Clinicopathological and Tissue Indicators of Para-Aminophenol Nephrotoxicity in Sprague–Dawley Rats

Abstract

A model of para-aminophenol (PAP) nephrotoxicity in Sprague–Dawley rats was utilized to characterize potential indicators of toxicity in the kidney and in biofluids, and to chronicle the progression of acute renal injury. Rats were administered PAP at a low or high dose and examined terminally at 6, 24 and 48 hours (4 animals/group with matching controls). Acute tubular necrosis was observed in the medullary rays (low and high doses) and the outer stripe of outer medulla (high dose only) as early as 6 hours postdosing. Starting at 24 hours, regeneration of the tubular epithelium was evident in both low and high dose studies. Associated with the tubular lesions, we observed elevation of urinary α-glutathione S-transferase levels, an indicator of proximal tubular injury. By immunohistochemistry of the kidney, decreased γ-glutamylcysteine synthetase expression correlated with tubular injury, especially at high dose, whereas elevation of vimentin, osteopontin, and Ki-67 expression was concurrent with tubular regeneration. Clusterin and kidney injury molecule-1 displayed expression patterns characteristic of both renal injury and regeneration. Taken together, this study provided insight into the progression of nephrotoxicity, and allowed the evaluation of potential urinary and tissue protein biomarkers that could complement the early detection of acute tubular injury.

Introduction

Drug-induced nephrotoxicity is one of the major safety concerns in drug development due to the risk of clinical renal complications associated with drug therapy (Racusen and Solez 1986; Hoitsma et al., 1991; Perazella 2005). Moreover, drug-induced toxicosis is an important cause of acute renal failure in humans (Brady et al., 2004). Based on these premises, the early prediction of nephrotoxicity in preclinical models is of great interest to the pharmaceutical industry, since the characterization of reliable translational biomarkers of nephrotoxicosis would improve the monitoring of renal function in clinical trials (Hewitt et al., 2004).

Para-aminophenol (PAP; also known as 4-aminophenol or p-aminophenol) is a well-known metabolite of acetaminophen and phenacetin, and a derivative of anilines in industrial use (Green et al., 1969; Newton et al., 1983; McCarthy et al., 1985; Kanbak et al., 1996). It induces acute tubular necrosis (ATN) affecting the proximal straight tubule (PST; S3 segment of the proximal tubule), a property that justifies the use of PAP as a model for acute analgesic nephropathy (Green et al., 1969). The lesions produced by PAP in the kidney are dose-responsive: at low dose they affect only the PST within the medullary rays (MRs), with a selectivity for cortical nephrons, while at higher concentrations they extend to the PSTs in the outer stripe of outer medulla (OSOM; Fowler et al., 1991, 1994). This selectivity might be related to the differential perfusion of the tubular areas (Fowler et al., 1991). The PAP-induced tubular damage appears early after administration (Green et al., 1969; Davis et al., 1983), allowing the use of early time points to study the progress of the lesions and the changes in the expression of relevant proteins.

There is currently an ongoing search for indicators of renal toxicity in animal models that could translate into human medicine. Some of the most promising molecules indicative of tubular damage are clusterin and kidney injury molecule-1 (KIM-1) (Hidaka et al., 2002; Han and Bonventre, 2004). Others, such as osteopontin, vimentin and Ki-67, may act as indicators of regeneration (Chevalier et al., 1998; Persy et al., 1999; Endl and Gerdes, 2000). The changes in expression levels of these and a few other proteins following PAP administration were evaluated by immunohistochemistry (IHC) to assess their utility as potential nephrotoxicity markers.

Of major interest in clinical settings is the assessment of renal function with non-invasive methods. Traditional markers of renal damage like blood urea nitrogen (BUN) and creatinine are often of low sensitivity or capture only irreversible lesions (Brady et al., 2004). Currently many scientists and physicians are collaborating in an attempt to obtain more reliable biomarkers of early renal damage (Han and Bonventre, 2004; Hewitt et al., 2004). In this sense, the glutathione S-transferase (GST) molecules have shown great potential as indicators of tubular injury due to their enhanced sensibility and specific distribution (Rozell et al., 1993; Han and Bonventre, 2004). We chose α-GST and μ-GST, markers of proximal and distal tubular damage, respectively, to evaluate the chronological variation in the urine of rats administered PAP.

In summary, the goals of this study were two-fold: to characterize the time-dependent changes of the microscopic lesions that occur in the kidneys of rats after a single PAP administration, and to correlate these findings to clinical pathology parameters and potential tissue biomarkers of nephrotoxicity.

Materials and Methods

Animals and Samples

Forty-eight 7-week-old male Sprague–Dawley rats (Crl: CD (SD) IGS BR) from Charles River Laboratories (Boston, MA) were distributed between low (50 mg/kg) and high dose (150 mg/kg) groups using a computer-assisted randomization procedure to ensure similar distribution of body weights. Each group was subdivided among three terminal time points (6, 24, and 48 hours), with 4 treated and 4 control animals per time point. The rats were housed individually in metabolic cages (VWR, West Chester, PA) to obtain urine at selected time points. The rats were kept in air-conditioned rooms with controlled temperature 20–24°C, relative humidity 55 ± 15% and 12:12 hours dark:light cycles.

For all purposes, standard procedures and conditions were applied for animal care, feeding and maintenance. Water and food were provided ad libitum. Clinical signs were monitored for the duration of the study. Rats were sacrificed with CO₂ overexposure, followed by puncture of the diaphragm and cervical dislocation. All experimental procedures were reviewed and approved by the Pfizer Institutional Animal Care and Use Committee and followed the published guidelines for animal welfare (Guide for the Care and Use of Laboratory Animals, 1996; Animal Welfare Act, 1966, as amended in 1970, as amended in 1976, and 1985, 9 CFR Parts 1, 2, and 3).

The urine was collected via metabolic cages in chilled 15 ml standard conical polypropylene tubes (VWR, West Chester, PA) and stored at 4°C until processed. Blood collection for clinical biochemistry was done following euthanasia with CO₂ overexposure and previous to puncture of the diaphragm. The blood was collected into 3 ml tubes containing lithium heparin.

Postmortem examination was performed on each rat and the gross findings annotated. A central slice was obtained from each kidney and fixed in 10% buffered formalin for 48 hours.

Chemical Formulation and Administration

PAP (Sigma-Aldrich, St. Louis, MO) was prepared in sterile saline at a concentration of 15 mg/ml for the high dose and 5 mg/ml for the low dose. At time 0, each animal was given a single ip injection of PAP or saline at a final volume of 10 ml/kg.

Laboratory Tests

Urine samples were analyzed for the following parameters using standard laboratory procedures: urine volume (measured manually), specific gravity (Clinitek Atlas, Bayer Corporation, Elkhart, IN), total protein (Hitachi 911, Roche Diagnostics, Indianapolis, IN), glucose (Hitachi 911), microalbumin (Hitachi 911), β2-microglobulin (Immulite, Diagnostic Products Corporation, Los Angeles, CA), creatinine (Hitachi 911), N-acetyl-β-D-glucosaminidase (NAG) (Hitachi 911), and γ-glutamyl transpeptidase (GGT; Hitachi 911). α-GST and μ-GST were evaluated via enzyme immunoassay with commercial antibodies (Biotrin, Dublin, Ireland). Routine blood chemistry parameters were obtained utilizing the Hitachi 911.

Numerical and Statistical Analysis

All urinalysis and blood chemistry results were expressed as mean ± standard error of the mean (SEM). Since low dose and high-dose studies were conducted independently of each other, Student’st-test was performed between values from control and treated animals at each dose-time combination to determine whether statistical significance (p < 0.05) was reached.

Tissue Processing and Preparation

Formalin-fixed kidneys were embedded in paraffin blocks and sectioned at 4 μm following routine histological protocols. The sections were placed on frosted glass slides, stained with hematoxylin and eosin (H&E) and periodic acid-Schiff (PAS), and examined by light microscopy. Samples from treated rats and time-matched control rats were examined in parallel. The changes we took into consideration for evaluation purposes were: tubular degeneration/necrosis, tubular dilation, intratubular protein, interstitial infiltrates, and mineralization.

Immunohistochemistry (IHC)

The selection of antibodies was based on preliminary gene expression profiling analysis (manuscript in preparation), and the relevance of the proteins in nephrotoxicity based on literature reviews. The chosen proteins were clusterin, KIM-1, osteopontin, vimentin, γ-glutamylcysteine synthetase (GCS), Ki-67 and cleaved caspase 3. Clusterin, KIM-1, osteopontin and vimentin show modulation during renal damage (Chevalier et al., 1998; Ichimura et al., 1998; Persy et al., 1999; Hidaka et al., 2002). Ki-67 is a well-recognized proliferation marker. GCS is specifically localized to the PST (Shepherd et al., 2000), and its expression was decreased following PAP treatment as shown by gene expression analysis (manuscript in preparation). Cleaved caspase 3 is a well-known marker of apoptosis. The IHC procedures were optimized for each antibody. The antibodies chosen and the methods used are listed in Table 1.

Blocking steps for all techniques included quenching of endogenous peroxidase with 3% H₂O₂ in distilled water, avidinbiotin blocking from Ventana (Tucson, AZ) or from Biocare (Walnut Creek, CA) (except for the Omni-Map kit), and protein blocking (TNB, Perkin-Elmer, Boston, MA). The IHC signal for clusterin was amplified with the TSA kit (Perkin-Elmer). Each secondary antibody was chosen to match the primary antibody and used at concentrations between 1:200–1:500. All slides were counterstained with hematoxylin, and examined by light microscopy in order to describe endogenous antigen distribution and treatment- and time-dependent changes.

Results

Clinical Signs, Clinical Pathology, and Gross Findings

High-Dose

The administration of a single dose of PAP at 150 mg/kg produced significant decreases in body weight and increases in absolute and relative kidney weights (Table 2). Grossly, the kidneys were swollen, and upon cross-section presented a pale band between the cortex and medulla, corresponding to the outer stripe of the outer medulla (OSOM), and extending along the medullary rays (MRs) into the cortex. A minimal amount of clear, straw-colored fluid (less than 1 ml) was found in the peritoneum of all animals at 6 and 48 hours postdosing.

Results of the urinalysis are presented in Table 3. One of the most remarkable findings in the urinalysis was the elevation in α-GST levels, peaking at 6 hr post-dosing (856-fold increase) and decreasing over time (Table 3). Additional indicators of renal toxicity were GGT, NAG, urinary glucose, total urinary protein, and μ-GST, showing different chronological trends (Table 3). Detectable microalbuminuria was only present at 48 hr post-dosing (Table 3).

Serum biochemistry is shown in Table 4. BUN and creatinine at all time points were the most significant parameters indicative of renal damage. The maximum values for both BUN and creatinine were observed at 48 hours postdosing.

Low Dose

The ip administration of a single dose of PAP at 50 mg/kg did not produce any clinical signs or body weight loss (Table 5). In contrast to the high- dose study, there were no gross findings at necropsy and no changes in the terminal relative kidney weights (Table 5).

Due to the expected dose-response effect, both the urinalysis and the clinical chemistry displayed more subtle changes compared to the high dose study. Interestingly, the GST analysis in the urine showed a similar trend, though with a lower magnitude of change, compared to the high-dose study. α-GST peaked at 6 hours (99-fold increase) and decreased steadily over time to control levels (Table 6). μ-GST also had a peak at 6 hours (48-fold increase), but not as pronounced as the α-GST elevation, and it showed a similar steep decrease over time. Both the microalbumin and the total protein had a maximum at 12 hours (data not shown), with significant elevation of the microalbumin also at 6 and 24 hours (Table 6). No serum biochemistry abnormalities reflective of renal damage were observed between the control and treated groups at any time point (Table 7).

Histopathology

High Dose

The histopathological changes in the kidney were remarkable from 6 hours onward. The kidneys showed acute tubular necrosis (ATN) extending along the MRs towards the outer cortex and also into the OSOM (Figure 1A). These areas of necrosis corresponded to the proximal straight tubule (PST; Figure 1C). The lumen of the affected tubules was slightly wider as a consequence of the damage to the epithelial lining (compare Figure 1B, C). Abundant intratubular eosinophilic and pale basophilic proteinaceous material was seen in affected areas, mostly downstream of the damaged PST and extending into the medulla and outer cortex (within distal convoluted tubules) (Figure 1A).

In some cases there was evidence of mineralization within the lumen of necrotic tubules and occasionally in the corresponding basement membranes, more frequently in the outer medulla. At 24 and 48 hours, the progression of the lesions was characterized by an increased number of regenerating tubules (Figure 1C). These regenerating cells often lacked a defined brush border, as observed with the PAS stain (compare Figure 1B, C). At 6 hours postdosing, interstitial capillaries and small vessels in the areas of damage often had an increased number of leukocytes, progressing to mild infiltrates of lymphocytes and macrophages at 24 and 48 hours.

Other lesions found in the kidney with uncertain correlation with treatment were erosion and hyperplasia with disorganization of the papillary epithelium (4/4 at 24 hours and 2/4 at 48 hours; Figure 1D); and multifocal vasculopathy with mural deposition of PAS positive material in high-dose rats (3/4 at 24 hours and 2/4 at 48 hours; Figure 1E).

Low Dose

The only region of the nephron apparently affected by low-dose PAP corresponded to the PST in the MRs (Figures 1F-H). The most prominent and extensive initial lesion (at 6 hours) was expansion of the tubular lumen, accompanied by flattening of the epithelial cells (Figure 1F) and by variable degree of tubulo-epithelial degeneration. The degenerative cells often remained attached to the basement membrane at this time point. In addition, we observed a clear increase in the number of leukocytes within interstitial vessels in and around injured renal areas (Figure 1F). At 24 hours the majority of the degenerative cells were sloughed into the lumen of the PSTs, and initial regeneration was observed. These changes were accompanied by an increased number of lymphocytes and macrophages infiltrating the areas of damage (Figure 1G).

At 48 hours, the lumen of the affected tubules contained a variable amount of cellular debris lacking recognizable nuclei, and the tubular epithelial lining was continuous and actively engaged in regeneration (Figure 1H). At this time point mononuclear infiltrates remained around the areas of damage and were slightly more cellular compared to those at 24 hours (Figure 1H).

Immunohistochemistry

Clusterin

Positive staining in the control animals was of variable intensity and comprised the lumen of vessels, including glomerular capillaries (Figure 2A). In addition, there was positive staining of tubules in areas of nephropathy, rare tubules in the inner medulla, and epithelial cells of the transitional epithelium of the renal pelvis. After treatment with 150 mg/kg of PAP, positive staining was seen initially in necrotic tubules (6 hours); at later time points it was centered in areas of regeneration (OSOM, MRs) and intraluminal material (Figure 2C), and extended to the inner stripe of outer medulla (ISOM) and inner medulla. The parietal epithelium of some glomeruli was also positive. The staining varied from diffusely cytoplasmic to granular or apically oriented (Table 8).

Animals receiving a low dose (50 mg/kg) PAP at 6 hours presented scattered positive cells in damaged PSTs of the MRs often with cytoplasmic staining and staining of pyknotic nuclei (Figure 2B). This distribution was similar at 24 hours, although with fewer positive cells and weaker cytoplasmic staining. At 48 hours, positive cells were even less frequent and had decreased staining intensity (Table 8).

Kidney Injury Molecule-1 (KIM-1)

The normal distribution of KIM-1 in untreated animals was sparse and limited to small foci of rat nephropathy (Figure 2D). Following high-dose PAP treatment, increased KIM-1 expression was seen multifocally at the periphery of necrotic areas, later on extending along regenerating tubules into the MRs and OSOM, with a maximum intensity at 24 hours. A few glomeruli showed positive parietal epithelium. The positive staining was often apically oriented in the tubular cells (Table 8 and Figure 2F).

In the low dose treated animals, rare and usually isolated positive cells in tubules near the areas of damage were observed at 6 hours, extending to the periphery of the MRs and slightly to the OSOM (Table 8). The staining evolved from cytoplasmic to apical (Figure 2E).

Osteopontin

The normal expression of osteopontin in control rats was intensely cytoplasmic in the simple papillary epithelium. Faint cytoplasmic staining was observed in numerous tubular cells at the transition of ISOM and OSOM. Six hours following high-dose treatment, an increased number of osteopontin-positive tubules were seen mostly within the ISOM and in the viable tubules within necrotic areas of the MRs and OSOM. The staining was more intense over time (24 and 48 hours; Figure 2I) and progressed from diffusely cytoplasmic to apical. Some parietal cells in glomeruli, proximal convoluted tubules and the intratubular material were also positive. Interestingly, the positive staining of the papillary epithelium disappeared partially or completely in animals showing histological evidence of damage (Table 8).

Animals treated with the low-dose PAP showed a different pattern of osteopontin expression from those treated with the high-dose PAP. A few positive cells in non-PST tubules of the MRs were present. This pattern was slightly more intense at 24 and 48 hours, with scattered positivity in proximal convoluted tubules and parietal cells (Table 8 and Figure 2H).

Vimentin

Vimentin-positive areas in control kidneys were distributed in the mesangium of glomeruli, interstitium, and blood vessels (Figure 2J), and in the epithelium of some tubules in the ISOM and inner medulla. Tubules in areas of nephropathy were also positive. In animals treated with high dose PAP there were minor changes at 6 hours, evolving to tubular cytoplasmic expression in the OSOM (24 hours) and OSOM and MRs (48 hours) within areas of regeneration (Table 8 and Figure 2L). Animals treated with low-dose PAP showed minimal change in vimentin expression at 6 hr. At 24 hours, we observed rare positive tubular cells in the areas of regeneration.

However, at 48 hours, there were frequent clusters of positive epithelial cells in the PST of the MRs (Table 8 and Figure 2K). The expression pattern appeared more membranous, concentrating at the periphery of the cytoplasm, and in some cases faintly cytoplasmic.

Ki-67

In sections of control kidneys, we observed a moderate number of positive nuclei scattered uniformly throughout the cortex and outer medulla, and a lower number in the inner medulla (Figure 2M). At 6 hours, the high-dose PAP-treated kidneys showed an overall decrease in the total number of labeled nuclei in the tubules due to lack of positive staining within necrotic areas. At later time points (24 and 48 hours), there was a progressive increase in the number of positive cells concentrated at the periphery and inside of OSOM and MRs (Table 8 and Figure 2O). An increased number of positive cells were also present in injured areas of the papillary and pelvic epithelium.

The pattern of Ki-67 positive cells in the low-dose treated animals was normal at 6 hours, and concentrated as clusters within MRs and rarely within the OSOM at 24 and 48 hours postdosing (Table 8 and Figure 2N). Interestingly, 1 control animal showed an elevated number of positive nuclei in the transitional epithelium, suggesting that the same lesions seen in high-dose animals might not be directly associated with PAP administration.

γ-Glutamylcysteine synthetase (GCS)

By immunohistochemistry, GCS showed a well-defined pattern of labeling in controls comprising the PSTs both in the OSOM and MRs (Figure 2P). The staining was intense and diffusely cytoplasmic, with slight variation in the intensity among cells in the same tubule, and often with a darker staining at the brush border in the MRs. At the outermost portion of the MRs, positive cells within the tubular profiles became scarce to finally become completely absent at the mid cortex. With the high-dose PAP treatment, the normal pattern of staining was reduced to the inner OSOM or lacking at all time points (Table 8 and Figure 2R).

At 6 hours, the low-dose treatment decreased the intensity and number of positive cells at the outermost portion of the MRs, and the staining was variable throughout the periphery of the MRs within the cortex (Figure 2Q). The staining of the brush border was also reduced in distended tubules. At 24 hours, some intraluminal cellular debris (necrotic cells) showed positive staining in the MRs and outer medulla. The GCS staining was considered to be within the normal range at 48 hours (Table 8).

Cleaved Caspase 3

The staining for cleaved caspase 3 was negative for both low- and high-dose studies, while the positive control (neoplasm) showed numerous positive cells.

Discussion

To study the modulation of protein expression at the tissue level and in biofluids, we have chosen to evaluate a model of analgesic nephropathy utilizing PAP, a toxicant known to cause necrosis of the renal PST (Green et al., 1969; Davis et al., 1983; Fowler et al., 1994; Harmon et al., 2005). This work followed the dose- and time-dependent progression of ATN lesions and protein expression changes that correlated with morphological stages of injury. The changes in protein expression may also be useful indicators of renal toxicological injury. In addition, clinical pathology showed the relevance of sensitive parameters (GSTs) in the detection of early tubular damage.

Clinical Pathology

At the clinico-pathological level the traditional markers of urinary disease (urine volume, urine specific gravity, urinary glucose, urinary protein, BUN, and serum creatinine) were poor indicators of early renal damage due to their low sensitivity and low dynamic range. In particular, BUN and serum creatinine detected the extensive renal lesions seen in the high-dose rats, but their overall variation in magnitude was minimal compared with nonroutine tests, especially α-GST (Tables 3 and 4). Results from these nonroutine tests, including NAG, GGT, α-GST and μ-GST, indicated damage as early as 6 hours postdosing (Tables 3 and 6). Of the markers evaluated, α-GST was the best in terms of greater dynamic range to detect ATN, showing a higher elevation at the earliest time point (6 hours). The lysosomal enzyme NAG and the brush border enzyme GGT are both excreted into the lumen of the tubules and can be detected in the urine when tubules are damaged but are known to show diurnal variation (Kramer et al., 2004).

μ-GST, traditionally recognized as an indicator of distal tubular damage (Rozell et al., 1993; Han and Bonventre, 2004; Howard, 2005), was elevated both in the low and high dose studies. Since distal tubular damage was not readily seen by histopathology and there was a lack of dose-responsiveness in μ-GST urinary levels, possible damage to distal tubules might be secondary to PAP toxicity (Davis et al., 1983). Finally, microalbumin, a nonspecific marker of glomerular or proximal tubular damage (Lebeau et al., 2005), was elevated at early time points at low dose, but only at 48 hours at high dose. This paradoxical finding could be best explained by the prozone effect in the presence of excess antigen (Jury et al., 1990) following high dose PAP administration. Together, these results emphasize the need to expand the monitoring of the renal function with a wider and more sensitive array of indicators, as underscored previously by other authors (Han and Bonventre, 2004; Hewitt et al., 2004; Howard, 2005).

Histopathology

The earliest and most significant change observed in the low dose study was tubular dilation and occasional hypereosinophilic cells in the MRs, whereas in the high dose study there was diffuse necrosis with loss of epithelial lining in the MRs and OSOM. The rapid dilation of tubules in the low dose has been attributed to a sudden increase in intratubular pressure (Davis et al., 1983). This pattern of lesion distribution along with dose responsiveness is characteristic of the analgesic model of ATN (Newton et al., 1983; Fowler et al., 1991; Harmon et al., 2005).

Evidence of epithelial regeneration for both doses was present at as early as 24 hours and was more prominent at 48 hours. The regeneration started at the periphery of necrotic areas in the high dose study, due to the complete absence on viable cells in the damaged segments of the tubules. In clear contrast, viable epithelial cells within the injured segments were the initial source of regeneration in the low-dose study. A similar phenomenon has been shown for other models (Ghielli et al., 1998). In general, the regenerating cells showed features of immaturity, expressing vimentin and lacking mature brush border (Witzgall et al., 1994). The increased number and the change in distribution of Ki-67 positive cells within areas of damage further supported the morphological interpretation of augmented regeneration in the tubules.

The lesions of ATN were accompanied by the presence of minimal-to-mild interstitial mononuclear infiltrates for both dosage regimens. Initially (6 hours) the cells were contained within vessels, and at later time points (24–48 hours) they infiltrated the interstitium around injured regions. This observation is in contrast to the “paucity of inflammatory reaction” attributed to this model in other studies (Green et al., 1969; Kiese et al., 1975; Newton et al., 1983). This discrepancy may be a consequence of the different route of administration, or the strain of rats used, among other possible factors. In general, infiltrating cells appearing after ATN are believed to play a relevant role in the local delivery of cytokines, and they are known to appear rapidly after injury and disappear upon complete repair (Ghielli et al., 1996; Ysebaert et al., 2004).

Vasculopathic changes and changes to the papillary epithelium were seen with relatively high frequency in high-dose PAP-treated rats. To our knowledge, similar lesions have not been reported in association with PAP nephrotoxicity. The relevance or mechanistic relationship to the treatment is unclear at this time. Fibrinoid necrosis of the mural arterial wall can be seen sporadically in rats (M. Albassam, personal communications), but in this study it was observed exclusively in high-dose treated rats.

Experimentally, similar pathology was induced in rats in a model of hypertension, with lesions appearing in kidney, pancreas, and mesentery as early as 12 hours (Wolfgarten and Magarey, 1959). Other possible causes of this type of vascular damage include uremia and polyarteritis (Wolfgarten and Magarey, 1959). Unfortunately, vascular dynamics were not monitored during the course of the study; therefore the mechanism of these vascular lesions remains unknown. The injury to the papillary and transitional pelvic epithelia could be an indirect response to severe renal tubular injury since similar changes have been seen with other tubular toxicants (F. Ramiro-Ibáñez, personal observation).

Immunohistochemistry

Protein expression studies represent a useful complement to gene expression studies, since the correlation between mRNA and protein levels for a given gene is sometimes variable due, for example, to posttranslational protein modifications. Evaluation of the distribution and intensity of relevant protein expression by IHC therefore provides a direct means of assessing the impact ATN at the tissue level.

Clusterin is a secreted glycoprotein present in many physiologic fluids and often overexpressed in different organs subject to injury or tissue remodeling, including the kidney (Rosenberg and Silkensen, 1995). Clusterin is produced during renal development; however, its specific function remains unclear, with proposed roles including protection of cell membranes and cell-to-cell contact (Nath et al., 1994; Rosenberg and Silkensen, 1995; Chevalier et al., 1999; Hidaka et al., 2002). In high-dose treated rats the renal expression of clusterin was observed in necrotic tubules and adjacent tubules, often with higher intensity at the apical pole of the epithelium, in agreement with results from other models of renal injury (Witzgall et al., 1994; Rosenberg and Silkensen 1995; Hidaka et al., 2002). Intense positive signal was also seen in tubular luminal contents, as would be expected of a secreted protein (Witzgall et al., 1994).

Moreover, different expression patterns manifested in proximal tubules (granular and cytoplasmic) and distal tubules (intensely perinuclear, diffusely cytoplasmic and apical) 48 hours following high-dose PAP treatment, suggesting different cellular roles at different locations. At low-dose PAP, sometimes only single or a few cells within a tubular profile showed positive signal for clusterin, in contrast with the concept of “all or none” expressed by some authors (Nath et al., 1994; Rosenberg and Silkensen, 1995). This finding also suggests that an initial increase of clusterin level might be used as an indicator of tubular damage at a low level of ATN, and a reduction in clusterin staining over time as a sign of recovery (Chevalier et al., 1999).

KIM-1, initially identified in postischemic rat kidneys (Ichimura et al., 1998), is a type I transmembrane glycoprotein with Ig and mucin ectodomains. Its basal expression is very low in the kidney, but is highly upregulated in dedifferentiated and regenerating proximal tubular epithelial cells (Ichimura et al., 1998). In this study, the expression of KIM-1 after treatment was often seen in regenerating tubules with both apical and cytoplasmic patterns, as described by others (Ichimura et al., 2004). The intensity and quantity of KIM-1 positive signal also correlated with the amount of damage observed histologically, which could allow the use of KIM-1 as a tissue indicator of renal injury. In addition, KIM-1 was detected in the parietal epithelium of some glomeruli and in some proximal convoluted tubules, suggesting an additional function for the protein in areas distant from the direct injury.

Osteopontin is an acidic phosphoprotein and cell adhesion molecule isolated from tissues such as bone and body fluids (Rittling and Denhardt, 1999). Up-regulation of osteopontin has been observed in a variety of renal pathologies, but the precise role of the protein in nephropathies is currently unclear (Rittling and Denhardt, 1999). In this study, the normal expression in the kidney followed previously described patterns, being localized at the transition ISOM and OSOM (thin limb of the loop of Henle) and in the papillary epithelium (Persy et al., 1999). There was also a difference in the sub-cellular expression between the proximal and distal tubules after PAP treatment (Persy et al., 1999).

Osteopontin has been recognized in the kidney as an indicator of tubular regeneration, playing a role in the reepithelialization of the tubules (Persy et al., 1999), consistent with the time-dependent variation and distribution usually away from the obviously damaged tubules in this study. Interestingly, the normal pattern of expression of osteopontin in the papillary transitional epithelium diminished in those animals that showed hyperplasia of the transitional epithelium, suggesting that decreased osteopontin expression may indicate epithelial immaturity for that area.

Vimentin is an intermediate filament that is widely utilized as a protein tissue marker for mesenchymal cells (Franke et al., 1982). In the context of renal injury and regeneration, vimentin has been shown to be expressed in undifferentiated cells repopulating the tubular structures after injury (Witzgall et al., 1994; Chevalier et al., 1998). Similarly, in this study vimentin was expressed in areas of tubular regeneration, and the level of expression directly correlated with the extent of regeneration, with positive tubular cells detected as early as 24 hours in both high- and low-dose studies.

The Ki-67 antigen is reported as strongly associated with cell proliferation and expressed in the nucleus and nucleolus in all phases of the cell cycle (Endl and Gerdes, 2000). Here Ki-67 IHC was used to assess the proliferative status of the tubular cells after toxic damage. There was a clear concentration of labeled cells in and around injured renal areas as early as 24 hours post-PAP dosing, coinciding with the onset of regeneration. Animals with barely recognizable lesions (low-dose PAP at 48 hours) did not show the characteristic grouping of positive cells in areas of the MRs, indicating that regeneration and damage were tightly associated.

GCS catalyzes the rate-limiting reaction in the glutathione biosynthesis and is known to be feedback-inhibited by glutathione (Griffith, 1999; Wild and Mulcahy, 2000). Increased expression of genes encoding GCS is thought to provide cytoprotection against oxidative stress, and is frequently observed in response to inducers of Phase II metabolizing enzymes, including oxidants, heavy metals, and glutathione-conjugating agents (Griffith, 1999; Wild and Mulcahy, 2000). Since PAP and some of its metabolites have been shown previously to conjugate and deplete cellular glutathione levels (Crowe et al., 1979; Fowler et al., 1991; Klos et al., 1992; Fowler et al., 1994), the decline of GCS protein levels following PAP treatment, most apparent at high dose, was unexpected, and cannot be explained by glutathione feedback inhibition. The initial wave of GCS repression could be attributed to cellular necrosis and/or active transcriptional downregulation of the genes encoding the enzyme (Jeyapaul and Jaiswal, 2000; Jardine et al., 2002; Bakin et al., 2005). Later on, during regeneration, GCS protein levels could still remain low due to the immature status of the regenerating tubular cells.

A lack of apoptosis in this model of ATN was suggested by an absence of cleaved caspase 3 signal. Apoptosis is a recognized mechanism of injury to the renal tubular epithelium in some cases of ATN and is triggered by multiple factors (Brady et al., 2004). For some compounds, such as cisplatin, apoptosis is a major player in toxic damage (Lieberthal et al., 1996). Despite the lack of positive results in this model, a caspase 3-independent apoptotic mechanism participating in the toxic injury cannot be completely ruled out (Kroemer and Martin, 2005).

Summary

In this work we studied in detail the chronology of the renal lesions induced by a single ip PAP administration at low (50 mg/kg) and high doses (150 mg/kg). Tubular dilation in the MRs was the earliest change detectable at low dose, while lesions of ATN were obvious in the MRs and OSOM at high dose. Rapid regeneration of the damaged tubules occurred at both doses and was detectable by H&E and IHC as early as 24 hours. At the clinico-pathological level, urinary α-GST was the best marker of tubular injury by virtue of its wide dynamic range and good chronological correlation with pathology. The IHC analysis revealed significant trends in protein expression associated with time course and severity of the lesions. Clusterin IHC may be a useful auxiliary tool to detect minimal damage in tubular cells at early stages of ATN, since it correlates with tubular injury and could potentially be quantified. The analysis of protein levels by IHC or other proteomics tools is an important complement to gene expression studies, allowing for a comprehensive evaluation of changes during toxicity studies.