Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Discordant Protein and mRNA Expression in Lung Adenocarcinomas

Discordant Protein and mRNA Expression in Lung Adenocarcinomas Research Discordant Protein and mRNA Expression in Lung Adenocarcinomas* Guoan Chen‡, Tarek G. Gharib‡, Chiang-Ching Huang§, Jeremy M. G. Taylor§, David E. Misek¶, Sharon L. R. Kardia, Thomas J. Giordano**, Mark D. Iannettoni‡, Mark B. Orringer‡, Samir M. Hanash¶, and David G. Beer‡ ‡‡ lung cancer and are now the most common histologic type. The relationship between gene expression measured at the mRNA level and the corresponding protein level is not Functional genomics, broadly defined as the comprehensive well characterized in human cancer. In this study, we analysis of genes and their products, have become a recent compared mRNA and protein expression for a cohort of focus of the life sciences (1). Application of these approaches to genes in the same lung adenocarcinomas. The abun- lung adenocarcinomas has the potential to aid in the identifica- dance of 165 protein spots representing 98 individual tion of high risk patients with resectable early stage lung cancer genes was analyzed in 76 lung adenocarcinomas and nine that may benefit from adjuvant therapy, as well as to identify non-neoplastic lung tissues using two-dimensional poly- new therapeutic targets. In human lung cancer, however, little is acrylamide gel electrophoresis. Specific polypeptides currently understood regarding the relationship between gene were identified using matrix-assisted laser desorption/ expression as determined by measuring mRNA levels and the ionization mass spectrometry. For the same 85 samples, corresponding abundance of the protein products. mRNA levels were determined using oligonucleotide mi- A number of powerful techniques for analysis of gene ex- croarrays, allowing a comparative analysis of mRNA and protein expression among the 165 protein spots. Twenty- pression have been used including differential display (2), eight of the 165 protein spots (17%) or 21 of 98 genes serial analysis of gene expression (3), DNA microarrays (4), (21.4%) had a statistically significant correlation between and proteomics via two-dimensional polyacrylamide gel elec- protein and mRNA expression (r > 0.2445; p < 0.05); trophoresis and mass spectrometry (5). Bioinformatics tools however, among all 165 proteins the correlation coeffi- have also been developed to help determine quantitative cient values (r) ranged from 0.467 to 0.442. Correlation mRNA/protein expression profiles of all types of cells and coefficient values were not related to protein abundance. tissues (6) and now can be applied to benign and malignant Further, no significant correlation between mRNA and tumors. DNA microarrays (cDNA and oligonucleotide) permit protein expression was found (r 0.025) if the average the parallel assessment of thousands of genes and have been levels of mRNA or protein among all samples were applied utilized in gene expression monitoring (7), polymorphism anal- across the 165 protein spots (98 genes). The mRNA/ ysis (8), and DNA sequencing (9). Recent studies have fo- protein correlation coefficient also varied among pro- teins with multiple isoforms, indicating potentially sep- cused on classification or identification of subgroups of lung arate isoform-specific mechanisms for the regulation of tumors using DNA microarrays (10, 11). The use of mRNA protein abundance. Among the 21 genes with a signifi- expression patterns by themselves, however, is insufficient for cant correlation between mRNA and protein, five genes understanding the expression of protein products, as addi- differed significantly between stage I and stage III lung tional post-transcriptional mechanisms, including protein adenocarcinomas. Using a quantitative analysis of mRNA translation, post-translational modification, and degradation, and protein expression within the same lung adenocarci- may influence the level of a protein present in a given cell or nomas, we showed that only a subset of the proteins tissue. Proteomic analyses, a complementary technology to exhibited a significant correlation with mRNA abundance. DNA microarrays for monitoring gene expression, involves Molecular & Cellular Proteomics 1:304 –313, 2002. protein separation and quantitative assessment of protein spots using 2D -PAGE and protein identification using mass spectrometry. By combining proteomic and transcriptional Lung cancer is the leading cause of cancer death for both analyses of the same samples, however, it may be possible to men and women in the United States. Adenocarcinomas of understand the complex mechanisms influencing protein ex- the lung comprise 40% of all new cases of non-small cell pression in human cancer. In this study, we determined mRNA and protein levels for From the Departments of ‡Surgery, §Biostatistics, Epidemiology, 165 proteins (98 genes) in 76 lung adenocarcinomas and nine **Pathology, and ¶Pediatrics, University of Michigan, Ann Arbor, Michigan 48109 Received, January 21, 2002, and in revised form, March 4, 2002 Published, MCP Papers in Press, March 12, 2001, DOI The abbreviations used are: 2D, two-dimensional; MALDI-MS, 10.1074/mcp.M200008-MCP200 matrix-assisted laser desorption/ionization mass spectrometry. 304 Molecular & Cellular Proteomics 1.4 © 2002 by The American Society for Biochemistry and Molecular Biology, Inc. This paper is available on line at http://www.mcponline.org This is an Open Access article under the CC BY license. Protein and mRNA Correlation in Lung Adenocarcinomas TABLE I Correlation coefficients of protein and mRNA where only one spot was present on 2D gels r*, correlation coefficient value  0.2445; p  0.05. Values in boldface are significant at p  0.05. Spot Unigene Gene name r* Protein name 1104 Hs.184510 SFN 0.4337 14-3-3 0994 Hs.77840 ANXA4 0.4219 Annexin IV 1314 Hs.10958 DJ-1 0.3982 DJ-1 protein/MER5 1454 Hs.75428 SOD1 0.3863 Superoxide dismutase (Cu-Zn) 1638 Hs.227751 LGALS1 0.3318 Galectin 1 0264 Hs.129548 HNRPK 0.3034 Transformation up-regulated nuclear protein 1405 Hs.111334 FTL 0.2849 Ferritin light chain 0963 Hs.300711 ANXA5 0.2468 Annexin V 1252 Hs.4745 PSMC 0.2445 26 S proteasome p28 0906 Hs.234489 LDHB 0.4420 L-lactate dehydrogenase H chain (LDH-B) 1171 Hs.241515 COX11 0.2310 COX 11 1160 Hs.181013 PGAM1 0.2023 Phosphoglycerate mutase 0759 Hs.74635 DLD 0.1965 Dihydrolipoamide dehydrogenase precursor 1193 Hs.83383 AOE372 0.1932 Antioxidant enzyme AOE372 0172 Hs.3069 HSPA9B 0.1872 GRP75 0777 Hs.979 PDHB 0.1855 Pyruvate dehydrogenase E1- subunit precursor 1249 Hs.226795 GSTP1 0.1773 Glutathione S-transferase pi (GST-pi) 1685 Hs.76136 TXN 0.1732 Thioredoxin 1205 Hs.82314 HPRT1 0.1588 HG phosphoribosyltransferase 1230 Hs.279860 TPT1 0.1466 Translationally controlled tumor protein (TCTP) 0603 Hs.181357 LAMR1 0.1463 LAMR 1358 Hs.28914 APRT 0.1399 Adenine phosphoribosyl transferase 1410 Hs.82113 DUT 0.1213 dUTP pyrophosphatase (dUTPase) 1825 Hs.112378 LIMS1 0.1213 Pinch-2 protein 0871 Hs.250502 CA8 0.1122 Carbonic anhydrase-related protein; Syntaxin 0289 Hs.82916 CCT6A 0.1106 Chaperonin-like protein 1143 Hs.11465 GSTTLp28 0.0997 Glutathione S-transferase homolog (GST homolog) 1456 Hs.118638 NME1 0.0932 Nm23 (NDPKA) 1598 Hs.278503 RIG 0.0905 RIIG (U32331) 1354 Hs.89761 ATP5D 0.0904 FIFO-type ATP synthase subunit d 1445 Hs.155485 HIP2 0.0843 Huntingtin interacting protein 2 (HIP2) 1479 Hs.177486 APP 0.0746 Amyloid B4A 0608 Hs.182265 KRT19 0.0439 Cytokeratin 19 1071 Hs.10842 RAN 0.0277 GTP-binding nuclear protein RAN(TC4) 0991 Hs.297939 CTSB 0.0254 Cathepsin B 0842 Hs.77274 PLAU 0.0248 Urokinase plasminogen activator 0823 Hs.198248 B4GALT1 0.0183  1,4-galactosyl transferase 0613 Hs.1247 APOA4 0.0176 Apolipoprotein A4 (ApoA4) 1338 Hs.104143 CLTA 0.0123 Clathrin light chain A 0902 Hs.5123 SID6–306 0.0117 Cytosolic inorganic pyrophosphatase 1688 Hs.1473 GRP 0.0040 Preprogastrin-releasing peptide 0265 Hs.274402 HSPA1B 0.0071 Heat shock-induced protein 1414 Hs.77541 ARF5 0.0096 ADP-ribosylation factor 1 0710 Hs.97206 HIP1 0.0114 Huntingtin interacting protein 1 (HIP1) 0532 Hs.170328 MSN 0.0132 Moesin/E 0525 Hs.284255 ALPP 0.0148 Alkaline phosphate, placental 0513 Hs.76901 PDIR 0.0289 Protein disulfide isomerase-related protein 5 1659 Hs.256697 HINT 0.0312 Protein kinase C inhibitor 1262 Hs.7016 RAB7 0.0362 Rab 7 protein 0190 Hs.184411 ALB 0.0470 Albumin 0948 Hs.2795 LDHA 0.0549 Lactate dehydrogenase-A (LDHA) 0502 Hs.180532 GPI 0.0575 Hsp89 0152 Hs.75410 HSPA5 0.0640 GRP78 1054 Hs.74276 CLIC1 0.0686 Nuclear chloride channel (RNCC protein) 0709 Hs.253495 SFTPD 0.0936 Pulmonary surfactant protein D 0867 Hs.78996 PCNA 0.0982 PCNA 0165 Hs.180414 HSPA8 0.1014 Heat shock cognate protein, 71 kDa 1109 Hs.75103 YWHAZ 0.1018 14-3-3 / 0137 Hs.554 SSA2 0.1032 Ro/ss-A antigen Molecular & Cellular Proteomics 1.4 305 Protein and mRNA Correlation in Lung Adenocarcinomas ABLE I—continued Spot Unigene Gene name r* Protein name 0278 Hs.4112 TCP1 0.1237 T-complex protein I,  subunit 1769 Hs.9614 NPM1 0.1738 B23/numatrin 0089 Hs.74335 HSPCB 0.2049 Hsp90 2511 Hs.153179 FABP5 0.2109 E-FABP/FABP5 1739 Hs.16488 CALR 0.2344 Calreticulin 32 1138 Hs.301961 GSTM4 0.2438 Glutathione S-transferase M4 (GST m4) 2533 Hs.77060 PSMB6 0.2512 Macropain subunit were performed as described previously (12, 13). Our 2D-PAGE sys- non-neoplastic lung tissues. Protein levels were determined tem allows us to run 20 gels at one time (one batch). Spot detection using quantitative 2D-PAGE analysis, and the separated pro- and quantification were accomplished utilizing Bio Image Visage Sys- tein polypeptides were identified using matrix-assisted laser tem software (Bioimage Corp., Ann Arbor, MI). The integrated inten- desorption/ionization mass spectrometry (MALDI-MS). The sity of each spot was calculated as the measured optical density corresponding mRNA levels for the identified proteins within units  mm . Of the total possible 2000 spots detectable on each gel, 820 spots on the gel of each sample were matched using a Gel-ed the same samples were determined using oligonucleotide match program with the same spots on a chosen “master” gel. In microarrays. Correlation analyses showed that protein abun- each sample, 250 ubiquitously expressed reference spots were used dance is likely a reflection of the transcription for a subset of to adjust for variations between gels, such as that created by subtle proteins, but translation and post-translational modifications differences in protein loading or gel staining. Slight differences be- also appear to influence the expression levels of many indi- cause of batch were corrected after spot-size quantification. Mass Spectrometry and 2D Western Blotting—Preparative 2D gels vidual proteins in lung adenocarcinomas. were run using extracts from A549 lung adenocarcinoma cells (ob- tained from ATCC) and using the identical experimental conditions as EXPERIMENTAL PROCEDURES the analytical 2D gels, except 30% more protein was loaded. The Tissues—Fifty-seven stage I and 19 stage III lung adenocarcino- resolved protein gels were silver-stained using successive incuba- mas, as well as nine non-neoplastic lung tissue samples, were used tions in 0.02% sodium thiosulfate for 2 min, 0.1% silver nitrate for 40 for protein and mRNA analyses. Patient consent was obtained, and min, and 0.014% formaldehyde plus 2% sodium carbonate for 10 the project was approved by the Institutional Review Board. All tis- min. For protein identification, protein polypeptides underwent trypsin sues were obtained after resection at the University of Michigan digestion followed by MALDI-MS using a MALDI-TOF Voyager-DE Health System between May 1991 and July 1998. Tissues were all mass spectrometer (Perseptive Biosystems, Framingham, MA). The snap-frozen in liquid nitrogen and then stored at 80 °C. The patients masses were compared with known trypsin digest databases using included 46 females and 30 males ranging in age from 40.9 to 84.6 the MS-FIT database (University of California, San Francisco; (average 63.8) years. Most patients (66/76) demonstrated a positive prospector.ucsf.edu/ucsfhtml3.2/msfit.htm). Some of the polypep- smoking history. Sixty-one tumor samples were classified as bron- tides included in the analysis had been identified prior to this study on chial-derived, 14 were classified as bronchoalveolar, and one had the basis of sequencing (14). The identified protein spots used in this both features. Eighteen tumor samples were classified as well differ- paper are shown in Fig. 1A. The method for 2D-PAGE Western blot entiated, 38 were classified as moderate, and 19 were classified as verification was as described previously (15). The 2D Western blots of poorly differentiated adenocarcinomas. Hematoxylin-stained cryostat GRP58 and Op18 are shown in Fig. 1, C and E; the others, such as sections (5 m), prepared from the same tumor pieces to be utilized GRP78, GRP75, HSP70, HSC70, KRT8, KRT18, KRT19, Vimentin, for protein and mRNA isolation, were evaluated by a pathologist and ApoJ, 14 –3-3, Annexin I, Annexin II, PGP9.5, DJ-1, GST-pi, and compared with hematoxylin- and eosin-stained sections made from 2 PGAM, are described elsewhere. paraffin blocks of the same tumors. Specimens were excluded from Statistical Analysis—Missing values were replaced with the mean analysis if they showed unclear or mixed histology (e.g. adenosqua- value of the protein spot. The transform x 3 log (1  x) was applied mous), tumor cellularity less than 70%, potential metastatic origin as to normalize all protein expression values. The relationship between indicated by previous tumor history, extensive lymphocytic infiltration, protein and mRNA expression levels within the same samples was or fibrosis or if the patient had received prior chemotherapy or examined using the Spearman correlation coefficient analysis (16). To radiotherapy. identify potentially significant correlations between gene and protein Oligonucleotide Array Hybridization—The HuGeneFL oligonucleo- expression, we used an analytical strategy similar to SAM (signifi- tide arrays (Affymetrix, Santa Clara, CA) containing 6800 genes were cance analysis of microarrays) (17), which uses a permutation tech- used in this study. Total RNA was isolated from all samples using nique to determine the significance of changes in gene expression Trizol reagent (Invitrogen). The resulting RNA was then subjected to between different biological states. To obtain permuted correlation further purification using RNeasy spin columns (Qiagen). Preparation coefficients between gene and protein expression, genes were ex- of cRNA, hybridization, and scanning of the HuGeneFL arrays were changed first in such a way that permutated correlation coefficient performed according to the manufacturer’s protocol (Affymetrix, were calculated based on pseudo pairs of genes and proteins. The Santa Clara, CA). Data analysis was performed using GeneChip 4.0 distribution of permutated correlation coefficients became stable after software. The gene expression profile of each tumor was normalized 60 permutations. This procedure was then repeated 60 times to to the median gene expression profile for the entire sample. Details of obtain 60 sets of permutated correlation coefficients. For each of the data trimming and normalization are described elsewhere (11). 60 permutations, the correlations of genes and proteins were ranked 2D-PAGE and Quantitative Protein Analysis—Tissue for both pro- tein and mRNA isolation came from contiguous areas of each sample. Protein separation using 2D-PAGE, silver staining, and digitization Chen et al., submitted for publication. 306 Molecular & Cellular Proteomics 1.4 Protein and mRNA Correlation in Lung Adenocarcinomas TABLE II Correlation coefficients of protein and mRNA where multiple isoforms were present on 2D gels r*, correlation coefficient value  0.2445; p  0.05. Values in boldface are significant at p  0.05. Spot Unigene Gene name r* Protein name 1494 Hs.81915 LAP18 0.4003 OP18 (Stathmin) 0957 Hs.77899 TPM1 0.3930 Tropomyosins 1–5 0353 Hs.289101 GRP58 0.3802 Protease disulfide isomerase (GRP58) 0855 Hs.169476 GAPD 0.3693 Glyceraldehyde-3-phosphate dehydrogenase 1198 Hs.41707 HSPB3 0.3668 Hsp27 1203 Hs.83848 TPI1 0.3395 Triose phosphate isomerase (TPI) 0523 Hs.65114 KRT18 0.3335 Cytokeratin 18 1492 Hs.81915 LAP18 0.3234 OP18 (Stathmin) 1493 Hs.81915 LAP18 0.3154 OP18 (Stathmin) 1181 Hs.78225 ANXA1 0.3102 Annexin variant I 0439 Hs.242463 KRT8 0.3049 Cytokeratin 8 0505 Hs.297753 VIM 0.2939 Vimentin 0593 Hs.297753 VIM 0.2809 Vimentin 1874 Hs.75313 AKR1B1 0.2790 Aldose reductase 0935 Hs.75544 YWHAH 0.2775 14-3-3 2524 Hs.78225 ANXA1 0.2612 Annexin I 2324 Hs.65114 KRT18 0.2601 Cytokeratin 18 1192 Hs.41707 HSPB3 0.2558 Hsp27 0350 Hs.289101 GRP58 0.2516 Phospholipase C (GRP58) 0992 Hs.75313 AKR1B1 0.2460 Aldose reductase 0861 Hs.75313 AKR1B1 0.0761 Aldose reductase 0853 Hs.75313 AKR1B1 0.0675 Aldose reductase 2503 Hs.76392 ALDH1 0.0565 Aldehyde dehydrogenase 0381 Hs.76392 ALDH1 0.0371 Aldehyde dehydrogenase 0371 Hs.76392 ALDH1 0.0680 Aldehyde dehydrogenase 1179 Hs.78225 ANXA1 0.2052 Annexin variant I 0762 Hs.78225 ANXA1 0.0739 Annexin I 0760 Hs.78225 ANXA1 0.0228 Annexin I 2506 Hs.217493 ANXA2 0.2223 Lipocotin (annexin II) 0772 Hs.217493 ANXA2 0.2080 Lipocotin (annexin II) 0723 Hs.217493 ANXA2 0.0701 Lipocotin 1239 Hs.93194 APOA1 0.1133 Apolipoprotein A1 (ApoA1) 1237 Hs.93194 APOA1 0.0373 Apolipoprotein A1 (ApoA1) 1234 Hs.93194 APOA1 0.0894 Apolipoprotein A1 (ApoA1) 0428 Hs.25 ATP5B 0.0080 ATP synthase  subunit precursor 0427 Hs.25 ATP5B 0.0122 ATP synthase  subunit precursor 0424 Hs.25 ATP5B 0.0992 ATP synthase  subunit precursor 0863 Hs.75106 CLU 0.0483 Apolipoprotein J (ApoJ) 0780 Hs.75106 CLU 0.0443 Apolipoprotein J (ApoJ) 1527 Hs.119140 EIF5A 0.0726 eIF-5A 1484 Hs.119140 EIF5A 0.0376 eIF-5A 1728 Hs.5241 FABP1 0.1916 L-FABP 1712 Hs.5241 FABP1 0.0473 L-FABP 0947 Hs.169476 GAPD 0.1745 Glyceraldehyde-3-phosphate dehydrogenase 1232 Hs.75207 GLO1 0.2249 Glyoxalase-I 1229 Hs.75207 GLO1 0.0450 Glyoxalase-1 1595 Hs.158300 HAP1 0.0137 Huntingtin-associated protein 1 (neuroan 1) 1810 Hs.75990 HP 0.4672 -Haptoglobin 1459 Hs.75990 HP 0.0802 -Haptoglobin 1458 Hs.75990 HP 0.0305 -Haptoglobin 0619 Hs.75990 HP 0.0461 B-haptoglobin 0615 Hs.75990 HP 0.0034 B-haptoglobin 1250 Hs.41707 HSPB3 0.1024 Hsp27 0549 Hs.79037 HSPD1 0.1074 Hsp60 0338 Hs.79037 HSPD1 0.2265 Hsp60 0333 Hs.79037 HSPD1 0.1383 Hsp60 0331 Hs.79037 HSPD1 0.1603 Hsp60 2381 Hs.65114 KRT18 0.2016 Cytokeratin 18 0535 Hs.65114 KRT18 0.1106 Cytokeratin 18 Molecular & Cellular Proteomics 1.4 307 Protein and mRNA Correlation in Lung Adenocarcinomas TABLE II—continued Correlation coefficients of protein and mRNA where multiple isoforms were present on 2D gels r*, correlation coefficient value  0.2445; p  0.05. Values in boldface are significant at p  0.05. Spot Unigene Gene name r* Protein name 0529 Hs.65114 KRT18 0.1279 Cytokeratin 18 0528 Hs.65114 KRT18 0.0414 Cytokeratin 18 0527 Hs.65114 KRT18 0.0436 Cytokeratin 18 0514 Hs.65114 KRT18 0.0733 Cytokeratin 18 0451 Hs.242463 KRT8 0.0111 Cytokeratin 8 0446 Hs.242463 KRT8 0.0347 Cytokeratin 8 0444 Hs.242463 KRT8 0.1311 Cytokeratin 8 0443 Hs.242463 KRT8 0.0942 Cytokeratin 8 1488 Hs.81915 LAP18 0.0495 OP18 (Stathmin) 0321 Hs.75655 P4HB 0.0546 PDI (proly-4-OH-B) 0320 Hs.75655 P4HB 0.0041 PDI (proly-4-OH-B) 1063 Hs.75323 PHB 0.0441 Prohibitin 0837 Hs.75323 PHB 0.1402 Prohibitin 0326 Hs.297681 SERPINA1 0.0227 -1-Antitripsin 0322 Hs.297681 SERPINA1 0.0277 -1-Antitripsin 0241 Hs.297681 SERPINA1 0.0148 -1-Antitripsin 1280 Hs.301254 SFTPA1 0.1488 Pulmonary surfactant-associated protein 1278 Hs.301254 SFTPA1 0.2040 Pulmonary surfactant-associated protein 0866 Hs.73980 TNNT1 0.1162 Troponin T 0778 Hs.73980 TNNT1 0.0740 Troponin T 1213 Hs.83848 TPI1 0.0024 Triose phosphate isomerase (TPI) 1210 Hs.83848 TPI1 0.0490 Triose phosphate isomerase (TPI) 1207 Hs.83848 TPI1 0.1615 Triose phosphate isomerase (TPI) 1204 Hs.83848 TPI1 0.0209 Triose phosphate isomerase (TPI) 1202 Hs.83848 TPI1 0.0721 Triose phosphate isomerase (TPI) 1161 Hs.83848 TPI1 0.2265 Triose phosphate isomerase (TPI) 1052 Hs.77899 TPM1 0.1040 Tropomysin clean-product 1039 Hs.77899 TPM1 0.2999 Cytoskeletal tropomyosin 1035 Hs.77899 TPM1 0.3821 Tropomyosin 0783 Hs.77899 TPM1 0.0757 Tropomyosins 1–5 1574 Hs.194366 TTR 0.0065 Transthyretin 0809 Hs.194366 TTR 0.0399 Transthyretin multimere 2202 Hs.76118 UCHL1 0.0220 Ubiquitin carboxyl-terminal hydrolase isozyme L1 1246 Hs.76118 UCHL1 0.1261 Ubiquitin carboxyl-terminal hydrolase isozyme L1 1242 Hs.76118 UCHL1 0.1473 Ubiquitin carboxyl-terminal hydrolase isozyme L1 0606 Hs.297753 VIM 0.0951 Vimentin 0594 Hs.297753 VIM 0.2664 Vimentin-derived protein (vid4) 0508 Hs.297753 VIM 0.1008 Vimentin-derived protein (vid2) 0419 Hs.297753 VIM 0.0032 Vimentin-derived protein (vid1) 1279 Hs.75544 YWHAH 0.0059 14-3-3 such that  (i) denotes the ith largest correlation coefficient for pth p protein spots on 2D gels representing 98 genes and com- permutation. Hence, the expected correlation coefficient,  (i), was the pared protein levels with mRNA levels for a cohort of 85 lung average over the 60 permutations,  (i)  (i)/60. A scatter plot of E p  1 p adenocarcinomas and normal lung samples. Of the 165 pro- observed correlations ((i)) versus the expected correlations is shown in tein spots, 69 proteins were represented by only one known Fig. 2D. For this study, we chose threshold 0.115 so that correlation would be considered significant if absolute value of difference between spot on 2D gels for an individual gene, whereas 96 protein (i) and  (i) was greater than the threshold. Twenty-nine (including one E spots showed multiple protein products from 29 different with observed correlation coefficient 0.4672) of 165 pairs of gene and genes. 2D Western blotting verified the proteins identified by protein expression were called significant in such criteria, and the mass spectrometry when specific antibodies were available. permuted data generated an average of 5.1 falsely significant pairs of Spearman correlation coefficients of the proteins and their gene and protein expression. This provided an estimated false dis- covery rate (the percentage of pairs of gene and protein expression associated mRNA for each protein spot were generated using identified by chance) for our data set. all 76 lung adenocarcinomas and nine non-neoplastic lung tissues (see Tables I and II, and see Figs. 1 and 2). The RESULTS correlation coefficients (r) ranged from 0.467 to 0.442 (Fig. Correlation of Individual Proteins and mRNA Expression 2D). A total of 28 protein spots (21 genes) were found to have within Each Tumor—We have examined quantitatively 165 a statistically significant correlation between expression of 308 Molecular & Cellular Proteomics 1.4 Protein and mRNA Correlation in Lung Adenocarcinomas FIG.1. A, digital image of a silver-stained 2D-PAGE separation of a stage I lung adenocarcinoma showing protein spots separated by molecular mass (MW) and isoelectric point (PI). Twenty-eight protein spots whose expression levels are correlated with mRNA abundance are indicated by the black arrows. B, the outlined areas of A showing protein GRP58. C, 2D Western blot of GRP58 from the A549 lung adenocarcinoma cell line. D, the outlined areas of A showing the protein isoforms of Op18. E, 2D Western blot of Op18 from A549 cells. their protein and mRNA (r  0.2445; p  0.05). This accounts tween protein and mRNA expression (r  0.0495). Similarly, just for 17% (28/165) of the 165 protein spots. Among the 69 one of five quantified isoforms of cytokeratin 8 (spot 439) dem- genes for which only a single protein spot was known (Table onstrated a statistically significant correlation between protein I), nine genes (9/69, 13%) were observed to show a statisti- and mRNA abundance (r  0.3049; p  0.05) (Table II). cally significant relationship between protein and mRNA In addition to differences in the relationship between mRNA abundance (r  0.2445; p  0.05). The proteins whose ex- levels and protein expression among separate isoforms, some pression levels were correlated with their mRNA abundance genes with very comparable mRNA levels showed a 24-fold included those involved in signal transduction, carbohydrate difference in their protein expression. Genes with comparable metabolism, apoptosis, protein post-translational modifica- protein expression levels also showed up to a 28-fold vari- tion, structural proteins, and heat shock proteins (Table III). ance in their mRNA levels. Individual Isoforms of the Same Protein Have Different Lack of Correlation for mRNA and Protein Expression when Protein/mRNA Correlation Coefficients—Of the 165 protein Using Average Tumor Values across All 165 Protein Spots (98 spots, 96 represent protein products of 29 genes with at least Genes)—The relationship between mRNA and protein expres- two isoforms. Among these 96 protein spots, 19 (19/96 pro- sion was also examined by using the average expression tein spots, 20%) showed a statistically significant correlation values for all samples. To analyze this relationship using this between their protein and mRNA expression (r  0.2445; p approach, the average value for each protein or mRNA was 0.05) (Table II) and represented 12 genes (12/29, 41%). Individ- generated using all 85 lung tissue samples. The range of ual isoforms of the same protein demonstrated different normalized average protein values ranged from 0.0646 to protein/mRNA correlation coefficients. For example, 2D-PAGE/ 0.0979 (raw value 0.0036 to 4.1947), and the range for mRNA Western analysis revealed four isoforms of OP18 differing in was from 0 to 15260.5 for all 165 individual protein spots. The regards to isoelectric point but similar in molecular weight. Spearman correlation coefficient for the whole data set (165 Three of the four isoforms (spots 1492, 1493, and 1494) showed protein spots/98 genes) was 0.025 (Fig. 3A). Even for the 28 a statistically significant correlation between their protein and mRNA abundance (r  0.3234, 0.3154, and 0.4003, respective- protein spots (Fig. 2D) that were found to have a statistically ly). The forth isoform (spot 1488) showed no correlation be- significant correlation between their mRNA and protein, use of Molecular & Cellular Proteomics 1.4 309 Protein and mRNA Correlation in Lung Adenocarcinomas FIG.2. A–C, plots showing the correlation between mRNA and protein for the three selected genes Op18, Annexin IV, and GAPD for all 76 lung adenocarcinomas and nine non-neoplastic lung samples (p  0.05). D, distribution of all 165 Spearman correlation coefficients (r) and verification analysis using SAM. A more detailed description of the method is provided under “Experimental Procedures.” Approximately 17% of the 165 proteins demonstrate a significant correlation between mRNA and protein levels as demonstrated by the values shown beyond the outer range of threshold  0.115. Normalized protein values were used, thus negative values for some proteins are observed. the average value resulted in a correlation coefficient value of stage III (n  19) lung adenocarcinomas (Table III). The num- 0.035, which was not significant (Fig. 3B). ber of non-neoplastic lung samples (n  9) was insufficient for Lack of a Relationship between Protein/mRNA Correlation a separate correlation analysis of this group. Many of the Coefficients and Average Protein Abundance—To determine protein spots represent one of several known protein isoforms whether an absolute protein level might influence the corre- for a given gene. The majority of genes (16/21) did not differ in lation with mRNA, the mean value of each protein (relative the protein/mRNA correlation between stage I and stage III abundance) and the Spearman protein/mRNA correlation co- tumors indicating a similar regulatory relationship between the efficients among all 85 samples were examined. No relation- mRNA and protein spot. GRP-58, PSMC, SOD1, TPI1, and ship between the protein abundance and the correlation co- VIM, however, were found to demonstrate significant differ- efficients was observed (r  0.039; p  0.05). A detailed ences in the correlation coefficients between stage I and analysis of separate subsets of proteins with differing levels of stage III lung adenocarcinomas. For GRP-58, PSMC, and VIM abundance (less than 0.0014, larger than 0.0014, or larger the change in the correlation coefficient was because of a than 0.0077) also showed a lack of correlation between mRNA relative increase in protein expression in stage III tumors. For and protein expression among the 83 (50%), 82 (50%), and 41 SOD and TPI the change resulted from a relative decrease in (25%) of 165 total protein spots, respectively (r  0.016, 0.08, expression of this specific protein in stage III tumors. and 0.172, respectively). Stage-related Changes in the Protein/mRNA Correlation DISCUSSION Coefficients—To determine whether the 21 genes (28 protein spots) showing a significant correlation between the protein Relatively little is known about the regulatory mechanisms and mRNA expression among all samples demonstrate controlling the complex patterns of protein abundance and changes in this relationship during tumor progression, the post-translational modification in tumors. Most reports con- correlations were examined separately for stage I (n  57) and cerning the regulation of protein translation have focused on 310 Molecular & Cellular Proteomics 1.4 Protein and mRNA Correlation in Lung Adenocarcinomas TABLE III Stage-dependent analysis of protein-mRNA correlation coefficients r, correlation coefficient. Values in boldface indicate a significant difference between stage I and stage III. Spot Gene name r (Stage I) r (Stage III) Function 1874 AKR1B1 0.269 0.106 Carbohydrate metabolism; electron transporter 2524 ANXA1 0.184 0.572 Phospholipase inhibitor; signal transduction 0994 ANXA4 0.660 0.362 Phospholipase inhibitor 0963 ANXA5 0.241 0.390 Phospholipase inhibitor; calcium binding; phospholipid binding 1314 DJ-1 0.363 0.354 Signal transduction 1405 FTL 0.126 0.358 Iron storage protein 0855 GAPD 0.243 0.581 Carbohydrate metabolism (glycolysis regulation) 0350 GRP58 0.327 0.087 Signal transduction; protein disulfide isomerase 0264 HNRPK 0.360 0.243 RNA-binding protein (RNA processing/modification) 1192 HSPB3 0.457 0.633 Heat shock protein 0523 KRT18 0.115 0.371 Structural protein 0439 KRT8 0.323 0.436 Structural protein 1492 LAP18 0.483 0.663 Signal transduction; cell growth and maintenance 1638 LGALS1 0.200 0.528 Apoptosis; cell adhesion; cell size control 1252 PSMC 0.253 0.060 Protein degradation 1104 SFN 0.465 0.475 Signal transduction (protein kinase C inhibitor) 1454 SOD1 0.352 0.079 Oxidoreductase 1203 TPI1 0.378 0.009 Carbohydrate metabolism 0957 TPM1 0.475 0.225 Structural protein (muscle); control of heart 0593 VIM 0.054 0.556 Structural protein 0935 YWHAH 0.283 0.210 Signal transduction one or several protein products (18). Celis et al. (19) found a anisms of protein translation among different cells or as good correlation between transcript and protein levels among measured in different laboratories (23). 40 well resolved, abundant proteins using a proteomic and In this study, we examined 165 protein spots identified in microarray study of bladder cancer. By comparing the mRNA lung adenocarcinomas. Ninety-six protein spots, representing and protein expression levels within the same tumor samples, the products of 29 genes, contained at least two protein we found that 17% (28/165) of the protein spots (21/98 genes) isoforms. Nineteen of 96 protein spots, representing 12 show a statistically significant correlation between mRNA and genes, were shown to have a statistically significant correla- protein. These proteins appear to represent a diverse group of tion between their protein and mRNA expression, suggesting gene products and include those involved in signal transduc- that the levels of these proteins reflects the transcription of the tion, carbohydrate metabolism, protein modification, cell struc- corresponding genes. Differences in protein/mRNA correlations ture, heat shock, and apoptosis. These results suggest that were found among the individual isoforms of a given protein. For expression of this subset of 165 proteins is likely to be regulated example, of the four OP18 isoforms, three showed a statistically at the transcriptional level in these tissues. The majority of the significant correlation between the protein and mRNA expres- protein isoforms, however, did not correlate with mRNA levels, sion levels. The lack of relationship for the one isoform, how- and thus their expression is regulated by other mechanisms. We ever, indicates that individual protein isoforms of the same gene also observed a subset of proteins that demonstrated a nega- product can be regulated differentially. This is not unexpected tive correlation with the mRNA expression values; for example and likely reflects other post-translational mechanisms that can -haptoglobin demonstrated a strong negative correlation with influence isoform abundance in tissues and cancer. its mRNA expression values. This may reflect negative feedback In addition to the analyses of the correlation of mRNA/ on the mRNA or the protein or the presence of other regulatory protein within the same tumor samples, we also tested the influences that are not understood currently. global relationship between mRNA and the corresponding Post-translational modification or processing will result in protein abundance across all 165 protein spots in the lung individual protein products of the same gene migrating to samples. A protein and mRNA average value for each gene different locations on 2D-PAGE gels (20). Because the identity was generated using all 85 lung tissues samples. We ob- of all possible isoforms for each protein examined has not served a very wide range of normalized average protein and been characterized completely, this may influence the corre- mRNA values. The correlation coefficient generated using this lation analyses performed in this study. This is partly because average value data set was 0.025, and even for the 28 of limitations of the 2D-PAGE and mass spectrometry tech- protein spots that showed a statistically significant correlation nologies (21, 22). Potential inconsistencies between mRNA between individual mRNA and proteins, the correlation value and protein correlations that have been reported may also be was only 0.035. This suggests that it is not possible to because of differences, even in the same gene, in the mech- predict overall protein expression levels based on average Molecular & Cellular Proteomics 1.4 311 Protein and mRNA Correlation in Lung Adenocarcinomas FIG.3. The overall correlation of mRNA and protein levels across all 165 protein spots (A) and across 28 protein spots that contained individ- ual r values larger than 0.244 (B) are shown. Each protein or mRNA mean value was calculated based on all 76 lung adenocarcinomas and nine non- neoplastic lung samples using quantita- tive 2D-PAGE and Affymetrix oligonu- cleotide microarrays. The Spearman correlation coefficients for the two data sets (A and B) were 0.025 and 0.035, respectively, indicating a lack of correla- tion if mean values for mRNA and protein for all samples is used. mRNA abundance in lung cancer samples. This conclusion is A comparison between the mean value of each protein and also supported by previous results from Anderson and Seil- the correlation coefficient generated using all 85 tissue sam- hamer (24), who examined 19 genes in human liver cells, and ples did not reveal a strong relationship between the overall by Gygi et al. (25), who examined 106 genes in yeast. Both protein abundance and the correlation coefficients (r  0.039; studies found a lack of correlation between mRNA and protein p  0.05). Detailed analysis of different subsets of protein abun- expression when average or overall levels were used. dance also failed to show a correlation between mRNA and A good correlation was reported when the 11 most abun- protein expression. Thus in contrast to yeast, a relationship dant proteins were examined in yeast (25), suggesting that the between mRNA/protein correlation coefficient and protein level of protein abundance may be a factor that may influence abundance in human lung adenocarcinomas was not observed. the correlation between mRNA and protein. In the present The results of this study indicate that the level of protein study, a fairly wide range of mean protein values among 165 abundance in lung adenocarcinomas is associated with the protein spots in lung adenocarcinomas was observed, and corresponding levels of mRNA in 17% (28 proteins) of the the correlation coefficients also varied from 0.467 to 0.442. total 165 protein spots examined. This was substantially 312 Molecular & Cellular Proteomics 1.4 Protein and mRNA Correlation in Lung Adenocarcinomas higher than the amount predicted to result by chance alone 8. Wang, D. G., Fan, J. B., Siao, C. J., Berno, A., Young, P., Sapolsky, R., Ghandour, G., Perkins, N., Winchester, E., Spencer, J., Kruglyak, L., (which was 5.1) and suggests that a transcriptional mecha- Stein, L., Hsie, L., Topaloglou, T., Hubbell, E., Robinson, E., Mittmann, nism likely underlies the abundance of these proteins in lung M., Morris, M. S., Shen, N., Kilburn, D., Rioux, J., Nusbaum, C., Rozen, adenocarcinomas. We also demonstrate that the expression S., Hudson, T. J., Lander, E. S. (1998) Large-scale identification, map- ping, and genotyping of single-nucleotide polymorphisms in the human of individual isoforms of the same protein may or may not genome. Science 280, 1077–1082 correlate with the mRNA, indicating that separate and likely 9. Pease, A. C., Solas, D., Sullivan, E. J., Cronin, M. T., Holmes, C. P., Fodor, post-translational mechanisms account for the regulation of S. P. (1994) Light-generated oligonucleotide arrays for rapid DNA se- quence analysis. Proc. Natl. Acad. Sci. U. S. A. 91, 5022–5026 isoform abundance. These mechanisms may also account for 10. Bhattacharjee, A., Richards, W. G., Staunton, J., Li, C., Monti, S., Vasa, P., the differences in the correlation coefficients observed between Ladd, C., Beheshti, J., Bueno, R., Gillette, M., Loda, M., Weber, G., Mark, stage I and stage III tumors, indicating that specific protein E. J., Lander, E. S., Wong, W., Johnson, B. E., Golub, T. R., Sugarbaker, D. J., Meyerson, M. (2001) Classification of human lung carcinomas by isoforms show regulatory changes during tumor progression. mRNA expression profiling reveals distinct adenocarcinoma subclasses. Further studies in lung adenocarcinomas will examine the rela- Proc. Natl. Acad. Sci. U. S. A. 98, 13790 –13795 tionship between the expression of individual protein isoforms 11. Giordano, T. J., Shedden, K. A., Schwartz, D. R., Kuick, R., Taylor, J. M. G., and specific clinical-pathological features of these tumors, such Lee, N., Misek, D. E., Greenson, J. K., Kardia, S. L. R., Beer, D. G., Rennert, G., Cho, K. R., Gruber, S. B., Fearon, E. R., Hanash, S. (2001) as the presence of angiolymphatic invasion, and nodal or pleu- Organ-specific molecular classification of lung, colon and ovarian adeno- ral surface involvement. The potential to identify specific protein carcinomas using gene expression profiles. Am. J. Pathol. 159, 1231–1238 isoforms associated with biological behavior in lung adenocar- 12. Strahler, J. R., Kuick, R., Hanash, S. M. (1989) in Protein Structure: A Practical Approach (Creighton, T., ed) pp. 65–92, IRL Press, Oxford cinomas would be of considerable interest and will add to our 13. Merril, C. R., Dunau, M. L., Goldman, D. (1981) A rapid sensitive silver stain understanding of the regulation of gene products by transcrip- for polypeptides in polyacrylamide gels. Anal. Biochem. 101, 201–207 tional, translational, and post-translational mechanisms. 14. Hanash, S. M., Strahler, J. R., Chan, Y., Kuick, R., Teichroew, D., Neel, J. V., Hailat, N., Keim, D. R., Gratiot-Deans, J., Ungar, D., Richardson, B. C. Acknowledgments—We thank Kerby A. Shedden, Rork D. Kuick, (1993) Data base analysis of protein expression patterns during T-cell Eric Puravs, Robert Hinderer, Melissa C. Krause, and Christopher ontogeny and activation. Proc. Natl. Acad. Sci. U. S. A. 90, 3314 –3318 15. Brichory, F. M., Misek, D. E., Yim, A. M., Krause, M. C., Giordano, T. J., Wood for assistance in this study. Beer, D. G., Hanash, S. M. (2001) An immune response manifested by the common occurrence of annexins I and II autoantibodies and high * This work was supported by NCI, National Institutes of Health circulating levels of IL-6 in lung cancer. Proc. Natl. Acad. Sci. U. S. A. 98, Grant U19 CA-85953. The costs of publication of this article were 9824 –9829 defrayed in part by the payment of page charges. This article must 16. Lavens-Phillips, S. E., MacGlashan, D. W., Jr. (2000) The tyrosine kinases therefore be hereby marked “advertisement” in accordance with 18 p53/56lyn and p72syk are differentially expressed at the protein level but U.S.C. Section 1734 solely to indicate this fact. not at the messenger RNA level in nonreleasing human basophils. Am. J. ‡‡ To whom correspondence should be addressed: General Tho- Respir. Cell Mol. Biol. 23, 566 –571 racic Surgery, SRB II, B560, Box 0686, University of Michigan Medical 17. Tusher, V. G., Tibshirani, R., Chu, G. (2001) Significance analysis of mi- School, Ann Arbor, MI 48109-086; E-mail: [email protected]. croarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. U. S. A. 98, 5116 –5121 REFERENCES 18. Tew, K. D., Monks, A., Barone, L., Rosser, D., Akerman, G., Montali, J. A., Wheatley, J. B., Schmidt, D. E., Jr. (1996) Glutathione-associated en- 1. Ideker, T., Thorsson, V., Ranish, J. A., Christmas, R., Buhler, J., Eng, J. K., zymes in the human cell lines of the National Cancer Institute Drug Bumgarner, R., Goodlett, D. R., Aebersold, R., Hood, L. (2001) Integrated Screening Program. Mol. Pharmacol. 50, 149 –159 genomic and proteomic analyses of a systematically perturbed meta- 19. Celis, J. E., Kruhoffer, M., Gromova, I., Frederiksen, C., Ostergaard, M., bolic network. Science 292, 929 –934 Thykjaer, T., Gromov, P., Yu, J., Palsdottir, H., Magnusson, N., Orntoft, 2. Liang, P., Pardee, A. B. (1998) Differential display. A general protocol. Mol. T. F. (2000) Gene expression profiling: monitoring transcription and Biotechnol. 10, 261–267 translation products using DNA microarrays and proteomics. FEBS Lett. 3. Porter, D. A., Krop, I. E., Nasser, S., Sgroi, D., Kaelin, C. M., Marks, J. R., 480, 2–16 Riggins, G., Polyak, K. (2001) A sage (serial analysis of gene expression) 20. Anderson, N. L., Anderson, N. G. (1998) Proteome and proteomics: new view of breast tumor progression. Cancer Res. 61, 5697–5702 technologies, new concepts, and new words. Electrophoresis 19, 4. Bittner, M., Meltzer, P., Chen, Y., Jiang, Y., Seftor, E., Hendrix, M., 1853–1861 Radmacher, M., Simon, R., Yakhini, Z., Ben-Dor, A., Sampas, N., 21. Gygi, S. P., Corthals, G. L., Zhang, Y., Rochon, Y., Aebersold, R. (2000) Dougherty, E., Wang, E., Marincola, F., Gooden, C., Lueders, J., Glat- Evaluation of two-dimensional gel electrophoresis-based proteome anal- felter, A., Pollock, P., Carpten, J., Gillanders, E., Leja, D., Dietrich, K., ysis technology. Proc. Natl. Acad. Sci. U. S. A. 97, 9390 –9395 Beaudry, C., Berens, M., Alberts, D., Sondak, V. (2000) Molecular 22. Fey, S. J., Larsen, P. M. (2001) 2D or not 2D. Two-dimensional gel elec- classification of cutaneous malignant melanoma by gene expression trophoresis. Curr. Opin. Chem. Biol. 5, 26 –33 profiling. Nature 406, 536 –540 23. McBride, S., Walsh, D., Meleady, P., Daly, N., Clynes, M. (1999) Bromode- 5. Fung, E. T., Wright, G. L., Jr., Dalmasso, E. A. (2000) Proteomic strategies oxyuridine induces keratin protein synthesis at a posttranscriptional level for biomarker identification: progress and challenges. Curr. Opin. Mol. in human lung tumor cell lines. Differentiation 64, 185–193 Ther. 2, 643– 650 24. Anderson, L., Seilhamer, J. (1997) A comparison of selected mRNA and 6. Davidson, D., Baldock, R. (2001) Bioinformatics beyond sequence: map- ping gene function in the embryo. Nat. Rev. Genet. 2, 409 – 417 protein abundances in human liver. Electrophoresis 18, 533–537 7. Chee, M., Yang, R., Hubbell, E., Berno, A., Huang, X. C., Stern, D., Winkler. 25. Gygi, S. P., Rochon, Y., Franza, B. R., Aebersold, R. (1999) Correlation J., Lockhart, D. J., Morris, M. S., Fodor, S. P. (1996) Accessing genetic between protein and mRNA abundance in yeast. Mol. Cell. Biol. 19, information with high-density DNA arrays. Science 274, 610 – 614 1720 –1730 Molecular & Cellular Proteomics 1.4 313 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Molecular & Cellular Proteomics Unpaywall

Discordant Protein and mRNA Expression in Lung Adenocarcinomas

Download PDF
10 pages

Loading next page...
 
/lp/unpaywall/discordant-protein-and-mrna-expression-in-lung-adenocarcinomas-hwsgWY7C0x

References (28)

Publisher
Unpaywall
ISSN
1535-9476
DOI
10.1074/mcp.m200008-mcp200
Publisher site
See Article on Publisher Site

Abstract

Research Discordant Protein and mRNA Expression in Lung Adenocarcinomas* Guoan Chen‡, Tarek G. Gharib‡, Chiang-Ching Huang§, Jeremy M. G. Taylor§, David E. Misek¶, Sharon L. R. Kardia, Thomas J. Giordano**, Mark D. Iannettoni‡, Mark B. Orringer‡, Samir M. Hanash¶, and David G. Beer‡ ‡‡ lung cancer and are now the most common histologic type. The relationship between gene expression measured at the mRNA level and the corresponding protein level is not Functional genomics, broadly defined as the comprehensive well characterized in human cancer. In this study, we analysis of genes and their products, have become a recent compared mRNA and protein expression for a cohort of focus of the life sciences (1). Application of these approaches to genes in the same lung adenocarcinomas. The abun- lung adenocarcinomas has the potential to aid in the identifica- dance of 165 protein spots representing 98 individual tion of high risk patients with resectable early stage lung cancer genes was analyzed in 76 lung adenocarcinomas and nine that may benefit from adjuvant therapy, as well as to identify non-neoplastic lung tissues using two-dimensional poly- new therapeutic targets. In human lung cancer, however, little is acrylamide gel electrophoresis. Specific polypeptides currently understood regarding the relationship between gene were identified using matrix-assisted laser desorption/ expression as determined by measuring mRNA levels and the ionization mass spectrometry. For the same 85 samples, corresponding abundance of the protein products. mRNA levels were determined using oligonucleotide mi- A number of powerful techniques for analysis of gene ex- croarrays, allowing a comparative analysis of mRNA and protein expression among the 165 protein spots. Twenty- pression have been used including differential display (2), eight of the 165 protein spots (17%) or 21 of 98 genes serial analysis of gene expression (3), DNA microarrays (4), (21.4%) had a statistically significant correlation between and proteomics via two-dimensional polyacrylamide gel elec- protein and mRNA expression (r > 0.2445; p < 0.05); trophoresis and mass spectrometry (5). Bioinformatics tools however, among all 165 proteins the correlation coeffi- have also been developed to help determine quantitative cient values (r) ranged from 0.467 to 0.442. Correlation mRNA/protein expression profiles of all types of cells and coefficient values were not related to protein abundance. tissues (6) and now can be applied to benign and malignant Further, no significant correlation between mRNA and tumors. DNA microarrays (cDNA and oligonucleotide) permit protein expression was found (r 0.025) if the average the parallel assessment of thousands of genes and have been levels of mRNA or protein among all samples were applied utilized in gene expression monitoring (7), polymorphism anal- across the 165 protein spots (98 genes). The mRNA/ ysis (8), and DNA sequencing (9). Recent studies have fo- protein correlation coefficient also varied among pro- teins with multiple isoforms, indicating potentially sep- cused on classification or identification of subgroups of lung arate isoform-specific mechanisms for the regulation of tumors using DNA microarrays (10, 11). The use of mRNA protein abundance. Among the 21 genes with a signifi- expression patterns by themselves, however, is insufficient for cant correlation between mRNA and protein, five genes understanding the expression of protein products, as addi- differed significantly between stage I and stage III lung tional post-transcriptional mechanisms, including protein adenocarcinomas. Using a quantitative analysis of mRNA translation, post-translational modification, and degradation, and protein expression within the same lung adenocarci- may influence the level of a protein present in a given cell or nomas, we showed that only a subset of the proteins tissue. Proteomic analyses, a complementary technology to exhibited a significant correlation with mRNA abundance. DNA microarrays for monitoring gene expression, involves Molecular & Cellular Proteomics 1:304 –313, 2002. protein separation and quantitative assessment of protein spots using 2D -PAGE and protein identification using mass spectrometry. By combining proteomic and transcriptional Lung cancer is the leading cause of cancer death for both analyses of the same samples, however, it may be possible to men and women in the United States. Adenocarcinomas of understand the complex mechanisms influencing protein ex- the lung comprise 40% of all new cases of non-small cell pression in human cancer. In this study, we determined mRNA and protein levels for From the Departments of ‡Surgery, §Biostatistics, Epidemiology, 165 proteins (98 genes) in 76 lung adenocarcinomas and nine **Pathology, and ¶Pediatrics, University of Michigan, Ann Arbor, Michigan 48109 Received, January 21, 2002, and in revised form, March 4, 2002 Published, MCP Papers in Press, March 12, 2001, DOI The abbreviations used are: 2D, two-dimensional; MALDI-MS, 10.1074/mcp.M200008-MCP200 matrix-assisted laser desorption/ionization mass spectrometry. 304 Molecular & Cellular Proteomics 1.4 © 2002 by The American Society for Biochemistry and Molecular Biology, Inc. This paper is available on line at http://www.mcponline.org This is an Open Access article under the CC BY license. Protein and mRNA Correlation in Lung Adenocarcinomas TABLE I Correlation coefficients of protein and mRNA where only one spot was present on 2D gels r*, correlation coefficient value  0.2445; p  0.05. Values in boldface are significant at p  0.05. Spot Unigene Gene name r* Protein name 1104 Hs.184510 SFN 0.4337 14-3-3 0994 Hs.77840 ANXA4 0.4219 Annexin IV 1314 Hs.10958 DJ-1 0.3982 DJ-1 protein/MER5 1454 Hs.75428 SOD1 0.3863 Superoxide dismutase (Cu-Zn) 1638 Hs.227751 LGALS1 0.3318 Galectin 1 0264 Hs.129548 HNRPK 0.3034 Transformation up-regulated nuclear protein 1405 Hs.111334 FTL 0.2849 Ferritin light chain 0963 Hs.300711 ANXA5 0.2468 Annexin V 1252 Hs.4745 PSMC 0.2445 26 S proteasome p28 0906 Hs.234489 LDHB 0.4420 L-lactate dehydrogenase H chain (LDH-B) 1171 Hs.241515 COX11 0.2310 COX 11 1160 Hs.181013 PGAM1 0.2023 Phosphoglycerate mutase 0759 Hs.74635 DLD 0.1965 Dihydrolipoamide dehydrogenase precursor 1193 Hs.83383 AOE372 0.1932 Antioxidant enzyme AOE372 0172 Hs.3069 HSPA9B 0.1872 GRP75 0777 Hs.979 PDHB 0.1855 Pyruvate dehydrogenase E1- subunit precursor 1249 Hs.226795 GSTP1 0.1773 Glutathione S-transferase pi (GST-pi) 1685 Hs.76136 TXN 0.1732 Thioredoxin 1205 Hs.82314 HPRT1 0.1588 HG phosphoribosyltransferase 1230 Hs.279860 TPT1 0.1466 Translationally controlled tumor protein (TCTP) 0603 Hs.181357 LAMR1 0.1463 LAMR 1358 Hs.28914 APRT 0.1399 Adenine phosphoribosyl transferase 1410 Hs.82113 DUT 0.1213 dUTP pyrophosphatase (dUTPase) 1825 Hs.112378 LIMS1 0.1213 Pinch-2 protein 0871 Hs.250502 CA8 0.1122 Carbonic anhydrase-related protein; Syntaxin 0289 Hs.82916 CCT6A 0.1106 Chaperonin-like protein 1143 Hs.11465 GSTTLp28 0.0997 Glutathione S-transferase homolog (GST homolog) 1456 Hs.118638 NME1 0.0932 Nm23 (NDPKA) 1598 Hs.278503 RIG 0.0905 RIIG (U32331) 1354 Hs.89761 ATP5D 0.0904 FIFO-type ATP synthase subunit d 1445 Hs.155485 HIP2 0.0843 Huntingtin interacting protein 2 (HIP2) 1479 Hs.177486 APP 0.0746 Amyloid B4A 0608 Hs.182265 KRT19 0.0439 Cytokeratin 19 1071 Hs.10842 RAN 0.0277 GTP-binding nuclear protein RAN(TC4) 0991 Hs.297939 CTSB 0.0254 Cathepsin B 0842 Hs.77274 PLAU 0.0248 Urokinase plasminogen activator 0823 Hs.198248 B4GALT1 0.0183  1,4-galactosyl transferase 0613 Hs.1247 APOA4 0.0176 Apolipoprotein A4 (ApoA4) 1338 Hs.104143 CLTA 0.0123 Clathrin light chain A 0902 Hs.5123 SID6–306 0.0117 Cytosolic inorganic pyrophosphatase 1688 Hs.1473 GRP 0.0040 Preprogastrin-releasing peptide 0265 Hs.274402 HSPA1B 0.0071 Heat shock-induced protein 1414 Hs.77541 ARF5 0.0096 ADP-ribosylation factor 1 0710 Hs.97206 HIP1 0.0114 Huntingtin interacting protein 1 (HIP1) 0532 Hs.170328 MSN 0.0132 Moesin/E 0525 Hs.284255 ALPP 0.0148 Alkaline phosphate, placental 0513 Hs.76901 PDIR 0.0289 Protein disulfide isomerase-related protein 5 1659 Hs.256697 HINT 0.0312 Protein kinase C inhibitor 1262 Hs.7016 RAB7 0.0362 Rab 7 protein 0190 Hs.184411 ALB 0.0470 Albumin 0948 Hs.2795 LDHA 0.0549 Lactate dehydrogenase-A (LDHA) 0502 Hs.180532 GPI 0.0575 Hsp89 0152 Hs.75410 HSPA5 0.0640 GRP78 1054 Hs.74276 CLIC1 0.0686 Nuclear chloride channel (RNCC protein) 0709 Hs.253495 SFTPD 0.0936 Pulmonary surfactant protein D 0867 Hs.78996 PCNA 0.0982 PCNA 0165 Hs.180414 HSPA8 0.1014 Heat shock cognate protein, 71 kDa 1109 Hs.75103 YWHAZ 0.1018 14-3-3 / 0137 Hs.554 SSA2 0.1032 Ro/ss-A antigen Molecular & Cellular Proteomics 1.4 305 Protein and mRNA Correlation in Lung Adenocarcinomas ABLE I—continued Spot Unigene Gene name r* Protein name 0278 Hs.4112 TCP1 0.1237 T-complex protein I,  subunit 1769 Hs.9614 NPM1 0.1738 B23/numatrin 0089 Hs.74335 HSPCB 0.2049 Hsp90 2511 Hs.153179 FABP5 0.2109 E-FABP/FABP5 1739 Hs.16488 CALR 0.2344 Calreticulin 32 1138 Hs.301961 GSTM4 0.2438 Glutathione S-transferase M4 (GST m4) 2533 Hs.77060 PSMB6 0.2512 Macropain subunit were performed as described previously (12, 13). Our 2D-PAGE sys- non-neoplastic lung tissues. Protein levels were determined tem allows us to run 20 gels at one time (one batch). Spot detection using quantitative 2D-PAGE analysis, and the separated pro- and quantification were accomplished utilizing Bio Image Visage Sys- tein polypeptides were identified using matrix-assisted laser tem software (Bioimage Corp., Ann Arbor, MI). The integrated inten- desorption/ionization mass spectrometry (MALDI-MS). The sity of each spot was calculated as the measured optical density corresponding mRNA levels for the identified proteins within units  mm . Of the total possible 2000 spots detectable on each gel, 820 spots on the gel of each sample were matched using a Gel-ed the same samples were determined using oligonucleotide match program with the same spots on a chosen “master” gel. In microarrays. Correlation analyses showed that protein abun- each sample, 250 ubiquitously expressed reference spots were used dance is likely a reflection of the transcription for a subset of to adjust for variations between gels, such as that created by subtle proteins, but translation and post-translational modifications differences in protein loading or gel staining. Slight differences be- also appear to influence the expression levels of many indi- cause of batch were corrected after spot-size quantification. Mass Spectrometry and 2D Western Blotting—Preparative 2D gels vidual proteins in lung adenocarcinomas. were run using extracts from A549 lung adenocarcinoma cells (ob- tained from ATCC) and using the identical experimental conditions as EXPERIMENTAL PROCEDURES the analytical 2D gels, except 30% more protein was loaded. The Tissues—Fifty-seven stage I and 19 stage III lung adenocarcino- resolved protein gels were silver-stained using successive incuba- mas, as well as nine non-neoplastic lung tissue samples, were used tions in 0.02% sodium thiosulfate for 2 min, 0.1% silver nitrate for 40 for protein and mRNA analyses. Patient consent was obtained, and min, and 0.014% formaldehyde plus 2% sodium carbonate for 10 the project was approved by the Institutional Review Board. All tis- min. For protein identification, protein polypeptides underwent trypsin sues were obtained after resection at the University of Michigan digestion followed by MALDI-MS using a MALDI-TOF Voyager-DE Health System between May 1991 and July 1998. Tissues were all mass spectrometer (Perseptive Biosystems, Framingham, MA). The snap-frozen in liquid nitrogen and then stored at 80 °C. The patients masses were compared with known trypsin digest databases using included 46 females and 30 males ranging in age from 40.9 to 84.6 the MS-FIT database (University of California, San Francisco; (average 63.8) years. Most patients (66/76) demonstrated a positive prospector.ucsf.edu/ucsfhtml3.2/msfit.htm). Some of the polypep- smoking history. Sixty-one tumor samples were classified as bron- tides included in the analysis had been identified prior to this study on chial-derived, 14 were classified as bronchoalveolar, and one had the basis of sequencing (14). The identified protein spots used in this both features. Eighteen tumor samples were classified as well differ- paper are shown in Fig. 1A. The method for 2D-PAGE Western blot entiated, 38 were classified as moderate, and 19 were classified as verification was as described previously (15). The 2D Western blots of poorly differentiated adenocarcinomas. Hematoxylin-stained cryostat GRP58 and Op18 are shown in Fig. 1, C and E; the others, such as sections (5 m), prepared from the same tumor pieces to be utilized GRP78, GRP75, HSP70, HSC70, KRT8, KRT18, KRT19, Vimentin, for protein and mRNA isolation, were evaluated by a pathologist and ApoJ, 14 –3-3, Annexin I, Annexin II, PGP9.5, DJ-1, GST-pi, and compared with hematoxylin- and eosin-stained sections made from 2 PGAM, are described elsewhere. paraffin blocks of the same tumors. Specimens were excluded from Statistical Analysis—Missing values were replaced with the mean analysis if they showed unclear or mixed histology (e.g. adenosqua- value of the protein spot. The transform x 3 log (1  x) was applied mous), tumor cellularity less than 70%, potential metastatic origin as to normalize all protein expression values. The relationship between indicated by previous tumor history, extensive lymphocytic infiltration, protein and mRNA expression levels within the same samples was or fibrosis or if the patient had received prior chemotherapy or examined using the Spearman correlation coefficient analysis (16). To radiotherapy. identify potentially significant correlations between gene and protein Oligonucleotide Array Hybridization—The HuGeneFL oligonucleo- expression, we used an analytical strategy similar to SAM (signifi- tide arrays (Affymetrix, Santa Clara, CA) containing 6800 genes were cance analysis of microarrays) (17), which uses a permutation tech- used in this study. Total RNA was isolated from all samples using nique to determine the significance of changes in gene expression Trizol reagent (Invitrogen). The resulting RNA was then subjected to between different biological states. To obtain permuted correlation further purification using RNeasy spin columns (Qiagen). Preparation coefficients between gene and protein expression, genes were ex- of cRNA, hybridization, and scanning of the HuGeneFL arrays were changed first in such a way that permutated correlation coefficient performed according to the manufacturer’s protocol (Affymetrix, were calculated based on pseudo pairs of genes and proteins. The Santa Clara, CA). Data analysis was performed using GeneChip 4.0 distribution of permutated correlation coefficients became stable after software. The gene expression profile of each tumor was normalized 60 permutations. This procedure was then repeated 60 times to to the median gene expression profile for the entire sample. Details of obtain 60 sets of permutated correlation coefficients. For each of the data trimming and normalization are described elsewhere (11). 60 permutations, the correlations of genes and proteins were ranked 2D-PAGE and Quantitative Protein Analysis—Tissue for both pro- tein and mRNA isolation came from contiguous areas of each sample. Protein separation using 2D-PAGE, silver staining, and digitization Chen et al., submitted for publication. 306 Molecular & Cellular Proteomics 1.4 Protein and mRNA Correlation in Lung Adenocarcinomas TABLE II Correlation coefficients of protein and mRNA where multiple isoforms were present on 2D gels r*, correlation coefficient value  0.2445; p  0.05. Values in boldface are significant at p  0.05. Spot Unigene Gene name r* Protein name 1494 Hs.81915 LAP18 0.4003 OP18 (Stathmin) 0957 Hs.77899 TPM1 0.3930 Tropomyosins 1–5 0353 Hs.289101 GRP58 0.3802 Protease disulfide isomerase (GRP58) 0855 Hs.169476 GAPD 0.3693 Glyceraldehyde-3-phosphate dehydrogenase 1198 Hs.41707 HSPB3 0.3668 Hsp27 1203 Hs.83848 TPI1 0.3395 Triose phosphate isomerase (TPI) 0523 Hs.65114 KRT18 0.3335 Cytokeratin 18 1492 Hs.81915 LAP18 0.3234 OP18 (Stathmin) 1493 Hs.81915 LAP18 0.3154 OP18 (Stathmin) 1181 Hs.78225 ANXA1 0.3102 Annexin variant I 0439 Hs.242463 KRT8 0.3049 Cytokeratin 8 0505 Hs.297753 VIM 0.2939 Vimentin 0593 Hs.297753 VIM 0.2809 Vimentin 1874 Hs.75313 AKR1B1 0.2790 Aldose reductase 0935 Hs.75544 YWHAH 0.2775 14-3-3 2524 Hs.78225 ANXA1 0.2612 Annexin I 2324 Hs.65114 KRT18 0.2601 Cytokeratin 18 1192 Hs.41707 HSPB3 0.2558 Hsp27 0350 Hs.289101 GRP58 0.2516 Phospholipase C (GRP58) 0992 Hs.75313 AKR1B1 0.2460 Aldose reductase 0861 Hs.75313 AKR1B1 0.0761 Aldose reductase 0853 Hs.75313 AKR1B1 0.0675 Aldose reductase 2503 Hs.76392 ALDH1 0.0565 Aldehyde dehydrogenase 0381 Hs.76392 ALDH1 0.0371 Aldehyde dehydrogenase 0371 Hs.76392 ALDH1 0.0680 Aldehyde dehydrogenase 1179 Hs.78225 ANXA1 0.2052 Annexin variant I 0762 Hs.78225 ANXA1 0.0739 Annexin I 0760 Hs.78225 ANXA1 0.0228 Annexin I 2506 Hs.217493 ANXA2 0.2223 Lipocotin (annexin II) 0772 Hs.217493 ANXA2 0.2080 Lipocotin (annexin II) 0723 Hs.217493 ANXA2 0.0701 Lipocotin 1239 Hs.93194 APOA1 0.1133 Apolipoprotein A1 (ApoA1) 1237 Hs.93194 APOA1 0.0373 Apolipoprotein A1 (ApoA1) 1234 Hs.93194 APOA1 0.0894 Apolipoprotein A1 (ApoA1) 0428 Hs.25 ATP5B 0.0080 ATP synthase  subunit precursor 0427 Hs.25 ATP5B 0.0122 ATP synthase  subunit precursor 0424 Hs.25 ATP5B 0.0992 ATP synthase  subunit precursor 0863 Hs.75106 CLU 0.0483 Apolipoprotein J (ApoJ) 0780 Hs.75106 CLU 0.0443 Apolipoprotein J (ApoJ) 1527 Hs.119140 EIF5A 0.0726 eIF-5A 1484 Hs.119140 EIF5A 0.0376 eIF-5A 1728 Hs.5241 FABP1 0.1916 L-FABP 1712 Hs.5241 FABP1 0.0473 L-FABP 0947 Hs.169476 GAPD 0.1745 Glyceraldehyde-3-phosphate dehydrogenase 1232 Hs.75207 GLO1 0.2249 Glyoxalase-I 1229 Hs.75207 GLO1 0.0450 Glyoxalase-1 1595 Hs.158300 HAP1 0.0137 Huntingtin-associated protein 1 (neuroan 1) 1810 Hs.75990 HP 0.4672 -Haptoglobin 1459 Hs.75990 HP 0.0802 -Haptoglobin 1458 Hs.75990 HP 0.0305 -Haptoglobin 0619 Hs.75990 HP 0.0461 B-haptoglobin 0615 Hs.75990 HP 0.0034 B-haptoglobin 1250 Hs.41707 HSPB3 0.1024 Hsp27 0549 Hs.79037 HSPD1 0.1074 Hsp60 0338 Hs.79037 HSPD1 0.2265 Hsp60 0333 Hs.79037 HSPD1 0.1383 Hsp60 0331 Hs.79037 HSPD1 0.1603 Hsp60 2381 Hs.65114 KRT18 0.2016 Cytokeratin 18 0535 Hs.65114 KRT18 0.1106 Cytokeratin 18 Molecular & Cellular Proteomics 1.4 307 Protein and mRNA Correlation in Lung Adenocarcinomas TABLE II—continued Correlation coefficients of protein and mRNA where multiple isoforms were present on 2D gels r*, correlation coefficient value  0.2445; p  0.05. Values in boldface are significant at p  0.05. Spot Unigene Gene name r* Protein name 0529 Hs.65114 KRT18 0.1279 Cytokeratin 18 0528 Hs.65114 KRT18 0.0414 Cytokeratin 18 0527 Hs.65114 KRT18 0.0436 Cytokeratin 18 0514 Hs.65114 KRT18 0.0733 Cytokeratin 18 0451 Hs.242463 KRT8 0.0111 Cytokeratin 8 0446 Hs.242463 KRT8 0.0347 Cytokeratin 8 0444 Hs.242463 KRT8 0.1311 Cytokeratin 8 0443 Hs.242463 KRT8 0.0942 Cytokeratin 8 1488 Hs.81915 LAP18 0.0495 OP18 (Stathmin) 0321 Hs.75655 P4HB 0.0546 PDI (proly-4-OH-B) 0320 Hs.75655 P4HB 0.0041 PDI (proly-4-OH-B) 1063 Hs.75323 PHB 0.0441 Prohibitin 0837 Hs.75323 PHB 0.1402 Prohibitin 0326 Hs.297681 SERPINA1 0.0227 -1-Antitripsin 0322 Hs.297681 SERPINA1 0.0277 -1-Antitripsin 0241 Hs.297681 SERPINA1 0.0148 -1-Antitripsin 1280 Hs.301254 SFTPA1 0.1488 Pulmonary surfactant-associated protein 1278 Hs.301254 SFTPA1 0.2040 Pulmonary surfactant-associated protein 0866 Hs.73980 TNNT1 0.1162 Troponin T 0778 Hs.73980 TNNT1 0.0740 Troponin T 1213 Hs.83848 TPI1 0.0024 Triose phosphate isomerase (TPI) 1210 Hs.83848 TPI1 0.0490 Triose phosphate isomerase (TPI) 1207 Hs.83848 TPI1 0.1615 Triose phosphate isomerase (TPI) 1204 Hs.83848 TPI1 0.0209 Triose phosphate isomerase (TPI) 1202 Hs.83848 TPI1 0.0721 Triose phosphate isomerase (TPI) 1161 Hs.83848 TPI1 0.2265 Triose phosphate isomerase (TPI) 1052 Hs.77899 TPM1 0.1040 Tropomysin clean-product 1039 Hs.77899 TPM1 0.2999 Cytoskeletal tropomyosin 1035 Hs.77899 TPM1 0.3821 Tropomyosin 0783 Hs.77899 TPM1 0.0757 Tropomyosins 1–5 1574 Hs.194366 TTR 0.0065 Transthyretin 0809 Hs.194366 TTR 0.0399 Transthyretin multimere 2202 Hs.76118 UCHL1 0.0220 Ubiquitin carboxyl-terminal hydrolase isozyme L1 1246 Hs.76118 UCHL1 0.1261 Ubiquitin carboxyl-terminal hydrolase isozyme L1 1242 Hs.76118 UCHL1 0.1473 Ubiquitin carboxyl-terminal hydrolase isozyme L1 0606 Hs.297753 VIM 0.0951 Vimentin 0594 Hs.297753 VIM 0.2664 Vimentin-derived protein (vid4) 0508 Hs.297753 VIM 0.1008 Vimentin-derived protein (vid2) 0419 Hs.297753 VIM 0.0032 Vimentin-derived protein (vid1) 1279 Hs.75544 YWHAH 0.0059 14-3-3 such that  (i) denotes the ith largest correlation coefficient for pth p protein spots on 2D gels representing 98 genes and com- permutation. Hence, the expected correlation coefficient,  (i), was the pared protein levels with mRNA levels for a cohort of 85 lung average over the 60 permutations,  (i)  (i)/60. A scatter plot of E p  1 p adenocarcinomas and normal lung samples. Of the 165 pro- observed correlations ((i)) versus the expected correlations is shown in tein spots, 69 proteins were represented by only one known Fig. 2D. For this study, we chose threshold 0.115 so that correlation would be considered significant if absolute value of difference between spot on 2D gels for an individual gene, whereas 96 protein (i) and  (i) was greater than the threshold. Twenty-nine (including one E spots showed multiple protein products from 29 different with observed correlation coefficient 0.4672) of 165 pairs of gene and genes. 2D Western blotting verified the proteins identified by protein expression were called significant in such criteria, and the mass spectrometry when specific antibodies were available. permuted data generated an average of 5.1 falsely significant pairs of Spearman correlation coefficients of the proteins and their gene and protein expression. This provided an estimated false dis- covery rate (the percentage of pairs of gene and protein expression associated mRNA for each protein spot were generated using identified by chance) for our data set. all 76 lung adenocarcinomas and nine non-neoplastic lung tissues (see Tables I and II, and see Figs. 1 and 2). The RESULTS correlation coefficients (r) ranged from 0.467 to 0.442 (Fig. Correlation of Individual Proteins and mRNA Expression 2D). A total of 28 protein spots (21 genes) were found to have within Each Tumor—We have examined quantitatively 165 a statistically significant correlation between expression of 308 Molecular & Cellular Proteomics 1.4 Protein and mRNA Correlation in Lung Adenocarcinomas FIG.1. A, digital image of a silver-stained 2D-PAGE separation of a stage I lung adenocarcinoma showing protein spots separated by molecular mass (MW) and isoelectric point (PI). Twenty-eight protein spots whose expression levels are correlated with mRNA abundance are indicated by the black arrows. B, the outlined areas of A showing protein GRP58. C, 2D Western blot of GRP58 from the A549 lung adenocarcinoma cell line. D, the outlined areas of A showing the protein isoforms of Op18. E, 2D Western blot of Op18 from A549 cells. their protein and mRNA (r  0.2445; p  0.05). This accounts tween protein and mRNA expression (r  0.0495). Similarly, just for 17% (28/165) of the 165 protein spots. Among the 69 one of five quantified isoforms of cytokeratin 8 (spot 439) dem- genes for which only a single protein spot was known (Table onstrated a statistically significant correlation between protein I), nine genes (9/69, 13%) were observed to show a statisti- and mRNA abundance (r  0.3049; p  0.05) (Table II). cally significant relationship between protein and mRNA In addition to differences in the relationship between mRNA abundance (r  0.2445; p  0.05). The proteins whose ex- levels and protein expression among separate isoforms, some pression levels were correlated with their mRNA abundance genes with very comparable mRNA levels showed a 24-fold included those involved in signal transduction, carbohydrate difference in their protein expression. Genes with comparable metabolism, apoptosis, protein post-translational modifica- protein expression levels also showed up to a 28-fold vari- tion, structural proteins, and heat shock proteins (Table III). ance in their mRNA levels. Individual Isoforms of the Same Protein Have Different Lack of Correlation for mRNA and Protein Expression when Protein/mRNA Correlation Coefficients—Of the 165 protein Using Average Tumor Values across All 165 Protein Spots (98 spots, 96 represent protein products of 29 genes with at least Genes)—The relationship between mRNA and protein expres- two isoforms. Among these 96 protein spots, 19 (19/96 pro- sion was also examined by using the average expression tein spots, 20%) showed a statistically significant correlation values for all samples. To analyze this relationship using this between their protein and mRNA expression (r  0.2445; p approach, the average value for each protein or mRNA was 0.05) (Table II) and represented 12 genes (12/29, 41%). Individ- generated using all 85 lung tissue samples. The range of ual isoforms of the same protein demonstrated different normalized average protein values ranged from 0.0646 to protein/mRNA correlation coefficients. For example, 2D-PAGE/ 0.0979 (raw value 0.0036 to 4.1947), and the range for mRNA Western analysis revealed four isoforms of OP18 differing in was from 0 to 15260.5 for all 165 individual protein spots. The regards to isoelectric point but similar in molecular weight. Spearman correlation coefficient for the whole data set (165 Three of the four isoforms (spots 1492, 1493, and 1494) showed protein spots/98 genes) was 0.025 (Fig. 3A). Even for the 28 a statistically significant correlation between their protein and mRNA abundance (r  0.3234, 0.3154, and 0.4003, respective- protein spots (Fig. 2D) that were found to have a statistically ly). The forth isoform (spot 1488) showed no correlation be- significant correlation between their mRNA and protein, use of Molecular & Cellular Proteomics 1.4 309 Protein and mRNA Correlation in Lung Adenocarcinomas FIG.2. A–C, plots showing the correlation between mRNA and protein for the three selected genes Op18, Annexin IV, and GAPD for all 76 lung adenocarcinomas and nine non-neoplastic lung samples (p  0.05). D, distribution of all 165 Spearman correlation coefficients (r) and verification analysis using SAM. A more detailed description of the method is provided under “Experimental Procedures.” Approximately 17% of the 165 proteins demonstrate a significant correlation between mRNA and protein levels as demonstrated by the values shown beyond the outer range of threshold  0.115. Normalized protein values were used, thus negative values for some proteins are observed. the average value resulted in a correlation coefficient value of stage III (n  19) lung adenocarcinomas (Table III). The num- 0.035, which was not significant (Fig. 3B). ber of non-neoplastic lung samples (n  9) was insufficient for Lack of a Relationship between Protein/mRNA Correlation a separate correlation analysis of this group. Many of the Coefficients and Average Protein Abundance—To determine protein spots represent one of several known protein isoforms whether an absolute protein level might influence the corre- for a given gene. The majority of genes (16/21) did not differ in lation with mRNA, the mean value of each protein (relative the protein/mRNA correlation between stage I and stage III abundance) and the Spearman protein/mRNA correlation co- tumors indicating a similar regulatory relationship between the efficients among all 85 samples were examined. No relation- mRNA and protein spot. GRP-58, PSMC, SOD1, TPI1, and ship between the protein abundance and the correlation co- VIM, however, were found to demonstrate significant differ- efficients was observed (r  0.039; p  0.05). A detailed ences in the correlation coefficients between stage I and analysis of separate subsets of proteins with differing levels of stage III lung adenocarcinomas. For GRP-58, PSMC, and VIM abundance (less than 0.0014, larger than 0.0014, or larger the change in the correlation coefficient was because of a than 0.0077) also showed a lack of correlation between mRNA relative increase in protein expression in stage III tumors. For and protein expression among the 83 (50%), 82 (50%), and 41 SOD and TPI the change resulted from a relative decrease in (25%) of 165 total protein spots, respectively (r  0.016, 0.08, expression of this specific protein in stage III tumors. and 0.172, respectively). Stage-related Changes in the Protein/mRNA Correlation DISCUSSION Coefficients—To determine whether the 21 genes (28 protein spots) showing a significant correlation between the protein Relatively little is known about the regulatory mechanisms and mRNA expression among all samples demonstrate controlling the complex patterns of protein abundance and changes in this relationship during tumor progression, the post-translational modification in tumors. Most reports con- correlations were examined separately for stage I (n  57) and cerning the regulation of protein translation have focused on 310 Molecular & Cellular Proteomics 1.4 Protein and mRNA Correlation in Lung Adenocarcinomas TABLE III Stage-dependent analysis of protein-mRNA correlation coefficients r, correlation coefficient. Values in boldface indicate a significant difference between stage I and stage III. Spot Gene name r (Stage I) r (Stage III) Function 1874 AKR1B1 0.269 0.106 Carbohydrate metabolism; electron transporter 2524 ANXA1 0.184 0.572 Phospholipase inhibitor; signal transduction 0994 ANXA4 0.660 0.362 Phospholipase inhibitor 0963 ANXA5 0.241 0.390 Phospholipase inhibitor; calcium binding; phospholipid binding 1314 DJ-1 0.363 0.354 Signal transduction 1405 FTL 0.126 0.358 Iron storage protein 0855 GAPD 0.243 0.581 Carbohydrate metabolism (glycolysis regulation) 0350 GRP58 0.327 0.087 Signal transduction; protein disulfide isomerase 0264 HNRPK 0.360 0.243 RNA-binding protein (RNA processing/modification) 1192 HSPB3 0.457 0.633 Heat shock protein 0523 KRT18 0.115 0.371 Structural protein 0439 KRT8 0.323 0.436 Structural protein 1492 LAP18 0.483 0.663 Signal transduction; cell growth and maintenance 1638 LGALS1 0.200 0.528 Apoptosis; cell adhesion; cell size control 1252 PSMC 0.253 0.060 Protein degradation 1104 SFN 0.465 0.475 Signal transduction (protein kinase C inhibitor) 1454 SOD1 0.352 0.079 Oxidoreductase 1203 TPI1 0.378 0.009 Carbohydrate metabolism 0957 TPM1 0.475 0.225 Structural protein (muscle); control of heart 0593 VIM 0.054 0.556 Structural protein 0935 YWHAH 0.283 0.210 Signal transduction one or several protein products (18). Celis et al. (19) found a anisms of protein translation among different cells or as good correlation between transcript and protein levels among measured in different laboratories (23). 40 well resolved, abundant proteins using a proteomic and In this study, we examined 165 protein spots identified in microarray study of bladder cancer. By comparing the mRNA lung adenocarcinomas. Ninety-six protein spots, representing and protein expression levels within the same tumor samples, the products of 29 genes, contained at least two protein we found that 17% (28/165) of the protein spots (21/98 genes) isoforms. Nineteen of 96 protein spots, representing 12 show a statistically significant correlation between mRNA and genes, were shown to have a statistically significant correla- protein. These proteins appear to represent a diverse group of tion between their protein and mRNA expression, suggesting gene products and include those involved in signal transduc- that the levels of these proteins reflects the transcription of the tion, carbohydrate metabolism, protein modification, cell struc- corresponding genes. Differences in protein/mRNA correlations ture, heat shock, and apoptosis. These results suggest that were found among the individual isoforms of a given protein. For expression of this subset of 165 proteins is likely to be regulated example, of the four OP18 isoforms, three showed a statistically at the transcriptional level in these tissues. The majority of the significant correlation between the protein and mRNA expres- protein isoforms, however, did not correlate with mRNA levels, sion levels. The lack of relationship for the one isoform, how- and thus their expression is regulated by other mechanisms. We ever, indicates that individual protein isoforms of the same gene also observed a subset of proteins that demonstrated a nega- product can be regulated differentially. This is not unexpected tive correlation with the mRNA expression values; for example and likely reflects other post-translational mechanisms that can -haptoglobin demonstrated a strong negative correlation with influence isoform abundance in tissues and cancer. its mRNA expression values. This may reflect negative feedback In addition to the analyses of the correlation of mRNA/ on the mRNA or the protein or the presence of other regulatory protein within the same tumor samples, we also tested the influences that are not understood currently. global relationship between mRNA and the corresponding Post-translational modification or processing will result in protein abundance across all 165 protein spots in the lung individual protein products of the same gene migrating to samples. A protein and mRNA average value for each gene different locations on 2D-PAGE gels (20). Because the identity was generated using all 85 lung tissues samples. We ob- of all possible isoforms for each protein examined has not served a very wide range of normalized average protein and been characterized completely, this may influence the corre- mRNA values. The correlation coefficient generated using this lation analyses performed in this study. This is partly because average value data set was 0.025, and even for the 28 of limitations of the 2D-PAGE and mass spectrometry tech- protein spots that showed a statistically significant correlation nologies (21, 22). Potential inconsistencies between mRNA between individual mRNA and proteins, the correlation value and protein correlations that have been reported may also be was only 0.035. This suggests that it is not possible to because of differences, even in the same gene, in the mech- predict overall protein expression levels based on average Molecular & Cellular Proteomics 1.4 311 Protein and mRNA Correlation in Lung Adenocarcinomas FIG.3. The overall correlation of mRNA and protein levels across all 165 protein spots (A) and across 28 protein spots that contained individ- ual r values larger than 0.244 (B) are shown. Each protein or mRNA mean value was calculated based on all 76 lung adenocarcinomas and nine non- neoplastic lung samples using quantita- tive 2D-PAGE and Affymetrix oligonu- cleotide microarrays. The Spearman correlation coefficients for the two data sets (A and B) were 0.025 and 0.035, respectively, indicating a lack of correla- tion if mean values for mRNA and protein for all samples is used. mRNA abundance in lung cancer samples. This conclusion is A comparison between the mean value of each protein and also supported by previous results from Anderson and Seil- the correlation coefficient generated using all 85 tissue sam- hamer (24), who examined 19 genes in human liver cells, and ples did not reveal a strong relationship between the overall by Gygi et al. (25), who examined 106 genes in yeast. Both protein abundance and the correlation coefficients (r  0.039; studies found a lack of correlation between mRNA and protein p  0.05). Detailed analysis of different subsets of protein abun- expression when average or overall levels were used. dance also failed to show a correlation between mRNA and A good correlation was reported when the 11 most abun- protein expression. Thus in contrast to yeast, a relationship dant proteins were examined in yeast (25), suggesting that the between mRNA/protein correlation coefficient and protein level of protein abundance may be a factor that may influence abundance in human lung adenocarcinomas was not observed. the correlation between mRNA and protein. In the present The results of this study indicate that the level of protein study, a fairly wide range of mean protein values among 165 abundance in lung adenocarcinomas is associated with the protein spots in lung adenocarcinomas was observed, and corresponding levels of mRNA in 17% (28 proteins) of the the correlation coefficients also varied from 0.467 to 0.442. total 165 protein spots examined. This was substantially 312 Molecular & Cellular Proteomics 1.4 Protein and mRNA Correlation in Lung Adenocarcinomas higher than the amount predicted to result by chance alone 8. Wang, D. G., Fan, J. B., Siao, C. J., Berno, A., Young, P., Sapolsky, R., Ghandour, G., Perkins, N., Winchester, E., Spencer, J., Kruglyak, L., (which was 5.1) and suggests that a transcriptional mecha- Stein, L., Hsie, L., Topaloglou, T., Hubbell, E., Robinson, E., Mittmann, nism likely underlies the abundance of these proteins in lung M., Morris, M. S., Shen, N., Kilburn, D., Rioux, J., Nusbaum, C., Rozen, adenocarcinomas. We also demonstrate that the expression S., Hudson, T. J., Lander, E. S. (1998) Large-scale identification, map- ping, and genotyping of single-nucleotide polymorphisms in the human of individual isoforms of the same protein may or may not genome. Science 280, 1077–1082 correlate with the mRNA, indicating that separate and likely 9. Pease, A. C., Solas, D., Sullivan, E. J., Cronin, M. T., Holmes, C. P., Fodor, post-translational mechanisms account for the regulation of S. P. (1994) Light-generated oligonucleotide arrays for rapid DNA se- quence analysis. Proc. Natl. Acad. Sci. U. S. A. 91, 5022–5026 isoform abundance. These mechanisms may also account for 10. Bhattacharjee, A., Richards, W. G., Staunton, J., Li, C., Monti, S., Vasa, P., the differences in the correlation coefficients observed between Ladd, C., Beheshti, J., Bueno, R., Gillette, M., Loda, M., Weber, G., Mark, stage I and stage III tumors, indicating that specific protein E. J., Lander, E. S., Wong, W., Johnson, B. E., Golub, T. R., Sugarbaker, D. J., Meyerson, M. (2001) Classification of human lung carcinomas by isoforms show regulatory changes during tumor progression. mRNA expression profiling reveals distinct adenocarcinoma subclasses. Further studies in lung adenocarcinomas will examine the rela- Proc. Natl. Acad. Sci. U. S. A. 98, 13790 –13795 tionship between the expression of individual protein isoforms 11. Giordano, T. J., Shedden, K. A., Schwartz, D. R., Kuick, R., Taylor, J. M. G., and specific clinical-pathological features of these tumors, such Lee, N., Misek, D. E., Greenson, J. K., Kardia, S. L. R., Beer, D. G., Rennert, G., Cho, K. R., Gruber, S. B., Fearon, E. R., Hanash, S. (2001) as the presence of angiolymphatic invasion, and nodal or pleu- Organ-specific molecular classification of lung, colon and ovarian adeno- ral surface involvement. The potential to identify specific protein carcinomas using gene expression profiles. Am. J. Pathol. 159, 1231–1238 isoforms associated with biological behavior in lung adenocar- 12. Strahler, J. R., Kuick, R., Hanash, S. M. (1989) in Protein Structure: A Practical Approach (Creighton, T., ed) pp. 65–92, IRL Press, Oxford cinomas would be of considerable interest and will add to our 13. Merril, C. R., Dunau, M. L., Goldman, D. (1981) A rapid sensitive silver stain understanding of the regulation of gene products by transcrip- for polypeptides in polyacrylamide gels. Anal. Biochem. 101, 201–207 tional, translational, and post-translational mechanisms. 14. Hanash, S. M., Strahler, J. R., Chan, Y., Kuick, R., Teichroew, D., Neel, J. V., Hailat, N., Keim, D. R., Gratiot-Deans, J., Ungar, D., Richardson, B. C. Acknowledgments—We thank Kerby A. Shedden, Rork D. Kuick, (1993) Data base analysis of protein expression patterns during T-cell Eric Puravs, Robert Hinderer, Melissa C. Krause, and Christopher ontogeny and activation. Proc. Natl. Acad. Sci. U. S. A. 90, 3314 –3318 15. Brichory, F. M., Misek, D. E., Yim, A. M., Krause, M. C., Giordano, T. J., Wood for assistance in this study. Beer, D. G., Hanash, S. M. (2001) An immune response manifested by the common occurrence of annexins I and II autoantibodies and high * This work was supported by NCI, National Institutes of Health circulating levels of IL-6 in lung cancer. Proc. Natl. Acad. Sci. U. S. A. 98, Grant U19 CA-85953. The costs of publication of this article were 9824 –9829 defrayed in part by the payment of page charges. This article must 16. Lavens-Phillips, S. E., MacGlashan, D. W., Jr. (2000) The tyrosine kinases therefore be hereby marked “advertisement” in accordance with 18 p53/56lyn and p72syk are differentially expressed at the protein level but U.S.C. Section 1734 solely to indicate this fact. not at the messenger RNA level in nonreleasing human basophils. Am. J. ‡‡ To whom correspondence should be addressed: General Tho- Respir. Cell Mol. Biol. 23, 566 –571 racic Surgery, SRB II, B560, Box 0686, University of Michigan Medical 17. Tusher, V. G., Tibshirani, R., Chu, G. (2001) Significance analysis of mi- School, Ann Arbor, MI 48109-086; E-mail: [email protected]. croarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. U. S. A. 98, 5116 –5121 REFERENCES 18. Tew, K. D., Monks, A., Barone, L., Rosser, D., Akerman, G., Montali, J. A., Wheatley, J. B., Schmidt, D. E., Jr. (1996) Glutathione-associated en- 1. Ideker, T., Thorsson, V., Ranish, J. A., Christmas, R., Buhler, J., Eng, J. K., zymes in the human cell lines of the National Cancer Institute Drug Bumgarner, R., Goodlett, D. R., Aebersold, R., Hood, L. (2001) Integrated Screening Program. Mol. Pharmacol. 50, 149 –159 genomic and proteomic analyses of a systematically perturbed meta- 19. Celis, J. E., Kruhoffer, M., Gromova, I., Frederiksen, C., Ostergaard, M., bolic network. Science 292, 929 –934 Thykjaer, T., Gromov, P., Yu, J., Palsdottir, H., Magnusson, N., Orntoft, 2. Liang, P., Pardee, A. B. (1998) Differential display. A general protocol. Mol. T. F. (2000) Gene expression profiling: monitoring transcription and Biotechnol. 10, 261–267 translation products using DNA microarrays and proteomics. FEBS Lett. 3. Porter, D. A., Krop, I. E., Nasser, S., Sgroi, D., Kaelin, C. M., Marks, J. R., 480, 2–16 Riggins, G., Polyak, K. (2001) A sage (serial analysis of gene expression) 20. Anderson, N. L., Anderson, N. G. (1998) Proteome and proteomics: new view of breast tumor progression. Cancer Res. 61, 5697–5702 technologies, new concepts, and new words. Electrophoresis 19, 4. Bittner, M., Meltzer, P., Chen, Y., Jiang, Y., Seftor, E., Hendrix, M., 1853–1861 Radmacher, M., Simon, R., Yakhini, Z., Ben-Dor, A., Sampas, N., 21. Gygi, S. P., Corthals, G. L., Zhang, Y., Rochon, Y., Aebersold, R. (2000) Dougherty, E., Wang, E., Marincola, F., Gooden, C., Lueders, J., Glat- Evaluation of two-dimensional gel electrophoresis-based proteome anal- felter, A., Pollock, P., Carpten, J., Gillanders, E., Leja, D., Dietrich, K., ysis technology. Proc. Natl. Acad. Sci. U. S. A. 97, 9390 –9395 Beaudry, C., Berens, M., Alberts, D., Sondak, V. (2000) Molecular 22. Fey, S. J., Larsen, P. M. (2001) 2D or not 2D. Two-dimensional gel elec- classification of cutaneous malignant melanoma by gene expression trophoresis. Curr. Opin. Chem. Biol. 5, 26 –33 profiling. Nature 406, 536 –540 23. McBride, S., Walsh, D., Meleady, P., Daly, N., Clynes, M. (1999) Bromode- 5. Fung, E. T., Wright, G. L., Jr., Dalmasso, E. A. (2000) Proteomic strategies oxyuridine induces keratin protein synthesis at a posttranscriptional level for biomarker identification: progress and challenges. Curr. Opin. Mol. in human lung tumor cell lines. Differentiation 64, 185–193 Ther. 2, 643– 650 24. Anderson, L., Seilhamer, J. (1997) A comparison of selected mRNA and 6. Davidson, D., Baldock, R. (2001) Bioinformatics beyond sequence: map- ping gene function in the embryo. Nat. Rev. Genet. 2, 409 – 417 protein abundances in human liver. Electrophoresis 18, 533–537 7. Chee, M., Yang, R., Hubbell, E., Berno, A., Huang, X. C., Stern, D., Winkler. 25. Gygi, S. P., Rochon, Y., Franza, B. R., Aebersold, R. (1999) Correlation J., Lockhart, D. J., Morris, M. S., Fodor, S. P. (1996) Accessing genetic between protein and mRNA abundance in yeast. Mol. Cell. Biol. 19, information with high-density DNA arrays. Science 274, 610 – 614 1720 –1730 Molecular & Cellular Proteomics 1.4 313

Journal

Molecular & Cellular Proteomics – Unpaywall

Published: Apr 1, 2002

There are no references for this article.