| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
INVITED REVIEW
Medical University of South Carolina and Ralph H. Johnson Veterans Affairs Medical Center, Charleston, South Carolina
Submitted 6 November 2006 ; accepted in final form 6 November 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT GRANTSREFERENCES |
|---|
two-dimensional gel electrophoresis; DIGE; mass spectrometry; nephron; glomerulus; renal tubule
|
|
Proteomic techniques have been used to examine protein expression in the entire kidney or to compare protein expression between different regions of the kidney. Several reference maps of protein expression in the kidney have been published. Magni and colleagues (84) separated human renal cortex by 2DE and identified 89 proteins and 74 isoforms by peptide mass fingerprinting. Bovine kidney proteins were separated by 2DE, and 85 protein gene products were identified in 170 spots (28). Two studies have compared expression of proteins between renal cortex and medulla in rat by 2DE. Witzmann and colleagues (134) showed differences in 127 protein spots between the cytosolic fraction of cortex and medulla and identified 20 of the proteins. We have published a reference map of whole cell lysates containing rat renal cortical and medullary proteins with 72 proteins, of which 16 were differentially expressed (9). Interestingly, only two proteins were found to be changed in both studies.
Differential Expression in Whole Kidney or Kidney Cortex
One of the more powerful capabilities of proteomic techniques is the ability to compare protein abundance between tissues or conditions. This allows comparison of many proteins simultaneously without prior knowledge of what proteins might be different between conditions. This approach allows unexpected hypotheses to be formed and tested. An example is an episodic hypoxia model of obstructive sleep apnea in rats. Proteomic analysis demonstrated that alterations in the kallikrein-kallistatin pathway in the kidney were associated with hypertension in this model. The hypothesis was tested and confirmed by demonstrating that rats overexpressing human kallikrein were protected from hypertension induced by episodic hypoxia. (112). Similarly, 41 proteins were found to be differentially expressed between diabetic and normal rat kidney homogenates by 2DE (110). The pattern of protein changes suggested that elastin would be increased (decreased elastase and increased elastase inhibitor). The hypothesis was confirmed in diabetic mice and humans by immunohistochemistry. Changes in expression of renal proteins associated with hypokalemic nephropathy have also been described (111). The changes consisted of both upregulation and downregulation of a number of metabolic enzymes, signaling proteins and cytoskeletal proteins.
Proteomic approaches can also be used to demonstrate differences in posttranslational modifications. Differences in protein expression in renal cortical mitochondria in rats with streptozotocin-induced diabetes were shown to be caused by methylglyoxal-induced modification of proteins belonging to the oxidative phosphorylation and fatty acid 
-oxidation pathways (97). Protein-protein interactions can be determined under the appropriate conditions. Interacting proteins can be isolated by creating constructs that contain the protein of interest attached to a molecular handle to isolate the proteins. A typical handle is glutathione S-transferase (GST). The GST-fusion protein is incubated with a soluble lysate, and the glutathione attached to a bead is used to separate a protein complex. This technique has been used in NRK cells to identify proteins that interact with the gap junction protein connexin-43 (Cx43). A fusion protein of GST with the COOH-terminal tail of Cx43 was used as bait, and tandem MS was done on the isolated proteins to identify 19 interacting proteins (105). The identified proteins included kinases, phosphatases, a membrane receptor, cell signaling molecules, and scaffolding proteins. Proteomic approaches can be used to identify differences in protein expression in genetic models of diseases. The renal proteome of a mouse model of autosomal recessive polycystic kidney disease (jck) was examined by 2DE. Galectin, sorcin, and vimentin were increased in the kidneys of mice with polycystic kidney disease compared with their wild-type control littermates. The authors propose that these differences may be related to the signaling and structural processes in the primary cilium (121). In another study, postmitochondrial fractions (membrane fraction after mitochondria are removed) from young and old rat kidneys have been compared by 2DE, and 167 proteins were differentially expressed. Among the proteins increased with aging were a number of antioxidative and proteolytic proteins (59). A different approach to the measurement of multiple proteins simultaneously was used by Knepper and colleagues. Using a collection of antibodies in a technique termed targeted proteomics (31, 32, 65), differential expression of renal transport proteins was measured in diabetic rats (14) and in ANG II, Na+/H+ exchanger isoform 3, and Na+-Cl– cotransporter knockout mice (17, 18). This approach has also been used to show that the Na+-Cl– cotransporter is downregulated during aldosterone escape (125). This technique is similar to antibody arrays but on a smaller scale.
Since proteins are the result of transcription and translation, it seems logical that changes in abundance of proteins should parallel that of mRNAs coding for the protein. Previous studies in yeast and liver have not shown this to be true (5, 42). Recently, however, a comparison of transcriptomics and proteomics in mouse kidney and liver revealed good correlation between transcription and translation among the majority of high-abundance proteins (77). Proteins which do not have a good correlation may have alterations in mRNA stability, translation, trafficking, or degradation. The degree of correlation between mRNA and protein levels is most likely dependent on the tissue, the gene, and the setting.
Effect of Toxins on Kidney Protein Expression
Proteomic approaches are useful for determining effects of toxicity on tissue protein expression. Changes in protein expression can be used to determine mechanisms or to identify markers of toxicity that can be used in screening new compounds. Protein markers of toxicity could dramatically increase the speed and sensitivity of screening compounds. The use of proteomics in toxicological studies has been reviewed (58, 76). The effect of gentamicin treatment on renal cortical protein expression has been studied by 2DE. The expression of over 20 proteins was altered. Many of the altered proteins were involved in energy production (21). The effect of lead exposure on renal protein expression has also been studied using proteomics. Lead toxicity may be present in as many as 56% of patients with chronic renal failure of unclear etiology, hypertension, and/or gout (99). Lead exposure altered the expression on 2D gels of 76 proteins in the renal cortex and 13 in the medulla (135). Lead exposure alters both the abundance and the isoform expression of the detoxification enzyme GST in a dose-dependent manner (56, 133). These findings could lead to markers of lead toxicity which could aid in the diagnosis of this common but underdiagnosed disease. Several other studies have used 2DE to identify changes in renal protein expression caused by toxins including perfluoro-n-octanoic acid and perfluoro-n-decanoic acid (136), jet fuel (131, 132), and fluoride (137). These studies demonstrate the potential for proteomic analysis to improve an understanding of the mechanisms of toxicity and to identify potential tissue biomarkers. These markers can be used to screen for toxicity among new agents and identify toxicity in patients with early renal manifestations.
Proximal Tubule
Fewer studies have used proteomic tools to analyze specific nephron segments in the kidney. Not surprisingly, many proteomic investigations that focus on kidney tubule function rely on cell culture models or brush-border membrane (BBM) vesicles to identify protein components involved in normal and pathophysiological conditions. Protein maps for the human proximal tubule cell line (HK-2) have been established by 2DE for chaperone proteins (79). In the same cell line, high-abundance hypothetical proteins for amino acid transport, cell cycle regulation (MAD1-homolog), cell cycle arrest/proliferation (p38–2g4 homolog), inorganic phosphatase, and oxidoreductase were identified by 2DE (1, 85).
More recent studies have focused on the identification of protein components in subcellular compartments of the proximal tubule. An intricate survey of abundant proteins from BBM vesicles or basolateral plasma membrane vesicles was identified from rat renal cortex by either free-flow electrophoresis (FFE) or magnesium precipitation followed by SDS-PAGE or strong cation exchange chromatography (SCX) and LC/MS/MS (27). The number of protein identifications from membrane vesicles separated by FFE was greater than from vesicles prepared by magnesium centrifugation, but both techniques resulted in the identification of several hundred proteins from both compartments. As expected, this technique found numerous proteins involved in cell signaling as well as several proteins for which expression has yet to be reported in the kidney. These studies demonstrate the utility of combining prefractionation techniques with proteomic methodologies to look at a specific component.
A less comprehensive proteomic approach was taken to identify proteins that redistribute in cortical apical membrane vesicles following administration of the angiotensin-converting enzyme inhibitor captopril (74). Separation of microvilli proteins by 1D SDS-PAGE showed a reduction in eight major membrane-associated proteins implicated in the reduction of sodium and fluid reabsorption. The extent to which protein abundances differed in rat cortical microvilli, compared with control animals, were large enough to resolve by 1D SDS-PAGE, making this a powerful technique for quickly discovering proteins involved in vesicular trafficking following drug administration. Although the number of proteins is much less than that found by Cutillas et al. (27), the straightforward methods of Leong et al. (74) are more adaptable to a physiology lab.
Because the proximal tubule is a major site of phosphate reabsorption in the kidney, several studies have attempted to identify proteins involved in proximal tubule phosphate handling. The effect of dietary phosphate restriction was examined in isolated proximal tubules from rats by 2DE (22). A low-phosphate diet resulted in an overall reduction in protein diversity as estimated by a decrease in the number of spots on the gel compared with rats given a normal-phosphate diet. Six spots representing differentially expressed proteins were found, most of which could be implicated in cell differentiation and/or actin remodeling. An earlier study found that long-term exposure of cultured opossum kidney proximal tubule cells to parathyroid hormone did not appear to cause a reduction in protein number but did result in the transient phosphorylation of five proteins determined by radiolabeled phosphate incorporation and subsequent separation by 2DE (96). Interestingly, 2DE of proximal tubule BBM from mice with X-linked hypophosphatemic rickets (HYP) showed the expression of a 56-kDa protein that was not found in normal mice (37). Although this protein was not identified, it has been suggested that the differential protein was an incomplete degradation product resulting from the defective zinc metalloproteinase PHEX (98).
Several proteomic strategies have been used to screen proteins involved in cytoprotection following exposure to stressors in proximal tubule cell lines. Primary rat kidney proximal tubule cells exposed to either gallium chloride or sodium arsenite resulted in the induction of several proteins ranging between 28 and 85 kDa, as resolved by 2DE and visualized by [35S]methionine labeling (6). Similar to the study of HYP mice (mentioned above), only protein spot abundance and size were reported, since these studies were conducted before the modern proteomic era.
More recent studies have utilized modern techniques to investigate the cytoprotective effect of the prostaglandin E2 analog DDM-PGE2 against oxidative stress induced by 2,3,5-tris-(glutathione-S-yl)hydroquinone, a reactive oxygen species-generating compound. 1D SDS-PAGE followed by ESI/MS/MS or matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) MS was utilized to identify downstream effectors of prostaglandin signaling in cultured porcine renal proximal tubule cells (LLC-PK1) labeled with [35S]methionine and treated with DDM-PGE2 (88, 117). Peptide sequencing of upregulated proteins led to the identification of glucose-regulated protein/BiP (GRP78), an endoplasmic reticular protein related to the heat shock protein 70 class of proteins. Interestingly, targeted knockdown of this protein by transfection of an antisense GRP78 plasmid reduced cytoprotection in these cells (52). Subsequent analysis by 2DE/MALDI-TOF of control and antisense GRP78 cells implicated three additional proteins in the GRP78 regulatory axis; retinol-binding protein, myosin light chain, and heat shock protein 27.
In a recent study, metabolic acidosis adaptive response proteins from isolated rat proximal tubules were identified using DIGE (25). In addition to the finding of enzymes involved in glutamine metabolism (glutaminase, glutamate dehydrogenase, and phosphoenolpyruvate carboxykinase) that lead to HCO3– and NH4+ production, a total of 23 enzymes were found to change during the adaptive response to metabolic acidosis. Interestingly, time series analysis by Western blotting revealed a set of proteins that follows changes exhibited by glutaminase and glutamate dehydrogenase, most of which are encoded by mRNAs that contain a putative pH-responsive element shared with the glutaminase mRNA, which in turn serves to stabilize the RNA. Given the correlation between changes in protein abundance and mRNA-response elements, the authors suggested that mRNA stability may play a major role in regulating protein abundance in the proximal tubule during metabolic acidosis.
Thick Ascending Limb of Henle
The thick ascending limb of Henle (TALH) and early distal convoluted tubule (DCT) are the primary sites of urinary dilution, making this region of the nephron particularly interesting for those studying cell-signaling mechanisms of loop and thiazide diuretics. Unfortunately, cell lines derived from the loop of Henle are not widely available nor can they be easily procured from commercial sources at present; therefore, proteomic studies relevant to this segment are lacking compared with other parts of the tubule. Despite this difficulty, some cell lines that can withstand osmolar stress have been propagated from the outer medulla and appear to represent those derived from the TALH. Proteomic analysis of a TALH cell line by 2DE identified 40 differentially expressed proteins that change expression following osmotic stress (30). Hypertonic stress resulted in the upregulation of sorbitol pathway enzymes, cytoskeletal proteins, and several heat shock proteins, whereas calcium-sequestering proteins were found to be downregulated, possibly implicating an important role for intracellular calcium homeostasis in hypertonic adaptation. A similar methodology could potentially be used to implicate proteins involved in the diuretic responses or models of abnormal TALH cation transport like paracellin or Na+-K+-Cl– cotransporter knockout mice.
DCT
The DCT has been studied using the Madin-Darby canine kidney (MDCK) cell model. Although an immortalized mouse DCT cell line is available which expresses many DCT-specific proteins (92), it has not been used for proteomic studies. Although the MDCK cell line is considered heterogeneous, the ability of these cells to form tight junctions and retain a transporting phenotype of the late distal tubule makes this cell line a good choice for distal tubule research. There appear to be numerous cell subtypes that exist within this line (7), and the selection of subtypes could be identified using lectins. Using lectin affinity columns and LC/MS, the MDCK cell surface glycoprotein fibronectin was identified as the target of peanut agglutinin (a selector of intercalated cells) (93).
Proteomic studies have largely utilized the MDCK cell line to focus on elucidating mechanisms of protein trafficking. Detergent-insoluble fractions were compared by 2DE from MDCK cells and revealed differences in protein composition which inferred the existence of an apical/basolateral "sorting platform" in epithelial cells (35). Further studies identified a number of small GTP-binding proteins by 2DE and [32P]GTP overlay and immunoblotting (50). One of the GTP-binding proteins localized Rab8 to the basolateral plasma membrane, suggesting a role for this protein in vesicle trafficking (49). A more detailed analysis of the proteins involved in trans-Golgi network vesicular trafficking was conducted in MDCK cells using 2DE and peptide sequencing or immunoblotting (34) and showed that members of the 14–3-3, SNARE, and two -subunits of G proteins exist in this complex. The importance of endosome sorting in MDCK cells was made evident when kRas-expressing MDCK cells failed to polarize (48). Further analysis by 2DE revealed differences in the endosomal protein pattern of expression that probably affected sorting. Since protein identification from many of these studies was done primarily by immunoblotting or GTP overlay, it would be interesting to repeat these experiments using current methods of protein identification to obtain a more comprehensive picture of trafficking in these cells.
Inner Medullary Collecting Duct
The inner medullary collecting duct (IMCD) is the final segment where urine concentration is determined before exiting the kidney and is the major site involved in vasopressin-mediated water and urea reabsorption. Of all the nephron segments, the IMCD appears to have the best foundation from which to establish a proteomic strategy for investigating protein expression. The groundwork for IMCD proteomics includes a database for collecting duct proteins (72) and methods for enriching IMCD cells from rat inner medulla that have been characterized by 2DE and peptide mass fingerprinting (45). Differential 2DE (DIGE) revealed several high-abundance proteins that appear to be differentially expressed in IMCD cells vs. non-IMCD cells residing in the inner medulla, some of which may be useful as selection markers for the IMCD subpopulation (45).
Because water permeability in the IMCD is mediated through vasopressin-induced trafficking of aquaporin-2 (AQP2) to the apical membrane, an LC/MS/MS approach was used to identify protein components of immunopurified AQP2 vesicles (13). Interestingly, this approach was successful in identifying numerous SNARE, Rab GTPase, cytoskeletal, and membrane transport proteins and also revealed proteins associated with specific subcellular compartments, making this a nice technique for identifying protein/organelle associations. Vasopressin signaling and intracellular trafficking in the IMCD are largely mediated through the phosphorylation of target proteins, including AQP2. High-throughput techniques that can identify and measure changes in the phosphorylation state of the phosphoproteome can serve to highlight proteins targeted during vasopressin administration or other mediators of IMCD trafficking. One of the major obstacles in phosphoproteomics is enrichment and detection of phosphorylated amino acid residues, due to the fact that phosphorylated residues are often transient and constitute a small subset of the protein population. Both of these obstacles can be overcome using metal affinity columns that preferentially bind phosphopeptides (IMAC columns) and sophisticated LC/MSn techniques. The recent study by Hoffert et al. (44) is an elegant demonstration of these techniques using IMCD preparations. The authors identified over 220 phosphoproteins in the IMCD and, using a novel MS quantification strategy called QUOIL (124), were able to estimate changes in relative abundance of specific phosphopeptides following vasopressin administration. This approach led to the discovery of several new phosphorylation sites for AQP2, AQP4, and the urea transporters UT-A1 and UT-A3, further emphasizing the power of this technique over traditional procedures for identifying phosphorylation sites.
DIGE has been particularly useful in identifying proteins that are involved in the adaptation of chronic vasopressin administration in the IMCD, where the effects of vasopressin are decreased over time to prevent excessive plasma dilution. Chronic vasopressin administration in Brattleboro rats resulted in the altered expression of 43 proteins (122). Some of the more interesting proteins identified were involved in nitric oxide synthesis or have been implicated in vasopressin escape. Since 2DE is limited in its ability to detect low-abundance proteins, the investigators used biochemical pathway analysis to implicate proteins involved in vasopressin escape that could not be seen with DIGE (46). Pathway analysis is a bioinformatics software tool provided by Ingenuity Systems that allows the researcher to construct complex protein regulatory networks from a small subset of protein expression information. From this network, one can better understand biological relationships from a limited dataset. Although there are limitations to computer-generated biological relationships, Hoorn et al. (46) successfully utilized this approach, coupled with immunoblotting, to discover changes in IMCD protein expression for low-abundance proteins, which included several transcription factors, e.g., c-myc, c-fos, c-jun, and p53.
With established IMCD isolation and proteomic techniques in place, the field of IMCD proteomics appears to be one of the most promising for elucidating mechanisms of water and sodium regulation in the kidney tubule. Future studies that focus on identifying proteins involved in sodium regulation within the IMCD by mineralocorticoids and peptide hormones such as atrial natriuretic peptide, endothelin, interleukin-1, and urodilatin will allow a better understanding of the interrelatedness of water and sodium transport in the collecting duct.
Glomerular Proteins
The ability to isolate glomeruli using sieving techniques provides a powerful tool to identify proteins that change in glomeruli in disease. A database of human glomerular proteins has been published (138). The database is available on the World Wide Web (http://www.sw.nec.co.jp/bio/rd/hgldb/index.html) and currently contains 350 identified proteins and the location of the protein spots on a 2D gel image. This is a valuable resource both as a list of proteins found in the glomerulus and as a reference map to aid in identification of proteins. In addition to identifying proteins that are present or that change with an intervention, proteomic tools can be used to identify protein interactions by precipitating a protein and identifying proteins that are associated with it in the lysate. When the identity of the associated protein is suspected, Western blotting of lysates can identify the protein. When it is not known, proteomic methodologies can be used to identify these proteins. Ahola and colleagues (2) precipitated nephrin, a key component of the glomerular filtration barrier, from lysates of sieved glomeruli and analyzed proteins by MALDI-TOF MS. Densin was identified as a protein that colocalizes with nephrin. Immunoelectron microscopy demonstrated that densin is present in the slit diaphragm (2). A similar approach with a fusion protein of the GST-nephrin tail was used to show that nephrin forms a multiprotein complex with cadherins, p120 catenin, and the scaffolding proteins ZO-1, CD2AP, and CASK in glomeruli (73). Renal afferent arterioles can also be isolated by sieving and magnetic separation after perfusion with a magnetized iron oxide suspension. Arteriolar proteins were compared by 2DE between normal and hypertensive rats. Fourteen proteins were differentially expressed including troponin T (90). Using a similar ferromagnetic methodology, Sitek et al. (106) were able to detect roughly 2,900 spots by DIGE from isolated glomeruli with as little as 3 µg protein. A comparison between cortical and glomerular proteins revealed 48 differentially expressed proteins between the two tissues. Despite these differences, glomerular-specific proteins (e.g., podocin) were not found. In the same study, laser capture microscopy followed by 2DE yielded 
900 protein spots from 10 human kidney glomeruli. As such, proteomic analysis of glomeruli from human biopsy samples by 2DE could be a real possibility for diagnostics in the future.
Proteomic analysis of cultured glomerular cells has been limited to just a few studies. 2DE has been used to compare protein expression in cultured cells of glomerular origin. The expression of six proteins was changed by glucocorticoids in cultured podocytes (94, 95). Mesangial proteomics have also been studied. Nitric oxide increases expression of Mn-SOD in mesangial cells, as shown by 2DE. These findings were confirmed by immunoblotting and correlated with mRNA expression (57). Studies done in cultured mesangial cells demonstrate the possibilities of combining proteomic technologies with other biochemical analyses. In an interesting approach to identification of kinase substrates in cultured mesangial cells, Barati and colleagues (12) combined an in vitro kinase assay with proteomic analysis. Mesangial cell lysates were incubated with [32P]ATP and the kinase AKT and separated by 2DE. The stained gel was compared with its own autoradiogram, and AKT phosphorylated proteins were identified by peptide mass fingerprinting. The results were confirmed and expanded by Western blotting of mesangial cells separated by 2DE (12). This approach is an example of how a combination of approaches can expand the information obtained. Another useful example of a combination of techniques which allow proteomic methods to provide important functional information about a system was demonstrated by Kuncewicz and coauthors (68) where S-nitrosylated proteins were identified in cultured mesangial cells using a "biotin-switch." In this technique, S-nitrosylated proteins are selectively biotinylated and purified by streptavidin affinity chromatography. The purified S-nitrosylated proteins were then separated by 2DE and identified. Using this approach, 35 S-nitrosylated proteins were identified including proteins involved in signaling and transcription and proteins located in the cytoplasm, cytoskeleton, and membranes (68). Finally, SELDI-TOF MS was used to demonstrate interaction of the mesangial cell-specific protease inhibitor megsin with plasmin (86).
Urological Cancers
Proteomics techniques have been used extensively in the studies of urological cancers. The subject has been previously reviewed (119). We will not review this area in great depth but will use examples of specific techniques that have been used. A number of studies have used 2D gels to compare proteins expressed in tumors to normal tissue (10, 51, 61, 100, 101, 104).
Laser capture microscopy (LCM) is a technique that permits isolation of small areas of tissue under direct visualization (33). LCM has been used extensively to characterize tissues at the DNA or RNA level. It has also been combined with proteomics to allow characterization of proteins in a small subset of cells. Tissue collected by LCM has been separated by 2DE in renal and cervical tumors (11, 24). A difference in protein expression could be seen, and the LCM procedure (when used with appropriate stain and fixation) did not change the appearance of the proteins on the gels or the spectra obtained by MS. LCM of cervical tumors resulted in enrichment of specific proteins, although the authors did not observe any enrichment of proteins after laser dissection of proximal tubule cells compared with renal cortical lysates. The length of time needed to acquire tissue to run a single gel could be prohibitive, however. Four days and 18,000 laser shots were required to obtain sufficient proximal tubule tissue. LCM has great potential for application in proteomic renal research, but it is time and labor intensive. Future advances in image recognition which allow automated collection of specific cell types may facilitate this type of analysis.
An interesting approach to identify tumor-associated antigens has been used in patients with renal carcinoma. It has been called serological proteome analysis. Protein expression was compared between renal cell cancer and the normal kidney by 2DE. Several proteins were found to be differentially expressed. Two of the proteins were used to screen sera from patients with cancer. None of the sera from 13 normal patients and 1/13 sera from patients with colon cancer reacted to the antigens, but 5/11 sera from patients with renal cell carcinoma reacted to one of the proteins (SM22-
) and 3/11 reacted to the other (carbonic anhydrase) (61). These findings suggest that comparison of protein expression may identify target proteins for immunological screening of cancers.
Urine Proteomics
One of the most promising uses of proteomic technologies will be to look for biomarkers of disease in urine. A large number of studies have been performed, and the area has been recently reviewed (115).We will discuss only a few studies to highlight the potential of the area. Urine proteins have been identified by 2DE (4, 19, 89, 113, 114) and by LC/MS (29, 108). The technique of CE/MS is one of the more promising approaches for biomarker discovery. It has been applied to urine to produce patterns of low-molecular-weight proteins and peptides separated in two dimensions: their mobility (by CE) and their size (by MS) (36). The technique has been used to identify diagnostic patterns of proteins associated with acute tubulointerstitial rejection in renal allograft nephropathy (130). Similarly, SELDI has been used to identify patterns of proteins associated with acute rejection (102, 103). The techniques of SELDI and CE/MS have been compared in one report, and the resolution of peaks as well as the pattern of polypeptides seen with CE/MS were much richer (81). 2DE combined with informatic analysis has also been used to identify urinary protein markers that can predict the class of lupus nephritis present (83). This study highlights an important advantage of 2DE in urine biomarker discovery, the ability to differentiate between posttranslationally modified forms of the protein. As can be seen in Fig. 2, multiple charge forms of the same protein can be separated on a 2D gel. In urine, these charge forms are caused by glycosylation and other posttranslational modifications. A very intriguing approach has been advanced by Knepper and colleagues (47, 91) to purify low-density urinary structures called exosomes that contain membrane proteins from several regions of the kidney. An analysis of exosomes in disease states and disease models will provide a rich source of information about changes in protein structure and modifications in disease. Recently, the group from the National Institute of Diabetes and Digestive and Kidney Diseases has described conditions for collecting, storing, and preserving urinary exosomes (140). Following methodologies that will allow the maximal amount of information to be obtained from these specimens is fundamentally important when studies for biomarker discovery are being designed. Analysis of urinary exosomes by DIGE in a cisplatin and an ischemia-reperfusion model of renal injury as well as in intensive care unit patients with acute kidney injury demonstrated an increase in fetuin-A (139). This is an exciting discovery of a protein that may be a diagnostic biomarker of structural injury. Another promising approach is biomolecular interaction analysis MS (BIA/MS), in which surface plasmon resonance sensing is used to detect interactions between surface-immobilized ligands and their solute-born analytes. This technique has been used to detect a subset of urine biomarkers (80).
|
Relatively little is known about proteins that are lost during dialysis. The availability of fluid from dialysis for study and the potential importance of specific protein loss during dialysis make this a potentially fruitful area for further observation. Several groups have looked at proteins that either appear in the dialysis filtrate or are bound to dialysis membranes. Capillary electrophoresis coupled to MS has been used to profile polypeptides in dialysis fluid. Comparisons were made between dialysate from patients on high- and low-flux dialysis membranes (54). The results were that 1,639 different polypeptides (based on size and elution time) were found in the low-flux membrane filtrate and 2,515 in the high-flux membrane filtrate. A relatively small fraction of the peptides were present in multiple samples even within the same group of patients. Polypeptides were smaller in the low-flux than in the high-flux ultrafiltrate, as would be expected. Since this study compared polypeptides by elution time and size only and did not identify any of them, it is of limited utility in the importance of protein loss across the membranes. Two studies have identified proteins in dialysate or ultrafiltrate. These studies can help in an understanding of the importance of protein removal in renal replacement therapy.
A combination of reverse phase chromatography and 2DE was used to identify proteins in ultrafiltrate from patients on continuous venovenous hemofiltration (71). Interestingly, the 47 protein spots identified represented multiple charge forms of only 10 different proteins. The proteins identified in the ultrafiltrate were albumin, apolipoprotein A-IV, 2-microglobulin, lithostathine, mannose-binding lectin-associated serine protease 2-associated protein, plasma retinol-binding protein, transferrin, transthyretin, vitamin D-binding protein, and Zn 2-glycoprotein. In a study of hemodialysate using gel electrophoresis in combination with LC/MS/MS, 292 proteins were identified (78). Surprisingly, over 70% had not been previously identified in serum or plasma. More than one-half of the proteins were smaller than 40 kDa. A number of other approaches have been used to identify proteins associated with chronic renal failure in dialysate (126, 129), serum, and urine (70, 75).
Nephrolithiasis
Although it is a potentially productive area for proteomic studies, relatively little has been done in proteomics in the area of nephro-or ureterolithiasis. Proteins from crushed stones were extracted by microdialysis, separated by 2DE, and identified (15). Urinary biomarkers of nephrolithiasis have been examined by SELDI-TOF (20). Unique proteins have been visualized by 2DE in urine of patients who are recurrent stone formers but have not been identified (39). Proteins extracted from stones have been studied by 2DE. Stone proteins were similar to urine proteins and characteristic of the type of stones present (53). LC/MS has also been used on proteins extracted from stones (55). Recently, an elegant approach was used to identify an inhibitor of crystal formation in urine (23). The investigators used gel filtration chromatography of anionic proteins to identify a fraction of urine which inhibited CaOx crystal formation. The protein was purified by anion exchange chromatography and identified by MALDI-TOF MS as trefoil factor 1. This is an example of the combination of previously used biochemical techniques with proteomic technologies to identify a protein with a specific function.
Summary
Proteomics of the kidney and urine provides a powerful tool for understanding issues in renal physiology and pathophysiology. The use of these tools has increased dramatically over the last 5 years and promises to facilitate studies of renal physiology and pathophysiology. Clever uses of the technologies have dramatically improved their usefulness. As proteomic tools become more widely available, the next 5 years promise to be even more productive. More complete lists of proteins in anatomically distinct regions of the kidney, cultured cells, and organelles will be made as well as determinations of proteins in signaling complexes. These tools will be used to elucidate protein-signaling events by identifying not just the proteins involved in the pathway but also the modifications of these proteins over time. Proteins identified in tissues associated with disease models will lead to a better understanding of glomerular and other renal diseases. The discovery of biomarkers in tissue and fluids will lead to improved diagnosis of renal diseases.
|
GRANTS |
|---|
|
TOPABSTRACTGRANTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTGRANTSREFERENCES |
|---|
) has been proportional to the total number of proteomics-related publications (
). The number of publications was determined by a PubMed search of the word "proteomics" with or without the terms "nephrology OR kidney OR renal OR urine."