UBC Faculty Research and Publications

Genes related to emphysema are enriched for ubiquitination pathways Stepaniants, Sergey; Wang, I-Ming; Boie, Yves; Mortimer, James; Kennedy, Brian; Elliott, Mark; Hayashi, Shizu; Luo, Honglin; Wong, Jerry; Loy, Leanna; Coulter, Silvija; Roberts, Christopher J; Hogg, James C; Sin, Don D; O’Neill, Gary; Crackower, Michael; Morris, Melody; Paré, Peter D; Obeidat, Ma’en Nov 29, 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12890_2014_Article_636.pdf [ 1.48MB ]
JSON: 52383-1.0221552.json
JSON-LD: 52383-1.0221552-ld.json
RDF/XML (Pretty): 52383-1.0221552-rdf.xml
RDF/JSON: 52383-1.0221552-rdf.json
Turtle: 52383-1.0221552-turtle.txt
N-Triples: 52383-1.0221552-rdf-ntriples.txt
Original Record: 52383-1.0221552-source.json
Full Text

Full Text

RESEARCH ARTICLEGenes related to emphyseoChDastultimately obliteration of small airways, and emphysema tribute to airway remodeling and to emphysema may beStepaniants et al. BMC Pulmonary Medicine 2014, 14:187http://www.biomedcentral.com/1471-2466/14/187tions we examined the relationship between expressionPaul’s Hospital, 1081 Burrard St, Vancouver V6Z 1Y6, BC, CanadaFull list of author information is available at the end of the articlewith enlargement of airspaces and destruction of lungparenchyma which causes loss of lung elasticity [1]. Mostpatients with COPD have both lesions although onemay predominate in individual patients. Although theprocesses differ substantially they both lead to a com-mon phenotype, reduced maximal expiratory flow.different, correlation of gene expression with the individ-ual lesions, rather than with lung function, may be morerevealing. The degree of emphysema can be measuredmicroscopically as the lung surface area to volume ratio(SA/V) [2]. In the current study we identified 181 geneswhose expression pattern correlated with the severity ofemphysema using two samples of lung tissue with differ-ent SA/V from 21 smokers with variable degrees of air-flow obstruction whose lungs were resected for smallperipheral lung lesions (P < 0.05). To validate associa-* Correspondence: Peter.pare@hli.ubc.ca†Equal contributors3University of British Columbia Centre for Heart and Lung Innovation, Stbronchiolitis with fibrosis which causes narrowing andthe small airway obstruction is due to inflammatory narrowing and obliteration. Despite their convergence inaltered physiology, different mechanisms contribute to these processes. The relationships between gene expressionand these specific phenotypes may be more revealing than comparison with lung function.Methods: We measured the ratio of alveolar surface area to lung volume (SA/V) in lung tissue from 43 smokers.Two samples from 21 subjects, in which SA/V differed by >49 cm2/mL were profiled to select genes whoseexpression correlated with SA/V. Significant genes were tested for replication in the 22 remaining subjects.Results: The level of expression of 181 transcripts was related to SA/V ( p < 0.05). When these genes were tested inthe 22 remaining subjects as a replication, thirty of the 181 genes remained significantly associated with SA/V (P < 0.05)and the direction of association was the same in 164/181. Pathway and network analysis revealed enrichment of genesinvolved in protein ubiquitination, and western blotting showed altered expression of genes involved in proteinubiquitination in obstructed individuals.Conclusion: This study implicates modified protein ubiquitination and degradation as a potentially important pathwayin the pathogenesis of emphysema.Keywords: Pulmonary emphysema, Surface area to lung volume ratio, Gene expression, Transcriptional analysis, mRNA,Cigarette smoking, Protein ubquitinationBackgroundThe pathology of chronic obstructive pulmonary disease(COPD) includes 2 main components; chronic obstructiveOne method to discover the molecular mechanismsleading to disease is to compare gene expression patternin the organs of affected and non-affected individuals.However since the gene expression patterns that con-ubiquitination pathwaysSergey Stepaniants1†, I-Ming Wang2†, Yves Boie2, James MHonglin Luo3, Jerry Wong3, Leanna Loy3, Silvija Coulter2,Gary O’Neill2, Michael Crackower2, Melody Morris2, PeterAbstractBackground: Increased small airway resistance and decrechronic obstructive pulmonary disease (COPD). The lesion© 2014 Stepaniants et al.; licensee BioMed CeCreative Commons Attribution License (http:/distribution, and reproduction in any mediumDomain Dedication waiver (http://creativecomarticle, unless otherwise stated.Open Accessma are enriched forrtimer2, Brian Kennedy2, Mark Elliott3, Shizu Hayashi3,ristopher J Roberts2, James C Hogg3, Don D Sin3,Paré3* and Ma’en Obeidat3ed lung elasticity contribute to the airflow limitation inhat corresponds to loss of lung elasticity is emphysema;ntral Ltd. This is an Open Access article distributed under the terms of the/creativecommons.org/licenses/by/4.0), which permits unrestricted use,, provided the original work is properly credited. The Creative Commons Publicmons.org/publicdomain/zero/1.0/) applies to the data made available in thisof these genes and SA/V in a separate set of 22 lungsamples. 30 of the 181 genes were significantly associ-ated with SA/V in the replication set. Gene set annota-tion and pathway analyses showed that genes involved inprotein ubiquitination and the proteasome pathway weredifferentially expressed in emphysematous lung tissue.Protein analysis confirmed the abnormal accumulationof protein-ubiquitin conjugates and the altered expres-sion of genes involved in protein ubiquitination in thelungs of obstructed smokers.MethodsSubject selection and experimental designPreoperatively, patients provided informed written con-sent that a portion of their lung tissue be used for re-search. The experimental protocols were approved by theethics review board of the University of British Columbiaand St Paul’s Hospital. Ethics approval numbers are H00-50110 and H09-00801. The strategy for selecting patientsfor gene profiling has been described in a previous publi-cation [3] and is detailed in the online supplement. Thefrozen lung tissue samples from 43 subjects with a rangeof lung function were profiled. In 21 of the subjects 2 sam-Lung function measurementDetails of lung function were described previously [3]and are contained in the online supplement. Briefly,measures of lung volumes, spirometry and the single breathdiffusing capacity were measured as previously describedand according to ATS standards [4]. Table 1 shows thenumber of patients in each GOLD category and their meansmoking history and lung function.Tissue processingDetails of tissue processing and determination of SA/Vwere described previously [3] and are contained in theonline supplement. Briefly, immediately following resec-tion the lung or lobe was frozen in liquid nitrogen fumes.Frozen cores of tissue (1.5 × 2 cm) were obtained fromslices of frozen lung using a power driven hole saw. Fro-zen sections were obtained from the surface of the ½ coreimmediately adjacent to the portion to be used for RNAextraction. The 10 micron sections were stained withhematoxylin and eosin and the severity of emphysema ineach core was measured on digital images of the sampleusing an in-house point counting program.%9098896646ryStepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 2 of 11http://www.biomedcentral.com/1471-2466/14/187ples, which showed widely different values for SA/V, wereprofiled to derive relationships between gene expressionand SA/V. For the remaining 22 subjects, one sample oflung tissue was profiled and served as a validation data setfor the relationships between gene expression and SA/V.Table 1 Subject characteristicsGOLD stage N Age (yrs) Male/Female Pack years FEV1Non-smokers 4 59 ± 8 2/2 OControl Smokers 18 67 ± 8 11/7 38 ± 191 9 61 ± 11 6/3 50 ± 142 9 63 ± 9 6/3 50 ± 213 3 67 ± 5 2/1 49 ± 12Means ± SD. An ANOVA showed no significant difference in age by GOLD categoDLCO % Predicted. (p < 0.0001 for all comparisons) Among smokers there was no sCurrent smokers, GOLD 1 - 6/9 Current smokers. GOLD 2 – 9/10 Current smokers and G6 months of surgical resection.Microarray study design and methodologyRNA from 8 GOLD 0 subjects (non-obstructed smokers)was pooled to form the reference RNA as described [3]and as detailed in the online supplement. mRNA fromthe specimens was profiled using Agilent’s Functionalpredicted FEV1/FVC % DLCO % predicted Diagnoses± 5 82 ± 3 71 ± 4 4 Carcinoid± 13 76 ± 4 80 ± 15 6 Squamous7 Adeno4 Large cell1 Small cell± 6 66 ± 4 78 ± 21 4 Squamous3 Adeno1 Large cell1 Small cell± 7 58 ± 4 68 ± 17 1 Squamous5 Adeno1 Large cell1 Poorly differentiated1 Amyloid± 1 53 ± 9 56 ± 14 1 Adeno (BAC)1 Squamous1 Carcinoidbut significant differences for pack years, FEV1 % predicted, FEV1/FVC % andignificant difference in pack years by GOLD category. Control smokers – 14/21OLD 3 – 2 Current smokers and one status unknown. Current smoker =withinStepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 3 of 11http://www.biomedcentral.com/1471-2466/14/187IDv2.0 array (Agilent, Santa Clara,CA). Microarray profil-ing was done for 23,757 probe-sets as previously described[3]. The gene expression data was deposited in GeneExpression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/projects/geo/), under the accession number GSE63073.Since a “batch effect” confounded the results, we per-formed a very conservative phase/batch adjustment stepbefore commencing with statistical analysis (see onlinesupplement).Statistical analysisIn order to identify genes correlated to SA/V, MatlabStatistical Toolbox Analysis of Covariance (ANCOVA)model ( [R14] Service Pack 1) with parallellines was used to fit each transcript to the SA/V ratio andsimultaneously account for patient effect.yij ¼ β0 þ βi þ α SA=Vð Þij þ εijWhere yij is gene expression with i =1, …21 represent-ing the patient; and j =1, 2 the two samples of lung fromeach patient respectively. β0 stands for the overall inter-cept of the model, βi is the intercept for each patient, αrepresent the common slope for the parallel lines for allpatients, and εij the model residual respectively. In thismodel the significance of the patient effect is accountedfor by the variation between individual intercepts βi andits corresponding patient p-value (Ppatient), while slope αand its corresponding p-value (Pslope) represent theregression coefficient and its significance with respect toSA/V, respectively. This model searches for the correla-tive pattern between gene expression and SA/V commonto all patients as slope α is the same for all patients. Itdoes not model individual patient SA/V association,which would be given by the corresponding interactioneffect.Genes that have significant patient effect, i.e., vary morebetween than within patients, can be selected by requiringPpatient <0.01. Conversely, those that do not vary sig-nificantly between the patients can be picked by settingPpatient >0.1, for example. Genes that significantly associatewith SA/V can be picked by establishing a p value cut offfor the slope α. It is these latter genes which are of interestsince they relate to the degree of emphysema.This model was fit individually to all 23,757 probe-sets on the microarray. Multiplicity adjustments usingBenjamini-Hochberg false discovery rate (FDR) were usedto adjust raw p-values for the slope α. Although 181 genesshowed nominal association with SA/V (P < 0.05), nonepassed FDR <0.1. However, real associations could be arti-ficially diluted by the large number of probe sets tested inthe course of the statistical analysis. One way to overcomethis problem is to perform a conservative selection ofcandidate markers and then assess their performance inan independent replication set. This latter step is alsorequired even for those markers which pass the FDRcutoff, and constitutes the ultimate check for the validityof the findings.Thus, in order to identify markers for subsequent rep-lication a combination of statistical cutoffs was used.First, more highly expressed genes that satisfied an aver-age logIntensity > −1 cutoff were selected. Secondly, thegenes were selected to be significantly associated withSA/V by requiring Pslope <0.05. Finally, the set was re-fined by filtering out those with significant patient effect,by requiring Ppatient >0.1. The reason that the final cutoffwas applied is to allow a replication of the SA/V relatedgenes in the independent set of 22 patient samples. Forthese patients only a single lung sample was availableand adjusting for the patient effect is not possible.Minimizing the patient effect allows a simple linearregression model to approximate the slope of the fitbetween each of the 181 genes and the correspondingSA/V ratios in the replication set.Networks and pathways enrichment analysesTo determine whether the 181 genes which were relatedto lung SA/V were enriched for specific biologicalprocesses or were suggestive of specific disease states weundertook pathway and network analysis using Meta-Core® from GeneGo Inc. Gene IDs of the 181 transcriptswere used as input for MetaCore®.Ubuquitination pathway analysisIn this study, we found that a number of deregulatedgenes are related to the process of ubiquitination or de-ubiquitination. To test if there was concordant dysregu-lation of this pathway at the protein level, we performedWestern analyses on extracts from an additional groupof lung samples. Based on commercial availability of theantibodies (see below), we selected 9 genes that weredifferentially regulated as a function of SA/V from thelist. The genes that were included in this analysis areshown in Additional file 1: Table S6 with a brief descrip-tion of their potential role. For the protein quantificationa separate subset of 20 individual lung samples were ex-amined (See online supplement for details). In additionto the measurement of protein levels, the proteosomeactivity of the lung tissue from 5 of the cases and 5controls was measured.Details of the Western blotting and proteosome activ-ity assays are available in the online supplement.ResultsThe analysis identified 181 genes whose level of ex-pression was related to SA/V in the discovery data set.Additional file 1: Table S4 lists these genes and the slope,intercept and p value for their relationship to SA/V. Avisual appreciation of the expression of the 181 genes re-lated to SA/V in the discovery set is shown in Figure 1.The data are arranged by SA/V. Although the data shownare the original data without model fitting, a SA/V-correlatedtrend is observed when all the samples are combined andsorted by increasing SA/V values. The level of expressionof these transcripts was then related to the SA/V in the 22samples which constitute the replication data set. Figure 2shows that when the genes from the discovery set arearranged in the same order, the SA/V-correlated trendis preserved in the replication set.The slopes relating gene expression to SA/V from thediscovery and the replication sets are compared in Figure 3which shows that the slopes and direction of effectobtained from both discovery and replication sets areconsistent, with an overall correlation of 0.6. If we use aslope p-value cutoff <0.05 in the replication set as a cri-teria to indicate replication of the initial relationship, 30 ofthe 181 genes achieve this level of significance (Table 2).Figure 4 is identical to Figure 3 except that it only showsthe 30 genes that replicated based on the p value cut off inthe replication sample set. These genes represent potentialmarkers and candidates for subsequent follow up.The results of the network and pathways analyses areshown in Table 3. Using MetaCore, the top significantlyenriched pathways (Minimum P = 2.786E-05) were cellcycle regulation of G1/S transition, followed by the Immuneresponse HMGB1/TLR signaling pathway (P = 6.224E-04).In terms of disease processes, “Starvation” ranked first(P = 1.042E-04). Gene Ontology (GO) processes’ enrich-ment presented in Additional file 1: Table S5 identifiedprocesses related to cellular macromolecule metabolism(P = 1.417E-10), and ubiquitination (P = 2.103E-08) path-ways as being significantly enriched. Many of thesegenes are related to the process of ubiquitination ordeubiquitination.Interestingly, analysis of the 30 replicated genes showedsignificant enrichment in disease biomarkers for muscularatrophy (p = 4.239E-04) and proteostasis (p = 4.829E-05).Protein replicationThe protein level of five of the 9 ubiquitination-associatedgenes (FBXL3, FBXO30, USP38, UBB, and RNF6) was sig-nificantly different between control and COPD lung tissueby western analysis (Figure 5). FBXL3, FBXO30, and USP38were upregulated at the protein level (Figure 5A), which isopposite in direction to the changes in mRNA. This couldbe a reflection of increased protein stability. HoweverUBB and RNF6 protein were downregulated consistentwith the gene expression profile (Figure 5B). The proteinxpStepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 4 of 11http://www.biomedcentral.com/1471-2466/14/187Figure 1 A heat map shows the relationship between the level of ethe samples in the discovery set. The samples are shown in rows ordereto SA/V are shown by the yellow bar while those whose expression was poression of the 181 SA/V-correlated genes (columns) for all 42 ofd by increasing SA/V. Genes whose expression was negatively relatedsitively related to SA/V are indicated with the blue bar.Figure 2 A heat map shows the relationship between the level of expresin the replication set. The samples are shown in rows ordered by increasing Sthe yellow bar while those whose expression was positively related to SA/V aFigure 3 Slopes of regression for 181 selected genes arecompared between the discovery and replication sets.The comparison shows a good concordance in the findings.Stepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 5 of 11http://www.biomedcentral.com/1471-2466/14/187abundance of the remaining four genes (UBE4A, RNF184,TBLR1, and UHRF2) was not significantly altered (datanot shown).DiscussionMost studies of gene expression in the lungs of patientswho have COPD have compared tissue from subjects withand without COPD based on lung function [3,5-7].Although COPD is defined by abnormalities of lungfunction there is convincing evidence that there are twopathological processes that contribute to the functionalimpairment [8]; loss of lung elasticity characterized patho-logically by emphysema and inflammatory narrowing andobliteration of the small conducting airways [9]. We useda morphometric measure of emphysema, derived from thesame sample of lung tissue that was used for transcrip-tomic analysis, as a continuous variable to relate geneexpression to structural changes.sion of the 181 SA/V-correlated genes (columns) for the 22 samplesA/V. Genes whose expression was negatively related to SA/V are shown byre indicated with the blue bar. The pattern is similar to the discovery set.ateereStepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 6 of 11http://www.biomedcentral.com/1471-2466/14/187Table 2 List of 30 replicated emphysema genesmRNA GenesymbolGene nameNM_005706 TSSC4 Tumor suppressing subtransferable candidAB067498 Esco1 Establishment of cohesion 1 homolog 1 (S. cNM_018559 kiaa1704 KIAA1704NM_004897 MINPP1 Multiple inositol polyphosphate histidinephosphatase, 1The design of the discovery study using a pair of lungsamples from each individual, one with the highest andone with the lowest SA/V ratios, allowed us to accountfor patient-to-patient variability. An advantage of thisdesign is that the demographic distribution of patients isof much lesser concern than for between- subject com-parisons, where serious attention must be given tostratification and balancing of patients with respect topotentially confounding demographic attributes. Tran-scripts that significantly correlated with SA/V ratio afterContig55580_RCNM_012180 FBXO8 F-box protein 8NM_153044 MORC2-AS1MORC2 antisense RNA 1NM_022876 SMN2 Survival of motor neuron 2, centromericNM_003838 FPGT Fucose-1-phosphate guanylyltransferaseAF055030 PHF10 PHD finger protein 10NM_007342 NUPL2 Nucleoporin like 2Contig45624_RC CBLL1 Cbl proto-oncogene-like 1, E3 ubiquitin proteinNM_002103 GYS1 Glycogen synthase 1 (muscle)AL359938 MEIS3 Meis homeobox 3; Meis homeobox 3 pseudogNM_014771 RNF40 Ring finger protein 40NM_022877 SMN2 Survival of motor neuron 2, centromericContig41498_RC PTPN4 Protein tyrosine phosphatase, non-receptor tNM_003084 SNAPC3 Small nuclear RNA activating complex, polypepNM_013234 EIF3K Eukaryotic translation initiation factor 3, subuNM_006852 TLK2 Tousled-like kinase 2X68560 SP3 Sp3 transcription factorAY007149 CEP350 Centrosomal protein 350 kDaContig51940_RC GABPA GA binding protein transcription factor, alphasubunit 60 kDaNM_032557 USP38 Ubiquitin specific peptidase 38NM_015153 PHF3 PHD finger protein 3NM_004162 RAB5A RAB5A, member RAS oncogene familyNM_022875 SMN2 Survival of motor neuron 2, centromericNM_005316 GTF2H1 General transcription factor IIH, polypeptide 162 kDaContig53191_RC GPD2 Glycerol-3-phosphate dehydrogenase 2(mitochondrial)NM_017411 SMN2 Survival of motor neuron 2, centromeric*All FDR values were >0.1 in the discovery samples.DiscoveryslopeDiscoveryslope P value*ReplicationslopeReplicationslope P value4 −0.00037 0.039 −0.00096 0.003visiae) 0.00038 0.018 0.00045 0.0050.00040 0.023 0.00065 0.0060.00035 0.033 0.00070 0.006adjusting for patient gene expression variation constituterelevant biological candidates.Although most studies of transcriptional profiles ofCOPD patients have focused on expression changes as-sociated with disease status or level of lung function,Campbell et al. recently reported the results of a studyexamining gene expression as related to alterations inlocal lung architecture [10]. This study examined thecorrelation of gene expression with Lm (Lm =mean linearintercept, a microscopic measure of emphysema directly0.00049 0.029 0.00139 0.0060.00040 0.010 0.00067 0.014−0.00040 0.022 −0.00054 0.0150.00046 0.019 0.00083 0.0170.00035 0.031 0.00054 0.0180.00036 0.011 0.00067 0.0190.00037 0.047 0.00081 0.019ligase 0.00052 0.018 0.00091 0.020−0.00033 0.017 −0.00052 0.020ene 2 −0.00029 0.040 −0.00054 0.022−0.00034 0.027 −0.00067 0.0240.00046 0.018 0.00077 0.024ype 4 0.00035 0.027 0.00049 0.024tide 3 0.00044 0.015 0.00075 0.025nit K −0.00025 0.044 −0.00045 0.0270.00021 0.037 0.00032 0.0280.00065 0.012 0.00113 0.0300.00034 0.029 0.00058 0.0300.00046 0.035 0.00093 0.0330.00048 0.017 0.00073 0.0340.00060 0.034 0.00057 0.0360.00030 0.045 0.00048 0.0370.0004843 0.015 0.00071 0.039, 0.00042 0.010 0.00069 0.0390.00043 0.043 0.00069 0.0410.00048 0.029 0.00071 0.0444 Neoplasms, Cystic, Mucinous, and Serous 1.401E-035 Neoplasms, Complex and Mixed 1.497E-036 Motor Neuron Disease 3.125E-03Process Networks# Networks p-value1 Proteolysis_Proteolysis in cell cycle and apoptosis* 7.634E-042 Cell cycle_Mitosis* 9.314E-043 Cell cycle_S phase 1.896E-034 Transcription_mRNA processing 2.716E-035 Transcription_Chromatin modification 5.222E-036 Signal Transduction_BMP and GDF signaling 8.269E-037 Cell cycle_G1-S 1.407E-02Stepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 7 of 11http://www.biomedcentral.com/1471-2466/14/187related to lung SA/V). Despite the apparent similarity inthe design of Campbell et al. and our study, we found nogenes in common. While this is undoubtedly due, at leastin part, to differences in patient populations, the specificsof sample preparation and/or the expression platform(Agilent versus Affymetrix), another factor may have beenexperimental design. Two of the eight lungs in the studyof Campbell et al. were from normal individuals, and theseFigure 4 Slopes of regression for the 30 genes that replicatedat p < 0.05 in the replication data set. The slopes for the replicationdata are on the x axis and those from the discovery set are onthe y axis.had substantially lower Lm than in COPD. This creates apotential bias such that many genes reported as correlat-ing with Lm were also dependent on COPD status. Whilethis covariate was technically accounted for with the in-clusion of a random patient effect in the fitted mixedmodels, the authors did not remove genes that had asignificant patient effect as we did. Indeed, 78 of 126 genesreported by Campbell et al. were differentially expressed innormal donors compared to COPD patients (p-value <0.001by student’s T-Test) and 65 of these genes also had a signifi-cant patient effect in our study (p-value of alpha <0.05).Thus, differences in results may indicate that the resultsof Campbell et al. reflect a mix of gene expression changesdue to emphysema and disease status.The most significantly enriched pathways we identifiedrelate to the regulation of G1/S transition, and to highmobility group box protein 1/toll like receptor (HMGB1/TLR) signaling. Further discussion of the potential role ofthese genes is included in the online supplement.Many of the genes identified as being related to SA/Vare involved in the handling of proteins. Cellular proteinhomeostasis, also known as proteostasis, involves thecontrol of the conformation, concentration, binding in-teractions, and localization of individual proteins withinTable 3 MetaCore pathway and network analysis resultsfor 181 genes related to emphysemaPathway analysis# Maps p-value1 Cell cycle_Regulation of G1/S transition* 2.786E-052 Cell cycle_Role of SCF complex in cell cycle regulation* 3.259E-043 Cell cycle_ESR1 regulation of G1/S transition* 4.803E-044 Immune response_ HMGB1/TLR signaling pathway* 6.224E-045 Immune response_HSP60 and HSP70/TLR signaling pathway 2.039E-03Diseases by Biomarker# Diseases p-value1 Starvation* 1.042E-042 Cystadenocarcinoma, Serous 1.066E-033 Cystadenocarcinoma 1.328E-03and outside the cell. The proteostasis network refers tothe >2000 genes in mammals encoding proteins thatwork together as a system to control protein concentra-tion and conformation through interactions of the prote-ome with chaperone systems and folding enzymes as wellas via protein degradation mediated by the ubiquitinproteasome system [11].There is evidence that an imbalance in proteostasiscontributes to certain diseases [12]. Our data implicatinggenes involved in the handling of protein suggests thatan imbalance in the synthesis and degradation of proteinmay be involved in the genesis of emphysema. This ideais not new; Min et al. used immunohistchemistry toexamine the expression of proteins involved in proteinprocessing and apoptosis in the lungs of individuals withvarying severity of COPD. They show accumulation ofpoly-ubiquitinated proteins in insoluble aggregates inthe lungs of subjects with emphysema [13]. Cantin and8 Cytoskeleton_Spindle microtubules 1.531E-029 Signal transduction_NOTCH signaling 1.720E-0210 Cell cycle_G2-M 3.453E-02*Donates that the enrichment for pathways, diseases and networks issignificant at FDR <0.1.Figure 5 (See legend on next page.)Stepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 8 of 11http://www.biomedcentral.com/1471-2466/14/187Richter also suggested that proteostasis imbalance is im-portant in obstructive lung diseases [14].In particular, genes involved in the ubiquitin-proteosomepathway were significantly over-represented in the presentstudy. The ubiquitin-proteasome system (UPS) serves acrucial function in protein quality control to maintain cel-lular proteostasis through the degradation of misfolded/damaged proteins and the turnover of normal short-livedregulatory proteins. Recently, impairment of the UPS hasbeen reported in several lung diseases, including COPD/emphysema and pulmonary fibrosis [15,16]. It was shownthat lung tissue from COPD patients with severe emphy-sema has aberrant accumulation of ubiquitinated proteinstissues as compared to control lung (Additional file 2:Figure S3), although the latter two changes did not reachstatistical significance probably due to small sample size.In addition to disposal of misfolded/damaged proteins,the UPS also plays a key role in the regulation of manyfundamental cellular functions, including apoptosis, cellcycle regulation, antigen processing, and transcriptionalregulation via controlling the degradation of normal regu-latory proteins [21-23]. Both FBXL3 and FBXO30 areF-box proteins, critical components of the SCF (SKP1-cullin-F-box) E3 ligases which are involved in the degradationof signal and cell cycle proteins [24]. Upregulation of theseproteins suggest a role for apoptosis and cell cycle arrest(See figure on previous page.)Figure 5 Protein expression of ubiquitination genes in lung tissues from control and COPD patients. Lung homogenates were preparedand Western blotting was performed to examine protein levels with the antibodies specified. β-actin was probed as a protein loading control.Protein levels were quantified by densitometric analysis wit the NIH ImageJ program and normalized to β-actin expression. A: results for FBXL3,FBXO30, and USP38 showing upregulation in COPD. B: results for UBB and RNF6 showing downregulation in COPD. Data are means ± SE, andsignificance was determined by Students’ t-tests.Stepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 9 of 11http://www.biomedcentral.com/1471-2466/14/187[13], a phenomenon commonly observed in proteinconformational diseases, including neurodegenerativeand heart diseases [17,18].Increased abundance of ubiquitinated conjugates couldbe the result of decreased proteasome activity and/or in-creased protein ubiquitination. Upon exposure to cigarettesmoke, the proteolytic activities of the proteasome in hu-man alveolar epithelial cells and mouse lung were mark-edly decreased [19,20]. We also demonstrated a consistentreduction of all three proteasome activities (chymotryp-sin-, trypsin-, and caspase-like) in human COPD lungFigure 6 Proposed model of dysregulation of the ubiquitin-proteasemphysema/COPD. Oxidative stress and inflammation induced by smoubiquitination-related genes and impairment of the proteasome function. Acproduction and decreased degradation causes further damage of the proteasoof the UPS results in apoptosis, inflammation, and matrix remodeling, pathogecan also cause compensatory upregulation of genes associated with the UPS.in the pathogenesis of emphysema. USP38 is a deubiquiti-nating enzyme and was recently identified as a susceptibil-ity gene for asthma [25]. Although the exact targets ofUSP38 remain unclear, it is speculated that upregulationof USP may reflect a cellular response to inflammationand injury. The gene UBB encodes polyubiquitin precur-sor protein which is processed by deubiquitinating en-zymes to produce a single ubiquitin moiety and ribosomalproteins [26]. Down-regulation of UBB will lead to de-creased availability of the free ubiquitin pool, therebyresulting in reduced protein ubiquitination and subsequentome system leading to the pathogenesis and progression ofke, genetic or environmental insults result in dysregulation ofcumulation of abnormal proteins in the lung as a result of increasedme function and dysregulation of UPS-related genes. Aberrant regulationnic characteristics of emphysema/COPD. Damaged proteasome functionFigure 6 summarizes the proposed mechanisms byStepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 10 of 11http://www.biomedcentral.com/1471-2466/14/187which smoke and ROS-mediated modulation of the UPScould contribute to the pathogenesis and progression ofCOPD by participating in the regulation of apoptosis, theinflammatory response, and interstitial fibrosis. Futurestudies are required to address the function and regulationof this system in COPD/emphysema and to furtherexplore the specific mechanisms involved.There are several limitations to this study. The samplesize is relatively small and thus the possibility of falsenegative or positive associations is real. We addressed this,in part, by using one portion of the samples as a derivationset and one portion as a replication set. Figure 3 showsthat the level of expression of 164 of the 181 genes wereassociated with SA/V in the same direction in the replica-tion set as in the derivation set and 30 of the 181 genesshowed significant (p < 0.05) association with SA/V inthe replication set (with the same direction of effect –see Figure 4). We used all 181 genes, not only the 30that replicated, to examine for enriched networks usingMetaCore®. We feel we are justified in this approachgiven that the vast majority of expression changes werein the same direction. Another limitation is the fact thatsamples were done in two batches which had a substan-tial systematic effect on gene expression. Although weadjusted for this bias statistically, it is possible that sometrue positive associations could have been missed due tothis adjustment.Finally, changes in gene expression do not confirmthat these genes are causal to emphysema. Some of thechanges could be a consequence of the molecular pro-cesses that result from the development of emphysema.ConclusionsThe analysis approach adopted in this study allowedexamination of the relationship between gene expres-sion and SA/V, a measure of tissue destruction by em-physema. Gene set annotation and pathway analysesimplicated genes involved in protein ubiquitinationand the proteasome pathways and protein analysis con-firmed the abnormal accumulation of protein-ubiquitinconjugates and the altered expression of genes involveddegradation. RNF6 is a RING (Really Interesting NewGene) finger domain E3 ligase and LIM kinase 1 (LIMK1)has been identified as a substrate of RNF6 [27]. LIMK1, aserine/threonine kinase involved in the regulation of actinpolymerization and microtubule disassembly, was shownto promote the disruption of endothelial barrier andinflammatory infiltration in mouse lung [28]. It is there-fore plausible to assume that downregulation of RNF6plays a role in the chronic inflammation of COPD.in protein ubiquitination in the lungs of obstructedsmokers.Availability of supporting dataAdditional data is available in a supplementary materialdocument. The microarray data was deposited on GeneExpression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/projects/geo/). The GEO accession number is GSE63073.Additional filesAdditional file 1: Online supplementary contacting detailedmethods, additional results and discussions.Additional file 2: Figure S3. Proteasome activities in lung tissues fromcontrol and COPD patients. Lung homogenates were prepared andchymotrypsin-, trypsin-, and caspase-like proteasome activities weremeasured as described in the Materials and Methods. Results are expressedas means ± SE. Significance was determined by Student's t-test.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsSS and I-MW conceived and designed the study, preformed the bulk of thedata analysis and drafted the manuscript. PDP conceived and designed thestudy, collected and integrated the phenotypic data, coordinated elementsof the study and drafted the manuscript. MO coordinated elements of thestudy, made a significicant intellectual contribution and drafted the manuscript.YB, JM, BK SC, CJR, GO, DDS and MC made significant intellectual contributionto the experimental design. SH provided critical review of the manuscript. ME,LL and JW made key technical and methodological contributions. JCH designedand implemented the biobank on which the study is based and contributed toexperimental design. HL supervised and conducted a key element of the study,the protein analysis. MM and MO performed specific specialized analyses.All co-authors contributed to the writing of the manuscript and signed off onthe final version.AcknowledgementsFunding for the conduct of this study came from Merck Frosst Canada andthe Canadian Institutes of Health Research. Merck Frosst Canada providedthe funds to do the gene expression and reviewed the manuscript andapproved it for publication. Scientists working with Merck ResearchLaboratory contributed to the conception and design of the study and asscientists were involved in the analysis and interpretation of the data. Noone within the company influenced the study in any other way. Specificallythe funders within the company did not influence the design of the study,the collection of the data, the analysis or interpretation of data; nor werethey involved in the writing of the manuscript.Author details1Covance Genomics Laboratory, Indianapolis, USA. 2Merck ResearchLaboratory, Rahway, USA. 3University of British Columbia Centre for Heartand Lung Innovation, St Paul’s Hospital, 1081 Burrard St, Vancouver V6Z 1Y6,BC, Canada.Received: 5 April 2014 Accepted: 19 November 2014Published: 29 November 2014References1. Barnes PJ: New concepts in chronic obstructive pulmonary disease. AnnuRev Med 2003, 54(1):113–129.2. Coxson HO, Rogers RM, Whittall KP, D'Yachkova Y, Paré PD, Sciurba FC,Hogg JC: A quantification of the lung surface area in emphysema usingcomputed tomography. Am J Respir Crit Care Med 1999, 159(3):851–856.3. Wang IM, Stepaniants S, Boie Y, Mortimer JR, Kennedy B, Elliott M, Hayashi S,Loy L, Coulter S, Cervino S, Harris J, Thornton M, Raubertas R, Roberts C,Hogg JC, Crackower M, O'Neill G, Paré PD: Gene expression profiling inpatients with chronic obstructive pulmonary disease and lung cancer.Am J Respir Crit Care Med 2008, 177(4):402–411.4. Retamales I, Elliott WM, Meshi B, Coxson HO, Paré PD, Sciurba FC, RogersRM, Hayashi S, Hogg JC: Amplification of inflammation in emphysemaStepaniants et al. BMC Pulmonary Medicine 2014, 14:187 Page 11 of 11http://www.biomedcentral.com/1471-2466/14/187and its association with latent adenoviral infection. Am J Respir Crit CareMed 2001, 164(3):469–473.5. Bhattacharya S, Srisuma S, Demeo DL, Shapiro SD, Bueno R, Silverman EK,Reilly JJ, Mariani TJ: Molecular biomarkers for quantitative and discreteCOPD phenotypes. Am J Respir Cell Mol Biol 2009, 40(3):359–367.6. Ning W, Li C-J, Kaminski N, Feghali-Bostwick CA, Alber SM, Di YP, Otterbein SL,Song R, Hayashi S, Zhou Z, Pinsky DJ, Watkins SC, Pilewski JM, Sciurba FC,Peters DG, Hogg JC, Choi AMK: Comprehensive gene expression profilesreveal pathways related to the pathogenesis of chronic obstructivepulmonary disease. Proc Natl Acad Sci U S A 2004, 101(41):14895–14900.7. Spira A, Beane J, Pinto-Plata V, Kadar A, Liu G, Shah V, Celli B, Brody JS:Gene expression profiling of human lung tissue from smokers withsevere emphysema. Am J Respir Cell Mol Biol 2004, 31(6):601–610.8. Patel BD, Coxson HO, Pillai SG, Agustí AGN, Calverley PMA, Donner CF, Make BJ,Müller NL, Rennard SI, Vestbo J, Wouters EFM, Hiorns MP, Nakano Y, Camp PG,Nasute Fauerbach PV, Screaton NJ, Campbell EJ, Anderson WH, Paré PD,Levy RD, Lake SL, Silverman EK, Lomas DA: Airway wall thickening andemphysema show independent familial aggregation in chronic obstructivepulmonary disease. Am J Respir Crit Care Med 2008, 178(5):500–505.9. McDonough JE, Yuan R, Suzuki M, Seyednejad N, Elliott WM, Sanchez PG,Wright AC, Gefter WB, Litzky L, Coxson HO, Paré PD, Sin DD, Pierce RA,Woods JC, McWilliams AM, Mayo JR, Lam SC, Cooper JD, Hogg JC: Small-airwayobstruction and emphysema in chronic obstructive pulmonary disease.N Engl J Med 2011, 365(17):1567–1575.10. Campbell J, McDonough J, Zeskind J, Hackett T, Pechkovsky D, BrandsmaC-A, Suzuki M, Gosselink J, Liu G, Alekseyev Y, Xiao J, Zhang X, Hayashi S,Cooper J, Timens W, Postma D, Knight D, Marc L, James H, Avrum S:A gene expression signature of emphysema-related lung destructionand its reversal by the tripeptide GHK. Genome Med 2012, 4(8):67.11. Hartl FU, Bracher A, Hayer-Hartl M: Molecular chaperones in protein foldingand proteostasis. Nature 2011, 475(7356):324–332.12. Mu T-W, Ong DST, Wang Y-J, Balch WE, Yates JR, Segatori L, Kelly JW:Chemical and biological approaches synergize to ameliorateprotein-folding diseases. Cell 2008, 134(5):769–781.13. Min T, Bodas M, Mazur S, Vij N: Critical role of proteostasis-imbalance inpathogenesis of COPD and severe emphysema. J Mol Med 2011,89(6):577–593.14. Cantin AM, Richter MV: Cigarette smoke-induced proteostasis imbalancein obstructive lung diseases. Curr Mol Med 2012, 12(7):836–849.15. Bodas M, Tran I, Vij N: Therapeutic strategies to correct proteostasis-imbalancein chronic obstructive lung diseases. Curr Mol Med 2012, 12(7):807–814.16. Meiners S, Eickelberg O: What shall we do with the damaged proteins inlung disease? ask the proteasome! Eur Respir J 2012, 40(5):1260–1268.17. Hegde AN, Upadhya SC: Role of ubiquitin-proteasome-mediated proteolysisin nervous system disease. Biochim Biophys Acta 2011, 1809(2):128–140.18. Luo H, Wong J, Wong B: Protein degradation systems in viral myocarditisleading to dilated cardiomyopathy. Cardiovasc Res 2010, 85(2):347–356.19. Somborac-Bacura A, van der Toorn M, Franciosi L, Slebos DJ, Zanic-Grubisic T,Bischoff R, van Oosterhout AJ: Cigarette smoke induces endoplasmicreticulum stress response and proteasomal dysfunction in human alveolarepithelial cells. Exp Physiol 2013, 98(1):316–325.20. van Rijt SH, Keller IE, John G, Kohse K, Yildirim AO, Eickelberg O, Meiners S:Acute cigarette smoke exposure impairs proteasome function in thelung. Am J Physiol Lung Cell Mol Physiol 2012, 303(9):L814–823.21. Ciechanover A: The ubiquitin-proteasome pathway: on protein death andcell life. Embo J 1998, 17(24):7151–7160.22. Glickman MH, Ciechanover A: The ubiquitin-proteasome proteolytic pathway:destruction for the sake of construction. Physiol Rev 2002, 82(2):373–428.23. Myung J, Kim KB, Crews CM: The ubiquitin-proteasome pathway andproteasome inhibitors. Med Res Rev 2001, 21(4):245–273.24. Jin J, Cardozo T, Lovering RC, Elledge SJ, Pagano M, Harper JW:Systematic analysis and nomenclature of mammalian F-box proteins.Genes Dev 2004, 18(21):2573–2580.25. Hirota T, Takahashi A, Kubo M, Tsunoda T, Tomita K, Doi S, Fujita K, Miyatake A,Enomoto T, Miyagawa T, Adachi M, Tanaka H, Niimi A, Matsumoto H, Ito I,Masuko H, Sakamoto T, Hizawa N, Taniguchi M, Lima JJ, Irvin CG, Peters SP,Himes BE, Litonjua AA, Tantisira KG, Weiss ST, Kamatani N, Nakamura Y, Tamari M:Genome-wide association study identifies three new susceptibility loci foradult asthma in the Japanese population. Nat Genet 2011, 43(9):893–896.26. Kimura Y, Tanaka K: Regulatory mechanisms involved in the control ofubiquitin homeostasis. J Biochem 2010, 147(6):793–798.Submit your next manuscript to BioMed Centraland take full advantage of: • Convenient online submission• Thorough peer review• No space constraints or color figure charges• Immediate publication on acceptance• Inclusion in PubMed, CAS, Scopus and Google Scholar• Research which is freely available for redistribution27. Tursun B, Schluter A, Peters MA, Viehweger B, Ostendorff HP, Soosairajah J,Drung A, Bossenz M, Johnsen SA, Schweizer M, Bernard O, Bach I:The ubiquitin ligase Rnf6 regulates local LIM kinase 1 levels in axonalgrowth cones. Genes Dev 2005, 19(19):2307–2319.28. Gorovoy M, Han J, Pan H, Welch E, Neamu R, Jia Z, Predescu D, Vogel S,Minshall RD, Ye RD, Malik AB, Voyno-Yasenetskaya T: LIM kinase 1 promotesendothelial barrier disruption and neutrophil infiltration in mouse lungs.Circ Res 2009, 105(6):549–556.doi:10.1186/1471-2466-14-187Cite this article as: Stepaniants et al.: Genes related to emphysema areenriched for ubiquitination pathways. BMC Pulmonary Medicine 2014 14:187.Submit your manuscript at www.biomedcentral.com/submit


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items