BRIEF REVIEW www.jasn.org

Diagnostic Potential of Proteome: A Broken Mirror of Renal Diseases

Michael S. Goligorsky, Francesco Addabbo, and Edmond O’Riordan

Departments of Medicine and Pharmacology and Renal Research Institute, New York Medical College, Valhalla, New York

ABSTRACT This brief overview of studies into the urine proteome illustrates its potential value for highly controversial issue,1 a complex map diagnostic, prognostic, and pathophysiologic discovery. Hypothesis-targeted investi- of the slit diaphragm and podocyte with gations of individual as well as proteome-wide searches for urinary biomar- their makeup is rapidly emerging. kers of various diseases and their progression are reviewed. The majority of urine This map should help in the understanding proteins appear as cleavage products that are found not only as free solutes but also of the breadth of glomerular proteins that in secreted membrane vesicles called exosomes. Described are several recent exam- are contributed to the tubular fluid after ples of important diagnostic findings using urine proteomics along with the idea that glomerular injury. Since the discovery of signature profiles of injury to individual segments can be measured by this nephrin, a host of protein components of technology. Shared are some thoughts on the most challenging step: Integration of the filtration barrier are known.8 Nephrin- seemingly unrelated findings of various protein fragments into a rational pathogenetic uria, for example, has been recognized as a pathway(s). The future chance that the centuries-old technique of uroscopy will reveal feature of diabetes and diabetic nephropa- its secrets using modern proteomic approaches makes gradual improvement. thy.9,10 Among 40 normoalbuminuric pa- tients with type 1 diabetes, nephrinuria, as J Am Soc Nephrol 18: 2233–2239, 2007. doi: 10.1681/ASN.2006121399 judged by immunorecognition of protein bands with molecular weight of 18-, 32-, 40-, 60-, 75-, and full-length 185-kD pro- , a cardinal symptom of renal dis- in our understanding of molecular mecha- teins, was detected in 30% of cases, whereas ease, has long been considered as a potential nisms of renal handling of proteins, espe- nephrin was undetectable by Western blot “black box” for diagnostic and even prognos- cially albumin,1,2 combined with the ongo- analysis in the urine of healthy individuals. tic information. Urine proteins arise from ing revolution in the technological tools for Is it possible that nephrinuria may serve as various sources, including filtration of peptide detection, quantification, and an early warning of impending nephropa- plasma proteins; impaired reabsorption of identification,3–7 have re-energized at- thy, and studies to address this question are filtered proteins; and appearance of proteins tempts to obtain proteomic footprints of in progress in Holthofer’s laboratory. that originate from injured glomeruli, tu- renal disease. In this brief review, we re- In parallel, a more detailed picture of bules, infiltrating inflammatory cells, or con- count the most important findings made the machinery for tubular protein reab- nective tissue as well as those that enter the so far in the field of urine biomarker pro- sorption is emerging. Nonspecific recep- urine in the urinary tract below the . teins, reflect on the capabilities and limita- tors megalin and cubilin and their internal- Not surprising, many traditional studies have tions of these biomarkers—frequently ization, lysosomal degradation of protein focused on the excretion of (1) individual protein fragments—in diagnosing the dis- cargo, and recycling to the luminal plasma proteins such as enzymes or albumin, (2) ease, and discuss the tortuous intellectual membrane in the proximal epithelium selectivity of proteinuria, and (3) attempts routes that lead to the faithful reconstruc- provided important insights into the ori- at global characterization of urine proteins tion of pathogenetic mechanisms of the gins of tubular proteinuria.11 In fact, it has using two-dimensional electrophoresis. disease from these seemingly unrelated become increasingly clear that most pro- Insufficient resolution of many tech- protein fragments (integrating the image from pieces of the broken mirror). niques, the lack of understanding of the Published online ahead of print. Publication date physiologic and pathologic underpinnings available at www.jasn.org. of proteinuria, and frequent appearance of Correspondence: Dr. Michael S. Goligorsky, New FROM FOOT PROCESSES TO protein fragments that are undetectable York Medical College, 26 Research Way, Valhalla, NY FOOTPRINTS OF DISEASE 10595. Phone: 914-594-4731; Fax: 914-594-4732; using standard approaches all hampered E-mail: [email protected] early attempts to extract information from Although the actual amount of protein and Copyright © 2007 by the American Society of this ubiquitous resource. Recent progress albumin filtered by glomeruli remains a

J Am Soc Nephrol 18: 2233–2239, 2007 ISSN : 1046-6673/1808-2233 2233 BRIEF REVIEW www.jasn.org teins are excreted as proteolytic fragments Ci rather than intact molecules and, there- CM10-2200_63 fore, are poorly detectable using today’s CM10-8324_86 standard approaches.12 For instance, 90% CM10-14574_7 of albumin is degraded to small, Ͻ10-kD CM10-16528_4 Q10-4161_17 fragments. Similar degradation has been IMAC30-13913_6 documented for other proteins, including transferrin, apolipoprotein, IgG, glucose oxidase, and lactate dehydrogenase.13,14 CM10-3463_48 CM10-10003_9 CM10-15870_9 CM10-50159_6 GENERIC AND DISEASE-SPECIFIC CM10-5874_54 CM10-11673_1 CM10-8565_69 CM10-132999 MARKERS OF INJURY CM10-11980_5 Q10-33612_1 IMAC30-5863_98 Q10-66882_9 Commonality in the excretion patterns of IMAC-11905_1 IMAC30-123657_ certain proteins or their fragments in vari- IMAC30-147781 ous diseases offers critical insights into IMAC30-3620_96 H50-2511_86 IMAC30-3751_83 H50-4866_54 the default mechanisms of pathologic IMAC30-10823 H50-5026_14 H50-6449_4 processes. For instance, there are well- established markers of tubular injury (e.g., ␣ 2-microglobulin, prealbumin, lactate Ct Cv aminopeptidase) or infectious processes, such as the release of defensins.15 Mega- lin itself is normally processed via in- Proteomics correlation with Banff Score using Gene@work tramembrane proteolysis characterized by the protein kinase C–regulated metal- Figure 1. Segregation of protein peaks (shown as the type of protein chip and molecular loprotease-mediated shedding of its mass) with the specific histologic manifestations of chronic allograft nephropathy. Ci, 16 large ectodomain, and it remains to be interstitial fibrosis; Ct, tubular atrophy; Cv, chronic vascular injury. Note that a subset of elucidated how this process may be al- proteins is common for all three, some protein peaks segregate with only two patterns, tered in . and several protein peaks are specific for each pattern. One of the techniques used to distin- guish between generic and the site- or dis- definable, whereas the other two categories alkaline phosphatase and ␲-glutathione S- ease-specific excretory proteins during provide a basis for believing that certain transferase markers of the brush border proteome-wide analyses is by stratifying proteomic markers are common to all pa- and distal tubular epithelia, respectively, overlapping occurrence of the proteins in tients with CAN. have predictive value for development of the former category among patients with acute renal failure (sensitivity 100%, spec- diverse diseases and contrasting it with ificity 91%) in patients who are admitted to unique occurrences in the latter. This type INTUITIVE SELECTION OF a general intensive care unit. The major of analysis can be represented by a Venn URINARY CANDIDATE limiting factor for these markers consists of diagram, as shown in Figure 1. The figure BIOMARKER PROTEINS the high sensitivity of release by tubular ep- demonstrates a representative analysis per- ithelia even in mild injury, which does not formed by Gene@work-based software of Enzymuria necessarily portend development of acute protein peaks characteristic of interstitial Enzymuria has long been considered as a renal failure. This is exemplified by a study fibrosis, tubular injury/atrophy, and vas- potentially informative marker for the de- of patients who underwent coronary by- culopathy in patients with biopsy-graded tection of tubular injury on the basis of the chronic allograft nephropathy (CAN). The prediction that the enzymes should leak pass and almost uniformly had presentation is segregated into classes of into the urine from damaged tubular epi- postoperative increases in urine excretion protein peaks that are characteristic for the thelia. Increased urinary excretion of a ly- of N-acetyl-glucosaminidase but did not individual histologic feature, peaks that are sosomal enzyme N-acetyl-glucosamini- succumb to acute renal failure.20 shared by two histologic features, and dase has been reported in contrast Devarajan’s group21 screened 71 pa- peaks that are common to all three major nephropathy.17 Brush border enzymes, tients who were undergoing cardiopulmo- histologic presentations of CAN. The pres- such as alkaline phosphatase, ␥-glutamyl- nary bypass (a carefully selected group of ence of peptides in the first group suggests transpeptidase, and ala(leu-gly)-amino- patients without additional confounding that there is a theoretical possibility that peptidase, are elevated in acute renal inju- problems) for the urinary excretion of neu- each morphologic abnormality of CAN is ry.18 Westhuyzen et al.19 demonstrated that trophil gelatinase–associated lipocalin

2234 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 2233–2239, 2007 www.jasn.org BRIEF REVIEW

(NGAL). NGAL excretion 2 h after cardio- congenital ureteral obstruction.30 Martin MS pulmonary bypass was a powerful inde- et al.31 also provided evidence that the ex- There are several types of mass spec- pendent predictor of , cretion of AQP2 in the urine of patients trometers35 that use the assortment of diagnosed in this study as a 50% rise in who had congestive heart failure (NYHA electrical, radio frequency, and magnetic plasma creatinine from baseline, showing class II to III) and received a V2-vasopres- fields: Time-of-flight, quadrupole, ion the sensitivity of 100% and specificity of sin receptor antagonist was decreased in trap, and Fourier transform ion cyclo- 98% for a cutoff value of 50 ␮g/L (only one proportion with the resulting increase in tron resonance.34 Mass spectrometers of 51 patients without acute kidney injury free water clearance. These findings intro- lack the ability to quantify the protein had urinary NGAL level Ͼ50 ␮g/L). This duce the idea that AQP2 may serve as a detected, the drawback that can be cir- successful strategy capitalized on the previ- marker of therapeutic efficacy. cumvented using isotope-coded affinity ous unbiased experimental findings ob- tagging. Because of the competition be- tained through cDNA microarray screen- tween proteins and peptides for capture ing and put forward a role for infiltrating PROTEINS EXCRETED WITH of charges, the detection of low-abun- inflammatory cells as important contribu- EXOSOMES dance species may be compromised. tors to biomarker profiles from urine. An- Therefore, additional separation strate- Recently it was demonstrated that AQP2 other inflammatory mediator, IL-18, at gies are required. is also excreted in the urine as part of levels Ͼ100 pg/ml in the urine of 52 pa- secreted exosomes.32 Exosomes are small tients with acute respiratory distress syn- Two-Dimensional Electrophoresis. vesicles that are derived indirectly from drome was associated with a 6.5-fold in- The cartesian position of a protein is a the apical endosomal system. Among 295 creased odds for acute kidney injury in the product of its pI and molecular mass. This proteins identified, AQP2 (potential next 24 h.22 Importantly, elevated IL-18 technique has the ability to be quantitative marker of nephrogenic ), was detected 24 to 48 h before develop- either through comparison of spot size be- polycystin-1 (autosomal dominant poly- ment of this acute kidney injury. tween two gels or by a differential in-gel cystic kidney disease type 1), podocin (au- Screening of subtraction libraries in an electrophoresis. The shortcomings of two- tosomal recessive steroid-resistant ne- animal model of acute renal failure re- dimensional electrophoresis include that phrotic syndrome), nonmuscle myosin II vealed an early upregulation of a secreted more than one protein per spot can exist (potential marker of Fechner syndrome cysteine-rich protein 61 (CYR61), a and that multiple spots can contain the and Epstein syndrome), angiotensin-con- growth factor–inducible immediate early ϩ ϩ same posttranslationally modified protein, verting enzyme (hypertension), Na K / gene, and was found to be enriched in the Ϫ at times making abundance profiling 2Cl co-transporter ( urine of rats 4 to 8 h after renal ischemia.23 ϩ Ϫ problematic. Furthermore, very small or type 1), thiazide-sensitive Na /Cl co- Similarly, kidney injury molecule-1, found large proteins and very acidic or basic pro- transporter (), and ep- to be highly upregulated in postischemic teins are not visualized. ithelial sodium channel (Liddle syndrome rat kidney,24 has been detected in the urine and autosomal recessive pseudohypoaldo- of patients with ischemic acute renal failure HPLC. steronism type 1) are excreted on exo- confirmed by renal biopsy,25 as was the el- HPLC is applied to the urine and in- somes. The studied exosomes were ob- evated urinary excretion of liver-type fatty cludes size exclusion, reverse phase, tained by ultracentrifugation of a large acid–binding protein.26 strong and weak cation binding and af- volume of urine (400 ml), and proteins finity binding (i.e., Ig adsorbing to pro- were initially separated by electrophoresis Aquaporinuria in the Diagnosis of tein A). One of the disadvantages of this followed by in-gel trypsin digestion. The Hypo-osmolar and Other Syndromes system is that a buffer dilutes the pro- tryptic peptides were analyzed by nanos- Aquaporin-2 (AQP2) is one of the most teins, thus requiring repeated re-concen- pray liquid chromatography–tandem studied proteins excreted in urine; it rep- tration of the sample and making quan- mass spectrometry (MS). A modification resents a reliable example of a well-estab- tification problematic. of this technique was recently used to de- lished urinary biomarker that is critical tect more than a 50-fold increase in fe- for diagnosis of several disorders of renal Capillary Electrophoresis. tuin-A excretion in experimental animals water balance. Kanno et al.27 studied the Capillary electrophoresis is a powerful sep- and patients with acute kidney injury.33 urinary excretion of AQP2 in diabetes in- aration technology, but it lacks the ability sipidus using a RIA. Elliot et al.28 demon- to quantify absolutely the proteome.36 strated urinary AQP2 as a marker of col- TECHNOLOGICAL TOOLS FOR lecting duct responsiveness to vasopressin. Protein Array Technologies PROTEOME-WIDE RESEARCH Valenti et al.29 showed changes in AQP2 This concept is similar to gene microarray expression and excretion in conditions Technological platforms that are used in with or tissue sections roboti- such as nocturnal enuresis characterized by the discovery phase of proteome-wide cally placed on a glass slide. Techniques for hypercalciuria or during follow-up in pa- research have been reviewed elsewhere,34 protein–protein (including antigen–anti- tients who underwent surgical relief of and we only recount them here. body), protein–DNA, protein–lipid, and

J Am Soc Nephrol 18: 2233–2239, 2007 Urine Proteome in Health and Disease 2235 BRIEF REVIEW www.jasn.org protein–drug interactions exist.37 Limita- tive approaches that have also proved to be useful tatively. They found that carrier proteins, tions include accurate quantification and in classification that is based on spectral data. complement components, and bioactive dependence on the availability of antibod- In summary, it is critical to appreciate peptides were excreted at higher concen- that there is no single proteomic or infor- ies. This technology has yet to be applied to trations in patients with Dent disease. renal proteomics. A variation of this tech- matics technique to fit the diverse re- quirements of analyses; therefore, com- nique is Luminex multiplex analysis, Proteome of the Transplanted bination of several approaches offers the which is based on the array of polystyrene Kidney optimal solution to the problem. microspheres with two spectrally distinct Three investigative groups have reported fluorochromes. Using the precise ratio of proteomic diagnosis of acute renal allo- these fluorochromes, 100 bead sets have PROTEOME-WIDE SEARCH AND graft rejection. Clarke et al.41 found that been created, each with its unique color- SELECTION OF BIOMARKERS proteins of 6.5, 6.6, 6.7, 7.1, and 13.4 kD coded signature. Each signature bead is performed well as biomarkers of acute re- Using the arsenal of technological and conjugated to an analyte-specific jection. Proteins in the mass ranges 5270 to bioinformatics tools already discussed, the and combined in a single assay to measure 5550 and 10,530 to 11,000 Da were re- following findings have been reported. up to 100 analytes in up to 96 samples si- ported to be good biomarkers by Schaub et multaneously. The assays are based on the al.,42,43 and a subsequent report identified conventional two-site sandwich method. Nephropathies ␤ In a study of 57 control urine samples these proteins as -2 microglobulin and its After conjugation reaction, a mixture of fragments. O’Riordan et al.44 identified beads is analyzed using a dual wavelength compared with samples from patients ϭ urine proteins with masses of 4756.3, laser flow cytometer–like apparatus. One with minimal-change disease (n 16), membranous nephropathy (n ϭ 18), and 25,665.7, and 19,018.8 Da as candidate laser beam detects color-coded beads, and FSGS (n ϭ 10), a group of 690 polypep- markers of acute kidney transplant rejec- another quantifies the reporter signal on tides were present in Ͼ50% of all normal tion compared with recipients with stable each bead. These techniques may be better samples. Plots of Ͼ500 polypeptides typ- transplants. Multiple protein peaks pro- suited for a more targeted analysis of iso- ical of each disease were compiled. The vided a more accurate assessment than lated proteins. rates of correct classification were 71.4% relying on only single biomarkers. In a for minimal-change disease/FSGS and more recent study, O’Riordan et al.45 Bioinformatics 92.9% for membranous nephropathy.36 chemically identified ␤1-defensin and Several databases are used to identify Woroniecki et al.38 studied steroid-re- anti-chymotrypsin as valuable candidate protein fragments by peptide mass fin- sistant nephrotic syndrome in a pediatric biomarkers of acute rejection. We also gerprinting after trypsin digestion or af- population with idiopathic nephritic obtained preliminary data suggesting ter MS/MS sequencing. Databases com- syndrome. A protein of mass 4144 Da that chronic allograft nephropathy is as- pare the size of the fragments recorded was identified as the single most impor- sociated with the increased degradation by the mass spectrometer with the trans- tant marker for distinguishing steroid- of perlecan and urinary excretion of its lated DNA sequence; search for trypsin sensitive and steroid-resistant patients digestion sites, comparing the theoretical fragment endorepellin (unpublished ob- with a high level of confidence. servations, M.S.G. and E.O.). with the measured values; and calculate Distinct polypeptide signatures also the probability of a correct match. Logis- seem to be associated with IgA nephrop- tical analysis of the findings is conducted Diabetic Nephropathy athy. In a study of 45 patients—includ- When 29 healthy individuals were com- using the following: ing those whose total urinary protein lev- pared with 112 patients with type 2 dia- els were within normal ranges39—the 1. Tree-based technologies such as RandomForest: betes, a distinct “no albuminuria/dia- urinary peptide patterns had a sensitivity As the samples are classified according to the pro- betic pattern” was detected. Another teins that most accurately classify the whole popu- of 100% and a specificity of 90%. Three distinct pattern was observed in patients lation, an inverted tree is formed. of the most promising polypeptides were Ͼ sequenced and shown to be albumin with albuminuria 100 mg/L. Found in 2. Adaboost: A method for combining weak classifi- fragments. IgA nephropathy could be 35% of patients with elevated urinary al- ers to create a summary and stronger classifier: The differentiated from membranous ne- bumin excretion rates and only 4% of basic principle is that after selection of the variable healthy volunteers, this pattern identi- that is most likely to predict correctly the class of a phropathy with a sensitivity of 77% and a sample, the samples are reweighted with increas- specificity of 100% and from minimal- fied individuals who were more likely to ing weight applied to the misclassified samples; the change disease, FSGS, and diabetic ne- have retinopathy. The characteristic next best classifier variable is then selected. This phropathy with a sensitivity and a speci- polypeptides were insulin-like peptide 3, process is then repeated with summation of the ficity of 100%. uromodulin, and an albumin frag- classifiers to create a robust and accurate classifier. Cutillas et al.40 applied three different ment.46 The putative role of nephrinuria 3. Genetic algorithms, neural networks, and unified techniques to examine the Dent disease as a potential biomarker of diabetic ne- maximum separability analysis: These are alterna- proteome both qualitatively and quanti- phropathy has been mentioned.

2236 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 2233–2239, 2007 www.jasn.org BRIEF REVIEW

Renal Cancer FROM DEDUCING PROTEIN tide/protein component of a pathway in Rogers et al.47 analyzed urinary proteome IDENTITY TO INTEGRATING THE the excretory compartment, this analysis in patients with clear cell renal carcinoma FINDINGS INTO THE becomes even more complicated. For in- and compared it with that of healthy vol- PATHOGENETIC PATHWAY: A stance, we recently identified a group of unteers and patients with other urogenital LONG WAY TO GO matricellular proteins—all products of diseases. In another study, kininogen levels the enzymatic cleavage of larger mole- were found to be elevated in the urine of a Let’s consider an optimistic scenario: cules—that appear in the urine of patients patient with renal cancer; the concentra- With multiple databases developed for with CAN; importantly, the same enzyme tion fell after nephrectomy.7 Urinary pro- assistance in chemical identification of is responsible for the activation of the la- teomic analysis has identified several bi- proteins, the investigator, after complet- tent TGF-␤ (unpublished observation, omarkers of bladder cancer: ␥-Synuclein, a ing cross-sectional and prospective anal- M.S.G. and E.O.), thus potentially linking soluble isoform of catechol-O-methyl- yses, eventually confirms the diagnostic several pathways leading to fibrosis. Inte- transferase, and calreticulin, which, when value of an ensemble of biomarker pep- gration of proteomic findings into patho- tested prospectively, were found to have a tides/proteins. As important as it is by genetically rational pathways requires sub- combined sensitivity of 76.8% and a spec- itself, the actual understanding of the de- stantial investiture. Integration can be ificity of 77.4%.48 fined markers can be attained only when assisted by the analysis of cDNA microar- the mechanisms of their appearance/ rays (e.g., detection of increased expression disappearance in the urine become eluci- of defensin and anti-chymotrypsin mRNA DYNAMICS OF DISEASE AND dated. Reconstruction of molecular corroborated the findings made in acute METAMORPHOSIS OF MARKERS pathways that are involved in these pro- renal allograft rejection and highlighted It would be unrealistic to expect that cesses and their integration into systems the importance of pro- and anti-inflam- each disease process is identifiable by biology are aided in silico by the growing matory pathways).45 Tissue analysis of unique signature proteins in the urine, collection of publications dealing with proteins in question could be best met us- which have been obtained during a pre- components of individual pathways, as ing immunohistochemical analyses of vious validating snapshot analysis. Dif- well as by the emergence of several soft- multiple intermediates and/or through the ferent stages of disease, variations in ware suits for pathway analysis, yet the recently described MS analysis of tissue mechanisms, and other comorbidities connectivity of diverse pathways sections. The latter permits the acquisition are likely to modify urine proteome. An through multiple components of an in- of protein profiles from individual example of the dynamics among signa- dividual pathway results in a complex ar- nephron segments under microscopic ture proteins in the urine is presented in borized structure that requires actual guidance,53 an excellent example of fusion Figure 2, where the results of proteomic analysis of multiple components to pro- between histology and proteomics. analyses are plotted against the Banff-de- file the correct one, as illustrated in Fig- fined stages of CAN. Knowledge of pro- ure 3. Taking into account that urinary tein dynamics in the course of disease proteins reflect not only the pathway(s) FUTURE DIAGNOSTIC processes may be helpful in their staging but also the mechanism(s) of the appear- ALGORITHMS and in monitoring response to therapy. ance or disappearance of a certain pep- This brief overview was intended to pros- elytize for proteomics approaches to re- nal disease and intentionally overlook the abounding problems—these are ad- dressed in several recent reviews.49–52 In short, many technical and bioinformat- ics issues await resolution, yet initial im- portant findings are emerging, and the whole field of inquiry is undergoing ex- ponential growth. Through accumula- tion of candidate biomarkers and their validation in larger patient populations, various signature combinations of pro- teins and their fragments should eventu- ally become available for many diseases Figure 2. Exemplary proteins identified by mass:charge ratio show the dynamic patterns of urinary excretion in patients with differing Banff grades of chronic vascular injury in and their stages. Studies of the urinary chronic allograft nephropathy. Note that individual peptides show a relative grade proteome are inseparable from in-depth dependence in the level of urinary expression. morphologic analysis of the kidney, pro- teomic mapping of different nephron

J Am Soc Nephrol 18: 2233–2239, 2007 Urine Proteome in Health and Disease 2237 BRIEF REVIEW www.jasn.org

A betic nephropathy of type 1 diabetes. Dia- betes 52: 2969–2974, 2003 10. Aaltonen P, Luimula P, Astrom E, Palmen T, Gronholm T, Palojoki E, Jaakkola I, Ahola H, Tikkanen I, Holthofer H: Changes in the ex- pression of nephrin gene and protein in ex- B perimental diabetic nephropathy. Lab Invest 81: 1185–1190, 2001 11. Gekle M: Renal tubule albumin transport. Annu Rev Physiol 67: 573–594, 2005 12. Osicka TM, Panagiotopoulos S, Jerums G, C Comper W: Fractional clearance of albumin is influenced by its degradation during renal passage. Clin Sci (Lond) 93: 557–564, 1997 13. Osicka TM, Houlihan CA, Chan JG, Jerums G, Comper W: Albuminuria in patients with Figure 3. Blueprints for integration of protein/peptide findings into metabolic pathways: type 1 diabetes is directly linked to changes Systems biology approach. Proteins/peptides that belong to diverse pathways (depicted in the lysosome-mediated degradation of as separate planes) may be linked via cross-talk between the pathways (blue lines) or albumin during renal passage. Diabetes 49: 1579–1584, 2000 promoter elements (entries into each pathway), common enzymes (brown semicircles) 14. Burne MJ, Osicka TM, Comper WD: Fractional acting on individual elements from different pathways, or comprising feedbacks between clearance of high molecular weight proteins in pathways (red lines). Proof of such links may require additional targeted studies to conscious rats using a continuous infusion reconstitute connectivity, enzymes, or regulatory elements (blue). In addition, tissue method. Kidney Int 55: 261–270, 1999 analysis could unveil whether changes in the urinary proteome are a result of parallel or 15. Ganz T: Defensins in the urinary tract and reciprocal changes in the abundance of biomarker proteins. other tissues. J Infect Dis 183[Suppl 1]: S41– S42, 2001 segments (as has been pioneered by Fogo REFERENCES 16. Biemesderfer D: Regulated intramembrane 53 proteolysis of megalin: Linking urinary pro- and Caprioli’s laboratories ), and intra- tein and gene regulation in proximal tubule? 1. Birn H, Christensen EI: Renal albumin ab- cellular organelles (e.g., mitochondrial Kidney Int 69: 1717–1721, 2006 sorption in physiology and pathology. Kid- 17. Hartmann HG, Braedel HE, Jutzler GA: De- proteome, lysosomal proteome in health ney Int 69: 440–449, 2006 tection of renal tubular lesions after abdom- in various diseases54). With further tech- 2. Russo LM, Bakris GL, Comper WD: Renal inal aortography and selective renal arte- handling of albumin: A critical review of ba- nological developments, more precise riography by quantitative measurements of sic concepts and perspective. Am J Kidney mapping of the urinary proteome in brush-border enzymes in the urine. Nephron Dis 39: 899–919, 2002 39: 95–101, 1985 health and disease, with in-depth analy- 3. Aebersold R, Mann M: Mass spectrometry- 18. Chew SL, Lins RL, Daelemans R, Nuyts G, De- based proteomics. Nature 422: 198–207, sis of the proteome of different nephron Broe M: Urinary enzymes in acute renal failure. 2003 segments, individual cells and intracellu- Nephrol Dial Transplant 8: 507–511, 1993 4. Knepper MA: Proteomics and the kidney. lar compartments, and accumulation of 19. Westhuyzen J, Endre ZH, Reece G, Reith D, J Am Soc Nephrol 13: 1398–1408, 2002 knowledge of posttranslational modifica- Saltissi D, Morgan T: Measurement of tubular 5. Hewitt SM, Dear J, Star RA: Discovery of enzymuria facilitates early detection of acute tions, the ancient art of uroscopy hopefully protein biomarkers for renal diseases. JAm renal impairment in the intensive care unit. will become a modern tool for physiologic, Soc Nephrol 15: 1677–1689, 2004 Nephrol Dial Transplant 18: 543–551, 2003 6. Anderson NL, Polanski M, Pieper R, Gatlin T, pathologic, and diagnostic inquiries. 20. Westhuyzen J, Cochrane AD, Tesar PJ, Mau Tirumalai RS, Conrads TP, Veenstra TD, Ad- T, Fleming S: Effect of preoperative supple- kins JN, Pounds JG, Fagan R, Lobley A: The mentation with alpha-tocopherol and ascor- human plasma proteome: A nonredundant bic acid on myocardial injury in patients un- list developed by combination of four sepa- dergoing cardiac operations. J Thorac ACKNOWLEDGMENTS rate sources. Mol Cell Proteomics 3: 311– Cardiovasc Surg 113: 942–948, 1997 326, 2004 21. Mishra J, Dent C, Tarabishi R, Mitsnefes MM, These studies were supported in part by Na- 7. Pieper R, Gatlin CL, McGrath AM, Makusky Ma Q, Kelly C, Ruff SM, Zahedi K, Shao M, tional Institutes of Health grants DK45462, AJ, Mondal M, Seonarain M, Field E, Schatz Bean J, Mori K, Barasch J, Devarajan P: Neu- CR, Estock MA, Ahmed N, Anderson NG, trophil gelatinase-associated lipocalin (NGAL) DK54602, and DK52783 (M.S.G.) and the Steiner S: Characterization of the human uri- as a biomarker for acute renal injury after car- Westchester Artificial Kidney Foundation. nary proteome: A method for high-resolu- diac surgery. Lancet 365: 1231–1238, 2005 E.O.R. was a recipient of the Kevin J. and Glo- tion display of urinary proteins on two-di- 22. Parikh CR, Abraham E, Ancukiewicz M, Edel- ria B. Kiely National Kidney Foundation Fel- mensional electrophoresis gels with a yield stein C: Urine IL-18 is an early diagnostic of nearly 1400 distinct protein spots. Pro- lowship. marker for acute kidney injury and predicts teomics 4: 1159–1174, 2004 mortality in the intensive care unit. JAmSoc 8. Johnstone DB, Holzman LB: Clinical impact Nephrol 16: 3046–3052, 2005 of research on the podocyte slit diaphragm. 23. Muramatsu Y, Tsujie M, Kohda Y, Pham B, Nat Clin Pract Nephrol 2: 271–282, 2006 Perantoni AO, Zhao H, Jo SK, Yuen PS, DISCLOSURES 9. Patari A, Forsblom C, Havana M, Taipale H, Craig L, Hu X, Star RA: Early detection of None. Groop PH, Holthofer H: Nephrinuria in dia- cysteine rich protein 61 (CYR61, CCN1) in

2238 Journal of the American Society of Nephrology J Am Soc Nephrol 18: 2233–2239, 2007 www.jasn.org BRIEF REVIEW

urine following renal ischemic reperfusion VM: High throughput two-dimensional blue- 45. O’Riordan E, Orlova TN, Podust VN, injury. Kidney Int 62: 1601–1610, 2002 native electrophoresis: A tool for functional Chander PN, Yanagi S, Nakazato M, Hu R, 24. Ichimura T, Bonventre JV, Bailly V, Wei H, proteomics of mitochondria and signaling Butt K, Delaney V, Goligorsky MS: Charac- Hession CA, Cate RL, Sanicola M: Kidney complexes. Proteomics 2: 969–977, 2002 terization of urinary peptide biomarkers of injury molecule-1 (KIM-1), a putative epithe- 35. O’Riordan E, Gross SS, Goligorsky MS: Tech- acute rejection in renal allografts. Am J lial cell adhesion molecule containing a nology insight: Renal proteomics—At the Transplant 7: 930–940, 2007 novel immunoglobulin domain, is up-regu- crossroads between promise and problems. 46. Mischak H, Kaiser T, Walden M, Hillmann M, lated in renal cells after injury. J Biol Chem Nat Clin Pract Nephrol 2: 445–458, 2006 Wittke S, Herrmann A, Knueppel S, Haller H, 273: 4135–4142, 1998 36. Weissinger EM, Wittke S, Kaiser T, Haller H, Fliser D: Proteomic analysis for the assess- 25. Han WK, Bailly V, Abichandani R, Thadhani R, Bartel S, Krebs R, Golovko I, Rupprecht HD, ment of diabetic renal damage in humans. Bonventre J: Kidney injury molecule-1 (KIM-1): Haubitz M, Hecker H, Mischak H, Fliser D: Clin Sci (Lond) 107: 485–495, 2004 A novel biomarker for human renal proximal Proteomic patterns established with capil- 47. Rogers MA, Clarke P, Noble J, Munro NP, tubule injury. Kidney Int 62: 237–244, 2002 lary electrophoresis and mass spectrometry Paul A, Selby PJ, Banks RE: Proteomic pro- 26. Nakamura T, Sugaya T, Node K, Ueda Y, for diagnostic purposes. Kidney Int 65: filing of urinary proteins in renal cancer by Koide H: Urinary excretion of liver-type fatty 2426–2434, 2004 surface enhanced laser desorption ioniza- acid-binding protein in contrast medium-in- 37. Hall DA, Ptacek J, Snyder M: Protein mi- tion and neural-network analysis: Identifica- duced nephropathy. Am J Kidney Dis 47: croarray technology. Mech Ageing Dev 128: tion of key issues affecting potential clinical 439–444, 2006 161–167, 2006 utility. Cancer Res 63: 6971–6983, 2003 27. Kanno K, Sasaki S, Hirata Y, Ishikawa S, 38. Woroniecki RP, Orlova TN, Mendelev N, Fushimi K, Nakanishi S, Bichet DG, Marumo Shatat IF, Hailpern SM, Kaskel FJ, Goligor- 48. Iwaki H, Kageyama S, Isono T, Wakabayashi Y, F: Urinary excretion of aquaporin-2 in pa- sky MS, O’Riordan E: Urinary proteome of Okada Y, Yoshimura K, Terai A, Arai Y, Iwamura tients with diabetes insipidus. N Engl J Med steroid-sensitive and steroid-resistant idio- H, Kawakita M, Yoshiki T: Diagnostic potential in 332: 1540–1545, 1995 pathic nephrotic syndrome of childhood. bladder cancer of a panel of tumor markers (cal- 28. Elliot S, Goldsmith P, Knepper M, Haughey Am J Nephrol 26: 258–267, 2006 reticulin, gamma-synuclein, and catechol-o- M, Olson B: Urinary excretion of aqua- 39. Haubitz M, Wittke S, Weissinger EM, Wal- methyltransferase) identified by proteomic anal- porin-2 in humans: A potential marker of den M, Rupprecht HD, Floege J, Haller H, ysis. Cancer Sci 95: 955–961, 2004 collecting duct responsiveness to vasopres- Mischak H: Urine protein patterns can serve 49. O’Riordan E, Goligorsky MS: Emerging sin. J Am Soc Nephrol 7: 403–409, 1996 as diagnostic tools in patients with IgA ne- studies of the urinary proteome: The end of 29. Valenti G, Laera A, Pace G, Aceto G, Lospalluti phropathy. Kidney Int 67: 2313–2320, 2005 the beginning? Curr Opin Nephrol Hyper- ML, Penza R, Selvaggi FP, Chiozza ML, Svelto 40. Cutillas PR, Chalkley RJ, Hansen KC, Cramer tens 14: 579–585, 2005 M: Urinary aquaporin 2 and calciuria correlate R, Norden AG, Waterfield MD, Burlingame 50. Oh J, Pyo JH, Jo EH, Hwang SI, Kang SC, with the severity of enuresis in children. JAm AL, Unwin RJ: The urinary proteome in Fan- Jung JH, Park EK, Kim SY, Choi JY, Lim J: Soc Nephrol 11: 1873–1881, 2000 coni syndrome implies specificity in the re- Establishment of a near-standard two-di- 30. Murer L, Addabbo F, Carmosino M, Procino absorption of proteins by renal proximal tu- mensional human urine proteomic map. Pro- G, Tamma G, Montini G, Rigamonti W, Zuc- bule cells. Am J Physiol Renal Physiol 287: teomics 4: 3485–3497, 2004 chetta P, Della Vella M, Venturini A, Zacch- F353–F364, 2004 51. Smith G, Barratt D, Rowlinson R, Nickson J, ello G, Svelto M, Valenti G: Selective de- 41. Clarke W, Silverman BC, Zhang Z, Chan DW, Tonge R: Development of a high-through- crease in urinary aquaporin 2 and increase in Klein AS, Molmenti EP: Characterization of put method for preparing human urine for prostaglandin E2 excretion is associated renal allograft rejection by urinary proteomic two-dimensional electrophoresis. Proteom- with postobstructive polyuria in human con- analysis. Ann Surg 237: 660–664, discussion ics 5: 2315–2318, 2005 genital . J Am Soc Nephrol 664–665, 2003 52. Spahr CS, Davis MT, McGinley MD, Robin- 15: 2705–2712, 2004 42. Schaub S, Rush D, Wilkins J, Gibson IW, son JH, Bures EJ, Beierle J, Mort J, 31. Martin PY, Abraham WT, Lieming X, Olson Weiler T, Sangster K, Nicolle L, Karpinski M, Courchesne PL, Chen K, Wahl RC, Yu W, BR, Oren RM, Ohara M, Schrier RW: Selec- Jeffery J, Nickerson P: Proteomic-based de- Luethy R, Patterson SD: Towards defining tive V2-receptor vasopressin antagonism tection of urine proteins associated with the urinary proteome using liquid chroma- decreases urinary aquaporin-2 excretion in acute renal allograft rejection. JAmSoc tography-tandem mass spectrometry. I. Pro- patients with chronic heart failure. JAmSoc Nephrol 15: 219–227, 2004 filing an unfractionated tryptic digest. Pro- Nephrol 10: 2165–2170, 1999 43. Schaub S, Wilkins JA, Antonovici M, Krokhin teomics 1: 93–107, 2001 32. Pisitkun T, Shen RF, Knepper MA: Identifica- O, Weiler T, Rush D, Nickerson P: Pro- 53. Xu BJ, Shyr Y, Liang X, Ma LJ, Donnert EM, tion and proteomic profiling of exosomes in teomic-based identification of cleaved uri- human urine. Proc Natl Acad Sci U S A 101: nary beta2-microglobulin as a potential Roberts JD, Zhang X, Kon V, Brown NJ, 13368–13373, 2004 marker for acute tubular injury in renal allo- Caprioli RM, Fogo AB: Proteomic patterns 33. Zhou H, Pisitkun T, Aponte A, Yuen PS, Hof- grafts. Am J Transplant 5: 729–738, 2005 and prediction of glomerulosclerosis and its fert JD, Yasuda H, Hu X, Chawla L, Shen RF, 44. O’Riordan E, Orlova TN, Mei JJ, Butt K, mechanisms. J Am Soc Nephrol 16: 2967– Knepper MA, Star RA: Exosomal fetuin-A Chander PM, Rahman S, Mya M, Hu R, Mo- 2975, 2005 identified by proteomics: A novel urinary bi- min J, Eng EW, Hampel DJ, Hartman B, 54. Bagshaw RD, Mahuran DJ, Callahan JW: A omarker for detecting acute kidney injury. Kretzler M, Delaney V, Goligorsky MS: Bioin- proteomic analysis of lysosomal integral Kidney Int 70: 1847–1857, 2006 formatic analysis of the urine proteome of membrane proteins reveals the diverse com- 34. Brookes PS, Pinner A, Ramachandran A, acute allograft rejection. J Am Soc Nephrol position of the organelle. Mol Cell Proteom- Coward L, Barnes S, Kim H, Darley-Usmar 15: 3240–3248, 2004 ics 4: 133–143, 2005

J Am Soc Nephrol 18: 2233–2239, 2007 Urine Proteome in Health and Disease 2239