Immunolocalization of Kim-1,RPA-1,and RPA-2 in Kidney of Gentamicin-,Mercury-,or Chromium-Treated Rats: Relationship to Renal Distributions of iNOS and Nitrotyrosine

Study 1 (Time Response Study)

Adult male SD rats (10–12 weeks old, weighing 270–300 g) were purchased from Harlan (Indianapolis, IN) and divided into 9 groups. Rats in groups 1 and 2 (n = 6) each received a single sc injection of 100 mg Gen/kg daily for 3 days. Rats in groups 3 and 4 (n = 6) each received a single iv injection of 6) each 0.25 mg Hg/kg (HgCl₂). Rats in groups 5 and 6 (n = received a single sc injection of 5 mg Cr/kg (K₂Cr₂O₇). All agents were purchased from Sigma Chemical Company (St. Louis, MO). Control rats in groups 7–9 (n =3 each) were injected with saline (vehicle for Hg and Cr) or deionized water (vehicle for Gen). Rats were euthanized 24 hours or 72 hours after the single injection of Hg and Cr and 24 or 72 hours after the third and final daily dose of Gen.

Study 2 (Dose Response Study)

Rats were divided into 6 groups, with n = 6 for groups 1, 2, 3, and 6 and n = 4 for groups 4 and 5. The rats in groups 1 to 5 received a single injection of Gen daily at doses of 50, 100, 150, 200, or 300 mg/kg/day, sc, for 3 days, respectively. Control rats (n = 6) in group 6 received sc injections with dH₂O. All rats were euthanized 24 hours after the third and final daily dose of Gen.

In studies 1 and 2, kidneys were collected at necropsy and fixed in 10% zinc-formalin for about 48 hours, embedded in paraffin, cut at a thickness of 5 μm, and then stained with hemotoxylineosin for histopathological study. Unstained, formalin-fixed, paraffin-embedded tissue sections were used for immunohistochemical studies.

Grading System for Renal Lesions

Renal lesions were assessed by light microscopic examination in blinded fashion using a scale of 0 to 5, according to the severity of tubular cell necrosis, apoptosis, degeneration, regeneration, tubular dilatation and protein casts, glomerular vacuolization, glomerular mesangial cell proliferation, and interstitial lymphocytic infiltration in the S1/S2 segments or S3 segments: 0 = normal histology; 1 = tubular epithelial cell degeneration, without significant necrosis or apoptosis; 2–5 = <25%, <50%, <75% and ≥75% of the tubules showing tubular epithelial cell necrosis and apoptosis accompanied with other concomitant alterations, respectively.

Immunohistochemical Studies

Renal tubular epithelial cell apoptosis was detected using the TUNEL procedure, according to the manufacturer’s instructions (Trevigen, Inc, Gaithersburg, MD).

For detection of Kim-1, RPA-1, and RPA-2, indirect immunoperoxidase staining procedures were used. Serial sections of formalin-fixed, paraffin-embedded renal tissue were mounted on glass slides coated with poly-L-Lysine (American HistoLabs, Inc, Gaithersburg, MD). Preparation of these sections included microwave irradiation in a pressure cooker with antigen retrieval Citra solution. After microwave treatment, slides were cooled in the antigen retrieval Citra solution for 20 minutes and then rinsed with dH₂O. For blocking endogenous peroxidase activity, sections were incubated with 0.3% hydrogen peroxide in methanol for 30 minutes and then with 5% normal horse serum for 30 minutes. Sections were incubated overnight at 4°C with the following primary mAb at 1:50 dilutions: mouse anti-rat Kim-1 ectodomain (MARKE) mAb (Harvard Medical School), RPA-1 purified mAb (Code: BI087CD, Biotrin International Ltd), and RPA-2 purified mAb (Code: BI088LH, Biotrin International Ltd). Subsequently, sections were incubated with a biotinylated secondary antibody (Vector, Burlingame, CA) for 1 hour and then incubated with avidin-biotinylated horseradish peroxidase complex (Vector) for 30 minutes. The peroxidase reaction was carried out with 0.05% 3’3’-diaminobenzidine in 0.1 M Tris-HCl buffer and 0.01% hydrogen peroxide for 10 minutes. The immunostained sections were counterstained with hematoxylin for 1 minute. For negative control staining, the primary mAb was omitted from the incubation step. The isotype mouse IgGs matched for the primary mAbs for Kim-1, RPA-1, and RPA-2 are not available from the manufacturers.

Immunoperoxidase stains for iNOS (1:100 dilution; BD Biosciences, San Diego, CA) and nitrotyrosine (1:100 dilution; Cell Sciences, Norwood, MA) were performed on 5-μm sections of formalin-fixed, paraffin-embedded tissues utilizing the avidinbiotin complex immunoperoxidase technique (Vectastain Elite kit, Vector Laboratories, Burlingame, CA) with antigen retrieval by microwave for 30 minutes in a 10 mM Citra solution. For negative controls, the primary antibodies recognizing iNOS and nitrotyrosine were omitted or replaced by the isotype-matched mouse IgG_2a (BD Biosciences) for iNOS and mouse IgG_2b (Cell Sciences) for nitrotyrosine, respectively (Zhang et al., 2006).

Statistics

A Kruskal-Wallis rank test was used to assess differences in renal lesion scores. A p value < 0.05 was considered statistically significant.

Histopathology

In Gen-, Hg-, or Cr-treated rats (study 1, time response study), renal lesion scores were greater at 72 hours after the last dose (scores of 3.3, 3.8, 4.8, for Gen, Hg, and Cr, respectively) than at 24 hours (scores of 1.5, 3.0, 2.8, for Gen, Hg, and Cr, respectively) (Table 2). In the dose-response Gen study (study 2), renal lesion scores at 24 hours after the last dose increased with dose (scores of 1.2, 1.5, 2.3, 3.5, and 4.8 at doses of 50, 100, 150, 200, and 300 mg/kg, respectively) (Table 3).

In studies 1 and 2, the patterns of tubular injury were similar. Epithelial cell necrosis was more prominent in the S1/S2 segments compared to the S3 segments in Gen- or Cr-treated rats. No significant changes were seen in any segments of LH and CD in the outer and inner medulla. Epithelial cell necrosis was usually accompanied with other less severe morphological alterations: cytoplasmic vacuolization, regeneration, and hyperplasia. Vacuolization of tubular epithelial cells of the S3 segments was prevalent in Hg-treated rats, whereas epithelial regeneration and hyperplasia of the S1/S2 and S3 segments were more evident in Gen- or Cr-treated rats.

Immunolocalization of Kim-1

Control rats exhibited no positive immunoperoxidase staining for Kim-1 in any segments of the nephron (Figure 1A), which is consistent with several published reports (Table 1). In the time course study, a positive reaction for Kim-1 was detected at the apical surface of individual cells of the S1/S2 segments 24 hours after the final injection in rats treated with Gen 100 mg/kg, sc (Figure 1B), whereas at 72 hours after the final injection, the apical positive staining became diffuse within the entire cytoplasm of tubular epithelial cells of the S1/S2 segments at sites of necrosis and apoptosis, particularly in the cytoplasm of desquamated necrotic cells in the lumen of these tubules (Figure 1C). The apical and cytoplasmic immunostaining pattern for Kim-1 was also visualized at 24 hours and at 72 hours, respectively, in Hg- (Figures 1E, 1F) or Cr-treated rats (Figures 1G, 1H).

In the dose-response study, 24 hours after the last administration of Gen injected sc for 3 days, Kim-1 immunoreactivity increased with the dose. At 24 hours after Gen 50 or 100 mg/kg, sc for 3 days, Kim-1 immunoreactivity was localized to individual S1/S2 segments. At doses of 150 and 200 mg/kg Gen, Kim-1 immunoreactivity extended to clusters of the S1/S2 segments. At the highest dose of 300 mg/kg of Gen, Kim-1 expression was localized in most of the S1/S2 segments and in a few S3 segments in the medullary rays (Figure 1D).

A lower magnification photomicrograph of a Kim-1-stained kidney section clearly showed the topographic distribution of Kim-1 expression; immunoreactivity was higher in the S1/S2 segments of the cortex but lower in the S3 segments of the outer medulla and medullary rays 72 hours posttreatment in rats injected with Gen 100 mg/kg, sc (Figure 2). A summary of time- and dose-dependent Kim-1 expression is presented in Tables 2 and 3, respectively.

Immunolocalization of RPA-1

Control rats exhibited positive immunostaining for RPA-1, which was mainly concentrated in the CD epithelial cells in the outer and inner medulla and papilla (Figure 3A) and expressed in the cytoplasm of CD cells (cover illustration). RPA-1 immunoreactivity was also diffusely localized in the cytoplasm of the cortical CD epithelial cells (Figure 3B) in control rats.

The immunostaining pattern and intensity of RPA-1 in the cortical, medullary, and papillary CD were similar between control (Figure 3A) and Gen-treated rats (Figures 3C, 3E). However, additional RPA-1 expression was noted in necrotic and apoptotic epithelial cells of the S1/S2 segments in rats 24 hours after the final of 3 daily injections of 50 mg Gen/kg (Figure 3D). In rats treated with 300 mg Gen/kg for 3 days, RPA-1 expression in the S1/S2 segments became intense (Figure 3F). A lower magnification photomicrograph showed that constitutive, basal RPA-1 expression in the cortical CD cells coexisted with nascent RPA-1 in the S1/S2 segments of the cortex and in the S3 segments of the outer medulla in rats 72 hours after the final of 3 daily injections of 100 mg Gen/kg, sc (Figure 2).

In Hg-treated rats, RPA-1 expression was observed in the S3 segments 24 hours posttreatment (Figure 3G). In contrast, in Cr-treated rats, RPA-1 immunoreactivity was expressed in the S1/S2 segments 24 hours posttreatment (Figure 3H).

Immunolocalization of RPA-2

Control rats exhibited a robust basal level of RPA-2 expression in the medullary LH (Figure 4A). The immunoperoxidase staining for RPA-2 in control rats showed cytoplasmic localization in the epithelial cells of the descending and ascending thin LH segments in the medulla (Figure 4B). RPA-2 expression was sporadically located in the epithelial cells of thick ascending LH in the cortex (Figure 4C). No significant visible differences of RPA-2 immunoreactivity were apparent in medullary LH cells between control (Figure 4A) and Gen-treated rats (Figure 4D). In contrast, in addition to the high basal expression of RPA-2 in the epithelial cells in the ascending thick LH segments of the cortex, nascent RPA-2 expression appeared in the proximal tubules (Figure 2) in Gen-treated rats. In higher magnification photomicrographs in Gen-treated rats, RPA-2 expression was observed in the S1/S2 segments (Figures 4E, 4F) adjacent to LH segments. The immunostaining intensity in the S1/S2 segments was stronger at 24 hours than at 72 hours after completion of 3 Gen injections, and the cytoplasmic coarse granules at 24 hours (Figure 4E) became a weaker and homogenous cytoplasmic stain after 72 hours posttreatment (Figure 4F).

In Hg-treated rats, RPA-2 expression was not increased above basal levels in the cortical LH epithelial cells (Figure 4G) and no RPA-2 was detected in the S1/S2 segments. In Cr-treated rats, basal RPA-2 expression was localized to the cortical LH epithelial cells; however, in contrast to Hg treatment, RPA-2 expression was also observed in the S1/S2 segments (Figure 4H).

Immunolocalization of Apoptosis, iNOS, and Nitrotyrosine

No apoptotic cells and no expression of iNOS or nitrotyrosine were found in the cortex and medulla in control rats (Figures 5A–5C). However, in Gen- and Cr-treated rats, a significant number of apoptotic cells were detected in the epithelia of the S1/S2 segments (Figures 5D, 5J), whereas in Hg-treated rats, apoptotic cells were located in S3 tubular cells (Figure 5G). The expression of iNOS was apparent in S1/S2 segments and in S3 segments (Figure 5E) in Gen-treated rats, in S3 segments in Hg-treated rats (Figure 5H), and in the S1/S2 segments in Cr-treated rats (Figure 5K). Nitrotyrosine expression was up-regulated in the S1/S2 segments and a few S3 segments in Gen-treated rats (Figure 5F), in the S3 segments in Hg-treated rats (Figure 5I), and in the S1/S2 segments and a few S3 segments in Cr-treated rats (Figure 5L).

Increased Immunoreactivity of Kim-1, RPA-1, and RPA-2 in Necrotic and Apoptotic Proximal Tubular Cells

The present study demonstrated that in Gen-, Hg-, and Cr-treated rats, Kim-1 immunoreactivity increased significantly compared to that of control kidneys. Kim-1 immunoreactivity correlated well with the severity of tubular epithelial cell necrosis and apoptosis, and the increases in Kim-1 immunoreactivity were time and dose dependent. For Gen and Cr exposures, Kim-1 immunoreactivity was observed to be predominantly localized in the S1/S2 segments and less markedly in the S3 segments. In contrast, for Hg exposure, Kim-1 immunoreactivity was observed predominately in the S3 segments, which are the primary targets where mercuric ions accumulate and express toxicity after exposure to elemental or inorganic forms of Hg (Diamond and Zalups, 1998). For all 3 nephrotoxicants, Kim-1 did not appear to be localized in any other nephron segments, that is, CD or LH. These findings are consistent with previous observations of undetectable Kim-1 in normal healthy kidneys, and its marked increase under pathological conditions (Amin et al., 2004; de Borst et al., 2007; Han et al., 2002; Ichimura et al., 1998, 2004; Lin et al., 2007; Vaidya et al., 2006; van Timmerren et al., 2006, 2007; Vaidya and Bonventre, 2006; Zhou et al., 2008).

Interestingly, RPA-1 and RPA-2 immunoreactivity above constitutive levels was also detected in Gen-, Hg-, and Cr-treated rats. Differences in RPA-1 and RPA-2 immunolocalization were also observed between rats treated with Gen or Cr (the S1/S2 segments, or the S3 segments) and rats treated with Hg (the S3 segments). An explanation for why Gen, Hg, and Cr induced different immunolocalization of Kim-1, RPA-1, and RPA-2 is unknown.

Immunoreactivity of RPA-1 and RPA-2 in CD and LH

In control and all rats treated with nephrotoxicants, no Kim-1 immunoreactivity was identified in the CD and LH of the medulla and papilla. The negative immunoreactivity of Kim-1 in the LH and CD highlights its specificity as a biomarker of proximal tubular epithelial cell injury. In contrast, RPA-1 and RPA-2 immunoreactivity was detected in the CD and LH, respectively, in the cortex, medulla, and papilla from control and all treated rats. The localization of RPA-1 in the medulla and cortex is consistent with previous observations (Falkenberg et al., 1996). Moreover, the immunoreactivity patterns of RPA-1 and RPA-2 were similar in both control and all treated rats. These findings were consistent with the expected results, because no histopathologic injury in the medulla and papilla was observed in any treated rats.

Relation of the Site Predilection of Kim-1, RPA-1, and RPA-2 to Renal Distribution of iNOS and Nitrotyrosine

Little or no Kim-1 mRNA and Kim-1 protein is expressed in the proximal tubular epithelial cells of the normal adult rat kidney; however, a variety of pathological conditions in the kidney up-regulate its expression (Ichimura et al., 1998, 2004; Vaidya and Bonventre, 2006; van Timmerren et al., 2006, 2007; Kramer et al., 2007; Zhou et al., 2008). In contrast, little is known about mRNA and proteins of RPA-1 and RPA-2 in normal and injured cells of the S1/S2 segments. In the present study, necrotic and apoptotic cells of the S1/S2 and S3 segments were immunoreactive to Kim-1, RPA-1, and RPA-2, but normal epithelial cells of these segments were not. Moreover, the same sites of predilection were also immunoreactive for iNOS and nitrotyrosine antigens. These findings suggest that the intracellular, immunoreactive Kim-1, RPA-1, and RPA-2 proteins were more abundant in injured, necrotic, and apoptotic cells of the S1/S2 and S3 segments than in their normal healthy counterparts. Kim-1, RPA-1, and RPA-2 immunoreactivities appear to be enhanced in nephron locations where iNOS and nitrotyrosine immunoreactivities are increased. Whether increased immunoreactive proteins represent enhanced de novo protein synthesis of RPA-1 and RPA-2 or reduced degradation of constitutive protein remains to be determined. Kim-1 mRNA is not expressed in healthy cells, and the cellular processes responsible for enhanced Kim-1 expression after nephrotoxicant administration are not entirely known.

Up-Regulation of Peroxynitrite as a Potential Mechanism of Renal Injury

Many renal pathological conditions, for example, ischemia/reperfusion injury with or without endotoxin insult and diabetic nephropathy, can significantly promote NO production and peroxynitrite formation. There have been several reports concerning the beneficial or detrimental effects of NO on renal injury. It was suggested that NO formation prevents Gen-induced nephrotoxicity in a model of rat acute renal failure (Ghaznavi et al., 2005) and that iNOS contributes to the prolonged protective effects of ischemic preconditioning in the mouse kidney (Park et al., 2003). In contrast, iNOS gene knockout mice are protected against hypoxic injury in renal proximal tubules (Ling et al., 1998). At present, it is difficult to reconcile these disparate findings involving opposite effects of iNOS; however, the concept of an initial beneficial and late detrimental effect seems reasonable and acceptable as an explanation (Amin et al., 2004).

In this context, the present study is consistent with the reported cytotoxic effects of NO generated from iNOS on acute tubular necrosis and apoptosis in Gen-, Hg-, and Cr-induced nephrotoxicity (Kim et al., 1997; Yanagisawa et al., 1998; Baliga et al., 1999; Bagchi et al., 2001; Basnakian et al., 2002; Maldonado et al., 2003; Pedraza-Chaverri et al., 2005; Parlakpinar et al., 2005). Furthermore, it was reported that NO results in oxidative damage and contributes to the “point of no return” of host cell injury, necrosis, and apoptosis (Thuraisingham et al., 2000; Goligorsky et al., 2002; Yokozawa et al., 2002; Forbes et al., 2002; Nakajima et al., 2006; Pacher et al., 2007). A recent review suggests that NO does not have as many potential toxic effects as previously reported. Most of the cytotoxicity attributed to NO is rather due to the peroxynitrite and peroxynitrite anion, which is more reactive and toxic than NO (Pacher et al., 2007). Peroxynitrite is capable of causing protein tyrosine nitration (Radi, 2004) and is also capable of mediating damage in renal mitochondria and apoptosis (Yokozawa et al., 2002). Radi (2004) suggested that peroxynitrite can cause protein tyrosine nitration in vitro and in vivo. NO does not directly produce nitrotyrosine, but it reacts rapidly with superoxide anion and forms peroxynitrite, a potent nitrating and oxidizing agent (Kim et al., 1997). Peroxynitrite can induce oxidative damage to proteins, DNA, and lipids (Bian et al., 1999; Schnellmann, 2001). Nitrotyrosine can be detected immunohistochemically using an antibody against nitrotyrosine (Bosse and Bachmann, 1997). Immunohistochemical localization of nitrotyrosine in the proximal tubular epithelial cells was detected in Gen-treated rats (Maldonado et al., 2003). Enhanced expression of nitrotyrosine correlated with necrotic tubular cells in Cr-treated rats (Pedraza-Chaverri et al., 2005). Recent reports suggest that nitrotyrosine represents the nitration of protein-bound tyrosine residues by peroxynitrite (Forbes et al., 2002) and that nitrotyrosine can be used as a marker for endogenous production of peroxynitrite (Maldonado et al., 2003).

Kim-1: A Possible Link to the Reaction of Peroxynitrite with Cysteine

Kim-1 is a type 1 membrane protein containing a 6-cysteine immunoglobulin (Ig)-like domain. Kim-1 is different from other members of the Ig superfamily in containing 6 cysteines, compared with 2 cysteines in most members of the Ig superfamily and 4 in mucosal dressin cell adhesion molecule-1 (MAdCAM-1); however, the reason for the large number of cysteines remains unclear. Kim-1 is labeled as a so-called novel Ig domain (Ichimura et al., 1998). A recent review suggests that peroxynitrite reacts with various amino acids (Pacher et al., 2007). Of these reactions, that of peroxynitrite with cysteine is the most prevalent. In addition, cysteine oxidation by peroxynitrite can promote enzyme activation and modify proteins (Pacher et al., 2007). Based on these findings, we hypothesize that the reaction of peroxynitrite and cysteine may lead to alterations of Kim-1 primary sequence and function. It is tempting to consider the possibility of S-nitrosylation of cysteine thiol residues, a posttranslation protein modification by the covalent addition of a nitrogen monoxide group to the thiol side chain of a protein or peptide cysteine residue (Fukumura et al., 2006). S-nitrosylation is used in mammalian cells to convey several specific signals that are elicited by NO (Fukumura et al., 2006). Further studies are needed to evaluate this hypothesis.

Spatial Relationship between the Apical Location of Kim-1 and the Subapical Location of Nitrotyrosine

In endotoxintreated rats, enhanced immunoreactivity of nitrotyrosine was attributed to reabsorption of hydrolytic nitrated proteins by proximal tubule epithelial cells (Bian et al., 1999). The marked staining of nitrotyrosine in proximal epithelial cells was in the subapical compartment where the endocytic lysosomes are located (Bian et al., 1999). A key finding is that nitrotyrosine is located at the apical cytoplasm beneath the luminal brush border of the S1/S2 segments (Bian et al., 1999). Similarly, immunohistochemical staining for renal nitrotyrosine in experimental diabetic nephropathy revealed that nitrotyrosine was mostly localized to the apical region of proximal tubular cells (Forbes et al., 2002). The present and previous studies (Han et al., 2002; Ichimura et al., 1998, 2004; Lin et al., 2007; Vaidya et al., 2006; van Timmerren et al., 2006, 2007; Vaidya and Bonventre, 2006; Zhou et al., 2008) have localized Kim-1 at the apical membrane of epithelial cells of proximal tubule segments. There appears to be a spatial relationship between the site-specific nitrosative injury at the subapical location and Kim-1 immunoreactivity.

Kim-1, RPA-1, and RPA-2: Potential Tissue Markers for Tubular Cell Injury

Previous studies in the laboratory have demonstrated that urinary Kim-1 is a sensitive nephrotoxicity biomarker in Gen-, Hg-, and Cr-treated rats (Zhou et al., 2008) and that urinary RPA-1 is a nephrotoxicity biomarker in Gen-treated rat (Shaw et al., 2007). Whether Kim-1, RPA-1, and RPA-2 could be used as sensitive, specific, and reliable tissue markers remains to be assessed in detail using IHC compared with the gold standard of histopathologic evaluation of hematoxylin and eosin-stained sections (Gillett et al., 2002). Previous IHC studies have demonstrated that the negative findings for RPA-1 and RPA-2 induction with proximal tubular toxicants reinforce their potential use as biomarkers for renal papillary injury and as negative controls for proximal tubular injury (Falkenberg et al., 1996; Hilderbrand et al., 1999; Shaw, 2005; Kilty et al., 2007). In the present study, induction of RPA-1 and RPA-2 in proximal tubular cells is unusual and warrants further study.

Control experiments with isotype-matched control antibodies on a serial section are important (Gillett et al., 2002). At present, these immunoglobulins are not commercially available. To compensate for the lack of these sera, immunostaining for Kim-1, RPA-1, and RPA-2 has been performed in a study using a non-nephrotoxic agent. SD rats (10-, 25-, 40-, 80-day-old) treated with valproic acid (a hepatotoxic agent) were used as animal experimental controls and compared with Gen-treated rats. No tubular epithelial injury of the kidney was found, nor was any positive reaction of Kim-1, RPA-1, and RPA-2 in renal tissues obtained from valproic acid–treated rats, compared to the positive immunostaining for these markers detected in the kidney of Gen-treated rats (manuscript, in preparation). In the present study, omission of primary antibodies of Kim-1, RPA-1, and RPA-2 resulted in a negative reaction in renal tissue sections obtained from Gen-, Hg-, and Cr-treated rats. Furthermore, no immunostaining for Kim-1, RPA-1, and RPA-2 was found in any of the proximal epithelial cells and glomerular cells in the kidney from vehicle-treated rats (Figures 1A, 3B, 4C).