Alterations in the Renal Elastin-Elastase System in Type 1 Diabetic Nephropathy Identified by Proteomic Analysis

J Am Soc Nephrol 15: 650–662, 2004 Alterations in the Renal Elastin-Elastase System in Type 1 Diabetic Nephropathy Identified by Proteomic Analysis

VISITH THONGBOONKERD,*1 MICHELLE T. BARATI,* KENNETH R. MCLEISH,*†ʈ CHARAF BENARAFA,¶ EILEEN REMOLD-O’DONNELL,¶ SHIRONG ZHENG,‡ BRAD H. ROVIN,# WILLIAM M. PIERCE,§ PAUL N. EPSTEIN,‡§ and JON B. KLEIN*†ʈ *Core Proteomics Laboratory, Kidney Disease Program, Department of Medicine, and Departments of †Biochemistry and Molecular Biology, ‡Pediatrics, and §Pharmacology and Toxicology, University of Louisville, Louisville, Kentucky; ʈVeterans Affairs Medical Center, Louisville, Kentucky; ¶Center for Blood Research and Department of Pediatrics, Harvard Medical School, Boston, Massachusetts; and #Department of Medicine, Ohio State University School of Medicine, Columbus, Ohio

Abstract. Diabetes now accounts for Ͼ40% of patients with elastase IIIB was decreased, leading to the hypothesis that ESRD. Despite significant progress in understanding diabetic elastin expression would be increased in diabetic kidneys. nephropathy, the cellular mechanisms that lead to diabetes- Renal immunohistochemistry for elastin of 325-d-old FVB and induced renal damage are incompletely defined. For defining OVE26 mice demonstrated marked accumulation of elastin in changes in protein expression that accompany diabetic ne- the macula densa, collecting ducts, and pelvicalyceal epithelia phropathy, the renal proteome of 120-d-old OVE26 transgenic of diabetic kidneys. Elastin immunohistochemistry of human mice with hypoinsulinemia, hyperglycemia, hyperlipidemia, renal biopsies from patients with type 1 diabetes (n ϭ 3) and proteinuria were compared with those of background FVB showed increased elastin expression in renal tubular cells and nondiabetic mice (n ϭ 5). Proteins derived from whole-kidney the interstitium but not glomeruli. These results suggest that lysate were separated by two-dimensional PAGE and identified coordinated changes in elastase inhibitor and elastase expres- by matrix-assisted laser desorption ionization–time-of-flight sion result in increased tubulointerstitial deposition of elastin in (MALDI-TOF) mass spectrometry. Forty-one proteins from diabetic nephropathy. The identification of these coordinated 300 visualized protein spots were differentially expressed in changes in protein expression in diabetic nephropathy indicates diabetic kidneys. Among these altered proteins, expression of the potential value of proteomic analysis in defining monocyte/neutrophil elastase inhibitor was increased, whereas pathophysiology.

Diabetes now accounts for Ͼ40% of patients with ESRD, and diabetic nephropathy, the cellular mechanisms that lead to the number of renal failure patients with diabetes is expected to diabetes-induced renal damage are incompletely defined. increase in the coming years (1). Renal pathologic changes that Recently, Clarkson et al. (6) demonstrated that at least 200 lead to decreased renal function are observed in all intrarenal genes were differentially expressed in mesangial cells after structures, including glomeruli, tubulointerstitium, and blood exposure to high-glucose media. These findings indicate the vessels (2, 3). These morphologic changes, coupled with ele- complexity of the development of diabetic nephropathy. How- vated intraglomerular pressure and hormonal dysregulation, ever, proteins, not genes, govern cellular functions. The study lead to glomerulosclerosis, interstitial fibrosis, and ultimately of changes in renal protein expression is necessary to under- renal failure (4, 5). Despite recent progress in understanding stand better the complex pathogenic mechanisms of diabetic nephropathy. Conventional protein studies—Western blotting and other immunologic methods—are limited to a relatively small number of proteins that can be studied in each experiment and to Received August 1, 2003. Accepted December 6, 2003. 1 Current address: Proteomics Center, Medical Molecular Biology Unit, Office previously identified proteins for which specific antibodies are for Research and Development, Faculty of Medicine at Siriraj Hospital, Ma- available. Proteomic analysis is an innovative approach that over- hidol University, Bangkok, Thailand. comes the limitations of immunology-based protein analyses. We Correspondence to Dr. Visith Thongboonkerd, Proteomics Center, Medical used proteomic analysis in the present study to evaluate global th Molecular Biology Unit, Office for Research and Development, 12 Floor— changes of renal protein expression in diabetic kidneys. The Adulyadej Vikrom Building, Siriraj Hospital, Prannok Road, Bangkoknoi, Bangkok 10700, Thailand. Phone: 66-2-4184793; Fax: 66-2-4184793; E-mail: diabetic animal model used in this study was the OVE26 trans- [email protected] genic mouse model. OVE26 mice at 120 d of age displayed many 1046-6673/1503-0650 characteristics of early-onset type 1 diabetic nephropathy, includ- Journal of the American Society of Nephrology ing hyperglycemia, hypoinsulinemia, hyperlipidemia, mesangial Copyright © 2004 by the American Society of Nephrology expansion, and thickening of glomerular basement membrane DOI: 10.1097/01.ASN.0000115334.65095.9B (GBM) (7, 8). Because a protein database for mouse kidney was J Am Soc Nephrol 15: 650–662, 2004 Proteomics and Diabetic Nephropathy 651

not available, we created an initial renal proteome map for FVB cusing using 100 mM sodium hydroxide as the cathode buffer and 10 nondiabetic mice (the background strain of the OVE26 line) and mM phosphoric acid as the anode buffer. Precast carrier ampholyte used this map as a reference to analyze protein expression in tube gels (pH 3 to 10), 1 mm ϫ 18 cm, were prefocused with maximal ␮ ␮ diabetic kidneys. 1500 V and 110 A per tube. Protein samples containing 100 g from Comparison of protein expression in kidneys from OVE26 individual animals were loaded into individual tube gels and were focused for 17 h and 30 min to reach 18,000 volt-hours. and FVB mice by two-dimensional (2-D) PAGE demonstrated significant differences in expression levels of eight groups of proteins: proteases, protease inhibitors, apoptosis-associated Second Dimension of 2-D PAGE proteins, regulators for oxidative tolerance, calcium-binding The gels were extruded from the tubes after completion of focusing and were incubated in premixed Tris/acetate equilibration buffer with proteins, transport regulators, cell signaling proteins, and 0.01% bromophenol blue and 50 mM DTT for 2 min before loading smooth muscle contractile elements. Our results showed coor- onto precast 10% homogeneous, 22 ϫ 22-cm, slab gels (Genomic dinated changes in expression of monocyte/neutrophil elastase Solutions). The upper running buffer contained 0.2 M Tris base, 0.2 M inhibitor (MNEI), which was increased, and elastase IIIB, tricine, and 0.4% SDS, and the lower running buffer contained 0.625 which was decreased. These findings suggested the hypothesis M Tris/acetate. Protein separation was performed with a maximum of that elastin, an extracellular matrix (ECM) protein, accumu- 500 V and 20,000 mW per gel. lates in diabetic kidneys and may participate in the development of diabetic nephropathy. This hypothesis was supported SYPRO Ruby Staining and Visualization by immunohistochemical studies in the OVE26 diabetic mice The gel slabs were fixed in 10% methanol and 7% acetic acid for and patients with type 1 diabetes demonstrating increased 30 min. The fixed solution was removed, and 500 ml of SYPRO Ruby elastin deposition in renal tubular cells. gel stain (Bio-Rad Laboratories) was added to each gel and incubated on gently continuous rocker at room temperature for 18 h. A high- Materials and Methods resolution 12-bit camera with ultraviolet light box system (Genomic The initial production, characterization, and maintenance of the Solutions) was used to visualize the gel images. diabetic OVE26 line were performed at the University of Louisville as described previously (7, 8). Control mice were nontransgenic animals Quantitative Analysis of Protein Expression from the same strain (FVB). Ten animals (five in each group) were Investigator HT analyzer (Genomic Solutions) software was used studied. All animal studies were approved by the University of Lou- for matching and analysis of protein spots. The principles of mea- isville Institutional Animal Care and Use Committee and were in surement of intensity value by 2-D analysis software were similar to accordance with NIH Guide for the Care and Use of Laboratory those of densitometric measurement. Average mode of background Animals. subtraction was used to normalize intensity value that represents the amount of protein per spot. The normalized intensity values of indi- Urine Albumin Assay vidual protein spots were then used to determine differential protein Mice (120 d old) were housed in metabolic cages (Nalgene, Brain- expression between groups by statistical analyses. tree, MA) with free access to solid laboratory food and feeding water. For obtaining adequate urine volume, the feeding water contained 10% (vol/vol) Glucerna liquid diet (Abbotts Laboratories, Columbus, Statistical Analyses OH). Urinary albumin excretion on 24-h collections was measured by After completion of spot matching, spot intensities of each protein a commercial ELISA kit using goat anti-mouse albumin antibody spot from individual animals were compared between control and (Bethyl Laboratories Inc., Montgomery, TX). diabetic groups. Because the sample size was relatively small, both unpaired t test and Mann-Whitney U test by SPSS software v. 10.0 Extraction of Renal Proteins for 2-D PAGE were used for statistical analyses to avoid spurious results. P Ͻ 0.05 was considered statistically significant. Only significant differences Mice were killed at 120 d of age by injection with Ketamine that were in agreement between the t test and the Mann-Whitney HCl/Xylazine HCl solution (Sigma Chemical Co., St. Louis, MO). U test were included, and the data were reported as mean Ϯ SEM. Protein extraction of the whole kidney was performed as described previously (9, 10). Kidneys were frozen in liquid nitrogen; ground to powder; resuspended in a buffer containing 50 mM Tris, 0.3% SDS, In-Gel Tryptic Digestion, MALDI-TOF Mass and 200 mM DTT; and incubated at 100°C for 5 min. DNA and RNA Spectrometry, and Peptide Mass Fingerprinting

were removed by a buffer containing 500 mM Tris, 50 mM MgCl2, 1 In-gel tryptic digestion and matrix-assisted laser desorption ioniza- mg/ml DNAse I, and 0.25 mg/ml RNAse A. Excess salts were tion–time-of-flight (MALDI-TOF) mass spectrometry (MS) were per- removed by acetone precipitation, and the protein pellet was finally formed using techniques described previously by our laboratory (11). resuspended in a buffer containing 40 mM Tris, 7.92 M urea, 0.06% Protein identification of peptide fragments was performed by us- SDS, 1.76% ampholytes, 120 mM DTT, and 3.2% Triton X-100. ing the “ProFound” search engine (129.85.19.192/profound_bin/ Protein concentration levels were measured by spectrophotometry WebProFound.exe). The National Center for Biotechnology Informa- using HP 8453 UV-visible system (Hewlett-Packard Company, Palo tion (NCBI) protein database was restricted to mammalian entries, and Alto, CA) and Bio-Rad Protein Assay (Bio-Rad Laboratories, Her- peptides were assumed to be monoisotopic, oxidized at methionine cules, CA). residues, and carbamidomethylated at cysteine residues. Up to one missed trypsin cleavage was allowed, although most matches did not First Dimension of 2-D PAGE contain any missed cleavages. A mass tolerance error of 150 ppm was A mobile ampholyte tube gel running system (Genomic Solutions allowed for matching peptide mass values. Z scores were estimated by Inc., Ann Arbor, MI) was used for first-dimensional isoelectric fo- comparison of search results against estimated random-match popu- 652 Journal of the American Society of Nephrology J Am Soc Nephrol 15: 650–662, 2004 lation and were the distances to the population mean in units of SD. insulinemia, and hypertriglyceridemia. Hyperglycemia in Scores Ͼ1.65 were considered statistically significant (P Ͻ 0.05). OVE26 mice occurred within 1 wk after birth, and random Identities of protein spots that did not reach this significant level were plasma glucose levels were Ͼ600 mg/dl from 10 wk of age. not reported. Plasma insulin levels of adult OVE26 mice were approximately 30% of normal values as a result of the ␤ cell–specific, Bioinformatic Analyses calmodulin transgene (8). The mice typically survived without To examine potential protein function and sequence homology, we insulin therapy or any other treatment for at least 1 yr. Mild performed bioinformatic analysis using public protein databases. In- mesangial expansion and GBM thickening were observed in ferred protein functions were determined by using data in the NCBI 120-d-old OVE26 diabetic mice without azotemia. For further (www.ncbi.nlm.nih.gov/), Swiss-Prot, and TrEMBL protein databases characterizing the renal phenotype of the OVE26 line, 24-h (ca.expasy.org/sprot/). The similarity of amino acid sequences was urinary albumin excretion was measured. OVE26 diabetic determined using the protein BLAST format (BLink) of the NCBI protein database, as well as standard pairwise protein BLASTp mice at 120 d of age had significantly increased 24-h urinary Ϯ searches. albumin excretion compared with the FVB controls (743 461 versus 29 Ϯ 17 ␮g/24 h; P Ͻ 0.05). When the mice were Western Blotting killed at approximately 17 wk (120 d), OVE26 mice had been diabetic for approximately 16 wk. Renal proteins were mixed with 2ϫ Laemmli sample buffer and boiled for 5 min, and 30 ␮g of total proteins were loaded on 10% SDS-PAGE for MNEI Western blot. Proteins were transferred to a nitrocellulose membrane and blocked with 5% milk/TTBS. The mem- Proteome Map of Normal FVB Mouse Kidney brane was treated with rabbit polyclonal anti-MNEI serum (12) (1: As a database of mouse kidney proteins was not available, a 1000 in 0.1% milk/PBS-Tween) at room temperature for 90 min. This proteome map for FVB mouse kidney was produced. Renal serum cross-reacts with recombinant mouse elastase inhibitor A proteins from individual animals (n ϭ 5) were separated by (EIA), the ortholog of MNEI (data not shown). Immunoreactive 2-D PAGE. The protein spot pattern was reproducible from protein was detected by autoradiography using horseradish peroxida- each animal. Up to 300 protein spots were visualized on each se–conjugated antibody and chemiluminescent substrate. 2-D gel by SYPRO Ruby staining. Of these visualized protein spots, 150 spots were excised and subjected to MALDI-TOF Immunohistochemistry MS. The remaining spots were not excised, as their expression Immunohistochemistry was performed on kidneys from 120- and was likely below the threshold of detectability by MALDI- 325-d-old FVB and OVE26 mice (n ϭ 2) and on human renal biopsies TOF MS. A total of 92 proteins were identified in our initial from patients with type 1 diabetes compared with normal biopsies, mouse kidney proteome map (Figure 1A). Positions of all of ϭ which were from candidate donors for renal transplantation (n 3). these identified proteins on 2-D gels were in the expected range Five-micrometer-thick mouse kidney sections were deparaffinized of their theoretical isoelectric points (pI) and molecular and rehydrated. Antigen retrieval was performed by incubation in DAKO Target retrieval solution (DAKO Corp., Carpinteria, CA) at weights (Mw). All identified proteins in the proteome map are 95°C for 20 min. Endogenous tissue peroxidase activity was sup- summarized in Table 1. Figure 2A illustrates mass spectra representative of a single protein spot (spot 70 in Figure 1) pressed by an incubation in 3% H2O2 at room temperature for 5 min. Nonspecific bindings were blocked using 5% (vol/vol) goat serum obtained by MALDI-TOF MS, and Figure 2B demonstrates (Vector Laboratories, Burlingame, CA) in Tris-buffered saline at peptide mass fingerprint analysis that matched those mass room temperature for 1 h. The sections were then incubated with spectra with the mouse serine protease inhibitor EIA with a rabbit polyclonal anti-elastin antibody (#RDI-TP592; Research Diag- fingerprint z score of 2.43 (Ͼ99 percentile; P Ͻ 0.01). nostics Inc., Flanders, NJ), 1:400 in 1% goat serum at 4°C overnight. Slides were washed three times in Tris buffer before incubation with biotinylated secondary antibody (Vector Laboratories), 1:200 in 1% Differential Renal Protein Expression in the OVE26 goat serum at room temperature for 30 min. After three washes in Tris Mice during Early Type 1 Diabetic Nephropathy buffer, the sections were incubated with an avidin/biotinylated peroxidase complex (Vectastain Elite ABC kit; Vector Laboratories) for Figure 1B shows a representative 2-D gel of 120-d-old 30 min. Immunoreactive elastin was detected by color developing OVE26 mouse kidney proteins. A total of 41 protein spots with Chromagen 3-3' diaminobenzidine (Vector Laboratories) for 4 were differentially expressed in OVE26 diabetic kidneys com- min. All sections were counterstained with hematoxylin. A section of pared with the controls (n ϭ 5). Of these differentially ex- mouse aorta was used as the positive control. A negative control was pressed proteins, 30 proteins were identified in normal mouse performed by incubation with 1% goat serum without primary kidney proteome map. The altered proteins were classified into antibody. functional groups on the basis of their major cellular functions, including proteases, protease inhibitors, apoptosis-associated Results proteins, regulators of oxidative tolerance, calcium-binding Clinical Characteristics of OVE26 Diabetic Mice proteins, transport regulators, cell signaling proteins, and Type 1 diabetes in the OVE26 transgenic model occurs as a smooth muscle contractile elements (Table 2). Some proteins result of apoptosis of pancreatic ␤ cells caused by overexpres- had variability in their expression levels among individual sion of the calcium-binding protein calmodulin (8). Metabolic animals. Therefore, visible differences of a number of protein derangements in OVE26 mice included hyperglycemia, hypo- spots in Figure 1 were not included as significant changes. J Am Soc Nephrol 15: 650–662, 2004 Proteomics and Diabetic Nephropathy 653

Figure 1. The proteome maps of normal kidney from FVB mice (A) and diabetic kidney from OVE26 mice (B). Each map was created from a representative two-dimensional (2-D) gel image among five individuals in each group. Renal proteins were separated by 2-D PAGE on the basis of differential isoelectric points (x axis) and molecular weights (y axis). Protein spots were excised and identified by matrix-assisted laser desorption ionization–time-of-flight (MALDI-TOF) mass spectrometry (MS), followed by peptide mass fingerprinting. A total of 92 forms of 65 unique proteins were identified in the normal kidney (summarized in Table 1). Renal protein expression in OVE26 kidneys was compared with the controls (n ϭ 5). Protein spot pattern in diabetic kidneys was comparable to the normal kidneys. 2-D analysis software was used to match corresponding protein spots among gels, and the intensity of each spot was compared by statistical analyses described in the “Materials and Methods” section. Expression levels of 41 protein spots were significantly changed in diabetic kidneys (summarized in Table 2). Of these altered protein spots, 30 proteins were identified in the proteome map, whereas the other 11 proteins (spots 93 to 103) were not identified. The spots labeled with yellow-highlighted numbers were upregulated, whereas the spots labeled with blue-highlighted numbers were downregu- lated. Number labeling corresponds to the spot number in Tables 1 and 2. 654 Journal of the American Society of Nephrology J Am Soc Nephrol 15: 650–662, 2004

Table 1. The identified proteins in the proteome map for normal FVB mouse kidneya

No. Proteins NCBI I.D. Accession

1 3-Mercaptopyruvate sulfurtransferase (MST) giԽ3122930 P97532 2 Acidic nuclear phosphoprotein 32 giԽ730318 P39687 3 Alpha enolase giԽ13637776 P17182 4 Alpha-1-antitrypsin giԽ203063 AAA40788 5 Alpha-1-antitrypsin giԽ203063 AAA40788 6 Alpha-1-antitrypsin giԽ203063 AAA40788 7 Alpha-1-macroglobulin giԽ202857 AAA40723 8 Antithrombin-III giԽ416621 P32261 9 Antithrombin-III giԽ416621 P32261 10 Antithrombin-III giԽ416621 P32261 11 Antithrombin-III giԽ416621 P32261 12 Antithrombin-III giԽ416621 P32261 13 Apolipoprotein A-IV precursor giԽ109575 B40892 14 Apoliprotein A-I giԽ109571 JC1237 15 Apoliprotein A-I giԽ109571 JC1237 16 ATP synthase beta subunit giԽ1374715 AAB02288 17 ATP synthase beta subunit giԽ1374715 AAB02288 18 ATP synthase delta chain, mitochondrial precursor giԽ1352036 P35434 19 Calbindin giԽ115396 P07171 20 Calbindin giԽ115396 P07171 21 Calmodulin giԽ223872 1003191A 22 Calreticulin giԽ117505 P18418 23 Cellular retinol-binding protein giԽ809309 809309 24 Chain D, deoxyribonuclease I complex with actin giԽ229691 229691 25 Chloride intracellular channel protein 1 giԽ6685328 Q9Z1Q5 26 Clathrin, light polypeptide (Lca) giԽ7949023 NP_058040 27 Collagen alpha 3 (V) chain giԽ105709 S20375 28 Complement component 1, q subcomponent binding protein giԽ9506435 NP_062132 29 Contraception associated protein 1 (CAP 1) giԽ7429594 JE 0344 30 Crocalbin-like protein giԽ8515718 AAF76141 31 Cu/Zn superoxide dismutase giԽ226471 226471 32 Cu/Zn superoxide dismutase giԽ226471 226471 33 Cu/Zn superoxide dismutase giԽ226471 226471 34 Cytochrome b5 giԽ554539 AAA72420 35 Cytoskeletal tropomyosin giԽ37424 CAA28258 36 Cytoskeletal tropomyosin giԽ37424 CAA28258 37 Deoxyribonuclease I (Dnase I) giԽ494869 494869 38 Deoxyribonuclease I (Dnase I) giԽ494869 494869 39 Dna-K type molecular chaperone grp75 precursor giԽ2119726 I56581 40 Dna-K type molecular chaperone grp75 precursor giԽ2119726 I56581 41 Elastase IIIB (protease E) giԽ14195655 P08861 42 Ferritin heavy chain giԽ6753912 NP_034369 43 Ferritin heavy chain giԽ6753912 NP_034369 44 Ferritin heavy chain giԽ6753912 NP_034369 45 Ferritin light chain 1 giԽ6753914 NP_034370 46 Fructose 1,6-bisphosphatase giԽ119740 P19112 47 Glutamate cysteine ligase giԽ8393446 NP_059001 48 GRP78 (78 kDa glucose-regulated protein) giԽ4033392 Q90593 49 GRP78 (78 kDa glucose-regulated protein) giԽ4033392 Q90593 50 GTP-specific succinyl-CoA synthetase beta subunit giԽ3766203 AAC64399 51 Heat shock 60 kDa protein (HSP60) (60 kDa chaperone) giԽ1334284 CAA37654 52 Heat shock 60 kDa protein (HSP60) (60 kDa chaperone) giԽ1334284 CAA37654 53 High mobility group 1 protein giԽ600761 AAA57042 J Am Soc Nephrol 15: 650–662, 2004 Proteomics and Diabetic Nephropathy 655

Table 1. (continued)

No. Proteins NCBI I.D. Accession

54 Hippocampal cholinergic neurostimulating peptide precursor giԽ9453889 BAB03276 55 Histone H3.2 giԽ70755 HSXL32 56 HSPC207 giԽ7106804 AAF36127 57 Ig VL giԽ227745 227745 58 Lactate dehydrogenase 2, B chain giԽ6678674 NP_032518 59 Myosin, light chain, smooth muscle giԽ12737351 XP_012180 60 NADH-ubiquinone oxidoreductase 24 kDa subunit precursor giԽ128867 P19234 61 NADH-ubiquinone oxidoreductase 75 kDa subunit precursor giԽ128825 P15690 62 Na-H exchanger, isoform 3 regulator 1 giԽ6755566 NP_036160 63 Na-H exchanger, isoform 3 regulator 1 giԽ6755566 NP_036160 64 Na-H exchanger, isoform 3 regulator 1 giԽ6755566 NP_036160 65 Na-H exchanger, isoform 3 regulator 1 giԽ6755566 NP_036160 66 Nucleobindin giԽ16758210 NP_167582 67 Phosphatidylethanolamine binding protein giԽ8393910 NP_058932 68 Put. Beta actin giԽ49868 CAA27396 69 Put. Beta actin giԽ49868 CAA27396 70 Serine protease inhibitor EIA; serpin clade B giԽ22347578 AAM95933 71 Recombinant rat annexin V giԽ4139939 1BC1 72 Alpha 1-antitrypsin/serine protease inhibitor 1-1 giԽ6678079 NP_033269 73 Alpha 1-antitrypsin/serine protease inhibitor 1-1 giԽ6678079 NP_033269 74 Alpha 1-antitrypsin/serine protease inhibitor 1-1 giԽ6678079 NP_033269 75 Alpha 1-antitrypsin/serine protease inhibitor 1-1 giԽ6678079 NP_033269 76 Albumin giԽ5915682 P07724 77 Albumin giԽ5915682 P07724 78 Albumin giԽ5915682 P07724 79 Albumin giԽ5915682 P07724 80 Skeletal muscle tropomyosin giԽ339958 AAA61227 81 Sodium-hydrogen exchanger regulatory factor giԽ5732682 AAD49224 82 Syntaxin 11 giԽ7447077 JE0094 83 Thioredoxin giԽ135776 P11232 84 T-kininogen giԽ207341 AAA4225 85 Transthyretin giԽ7305599 NP_038725 86 Tropomyosin 1, smooth muscle giԽ136100 P10469 87 Tropomyosin 4 giԽ6981672 NP_036810 88 Tropomyosin 5 giԽ136097 P21107 89 Tropomyosin alpha chain, smooth muscle giԽ136101 P06469 90 Tropomyosin, fibroblast isoform 1 giԽ1174753 P46901 91 Tubulin, beta-3 giԽ12852060 BAB29257 92 Vimentin giԽ1353212 P48670

a The identified proteins were summarized with the NCBI identification and accession numbers. Only the identities with significant z scores (Ͼ1.65, P Ͻ 0.05) were included. All of the identified proteins were in the expected ranges of pI and Mw in the 2-D gel in Figure 1A.

Coordinated Changes in the Renal Elastin-Elastase the only functional counterpart of MNEI on the basis of se- System in Diabetic Kidneys quence, tissue expression, and inhibitory function (13). West- We identified increased expression of the serine protease ern blot analysis in 120-d-old mouse kidneys confirmed the inhibitor EIA (spot 70, Figures 1 and 2). EIA was recently presence of MNEI in normal kidneys and its upregulation by identified as one of the four mouse homologs of the human diabetes (Figure 3). Therefore, MNEI and EIA are used inter- elastase inhibitor MNEI (13). MNEI is one of the most effi- changeably in the rest of this article. cient inhibitors of elastase-like serine proteases and is the Coordinated changes in expression of elastase inhibitor product of a single gene (SERPINB1) in humans (14, 15). The MNEI, which was increased (Figure 4, A and C), and elastase characterization of the mouse homologs revealed that EIA is IIIB, which was decreased (Figure 4, B and D), suggested the 656 Journal of the American Society of Nephrology J Am Soc Nephrol 15: 650–662, 2004

Table 2. Quantitative analysis and functional classification of differentially expressed proteins in OVE26 diabetic mouse kidneya

Normal DM Alteration Proteins Spot No. (DM/Normal) Average SEM Average SEM Protease inhibitors antithrombin-IIIb 12 21061 1982 36561 2971 Up (1.74) hippocampal cholinergic neurostimulating peptide precursorc,d 54 184971 27935 285705 23368 Up (1.54) serine protease inhibitor EIA (Serpinb1)e 70 27525 7505 77069 16540 Up (2.80) T-kininogenb 84 14506 2930 45489 12234 Up (3.14) Proteases contraception associated protein 1 (CAP 1)f,g 29 6658 5300 32339 21017 Up (4.86) elastase IIIB (protease E) 41 3377 884 1889 785 Down (0.56) Apoptosis-associated proteins deoxyribonuclease I (Dnase I) 37 27650 4364 102206 18588 Up (3.70) histone H3.2 55 7848 3181 21115 7500 Up (2.69) Regulators of oxidative tolerance ferritin heavy chainb 42 6890 2164 27206 4661 Up (3.95) ferritin heavy chainb 44 89632 30177 172614 16286 Up (1.93) ferritin light chain 1b 45 373232 31457 561829 28061 Up (1.51) Calcium-binding proteins calbindinb 20 10538 9629 70654 53779 Up (6.70) calmodulinb 21 125072 43571 398045 139427 Up (3.18) crocalbin-like proteinh 30 2430 902 9598 4151 Up (3.95) recombinant rat annexin V 71 2978 465 7206 899 Up (2.42) Transport regulators cellular retinol-binding proteinb 23 33017 7247 69722 12705 Up (2.11) Na-H exchanger, isoform 3 regulator 1b 62 8095 1847 12810 937 Up (1.58) syntaxin 11 82 6142 2145 20802 7270 Up (3.39) transthyretinb 85 68362 4563 37306 7219 Down (0.55) Cell-signaling proteins complement component 1, q subcomponent binding proteini 28 39722 11846 108044 36701 Up (2.72) Smooth muscle contractile elements cytoskeletal tropomyosinb 35 383707 44003 730919 138816 Up (1.90) myosin, light chain, smooth muscleb 59 51242 12998 116637 39385 Up (2.28) skeletal muscle tropomyosinb 80 8244 1491 16826 1990 Up (2.04) tropomyosin 1, smooth muscleb 86 6614 1149 18950 4519 Up (2.86) tropomyosin 5b 88 4256 1011 12602 3543 Up (2.96) tropomyosin alpha chain, smooth muscleb 89 77722 18199 123246 16660 Up (1.59) tropomyosin, fibroblast isoform 1b 90 113846 29107 226989 79713 Up (1.99) vimentinb 92 8827 2112 24373 6454 Up (2.76) Miscellaneous apolipoprotein A-IV precursorb 13 14698 3514 39656 13772 Up (2.70) HSPC207j 56 27175 4031 38678 7869 Up (1.42) unidentified 93 1326 569 2129 457 Up (1.61) unidentified 94 4267 818 6969 1319 Up (1.63) unidentified 95 13487 3108 21497 5025 Up (1.59) unidentified 96 3020 394 4447 509 Up (1.47) unidentified 97 3893 1106 16322 5095 Up (4.19) unidentified 98 2179 834 4517 1837 Up (2.07) unidentified 99 3911 539 7872 1332 Up (2.01) unidentified 100 1924 983 14320 8936 Up (7.44) unidentified 101 17652 2633 8395 1417 Down (0.48) unidentified 102 2700 1091 11030 3410 Up (4.09) unidentified 103 4630 4192 15224 8128 Up (3.29)

a The spot intensity (in pixel units) representing the amount of protein per spot was analyzed using 2-D analysis software. Only statistically significant changes were included, and the altered proteins were classified into variable functional categories on the basis of their major functions in the protein databases. b Proteins that were previously shown to be regulated during diabetes. c 99% amino acid identity with phosphatidylethanolamine binding protein (giԽ8393910, NP_058932). d Synonym: Raf kinase inhibitor protein (RKIP). e 80% amino acid identity with human monocyte/neutrophil elastase inhibitor (MNEI) (giԽ266344, P30740). f 95% amino acid identity with DJ-1 (giԽ3256343, BAA29063). g Containing a protease domain (merops.sanger.ac.uk). h 99% amino acid identity with calumenin (giԽ14718453, AAK72908). i 95% amino acid identity with P32-RACK (receptor of activated protein kinase C) (giԽ18652991, AAL77246). j 97% amino acid identity with uncharacterized bone marrow protein. J Am Soc Nephrol 15: 650–662, 2004 Proteomics and Diabetic Nephropathy 657

performed immunohistochemical study for elastin on mouse kidneys and human renal biopsies. Kidneys from 120-d-old (n ϭ 2) and 325-d-old (n ϭ 2) mice were studied. Normal distribution of elastin in mouse kidneys was observed in the Bowman’s capsule, GBM, juxtaglomerular apparatuses, and vessels, with the most prominent staining in proximal tubular epithelial cells. Although there was no difference observed in diabetic mice versus normal at 120 d of age, a markedly increased accumulation of elastin was observed in the macula densa, collecting ducts, and pelvicalyceal epithelia in diabetic kidneys of older (325 d) mice (Figure 5). Figure 6 shows the elastin immunohistochemistry of human renal biopsies (n ϭ 3). Normal distribution of elastin in human kidneys was different from that in normal mice. In general, elastin expression in renal tubular cells and the glomeruli was greater in humans than in mice. Elastin expression was markedly increased in renal tubular epithelial cells in early-stage diabetic kidneys (Figure 6, B and D) and in the interstitium of late-stage diabetic kidney (Figure 6F). There was no obvious increase in the amount of elastin staining in the glomeruli of patients with diabetes, although the staining was more prominent in the periphery of glomerular tufts from patients with Figure 2. MALDI-TOF MS and peptide mass fingerprinting. (A) diabetes. Peptide masses (in mass per charge [m/z] units) were obtained by MALDI-TOF MS after in-gel tryptic digestion of spot 70 in Figure 1. Discussion (B) The peptide masses were queried to the theoretical masses of Current therapy aiming to halt the progression of renal mammalian entries in the National Center for Biotechnology Infor- mation protein database using the ProFound search engine. A maxi- damage in established diabetic nephropathy is limited to anti- mum 150-ppm error window and one missed tryptic cleavage were hypertensive drugs, especially angiotensin-converting enzyme allowed, although most of the matched masses had no missed cleav- inhibitors and angiotensin receptor blockers (16). Although age. After excluding autolytic trypsin masses, 10 of 12 sample masses this therapy effectively slows the rate of progression of diabetic matched with serine protease inhibitor (EIA), serpin clade B with z renal injury, renal failure remains a common complication. score of 2.43 (Ͼ99 percentile, P Ͻ 0.01). *Oxidation at methionine Defining the pathophysiologic mechanisms of diabetic ne- residue. phropathy is necessary to identify new targets for therapeutic intervention. Our approach to defining potential novel mechanisms of diabetic nephropathy was to identify proteins with altered expression in diabetic kidneys using proteomic techniques. Developments in proteomic techniques during the past decade allow simultaneous identification of a large number of proteins and comparison of expression of these proteins between groups. The present study compared renal protein expression of OVE26 mice, a transgenic mouse model that mim- ics many aspects of human type 1 diabetes, with that of FVB mice, a background nondiabetic strain. The renal lesion in Figure 3. Western blotting for monocyte/neutrophil elastase inhibitor OVE26 mice was characterized previously by showing renal (MNEI). Proteins derived from whole-kidney lysate of 120-d-old FVB histologic changes commonly seen in human diabetic nephrop- and OVE26 mice (total 30 ␮g) were separated by 1-D PAGE and athy, including mesangial expansion and increased thickness of transferred to a nitrocellulose membrane. Rabbit polyclonal anti- the GBM by 120 d of age (7). The present study also shows MNEI was used as a primary antibody. Only a single band at approx- that OVE26 mice had a dramatic increase in urinary albumin imately 42 kD, the expected molecular size, was observed in each excretion. Thus, OVE26 mice present many of the features lane. Recombinant human MNEI was used as a positive control. Western blot analysis confirmed that EIA, the murine ortholog of observed in human diabetic nephropathy. MNEI, was present in the mouse kidney and was upregulated in the We identified 92 proteins in our initial proteome map of OVE26 diabetic kidneys. normal FVB mouse kidney, 30 of which were differentially expressed in diabetic kidneys. The mechanism for altered expression of these proteins was not investigated but could be hypothesis that elastin expression would be increased in dia- induced by hyperglycemia and/or hypoinsulinemia, as the an- betic kidneys. To determine whether elastin accumulated in imals were both hyperglycemic and hypoinsulinemic. Nineteen diabetic kidneys and to define the location of this change, we of these altered proteins were previously shown to be regulated 658 Journal of the American Society of Nephrology J Am Soc Nephrol 15: 650–662, 2004

Figure 4. Coordinated changes of elastase inhibitor MNEI (A and C) and elastase IIIB (B and D). (A and B) Zoom-in images of spots 70 and 41, respectively, from individual animals. (C and D) The summary of intensity data of those two spots. The elastase inhibitor MNEI was increased, whereas the elastase IIIB was decreased in diabetic kidneys. *P Ͻ 0.05.

during diabetes (4, 17–24) and are marked with a superscript b densa. The changes in elastin expression in 325-d-old mice but in Table 2. Involvement of the other 11 altered proteins in not at 120 d suggested that a gradual increase in elastin diabetic nephropathy had not previously been reported, sug- expression required long-term changes in MNEI and elastase gesting their roles in novel mechanisms of diabetic IIIB. The pattern of elastin expression in normal human kid- nephropathy. neys differed from that in normal mice with greater elastin When we examined the list of proteins regulated in diabetic expression in both tubules and glomeruli in human kidneys. In kidneys (Table 2), the only apparent pathway in which multiple patients with diabetes, elastin expression was increased in renal proteins were regulated was the elastin-elastase pathway. Be- tubular epithelial cells. The mechanism of differences in nor- cause the role of the elastin ECM protein had not been well mal distribution and increased elastin expression in mice ver- characterized in diabetic nephropathy, changes in elastin-reg- sus humans remains unknown. ulating proteins focused our attention on this pathway. Expres- A role for elastin in the pathogenesis of diabetic nephropathy sion of the elastase inhibitor MNEI was increased approxi- has not previously been described. The major function of mately threefold in diabetic kidneys. MNEI is one of the most elastin is to provide vascular wall elasticity (26). Alterations in efficient inhibitors of elastase-like serine proteases by forming elastin expression and function are associated with vasculopa- stable covalent inhibitory complexes with target proteases. thy of large vessels induced by diabetes (27, 28). However, we Whereas the expression of elastase inhibitor MNEI was in- did not observe change in elastin deposition in the intrarenal creased in diabetic kidneys, elastase IIIB expression was de- vessels. Therefore, it is unlikely that altered elastin expression creased. Thus, both expression and activity of elastase would observed in our study is related to diabetic renal vasculopathy. be predicted to be decreased in diabetic nephropathy. These In the kidney, elastin plays an important role in stabilizing the coordinated changes defined by proteomic analysis suggested glomerular tuft (25). The increase in elastin expression, how- the hypothesis that elastin expression would be increased in ever, was located in renal tubular epithelial cells, not glomer- diabetic kidneys. Renal elastin immunohistochemistry was per- uli. These findings suggest that elastin may play a role in formed to address this hypothesis. tubular or interstitial changes, which accompany diabetic The elastin distribution in normal mouse kidney in the nephropathy. present study was similar to that observed by Sterzel et al. (25). Elastin deposition is a highly regulated process that occurs Immunohistochemical analysis confirmed an increase in elastin primarily during early development (29). Increased elastin expression in diabetic kidneys, particularly in the macula expression results from increased transcription and translation. J Am Soc Nephrol 15: 650–662, 2004 Proteomics and Diabetic Nephropathy 659

renal tubular epithelial cells, not in extracellular spaces, occurred in parallel with an increase of vimentin (Table 2), a marker for cells derived from mesenchymal tissues (31). Our data were consistent with the data reported by Rastaldi et al. (32) that renal tubular epithelial cells can produce ECM proteins and directly intervene in fibrotic processes via the epithelial-mesenchymal transdifferentiation. The increase in expression of a myofibroblast protein (fibroblast tropomyosin) and proteins associated with proliferation, modulation, and differentiation of myofibroblast and fibrogenesis (calmodulin and cellular retinol-binding protein; Table 2) provides further support for this process (33, 34). In addition, Figure 6F shows a marked increase of elastin staining in the tubulointerstitium of an end-stage diabetic kidney. Taken together, these data suggest that elastin may play a role in tubular disorders and interstitial fibrosis in diabetic nephropathy. However, elastin accumulation might reflect advanced tissue response to injury. Thus, alterations in the renal elastin-elastase system may be the cause of or the result of diabetic nephropathy. Further functional and time-course studies are required to evaluate the significance of disordered elastin deposition. Although our study is the first to examine global changes in protein expression in diabetic kidneys, genomic approaches were previously applied to identify genes in renal cells regulated by hyperglycemia. Clarkson et al. (6) showed that 200 genes were differentially expressed when mesangial cells were propagated in high ambient glucose in vitro. It is not surprising that our proteomic data and Clarkson’s genomic data are not completely concordant. Some genes, such as ferritin, were upregulated by hyperglycemia at the level of transcription and translation. However, other genes, such as myosin, were down- regulated in Clarkson’s study, but the protein products of those genes were upregulated in vivo in our OVE26 diabetic mice. Several differences in the experimental approaches might ac- count for the disparate findings. First, we studied the whole kidneys from diabetic animals, whereas Clarkson et al. studied mesangial cells in vitro. Second, changes in mRNA and protein do not always correlate (35–38). Third, differential protein expression can result from posttranslational modifications that are not detectable by genomic analysis. Finally, protein deg- radation is not measured by changes in mRNA expression. Figure 5. Renal immunohistochemistry for elastin in 325-d-old mice. Several limitations of 2-D proteomic analysis in the present (A) Positive control from FVB mouse aorta demonstrates that elastin study need to be noted. First, this approach identified high- is typically present in intimal and adventitial layers of the aorta. (B) abundance proteins, whereas detection of low-abundance pro- Negative control demonstrates the absence of elastin immunoreactive teins was limited. This may explain the failure to identify staining. C, E, G, and I were from normal (FVB) mice, and D, F, H, several proteins previously shown to be regulated in diabetes, and J were from diabetic (OVE26) mice. Normal distribution of for example, TGF-␤ and protein kinase C. Second, we did not elastin was most prominent in proximal tubular epithelial cells (C). In identify all of the proteins present in the 2-D gels. Using more the OVE26 diabetic kidneys, elastin expression was increased in the sensitive mass spectrometric techniques, such as tandem MS, macula densa (D), collecting ducts (F), and pelvicalyceal epithelia may permit identification of low-abundance proteins. Third, a (H). There was no change of elastin expression observed in the vessels (I and J). Magnification, ϫ40. larger number of visualized protein spots would be expected on individual gels than the 300 spots observed in the present study. This limitation likely resulted from the small amount of protein (100 ␮g) used for individual analytical gels. In addi- TGF-␤, which is upregulated in diabetic kidneys, was reported tion, the extraction protocol used did not solubilize some previously to stabilize elastin mRNA and promote elastin dep- protein components, especially membrane-associated and hy- osition (30). The increased elastin expression identified in drophobic proteins. 660 Journal of the American Society of Nephrology J Am Soc Nephrol 15: 650–662, 2004

Figure 6. Elastin immunohistochemistry of human renal biopsies. A, C, and E were from three normal biopsies, and B, D, and F were from three patients with type 1 diabetes. Elastin expression was markedly increased in renal tubular epithelial cells in early-stage (B and D) and in the interstitium of late-stage (F) diabetic kidneys. There was no obvious increase in the amount of elastin staining in the glomeruli of patients with diabetes, although the staining was more prominent in the periphery of glomerular tufts from patients with diabetes (B). Magnifications: ϫ40 in A through E; ϫ10 in F.

We examined protein expression from the whole kidneys in change in other structures or changes in other structures are in the present study. This approach cannot identify localized an opposite direction. Moreover, it should be noted that iden- changes such as those that may be confined to glomeruli, tification of the altered proteins in the kidney does not confirm tubules, or even podocytes. In additional, the magnitude of that they are kidney-specific changes, as a systematic study of changes is affected by degrees of changes in individual intra- other organs was not performed. Changes similar to those that renal structures. For example, mild changes in mesangial cells we observed in the kidney may be present in liver, muscles, or may not be detected in the whole-kidney analysis if there is no other organs. Finally, we examined renal protein expression J Am Soc Nephrol 15: 650–662, 2004 Proteomics and Diabetic Nephropathy 661 only in 120-d-old animals. Defining alterations in renal protein 12. Rees DD, Rogers RA, Cooley J, Mandle RJ, Kenney DM, expression at a single time point does not represent the entire Remold-O’Donnell E: Recombinant human monocyte/neutrophil dynamic process of diabetic nephropathy. We are currently elastase inhibitor protects rat lungs against injury from cystic performing a serial study of other time points of isolated fibrosis airway secretions. Am J Respir Cell Mol Biol 20: 69–78, glomeruli and blood vessels to characterize more thoroughly 1999 13. Benarafa C, Cooley J, Zeng W, Bird PI, Remold-O’Donnell E: the renal subproteome during diabetes. Characterization of four murine homologs of the human ov- In summary, proteomic analysis revealed coordinated serpin monocyte neutrophil elastase inhibitor MNEI (SER- changes of elastase inhibitor MNEI and elastase IIIB, leading PINB1). J Biol Chem 277: 42028–42033, 2002 to the identification of increased elastin expression in mouse 14. Remold-O’Donnell E, Chin J, Alberts M: Sequence and molec- and human diabetic kidneys. Application of proteomic analysis ular characterization of human monocyte/neutrophil elastase in- may yield new insights into the pathogenic mechanisms of hibitor. Proc Natl Acad SciUSA89: 5635–5639, 1992 diabetic nephropathy. 15. Cooley J, Takayama TK, Shapiro SD, Schechter NM, Remold- O’Donnell E: The serpin MNEI inhibits elastase-like and chy- motrypsin-like serine proteases through efficient reactions at two Acknowledgments active sites. Biochemistry 40: 15762–15770, 2001 This study was supported by National Institutes of Health Grants 16. Cooper ME: Pathogenesis, prevention, and treatment of diabetic R21 DK62086-01 (J.B.K.), R01 HL66358-01 (J.B.K.), R21 DK06289 nephropathy. Lancet 352: 213–219, 1998 (K.R.M.), and HL66548 (E.R.O.) and Department of Veterans Affairs 17. Asakawa H, Tokunaga K, Kawakami F: Elevation of fibrinogen (J.B.K. and K.R.M.). We gratefully acknowledge the assistance of and thrombin-antithrombin III complex levels of type 2 diabetes Patricia M. Kralik, Jian Cai, Xia Shen, Naira Meterveli, and Michael mellitus patients with retinopathy and nephropathy. J Diabetes E. Brier. Complications 14: 121–126, 2000 18. Clodi M, Oberbauer R, Bodlaj G, Hofmann J, Maurer G, Kostner References K: Urinary excretion of apolipoprotein(a) fragments in type 1 1. USRDS. United States Renal Data System. Incidence and prev- diabetes mellitus patients. Metabolism 48: 369–372, 1999 alence of ESRD. Am J Kidney Dis 30: S40–S53, 1997 19. Ambuhl P, Amemiya M, Preisig PA, Moe OW, Alpern RJ: 2. Schleicher E, Kolm V, Ceol M, Nerlich A: Structural and func- Chronic hyperosmolality increases NHE3 activity in OKP cells. tional changes in diabetic glomerulopathy. Kidney Blood Press J Clin Invest 101: 170–177, 1998 Res 19: 305–315, 1996 20. Basu TK, Basualdo C: Vitamin A homeostasis and diabetes 3. Mauer SM, Lane P, Hattori M, Fioretto P, Steffes MW: Renal mellitus. Nutrition 13: 804–806, 1997 structure and function in insulin-dependent diabetes mellitus and 21. Rowe DJ, Anthony F, Polak A, Shaw K, Ward CD, Watts GF: type I membranoproliferative glomerulonephritis in humans. Retinol binding protein as a small molecular weight marker of J Am Soc Nephrol 2[Suppl 1]: S181–S184, 1992 renal tubular function in diabetes mellitus. Ann Clin Biochem 4. Lehmann R, Schleicher ED: Molecular mechanism of diabetic 24[Suppl]:477–482, 1987 nephropathy. Clin Chim Acta 297: 135–144, 2000 22. Tschope C, Reinecke A, Seidl U, Yu M, Gavriluk V, Riester U, 5. Marcantoni C, Ortalda V, Lupo A, Maschio G: Progression of Gohlke P, Graf K, Bader M, Hilgenfeldt U, Pesquero JB, Ritz E, renal failure in diabetic nephropathy. Nephrol Dial Transplant Unger T: Functional, biochemical, and molecular investigations 13[Suppl 8]: 16–19, 1998 of renal kallikrein-kinin system in diabetic rats. Am J Physiol 6. Clarkson MR, Murphy M, Gupta S, Lambe T, Mackenzie HS, 277: H2333–H2340, 1999 Godson C, Martin F, Brady HR: High glucose-altered gene 23. Essawy M, Soylemezoglu O, Muchaneta-Kubara EC, Shortland expression in mesangial cells. Actin-regulatory protein gene ex- J, Brown CB, el Nahas AM: Myofibroblasts and the progression pression is triggered by oxidative stress and cytoskeletal disas- of diabetic nephropathy. Nephrol Dial Transplant 12: 43–50, sembly. J Biol Chem 277: 9707–9712, 2002 1997 7. Carlson EC, Audette JL, Klevay LM, Nguyen H, Epstein PN: 24. Sanai T, Sobka T, Johnson T, el Essawy M, Muchaneta-Kubara Ultrastructural and functional analyses of nephropathy in calm- EC, Ben Gharbia O, el Oldroyd S, Nahas AM: Expression of odulin-induced diabetic transgenic mice. Anat Rec 247: 9–19, cytoskeletal proteins during the course of experimental diabetic 1997 nephropathy. Diabetologia 43: 91–100, 2000 8. Epstein PN, Overbeek PA, Means AR: Calmodulin-induced ear- 25. Sterzel RB, Hartner A, Schlotzer-Schrehardt U, Voit S, ly-onset diabetes in transgenic mice. Cell 58: 1067–1073, 1989 Hausknecht B, Doliana R, Colombatti A, Gibson MA, Braghetta 9. Thongboonkerd V, Gozal E, Sachleben LR, Arthur JM, Pierce P, Bressan GM: Elastic fiber proteins in the glomerular mesan- WM, Cai J, Chao J, Bader M, Pesquero JB, Gozal D, Klein JB: gium in vivo and in cell culture. Kidney Int 58: 1588–1602, 2000 Proteomic analysis reveals alterations in the renal kallikrein 26. Debelle L, Tamburro AM: Elastin: Molecular description and pathway during hypoxia-induced hypertension. J Biol Chem 277: function. Int J Biochem Cell Biol 31: 261–272, 1999 34708–34716, 2002 27. Tanno T, Yoshinaga K, Sato T: Alteration of elastin in aorta from 10. Arthur JM, Thongboonkerd V, Scherzer JA, Cai J, Pierce WM, diabetics. Atherosclerosis 101: 129–134, 1993 Klein JB: Differential expression of proteins in renal cortex and 28. Kwan CY, Wang RR, Beazley JS, Lee RM: Alterations of elastin medulla: A proteomic approach. Kidney Int 62: 1314–1321, and elastase-like activities in aortae of diabetic rats. Biochim 2002 Biophys Acta 967: 322–325, 1988 11. Thongboonkerd V, Luengpailin J, Cao J, Pierce WM, Cai J, 29. Pasquali-Ronchetti I, Baccarani-Contri M: Elastic fiber during Klein JB, Doyle RJ: Fluoride exposure attenuates expression of development and aging. Microsc Res Tech 38: 428–435, 1997 Streptococcus pyogenes virulence factors. J Biol Chem 277: 30. Kucich U, Rosenbloom JC, Abrams WR, Bashir MM, Rosen- 16599–16605, 2002 bloom J: Stabilization of elastin mRNA by TGF-beta: Initial 662 Journal of the American Society of Nephrology J Am Soc Nephrol 15: 650–662, 2004

characterization of signaling pathway. Am J Respir Cell Mol Biol 35. Tooyama I, Sato H, Yasuhara O, Kimura H, Konishi Y, Shen 17: 10–16, 1997 Y, Walker DG, Beach TG, Sue LI, Rogers J: Correlation of 31. Geisler N, Plessmann U, Weber K: Amino acid sequence char- the expression level of C1q mRNA and the number of C1q- acterization of mammalian vimentin, the mesenchymal interme- positive plaques in the Alzheimer disease temporal cor- diate filament protein. FEBS Lett 163: 22–24, 1983 tex. Analysis of C1q mrna and its protein using adjacent or 32. Rastaldi MP, Ferrario F, Giardino L, Dell’Antonio G, Grillo C, nearby sections. Dement Geriatr Cogn Disord 12: 237–242, Grillo P, Strutz F, Muller GA, Colasanti G, D’Amico G: Epithe- 2001 lial-mesenchymal transition of tubular epithelial cells in human 36. Tong X, Kawabata H, Koeffler HP: Iron deficiency can upregu- renal biopsies. Kidney Int 62: 137–146, 2002 late expression of transferrin receptor at both the mRNA and 33. Dalley A, Smith JM, Reilly JT, Neil SM: Investigation of cal- protein level. Br J Haematol 116: 458–464, 2002 modulin and basic fibroblast growth factor (bFGF) in idiopathic 37. Gygi SP, Rochon Y, Franza BR, Aebersold R: Correlation be- myelofibrosis: Evidence for a role of extracellular calmodulin in tween protein and mRNA abundance in yeast. Mol Cell Biol 19: fibroblast proliferation. Br J Haematol 93: 856–862, 1996 1720–1730, 1999 34. Uchio K, Tuchweber B, Manabe N, Gabbiani G, Rosenbaum J, 38. Ji P, Xuan JW, Onita T, Sakai H, Kanetake H, Gabril MY, Sun Desmouliere A: Cellular retinol-binding protein-1 expression Y, Moussa M, Chin JL: Correlation study showing no concor- and modulation during in vivo and in vitro myofibroblastic dance between EPAS-1/HIF-2␣ mRNA and protein expression differentiation of rat hepatic stellate cells and portal fibroblasts. in transitional cellcancer of the bladder. Urology 61: 851–857, Lab Invest 82: 619–628, 2002 2003