Pathophysiology of unilateral ischemia-reperfusion injury: importance of renal counterbalance and implications for the AKI-CKD transition

Abstract

Unilateral ischemia-reperfusion (UIR) injury leads to progressive renal atrophy and tubulointerstitial fibrosis (TIF) and is commonly used to investigate the pathogenesis of the acute kidney injury-chronic kidney disease transition. Although it is well known that contralateral nephrectomy (CNX), even 2 wk post-UIR injury, can improve recovery, the physiological mechanisms and tubular signaling pathways mediating such improved recovery remain poorly defined. Here, we examined the renal hemodynamic and tubular signaling pathways associated with UIR injury and its reversal by CNX. Male Sprague-Dawley rats underwent left UIR or sham UIR and 2 wk later CNX or sham CNX. Blood pressure, left renal blood flow (RBF), and total glomerular filtration rate were assessed in conscious rats for 3 days before and over 2 wk after CNX or sham CNX. In the presence of a contralateral uninjured kidney, left RBF was lower (P < 0.05) from 2 to 4 wk following UIR (3.6 ± 0.3 mL/min) versus sham UIR (9.6 ± 0.3 mL/min). Without CNX, extensive renal atrophy, TIF, and tubule dedifferentiation, but minimal pimonidazole and hypoxia-inducible factor-1α positivity in tubules, were present at 4 wk post-UIR injury. Conversely, CNX led (P < 0.05) to sustained increases in left RBF (6.2 ± 0.6 mL/min) that preceded the increases in glomerular filtration rate. The CNX-induced improvement in renal function was associated with renal hypertrophy, more redifferentiated tubules, less TIF, and robust pimonidazole and hypoxia-inducible factor-1α staining in UIR injured kidneys. Thus, contrary to expectations, indexes of hypoxia are not observed with the extensive TIF at 4 wk post-UIR injury in the absence of CNX but are rather associated with the improved recovery of renal function and structure following CNX.

INTRODUCTION

There is increasing recognition that acute kidney injury (AKI), particularly if severe, is often associated with incomplete recovery and residual structural and functional defects that contribute to chronic kidney disease (CKD) (i.e., the AKI-CKD transition) and eventually to progressive nephron loss and end-stage renal disease (ESRD) (3, 10, 66). Renal ischemia-reperfusion (IR) models are commonly used to investigate AKI-CKD transition mechanisms (26). Following IR, and depending on the initial severity of injury, the decline in glomerular filtration rate (GFR) is usually most pronounced between 24 and 72 h and is followed by recovery over the next 7–14 days (2). As dead epithelium is shed from injured tubules, repair and regeneration occur through a coordinated series of events that include initial dedifferentiation followed by proliferation and finally redifferentiation (65, 66). By the end of this phase (~14 days), most damaged tubules have recovered and regained essential function. However, depending on the severity of renal IR injury, recovery is often not complete, with residual areas and foci of inflammation, fibrogenic cytokine production, fibroblast proliferation, excessive extracellular matrix deposition, and eventually tubulointerstitial fibrosis (TIF). The loss of peritubular capillaries and consequent hypoxia have been postulated to play a major pathogenic role in the AKI-CKD transition and subsequent CKD progression (4, 62).

Given that the AKI-CKD transition pathology is proportional to the initial severity of IR injury, unilateral IR (UIR) models have gained increased investigative favor (13, 26, 37, 59). This is because the presence of an uninjured contralateral kidney prevents death of the animal, and, thus, longer ischemia times can be used, resulting in more severe and reproducible injury. Moreover, with UIR models, the progression to ESRD can be investigated within a relatively short time period. For these reasons, UIR injury has been recommended as one of the more useful models to investigate the pathogenesis of the AKI-CKD transition (26, 37). Indeed, some have equated the UIR and bilateral ischemia-reperfusion (BIR) models and considered the UIR model to represent a more severe form of BIR injury pathophysiology and accelerated AKI-CKD transition (68–72). However, there are data that indicate that the pathophysiological mechanisms contributing to the AKI-CKD transition may be very different between UIR and BIR models. Unlike an acute injury phase followed by recovery, albeit incomplete, that characterizes BIR models, Koletsky (36) and Finn (22) have shown that UIR models exhibit a phase of progressive renal parenchymal atrophy and TIF following apparent recovery. They also showed that if the normal and hypertrophied contralateral kidney was removed 2–3 wk after UIR injury, at a time when the ipsilateral kidney had started to exhibit progressive atrophy, the ipsilateral kidney exhibited significant recovery. This phenomenon was originally described in the context of renal structural recovery following relief of complete unilateral ureteral obstruction (UUO) of <2 wk duration by Hinman (32, 33) and termed “renal counterbalance.” In any event, such data clearly indicate that the degree of recovery after IR injury and the pathogenesis of AKI-CKD transition may be greatly influenced by the presence of a normal and hypertrophying contralateral kidney and thus likely to be different after UIR versus BIR. Nevertheless, few data exist regarding the time course of renal functional changes following contralateral nephrectomy (CNX) several weeks after UIR injury and/or the evolution of the phenotype or signaling characteristics of dormant tubules that can be rescued by CNX. Thus, the goal of the present study was to investigate the time course of hemodynamic changes in conscious rats as well as the renal tubule signaling pathways associated with the renal counterbalance phenomenon. We tested the hypothesis that CNX 2 wk after UIR injury improves renal functional and structural recovery and reduces indexes of hypoxia.

METHODS

Animals.

Experiments were performed on 6- to 8-wk-old male Sprague-Dawley rats obtained from Charles River Laboratory. Rats were fed standard 1% NaCl chow and provided water ad libitum. All animals were cared for in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and protocols approved by the Hines Veteran Affairs Institutional Animal Care and Use Committee.

Surgical procedures and experimental design.

As shown in Fig. 1, rats were anesthetized with pentobarbital sodium (50 mg/kg ip) and subjected to one of the following four IR procedures: 1) 40-min BIR, 2) 40-min left UIR, 3) 60-min left UIR, or 4) sham IR. Renal IR was performed by placing a vascular clamp on the renal pedicle(s) while core body temperature was maintained at 37°C via a servo-controlled heated surgical table, as previously described (45, 46). Kidneys were monitored for adequate reflow upon clamp release. One week following BIR, left UIR, or sham IR, all rats were surgically instrumented with a blood pressure (BP) transmitter (PA-C40, Data Sciences International) inserted into the abdominal aorta via the right femoral artery, a renal blood flow (RBF) probe (1PRB, Transonic) positioned on the left renal artery, and an osmotic minipump (2ML4, Alzet), containing FITC-inulin (40 mg/mL) dissolved in PBS implanted subcutaneously between the scapulae, as previously described (45). Beginning 5 days after surgical implantation of the BP transmitter, RBF probe, and osmotic minipump, BP and RBF were assessed (200 Hz) for 1–2 h over 3 consecutive days. Rats were then placed in metabolic cages overnight for urine collection, and a blood sample was obtained the next day for the assessment of GFR (45, 47). Two weeks following BIR, left UIR, or sham UIR, rats were anesthetized with pentobarbital sodium, and rats in the UIR or sham UIR groups were subjected to CNX (i.e., removal of the right uninjured kidney) or sham CNX. At 2, 5, 7 and 14 days following CNX or sham CNX, BP, left RBF, and GFR were assessed in a similar fashion as the pre-CNX or pre-sham CNX measurements. Two weeks following CNX or sham CNX (4 wk following BIR, left UIR, or sham UIR), rats were anesthetized, and kidneys were perfusion fixed using a periodate-lysine-paraformaldehyde solution.

Fig. 1.

Experimental groups used to investigate the hemodynamic and tubule signaling characteristics associated with the renal counterbalance phenomenon. Four groups of rats were subjected to the following various ischemia-reperfusion (IR) protocols: 1) 40-min bilateral IR (BIR), 2) sham unilateral IR (UIR), 3) 40-min left UIR, and 4) 60-min left UIR. One week following IR or sham IR, all rats were instrumented with a blood pressure (BP) radiotelemeter, left renal blood flow (RBF) probe, and osmotic minipump containing FITC-inulin for assessment of glomerular filtration rate (GFR). Five days later, BP, left RBF, and GFR were assessed in conscious rats for 3 days before contralateral nephrectomy (CNX) or sham CNX. At 2, 5, 7, and 14 days post-CNX or sham CNX, BP, left RBF, and GFR were again assessed in conscious rats. A subset of rats subjected to 60-min left UIR with and without CNX was administered pimonidazole (60 mg/kg ip) 2 days before tissue harvest to detect renal hypoxia. A different group of rats subjected to 60-min left UIR with and without CNX was administered Microfil to assess capillary density 4 wk postinjury.

One strength of the experimental design used in this study was the longitudinal assessment of BP and renal hemodynamics using gold standard techniques in conscious, chronically instrumented rats. This avoided the potentially confounding effects of anesthesia on systemic and renal hemodynamics. Indeed, previous studies have reported that anesthesia alters BP and causes the release of endocrine/paracrine factors (e.g., angiotensin II, etc.) (8, 35, 44) that can alter renal hemodynamics (8, 67). Moreover, anesthesia attenuates the spontaneous fluctuations in BP and RBF observed in conscious rats (6). Figure 2 shows these well-known effects of anesthesia on BP and RBF in a chronically instrumented rat.

Fig. 2.

Effects of anesthesia on blood pressure (BP) and renal blood flow (RBF) in a chronically instrumented rat. BP and RBF were recorded for 1 min before and for 1 min immediately after induction of anesthesia with isoflurane in a Sprague-Dawley rat that had been chronically instrumented with a BP radiotransmitter (PA-C40, Data Sciences International) and RBF probe (1PRB, Transonic) 1 wk prior. Note the decrease in absolute levels of mean arterial pressure (MAP) and RBF as well as their spontaneous fluctuations following anesthesia.

Assessment of renal hypoxia and capillary density.

Two days before rats were euthanized, a subset of rats in the 60-min left UIR group with (n = 4) or without CNX (n = 4) was administered pimonidazole (60 mg/kg ip, Hypoxyprobe) to detect hypoxic regions within the kidney where Po₂ levels were <10 mmHg (53). Pimonidazole was administered 2 days before harvesting of kidneys to avoid hypoxia artifacts due to anesthesia and surgery, which are commonly observed when it is administered 2 h before euthanasia (53).

In a separate group of rats that underwent 60-min left UIR with (n = 3) or without (n = 3) CNX, Microfil (Flow Tech) was administered 2 wk post-CNX or sham CNX (4 wk post-left UIR or sham UIR) to assess renal capillary density. As previously described (4), rats were anesthetized with pentobarbital sodium, and, following a midline incision, an aortic catheter was inserted into the abdominal aorta. A loose ligature was then placed on the aorta above the left and right renal arteries. Heparinized saline was then administered via the aortic catheter, and the animal was allowed to stabilize for 30 min. The renal circulation was then isolated by tying the proximal aortic ligature and Microfil [1:2 ratio of MV-130 red to MV diluent, 10% (vol/vol) catalyst, 5.5 mL/kidney] was infused with a syringe pump at a rate to maintain perfusion pressure at 150 mmHg. The left and right renal veins were then cut to ensure adequate perfusion of the kidney(s). Following administration of Microfil, kidneys were removed, cut into 1-cm pieces, and fixed/cleared according to the manufacturer’s instructions. Kidneys were then paraffin embedded, and 20-μm sections were made.

Morphology and immunohistochemistry.

Paraffin-embedded sections (5 μm) were deparaffinized and stained with Masson’s trichrome for the assessment of TIF. For immunohistochemistry, deparaffinized sections were heated to 99–100°C in 1 mM Tris-EDTA (pH 8.0) for 20–30 min, and endogenous peroxidase activity was neutralized with 3% H₂O₂ in water. Sections were blocked with 2.5% horse serum and incubated overnight with primary antibodies at 4°C. Primary antibodies included vimentin, aquaporin 1 (AQP1), phosphatase and tensin homolog (PTEN), VEGF, JG-12, hypoxia-inducible factor (HIF)-1α, and pimonidazole. Incubation with the primary antibodies (vimentin, AQP1, PTEN, VEGF, and JG-12) was followed by exposure to ImmPRESS horseradish peroxidase polymer-conjugated secondary antibodies (MP-7801 and MP-7802, Vector Laboratories) as previously described (46). Incubation with the primary antibody for pimonidazole and HIF-1α was followed by exposure to Dako and ImmPRESS (MP-7401) horseradish peroxidase polymer-conjugated secondary antibodies, respectively. Detailed information on all antibodies is shown in Table 1. The levels of pimonidazole and HIF-1α in renal tubules processed by immunohistochemistry were scored semiquantitatively in a blinded fashion by C. Rosenberger on a scale from 0 to 5, with 0 indicating the absence of HIF-1α or pimonidazole within tubules.

Table 1.

Details of antibodies used for immunohistochemistry experiments

Antibody	Company	Catalog No.	Clonality	Host Species	Immunogen Species	Dilution
Vimentin	ThermoFisher	MA5-11883	Monoclonal	Mouse	Pig	1:250
AQP1	Santa Cruz Biotechnology	sc-25287	Monoclonal	Mouse	Human	1:100
PTEN	Cell Signaling	9559	Monoclonal	Rabbit	Human	1:100
VEGF	Santa Cruz Biotechnology	sc-53462	Monoclonal	Mouse	Human	1:200
JG-12	Gift from Dontscho Kerjaschski		Monoclonal	Mouse	Rat	1:100
HIF-1α	Cayman Chemical	10006421	Polyclonal	Rabbit	Human	1:3,000
Pimonidazole	Hypoxyprobe	HP PAb2627	Polyclonal	Rabbit		1:50
ImmPRESS horse anti-mouse IgG PLUS	Vector Laboratories	MP-7401	Polyclonal	Horse	Rabbit	Ready to use
ImmPRESS horse anti-rabbit IgG PLUS	Vector Laboratories	MP-7801	Polyclonal	Horse	Rabbit	Ready to use
ImmPRESS horse anti-mouse IgG PLUS	Vector Laboratories	MP-7802	Polyclonal	Horse	Mouse	Ready to use
CSA II goat anti-rabbit	Agilent Dako	K1501	Polyclonal	Goat	Rabbit	Ready to use

AQP1, aquaporin 1; PTEN, phosphatase and tensin homolog; VEGF, vascular endothelial growth factor; HIF-1α, hypoxia-inducible factor-1α.

Statistical analyses.

Results are presented as means ± SE. A paired t test was used to compare left versus right kidney weight within each group. A Bonferroni correction was used to reduce the risk of a type I error for multiple comparisons. The adjusted significance level for the paired t tests was P < 0.0125. One-way ANOVA with Tukey’s post hoc comparison was used to assess differences in body weight and left kidney weight among the sham CNX + BIR groups and among the CNX groups. One-way ANOVA with Tukey’s post hoc comparison was also used to assess differences in right kidney weight across all groups as well as differences in HIF-1α and pimonidazole staining within various regions of the kidney across all groups. For data that were not normally distributed, a Kruskal-Wallis test with Dunn’s post hoc comparison was used. Two-way repeated-measures ANOVA with Tukey’s post hoc comparison was used to assess differences in RBF, GFR, mean arterial pressure (MAP), and renal vascular resistance (RVR) among groups of rats subjected to CNX or sham CNX. Unless stated otherwise, P < 0.05 was considered statistically significant.

RESULTS

Effects of CNX or sham CNX on body weight and kidney weight 4 wk following IR injury.

As shown in Table 2, there were no significant differences in body weight among all groups. Rats subjected to 60-min left UIR + sham CNX exhibited atrophy of the left kidney and hypertrophy of the right kidney with the left kidney being 40% smaller than the contralateral uninjured kidney. At 4 wk following a more modest injury produced by 40-min left UIR, the left kidney was 15% smaller than the contralateral kidney. Removal of the contralateral kidney 2 wk following 40- or 60-min left UIR led to the greatest degree of hypertrophy of the left kidney of all experimental groups at 4 wk postinjury. Of note, the magnitude of renal hypertrophy in both left UIR + CNX groups was significantly greater than that observed in rats subjected to sham UIR + CNX. BIR led to significant hypertrophy of both kidneys, with the left and right kidneys being 67% and 61% larger, respectively, compared with the left and right kidneys of rats subjected to sham UIR + sham CNX. Together, these results confirm the dramatic effects of a contralateral uninjured kidney to hinder renal growth and recovery following UIR injury, which differs from the hypertrophy response following BIR injury.

Table 2.

Body weight and kidney weight 4 wk following renal IR or sham IR

Experimental Group	Number of Animals/Group	Body Weight, g	Left Kidney Weight, g/kg body wt	Right Kidney Weight, g/kg body wt
Sham UIR + sham CNX	7	398 ± 11	4.2 ± 0.2	4.6 ± 0.2
Sham UIR + CNX	9	407 ± 13	5.7 ± 0.2
40 min BIR	6	391 ± 13	7.0 ± 0.7a	7.4 ± 0.6d
40-min left UIR + sham CNX	12	403 ± 8	4.7 ± 0.3	5.6 ± 0.2f
40-min left UIR + CNX	11	372 ± 9	10.4 ± 0.9c
60-min left UIR + sham CNX	10	414 ± 10	3.5 ± 0.2b	6.0 ± 0.3e,f
60-min left UIR + CNX	12	360 ± 16	10.2 ± 0.8c

Data are means ± SE. Body weight and kidney weight 4 wk following sham unilateral ischemia reperfusion (UIR), 40-min bilateral ischemia reperfusion (BIR), and 40-min left UIR or 60-min left UIR with and without contralateral nephrectomy (CNX) 2 wk post-UIR are shown. One-way ANOVA with Tukey’s post hoc comparison was used to compare body weight and left kidney weight among sham CNX + BIR groups and among CNX groups. This test was also used to compare right kidney weight among all groups. For data that were not normally distributed, a Kruskal-Wallis test with Dunn’s post hoc comparison was used. P < 0.05 was considered statistically significant for one-way ANOVA. A paired t test was used to compare left versus right kidney within the same group. Bonferroni correction was used to reduce the risk of a type I error for multiple comparisons. The adjusted significance rate for the paired t tests was P < 0.0125.

^aP < 0.05 vs. sham UIR + sham CNX and 60-min left UIR + sham CNX for left kidney weight;

^bP < 0.05 vs. 40-min left UIR + sham CNX for left kidney weight;

^cP < 0.05, 40-min and 60-min left UIR + CNX for left kidney weight vs. left kidney weight of all other groups;

^dP < 0.05 vs. right kidney weight of all other groups;

^eP < 0.05 vs. sham UIR + sham CNX for right kidney weight;

^fP < 0.0125 vs. the respective left kidney weight value.

Renal hemodynamics during recovery from UIR injury in rats with an intact contralateral kidney.

Striking differences in renal hemodynamics were observed between rats subjected to left UIR versus BIR. As shown in Fig. 3A, left, RBF was about twofold lower in both UIR groups compared with left RBF in rats subjected to BIR or sham UIR, which were not different from each other. The reduced left RBF following left UIR injury was entirely due to increased left RVR, as MAP was not statistically different among all groups without CNX. Left RVR was significantly higher and left RBF tended to be lower in rats subjected to 60- versus 40-min UIR injury. Moreover, left RVR progressively increased from 2 to 4 wk post-sham CNX in rats subjected to 60-min left UIR, reaching levels that were significantly higher at days 7 and 14 post-sham CNX compared with pre-sham CNX values. The increase in RVR from 2 to 4 wk post-sham CNX was associated with a progressive decrease in RBF during this time, although this did not reach statistical significance compared with pre-sham CNX values. There were no significant differences in GFR among all groups without CNX, likely due to the compensatory hypertrophy of the uninjured contralateral kidney in the UIR groups and of the uninjured or recovered hypertrophic nephrons in the BIR group. A summary of the two-way repeated-measures ANOVA main effects for the data shown in Fig. 3A is provided in Table 3.

Fig. 3.

Renal hemodynamics associated with the renal counterbalance phenomenon. Left renal blood flow (RBF), total glomerular filtration rate (GFR), mean arterial pressure (MAP), and calculated left renal vascular resistance (RVR) were assessed in conscious rats 11–14 days following sham unilateral ischemia-reperfusion (UIR), 40-min bilateral IR (BIR), 40-min left UIR, or 60-min left UIR and over 2 wk following sham contralateral nephrectomy (CNX; A) or CNX (B) 2 wk post-UIR. Data are mean ± SE; n = 7 sham UIR + sham CNX, n = 9 sham UIR + CNX, n = 6 40 min BIR, n = 12 40-min left UIR + sham CNX, n = 11 40-min left UIR + CNX, n = 10 60-min left UIR + sham CNX, and n = 12 60-min left UIR + CNX. Two-way repeated-measures ANOVA with Tukey’s post hoc comparison was used to assess differences in left RBF, GFR, MAP, and left RVR among groups of rats subjected to sham CNX or CNX. Main effects of the two-way repeated-measures ANOVA are shown in Table 3. #P < 0.05, sham UIR + sham CNX and 40 min BIR vs. 40- and 60-min left UIR + sham CNX groups; ##P < 0.05, sham UIR + CNX vs. 40- and 60-min left UIR + CNX groups; *P < 0.05 vs. the respective pre-sham CNX or pre-CNX value; **P < 0.05 vs. the respective 2-day post-CNX value; †P < 0.05, 40-min left UIR vs. respective 60-min left UIR sham CNX or CNX groups. Brackets and asterisks indicate all values 2–14 days post-CNX for all groups are significantly different from their respective pre-CNX values.

Table 3.

Main effects of two-way repeated-measures ANOVA for hemodynamic data shown in Fig. 3

	Left Renal Blood Flow	Glomerular Filtration Rate	Mean Arterial Pressure	Left Renal Vascular Resistance
Without CNX (Fig. 3A)
Group	P < 0.001	P = 0.329	P = 0.127	P < 0.001
Time	P = 0.952	P < 0.050	P = 0.685	P = 0.460
Interaction	P = 0.065	P = 0.130	P = 0.507	P < 0.050
With CNX (Fig. 3B)
Group	P < 0.001	P < 0.001	P < 0.010	P < 0.001
Time	P < 0.001	P < 0.001	P < 0.001	P < 0.001
Interaction	P = 0.601	P = 0.229	P < 0.001	P < 0.050

CNX, contralateral nephrectomy.

Effects of CNX 2 wk following left UIR on renal hemodynamics.

As shown in Fig. 3B, both left RBF and GFR increased within the first few days following CNX. In all groups, the compensatory increase in left RBF was complete within just 2 days following CNX. The compensatory increase in GFR in rats subjected to sham UIR was complete by day 5 post-CNX, whereas that of rats in the 40- and 60-min left UIR groups was complete by days 7 and 14, respectively, post-CNX. Two weeks following CNX, the percent increase in left RBF was 27% in the sham UIR group and 50% in both the 40- and 60-min left UIR groups compared with pre-CNX values. Since we could not assess GFR in the injured left kidney before CNX due to the presence of the uninjured contralateral kidney, the percent increase in GFR 2 wk post-CNX was compared with values obtained 2 days after CNX within groups. The percent increase in GFR was ~45% in both the sham UIR and 40-min left UIR groups after CNX. A 71% increase in GFR was observed in the 60-min left UIR group. Of note, these values likely underestimate the percent increase in GFR following CNX, because GFR could not be assessed in the left kidney before CNX. After CNX, significant differences in MAP were observed among groups, with 60-min left UIR > 40-min left UIR > sham UIR. These differences were expected given the increasing severity of UIR injury associated with the longer ischemia times. Finally, at 2 days post-CNX, left RVR decreased by 10%, 22%, and 29% in rats subjected to sham UIR, 40-min left UIR, and 60-min left UIR, respectively, compared with pre-CNX values. At 14 days post-CNX, left RVR was 21% and 29% lower than pre-CNX values in rats subjected to sham UIR and 40-min left UIR, respectively. In contrast, left RVR increased from 2 to 14 days post-CNX in rats subjected to 60-min left UIR. The elevation in left RVR was likely an autoregulatory response to the progressive increase in MAP observed during this time. A summary of the two-way repeated-measures ANOVA main effects for the data shown in Fig. 3B is provided in Table 3.

Effects of CNX 2 wk following 60-min left UIR on renal hypoxia.

Because 40-min left UIR was associated with more variability in renal injury at 4 wk postinjury, we examined the tubular signaling pathways of renal counterbalance in the 60-min left UIR model, which was associated with a more robust, consistent renal injury. We used both HIF-1α and pimonidazole staining within tubules of the cortex, outer stripe of the outer medulla (OSOM), inner stripe of the outer medulla (ISOM), and inner medulla (IM) 4 wk post-60-min left UIR to assess the effects of an uninjured contralateral kidney on renal hypoxia (Fig. 4). Interestingly, tubular HIF-1α expression was strikingly low within all regions of the left injured kidney of rats with an intact contralateral kidney. In contrast, abundant levels of HIF-1α were observed in tubules of all regions of the contralateral uninjured kidney. Moreover, CNX 2 wk after left UIR led to robust expression of HIF-1α. Indeed, HIF-1α levels in kidneys from the left UIR + CNX group reached values that were not statistically different from those in the OSOM and ISOM of the contralateral uninjured kidney. In all regions of the kidney except the OSOM, HIF-1α expression was significantly higher in the injured kidney of rats subjected to CNX compared with those with an intact contralateral kidney. In addition, the higher levels of HIF-1α expression were observed in hypertrophied, differentiated tubules that lacked vimentin expression, as shown in Fig. 5. In contrast, dedifferentiated tubules, identified by the presence of vimentin expression, exhibited minimal HIF-1α expression.

Fig. 4.

Indexes of hypoxia associated with the renal counterbalance phenomenon. Representative images of hypoxia inducible factor (HIF)-1α (A) and pimonidazole (B) staining in the right (n = 4) and left (n = 4) kidney from rats 4 wk following 60-min left unilateral ischemia reperfusion (UIR) + sham contralateral nephrectomy (CNX) and the left kidney (n = 4) from rats subjected to 60-min left UIR + CNX. CNX was performed 2 wk following UIR. Data are means ± SE. C: results of semiquantitative analysis of HIF-1α and pimonidazole staining. OSOM, outer stripe of the outer medulla; ISOM, inner stripe of the outer medulla; IM, inner medulla. One-way ANOVA with Tukey’s post hoc comparison was used to assess differences in HIF-1α and pimonidazole staining within various regions of the kidney across all groups. For data that were not normally distributed, a Kruskal-Wallis test with Dunn’s post hoc comparison was used. †P < 0.05 vs. the uninjured right kidney; *P < 0.05 vs. the left injured kidney in rats subjected to 60-min left UIR + sham CNX.

Fig. 5.

Coexpression of renal hypoxia inducible factor (HIF)-1α and vimentin 4 wk following 60-min left unilateral ischemia reperfusion (UIR) + right contralateral nephrectomy (CNX) 2 wk prior. Hypertrophied and differentiated tubules lacking vimentin expression (blue; A) had high levels of HIF-1α expression (brown; B). In contrast, atrophied, dedifferentiated tubules expressing vimentin generally lacked HIF-1α expression. In serial sections, symbols indicate the following: *, hypertrophied, differentiated tubules expressing HIF-1α but not vimentin; ♢, tubules with focal injury expressing both HIF-1α and vimentin; and ♦, atrophied, dedifferentiated tubules expressing vimentin but not HIF-1α.

In large part, levels of pimonidazole staining paralleled those of HIF-1α expression. Pimonidazole staining was lower or tended to be lower within the OSOM, ISOM, and IM of rats subjected to 60-min left UIR + sham CNX compared with the left kidney of rats subjected to 60-min left UIR + CNX. Similar differences in pimonidazole staining were observed between the left kidney of rats subjected to 60-min left UIR and the right uninjured contralateral kidney, except within the cortex and OSOM. Interestingly, minimal pimonidazole staining was observed in tubules in the cortex and OSOM within the contralateral uninjured kidney despite robust HIF-1α expression within these tubule segments. Such a discrepancy between HIF-1α and pimonidazole expression may be attributed to differences in the degree of hypoxia required to stimulate expression of HIF-1α (e.g., moderate hypoxia) versus pimonidazole accumulation (e.g., more pronounced or sustained hypoxia) (50). Nevertheless, these data clearly indicate that the improvement in renal function of the ipsilateral injured kidney following CNX is associated with robust expression of indexes of renal tubular hypoxia within the OSOM and ISOM, the regions of the kidney most susceptible to IR-induced AKI.

Effects of CNX 2 wk following 60-min left UIR on renal structural recovery.

As shown in Fig. 6, 60-min left UIR injury led to substantial levels of TIF in rats with an intact contralateral kidney. The extent of TIF was reduced in rats subjected to CNX 2 wk post-60-min left UIR injury. Moreover, a large number of tubules in injured kidneys of rats subjected to 60-min left UIR injury + sham CNX were in a dedifferentiated state. Serial sections showed that these tubules exhibited increased expression of the dedifferentiation marker vimentin and decreased expression of the differentiation markers AQP1 and PTEN. Interestingly, populations of tubules that exhibited both vimentin and AQP1 expression (red arrows) were seen in rats subjected to 60-min left UIR with an intact contralateral kidney. This expression phenotype might indicate tubules that would have recovered after UIR injury, but such recovery was hindered by the deleterious effects of a hypertrophied contralateral kidney. Indeed, CNX 2 wk after 60-min left UIR led to improved recovery of previously injured tubules, evident by decreased vimentin and increased AQP1 and PTEN expression. These data show that CNX 2 wk post-UIR injury improves recovery and redifferentiation of injured tubule epithelial cells and mitigates the development of TIF.

Fig. 6.

Renal fibrosis and tubule signaling characteristics associated with the renal counterbalance phenomenon. Representative images show renal tubulointerstitial fibrosis (TIF; trichrome staining) and serial sections of vimentin, aquaporin 1 (AQP1), and phosphatase and tensin homolog (PTEN) staining 4 wk following 60-min left unilateral ischemia reperfusion (UIR) + sham contralateral nephrectomy (CNX) or CNX 2 wk prior. Note the tubules in the left kidney of rats subjected to 60-min left UIR injury + sham CNX that are collapsed and positive for both vimentin and AQP1 (indicated by red arrows). These tubules may represent dormant tubules that can be rescued by CNX.

Effects of CNX 2 wk following 60-min left UIR on renal capillary density and tubular VEGF expression.

Capillary density, as shown by Microfil, was greatly diminished in injured kidneys 4 wk after 60-min left UIR + sham CNX (Fig. 7). This was consistent with the sparse JG-12 staining, another index of capillary density, observed in these kidneys (Fig. 7). Removal of the contralateral kidney 2 wk after 60-min left UIR partly restored capillary density. Moreover, VEGF expression in tubules paralleled the observed differences in capillary density among groups (Fig. 7). Thus, CNX 2 wk following 60-min left UIR mitigates the loss of peritubular capillaries and tubular VEGF expression that would have occurred had the contralateral kidney remained intact.

Fig. 7.

Renal capillary density and tubular vascular endothelial growth factor (VEGF) staining associated with the renal counterbalance phenomenon. Representative images show renal capillary density (Microfil and JG-12 staining) and tubular VEGF staining 4 wk following 60-min left unilateral ischemia reperfusion (UIR) + sham contralateral nephrectomy (CNX) or CNX 2 wk prior. Note the lack of tubule VEGF staining parallels microvascular rarefaction in the left injured kidney of rats subjected to 60-min left UIR + sham CNX.

DISCUSSION

Renal IR models are commonly used to investigate mechanisms of the AKI-CKD transition. The use of the UIR model, in which one kidney is subjected to IR in the presence of an uninjured contralateral kidney, has greatly increased in recent years because it results in more severe and reproducible injury compared with BIR models (26, 37). While it is generally acknowledged that the presence of an uninjured contralateral kidney contributes to worse outcomes following UIR versus BIR injury (18, 22, 25, 26, 68–72), the underlying mechanisms are not fully understood. Moreover, the mechanisms by which removal of the uninjured contralateral kidney, even several weeks after UIR injury, improves recovery and reduces the magnitude of the AKI-CKD transition remain even more unclear. The present study provides new insights into these issues by showing that striking differences in renal hemodynamics are observed between UIR and BIR models during the AKI-CKD transition phase. Moreover, upon removal of a contralateral uninjured kidney 2 wk after 60-min UIR injury, 1) compensatory increases in RBF occur early and precede that of GFR; 2) the magnitude of the compensatory increase in GFR is greater than RBF, indicating greater oxygen demand versus supply; and 3) extensive tubular hypoxia is observed in hypertrophic differentiated tubules that were rescued by CNX. These data strongly support the premise of the renal counterbalance theory proposed by Hinman (32, 33) almost a century ago and provide new insights into the mechanisms that may, in part, determine whether a tubule recovers from injury or undergoes atrophy and atresia.

Impaired acute and chronic renal hemodynamic responses have been suggested to contribute to the greater magnitude of the AKI-CKD transition following UIR versus BIR injury (18, 22, 26). Fernandez-Repollet and Finn (18) showed that, within the first 2 h following 60-min unilateral renal ischemia, the increases in RBF, GFR, and urine flow following clamp removal are enhanced if the contralateral kidney is removed before reperfusion. Such improvements in renal function were associated with reduced RVR and intratubular obstruction, two important modulators of the initial severity of AKI. Thus, it is possible that such factors may contribute to more severe AKI following a similar duration of ischemia in UIR versus BIR models. However, it has also been clearly established that the presence of a normal contralateral kidney negatively impacts recovery from UIR injury via factors that are independent of the severity of AKI (22, 36). Support for this concept is provided by evidence that removal of an uninjured contralateral kidney, or a reduction in its function, even 2–3 wk after UIR, drastically improves renal hemodynamics and structural recovery of the ipsilateral injured kidney (22, 36). Observations very similar to the UIR model have been made by Gobe et al. (27) in the two-kidney, one-clip model of unilateral renal artery stenosis. Indeed, CNX better mitigated the progressive atrophy of the stenotic kidney as opposed to removal of the clip without CNX. Thus, the injured kidney is capable of a more complete restoration of function and structure after UIR, but recovery is in some way hindered by the presence of a normal contralateral kidney. One of the deleterious effects of an uninjured contralateral kidney on recovery from UIR injury involves the redistribution of renal function (i.e., RBF and GFR) from the injured kidney to the uninjured kidney, as shown in the present study. Although such data are consistent with that originally proposed by Hinman (32, 33), the underlying mechanisms responsible for the redistribution of renal function between kidneys following UIR remain completely obscure.

Evidence that hemodynamic factors may be involved in the improved recovery following CNX 2 wk after 60-min UIR was first provided by Finn (22). He showed that at 4 wk post-UIR injury, renal plasma flow and GFR were twofold higher in previously injured kidneys in the absence versus presence of a normal contralateral kidney. Moreover, the higher GFR in rats subjected to CNX was due to the recruitment of previously dormant nephrons as opposed to hypertrophy of functioning nephrons, because single nephron GFR was similar whether the contralateral kidney was intact or not. The present data extend the findings of Finn by showing that the compensatory increase in RBF following CNX 2 wk post-UIR occurs rapidly, precedes that of GFR, and is associated with an increase in capillary density. These data indicate that early increases in RBF following CNX may be a critical upstream factor responsible for the recruitment of dormant nephrons and subsequent improvement in GFR and structural recovery after UIR injury.

In his counterbalance theory, Hinman suggested that the redistribution of renal function/work from an injured to an uninjured kidney contributed to a form of disuse atrophy or hypotrophy of tubules (31, 33). Similar observations were made by Koletsky (36), who noted that the progressive tubular atrophy following UIR injury was not a direct consequence of ischemic necrosis, because it was preceded by an initial period of repair of tubular necrosis that was evident by 4 days and essentially completed within the first 2 wk. A role for disuse atrophy in contributing to the worse renal outcomes following UIR injury with an intact contralateral kidney is logical, considering the relationship between renal function and size. Approximately 80% of total renal oxygen consumption is used to drive Na⁺-K⁺-ATPase for reabsorptive work (15, 58), which is dependent on the level of GFR. Moreover, reabsorptive work is a major stimulus for renal epithelial cell growth (19, 20, 61). The redistribution of RBF and GFR from the injured kidney to the uninjured kidney following UIR injury would be expected to decrease total renal Na⁺ reabsorption and oxygen consumption in the injured kidney and lead to renal atrophy. In contrast, the significant increase in RBF and GFR in the injured kidney following CNX would be expected to increase total renal Na⁺ reabsorption and oxygen consumption and lead to renal hypertrophy. Of note, the disuse atrophy concept discussed here is entirely consistent with observations following chronic obstructive single nephron wax blockade; that is, the tubule segments that exhibit the most pathology, both acutely and chronically, are those downstream from the obstruction rather than those upstream that are exposed to elevated intratubular pressures (14, 63). Collectively, these data suggest that some basal level of renal function may be necessary for normal recovery processes to occur.

Further support for a role of disuse atrophy in the context of renal counterbalance is demonstrated by the strikingly low levels of indexes of hypoxia observed in atrophic tubules in kidneys with significant TIF when the contralateral kidney was left intact. In contrast, robust expression of HIF-1α and pimonidazole staining was observed in hypertrophied, differentiated tubules with less TIF in rats subjected to 60-min UIR + CNX. Given that renal hypoxia is determined, in part, by the balance between renal oxygen delivery versus consumption (41), the vast presence of hypoxia in tubules of rats subjected to 60-min UIR that were rescued by CNX is consistent with the greater compensatory increase in GFR, and thus reabsorptive work, versus RBF observed in our study. Along the same lines, the extensive renal hypoxia noted in tubules in the uninjured, hypertrophied contralateral kidney can be explained by the continued redistribution of renal function from the injured to the uninjured kidney from 2 to 4 wk post-60 min UIR in rats without CNX.

Such dissociation between renal hypoxia and TIF may seem counterintuitive, given that hypoxia is widely believed to play a pivotal role in AKI, its transition to CKD, and the subsequent progression of CKD (4, 5, 7, 21, 26, 39, 42, 56, 62, 66). However, there is accumulating evidence showing a dissociation between hypoxia and renal injury. Rosenberger et al. (52) found decreased levels of renal tubular HIF-1α expression in patients with nonfunctional allografts as opposed to robust expression in patients with functional allografts. Similarly, lower levels of renal HIF-1α and pimonidazole expression have been observed in experimental models of severe versus moderate TIF (28, 40, 49, 50). Rosenberger et al. (51) have also shown that the degree of HIF-1α positivity is less extensive and intensive following renal IR injury compared with when renal oxygen delivery is reduced by the administration of carbon monoxide. Moreover, recent studies by Evans et al. failed to detect hypoxia in the subacute phase of severe renal IR injury (41) as well as in the early stages of adenine-induced CKD in rats (64). In their renal IR study, they also showed that renal HIF-1α levels were downregulated and suggested that the lack of hypoxia was due to greater levels of renal oxygen delivery versus consumption (i.e., GFR was suppressed more than RBF). Finally, Rempel et al. (48) recently demonstrated that chronic exposure to hypoxia attenuates renal injury in the remnant kidney model of CKD in rats. Collectively, these studies indicate that renal hypoxia and injury are not necessarily a cause-and-effect relationship and that the development of hypoxia is likely to be largely dependent on the balance between renal oxygen delivery and consumption. In any event, the role of hypoxia in contributing to AKI, AKI-CKD transition, and CKD progression remains to be established (16, 42).

Some caveats regarding the methods used in the present study to detect hypoxia warrant discussion. First, we did not directly measure renal oxygen tension using fluorescence optodes or Clark electrodes; thus, the actual levels of renal Po₂ were unknown. In addition to not directly assessing Po₂, pimonidazole adduct immunohistochemistry is prone to artifactual staining, especially in acutely damaged kidneys (1, 43). Finally, it is also possible that processing of pimonidazole and expression of HIF-1α are affected by factors other than oxygen levels, such as metabolic byproducts, in atrophic nonfunctioning versus hypertrophic functioning tubules (34, 38, 55). Nevertheless, the fact that HIF-1α and pimonidazole staining largely changed in parallel in the present study provides some degree of confidence in the observed differences in renal hypoxia among groups.

The present study indicates that hypoxia-induced HIF-1α expression improves renal function and structure in rats subjected to left UIR + CNX. The importance of renal hypoxia and upregulation of HIF-1α, the key transcription factor regulator of the canonical adaptive responses to hypoxia, on renal outcomes following injury are controversial (56). For example, in CKD models, experimental data support both renoprotective and injury-promoting roles of HIF-1α (40, 56). In contrast, the balance of evidence suggests that acutely injured kidneys benefit from the protective effect of hypoxia-mediated HIF-1α activation (17, 30, 40, 56, 57), consistent with the present study. This is congruent with the study by Conde et al. (11), who demonstrated that HIF-1α activation following renal IR injury in rats promoted proximal tubule epithelial cell survival and recovery. Similar to Rosenberger et al. (52), they also demonstrated that HIF1-α expression was primarily observed in healthy, functioning proximal tubules of human transplant recipients (11). Interestingly, the pattern of tubular HIF-1α staining in the present study paralleled that of VEGF, a major downstream target of HIF-1α. Tubular VEGF expression has been reported to signal to nearby peritubular capillaries (12). Thus, it is possible that the increased tubule HIF-1α expression following CNX led to an increase in VEGF expression and facilitated the repair or activation of injured/dormant capillaries that would have otherwise atrophied and undergone rarefaction (Fig. 8).

Fig. 8.

Proposed mechanisms whereby contralateral nephrectomy (CNX) 2 wk following left unilateral ischemia reperfusion (UIR) injury improves recovery of renal function and structure 2 wk later. If recovery from UIR occurs in the presence of a normal, hypertrophied kidney, our data show that decreases in glomerular filtration rate (GFR) exceed that of renal blood flow (RBF). This is expected to result in decreased renal oxygen consumption and minimal hypoxia, because oxygen delivery is greater than demand. Our data show that lack of hypoxia and reduced expression of hypoxia-inducible factor (HIF)-1α in such settings is associated with reduced expression of vascular endothelial growth factor (VEGF) and phosphatase and tensin homolog (PTEN) in tubules. This may contribute to progressive renal atrophy and fibrosis by promoting capillary rarefaction and preventing redifferentiation of tubules. In contrast, if CNX is performed 2 wk post-UIR injury, it results in greater increases in GFR versus RBF. As a result of the expected increase in renal oxygen consumption, hypoxia will develop and trigger upregulation of HIF-1α. We propose that the increase in HIF-1α promotes upregulation of VEGF and PTEN in tubules, which promotes capillary and tubule repair.

While the focus of the present study was primarily addressed to the effects of CNX after UIR injury in the context of the renal counterbalance phenomenon, there are data showing that factors other than CNX can improve renal outcomes after UIR injury. Finn et al. (24) administered either a low-salt (0.01%) diet or 1% NaCl as drinking water after UIR injury in which the contralateral kidney remained intact. Four weeks following UIR injury, GFR of the ipsilateral injured kidney was 20% versus 43% of the left kidney GFR of control rats in rats fed the low- versus high-salt diets, respectively. In addition, Zager and colleagues demonstrated that administration of an endothelin antagonist (70) as well as dexamethasone (71) improves recovery following UIR injury in mice despite the presence of a normal contralateral kidney. The beneficial effects of such treatments were attributed to a reduction in inflammation, lipid accumulation, and histone modification; however, it is also possible that hemodynamic and/or trophic effects, similar to those observed following CNX, contributed to the improved recovery following UIR injury. These studies provide evidence that interventions other than CNX that increase RBF, GFR, reabsorptive work, or epithelial cell growth can also improve recovery after UIR injury.

The present study has important implications regarding the use of the UIR model to investigate the pathogenesis of the AKI-CKD transition and subsequent progression to ESRD. Indeed, our study indicates that the pathophysiological mechanisms contributing to the AKI-CKD transition are very different between UIR and BIR models. In the setting of UIR, the deleterious effects of work redistribution from the injured kidney to the contralateral hypertrophied kidney is superimposed on an injured, recovering kidney. The major differences in RBF and kidney weight between the 40-min BIR versus UIR models in the present study reinforce such differences in the pathophysiology of AKI-CKD transition. Moreover, the mechanisms of CKD progression are also likely to be very different in the UIR versus BIR models. We (45) previously demonstrated that the development of hypertension and increased BP transmission to hypertrophied remnant glomeruli following IR-induced AKI of the entire renal mass is likely a major factor in the progression to ESRD. In contrast, the progression of CKD in the UIR model likely involves a progressive reduction in single nephron RBF and GFR, renal atrophy, and eventually TIF (29). Thus, the UIR model may be best suited to investigating the pathogenesis of AKI-CKD transition and CKD progression in clinical settings involving subtotal unilateral nephrectomy during renal artery clamping to remove cancerous tissue (60) as well as following the reversal of UUO or unilateral renal artery stenosis. Indeed, the UIR model may provide key insights into the enigma of capillary rarefaction observed in the poststenotic kidney in settings of unilateral renal artery stenosis (9, 54). Our data suggest that capillary rarefaction may be a normal response to greater deficits in GFR versus RBF in the poststenotic kidney, similar to that observed following UIR injury in the presence of an uninjured, hypertrophied contralateral kidney.

In summary, the present study provides new insights into the hemodynamic and tubule signaling pathways that contribute to impaired recovery and renal atrophy following UIR injury and the improvement in recovery and mitigation of atrophy and TIF following removal of the contralateral kidney. The data strongly support the basis of the “renal counterbalance” theory originally proposed by Hinman (31, 33). Moreover, the present study also has important implications regarding the potential role of renal counterbalance to operate in microenvironments of a single kidney, by which recovery from injury in subsets of nephrons is influenced by surroundings areas of normal hypertrophying nephrons, as previously suggested by Finn (23).

GRANTS

This work was supported by Veterans Administration Grant IK2BX001285, the National Kidney Foundation of Illinois, an American Society of Nephrology Carl Gottschalk Research Scholar Grant, and American Heart Association Grant 17AIREA33660433 (to A. J. Polichnowski), National Institute of Diabetes and Digestive and Kidney Diseases Grants DK-40426 (to A. K. Bidani), DK-61653 (to K. A. Griffin), and DK-104128 (to M. A. Venkatachalam), the Loyola University Medical Center (to A. K. Bidani), and the Else Kroener Fresenius Stiftung and The German Research Foundation (CRC Renoprotection) (to C. Rosenberger).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

A.J.P., K.A.G., C.R., M.A.V., and A.K.B. conceived and designed research; A.J.P., H.L.-V., R.L., C.R., and S.M. performed experiments; A.J.P., J.L., G.A.W., and C.R. analyzed data; A.J.P., K.A.G., M.M.P., G.A.W., C.R., M.A.V., and A.K.B. interpreted results of experiments; A.J.P. prepared figures; A.J.P. and A.K.B. drafted manuscript; A.J.P., K.A.G., G.A.W., C.R., M.A.V., and A.K.B. edited and revised manuscript; A.J.P., K.A.G., H.L.-V., R.L., M.M.P., J.L., G.A.W., C.R., S.M., M.A.V., and A.K.B. approved final version of manuscript.