A folic acid-induced rat model of renal injury to identify biomarkers of tubulointerstitial fibrosis from urinary exosomes

Animal care followed the U.S. National Institutes of Health criteria for the care and use of laboratory animals in research, under protocols approved by the National Institute of Diabetes and Digestive and Kidney Diseases Animal Care and Use Committee (ASP No. K087-KDB-11–14) following guidelines issued by the U.S. Department of Health and Human Services, the U.S. Department of Agriculture, the Association for the Assessment and Accreditation of Laboratory Animal Care, and Office of Laboratory of Animal Welfare (assurance No. A4149-01). Male Sprague Dawley rats (250-300 g) were purchased from Charles River Laboratories (Germantown, MD, USA) and had free access to water and standard laboratory rodent food. Rats were administered FA in 0.3 M NaHCO ₃ (FA group, 250 mg/kg body weight) or 0.3 M NaHCO₃ alone (vehicle (Veh) group) using a single intraperitoneal injection (0.9 to 1.1 mL) (see Discussion for additional details). While there might be a difference in alkali administration between these groups, (i) the volume of the single injection was small, (ii) any effect on plasma pH would be quite transient, (iii) there was no obvious abnormality in mice after loading of this dose, and (iv) this is the standard protocol for many published studies using the FA-induced model of renal injury. Only FA-injected rats with blood urea nitrogen (BUN) ≥100 mg/dL at day 2 after injection were selected for further study (for animals with BUN <100 mg/dL, the development of TI fibrosis was inconsistent (data not shown).

In standard experiments using this FA-induced model of renal injury, samples of blood (drawn from the tail vein) and urine (24 h, collected using rat metabolic cages; Hatteras Instruments, Cary, NC, USA) were obtained for analysis and body weight was measured, before administration of either FA or Veh (day 0) and at days 2, 7, 14, and 28 after administration. We used 35 rats in the experiments: 20 rats were used to explore the model and 9 were used for the proteomic analysis. The remaining 6 rats failed to produce BUN >100 mg/dL and were excluded from the study.

For quantitative comparisons of putative biomarkers in urine and/or urinary exosomes, both specific protein measurements and normalization based on the total amount of protein in a single urine sample (i.e., a sample collected over a short time) are generally unsatisfactory [16]. First, normal physiological variations in water excretion can lead to relative dilution or concentration of urinary proteins in such samples, thereby affecting the total amount of protein present. Second, total protein excretion in urine can vary broadly among various pathophysiological states. Therefore, for accurate quantitative measurements from urine samples, it is necessary to measure the actual rates of excretion of both specific proteins (e.g., putative biomarkers) and total protein, and to ask whether the excretion rates differ between control and experimental populations. To minimize the effects of the physiological variables mentioned above, rates of excretion are generally measured over a 24 h urine collection period, and this analytical approach was termed “time-normalization”.

All urine samples were collected (over 24 h) using metabolic cages, protease inhibitors were added, and samples were centrifuged at 1000 × g for 10 min to remove debris [5, 17, 18]. From the supernatant of each urine sample, two aliquots (100 μL) were reserved for creatinine and protein analyses, then the remainder of each urine sample was divided into two equal aliquots (for western blot and proteomic analyses) and stored at -80°C until used.

BUN, serum creatinine (SCr), and urine creatinine (UCr) were measured using colorimetric assays (QuantiChrom kits DIUR-500 for urea assay and DICT-500 for creatinine assay; BioAssay Systems, Haywood, CA, USA). Urine protein level (24 h urine albumin) was measured using enzyme-linked immunosorbent assay kits (Nephrat, Exocell, Philadelphia, PA, USA). Because SCr does not reach a steady state during acute kidney injury, we directly calculated creatinine clearance (CCr) from 24 h urine collection of UCr.

The CCr was calculated using the following equation: CCr = (UCr * 24 h urine volume)/SCr. For histology, kidneys were fixed in 10% neutral buffered formalin solution for paraffin embedding and sectioning (4 mm thickness), then stained using a Masson trichrome method. TI fibrosis was evaluated at 200 × magnification by the following semiquantitative criteria to estimate area of damage: 0, <5%; 1, 5%-10%; 2, 11%-25%; 3, 26%-50%; 4, >50%.

For western blot analysis, exosomes from each urine sample were isolated separately; for proteomic analysis, all urine samples from each group were pooled before exosome preparation. Rat urinary exosomes were isolated by differential centrifugation (17,000 × g for 15 min then 200,000 × g for 1 h) as described previously [5, 17, 18]. Next, the isolated exosomes (i.e., the low-density membrane pellet after 200,000 × g centrifugation) were resuspended in an isolation solution (10 mM triethanolamine, 250 mM sucrose, pH 7.6) and depleted of Tamm-Horsfall protein by incubating with dithiothreitol (200 mg/ml) at 95°C for 2 min, followed by recentrifugation (200,000 × g, 1 h, 4°C). The final exosome pellets were dissolved in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer for western blot and proteomic analysis.

After isolation of exosomes from individual rat urine samples, exosomal proteins were separated by one-dimensional SDS-PAGE (12% gel), transferred to nitrocellulose by electroblotting, and probed with the following primary antibodies: mouse monoclonal antibody to ALG-2-interacting protein 1 (ALIX; also known as programmed cell death 6-interacting protein or PDCD6IP) (BD Biosciences, Franklin Lakes, NJ, USA); rabbit polyclonal antibodies to tumor susceptibility gene 101 (TSG101) (Abcam, Cambridge, MA, USA), and clusterin (Sigma Aldrich, St. Louis, MO, USA). Immunoreactivity was visualized with horseradish peroxidase-conjugated secondary antibodies (Sigma Aldrich) and enhanced chemiluminescence substrate (Amersham Biosciences, Little Chalfont, Bucks, UK) using light-sensitive film (Kodak BioMax XAR, Rochester, NY, USA). Band intensities were determined by ImageJ version 1.36b (National Institutes of Health, Bethesda, MD, USA).

Solubilized exosomes from pooled rat urine samples (from 3 rats in each group) were fractionated by SDS-PAGE (12%), digested with trypsin, and analyzed by nanospray LC–MS/MS as described previously [5]. Briefly, each sample lane was cut into eight pieces (from top to bottom), and each gel piece was minced and subjected to in-gel tryptic digestion as follows: washed (25 mM NH₄HCO₃ in 50% acetonitrile); rehydrated (10 mM dithiothreitol, 25 mM NH₄HCO₃) at 56°C for 1 h; alkylated (55 mM iodoacetamide in 25 mM NH₄HCO₃) at room temperature for 1 h; washed twice and rehydrated again at 4°C for 0.5 h; then finally incubated with the sequencing-grade trypsin (product No. V5111; Promega, Fitchburg, WI, USA) at 37°C overnight. Subsequently, peptides were extracted from the gel in 50% acetonitrile containing 5% formic acid, then desalted and concentrated using a ZipTip (Merck Millipore, Darmstadt, Germany), eluted with 0.1% FA and transferred to LC-MS/MS auto-sampler vials. The LC-MS/MS analysis was carried out using a ProteomeX workstation LTQ-MS (Thermo Finnigan, San Jose, CA, USA). The NCBI RefSeq protein database (Rattus norvegicus), appended with a list of common contaminants (pig and bovine trypsin, and serum albumin, and human isoforms of keratin), was searched using BioWorks version 3.1 (Thermo Finnigan) based on the SEQUEST algorithm. False discovery rate (FDR) was calculated using the target-decoy search strategy described by Elias et al. [19]. Low quality matches were filtered using a FDR threshold of <1%. Protein was quantified by spectral counting of all corresponding peptides from a given protein at each experimental time point. For this study, normalized spectral counts (total spectral counts divided by the amino acid number of a given protein) were used as a rough approximation of absolute protein abundance.

All data are expressed as mean ± SEM. Differences between or among groups were analyzed for statistical significance by one-way ANOVA (SPSS for Windows, version 14.0, SPSS Inc., Chicago, IL, USA) P < 0.05 was accepted as significant.

An acute change in renal function, as determined by an increase in BUN concentration and a decrease in CCr, was detected at day 2 after FA injection (Figure 1A and B). By day 7 after injection, both BUN and CCr concentrations demonstrated spontaneous recovery of renal function to the control level, and there were no subsequent changes in either BUN or CCr observed after day 7. By contrast, urine volume was substantially elevated at day 7 (and returned to normal by days 14 and 28), indicating that at least some aspects of renal function remained (at least transiently) abnormal (Figure 1C). Nevertheless, there was no change in albuminuria or body weight (data not shown), suggesting a lack of glomerular involvement in this model of renal injury, which is consistent with findings of other studies [14, 15, 20]. The occurrence of acute tubular injury in the present model was demonstrated histologically, with interstitial infiltration by inflammatory cells observed at 1 week after FA injection, and significant interstitial fibrosis (as determined by Masson trichrome staining) detected at both 1 and 2 weeks after FA injection (Figure 1D and E); inflammatory cell infiltration appeared diminished at 2 weeks after FA injection. In addition, the histopathology at day 28 was not different from day 14 (data not shown). Despite histological differences between FA- and Veh-treated rats, proteinuria levels (as a TI injury indicator) were not different (Figure 1F). This supports the necessity for better biomarkers specific for TI injury and fibrosis, and a discovery of such new biomarkers by proteomic analysis could provide novel information compared to conventional proteinuria levels.

Figure 1

Next, we conducted a preliminary feasibility study to evaluate the use of urinary exosomes to identify novel protein biomarkers of TI fibrosis.

Isolation of exosomes from urine (24 h collection) was confirmed by immunoblot detection of ALIX and TSG101, marker proteins for urinary exosomes [5, 17] (Figure 2A). After injection of FA, the amounts of ALIX and TSG101 exosomal proteins in rat urine were increased compared with the Veh-injected controls, and were approximately 3-fold higher at week 2 after injection (Figure 2B and C). The samples analyzed by immunoblotting were time-normalized, representing exosomes produced by FA- and Veh-treated rats during the same 24 h (see “Materials and Methods: Animals, model, and sample collection” for a complete explanation). Because the amounts of urinary exosomal marker proteins increased 3-fold after FA-injection, then either (i) 3-fold more exosomes were produced by FA-treated rats (i.e., content of ALIX and TSG101 per exosome is constant) or (ii) the content of ALIX and TSG101 per exosome was 3-fold higher in FA-treated rats (i.e., number of exosomes produced is constant). To distinguish between these two possibilities, in future studies we would need to quantify the number of exosomes (e.g., using nanoparticle tracking analysis).

Figure 2

We performed proteomic analysis of urinary exosomes from rats injected with Veh alone, and at 1 and 2 weeks after FA injection. We identified a combined total of 372 proteins from the exosomes of all three samples; approximately one-third of these proteins (134) were common to all three samples (Figure 3; see Supplemental Data at http://sysbio.chula.ac.th/Supplemental_Data/Folic_acid-induced_rat_kidney_fibrosis_2016/Supplemental_Data.xlsx for a complete list of the proteins identified and their associated information). In the present study, an average of approximately 250 proteins was identified in each exosome sample (range about 200-320). The most abundant proteins identified in the exosome samples (as estimated from normalized spectral counts) are shown in Table 1. Among the 20 most abundant proteins identified in control exosomes, 19 were also found in both 1- and 2-week FA samples, and 5 are known to be membrane proteins predominantly or exclusively expressed in the kidney (Table 1). Furthermore, one of these proteins (aminopeptidase N) has been used previously as a urinary biomarker for renal damage [21]. Consistent with the immunoblotting results (Figure 2), the urine exosome marker protein ALIX was also detected exclusively in both exosome samples from FA-treated rats by proteomic analysis. TSG101, the other urine exosome marker protein confirmed by immuno-blotting, was only detected at very low abundance in a single exosome sample (FA-treated, 1 week) by proteomic analysis (Supplemental Data). Finally, 10 of the 25 proteins most frequently identified in exosomes [22] were also detected in all three samples from this study, and 17 of 25 were found in at least two samples.

Figure 3

Table 1

Comparison of the most abundant proteins (estimated from normalized spectral counts) identified in urine exosomal samples from vehicle-treated rats and rats 1 and 2 weeks after folic acid (FA) treatment.

Among the most abundant proteins identified uniquely in urinary exosomes from FA-treated rats (i.e., not present in control exosomes), the top 2 most abundant proteins (-1-antiproteinase and annexin A5) were common to both 1- and 2-week FA samples (Table 1). In addition, increased abundance in urine following acute kidney injury has been demonstrated previously for both of these proteins [23, 24]. With respect to potential use of urinary exosomes for biomarker discovery, the results for one of these proteins, annexin A5, were particularly interesting. In all, proteomic analysis identified 8 different annexin proteins, which could be grouped into 3 categories: (i) those found in relatively moderate abundance in all 3 samples; (ii) those found in low abundance and only in some samples; and (iii) annexin A5, which was absent in control exosomes, but was found in moderate abundance in both exosome samples from FA-treated rats (Table 2). This result is consistent with previous identification of annexin A5 as a potential urine biomarker for diseases of the kidney [24].

Table 2

Annexin family proteins identified in rat urine exosomes

Another interesting protein identified uniquely in urinary exosomes from FA-treated rats was clusterin (Table 1 and Supplemental Data, which has been previously reported as a biomarker for acute kidney injury [25] and renal fibrosis [26]. In the present study, we confirmed the proteomic result of significantly increased clusterin after FA administration by immunoblotting (Figure 4).

Figure 4