was used, the distribution ratio was broad and bimodal, with 17% of proteins exhibiting more than a 4-fold expression. When 5 cell lines (The super-SILAC mixture) were used, the distribution was much narrower and unimodal with only 10% of proteins showing more than a 4-fold ratio. Because the super-SILAC method focuses on tumor tissues, we introduced a normal liver cell line, L02, which can be helpful to mimic the proteome from non-HCC tissues. Due to biological complexity and the need for relative quantification between cancerous and normal state, we used two cell lines (Hepatoma cell line HepG2 and normal liver cell line L02) as global internal standards. The distribution ratio between HCC tissues and SILAC-labeled cell lines or between non-HCC tissues and SILAC-labeled cell lines was unimodal (Supplemental Figure 2AB, Supporting Information), with barely 6% (63/1030 or
a
0.115
0.355
43/16 (38) (0200)
37/19 (60) (0190)
38/6 (60) (0225)
31/5 (60) (0143)
16/0 (5) (0113)
19/4 (34) (095) 90/109 (93) (0190) 85/78 (94) (0190) Cytoplasm
0.726
84/75 (136) (0285) 38/8 (56) (0180) Nucleus
1.5 fold changes) was highlighted by the functional enrichment analysis using the bioinformatic resources David (data not shown). In summary, the HCC/non-HCC proteomic phenotype delineated in this work provides an unbiased global portrait of representative biological functions in clinical tissues on the proteomic level. The results showed that HCC shares common hallmarks of cancer that result from 881
dx.doi.org/10.1021/pr2006999 |J. Proteome Res. 2012, 11, 871–885
Journal of Proteome Research
ARTICLE
many stresses, including metabolic, mitotic, oxidative and DNA damage stresses.40
is mainly nuclear; whereas in carcinomas, nuclear, cytoplasmic, and nuclear/cytoplasmic stainings were observed.49 A similar phenomenon was found in our TMA data and confirmed by immunofluorescence staining in HepG2 and L02 cells (Supplemental Figure 4A, Supporting Information). Moreover, the compartmentalization of APEX1 in different subcellular regions, such as the nucleus, ER, or mitochondria, may explain its different functional specificities.49 In our TMA analysis, we observed significantly higher expression of APEX1 in both the nucleus and the cytoplasm and only in nucleus in HCC tumor cells (Figure 5A). Intriguingly, statistical analysis also showed a significant increase of APEX1 expression with progressive pathological grades in both the nucleus and cytoplasm and only in the nucleus, suggesting the potential of APEX1 as an HCC pathological grade marker (Figure 5C). Additionally, a truncated isoform of APEX1 (Approximately 33 kDa) was found only in the HCC tissues of 5 Edmondson grade III patients (Figure 4A) by Western blot, implying that such a cleaved isoform is related to HCC cancerogenesis. Consistently, cytoplasmic distribution of APEX1 has been observed in several tumors, including HCC,50 lung cancer,51 ovarian cancer,52 thyroid cancer,53,54 and breast cancer,55 and is correlated with poor prognosis. Moreover, the cytoplasmic APEX1 isoform (33 kDa) is derived from full-length 36 kDa APEX1, but its N-terminal 33 amino acids, which contain the nuclear localization signal (NLS), were removed by a specific mitochondria-associated N-terminal peptidase.35 Vittorio Di Maso reported that cytoplasmic APEX1 localization is associated with a survival time of patients shorter than for those with negative cytoplasmic reactivity and might be a predictor of survival for HCC patients.50 The above-described evidence suggests that the expression, subcellular localization, and trafficking of APEX1 are highly regulated processes in hepatocarcinogenesis and HCC pathological grades. ANP32A is a member of the ANP32 family of acidic, leucinerich, nuclear phosphoproteins found in cells capable of selfrenewal and in certain long-lived neuronal populations.56 ANP32A also functions in many cellular processes, such as proliferation, 57 differentiation,58 caspase-dependent and caspase-independent apoptosis, 5961 tumor suppression,62 and regulation of mRNA trafficking. 63 In contrast to its proapoptotic and transformation inhibition functions, ANP32A is highly expressed in cancer. 56,57,61,64 Augmented caspase activation is mediated by ANP32A, which is overexpressed in breast cancers.61 Higher levels of ANP32A mRNA were reported in prostatic adenocarcinomas and might contribute to increased malignancy.64 In this work, we reported and confirmed the overexpression of ANP32A in HCC tumorous tissues (Figures 4B and 5B). Nevertheless, the IHC staining of TMAs showed that ANP32A had both cytoplasmic and nuclear localization in HCC hepatocytes; this was validated by immunofluorescence in HepG2 and L02 cell lines (Supplemental Figure 4B, Supporting Information). Significantly higher expression of ANP32A at the whole-cell and nuclear levels in HCC hepatocytes suggestes that both expression and subcellular localization of ANP32A might be important features associated with hepaptocarcinogenesis. Based on ROC curves, we determined that the combination of the expression and nuclear accumulation of both APEX1 and ANP32Aexhibits a better discriminative power than either protein alone, indicating the importance of exploring the whole SET complex as an indicator for HCC oncogenesis.
Up-regulation of the SET Complex Associated with Hepatocarcinogenesis
As illustrated in Figure 2C and Figure 3, we found abnormal regulation of the GzmA pathway (p = 0.013), which is mainly attributed to the significant up-regulation of the SET complex (Four members) and is a central component in the GzmA pathway in HCC tissues. Granzyme A (GzmA), an abundant serine protease, induces a caspase-independent cell death pathway and recognizes either virally infected or tumor cells.34,41 GzmA triggers cell death by targeting the 270420 kDa endoplasmic reticulum (ER)-associated complex, termed the SET complex.34 It is well-known that the SET complex contains the GzmA-activated DNase NME1, the protein phosphatase PP2A inhibitor ANP32A, and three GzmA substrates---the nucleosome assembly protein SET, the base excision repair enzyme APEX1, and the DNA-binding protein HMGB2.42,43 Therefore, we used Western blotting to validate the overexpression of all five SET complex components in clinical HCC tissues (Figure 4). Similar results of a significant correlation between the expression of SET complex proteins and tumor differentiation have been observed in ovarian cancer.44 Herein, we described the first simultaneous study of all members of the SET complex in clinical HCC tissues and demonstrated the deregulation of the SET complex and the GzmA pathway using quantitative proteomics. Such up-regulation of the SET complex in HCC tissues is consistent with the fact that cellular transformation is generally associated with mutations in components of apoptosis pathways. Viruses including HBV have evolved mechanisms to interfere with apoptosis programs to prolong survival of their host cells.41 The onset of apoptosis in transformed cells can be triggered by the release of GzmA and reactive oxygen species (ROS) formation is a hallmark of apoptosis.45 Martinvalet et al. showed that GzmA rapidly penetrates the mitochondrial matrix and cleaves the NADH dehydrogenase FeS protein 3 (NDUFS3), disrupting the electron transport chain and leading to ROS production.46 An elevated oxidative state has been found in many types of cancer cells, and the introduction of chemical and enzymatic antioxidants can inhibit tumor cell proliferation, pointing to a critical role of ROS in mediating the loss of growth control.47 Increased ROS in the cytoplasm of cells promotes translocation of the SET complex into the nucleus. Recently, the exonuclease TREX1 was also reported to be in the SET complex, acting in concert with NM23-H1 to degrade DNA during GzmA-mediated cell death.48 In the nucleus, GzmA cleaves SET, APEX1, and HMGB2 and induces the release of ANP32A, NM23-H1 and TREX1, three proteins that are responsible for DNA fragmentation and that contribute to GzmA-induced cell death.4648 Therefore, ROS formation and oxidative DNA damage triggered by the deregulation of the SET complex and GzmA pathway may be involved in the process of hepatocarcinogenesis. Both the expression and subcellular localization of APEX1 and ANP32A were further validated in an independent sample set consisting of 155 HCC patients by TMAs (Figure 5, Supplemental Figure 5 (Supporting Information), and Table 1). APEX1 is a good example of the functional complexity of a biological macromolecule; it can act as a regulator of the cellular response to oxidative stress, a transcriptional coactivator, and a DNA repair enzyme.39 In normal tissue, the APEX1 localization 882
dx.doi.org/10.1021/pr2006999 |J. Proteome Res. 2012, 11, 871–885
Journal of Proteome Research
’ CONCLUSIONS In summary, we provide the first simultaneous study of all members of the SET complex in the context of HCC, demonstrating that expression of APEX1, ANP32A, SET, HMGB2 and NME1 is deregulated in clinical HCC tissues. This study highlights the advantage of looking at key members (The SET complex) of a pathway (The GzmA pathway) supported by bioinformatics phenotyping. In particular, APEX1 and ANP32A may not only offer novel functional insights into molecular oncology, but also may be potential markers in HCC tissues. Moreover, APEX1 could be associated with not only HCC malignancy, but also with HCC differentiation. Therefore, the combination of APEX1 and ANP32A will increase the sensitivity and specificity of diagnosis. Our work provides a systematic and quantitative analysis, combining culture-derived isotope tag strategy with highthroughput and accuracy MS technology. Using this method, we have demonstrated significantly changed proteins and key regulatory processes in clinical HCC/non-HCC tissues. These methods have potential for understanding pathogenesis, detecting markers and identifying new drug targets and could be expanded to perform research on all kinds of cancers.
ARTICLE
apoptosis pathway; APEX1, DNA-(apurinic or apyrimidinic site) lyase, Ape1; ANP32A, Acidic leucine-rich nuclear phosphoprotein 32 family member A, pp32; SET, isoform 1 of protein SET; HMGB2, high mobility group protein B2, HMG2; NME1, nucleoside diphosphate kinase A, NM23-H1; TMA, tissue microarray; IHC, immunohistochemistry; ROS, reactive oxygen species; ROC, receiver operating characteristic curve; AUC, area under ROC curve
’ REFERENCES (1) Arbuthnot, P.; Kew, M. Hepatitis B virus and hepatocellular carcinoma. Int. J. Exp. Pathol. 2001, 82 (2), 77–100. (2) Coronado, G. D.; Taylor, V.; Acorda, E.; Hoai, D. H.; Thompson, B. Development of an English as a second language curriculum for hepatitis B virus testing in Chinese Americans. Cancer 2005, 104 (12 Suppl), 2948–2951. (3) Vivekanandan, P.; Singh, O. V. High-dimensional biology to comprehend hepatocellular carcinoma. Expert Rev. Proteomics 2008, 5 (1), 45–60. (4) Coleman, W. B. Mechanisms of human hepatocarcinogenesis. Curr. Mol. Med. 2003, 3 (6), 573–588. (5) Roberts, L. R.; Gores, G. J. Hepatocellular carcinoma: molecular pathways and new therapeutic targets. Semin. Liver Dis. 2005, 25 (2), 212–225. (6) Ai, J.; Tan, Y.; Ying, W.; Hong, Y.; Liu, S.; Wu, M.; Qian, X.; Wang, H. Proteome analysis of hepatocellular carcinoma by laser capture microdissection. Proteomics 2006, 6 (2), 538–546. (7) Chaerkady, R.; Harsha, H. C.; Nalli, A.; Gucek, M.; Vivekanandan, P.; Akhtar, J.; Cole, R. N.; Simmers, J.; Schulick, R. D.; Singh, S.; Torbenson, M.; Pandey, A.; Thuluvath, P. J. A quantitative proteomic approach for identification of potential biomarkers in hepatocellular carcinoma. J. Proteome Res. 2008, 7 (10), 4289–4298. (8) Codarin, E.; Renzone, G.; Poz, A.; Avellini, C.; Baccarani, U.; Lupo, F.; di, M.; Croce, V; Tiribelli, S. L.; Arena, C.; Quadrifoglio, S.; Scaloni, F.; Tell, A. G. Differential Proteomic Analysis of Subfractioned Human Hepatocellular Carcinoma Tissues. J. Proteome Res. 2009, 8 (5), 2273–2284. (9) Minagawa, H.; Honda, M.; Miyazaki, K.; Tabuse, Y.; Teramoto, R.; Yamashita, T.; Nishino, R.; Takatori, H.; Ueda, T.; Kamijo, K.; Kaneko, S. Comparative proteomic and transcriptomic profiling of the human hepatocellular carcinoma. Biochem. Biophys. Res. Commun. 2008, 366 (1), 186–192. (10) Park, K. S.; Kim, H.; Kim, N. G.; Cho, S. Y.; Choi, K. H.; Seong, J. K.; Paik, Y. K. Proteomic analysis and molecular characterization of tissue ferritin light chain in hepatocellular carcinoma. Hepatology 2002, 35 (6), 1459–1466. (11) Sun, W.; Xing, B.; Sun, Y.; Du, X.; Lu, M.; Hao, C.; Lu, Z.; Mi, W.; Wu, S.; Wei, H.; Gao, X.; Zhu, Y.; Jiang, Y.; Qian, X.; He, F. Proteome analysis of hepatocellular carcinoma by two-dimensional difference gel electrophoresis: novel protein markers in hepatocellular carcinoma tissues. Mol. Cell. Proteomics 2007, 6 (10), 1798–1808. (12) Ding, S. J.; Li, Y.; Tan, Y. X.; Jiang, M. R.; Tian, B.; Liu, Y. K.; Shao, X. X.; Ye, S. L.; Wu, J. R.; Zeng, R.; Wang, H. Y.; Tang, Z. Y.; Xia, Q. C. From proteomic analysis to clinical significance: overexpression of cytokeratin 19 correlates with hepatocellular carcinoma metastasis. Mol. Cell. Proteomics 2004, 3 (1), 73–81. (13) Li, C.; Hong, Y.; Tan, Y. X.; Zhou, H.; Ai, J. H.; Li, S. J.; Zhang, L.; Xia, Q. C.; Wu, J. R.; Wang, H. Y.; Zeng, R. Accurate qualitative and quantitative proteomic analysis of clinical hepatocellular carcinoma using laser capture microdissection coupled with isotope-coded affinity tag and two-dimensional liquid chromatography mass spectrometry. Mol. Cell. Proteomics 2004, 3 (4), 399–409. (14) Li, C.; Tan, Y. X.; Zhou, H.; Ding, S. J.; Li, S. J.; Ma, D. J.; Man, X. B.; Hong, Y.; Zhang, L.; Li, L.; Xia, Q. C.; Wu, J. R.; Wang, H. Y.; Zeng, R. Proteomic analysis of hepatitis B virus-associated hepatocellular carcinoma: Identification of potential tumor markers. Proteomics 2005, 5 (4), 1125–1139. (15) Li, C.; Yi, H.; Tan, Y. X.; Ai, J. H.; Zhou, H.; Li, S. J.; Zhang, L.; Xia, Q. C.; Wu, J. R.; Wang, H. Y.; Zeng, R. Analysis of microdissected
’ ASSOCIATED CONTENT
bS
Supporting Information Supplementary tables and figures. This material is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION Corresponding Author
*Dr. Rong Zeng, Key Laboratory of Systems Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai 200031, China. Tel: +86-21-54920160. Fax: +86-21-54920171. E-mail:
[email protected]. Author Contributions ‡
These authors contributed equally to this work
’ ACKNOWLEDGMENT We thank Ph.D. Ai-ping lv for criticism and help with the manuscript. Supported by grants from Basic Research Foundation (2011CB910600, 2007CB947803, 2010CB912100, 2011CB910200). National Natural Science Foundation (30821065, 90713032, and 30700397), and SIBS Project (2007KIP308). ’ ABBREVIATIONS HCC tissues, hepatocellular carcinoma tumorous tissues; nonHCC tissues, adjacent paired nontumorous tissues; SCX, strong cation exchange chromatography; RP, reversed phase chromatography; 2D-LCMS/MS-CDIT, two-dimensional liquid chromatographytandem mass spectrometry coupled with culture-derived isotope tags; LTQ-Orbitrap, linear ion trap/orbitrap mass spectrometer; SILAC, stable isotope labeling by amino acids in cell culture; GO, Gene Ontology; KEGG, Kyoto encyclopedia of genes and genomes; EHCO, Encyclopedia of hepatocellular carcinoma genes online; DAVID, database for annotation, visualization and integrated discovery; UniProtKB, The UniProt Knowledgebase; GzmA pathway, Granzyme A-mediated 883
dx.doi.org/10.1021/pr2006999 |J. Proteome Res. 2012, 11, 871–885
Journal of Proteome Research
ARTICLE
markers’ discovery to functional pathways. Liver Int. 2007, 27 (8), 1021–1038. (34) Chakravarti, D.; Hong, R. SET-ting the stage for life and death. Cell 2003, 112 (5), 589–591. (35) Chattopadhyay, R.; Wiederhold, L.; Szczesny, B.; Boldogh, I.; Hazra, T. K.; Izumi, T.; Mitra, S. Identification and characterization of mitochondrial abasic (AP)-endonuclease in mammalian cells. Nucleic Acids Res. 2006, 34 (7), 2067–2076. (36) Kruger, M.; Moser, M.; Ussar, S.; Thievessen, I.; Luber, C. A.; Forner, F.; Schmidt, S.; Zanivan, S.; Fassler, R.; Mann, M. SILAC mouse for quantitative proteomics uncovers kindlin-3 as an essential factor for red blood cell function. Cell 2008, 134 (2), 353–364. (37) Faca, V. M.; Song, K. S.; Wang, H.; Zhang, Q.; Krasnoselsky, A. L.; Newcomb, L. F.; Plentz, R. R.; Gurumurthy, S.; Redston, M. S.; Pitteri, S. J.; Pereira-Faca, S. R.; Ireton, R. C.; Katayama, H.; Glukhova, V.; Phanstiel, D.; Brenner, D. E.; Anderson, M. A.; Misek, D.; Scholler, N.; Urban, N. D.; Barnett, M. J.; Edelstein, C.; Goodman, G. E.; Thornquist, M. D.; McIntosh, M. W.; DePinho, R. A.; Bardeesy, N.; Hanash, S. M. A mouse to human search for plasma proteome changes associated with pancreatic tumor development. PLoS Med. 2008, 5 (6), e123. (38) Chen, Y. T.; Chen, C. L.; Chen, H. W.; Chung, T.; Wu, C. C.; Chen, C. D.; Hsu, C. W.; Chen, M. C.; Tsui, K. H.; Chang, P. L.; Chang, Y. S.; Yu, J. S. Discovery of novel bladder cancer biomarkers by comparative urine proteomics using iTRAQ technology. J. Proteome Res. 2010, 9 (11), 5803–5815. (39) Wang, D.; Hincapie, M.; Guergova-Kuras, M.; Kadas, J.; Takacs, L.; Karger, B. L. Antigen identification and characterization of lung cancer specific monoclonal antibodies produced by mAb proteomics. J. Proteome Res. 2010, 9 (4), 1834–1842. (40) Luo, J.; Solimini, N. L.; Elledge, S. J. Principles of cancer therapy: oncogene and non-oncogene addiction. Cell 2009, 136 (5), 823–837. (41) Lucken-Ardjomande, S.; Martinou, J. C. Granzyme a, a stealth killer in the mitochondrion. Cell 2008, 133 (4), 568–570. (42) Martinvalet, D.; Zhu, P.; Lieberman, J. Granzyme A induces caspase-independent mitochondrial damage, a required first step for apoptosis. Immunity 2005, 22 (3), 355–370. (43) Beresford, P. J.; Zhang, D.; Oh, D. Y.; Fan, Z.; Greer, E. L.; Russo, M. L.; Jaju, M.; Lieberman, J. Granzyme A activates an endoplasmic reticulum-associated caspase-independent nuclease to induce single-stranded DNA nicks. J. Biol. Chem. 2001, 276 (46), 43285–43293. (44) Ouellet, V.; Le Page, C.; Guyot, M. C.; Lussier, C.; Tonin, P. N.; Provencher, D. M.; Mes-Masson, A. M. SET complex in serous epithelial ovarian cancer. Int. J. Cancer 2006, 119 (9), 2119–2126. (45) Cullen, S. P.; Martin, S. J. Mechanisms of granule-dependent killing. Cell Death Differ. 2008, 15 (2), 251–262. (46) Martinvalet, D.; Dykxhoorn, D. M.; Ferrini, R.; Lieberman, J. Granzyme A cleaves a mitochondrial complex I protein to initiate caspase-independent cell death. Cell 2008, 133 (4), 681–692. (47) Behrend, L.; Henderson, G.; Zwacka, R. M. Reactive oxygen species in oncogenic transformation. Biochem. Soc. Trans. 2003, 31 (Pt 6), 1441–1444. (48) Chowdhury, D.; Beresford, P. J.; Zhu, P.; Zhang, D.; Sung, J. S.; Demple, B.; Perrino, F. W.; Lieberman, J. The exonuclease TREX1 is in the SET complex and acts in concert with NM23-H1 to degrade DNA during granzyme A-mediated cell death. Mol. Cell 2006, 23 (1), 133–142. (49) Tell, G.; Damante, G.; Caldwell, D.; Kelley, M. R. The intracellular localization of APE1/Ref-1: more than a passive phenomenon? Antioxid Redox. Signal. 2005, 7 (34), 367–384. (50) Di, M., V; Avellini, C.; Croce, L. S.; Rosso, N.; Quadrifoglio, F.; Cesaratto, L.; Codarin, E.; Bedogni, G.; Beltrami, C. A.; Tell, G.; Tiribelli, C. Subcellular localization of APE1/Ref-1 in human hepatocellular carcinoma: possible prognostic significance. Mol. Med. 2007, 13 (12), 89–96. (51) Puglisi, F.; Aprile, G.; Minisini, A. M.; Barbone, F.; Cataldi, P.; Tell, G.; Kelley, M. R.; Damante, G.; Beltrami, C. A.; Di, L. C. Prognostic
cells by two-dimensional LC-MS approaches. Methods Mol. Biol. 2008, 428, 193–208. (16) Sechi, S.; Oda, Y. Quantitative proteomics using mass spectrometry. Curr. Opin. Chem. Biol. 2003, 7 (1), 70–77. (17) Ong, S. E.; Blagoev, B.; Kratchmarova, I.; Kristensen, D. B.; Steen, H.; Pandey, A.; Mann, M. Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Mol. Cell. Proteomics 2002, 1 (5), 376–386. (18) Geiger, T.; Cox, J.; Ostasiewicz, P.; Wisniewski, J. R.; Mann, M. Super-SILAC mix for quantitative proteomics of human tumor tissue. Nat. Methods 2010, 7 (5), 383–385. (19) Ishihama, Y.; Sato, T.; Tabata, T.; Miyamoto, N.; Sagane, K.; Nagasu, T.; Oda, Y. Quantitative mouse brain proteomics using culture-derived isotope tags as internal standards. Nat. Biotechnol. 2005, 23 (5), 617–621. (20) Franzen, B.; Hirano, T.; Okuzawa, K.; Uryu, K.; Alaiya, A. A.; Linder, S.; Auer, G. Sample preparation of human tumors prior to twodimensional electrophoresis of proteins. Electrophoresis 1995, 16 (7), 1087–1089. (21) Dai, J.; Shieh, C. H.; Sheng, Q. H.; Zhou, H.; Zeng, R. Proteomic analysis with integrated multiple dimensional liquid chromatography/mass spectrometry based on elution of ion exchange column using pH steps. Anal. Chem. 2005, 77 (18), 5793–5799. (22) Peng, J.; Elias, J. E.; Thoreen, C. C.; Licklider, L. J.; Gygi, S. P. Evaluation of multidimensional chromatography coupled with tandem mass spectrometry (LC/LCMS/MS) for large-scale protein analysis: the yeast proteome. J. Proteome Res. 2003, 2 (1), 43–50. (23) Jin, W. H.; Dai, J.; Li, S. J.; Xia, Q. C.; Zou, H. F.; Zeng, R. Human plasma proteome analysis by multidimensional chromatography prefractionation and linear ion trap mass spectrometry identification. J. Proteome Res. 2005, 4 (2), 613–619. (24) Tu, C. J.; Dai, J.; Li, S. J.; Sheng, Q. H.; Deng, W. J.; Xia, Q. C.; Zeng, R. High-sensitivity analysis of human plasma proteome by immobilized isoelectric focusing fractionation coupled to mass spectrometry identification. J. Proteome Res. 2005, 4 (4), 1265–1273. (25) Park, S. K.; Venable, J. D.; Xu, T.; Yates, J. R., III A quantitative analysis software tool for mass spectrometry-based proteomics. Nat. Methods 2008, 5 (4), 319–322. (26) Wu, Y. B.; Dai, J.; Yang, X. L.; Li, S. J.; Zhao, S. L.; Tang, J. S.; Zheng, G. Y.; Li, Y. X.; Wu, J. R.; Zeng, R. Concurrent quantification of proteome and phosphoproteome to reveal system-wide association of protein phosphorylation and gene expression. Mol. Cell. Proteomics 2009, 8 (12), 2809–2826. (27) Pan, C.; Kumar, C.; Bohl, S.; Klingmueller, U.; Mann, M. Comparative proteomic phenotyping of cell lines and primary cells to assess preservation of cell type-specific functions. Mol. Cell. Proteomics 2009, 8 (3), 443–450. (28) Huang, d. W.; Sherman, B. T.; Lempicki, R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. 2009, 4 (1), 44–57. (29) Sturn, A.; Quackenbush, J.; Trajanoski, Z. Genesis: cluster analysis of microarray data. Bioinformatics 2002, 18 (1), 207–208. (30) Goncalves, A.; Charafe-Jauffret, E.; Bertucci, F.; Audebert, S.; Toiron, Y.; Esterni, B.; Monville, F.; Tarpin, C.; Jacquemier, J.; Houvenaeghel, G.; Chabannon, C.; Extra, J. M.; Viens, P.; Borg, J. P.; Birnbaum, D. Protein profiling of human breast tumor cells identifies novel biomarkers associated with molecular subtypes. Mol. Cell. Proteomics 2008, 7 (8), 1420–1433. (31) Ong, S. E.; Kratchmarova, I.; Mann, M. Properties of 13Csubstituted arginine in stable isotope labeling by amino acids in cell culture (SILAC). J. Proteome Res. 2003, 2 (2), 173–181. (32) Hsu, C. N.; Lai, J. M.; Liu, C. H.; Tseng, H. H.; Lin, C. Y.; Lin, K. T.; Yeh, H. H.; Sung, T. Y.; Hsu, W. L.; Su, L. J.; Lee, S. A.; Chen, C. H.; Lee, G. C.; Lee, D. T.; Shiue, Y. L.; Yeh, C. W.; Chang, C. H.; Kao, C. Y.; Huang, C. Y. Detection of the inferred interaction network in hepatocellular carcinoma from EHCO (Encyclopedia of Hepatocellular Carcinoma genes Online). BMC. Bioinform. 2007, 8, 66. (33) Sun, S.; Lee, N. P.; Poon, R. T.; Fan, S. T.; He, Q. Y.; Lau, G. K.; Luk, J. M. Oncoproteomics of hepatocellular carcinoma: from cancer 884
dx.doi.org/10.1021/pr2006999 |J. Proteome Res. 2012, 11, 871–885
Journal of Proteome Research
ARTICLE
significance of Ape1/ref-1 subcellular localization in non-small cell lung carcinomas. Anticancer Res. 2001, 21 (6A), 4041–4049. (52) Moore, D. H.; Michael, H.; Tritt, R.; Parsons, S. H.; Kelley, M. R. Alterations in the expression of the DNA repair/redox enzyme APE/ref-1 in epithelial ovarian cancers. Clin. Cancer Res. 2000, 6 (2), 602–609. (53) Russo, D.; Arturi, F.; Bulotta, S.; Pellizzari, L.; Filetti, S.; Manzini, G.; Damante, G.; Tell, G. ApeI/Ref-I expression and cellular localization in human thyroid carcinoma cell lines. J. Endocrinol. Invest. 2001, 24 (3), RC10–RC12. (54) Tell, G.; Pellizzari, L.; Pucillo, C.; Puglisi, F.; Cesselli, D.; Kelley, M. R.; Di, L. C.; Damante, G. TSH controls Ref-1 nuclear translocation in thyroid cells. J. Mol. Endocrinol. 2000, 24 (3), 383–390. (55) Puglisi, F.; Barbone, F.; Tell, G.; Aprile, G.; Pertoldi, B.; Raiti, C.; Kelley, M. R.; Damante, G.; Sobrero, A.; Beltrami, C. A.; Di, L. C. Prognostic role of Ape/Ref-1 subcellular expression in stage I-III breast carcinomas. Oncol. Rep. 2002, 9 (1), 11–17. (56) Walensky, L. D.; Coffey, D. S.; Chen, T. H.; Wu, T. C.; Pasternack, G. R. A novel M(r) 32,000 nuclear phosphoprotein is selectively expressed in cells competent for self-renewal. Cancer Res. 1993, 53 (19), 4720–4726. (57) Malek, S. N.; Katumuluwa, A. I.; Pasternack, G. R. Identification and preliminary characterization of two related proliferation-associated nuclear phosphoproteins. J. Biol. Chem. 1990, 265 (22), 13400–13409. (58) Matsuoka, K.; Taoka, M.; Satozawa, N.; Nakayama, H.; Ichimura, T.; Takahashi, N.; Yamakuni, T.; Song, S. Y.; Isobe, T. A nuclear factor containing the leucine-rich repeats expressed in murine cerebellar neurons. Proc. Natl. Acad. Sci. U.S.A. 1994, 91 (21), 9670–9674. (59) Fan, Z.; Beresford, P. J.; Oh, D. Y.; Zhang, D.; Lieberman, J. Tumor suppressor NM23-H1 is a granzyme A-activated DNase during CTL-mediated apoptosis, and the nucleosome assembly protein SET is its inhibitor. Cell 2003, 112 (5), 659–672. (60) Jiang, X.; Kim, H. E.; Shu, H.; Zhao, Y.; Zhang, H.; Kofron, J.; Donnelly, J.; Burns, D.; Ng, S. C.; Rosenberg, S.; Wang, X. Distinctive roles of PHAP proteins and prothymosin-alpha in a death regulatory pathway. Science 2003, 299 (5604), 223–226. (61) Schafer, Z. T.; Parrish, A. B.; Wright, K. M.; Margolis, S. S.; Marks, J. R.; Deshmukh, M.; Kornbluth, S. Enhanced sensitivity to cytochrome c-induced apoptosis mediated by PHAPI in breast cancer cells. Cancer Res. 2006, 66 (4), 2210–2218. (62) Pan, W.; da Graca, L. S.; Shao, Y.; Yin, Q.; Wu, H.; Jiang, X. PHAPI/pp32 suppresses tumorigenesis by stimulating apoptosis. J. Biol. Chem. 2009, 284 (11), 6946–6954. (63) Brennan, C. M.; Gallouzi, I. E.; Steitz, J. A. Protein ligands to HuR modulate its interaction with target mRNAs in vivo. J. Cell Biol. 2000, 151 (1), 1–14. (64) Kadkol, S. S.; Brody, J. R.; Epstein, J. I.; Kuhajda, F. P.; Pasternack, G. R. Novel nuclear phosphoprotein pp32 is highly expressed in intermediate- and high-grade prostate cancer. Prostate 1998, 34 (3), 231–237.
885
dx.doi.org/10.1021/pr2006999 |J. Proteome Res. 2012, 11, 871–885
