Source: Adapted from ref. 5.

Source: Adapted from ref. 5.

• A preference for noninvasive sample collection.

• New biomarkers developed for preclinical use should optimally be translated to the clinic.

• Assays for new biomarkers should be robust and kits readily available for testing.

• Assays should be multiplexed to minimize cost and expedite sample analysis.

• Biomarkers should ideally predict or report out site-specific injury.

• Biomarkers must be more sensitive and specific of kidney injury than existing standards.

• Biomarkers should be predictive (prodromal) of kidney injury in the absence of histopathology.

The preference for noninvasive sample collection made urine the obvious choice of biofluid. Urine has proven to be a fertile substrate for the discovery of promising new biomarkers for the early detection of nephrotoxicity [5]. A number of these markers have been selected for further development and qualification by the ILSI-HESI and C-Path Nephrotoxicity Working Groups in both preclinical and clinical settings, with the exception of RPA-1 and GST Yb1 (Biotrin), which are markers developed specifically for the analysis of kidney effects in rats (Table 2). The utility and limitations of each marker used in the context of early and site-specific detection are discussed below.


Human p2-microglobulin (p2M) was isolated and characterized in 1968 [6]. p2M was identified as a small 11,815-Da protein found on the surface of human cells expressing the major histocompatibility class I molecule [7]. p2M is shed into the circulation as a monomer, from which it is normally filtered by the glomerulus and subsequently reabsorbed and metabolized within proximal tubular cells [8]. Twenty-five years ago, serum p2M was advocated for use as an index of renal function because of an observed proportional increase in serum p2M levels in response to decreased renal function [9]. It has since been abandoned due to a number of factors complicating the interpretation of the findings. More recently, increased levels of intact urinary p2M have been directly linked to impairment of tubular uptake. Additional work in rats and humans has demonstrated that increased urinary levels of p2M can be used as a marker for proximal tubular function when p2M production and glomerular filtration are normal in a setting of minimal proteinuria [10-13].

Urinary p2M has been shown to be superior to N-acetyl-p-glucosaminidase as a marker in predicting prognosis in idiopathic membranous neuropathy [14]. In this context p2M can be used to monitor and avoid unnecessary immunosuppressive therapy following renal transplantation. p2M is being considered for evaluation as an early predictor of proximal tubular injury in preclinical models of drug-induced nephrotoxicity. Although easily detected in urine, there are several factors that may limit its value as a biomarker. For example, p2M is readily degraded by proteolytic enzymes at room temperature and also degrades rapidly in an acidic environment at or below pH < 6.0 [15]. Therefore, great care must be taken to collect urine in an ice-cold, adequately buffered environment with the addition of stabilizers to preserve p2M levels during the period of collection and in storage. It is unlikely that p2M will be used as a stand-alone marker to predict or report proximal tubule injury pre-clinically or in the clinic. Rather, it is likely to be used in conjunction with other proximal tubule markers to support such a finding. A brief survey for commercially available antibodies used to detect p2M indicates that most are species specific ( Instances of cross-reactivity were noted for specific reagents between human and pig, chicken and turkey, and human and other primates. A single monoclonal reagent is reported to have cross-reactivity with bovine, chicken, rabbit, and mouse, and none were listed that specifically recognized dog p2M. Because both commonly used preclinical species rat and dog p2M proteins share only 69.7% and 66.7% amino acid identity with the human protein (, it would be prudent to develop and characterize antibody reagents specific to each species and cross-reacting antisera to specific amino acid sequences shared by all three proteins.


Clusterin is a highly glycosylated and sulfated secreted glycoprotein first isolated from ram rete testes fluid in 1983 [16] . It was named clusterin because of its ability to elicit clustering of Sertoli cells in vitro [17]. Clusterin is found primarily in the epithelial cells of most organs. Tissues with the highest levels of clusterin include testis, epididymus, liver, stomach, and brain. Metabolic and cell - specific functions assigned to clusterin include sperm maturation, cell transformation, complement regulation, lipid transport, secretion, apoptosis, and metastasis [18] . Clusterin is also known by a number of synonyms as a consequence of having been identified simultaneously in many parallel lines of inquiry. Names include glycoprotein III (GPIII), sulfated glycoprotein- 2 (SG-2), apolipoprotein J (apo J), testosterone-repressed message-2 (TRPM-2), complement associated protein SP-40, 40, and complement cytolysis inhibitor protein (see Table 1). Clusterin has been cloned from a number of species, including the rat [19]. The human homolog is 449 amino acids in length, coding for a protein with a molecular weight of 52,495 Da - 20] . However, due to extensive posttranslational modification, the protein migrates to an apparent molecular weight of 70 to 80 kDa following sodium dodecyl sulfate polyacryl-amide gel electrophoresis (SDS-PAGE). Amino acid identity between species is moderate. Human clusterin shares 70.3%, 76.6%, 71.7%, and 77% with the bovine, mouse, pig, and rat homologs, respectively (

Clusterin is a heterodimer comprised of an a and a p subunit, each having an apparent mass of 40 kDa by SDS - PAGE. The subunits result from the proteolytic cleavage of the translated polypeptide at amino acid positions 23 and 277. This eliminates the leader sequence and produces the mature 205-amino acid p subunit and the remaining 221-amino acid a subunit. The a and P subunits are held together by five sulfhydryl bonds afforded by cysteine residues clustered within each of the subunits [21] . In addition, each subunit has three N-linked carbohydrates that are also heavily sulfated, giving rise to the higher apparent molecular weight observed following SDS-PAGE. Considerable evidence has been provided suggesting that clusterin plays an important role in development. For example, clusterin mRNA expression has been observed at 12.5 days ' postgestation in mice, where it is present in all germ cell layers [22] . Furthermore, stage-specific variations of the transcript have been observed, as have changes in specific localization during development. Similarly, changes in the developmental expression of clusterin in kidney, lung, and nervous system have also been reported [23]. These observations suggest that clusterin might play a role in tissue remodeling.

In the developing murine kidney, clusterin is expressed in the tubular epithelium and later in development is diminished as tubular maturation progresses [24] . Interestingly, clusterin is observed in newly formed tubules but appears to be absent in glomeruli. Of interest to many investigators of renal function is the reemergence of clusterin observed following induction of a variety of kidney diseases and drug-induced renal injury. Clusterin induction has been observed following ureteral obstruction [25] and ischemia reperfusion injury [26]. Elevations in the levels of clusterin have also been observed in the peri-lnfarct region following subtotal nephrectomy [27] and in animal models of hereditary polycystic kidney disease [28]. Marked increases of clusterin released in urine have also been recorded in animal models of aminoglycoside-induced nephrotoxicity [29-31].

Authors have opined that clusterin functions in either a protective role by scavenging cell debris or may play a role in the process of tissue remodeling following cellular injury based on these observations. Collectively, the body of work linking elevated levels of urinary clusterin to kidney damage has suggested that measurement of urinary clusterin may be useful as a marker of renal tubular injury. Indeed, an early study comparing urinary levels of clusterin against N-acetyl-P-glucosaminidase (NAG) following chronic administration of gentamicin over a two-month period demonstrated that while the urinary levels of both proteins rose rapidly, peaked, and then declined, clus-terin levels remained significantly higher than control values over the duration of the experiment. By contrast, NAG levels dropped to within control values within 10 days of treatment even though evidence of tubulointerstitial disease persisted [30]. More recent work examining the levels of urinary clusterin in the autosomal-dominant polycystic kidney disease (cy/+) rat model compared to the FHH rat model of focal segmental glomerulosclerosis following bilateral renal ischemia demonstrated that clusterin levels correlated with the severity of tubular damage and suggested use as a marker for differentiating between tubular and glomerular damage [32]. Although the value of clusterin as an early marker of tubular epithelial injury has not yet been established clinically, preclinical findings suggest that it is an ideal candidate for translation to the clinic as an early marker of nephrotoxicity.


Cystatin C (Cys-C) is a 13-kDa nonglycosylated protein belonging to the superfamily of cysteine protease inhibitors [33]. Cys-C is produced by all nucleated cells and, unlike SC, is unaffected by muscle mass. Serum Cys- C was suggested to be closer to the "-deal" biomarker reporting GFR because although freely filtered by the glomerulus, it is not secreted. Instead, Cys-C is adsorbed by tubular epithelial cells, where it is catabolized and is not returned to the bloodstream, thus obviating the need to calculate urinary Cys- C to measure GFR [34]. Several studies have been designed to examine the usefulness of serum Cys- C as a measure or biomarker of GFR [ 35]. In one such study, serum Cys-C was shown to be a useful biomarker of acute renal failure and could be detected one to two days prior to the elevation in levels of SC, the accepted clinical diagnosis of AKI [36]. Although earlier in detection than

SC, serum Cys-C levels were not predictive of kidney disease and, like SC, reported out kidney injury long after serious damage had occurred. In another study, investigators monitored and compared the levels of serum Cys-C and urinary Cys-C in patients following cardiothoracic surgery with and without complicating AKI [ 37]. The results clearly demonstrated that while plasma Cys- C was not a useful predictor of AKI, early and persistent increases in urinary Cys-C correlated with the development and severity of AKI. Another interesting but unexplained observation in this study was that women had significantly higher postoperative levels of urinary Cys-C than did men even though preoperative Cys- C levels were similar. These data have prompted groups like ILSI-HESI and C-Path to examine the utility of urinary Cys-C as a preclinical biomarker of drug-induced renal injury in the hope that elevated levels of Cys - C can be detected in urine prior to the emergence of overt tubular dysfunction.

Glutathione S-Transferases

The glutathione S-transferases (GSTs) form a family of homo-and heterodi-meric detoxifying enzymes [38] identified originally as a group of soluble liver proteins that play a major role in the detoxification of electrophilic compounds [39]. They have since been shown to be products of gene superfamilies [40] and are classified into alpha, mu, pi, and theta subfamilies based on sequence identity and other common properties [41]- Tissue distribution and levels of GST isoform expression has been determined by immunohistochemical localization [42] , isoform - specific peptide antibody Western blotting, and mass spectrometry [40] - Analysis of GST subunit diversity and tissue distribution using peptide --pecific antisera has shown GST | isoforms to be the most widely distributed class of GSTs, with expression evident in brain, pituitary, heart, lung, adrenal gland, kidney, testis, liver, and pancreas, with the highest levels of GST |1 observed in adrenals, testis, and liver. Isoforms of the GSTa subfamily, also known by the synonyms glutathione S - tansferase - 1, glutathione S-transferase Ya-1, GST Ya1, ligandin, GST 1a-1a, GST B, GST 1- 1, and GST A1 - 1 ( ), are more limited in distribution, with highest levels of expression observed in hepatocytes and proximal tubular cells of the kidney [42].

Indeed, proximal tubular GSTa levels have been reported to approximate 2% of total cytosolic protein following exposure to xenobiotics or renal toxins [43]. In the Rowe study [40], GSTa was found to be rather evenly distributed between adrenals, kidney, and pancreas, with highest levels observed in liver, whereas isoforms of the GSTn subclass were expressed in brain, pituitary, heart, liver, kidney, and adrenals, with highest levels of expression observed in kidney. The high levels of expression and differential distribution of GST isoforms made them attractive candidates as biomarkers that could be used to indicate site-specific drug-induced nephrotoxicity. For example, development of a radioimmunoassay to quantify leakage of ligandin (GSTa)

into the urine as a measure of nephrotoxicity in the preclinical rat model was reported as early as 1979 [44] - Subsequent work described the development of a radioimmunoassay for the quantitation of GSTn in the urine [45a] later used as an indicator of distal tubular damage in the human kidney [ 45b] , Additional work described the development of a multiplexed ELISA for the simultaneous quantitation of GSTa and GSTn to discriminate between proximal and distal tubular injury, respectively [46].

In terms of sensitivity, a study examining the nephrotoxic effects of the sevoflurane degradation product,fluoromethyl-2,2-difluoro-1-(trifluoromethyl) vinyl ether, in rats showed urinary GSTa to be the most sensitive marker of mild proximal tubular damage compared to other urinary markers measured, including protein and glucose [47]. A second study in which four human volunteers were given sevoflurane demonstrated abnormalities in urinary glucose, albumin, GSTa, and GSTn, while levels of BUN or SC were unaffected, suggesting that the GSTs were more sensitive markers of site-specific dug-induced nephrotoxicity [48]. Immunohistochemical staining of the rat kidney with antibodies to different GST isoforms has shown that GSTa subunits are expressed selectively in the proximal tubule, whereas GST^ and n subnits are localized to the thin loop of Henle and proximal tubules, respectively [38]. An examination of the distribution of the rat GST^ equivalent, GSTYb1, in the kidney indicates that it is localized to the distal tubules. Simultaneous measurement of urinary GSTa and GSTYb1 has been used to discriminate between drug-induced proximal and distal tubular injury (cited by Kilty et al. [49]). The high levels of GSTs in the kidney and site-specific localization of different GST classes in addition to increased sensitivity in detecting drug-induced nephro-toxicity in humans make them ideal candidates for the development and testing of preclinical markers that predict or report early signs of nephro-toxicity to support preclinical safety studies and subsequent compound development.

Kidney Injury Molecule 1

Rat kidney injury molecule 1 (KIM-1) was discovered as part of an effort to identify genes implicated in kidney injury and repair [50] using the polymerase chain reaction (PCR) subtractive hybridization technique of representational difference analysis originally developed to look at differences in genomic DNA [51] but adapted to examine differences in mRNA expression [52]. Complementary DNA generated from poly(A+) mRNA purified from normal and 48-hour postischemic rat kidneys was amplified to generate driver and tester amplicons, respectively. The amplicons were used as templates to drive the subtractive hybridization process to generate designated differential products, three of which were ultimately gel purified and subcloned into the pUC18 cloning vector. Two of these constructs were used to screen ^ZapII cDNA libraries constructed from 48-hour postischemic rat kidneys. Isolation and purification of positively hybridizing plaques resulted in the recovery of a

2.5-kb clone that contained sequence information on all three designated differential products. A BLAST search of the NCBI database revealed that the rat KIM-1 sequence had limited (59.8%) amino acid homology to HAVcr-1, identified earlier as the monkey gene coding for the hepatitis A virus receptor protein [ 53]. The human homolog of KIM-1 was isolated by low-stringency screening of a human embryonic liver ^gt10 cDNA library using the same probe that yielded the rat clones [50]. The plaque of one of two clones purified from this exercise was shown to code for a 334-amino acid protein sharing 43.8% identity and 59.1% similarity to the rat KIM-1 protein. Comparison to the human HAVcr protein [54] revealed 85.3% identity demonstrating a clear relationship between the two proteins.

Subsequent work has demonstrated that KIM-1 and HAVcr are synonyms for the same protein, also known as T-cell immunoglobulin and mucin domain-containing protein 1 (TIMD-1) and TIM-1. The TIMD proteins are all predicted to be type I membrane proteins that share a characteristic immunoglobulin V, mucin, transmembrane, and cytoplasmic domain structure [55]. It is not clear what the function of KIM-1 (TIMD-1) is, but it is believed that TIMD- 1 is involved in the preferential stimulation of Th2 cells within the immune system [ 56]. In the rat, KIM-1 mRNA expression is highest in liver and barely detected in kidney [50]. KIM-1 mRNA and protein expression are dramatically up-regulated following ischemic injury. Immunohistochemical examination of kidney sections using a rat-specific KIM-1 antibody showed that KIM-1 is localized to regenerating proximal tubule epithelial cells. KIM-1 was proposed as a novel biomarker for human renal proximal tubule injury in a study that demonstrated that KIM-1 could be detected in the urine of patients with biopsy-proven acute tubular necrosis [57]. Human KIM- 1 occurs as two splice variants that are identical with respect to the extracellular domains but differ at the carboxy termini and are differentially distributed throughout tissues [58]. Splice-variant KIM-1b is 25 amino acids longer than the originally identified KIM-1a and is found predominantly in human kidney.

Interestingly, cell lines expressing endogenous KIM-1 or recombinant KIM-1b constitutively shed KIM-1 into the culture medium, and shedding of KIM-1 could be inhibited with metalloprotease inhibitors, suggesting a mechanism for KIM-1 release into the urine following the regeneration of proximal tubule epithelial cells as a consequence of renal injury. Evidence supporting KIM-1 -s potential as a biomarker for general kidney injury and repair was clearly demonstrated in another paper describing the early detection of urinary KIM-1 protein in a rat model of drug-induced renal injury. In this study increases in KIM- 1 were observed before significant increases in SC levels could be detected following injury with folic acid and prior to measurable levels of SC in the case of cisplatin-treated rats [59]. In later, more comprehensive studies examining the sensitivity and specificity of KIM-1 as an early biomarker of mechanically- [60] or drug-induced renal injury [61], KIM-1 was detected earlier than any of the routinely used biomarkers of renal injury, including BUN, SC, urinary NAG, glycosuria, and proteinuria. Certainly, the weight of evidence described above supports the notion that KIM- 1 is an excellent biomarker of AKI and drug-induced renal injury. The increasing availability of antibody-based reagents and platforms to rat and human KIM-1 proteins offer convenient and much needed tools for preclinical safety assessment of drug-induced renal toxicity and for aid in diagnosing or monitoring mild to severe renal injury in the clinic. Further work is required to determine if KIM-1 is a useful marker for long-term injury and whether it can be used in combination with other makers to determine site-specific kidney injury.


The examination of proteins excreted into urine provides useful information about renal function (reviewed in [62]). Tamm-Horsfall proteins that originate from renal tubular cells comprise the largest fraction of protein excreted in normal urine. The appearance of low- molecular- weight urinary proteins normally filtered through the basement membrane of the glomerulus, including insulin, parathormone, lysozyme, trypsinogen and p2-microglobulin indicate some form of tubular damage [63]. The detection of higher-molecular-weight (40- to 150-kDa) urinary proteins not normally filtered by the glomerulus, including albumin, transferrin, IgG, caeruloplasmin, a 1 - acid glycoprotein, and HDL, indicate compromised glomerular function [64]. Albumin is by far the most abundant protein constituent of proteinuria. Although gross increases in urinary albumin measured by the traditional dipstick method with a reference interval of 150 to 300 mg/mL have been used to indicate impairment of renal function, there are many instances of subclinical increases of urinary albumin within the defined reference interval that are predictive of disease [65-67].

The term microalbuminuria was coined to define this phenomenon, where such increases had value in predicting the onset of nephropathy in insulindependant diabetes mellitus [68]. The accepted reference interval defined for microalbuminuria is between 30 to 300 mg in 24 hours [69,70]. Because micro-albuminuria appears to be a sensitive indicator of renal injury, there is a growing interest in the nephrotoxicity biomarker community to evaluate this marker in the context of an early biomarker predictive of drug-induced renal injury. Although microalbuminuria has traditionally been used in preclinical drug development to assess glomerular function there is growing evidence to suggest that albuminuria is a consequence of impairment of the proximal tubule retrieval pathway [71]. Evidence that microalbuminuria might provide value in diagnosing drug-induced nephrotoxicity was reported in four of 18 patients receiving cisplatin, ifosamide, and methotrextate to treat osteosar-coma [72]. Because microalbuminuria can be influenced by other factors unrelated to nephrotoxicity, including vigorous exercise, hematuria, urinary tract infection, and dehydration [5] - it may have greater predictive value for renal injury in the context of a panel of markers with increased sensitivity and site specificity. Indeed, further evaluation of microglobulin as an early biomarker of site - specific or general nephrotoxicity is required before qualification for preclinical and clinical use.


Osteopontin (OPN) is a 44-kDa highly phosphorylated secreted glycoprotein originally isolated from bone [73] . It is an extremely acidic protein with an isoelectric point of 4.37 (, made even more acidic through phosphorylation on a cluster of up to 28 serine residues [74]. Osteopontin is widely distributed among different tissues, including kidney, lung, liver, bladder, pancreas, and breast [75] as well as macrophages [76], activated T-cells [77], smooth muscle cells [78], and endothelial cells [79]. Evidence has been provided demonstrating that OPN functions as a calcium oxalate crystal formation inhibitor in cultured murine kidney cortical cells [80]. Immunohistochemical and in situ hybridization examination of the expression and distribution of OPN protein and mRNA in the rat kidney clearly demonstrated that levels are highest in the descending thin loop of Henle and cells of the papillary surface epithelium [81]. Uroprontin, first described as a relative of OPN was among the first examples of OPN isolation from human urine [82].

Although normally expressed in kidney, OPN expression can be induced under a variety of experimental pathologic conditions [83,84], including tubu-lointerstitial nephritis [85] , cyclosporine - induced neuropathy [86] , hydrone-phrosis as a consequence of unilateral ureteral ligation [87], renal ischemia [88], nephropathy induced by cisplatin, and crescentric glomeulonephritis [89]. Up-regulation of OPN has been reported in a number of animal models of renal injury, including drug--nduced nephrotoxicity by puromycin, cylcoapo-rine, strptozotocin, phenylephrine, and gentamicin (reviewed in [90a]). In the rat, gentamicin-induced acute tubular necrosis model OPN levels were highest in regenerating proximal and distal tubules, leading the authors to conclude that OPN is related to the proliferation and regeneration of tubular epithelial cells following tubular damage - 90b] . Although osteopontin has been proposed as a selective biomarker of breast cancer - 91] and a useful clinical biomarker for the diagnosis of colon cancer [92] - OPN shows great promise and requires further evaluation as a clinical biomarker for renal injury. Certainly, the high levels of OPN expression following chemically or physically induced renal damage coupled with the recent availability of antibody-based reagents to examine the levels of mouse, rat, and human urinary OPN provide ample opportunity to evaluate OPN as an early marker of AKI in the clinic and a predictive marker of drug--nduced nephrotoxicity preclinically. Further planned work by the ILSI-HESI and C-Path groups hope to broaden our understanding regarding the utility of OPN in either capacity as an early predictor of renal injury.

Neutrophil Gelatinase-Associated Lipocalin

Neutrophil gelatinase-associated lipocalin (NGAL) was first identified as the small molecular-weight glycoprotein component of human gelatinase affinity purified from the supernatant of phorbol myristate acetate stimulated human neutrophils. Human gelatinase purifies as a 135-kDa complex comprised of the 92-kDa gelatinase protein and the smaller 25-kDa NGAL [93]. NGAL has subsequently been shown to exist primarily in monomeric or dimeric form free of gelatinase. A BLAST search of the 178-amino acid NGAL protein yielded a high degree of similarity to the rat a2 microglobulin-related protein and mouse protein 24p3, suggesting that NGAL is a member of the lipocalin family. Lipocalins are characterized by the ability to bind small lipophilic substances and are thought to function as modulators of inflammation [94] . More recent work has shown that NGAL, also known as siderocalin, complexes with iron and iron-binding protein to promote or accelerate recovery from proximal tubular damage (reviewed in [95]). RNA dot blot analysis of 50 human tissues revealed that NGAL expression is highest in trachea and bone tissue, moderately expressed in stomach and lung with low levels of transcript expression in the remaining tissues examined, including kidney [94]. Because NGAL is a reasonably stable small- molecular- weight protein, it is readily excreted from the kidney and can be detected in urine.

NGAL was first proposed as a novel urinary biomarker for the early prediction of acute renal injury in rat and mouse models of acute renal failure induced by bilateral ischemia [96] - Increases in the levels of urinary NGAL were detected in the first hour of postischemic urine collection and shown to be related to dose and length of exposure to ischemia. In this study the authors reported NGAL to be more sensitive than either NAG or p2M, underscoring its usefulness as an early predictor of acute renal injury. Furthermore, the authors proposed NGAL to be an earlier marker predictive of acute renal injury than KIM - 1, since the latter reports injury within 24 hours of renal injury compared to 1 hour for NGAL. Marked up-regulation of NGAL expression was observed in proximal tubule cells within 3 hours of ischemia-induced damage, suggesting that NGAL might be involved in postdamage reepitheli-alization. Additional work demonstrated that NGAL expression was induced following mild ischemia in cultured human proximal tubule cells. This paper also addressed the utility of NGAL as an early predictor of drug-induced renal injury by detecting increased levels of NGAL in the urine of cisplatin-treated mice.

Adaptation of the NGAL assay to address utility and relevance in a clinical setting showed that both urinary and serum levels of NGAL were sensitive, specific, and highly predictive biomarkers of acute renal injury following cardiac surgery in children [97]. In this particular study, multivariate analysis showed urinary NGAL to be the most powerful predictor in children that developed acute renal injury. Measurable increases in urinary NGAL concentrations were recorded within 2 hours of cardiac bypass surgery, whereas increases in SC levels were not observed until 1 to 3 days postsurgery. Other examples demonstrating the value of NGAL as a predictive biomarker of early renal injury include association of NGAL with severity of renal disease in proteineuric patients [98] and NGAL as an early predictor of renal disease resulting from contrast-induced nephropathy [99]. NGAL has been one of the most thoroughly studied new biomarkers predictive of AKI as a consequence of disease or surgical intervention, and to a lesser extent, drug-induced renal injury.

Sensitive and reliable antibody-based kits have been developed for a number of platforms in both humans and rodents (Table 2) and there is considerable interest in examining both the specificity and sensitivity of NGAL for acceptance as a fit - for- purpose predictive biomarker of drug -induced renal injury to support regulatory submissions. Certainly, because NGAL is such an early marker of renal injury, it will have to be assessed as a stand-alone marker of renal injury as well as in the context of a larger panel of markers that may help define site specific and degree of kidney injury.

Renal Papillary Antigen 1

Renal papillary antigen 1 (RPA-1) is an uncharacterized antigen that is highly expressed in the collecting ducts of the rat papilla and can be detected at high levels in rat urine following exposure to compounds that induce renal papillary necrosis [100] . RPA - 1 was identified by an IgG1 monoclonal antibody, designated Pap X 5C10, that was generated in mice immunized with pooled fractions of a cleared lysate of homogenized rat papillary tissue following crude DEAE anion-exchange chromatography. Immunohistochemical analysis of rat papillae shows that RPA- I is localized to the epithelial cells lining the luminal side of the collecting ducts and to a lesser extent in cortical collecting ducts. A second publication described the adaptation of three rat papilla-specific monoclonal antibodies, including Pap X 5C10 (PapA1), to an ELISA assay to examine antigen excretion in rat urine following drug-induced injury to the papillae using 2-bromoethanamine, propyleneimine, indomethicin, or ipsapirone [101]. Of the three antibodies evaluated, PapA1 was the only antigen released into the urine of rats following exposure to each of the toxicants. The authors concluded that changes in the rat renal papilla caused by xenobiotics could be detected early by urinary analysis and monitored during follow-up studies. This study also clearly demonstrated that the Pap X 5C10, PapA1, RPA-1 antigen had the potential for use as a site-specific biomarker predictive of renal papillary necrosis. Indeed, the Pap X 5C10 monoclonal antibody was adapted for commercial use as an RPA-1 ELISA kit marketed specifically to predict or monitor site-specific renal injury in the rat [49].

The specificity and sensitivity of the rat reagent has generated a great deal of interest in developing an equivalent reagent for the detection of human papillary injury. Identification of the RPA-1 antigen remains elusive. Early biochemical characterization of the antigen identified it as a large-

molecular-weight protein (150 to 200kDa) that could be separated into two molecular- weight species with isoelectric points of 7.2 and 7.3, respectively [100]. However, purification and subsequent protein identification of the antigen were extremely challenging. A recent attempt at the biochemical purification and identification of the RPA-1 antigen has been equally frustrating, with investigators providing some evidence that the antigen may be a large glycoprotein and suggesting that the carbohydrate moiety is the specific epitope recognized by the Pap X 5C10 monoclonal antibody [102]. This would be consistent with, and help to explain why, the rat reagent does not cross-react with a human antigen in the collecting ducts, as protein glycosylation of related proteins often differs dramatically between species, thereby precluding the likelihood of presenting identical epitopes. Nevertheless, continued efforts toward identifying a human RPA-1 antigen will provide investigators with a sorely needed clinical marker for the early detection of drug-i nduced renal papillary injury.

Project Management Made Easy

Project Management Made Easy

What you need to know about… Project Management Made Easy! Project management consists of more than just a large building project and can encompass small projects as well. No matter what the size of your project, you need to have some sort of project management. How you manage your project has everything to do with its outcome.

Get My Free Ebook

Post a comment