Urinary exosomes from a mouse model of chronic tubulointerstitial kidney disease induced by chronic renal ischemia–reperfusion injury and nephrectomy

Animal care followed the U.S. National Institutes of Health and local Thai criteria for the use and treatment of laboratory animals. We used 6-week-old CD-1 mice purchased from the National Laboratory Animal Center, Nakhon Pathom, Thailand. The animal protocols used were approved by the Institutional Animal Care and Use Committee (IACUC) of the Faculty of Medicine, Chulalongkorn University. Bangkok, Thailand. We divided 14 mice into equalsized sham and Chr IR surgery groups. The baseline blood chemistry tests were conducted 2 weeks before nephrectomy (–2 wk) using blood samples obtained from a tail vein. One week later a Chr IR injury was made as previously described [10], with slight modifications as follows. In brief, the surgery was conducted in 2 stages. In the first stage at 1 week after baseline blood collection (–1 wk), an abdominal incision was made under isoflurane anesthesia and the left renal artery was clamped. The abdominal wall was closed layer-by-layer with nylon suture thread (4-0). The mice were kept in a housing cage for 50 min. Subsequently, the abdominal wall was opened, and the vascular clamp removed (the total time of ischemia was 55 min). Again the abdominal wall was closed layer-by-layer with suture thread (4-0). After 1 week (0 wk), a right nephrectomy was performed under isoflurane anesthesia via an abdominal incision. Blood was collected from tail veins at 2 wk and 6 wk after nephrectomy. A 24 h urine sample was collected using a mouse metabolic cage (Hatteras Instruments, Cary, NC, USA) and preserved a protease inhibitor, then stored at –80°C until use. The urine collection and body weight measurements were done 1 day before blood collection. All mice were humanely killed at 12 wk after the nephrectomy by cardiac puncture under isoflurane anesthesia and the remaining kidney excised and stored in formalin for the renal histopathology. In sham-surgery control mice, the renal arteries were simply identified and the abdominal wall was closed layer-by-layer.

Blood urea nitrogen (BUN) and serum creatinine (Scr) were measured using colorimetric assays (QuantiChrom kits DIUR-500 for urea assay and DICT-500 for creatinine assay; BioAssay Systems, Haywood, CA, USA), 24 h urine protein levels were measured using a Bradford assay (Bio-Rad, Hercules, CA, USA). The 24 h urine protein levels (24 h U protein) were calculated using the following equation, 24 h U protein = urinary protein × 24 h urine volume. Hematocrit (Hct) was measured using a microhematocrit method with a Hitachi 917 automated biochemistry analyzer (Roche Diagnostics, Indianapolis, IN, USA). The single kidneys removed at 12 wk were immediately fixed in 10% neutral buffered formalin solution for paraffin embedding, and then 4 μm thick sections deparaffinized, rehydrated, and stained using a Masson trichrome method to determine the fibrosis. Tubulointerstitial fibrosis was estimated at 200× magnification using the following semiquantitative criteria: 0, area of damage <5%; 1, areas of damage 5%–10%; 2, damage involving 10%– 25%; 3, damage involving 25%–50%; 4, >50% of the area being affected [11,12,13,14,15]. The severity of the glomerular injury was determined by the percentage of the injured glomeruli [11,12,13,14,15].

Urine collected for 24 h at 12 wk after nephrectomy or sham surgery was centrifuged at 1000 ×g for 10 min to remove debris, and stored at –80°C until isolation of urinary exosomes. Urinary exosomes were isolated by differential centrifugation (17,000 ×g for 15 min then 200,000 ×g for 1 h) as described previously [16,17]. In short, the exosome-associated proteins isolated from urine samples were suspended using an isolation solution (10 mM triethanolamine, 250 mM sucrose, pH 7.6). Tamm–Horsfall protein was depleted by incubation with dithiothreitol (200 mg/ mL) at 95°C for 2 min then samples stored at –80°C for western blot analysis. Of note, the urine of mice in the sham-surgery control group was pooled because of the inadequate urine volume for the exosome extraction for western blot analysis. By contrast, urine volumes from individual Chr IR-injured mice were sufficient for exosome extraction.

Exosome-associated proteins, isolated from individual mouse urine samples, were separated by 1-dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis and then transferred to polyvinylidene difluoride membranes. Then the membranes were probed overnight at 4°C with rabbit polyclonal antibody to tumor susceptibility gene 101 (TSG101) (Abcam, Cambridge, MA, USA). Peroxidase conjugated, affinity-purified donkey anti-rabbit immunoglobulin G (Jackson Immuno Research Laboratories, West Grove, PA, USA) was used as the secondary antibody. The antigen-antibody reactions were visualized using enhanced chemiluminescence (Amersham Biosciences, Little Chalfont, Bucks, UK) and light-sensitive film (BioMax XAR, Kodak, Rochester, NY, USA). The density of western blotting results was determined by ImageJ, version 1.36b (National Institutes of Health, Bethesda, MD, USA).

All data are expressed as mean ± SD. Differences between groups were analyzed using a t test or ANOVA. P < 0.05 was accepted as significant.

We evaluated renal function of the Chr IR-induced model of renal injury using the time course of BUN and Scr (Figure 1A and B). At baseline, 1 week before IR injury of the left kidney, BUN and Scr were

13 ± 6 and 0.2 ± 0.1 mg/dL, respectively. Then renal function was rapid deteriorated at 2 wk consistent with the renal mass reduction from the two-stage Chr IR surgery. BUN and Scr at 2 wk were 58 ± 22 and 0.6 ± 0.2 mg/dL, respectively. Subsequently, elevated levels of both BUN and Scr persisted, BUN and Scr at 12 wk were 59 ± 14 and 0.6 ± 02 mg/dL, respectively. TI fibrosis in renal histopathology, as evidence of renal injury, was seen as areas of blue connective tissue after Masson trichrome staining (Figure 2). Of note, there was a mild TI injury score without different glomerular injury compared with mice in the sham surgery group (Figure 2) demonstrating that the Chr IR-induced model of CKD was predominantly a model of TI fibrosis. In addition, we found that the Chr IR-induced model of CKD showed anemia as early as 2 wk with approximately 50% reduction from baseline, Hct reduced from 50 ± 7% to 26 ± 5% (Figure 1D). Subsequently, Hct improved, but failed to return to baseline levels; Hct at 12 wk was 38 ± 7%.

Figure 1

Figure 2

To determine whether there was another presentation of uremia, we evaluated the time course of mouse body weight (Figure 3A). Despite ischemia-reperfusion-injury of the left kidney with right nephrectomy, there was no weight reduction in the mice. However, body weight of mice affected by Chr IR did not increase within 4 weeks of the perioperative period from –2 wk to 2 wk; 33 ± 4 g to 33 ± 2 g, respectively. By contrast, the body weight of sham-operated mice increased significantly from 32 ± 2 g to 37 ± 3 g at –2 wk and 2 wk, respectively. After 2 wk, there was weight gain by both Chr IR-injured and sham-operated mice, but the gain differed. Weight gain from 2 wk to 12 wk in mice with Chr IR-injury was approximately 7.8 g (from 33 ± 2 g to 41 ± 5 g; 23% increase) and was significantly less than the gain in the mice in the sham-surgery group, which was approximately 10 g (from 37 ± 3 g to 47 ± 2 g; 27% increase). To explore whether there was sufficient urine volume for exosome extraction using the double centrifugation method, we measured the time course of 24 h urine volume (Figure 3B).

Figure 3

At 2 wk and 12 wk after surgery, 24 h urine volume was greater than at baseline and sufficient for exosome extraction. We selected urine volume at 12 wk to extract urine exosomes. Of note, the urine volume at 2 wk postoperatively was also sufficient for exosome extraction, but no fibrosis was seen in renal histopathology at 2 wk (data not shown). We found TSG101, a structural molecule of exosomes, in urine from mice with Chr IR-injury at 12 wk (Figure 4).

Figure 4

We produced a Chr IR-injury in mice as previously described with a few modifications in the time of ischemia and method of anesthesia [10]. We describe the characteristics of the model of CKD in terms of renal injury with serum markers, urine protein, and renal histology. We determined whether the 24 h urine production at 12 wk after nephrectomy, when renal interstitial fibrosis could be seen in histopathology, was sufficient for urinary exosome extraction using a centrifugation method. We found that Chr IR-injury resulted in persistent renal injury (as measured by BUN, Scr, and anemia), tubular proteinuria, failure to thrive, and renal interstitial fibrosis resembling that in patients with early TI fibrosis injury. Additionally, 24 h urine from mice with Chr IR injury at 12 wk was sufficient to extract exosomes as demonstrated by detection of TSG101 using western blotting.

The Chr IR-induced model of CKD followed the strategy of the remnant kidney model by reducing kidney function to <50% that of normal function [5]. Among various rodent models of CKD, the remnant kidney models are the most favorable. Other models of CKD, such as podocyte injury models, are generally used for genetic manipulation to examine glomerulopathy [18,19,20]. Mouse models of TI CKD are limited. We recently developed a Ch IR-induced model of CKD [10]. In the present study, we modified our previous protocol to use a longer time of ischemia to ensure development of renal injury. After Chr IR, mouse renal function rapidly deteriorated as early as 2 wk with high BUN and Scr compared with baseline renal function (Figure 1A and B). However, the severity of the renal injury was just moderate compared with the 5/6 nephrectomy mouse model of CKD induced by the partial renal excision [13]. However, the renal injury was sufficiently severe to maintain renal injury (Figure 1A and B) and showed renal TI fibrosis at 12 wk after surgery (Figure 2). This model might therefore be appropriate for early TI fibrosis as seen in patients with early CKD. Moreover, we observed anemia (Figure 1D) and failure to thrive (Figure 3A) consistent with clinical presentations in patients with CKD. The body weight of mice with sham surgery increased more than it did in mice with Chr IR injury. In parallel, Hct was lowest at 2 wk after nephrectomy, which might the result of a perioperative injury affecting interstitial erythropoietin synthesis (Figure 3B). The Hct then improved at 6 wk, but was persistently lower than baseline at 12 wk. Therefore, the Chr IR-induced model of CKD might also be a good model of CKD induced anemia and malnutrition.

There was less glomerular injury, but more prominent TI injury in the Chr IR-induced model (Figure 2) and a compatible level of proteinuria. It is interesting that proteinuria was detectable as early as 2 wk after Chr IR. The proteinuria at 2 wk tended to be higher than at 12 wk after nephrectomy, which might be explained by a postischemic effect. The Bradford protein assay detects both low-molecular-weight protein and albumin. Therefore is difficult to determine whether the proteinuria is prominently of tubular or glomerular dysfunction. Further studies of tubular function in this model will be of interest.

Urinary exosomes are 35–40 nM membrane bound vesicles with unknown function, excreted by several types of cells along the urinary system [21]. Because the components of the exosome and cell membranes are the same hydrophobic molecules, urinary exosomes are a method of sampling the hydrophobic membrane molecules from hydrophilic urine [22]. Moreover, the exosome membrane protects internal cytosolic molecules. Not surprisingly, urine exosomes are a good source of mRNA, miRNA, and transcriptional factors that usually deteriorate rapidly in extracellular environments [23,24]. Therefore, urinary exosomes are a new source of biomarkers [25]. However, the discovery of biomarkers requires appropriate animal models that mimic specific diseases and conditions. The Chr IR-induced mouse model of CKD appears representative of early chronic TI fibrosis. We took advantage of the presentation of diuresis in the Chr IR-induced model to extract exosomes from urine. Mouse urine volume increased approximately 1.5 times from baseline at 12 wk after nephrectomy (Figure 3B). The standard double centrifugation method for exome extraction requires at least 4.5–5 mL urine [16,17]. To detect urinary exosomes, we chose TSG101, a molecule found in the exosome production pathway [21]. With time normalized to 24 h, the level of TSG101 was greater at 12 wk after nephrectomy than it was at baseline (Figure 4). However, we could not conclude that there were more exosomes at 12 wk after nephrectomy because there might be more TSG101 in individual exosomes at 12 wk than in urinary exosomes at baseline. Further experiments to determined exosome numbers are needed. However, urinary exosome TSG101 seems to increase with the progression of CKD in the Chr IR-induced model of CKD, and TSG101 itself is a candidate biomarker. Further studies are required to determine whether TSG101 is a biomarker.

We demonstrated that Chr IR injury in a male CD-1 mouse induces a model of TI fibrosis because: (1) there were persistent kidney injury as shown by an increase BUN and Scr, (2) TI fibrosis is seen in renal histopathology, (3) proteinuria is consistent with that in patients with TI injury, and (4) anemia and reduced growth rate resemble uremic symptoms. We found that 24 h urine from mice with the Chr IR-induced injury was sufficient for urinary exosome extraction and contained TSG101. The Chr IR-induced mouse model of CKD is appropriate for the study of several topics such as tubulointerstitial CKD, uremic malnutrition, and CKD induced anemia.