In this study, we compared the traditional murine model with renal pedicle clamp with models that clamped the renal artery or vein alone as well as to a whole body ischemia-reperfusion injury (WBIRI) model. Male C57BL/6J mice underwent either clamping of the renal artery, vein, or both (whole pedicle) for 30 or 45 min followed by reperfusion, or 10 min of cardiac arrest followed by resuscitation up to 24 h. After 30 min of ischemia, the mice with renal vein clamping showed the mostly increased serum creatinine and the most severe renal tubule injury. After 45 min of ischemia, all mice with renal vasculature clamping had a comparable increase in serum creatinine but the renal tubule injury was most severe in renal artery-clamped mice. Renal arterial blood flow was most decreased in mice with a renal vein clamp compared with a renal artery or pedicle clamp. A 30-or 45-min renal ischemia time led to a significant increase in the protein level of interleukin-6, keratinocyte-derived chemokine (KC), and granular colony-stimulating factor in the ischemic kidney, but the KC was the highest in the renal pedicle-clamped kidney and the lowest in the renal vein-clamped kidney. Of note, 10 min of WBIRI led to kidney dysfunction and structural injury, although less than longer time clamping of isolated renal vasculature. Our data demonstrate important differences in ischemic AKI models. Understanding these differences is important in designing future experimental studies in mice as well as clinical trials in humans.
- renal vasculature clamp
- whole body ischemia
- blood flow velocity
acute kidney injury (AKI) caused by ischemia-reperfusion injury (IRI) is a common and important problem in both native kidneys and transplanted kidneys (24). The mortality during native kidney AKI is close to 50% in the intensive care unit, and AKI in early transplants leads to more rejections and worse long-term outcomes (11). Small-animal studies of AKI have been invaluable in elucidating mechanisms, but most therapeutic successes in models have not translated to humans. One possible reason is that these models differ too much from human disease. There is an important need for the development of more appropriate “clinical” models that more accurately simulate the complexity of human disease. Most experimental work on murine ischemic AKI has used renal pedicle (both artery and vein) clamping. However, most patients with native kidney AKI have only renal artery hypoperfusion during cardiac surgery (16, 21), and relatively few patients only have renal venous flow problems like thrombosis (6, 8, 20, 22, 29) or whole pedicle disease. We hypothesized that isolated kidney ischemia by renal artery occlusion would differ in the outcome of renal injury from kidney ischemia by renal vein or by pedicle occlusion. We therefore performed a direct comparison of the traditional renal pedicle-clamping model in mice with renal artery clamping or renal vein clamping alone. We also compared the traditional isolated kidney ischemia model with a newer whole body ischemia-reperfusion injury (WBIRI) model caused by cardiac arrest followed by cardiopulmonary resuscitation (CPR). We examined renal function, kidney histology, renal arterial blood flow velocity, and proinflammatory cytokine and chemokine protein production among these AKI models. We found that at 24 h after 30 min of renal vascular clamping, both kidney functional and histological injury were the most severe in mice with a renal vein clamp and the least severe in mice with a renal artery clamp. Mice with renal vein clamping had the largest decrease in renal blood flow compared with mice with renal pedicle or artery clamping. Meanwhile, renal proinflammatory mediator keratinocyte-derived chemokine (KC) production increased most in mice with a renal pedicle clamp compared with mice with a renal vein clamp alone. The WBIRI mice, despite only 10 min of ischemia, had marked changes in the kidney structure and function, although less than the effects of 30 min of isolated kidney ischemia. Use of the WBIRI model could be an improved model coupled with traditional models to better simulate human AKI, particularly in transplant patients receiving kidneys from donors with cardiac death.
MATERIALS AND METHODS
Male C57BL/6 mice, 6–8 wk of age, were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were maintained under pathogen-free conditions in the Johns Hopkins Medical Institutions animal facilities with air conditioning and 14:10-h light-dark cycles. Mice had free access to food and water during the experiments. All experimental protocols used in this study were approved by the Institutional Animal Care and Use Committee.
Murine AKI models.
We used four distinct mouse models of AKI in this study. Three were based on the traditional kidney clamp, i.e., isolated kidney IRI models, and one was a recently developed model of WBIRI. Isolated kidney IRI was induced by clamping the renal pedicle (13), artery (25), or vein alone. WBIRI was induced by 10 min of cardiac arrest followed by CPR. Cardiac arrest beyond 10 min of time did not permit significant survival of mice after CPR, and fewer than 9 min of arrest did not lead to an appreciable rise in serum creatinine. Briefly, in the kidney IRI model, mice were anesthetized with an injection of pentobarbital sodium (75 mg/kg body wt ip). Following abdominal incisions, the left and right renal pedicles, arteries, or veins were bluntly dissected under a surgical microscope and a nontraumatic vascular clip (Roboz Surgical Instrument, Gaithersburg, MD) with a pressure of 70 g was placed on each dissected renal pedicle, artery, or vein by an expert microsurgeon for 30 or 45 min. During the procedure, animals were kept well hydrated with 2 ml of warm saline (1 ml each when the clips were placed and another 1 ml when removed) and at a constant body temperature (37°C). After 30 or 45 min of ischemia, the clips were removed, the wounds were sutured, and the animals were allowed to recover. In the WBIRI model, to induce cardiac arrest, 2.8 μl/g body wt of cold 0.5 M potassium chloride (KCI) was administered to the mice via the jugular venous catheter. Cardiac arrest was confirmed by EKG showing no cardiac activity. At 10 min after the induction of cardiac arrest, the ventilator was turned on to start mechanical ventilation with a respiratory rate of 190 breaths/min and 100% oxygen. At the same time, 20 μl/g body wt of warm epinephrine (diluted in 0.9% saline with a final concentration of 16 μg/ml) was injected via the jugular venous catheter, and simultaneously chest compressions were initiated at a rate of 300 compressions/min. When restoration of spontaneous circulation was achieved, defined as a spontaneous heartbeat which was confirmed by ECG, cardiac massage was terminated. Once spontaneous breathing was achieved, mechanical ventilation was stopped and the endotracheal tube was removed after gentle suctioning. Catheters were then removed, and wounds were closed (4). As controls to isolated kidney IRI or WBIRI AKI models, sham-operated animals underwent identical surgical procedure in each model; however, neither kidney IRI nor cardiac arrest was induced. All mice were euthanized at 24 h after the surgical procedure for renal function, kidney histology, and kidney proinflammatory molecule measurements. Renal artery flow velocity was measured just before euthanasia with laser-Doppler.
Assessment of renal function.
Blood samples were obtained via cardiac puncture at 24 h after ischemia. Blood was centrifuged to collect serum, which was used to measure creatinine concentration as a marker of renal function with a Roche Cobas Mira Plus automated analyzer system (Roche Diagnostics, Indianapolis, IN) by using a Creatinine 557A kit (Sigma Disagnostics, St. Louis, MO).
Kidneys were dissected from mice, and tissue slices were fixed in 10% formalin. For histological examination, formalin-fixed tissues were embedded in paraffin and 4-μm sections were cut and stained with hematoxylin and eosin. Each kidney section was examined under a light microscope for renal tubular injury in a blinded fashion by a renal pathologist. The percentage of necrotic tubules of total tubules in each of at least 10 high-power fields in the cortex or medulla was counted and calculated, and the average of the percentage of tubular necrosis in total fields in each section was presented as the renal tubular injury score of each mouse.
Measurement of renal blood flow velocity.
Before euthanasia at 24 h after sham surgery or AKI procedure, mice were lightly anesthetized with pentobarbital sodium at 50 mg/kg and the renal blood flow velocity was measured using a Sequoia C256 ultrasound (Siemens, Mountain View, CA) with a 15-MHz linear array transducer (23). The mouse kidney was first imaged in the two-dimensional long-axis view. Color flow and velocity time integral were then obtained at a sweep speed of 200 mm/s. Three to five blood flow-velocity recordings were obtained from each mouse, and the average was used as individual data.
Kidney cytokine/chemokine protein analysis.
Protein levels of IL-1a, IL-1b, IL-2,IL-6, IL-10, TNF-α, IFN-γ, KC, granular colony-stimulating factor (G-CSF), and monocyte chemoattractant protein-1 (MCP-1) were detected in mice kidneys by using a Bio-Plex Protein Array System (Bio-Rad Laboratories, Hercules, CA) described elsewhere in depth (13). Briefly, a portion of snap-frozen tissue was homogenized in cell lysis buffer, and the homogenates were centrifuged at 12,000 rpm for 15 min at 4°C. Total protein concentration of each supernatant was determined using a Bio-Rad Protein Assay Kit, and each sample was adjusted to 500 μg/ml with cell lysis buffer. Each sample was first incubated with a mixture of microbeads for 90 min at room temperature (RT) followed by incubation with biotinylated detection antibodies for 30 min, then with streptavidin-coupled phycoerythrin for 10 min (RT). Finally, the samples were subjected to a flow cytometric system. All acquired data was analyzed using Bio-Plex Manager 3.0 software (Bio-Rad).
Data are expressed as means ± SE and are first compared by one-way ANOVA comparing four group comparisons (sham, renal artery, renal vein, renal pedicle, and WBIRI). Any significant difference between any two of the four groups was confirmed with a standard Student t-test. Statistical significance of a difference is defined as P < 0.05.
Renal function after AKI.
At 24 h after 30 or 45 min of renal artery, vein, or pedicle clamping or 10 min of cardiac arrest, all AKI mice had a significant increase in serum creatinine level compared with sham-operated mice. After 30 min of kidney ischemia, serum creatinine was most increased in renal vein-clamped mice and least increased in renal artery-clamped mice (renal vein vs. pedicle vs. artery, in mg/dl: 2.42 ± 1.45 vs. 1.65 ± 0.37 vs. 0.55 ± 0.56, P < 0.05, n = 6). After 45 min of kidney ischemia, serum creatinine was comparably increased among three groups of renal vasculature clamping over the sham operation [renal vein vs. pedicle vs. artery, in mg/dl: 2.82 ± 0.26 vs. 2.71 ± 0.30 vs. 2.79 ± 0.70, P = not significant (NS), n = 5]. However, after 10 min of whole body ischemia, the level of serum creatinine was significantly less increased (0.81 ± 0.25 mg/dl, n = 6) compared with 30 or 45 min of kidney ischemia except 30 min of renal artery clamping (Fig. 1).
Kidney histology after AKI.
All AKI mice had a higher percentage of renal tubular necrosis at 24 h after ischemia compared with sham-operated mice (Fig. 2). With 30 min of renal vasculature clamping, the percentage of tubular necrosis in the cortex was comparable among three groups of kidney IRI mice (vein vs. artery vs. pedicle: 10.5 ± 16.5 vs. 10.2 ± 13 vs. 8.2% ± 4.5, P = NS) but was not significantly higher in the medullary region in the renal pedicle- or vein-clamped group compared with the renal artery-clamped group (pedicle vs. vein vs. artery: 49.5 ± 3.2 vs. 43.7 ± 17.3 vs. 28.1% ± 14.3, P = NS, n = 6). Interestingly, with 45 min of clamping, the percentage of the tubular necrosis in the cortex or medulla was significantly higher in the renal artery-clamped compared with the renal vein-clamped group (artery vs. pedicle vs. vein, cortex: 51.2 ± 13.1 vs. 21 ± 9.0 vs. 14% ± 9.0, P = 0.008; medulla: 77 ± 10.3 vs. 50.6 ± 7.1 vs. 46.2% ± 9.2, P = 0.001, n = 5/group). However, the WBIRI mice had a significantly lower percentage of tubular necrosis compared with the kidney IRI mice (cortex: 3.83% ± 1.54, P = 0.009 vs. renal artery, n = 6; medulla: 22.9% ± 2.5, P < 0.02 vs. all renal vasculatures. n = 6) (Fig. 3).
Renal arterial blood flow velocity.
Before euthanasia at 24 h after sham surgery or AKI, mice were anesthetized with pentobarbital sodium to measure renal arterial blood flow velocity by using a Sequoia C256 ultrasound. As shown in Figs. 4 and 5, renal blood flow was significantly decreased after 30-min clamping of the renal artery, vein, or pedicle compared with the sham-operated mice [renal blood flow (in m/s): artery vs. vein vs. pedicle vs. sham: 0.79 ± 0.05 vs. 0.31 ± 0.04 vs. 0.45 ± 0.04 vs. 1.34 ± 0.03, P < 0.01, n = 6]. However, renal blood flow was significantly higher in the WBIRI mice compared with those in kidney IRI mice. (1.46 ± 0.01, n = 6, P < 0.001 vs. all kidney IRI groups) (Fig. 5).
Cytokines and chemokines in murine renal tissue.
We utilized a cytokine multiplex assay to determine protein levels of 10 cytokines/chemokines in renal tissue from sham-operated mice, kidney IRI mice with pedicle, artery, or vein clamping, or WBIRI mice. We found that after 30 or 45 min of ischemia, renal KC, IL-6, and G-CSF levels in all three groups were significantly increased compared with sham-operated mice (P < 0.01). The rise of KC level was significantly greater in the 30-min pedicle-clamped (17.62 ± 2.84 pg/ml) compared with the 30-min vein-clamped group (6.92 ± 1.77 pg/ml, P < 0.05) but not statistically significant compared with the renal artery-clamped group (10.63 ± 2.56 pg/ml, P = 0.07). There was no significant difference in the levels of IL-1a (not shown), IL-1b, IL-2 (not shown), IL-6, IL-10, TNF-α, INF-γ, G-CSF, and MCP-1 among the three kidney ischemic groups. The WBIRI mice had significantly increased levels of IL-10 (P < 0.05 vs. 30- or 45-min pedicle clamping) and decreased levels of IL-6 (P < 0.05 vs. 45-min artery or vein clamping) compared with the kidney IRI mice. The cytokine IL-1b, chemokine KC, and G-CSF were not measured in WBIRI mice (Fig. 6).
In the present study, we have demonstrated that in a widely used mouse model of moderate renal ischemia for 30 min following by reperfusion for 24 h, there is a marked difference in the effects of clamping the renal artery compared with clamping the renal vein or both the renal artery and vein. We also found significant differences in AKI between the isolated clamp models and a newer WBIRI model. Renal artery clamping alone causes the least kidney injury, and this correlates with the least decrease in renal blood flow. However, at longer time periods of ischemia (45 min), the differences between interventions were reduced. Given that most experiments in mouse ischemic AKI use the renal pedicle (artery and vein) clamp model, while most patients have arterial hypoperfusion alone during hypovolemic shock or cardiac surgery, these results are important in developing better mouse models of AKI to simulate human disease. Small-animal studies of AKI have been invaluable in elucidating mechanisms of disease (12). Most murine work on ischemic AKI has used the renal pedicle clamp model, and this model is quite different from the disease in humans who have either renal arterial (26) or occasionally venous (10) problems alone. We therefore performed a direct comparison of the effects of the traditional renal pedicle clamp model with either renal artery or renal vein clamping alone on renal function, kidney histology, renal blood flow, and renal proinflammatory-mediator proteins. We previously published a new murine model of AKI where 10 min of cardiac arrest in mice followed by CPR causes AKI in survivors (4). In this study, we also compared the traditional isolated kidney IRI models with this new clinically relevant WBIRI AKI model.
First, we examined renal function and histology in mice with renal artery or vein clamping for 30 min and compared with that in mice with renal pedicle clamping. We found that, compared with mice with renal pedicle clamping, mice with a renal vein clamp had significant increases in serum creatinine concentration and kidney histological injury score. Meanwhile, the mice with a renal artery clamp had a significantly smaller rise in serum creatinine concentration, but comparable kidney histological injury. To test whether the differences in renal dysfunction and histological injury correspond to the degree of severity of ischemic injury, we performed renal vascular clamping for 45 min in mice and examined renal function and histology at 24 h after surgery. We found that 45 min of renal vascular clamping produced similar renal dysfunction and histological injury among mice with renal artery, vein, or pedicle clamping. The WBIRI mice with 10 min of warm ischemia had marked kidney structural and functional changes which would not be seen at 10 min of isolated renal ischemia (data not shown). However, kidney changes in the WBIRI mice were generally less severe than that of mice that underwent the much longer isolated renal clamp times.
Whole renal pedicle clamping is usually used in mouse models due to the technical ease in the difficult operating conditions in mice. Although the outcomes of kidney injury from different renal vasculature clamping are thought to be different, there are few studies to compare these different models, especially in mice. We clamped the bilateral renal pedicle, artery, or vein alone of mice for 30 min following by 24 h of reperfusion, and we found that mice with renal artery clamping had the least kidney injury, while we saw the most injury in renal vein-clamped mice. Similar to our findings, Orvieto et al. (19) clamped the pig renal pedicle or artery for 2 h followed by 1–3 days of reperfusion and found that the serum creatinine level in pigs with renal artery clamping showed a lower increase than that in pigs with renal pedicle clamping.
We also measured renal blood flow velocity by Doppler ultrasound in the different groups. At 24 h after surgery, all mice with renal vascular clamping had a significant decrease in renal blood flow compared with sham-operated mice. Of note, mice with renal vein clamping had the most pronounced decrease in renal artery flow compared with mice with renal artery or pedicle clamping. In our WBIRI model, renal blood flow did not change much at 24 h. Interestingly, in most cases of transplant-delayed graft function with minimal kidney function, there is good renal large arterial flow. In a canine model of kidney IRI, Neely et al. (17) clamped the dog renal artery or pedicle for 1 h or renal vein for 30 min and measured renal blood flow immediately and 1 h after clamp release. They found that pedicle- and vein-clamped kidneys had significantly decreased renal blood flow at both time points compared with controls; dogs with renal artery clamping showed an immediate decrease in renal blood flow, but this returned to control levels after 1 h. Additionally, venous clamping was accompanied by kidney congestion with increased kidney weights at both time points. Handley and colleagues (9) compared AKI from unilateral renal artery occlusion with a combination of renal artery occlusion and aortic occlusion above the renal artery level. They found that the combination produced more profound renal ischemic injury and more decreased renal blood flow compared with renal artery clamp alone (9). Interestingly, renal function and blood flow in the nonoccluded kidney actually improved under the combination of aortic and renal artery occlusion. In a porcine model, Tracy et al. (27) found that the percentage of HbO2 in the kidney was significantly less decreased from baseline in renal artery occlusion than that in renal pedicle occlusion for up to 35 min during 60 min of occlusion; the HbO2 decrease then became similar at the end of occlusion, indicating a possible “ischemic window” in which artery occlusion may provide benefit over pedicle occlusion. However, after 25–30 min of reperfusion, the decrease in HbO2 in the renal artery- and pedicle-clamped groups was similarly improved up to baseline. In our bilateral kidney vasculature clamp model, while we did not measure HbO2 in postischemic kidneys, renal blood flow was still significantly decreased even after 24 h of reperfusion.
A decrease in renal blood flow is of critical importance in initiating and extending the pathophysiology of ischemic AKI. Under normal physiological conditions, Po2 in the kidney decreases as one moves from the outer cortex to the inner medulla (3). Regional alterations in renal blood flow that persist after the initial ischemic event play an important role in the extension phase of renal ischemic injury. During reperfusion, a reduction in total renal blood flow to 40–50% of normal has been reported in both animal models of ischemic acute renal failure and in human ischemic AKI (11a). Studies have demonstrated that a persistent reduction in renal blood flow contributes significantly to the diminished glomerular filtration rate observed in human renal allografts following ischemic AKI (1). These persistent perfusion deficits have been demonstrated to be of greater magnitude in the outer medulla than in the outer cortex or inner medulla in an animal model of ischemic AKI (15, 18).
To study the underlying mechanisms by which kidney injury varied in different renal vascular clamping models, we measured proinflammatory cytokine/chemokine proteins in the kidney at 24 h after ischemia. We found that IL-6, KC, and G-CSF were significantly increased in the ischemic kidney compared with sham-operated mice. KC was most increased in the 30-min pedicle-clamped kidney compared with the 30-min vein-clamped kidney but not the 30-min artery-clamped kidney, indicating that both ischemia and congestion contribute to production of these chemokines. Meanwhile, IL-6 and G-CSF levels were comparable among the three 30-min clamping groups. With 45 min of clamping, all these mediators were comparable among three groups. However, the WBIRI mice showed significant increases in both anti-inflammatory IL-10 and proinflammatory IFN-γ compared with the kidney IRI mice.
In summary, in the present study we demonstrate that in a traditional mouse model of isolated kidney IRI by clamping different renal vasculatures, renal artery occlusion produces less tubular injury than renal vein occlusion due to less kidney congestion. We also demonstrate that whole body ischemia after controlled cardiac arrest has both important similarities as well as differences from the isolated clamp model. Understanding the strengths and limitations of different mouse models of kidney IRI should help us to better understand and develop treatments for human kidney IRI.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant R21DK071792 (H. Rabb and M. Liu). H. Rabb was supported by a research gift from Emanuel Gonzalez-Revilla (Panama).
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: X.L., D.B., and C.T. performed experiments; X.L., M.L., and L.C.R. analyzed data; X.L. drafted manuscript; M.L. and K.L.G. interpreted results of experiments; M.L. prepared figures; M.L. and H.R. edited and revised manuscript; H.R. provided conception and design of research; H.R. approved final version of manuscript.
Present address of X. Li: Dept. of Urology, Dongzhimen Hospital, Beijing Univ. of Chinese Medicine. No. 5. HaiYunChang, DongCheng District, Beijing 100700, China.
- Copyright © 2012 the American Physiological Society