CHCHD2 Is Coamplified with EGFR in NSCLC and Regulates Mitochondrial Function and Cell Migration

Published OnlineFirst March 17, 2015; DOI: 10.1158/1541-7786.MCR-14-0165-T

Genomics Molecular Cancer Research CHCHD2 Is Coampliﬁed with EGFR in NSCLC and Regulates Mitochondrial Function and Cell Migration Yuhong Wei1, Ravi N. Vellanki2,3, Etienne Coyaud3, Vladimir Ignatchenko3, Lei Li1,3, Jonathan R. Krieger1, Paul Taylor1, Jiefei Tong1, Nhu-An Pham2,3, Geoffrey Liu3,4, Brian Raught3,5, Bradly G. Wouters2,3,5,6, Thomas Kislinger3,5, Ming Sound Tsao2,3,7, and Michael F. Moran1,3,8

Abstract

Coiled-coil-helix-coiled-coil-helix domain-containing 2, a tary approaches of affinity purification mass spectrometry and mitochondrial protein, encoded by CHCHD2 is located at in vivo proximity ligation. The CHCHD2 interactome includes chromosome 7p11.2 and proximal to the EGFR gene. Here, the apparent hub proteins C1QBP (a mitochondrial protein) bioinformatic analyses revealed that CHCHD2 is consistently and YBX1 (an oncogenic transcription factor), and an over- coamplified with EGFR in non–smallcelllungcarcinoma lapping set of hub-associated proteins implicated in cell (NSCLC). In addition, CHCHD2 and EGFR protein expression regulation. levels were positively correlated and upregulated relative to normal lung in NSCLC tumor-derived xenografts. Knockdown Implications: CHCHD2 influences mitochondrial and nu- of CHCHD2 expression in NSCLC cells attenuated cell prolif- clear functions and contributes to the cancer phenotype asso- eration, migration, and mitochondrial respiration. CHCHD2 ciated with 7p11.2 amplification in NSCLC. Mol Cancer Res; protein–protein interactions were assessed by the complemen- 13(7); 1119–29. 2015 AACR.

Introduction NSCLC with the EGFR inhibitor erlotinib (8). Despite advances in the identification and pharmacologic modulation of cancer Lung cancer is the most common cause of cancer-related drivers such as the EGFR, and improved responses to combi- mortality in the world (1). Eighty-five percent of lung cancers nation therapies, NSCLC outcomes remain poor. Contributing are non–small cell lung carcinoma (NSCLC), a heterogeneous to this, tumor heterogeneity and genetic complexity are signif- group composed of two main subtypes, squamous cell carcinoma icant factors that confound NSCLC tumor classification and (SCC) and adenocarcinoma (ADC; ref. 2). NSCLC are character- treatment (2–4, 9). ized by somatic genetic alterations that recur according to histol- Regions of DNA amplification can be up to 1 Mb in size, and ogy, and which include oncogenic driver mutations that in therefore often encompass neighboring genes in addition to the some instances are associated with genomic amplification or suspected driver gene (10–12). It has been suggested that genes deletion (3–6). For example, in NSCLC, recurrent amplification coamplified within a given amplicon could be functionally rel- is observed in the 7p11 region encoding the drug target EGFR evant (12) and that coamplified genes may cooperate to promote (3, 6, 7). Indeed, EGFR gene copy number is a significant molec- tumor progression (2, 10). Indeed, in a series of NSCLC-derived ular predictor of a differential survival benefit from treatment of cell lines, several genes mapping proximal to EGFR, including CHCHD2, have been reported as coamplified and with elevated mRNA expression (13). Similarly, a recent integrated omic anal- 1 Program in Molecular Structure and Function, Hospital For Sick Chil- ysis of NSCLC indicated consistent upregulation of proteins dren, Toronto, Ontario, Canada. 2Campbell Family Institute for Cancer Research, Toronto, Ontario, Canada. 3Princess Margaret Cancer Cen- genetically linked to EGFR within the recurrent 7p11.2 amplicon, tre, Toronto, Ontario, Canada. 4Department of Medical Oncology, including the chaperonin subunit CCT6A, and CHCHD2 (14). 5 University Health Network, Toronto, Ontario, Canada. Department of However, the genes frequently coamplified with EGFR, including Medical Biophysics, University of Toronto, Toronto, Ontario, Canada. 6Department of Radiation Oncology, University of Toronto, Toronto, CHCHD2, and their protein products have not been systemati- Ontario, Canada. 7Department of Laboratory Medicine and Pathobi- cally investigated for roles in NSCLC. 8 ology, University of Toronto,Toronto, Ontario,Canada. Department of According to The Human Protein Atlas, the CHCHD2 gene Molecular Genetics, University of Toronto, Toronto, Ontario, Canada. product is a widely expressed 16.7-kDa mitochondrion-localized Note: Supplementary data for this article are available at Molecular Cancer protein that is upregulated in lung cancer (15). CHCHD2 consists Research Online (http://mcr.aacrjournals.org/). of an N-terminal mitochondrion localization sequence and Corresponding Author: Michael F. Moran, Hospital For Sick Children, 686 Bay a strongly conserved C-terminal CHCH domain, which is char- Street, Toronto, ON, M5G 0A4, Canada. Phone: 647-235-6435; Fax: 416-813- acterized by twin CX C motifs in a CX CXnCX C configuration. 5993; E-mail: [email protected] 9 9 9 Twin CX9C proteins, such as CHCHD2, have been proposed to doi: 10.1158/1541-7786.MCR-14-0165-T function as scaffolding proteins in the mitochondrion, and 2015 American Association for Cancer Research. have been identified as part of respiratory chain complexes,

www.aacrjournals.org 1119

Wei et al.

cytochrome c oxidase (COX) assembly factors, or involved in the duced CHCHD2 knockdown (KD) and nonsilencing control maintenance of mitochondrion structure and function (16–18). (NSC) cell lines were selected by growth in medium containing Depletion of CHCH proteins decreases cellular oxygen consump- 2 mg/mL puromycin. The rescue/overexpression cell lines (RES/ tion and ATP production (16, 18, 19). Recently, CHCHD2 was OE) were established by ectopic expression of an shRNA-resistant predicted through computational expression screening, and then human CHCHD2 cDNA in the knockdown lines: CHCHD2 full- experimentally validated, as a regulator of oxidative phosphory- length cDNA, resistant to pGIPZ-CHCHD2 shRNAmir by intro- lation (oxphos; ref. 20). It was further demonstrated that duction of a silent mutation, was subcloned into pLX303 CHCHD2 could bind to the promoter of COX subunit 4 isoform (from Dr. Sergio Grinstein, Hospital for Sick Children), and used 2 (COX4I2) and interact with two other transcription factors to prepare virus particles. Transduced, blasticidin-resistant (RBPJ, CXXCC5) in regulating COX4I2 gene expression under CHCHD2-KD cells (derived from HCC827 and LPC43) were the influence of oxygen concentration (21). The CHCHD2 cDNA selected and verified for CHCHD2 expression. was found to promote cell migration and alter cell adhesion when ectopically overexpressed in NIH3T3 fibroblasts (22). Dysregu- Affinity purification and Western blotting lated metabolism and altered migration/invasion are pivotal For the isolation of immunoprecipitates (IPs) containing factors for tumor initiation and progression (23). Therefore, the CHCHD2 protein complexes, HEK293 cells were transiently EGFR-linked CHCHD2 gene product is implicated as a factor transfected with Flag-tagged CHCHD2. Cells were lysed in NP- contributing to the cancer phenotype in NSCLC. 40 buffer [50 mmol/L Tris–HCl, pH 7.5, 150 mmol/L sodium In this study, the role of CHCHD2 in NSCLC was addressed by chloride, 1% (v/v) Nonidet P-40, 100 mmol/L NaF, 1 mmol/L analysis of primary NSCLC-derived xenografts, and tumor- sodium orthovanadate, and protease inhibitors; ref. 24], clar- derived cell lines. An integrative approach including measure- ified, and subject to immunoaffinity purification by using ment and modulation of CHCHD2 protein expression, affinity immobilized anti-FLAG M2 agarose beads (Sigma-Aldrich). purification-mass spectrometry (AP-MS), in vivo proximity liga- For coimmunoprecipitation (co-IP), immune complexes were tion, and computational methods was used to gain insight into eluted from beads by using 2X Laemmli sample buffer, fol- the interactions and function of CHCHD2. The results suggest that lowed by Western blotting. Western blotting of total cell pro- CHCHD2 functions in the regulation of NSCLC cell growth, tein extracts (15 mg protein/sample) or IP eluates involved migration, and mitochondrial function via complex protein– SDS–PAGE followed by electrophoretic transfer to nitrocellu- protein interaction networks. lose membranes (Amersham Bioscience), blocking, and prob- ing with primary and secondary antibodies as previously described (24). For IP followed by liquid chromatography- Materials and Methods tandem mass spectrometry (LC/MS-MS), isolated immune Cells, constructs, and reagents complexes were dissociated by using 0.15% trifluoroacetic acid The established NSCLC cell line HCC827 and the xenograft (TFA), neutralized, and then immediately processed for down- derived lung primary cell line LPC43 (Supplementary Table S1) stream proteomic analysis (25). were maintained in RPMI-1640 supplemented with 10% fetal bovine serum, 50 U/mL penicillin, and 50 mg/mL streptomycin Immunofluorescence staining and confocal microscopy (Invitrogen). HEK293 cells were grown in DMEM with 10% FBS LPC43 and Hela cells were grown on coverslips for 24 hours. and penicillin/streptomycin. All cells were cultured at 37 Cina For transient CHCHD2-flag expression, Hela cells were trans- 5% CO2-humidified incubator. fected by using Lipofectamine reagent (Life Technologies), and The plasmid containing human CHCHD2 full-length cDNA then incubated for 24 hours. For CHCHD2 imaging, cells were was obtained from the SPARC BioCentre (Hospital for Sick grown on coverslips submerged in a 24-well plate until subcon- Children, Toronto, ON, Canada). To generate a C-terminal fluent. The mitochondria of live cells were stained with Mito- fi Flag-tagged construct, CHCHD2 cDNA was ampli ed by PCR Tracker Red CMXRos (red color) for 30 minutes (37 C, 5% CO2). using specific primers with a linker encoding the Flag tag (Sup- After paraformaldehyde fixation, cells were permeabilized with plementary Table S1). The PCR product was subcloned into a 0.2% Triton X-100, incubated with rabbit antibody to CHCHD2 pcDNA3.1-Myc/His expression vector (Invitrogen). All expression (Sigma-Aldrich; HPA027407) and then detected with Alexa Fluor vectors were sequenced to confirm their authenticity. 647–conjugated secondary antibodies (far-red). Confocal micros- Antibodies used were rabbit polyclonal anti-CHCHD2 copy was performed by using a Zeiss LSM 510 META laser (HPA027407; Sigma-Aldrich), mouse monoclonal anti-FLAG scanning microscope equipped with a 64 objective. Green M2 (Sigma-Aldrich), rabbit polyclonal anti-EGFR (SC-03; Santa fluorescence was an indication of lentivirus infection. Counter- Cruz Biotechnology), rabbit monoclonal anti-C1QBP (6502S; staining of DNA with 40,6-diamidino-2-phenylindole (DAPI) was Cell Signaling Technology) and rabbit monoclonal anti-YB1 used to localize cell nuclei (blue). Additional details on experi- (9744; Cell Signaling Technology). All other reagents were pur- mental protocols were published previously (24, 26). chased from Sigma-Aldrich unless otherwise specified. Liquid chromatography-tandem mass spectrometry analysis Generation of CHCHD2 knockdown and rescue stable cell lines MS sample preparation was performed essentially as described Stable knockdown at the protein level was achieved by using previously (27). Tryptic peptides were concentrated and purified shRNAi: pGIPZ-GFP-human CHCHD2-shRNAmir (Clone ID: on homemade C18 columns or C18 StageTips (Thermo Fisher V2LHS_97752) and pGIPZ-GFP-scrambled shRNAmir lentiviral Scientific) before LC/MS-MS. For MS analysis, peptides were vectors were sourced from Open Biosystems. Lentiviral particles, separated by reverse-phase chromatography using a nanoflow prepared by SPARC BioCentre, were used to infect target cells UPLC system (Thermo Fisher Scientific) with a 240-minute linear (HCC827, LPC43) following standard protocols. The stably trans- gradient. The UPLC system was coupled to an Orbitrap Elite MS

1120 Mol Cancer Res; 13(7) July 2015 Molecular Cancer Research

Functional Roles of CHCHD2 in NSCLC

instrument (Thermo Fisher Scientific), and peptides were frag- according to the manufacturer's instructions. For growth curves, mented by collision-induced dissociation (CID). cells were seeded in triplicate into 96-well E-plates, and the cell index, reflecting electrical impedance was measured every 15 MS data processing and analysis minutes throughout the experiment. For migration, cells were Raw MS files acquired from the Orbitrap-Elite were processed by serum-starved for 16 hours, trypsinized, resuspended in serum- MaxQuant software (version 1.3.0.5) according to the standard free media and then seeded in triplicate on the top chamber of workflow (28). MS/MS spectra were searched against the UniProt CIM-16 plates. Serum-containing medium was added to the lower human proteome (release 2013_03_06) containing 87,656 entries chamber as a migratory stimulant. Cells were allowed to settle for (including common contaminants) by using the Andromeda 30 minutes before measurements were taken. The cell index was search engine (29). For statistical evaluation of the data, a false collected at 5-minute intervals. Proliferation and migration rates discovery rate of 0.01 was set for peptide and protein identification. were determined from the slope of cell index curve. MS data related to NSCLC primary tumors were accessed from ProteomeXchange (http://www.proteomexchange.org) with the Oxygen consumption and extracellular acidification dataset identifier PXD000853, and analyzed by using MaxQuant measurement as described in Li and colleagues (14). Protein LFQ (label-free To measure mitochondria function in intact cells, a Seahorse quantification) intensity obtained from MaxQuant was chosen as Bioscience XF96 Extracellular Flux Analyzer was used. Briefly, cells the quantitative value representing protein abundance, and used were seeded in XF96-well microplates (25,000/well) with RPMI- for calculation of protein differential expression. For quantitative 1640 complete medium, and incubated for 16 hours. The pH of proteomics, protein LFQ intensities were exported and statistical medium was adjusted to 7.4, and final concentrations of glucose analysis was performed by using the R software package. For and glutamine were 11 mmol/L and 2 mmol/L, respectively. Cells quantitative mass spectrometric analysis of anti-Flag IPs, protein were subsequently washed and restored with bicarbonate-free fi LFQ intensities across different samples were rst normalized medium and incubated in a CO2-free incubator for 2 hours. according to the intensities of the bait protein CHCHD2 in each Oxygen consumption rate and extracellular acidification rate were sample. Normalized LFQ intensities were then used for determi- measured under basal conditions, in the presence of the ATP- nation of specific protein–protein interactions by using Perseus synthase inhibitor oligomycin (1 mmol/L) and the mitochondri- software tools available in the MaxQuant environment. on uncoupling agent, FCCP (500 nmol/L). Data were normalized according to the number of seeded cells. Identification of CHCHD2 interactions by proximity- dependent biotinylation (BioID) and MS Statistical analysis and bioinformatics To identify proximate and interacting proteins in living cells, The significance of difference between groups was determined the BioID method (30) was used. By this approach, proteins in the by the Student t test using R software or by one-way ANOVA – vicinity of the FlagBirAR118G CHCHD2 fusion protein were cova- followed by Bonferroni post hoc test using GraphPad Prism 3.0 lently modified by biotin in vivo, followed by MS identification of (GraphPad Software Inc.), unless stated otherwise. biotinylated proteins isolated by streptavidin-based affinity cap- Gene ontology (GO) enrichment analysis was performed by ture in vitro. In brief, full-length CHCHD2 coding sequence was using DAVID bioinformatics resources (33). A P value less than subcloned into the pcDNA5/FRT/TO-FlagBirA expression vector. 0.05 was considered statistically significant. Protein interaction HEK293T-REx cells stably expressing FlagBirA -CHCHD2 or network analysis was based on literature investigations and the CHCHD2-BirA Flag were generated by using the Flp-In system Biological General Repository for Interaction Datasets (BioGRID; (Invitrogen). Cells were incubated for 24 hours in complete ref. 34). Cytoscape version 3.0 (35) was used for visualization of medium supplemented with 1 mg/mL tetracycline (Sigma- protein interaction networks. Aldrich) and 50 mmol/L biotin (Bioshop), and then lysed in RIPA buffer(50 mmol/L Tris–HCl pH 7.4, 150 mmol/L NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 1 mmol/L EDTA, Results protease inhibitors). Clarified cell lysates were subject to strepta- Correlated expression of EGFR and CHCHD2 in NSCLC vidin affinity purification followed by trypsin digestion and MS As EGFR overexpression in NSCLC is often a consequence of analysis, essentially as described previously (31). Detailed infor- gene amplification (36), we examined whether the EGFR-proxi- mation is included in Supplementary Methods. mal gene CHCHD2 is coamplified. Data on somatic copy number variation (SCNV) in lung adenocarcinoma (LADC) and lung SCC CHCHD2 quantification by selected reaction monitoring-MS (LSCC) were retrieved from The Cancer Genome Atlas (TCGA) Selected reaction monitoring (SRM) was performed as data portal (www.cancergenome.nih.gov) and through cBioPro- described (27) with slight modifications. SRM peak intensities tal (37). This indicated that CHCHD2 is indeed frequently coam- were calculated by using Skyline software (32), and signal inten- plified along with EGFR in both subtypes, and with a strong sity of each peptide was the average of two technical replicates. tendency toward cooccurrence (odds ratio >>10; Fig. 1A). As Total intensities from three CHCHD2 peptides (Supplementary shown in Fig. 1A, it is apparent that EGFR as expected, but not Table S2) were summed and multiplied by a correction factor, CHCHD2, is in some instances mutated in addition to being which was determined by normalization of the sum of MYH9 and amplified. We recently reported that EGFR and CHCHD2 protein ACTG total peak intensities across the different samples. levels are upregulated in primary NSCLC (14). Figure 1B shows that the levels of EGFR and CHCHD2 are consistently elevated Cell proliferation and migration assays relative to normal lung in a set of 11 NSCLC patient-derived Real-time cell proliferation and migration activities were mon- xenograft (PDX) tumors (listed in Supplementary Table S1), and itored by using the xCELLigence system (AECA Biosciences) show a strong correlation in their expression levels, with a

www.aacrjournals.org Mol Cancer Res; 13(7) July 2015 1121

Wei et al.

coefﬁcient of determination (Pearson R2) > 0.6 (Fig. 1C). Ectopic overexpression of EGFR in HEK293 cells, an established model system for EGFR regulation (e.g., 38), did not upregulate endogenous CHCHD2 protein expression concomitantly (Fig. 1D). This suggests that the correlated expression of the two gene products that occurs at the protein level, at least in NSCLC contexts, extends beyond their genetic linkage to the protein level, but, as shown by the HEK293 experiment (Fig. 1D), does not appear to reﬂect increased synthesis or stability of CHCHD2 as a product of elevated EGFR expression. Western blot analysis of four NSCLC cell lines, including LPC43 and LPC72, which were generated from NSCLC PDX sample A6 (see Supplementary Table S1), and the established cell lines RVH6849 and HCC827, showed that EGFR and CHCHD2 were coexpressed in vitro (Supplementary Fig. S1).

Modulation of CHCHD2 expression alters NSCLC cell proliferation and migration The coamplification and elevated protein expression of EGFR and CHCHD2 in NSCLC led us to consider whether CHCHD2 expression contributes to key aspects of the cancer phenotype, including cell proliferation and migration. The NSCLC cell lines HCC827 and PDX A6-derived LPC43 were used to generate cell lines stably expressing (i) NSC shRNA, (ii) shRNA directed against human CHCHD2 (knockdown), and (iii) CHCHD2 knockdown cells "rescued" with ectopic overexpression of an shRNA-resistant human CHCHD2 cDNA (RES/OE). CHCHD2 protein in LPC43- derived cells, was mainly localized in mitochondria, as determined by colocalization with MitoTracker, whereas CHCHD2 staining was minimal on knockdown cells (Fig. 2). Consistent with the cell imaging results, quantification of CHCHD2 by SRM- MS analysis of three different CHCHD2 peptide ions confirmed that protein expression was greatly reduced in knockdown and increased in RES/OE lines relative to NSC cells (Fig. 3A). In HCC827 knockdown cells, the level of CHCHD2 was 7% of the level measured in NSC, and in RES/OE the amount of CHCHD2 was approximately 2.4-fold greater than NSC. In LPC43 the Figure 1. amount of CHCHD2 in knockdown was 12% of that seen in CHCHD2 and EGFR protein expression in NSCLC. A, amplification NSC, and RES/OE expressed 6.7-fold more than RES (Fig. 3A). of EGFR and CHCHD2 genes in NSCLC subtypes. Shown are OncoPrint The impact of CHCHD2 depletion and overexpression on outputs (cBioPortal for Cancer Genomics; www.cbioportal.org) where each bar represents a tumor that was found to contain a DNA NSCLC cell proliferation and migration were considered. Cell alteration (amplification, deletion, mutation, as indicated) in EGFR and proliferation kinetics were measured in real-time by monitoring CHCHD2 in the indicated NSCLC subtypes; note that aligned bars the electrical impedance of the substrate surface, which is affected represent the same tumor. Data for LADC and LSCC are from TCGA by the area of cell contact, and hence cell number (Fig. 3B). In order resource, obtained through cBioPortal. For LADC, 40 of 230 cases (17%) to measure cell migration, a Transwell migration assay based on contained a DNA alteration in the EGFR gene, while the incidence for fi cell-surface impedance was used (Fig. 3C). In HCC827, there was CHCHD2 was 7%. In 13 of 15 cases, the genes were coampli ed; odds fi ratio ¼ 101.8, indicating a strong tendency toward cooccurrence not a signi cant change in cell proliferation when CHCHD2 was of EGFR and CHCHD2 amplification. For LSCC 16 of 178 cases (9%) knocked down, but there was a significant increase in the over- contained EGFR amplification or mutation, and an equal number of cases expressing RES/OE cells compared with knockdown. In LPC43, (13) contained amplification of both EGFR and CHCHD2. Odds ratio ¼ CHCHD2 knockdown significantly reduced the rate of prolifera- infinite, indicating a strong tendency toward cooccurrence of tion, but proliferation remained impaired in the RES/OE cells. In EGFR and CHCHD2 amplification.EGFRasexpected,butnotCHCHD2,is HCC827 and LPC43, migration was significantly decreased in in some instances mutated in addition to being amplified, as indicated by the green features. B, LC/MS-MS analysis of 11 human tumor- CHCHD2 knockdown cells. This effect was not recued by in derived NSCLC xenografts (as indicated; A, adenocarcinoma; S, HCC827 RES/OE, but was in LPC43. Therefore, in both cell types squamous cell carcinoma). C, EGFR and CHCHD2 expression were HCC827 and LPC43, measures of proliferation and migration highly correlated as indicated by correlation of determination analysis associated with CHCHD2 expression showed similar trends toward 2 (R ), as indicated. After removing sample S3 that was the only sample decreases in knockdown compared with NSC and/or RES/OE. with a cytogenetically defined amplification of EGFR/7p11.2, the two proteins remained moderately correlated in their expression level (R2 ¼ 0.35). D, Western blot analysis of CHCHD2 and EGFR expression Mitochondrial respiration as a function of CHCHD2 in NSCLC in HEK293 cells (control), or this cell type stably expressing ectopic Metabolic capacity in NSC, knockdown, and RES/OE cells was EGFR (EGFR). Transferrin receptor(TfR)wasdetectedasaninput assessed by using an extracellular flux analyzer. The parameters control. measured were basal oxygen consumption rate (OCR), an index

1122 Mol Cancer Res; 13(7) July 2015 Molecular Cancer Research

Functional Roles of CHCHD2 in NSCLC

The effect of CHCHD2 expression on NSCLC proteomes To elucidate the molecular changes that contribute to the phenotypes observed, MS-based quantitative proteome analyses were performed to identify proteins differentially expressed as a function of CHCHD2 expression in the LPC43- and HCC827- based, NSC, knockdown, and RES/OE cell lines. Three independent analyses of each cell line were completed and in aggregate a total of 4,308 nonredundant proteins were identified and quan- tified. An overlap of 70% to 75% (at the protein level) was observed between the independent replicate experiments (Sup- plementary Fig. S3). After pooling the respective NSC, knockdown, and RES/OE datasets derived from LPC43 and HCC827, a majority of proteins (3,249; 75%) was shared by the NSC, knockdown, and RES/OE groups (Fig. 5A). Statistical analysis (paired t test) identified 205 proteins as significantly differentially expressed (P < 0.05) between knockdown and NSC cells (Sup- plementary Table S3). GO analysis revealed that the differentially expressed proteins are implicated in diverse biologic processes (Fig. 5B), and distributed across various cellular compartments, primarily mitochondrion, nuclear lumen, ribonucleoprotein complex, and cytoskeleton (Supplementary Fig. S3). In order to more stringently identify candidates whose expression levels changed upon CHCHD2 depletion, the list of differentially expressed proteins was filtered to include only proteins showing 2-fold expression difference between NSC and knockdown, but not between NSC and RES/OE. This produced a smaller set of 13 proteins, excluding CHCHD2, regarded as CHCHD2 regulated proteins (Table 1). Ten proteins showed parallel expression changes with CHCHD2 expression: downregulation in CHCHD2 knockdown cells, and recovered expression in CHCHD2 RES/OE cells (Table 1). This set includes two protein kinases: the tyrosine-specific anaplastic lymphoma kinase (ALK), and the MAP2K4-encoded dual specificity enzyme known as Figure 2. Expression and localization of CHCHD2 in NSCLC cells. Fluorescence confocal mitogen-activated protein kinase kinase 4 (MKK4). Substrates microscopy was used to determine the localization of endogenous and activated by MKK4 include the MAP kinases JNK1 and -2 and the ectopic CHCHD2 in LPC43 cells. CHCHD2 knockdown, rescue/overexpressed stress-activated MAP kinase p38 (25). In contrast, the expression (RES/OE) and control (NSC) cells were grown on coverslips until sub- levels of three proteins, including ATP-dependent Clp protease confluent, and then live cells were stained for mitochondria by using (CLPP), WD repeat-containing protein 20 (WDR20) and Agrin fi MitoTracker (red). After xation, permeabilized cells were probed with anti- (AGRN), were increased upon CHCHD2 knockdown (Table 1). CHCHD2 and an Alexa Fluor 647 secondary antibody (far-red; imaged cyan). Green fluorescent protein (GFP), encoded by the shRNA pGIPZ lentiviral vector, was used as a marker of transduced cells (top left). DAPI staining was Proteomic analysis of the CHCHD2 protein interactome used to stain nuclei in some samples. Pink indicates colocalization of As a putative scaffolding protein and transcription factor, mitochondria and CHCHD2 in NSC and RES/OE cells (right-middle, and right- CHCHD2 is expected to function through protein–protein inter- bottom), which was largely absent in knockdown cells (bottom-left). Scale actions. To test this, CHCHD2-associated proteins were charac- m bars are 15 m. terized by affinity purification mass spectrometry (AM-MS). C- terminal Flag-tagged CHCHD2 constructs were ectopically of oxphos, along with FCCP-stimulated maximal oxygen con- expressed in transiently transfected cells, which were verified by sumption, and mitochondrial reserve capacity. In both cell types immunoblot and immunostaining analysis (Supplementary Fig. (HCC827 and LPC43) basal and maximal OCR and mitochon- S4). Anti-Flag IPs were prepared from cells expressing epitope- drial reserve capacity were significantly decreased in knockdown tagged CHCHD2 and control cells not expressing ectopic cells compared with NSC controls (Fig. 4A and B). In both of the CHCHD2. The criterion for a specific CHCHD2-interacting pro- RES/OE cell types, these parameters showed a general trend tein was differential recovery relative to the negative control (fold intermediate between control NSC and knockdown values, but enrichment 32, and P 0.01; Fig. 6A; Supplementary Fig. S4). were not significantly different from either. This suggests that in By this criterion, 58 proteins were identified as binding directly or terms of mitochondrial function, overexpression of CHCHD2 in indirectly to CHCHD2 (Supplementary Table S4). RES/OE cells did not achieve a true "rescue" effect. Analysis of To further validate the CHCHD2-interacting proteins, a recently additional RES clones with CHCHD2 expression closer to endog- developed proximity-dependent in vivo biotinylation method enous levels might address this possibility. Basal extracellular termed BioID (30) was applied in combination with MS to identify acidification rate (ECAR), a marker for glycolysis was not signif- proteins that interact or come into close proximity with CHCHD2 icantly changed as a function of CHCHD2 in the two cell lines within cells. Following the BioID methodology, CHCHD2 was (Supplementary Fig. S2). expressed in transfected cells as a fusion protein including the

www.aacrjournals.org Mol Cancer Res; 13(7) July 2015 1123

Wei et al.

Figure 3. SRM-MS quantiﬁcation of CHCHD2. A, CHCHD2 protein was measured in HCC827 and LPC43 NSC, knockdown, and RES/OE cells by using SRM-MS. The effect of CHCHD2 protein expression on cell proliferation (B) and cell migration was measured (C). Rates, expressed in arbitrary units (a. u.) per hour, were calculated from the slopes of cell index curves. Data are mean SD for three independent peptides (A), or SEM from n 3 independent experiments (B and C). , P < 0.05.

promiscuous biotin protein ligase domain BirA fused to the which were previously reported as CHCHD2-interacting proteins carboxyl end of CHCHD2. A total of 65 proteins that became (22, 41–43), five other CHCHD2-interacting candidates were biotinylated in vivo due to their proximity with CHCHD2-BirA identified by both AP-MS and proximity ligation including the passed stringent criteria applied by using SAINT software (Fig. 6A uncharacterized protein coiled-coil domain containing 71-like and Supplementary Table S5; ref. 39). Seven proteins were iden- (CCDC71L); cytidine monophosphate N-acetylneuraminic acid tified by both the AP-MS and proximity ligation methods, suggest- synthetase (CMAS), the inner nuclear membrane protein LBR; the ing they are central, or at least high-confidence CHCHD2-associ- putative RNA helicase DHXS7; and the nuclear chaperone THOC4 ated proteins (Fig. 6B). Among the remaining CHCHD2-interact- (encoded by ALYREF). Two previously reported CHCHD2-inter- ing proteins that were identified by only one of the methods (i.e., acting proteins RPL24 and polyubiquitin-C (UBC; ref. 41) also AP-MS or BioID), interrogation of BioGRID (34) verified eight as were detected by AP-MS, but did not pass the selection criteria. having interactions with one or more of the seven high-confidence Further support for the interaction of CHCHD2 with YBX1 and CHCHD2-associated proteins. These proteins were organized and C1QBP was obtained by demonstration of associations between visualized by using Cytoscape tools (Fig. 6B; ref. 35). This revealed ectopic CHCHD2 and endogenous YBX1 and C1QBP by Western two highly connected protein subclusters that were centered on blot analysis of CHCHD2-Flag IPs recovered from transfected proteins C1QBP and YBX1, suggesting that they may function as HEK293 cells (Fig. 6C). hubs within a CHCHD2 network. The protein CAND1 was common to both the C1QBP and YBX1 complexes. It is reportedly a transcription factor, and also an F-box protein exchange factor, Discussion influencing the association of diverse F-box proteins in SCF ubi- Recurrent cancer-associated amplifications and deletions are quitin ligase complexes (40). In addition to C1QBP and YBX1, logically thought to be driven by changes in expression of encoded

1124 Mol Cancer Res; 13(7) July 2015 Molecular Cancer Research

Functional Roles of CHCHD2 in NSCLC

Figure 4. Dependence of mitochondrial respiration on CHCHD2 protein expression. Control (NSC), knockdown, and rescue/ overexpression (RESOE) cells were seeded in specialized microplates and cultured for 16 hours. Cells were then switched to bicarbonate-free medium and mitochondrial function was assessed using sequential injection of oligomycin and FCCP. Shown are representative OCR curves and summarized quantiﬁcation of basal OCR, maximal OCR, and reserve capacity of (A) HCC827 and (B) LPC43. Results are mean SEM from three independent experiments. , P < 0.01; , P < 0.05.

oncogenes and tumor suppressors, respectively. Integrated anal- primary NSCLC xenografts. However, in HEK293 cells, CHCHD2 ysis of NSCLC proteomes and somatic gene copy number vari- protein expression was not increased as a product of elevated ation suggests that genomes are organized such that coamplified ectopic EGFR expression. This negates a model, at least in the genes are functionally linked to the cancer phenotype (14). HEK293 system, wherein CHCHD2 protein expression is itself Coamplification of EGFR and CHCHD2 has been reported in elevated as a product of EGFR protein expression. Although we do NSCLC (13) and other cancer types, including glioma (12). not yet understand the mechanisms regulating EGFR or CHCHD2 However, not all amplified genes in tumors are overexpressed protein levels, the observations that both proteins appear con- (12). Our analysis of TCGA datasets clearly indicates that EGFR sistently upregulated in NSCLC (Fig. 1B; ref. 14) provided a and CHCHD2 are indeed coamplified in both the ADC and SCC rationale for experiments to address whether CHCHD2 contri- subtypes of NSCLC. Consistent with this trend, we found that butes to the transformed cell phenotype. CHCHD2 and EGFR protein expression levels are elevated relative Cancer metabolism is known to affect cell migration (44). to normal lung and are positively correlated with each other in CHCHD2 was implicated as a regulator of oxphos (20), and

www.aacrjournals.org Mol Cancer Res; 13(7) July 2015 1125

Wei et al.

Figure 5. Proteome changes associated with CHCHD2 expression in NSCLC cells. A, Venn diagram depicting the number of nonredundant proteins identiﬁed in CHCHD2 knockdown, rescue/overexpression (RES/OE), and control (NSC) groups. Label-free quantitative proteome analysis was performed on six NSCLC cell lines (NSC, knockdown, and RES/OE derivatives of HCC827 and LPC43). B, enriched GO biologic process terms from proteins differentially expressed between CHCHD2 knockdown and control cells.

ectopic CHCHD2 promoted NIH3T3 cell migration (22). Con- cell proliferation, cell cycle, and cell projection morphogene- sistent with these observations, we found that cell migration and sis). For example, Agrin protein expression increased 2-fold proliferation rates in the NSCLC cell lines HCC827 and LPC43 upon CHCHD2 knockdown and this effect was reversed by were stimulated by CHCHD2 expression. Furthermore, knock- CHCHD2 rescue. Overexpression of Agrin increased substrate down of CHCHD2 significantly attenuated mitochondrial respi- adhesion in COS7 cells (47). We therefore speculate that ratory activity, typified by reduced basal and maximum OCR, and elevated Agrin may have stimulated cell adhesion, and conse- diminished mitochondrial reserve capacity. In breast cancer cells, quently hindered cell migration, in response to CHCHD2 loss of reserve capacity is linked to apoptosis (45), and mainte- knockdown. These results are consistent with observations of nance of reserve capacity positively correlates with rapid prolif- increased cell motility associated with EGFR amplification eration (46). Therefore, we speculate that the impaired NSCLC (48), and suggest that coamplification of CHCHD2 may con- cell proliferation associated with CHCHD2 knockdown was, at tribute to this cancer-associated phenomenon. least in part, due to the measured reduction in mitochondrial The receptor tyrosine kinase ALK is activated as a conse- reserve capacity. Consistent with this, our comprehensive analysis quence of somatic chromosomal rearrangements in a subset of NSCLC cell proteomes revealed proteins that changed in (3%–7%) of ADC-subtype NSCLC (49). Knockdown of abundance as a function of CHCHD2 knockdown and rescue, CHCHD2 was accompanied by reduced expression of ALK; and including significant enrichment for biologic processes however, to the best of our knowledge, wild-type ALK, which involved in cell proliferation and migration (i.e., GO categories isknowntofunctioninbraindevelopmentandneural

1126 Mol Cancer Res; 13(7) July 2015 Molecular Cancer Research

Functional Roles of CHCHD2 in NSCLC

Table 1. Proteins differentially expressed between CHCHD2 knockdown and NSC

Gene name Protein identifier Protein description P Log2[KD:NSC] CHCHD2 Q9Y6H1 Coiled-coil-helix-coiled-coil-helix domain-containing protein 2 0.000 4.652 HIP1R O75146 Huntingtin-interacting protein 1-related 0.043 2.422 UFL1 O94874 E3 UFM1-protein ligase 1 0.006 1.436 CTPS2 Q9NRF8 CTP synthase 2 0.036 1.326 GSTT1 P30711 Glutathione S-transferase theta-1 0.012 1.274 MFAP1 P55081 Microfibrillar-associated protein 1 0.012 1.216 ACOT1/2 P49753 Acyl-coenzyme A thioesterase (1-cytoplasm, 2-mitochondrial) 0.030 1.124 MAP2K4 P45985 Dual specificity mitogen-activated protein kinase kinase 4 0.023 1.123 QTRT1 Q9BXR0 Queuine tRNA-ribosyltransferase 0.048 1.094 AATF Q9NY61 Protein AATF 0.015 1.061 ALK Q9UM73 ALK receptor tyrosine kinase 0.014 1.024 AGRN O00468 Agrin 0.001 1.016 WDR20 E7EUY8 WD repeat-containing protein 20 0.047 1.529 CLPP Q16740 Putative ATP-dependent Clp protease proteolytic subunit, mitochondrial 0.027 1.610 function, has not been implicated as an oncogenic driver. study, our results suggest that coamplification of CHCHD2 and Interestingly, recent observations of concomitant EGFR muta- EGFRmaybelinkedtoALK. tion and ALK gene rearrangement has suggested that coordi- In addition to Agrin, the expression of other proteins was nated activation of the signaling networks of EGFR and ALK affected by CHCHD2 protein levels (Table 1). WDR20 displayed may be important in lung ADC (50). Although not tested in this increased expression upon CHCHD2 depletion and dropped back

Figure 6. Analysis of CHCHD2 interactome by MS approach. A, schematic depiction of the experimental plan for the identification of CHCHD2 associated proteins. In one arm, CHCHD2-associated proteins were identified by AP-MS analysis of proteins recovered by anti-Flag immunoprecipitation of ectopically expressed CHCHD2-Flag. In the other arm, the BioID method was used to characterize by MS proteins that became covalently modified by ubiquitin in cells expressing a fusion protein of CHCHD2 linked to the promiscuous ubiquitin ligase BirA*. Seven proteins in addition to the CHCHD2 bait proteins were identified by both approaches. B, network analysis of the CHCHD2 interactome. Proteins and their interactions are shown as nodes and edges. Interacting proteins were identified by AP-MS (light purple), BioID (orange), or both methods (pink), as indicated. Solid black lines indicate interactions also present in the BioGRID database search (http://thebiogrid.org). The dashed boxes encompass two potential CHCHD2-associated complexes. C, Western blot validation of CHCHD2 interacting proteins. Lysates and anti-Flag IPs from transfected HEK293 cells expressing ectopic CHCHD2-Flag and control cells not expressing CHCHD2-Flag were analyzed by SDS-PAGE and immuno-blotting. Expression of YBX1 and C1QBP was confirmed by analysis of lysates (lanes 1 and 2). Both proteins were recovered by co-IP with CHCHD2-Flag (lane 4), but not detected in anti-Flag IPs from control cells (lane 3).

www.aacrjournals.org Mol Cancer Res; 13(7) July 2015 1127

Wei et al.

toward control levels after CHCHD2 was reexpressed. WDR20 CHCHD2 gene copy number and protein levels are linked with interacts with and stimulates the activity of ubiquitin-specific EGFR in NSCLC. CHCHD2 participates in mitochondrial and protease USP12 and also interacts with USP46 (51). USP12 extra-mitochondrial protein–protein interactions and is an effec- regulates Notch signaling (52) and USP46 is a tumor suppressor tor of cell proliferation, migration, and respiration. Therefore, (53). However, this study did not address mechanisms through along with EGFR, it should be considered as a driver of 7p11.2 which CHCHD2 might affect protein production, modifications, amplification and the cancer phenotype. or stability. The proteins that changed in abundance in cells depleted of CHCHD2 relate to several different cellular functions Disclosure of Potential Conflicts of Interest and subcellular localizations other than mitochondrion. No potential conflicts of interest were disclosed. The CHCHD2 interactome was defined by the complementary approaches of AM-MS and proximity ligation. This analysis Authors' Contributions revealed two highly connected subnetworks centered on putative Conception and design: Y. Wei, B. Raught, B.G. Wouters, T. Kislinger, M.S. Tsao, hub proteins C1QBP and YBX1. The mitochondrial protein M.F. Moran C1QBP is upregulated in a variety of neoplasms, including lung Development of methodology: Y. Wei, R.N. Vellanki, E. Coyaud, J. Tong cancer (43, 54), and has promigration activity in cancer cell lines Acquisition of data (provided animals, acquired and managed patients, (43, 54). Moreover, C1QBP has been shown to promote cancer provided facilities, etc.): Y. Wei, E. Coyaud, J.R. Krieger, P. Taylor, N.-A. Pham, cell proliferation and resistance to cell death (54). YBX1 is an G. Liu, B. Raught, M.S. Tsao oncogenic transcription factor with roles in cell proliferation, Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Wei, R.N. Vellanki, V. Ignatchenko, L. Li, invasion, and metastasis as well as energy metabolism (55). It J.R. Krieger, P. Taylor, M.S. Tsao, M.F. Moran was suggested that nuclear translocation of YBX1 could induce Writing, review, and/or revision of the manuscript: Y. Wei, R.N. Vellanki, upregulation of EGFR expression and a more aggressive NSCLC N.-A. Pham, G. Liu, B.G. Wouters, M.S. Tsao, M.F. Moran phenotype (56). A recent report further revealed YBX1 as a Administrative, technical, or material support (i.e., reporting or organizing convergent hub for lysophosphatidic acid and EGF signaling, and data, constructing databases): Y. Wei, J. Tong, N.-A. Pham, G. Liu, M.S. Tsao activation of YBX1 contributed to ovarian cancer cell invasion Study supervision: M.S. Tsao, M.F. Moran (57). Considering the multifaceted roles of C1QBP and YBX1, Acknowledgments together with our observation that they present as "hubs" within The authors thank Drs. Sergio Grinstein and Robert Rottapel for DNA the CHCHD2 interactome, we propose that CHCHD2 functions, constructs. We gratefully acknowledge TCGA Research Network and their at least in part, through interactions with C1QBP and YBX1. Our research groups and specimen donors for allowing access to datasets. nonquantitative indirect immunofluorescence microscopic analysis most obviously depicted a mitochondrial localization of Grant Support CHCHD2, which may reflect its major subcellular localization This work was supported through funding provided by the Canada under the conditions of our analysis, but may also mean that non- Research Chairs Program (to T. Kislinger and M.F. Moran), the Ontario mitochondrial CHCHD2 was below our level of detection. The Research Fund (to M.S. Tsao), and the Canadian Institutes of Health Research changes in protein levels in response to CHCHD2 knockdown (M.F. Moran). The costs of publication of this article were defrayed in part by the payment of (Table 1), could be downstream effects stemming from the loss of page charges. This article must therefore be hereby marked advertisement in CHCHD2 function in mitochondria, and/or a consequence of accordance with 18 U.S.C. Section 1734 solely to indicate this fact. changes in gene expression related to the noted transcription factor activity of CHCHD2 (21), and possibly involving its inter- Received May 27, 2014; revised March 6, 2015; accepted March 7, 2015; actions with YBX1. In conclusion, our analyses indicate that published OnlineFirst March 17, 2015.

References 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin Cancer Institute of Canada Clinical Trials Group Study BR.21. J Clin Oncol 2013;63:11–30. 2008;26:4268–75. 2. Wang J, Qian J, Hoeksema MD, Zou Y, Espinosa AV, Rahman SM, et al. 9. Duffy MJ. The war on cancer: are we winning? Tumour Biol 2013;34:1275–84. Integrative genomics analysis identifies candidate drivers at 3q26-29 10. Kao J, Pollack JR. RNA interference-based functional dissection of the amplicon in squamous cell carcinoma of the lung. Clin Cancer Res 2013; 17q12 amplicon in breast cancer reveals contribution of coamplified genes. 19:5580–90. Genes Chromosomes Cancer 2006;45:761–9. 3. The Clinical Lung Cancer Genome Project (CLCGP) and Network Genomic 11. Scott D, Elsden J, Pearson A, Lunec J. Genes co-amplified with MYCN in Medicine (NGM). A genomics-based classification of human lung tumors. neuroblastoma: silent passengers or co-determinants of phenotype? Can- Sci Transl Med 2013;5:209ra153. cer Lett 2003;197:81–6. 4. Cardarella S, Johnson BE. The impact of genomic changes on treatment of 12. Vogt N, Gibaud A, Almeida A, Ourliac-Garnier I, Debatisse M, Malfoy B. lung cancer. Am J Respir Crit Care Med 2013;188:770–5. Relationships linking amplification level to gene over-expression in glio- 5. Ding L, Getz G, Wheeler DA, Mardis ER, McLellan MD, Cibulskis K, et al. mas. PLoS ONE 2010;5:e14249. Somatic mutations affect key pathways in lung adenocarcinoma. Nature 13. Campbell JM, Lockwood WW, Buys TP, Chari R, Coe BP, Lam S, et al. 2008;455:1069–75. Integrative genomic and gene expression analysis of chromosome 7 6. Sos ML, Michel K, Zander T, Weiss J, Frommolt P, Peifer M, et al. Predicting identified novel oncogene loci in non–small cell lung cancer. Genome drug susceptibility of non–small cell lung cancers based on genetic lesions. 2008;51:1032–9. J Clin Invest 2009;119:1727–40. 14. Li L, Wei Y, To C, Zhu CQ, Tong J, Pham NA, et al. Integrated Omic analysis 7. Weir BA, Woo MS, Getz G, Perner S, Ding L, Beroukhim R, et al. Charac- of lung cancer reveals metabolism proteome signatures with prognostic terizing the cancer genome in lung adenocarcinoma. Nature 2007;450: impact. Nat Commun 2014;5:5469. 893–8. 15. Uhlen M, Oksvold P, Fagerberg L, Lundberg E, Jonasson K, Forsberg M, 8. Zhu CQ, da Cunha Santos G, Ding K, Sakurada A, Cutz JC, Liu N, et al. Role et al. Towards a knowledge-based Human Protein Atlas. Nat Biotechnol of KRAS and EGFR as biomarkers of response to erlotinib in National 2010;28:1248–50.

1128 Mol Cancer Res; 13(7) July 2015 Molecular Cancer Research

Functional Roles of CHCHD2 in NSCLC

16. An J, Shi J, He Q, Lui K, Liu Y, Huang Y, et al. CHCM1/CHCHD6, novel 38. Tong J, Taylor P, Moran MF. Proteomic analysis of the EGFR inter- mitochondrial protein linked to regulation of mitofilin and mitochondrial actome and post-translational modifications associated with receptor cristae morphology. J Biol Chem 2012;287:7411–26. endocytosis in response to EGF and stress. Mol Cell Proteomics 2014; 17. Cavallaro G. Genome-wide analysis of eukaryotic twin CX9C proteins. Mol 13:1644–58. Biosyst 2010;6:2459–70. 39. Choi H, Larsen B, Lin ZY, Breitkreutz A, Mellacheruvu D, Fermin D, et al. 18. Darshi M, Mendiola VL, Mackey MR, Murphy AN, Koller A, Perkins GA, SAINT: probabilistic scoring of affinity purification-mass spectrometry et al. ChChd3, an inner mitochondrial membrane protein, is essential for data. Nat Methods 2011;8:70–3. maintaining crista integrity and mitochondrial function. J Biol Chem 40. Pierce NW, Lee JE, Liu X, Sweredoski MJ, Graham RL, Larimore EA, et al. 2011;286:2918–32. Cand1 promotes assembly of new SCF complexes through dynamic 19. Yang J, Staples O, Thomas LW, Briston T, Robson M, Poon E, et al. Human exchange of F box proteins. Cell 2013;153:206–15. CHCHD4 mitochondrial proteins regulate cellular oxygen consumption 41. Havugimana PC, Hart GT, Nepusz T, Yang H, Turinsky AL, Li Z, et al. rate and metabolism and provide a critical role in hypoxia signaling and A census of human soluble protein complexes. Cell 2012;150: tumor progression. J Clin Invest 2012;122:600–11. 1068–81. 20. Baughman JM, Nilsson R, Gohil VM, Arlow DH, Gauhar Z, Mootha VK. A 42. Tsofack SP, Garand C, Sereduk C, Chow D, Aziz M, Guay D, et al. NONO computational screen for regulators of oxidative phosphorylation implicates and RALY proteins are required for YB-1 oxaliplatin induced resistance in SLIRP in mitochondrial RNA homeostasis. PLoS Genet 2009;5:e1000590. colon adenocarcinoma cell lines. Mol Cancer 2011;10:145. 21. Aras S, Pak O, Sommer N, Finley R Jr, Huttemann M, Weissmann N, et al. 43. Zhang X, Zhang F, Guo L, Wang Y, Zhang P, Wang R, et al. Interactome Oxygen-dependent expression of cytochrome c oxidase subunit 4-2 gene analysis reveals that C1QBP (complement component 1, q subcomponent expression is mediated by transcription factors RBPJ, CXXC5 and binding protein) is associated with cancer cell chemotaxis and metastasis. CHCHD2. Nucleic Acids Res 2013;41:2255–66. Mol Cell Proteomics 2013;12:3199–209. 22. Seo M, Lee WH, Suk K. Identification of novel cell migration-promoting 44. Han T, Kang D, Ji D, Wang X, Zhan W, Fu M, et al. How does cancer cell genes by a functional genetic screen. FASEB J 2010;24:464–78. metabolism affect tumor migration and invasion? Cell Adh Migr 2013;7: 23. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 395–403. 2011;144:646–74. 45. Diers AR, Higdon AN, Ricart KC, Johnson MS, Agarwal A, Kalyanaraman B, 24. Tong J, Taylor P, Peterman S, Prakash A, Moran M. EGF receptor phos- et al. Mitochondrial targeting of the electrophilic lipid 15-deoxy-Del- phorylation sites S991 and Y998 are implicated in the regulation of ta12,14-prostaglandin J2 increases apoptotic efficacy via redox cell signal- receptor endocytosis and p38 MAP kinase-dependent phosphorylations ling mechanisms. Biochem J 2010;426:31–41. at S1039 and T1041. Mol Cell Proteomics 2009;8:2131–44. 46. Diers AR, Broniowska KA, Chang CF, Hogg N. Pyruvate fuels mitochondrial 25. Fleming Y, Armstrong CG, Morrice N, Paterson A, Goedert M, Cohen P. respiration and proliferation of breast cancer cells: effect of monocarbox- Synergistic activation of stress-activated protein kinase 1/c-Jun N-terminal ylate transporter inhibition. Biochem J 2012;444:561–71. kinase (SAPK1/JNK) isoforms by mitogen-activated protein kinase kinase 4 47. Lin L, McCroskery S, Ross JM, Chak Y, Neuhuber B, Daniels MP. Induction (MKK4) and MKK7. Biochem J 2000;352(Pt 1):145–54. of filopodia-like protrusions by transmembrane agrin: role of agrin gly- 26. Tong J, Sydorskyy Y, St-Germain JR, Taylor P, Tsao MS, Moran MF. Odin cosaminoglycan chains and Rho-family GTPases. Exp Cell Res 2010; (ANKS1A) modulates EGF receptor recycling and stability. PLoS ONE 316:2260–77. 2013;8:e64817. 48. Parker JJ, Dionne KR, Massarwa R, Klaassen M, Foreman NK, Niswander L, 27. Wei Y, Tong J, Taylor P, Strumpf D, Ignatchenko V, Pham NA, et al. Primary et al. Gefitinib selectively inhibits tumor cell migration in EGFR-amplified tumor xenografts of human lung adeno and squamous cell carcinoma human glioblastoma. Neuro Oncol 2013;15:1048–57. express distinct proteomic signatures. J Proteome Res 2011;10:161–74. 49. Soda M, Choi YL, Enomoto M, Takada S, Yamashita Y, Ishikawa S, et al. 28. Cox J, Mann M. MaxQuant enables high peptide identification rates, Identification of the transforming EML4-ALK fusion gene in non–small- individualized p.p.b.-range mass accuracies and proteome-wide protein cell lung cancer. Nature 2007;448:561–6. quantification. Nat Biotechnol 2008;26:1367–72. 50. Baldi L, Mengoli MC, Bisagni A, Banzi MC, Boni C, Rossi G. Concomitant 29. Cox J, Neuhauser N, Michalski A, Scheltema RA, Olsen JV, Mann M. EGFR mutation and ALK rearrangement in lung adenocarcinoma is more Andromeda: a peptide search engine integrated into the MaxQuant envi- frequent than expected: report of a case and review of the literature with ronment. J Proteome Res 2011;10:1794–805. demonstration of genes alteration into the same tumor cells. Lung Cancer 30. Roux KJ, Kim DI, Raida M, Burke B. A promiscuous biotin ligase fusion 2014;86:291–5. protein identifies proximal and interacting proteins in mammalian cells. 51. Kee Y, Yang K, Cohn MA, Haas W, Gygi SP, D'Andrea AD. WDR20 regulates J Cell Biol 2012;196:801–10. activity of the USP12 x UAF1 deubiquitinating enzyme complex. J Biol 31. Comartin D, Gupta GD, Fussner E, Coyaud E, Hasegan M, Archinti M, et al. Chem 2010;285:11252–7. CEP120 and SPICE1 cooperate with CPAP in centriole elongation. Curr 52. Moretti J, Chastagner P, Liang CC, Cohn MA, Israel A, Brou C. The Biol 2013;23:1360–6. ubiquitin-specific protease 12 (USP12) is a negative regulator of notch 32. MacLean B, Tomazela DM, Shulman N, Chambers M, Finney GL, Frewen B, signaling acting on notch receptor trafficking toward degradation. J Biol et al. Skyline: an open source document editor for creating and analyzing Chem 2012;287:29429–41. targeted proteomics experiments. Bioinformatics 2010;26:966–8. 53. Li X, Stevens PD, Yang H, Gulhati P, Wang W, Evers BM, et al. The 33. Huang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis deubiquitination enzyme USP46 functions as a tumor suppressor by of large gene lists using DAVID bioinformatics resources. Nat Protoc 2009; controlling PHLPP-dependent attenuation of Akt signaling in colon cancer. 4:44–57. Oncogene 2013;32:471–8. 34. Stark C, Breitkreutz BJ, Reguly T, Boucher L, Breitkreutz A, Tyers M. 54. McGee AM, Douglas DL, Liang Y, Hyder SM, Baines CP. The mitochondrial BioGRID: a general repository for interaction datasets. Nucleic Acids Res protein C1qbp promotes cell proliferation, migration and resistance to cell 2006;34:D535–9. death. Cell Cycle 2011;10:4119–27. 35. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. 55. Lasham A, Print CG, Woolley AG, Dunn SE, Braithwaite AW. YB-1: Cytoscape: a software environment for integrated models of biomolecular oncoprotein, prognostic marker and therapeutic target? Biochem J 2013; interaction networks. Genome Res 2003;13:2498–504. 449:11–23. 36. Hirsch FR, Varella-Garcia M, Bunn PA Jr, Di Maria MV, Veve R, Bremmes 56. Hyogotani A, Ito K, Yoshida K, Izumi H, Kohno K, Amano J. Association of RM, et al. Epidermal growth factor receptor in non–small-cell lung carci- nuclear YB-1 localization with lung resistance-related protein and epider- nomas: correlation between gene copy number and protein expression and mal growth factor receptor expression in lung cancer. Clin Lung Cancer impact on prognosis. J Clin Oncol 2003;21:3798–807. 2012;13:375–84. 37. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio 57. Panupinthu N, Yu S, Zhang D, Zhang F, Gagea M, Lu Y, et al. Self- cancer genomics portal: an open platform for exploring multidimensional reinforcing loop of amphiregulin and Y-box binding protein-1 contributes cancer genomics data. Cancer Discov 2012;2:401–4. to poor outcomes in ovarian cancer. Oncogene 2014;33:2846–56.

www.aacrjournals.org Mol Cancer Res; 13(7) July 2015 1129

CHCHD2 Is Coamplified with EGFR in NSCLC and Regulates Mitochondrial Function and Cell Migration

Yuhong Wei, Ravi N. Vellanki, Étienne Coyaud, et al.

Mol Cancer Res 2015;13:1119-1129. Published OnlineFirst March 17, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-14-0165-T

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2015/03/18/1541-7786.MCR-14-0165-T.DC1

Cited articles This article cites 57 articles, 18 of which you can access for free at: http://mcr.aacrjournals.org/content/13/7/1119.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/13/7/1119.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/13/7/1119. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.