Contrast-induced acute kidney injury (CI-AKI) can occur in 3–25% of patients receiving radiocontrast material (RCM) despite appropriate preventive measures. Often patients with an atherosclerotic vasculature have to receive large doses of RCM. Thus, animal studies to uncover the exact pathomechanism of CI-AKI are needed. Sensitive and specific histologic end-points are lacking; thus in the present review we summarize the histologic appearance of different rodent models of CI-AKI. Single injection of RCM causes overt renal damage only in rabbits. Rats and mice need an additional insult to the kidney to establish a clinically manifest CI-AKI. In this review we demonstrate that the concentrating ability of the kidney may be responsible for species differences in sensitivity to CI-AKI. The most commonly held theory about the pathomechanism of CI-AKI is tubular cell injury due to medullary hypoxia. Thus, the most common additional insult in rats and mice is some kind of ischemia. The histologic appearance is tubular epithelial cell (TEC) damage; however severe TEC damage is only seen if RCM is combined by additional ischemia. TEC vacuolization is the first sign of CI-AKI, as it is a consequence of RCM pinocytosis and lysosomal fusion; however it is not sensitive as it does not correlate with renal function and is not specific as other forms of TEC damage also cause vacuolization. In conclusion, histopathology alone is insufficient and functional parameters and molecular biomarkers are needed to closely monitor CI-AKI in rodent experiments.
Contrast-induced acute kidney injury (CI-AKI) is defined as an increase of >25% or >0.5 mg/dL (44
Use of contrast materials in the USA. The most common cause of intravenous iodinated contrast media [
Although hypoxia of the renal medulla [
The
Viscosity and osmolality of the 3 generations of radiocontrast materials (RCM). The iodine/molecule ratio is 1.5 : 1 in high-osmolality contrast media (HOCM), 3 : 1 in LOCM (tri-iodinated molecules), and 6 : 1 in IOCM dimers [
Osmol. group | Name | Chem struct | Viscosity (mPa) | Osmolality (m) | Year intro. | ||
---|---|---|---|---|---|---|---|
(intro.) | Chemical | Brand | (20°C) | (37°C) | mOsm/kg H2O | ||
Isoosmotic (IOCM) (1990s) | Iodixanol | Visipaque | Nonionic dimer | 26.6 | 11.1 | 290 | 1996 |
Iotrolan | Iovist | 6.8 | 9.5 | 320 | 1989 | ||
Low (LOCM) (1980s) | Ioxaglate | Hexabrix | Ionic dimer | 15.7 | 7.5 | 600 | 1985 |
Ioxilan | Oxilan | Nonionic monomer | 16.3 | 7.8 | 695 | 1995 | |
Iomeprol | Imeron | 15.6 | 8.1 | 726 | 1994 | ||
Iopromide | Ultravist | 22 | 9.5 | 770 | 1995 | ||
Iohexol | Omnipaque | 20.4 | 11.2 | 780 | 1985 | ||
Ioversol | Optiray | 18 | 8.5 | 792 | 1988 | ||
Iopamidol | Isovue | 20.9 | 9.8 | 796 | 1997 | ||
Iobitridol | Xenetix | 21 | 10 | 915 | 1994 | ||
High (HOCM) (1950s) | Diatrizoate | Crystographin Hypaque | Ionic monomer | 18.5 | 8.4 | 2000, 1550 | 1955 |
Metrizoate | Isopaque | NA | 3.4 | 2100 | 1959 | ||
Iothalamate | Conray | NA | 9 | 2400 | 1962 |
Anatomically, rodents generally have a one-papilla kidney compared to the multi-papilla (4–18) human kidneys. The anatomic zones are similar [
Comparison of mouse and human kidney. The one-papilla mouse kidney has a well developed outer stripe (a) (own picture), whereas this zone is much less prominent in the multipapilla human kidney (b) (courtesy of Attila Fintha, Semmelweis University, 2nd Department of Pathology) (magnification: 10x, PAS staining).
Functionally, the renal concentrating ability is higher in rodents than in humans, especially in mice (4000 mOsm/kg) [
Differences between human and rodent kidney, summarized from [
Human | Rabbit | Rat | Mouse | |
---|---|---|---|---|
Number of papillas | 7–9 | 1 | 1 | 1 |
Number of nephrons | 0.2–2 million | 30 000 | 25–35 000 | 10–14 000 |
Concentrating ability (mOsmol/kg) | 1200 | 1400 | 3000 | 4000 |
Glomerular diameter ( | 200 | 140 | 120 | 73 |
A single injection of iodine containing RCM (5 g/kg ioxilan) induces CI-AKI only in rabbits [
Rat and mouse models of CI-AKI.
Injury type (besides RCM injection) | Species | Advantage | Disadvantage | Ref. |
---|---|---|---|---|
| Pronounced medullary hypoxia | Multiple insults | ||
Indomethacin (+salt depletion ± UNX) | Rat | Complex, clin. relevant | CPN for all rat models | [ |
Indomethacin + L-NAME | Rat | Medullary hypoxia | ||
Indomethacin + L-NAME | Mouse | pathomechanistic | High drug dose needed | [ |
| Dehydration amplifies injury | Hydration state affects CI-AKI progression | ||
Dehydration (24 h) | Rat | [ | ||
Dehydration (72 h) | Mouse, Rat | [ | ||
Dehydration (24 h) + eNOS deficiency (KO) | Mouse | [ | ||
Dehydration (24 h) + Indomethacin + furosemide | Rat | [ | ||
Dehydration (24 h) + glycerol rhabdomyolysis | Rat | [ | ||
| Reliable models | Microsurgery experience | ||
| Short duration | Species differences | ||
Ischemia-reperfusion | Mouse | [ | ||
| Clinical relevance | Chronic protocol | ||
Diabetes (streptozotocin: STZ) | Rat | [ | ||
5/6 nephrectomies + dehydration (48 h) | Rat | [ | ||
Long term cholesterol feeding | Rat | [ |
clin.: clinically, UNX: Uninephrectomy, CPN: chronic progressive nephropathy, and eNOS: endothelial nitrogen monoxide synthase.
The classic rat model of CI-AKI includes inhibition of vasodilators with nitric oxide synthase (NOS) inhibition by 10 mg/kg N
Taken together, it is easier to induce CI-AKI in rabbits (single injection of RCM without any additional injury) than in rats and the most severe additional injury is required in mice (Table
Iodinated contrast media are eliminated almost entirely by glomerular filtration [
A general histopathological feature of CI-AKI is
Vacuolization in different rodent models of CI-AKI. (a, b) Tubular cell vacuolization in a CI-AKI rat (Sprague-Dawley) model. (a) Indomethacin + L-NAME + ioversol. (b) Normal rat kidney cortex (PAS, 400x, [
Vacuolization is often
tubular vacuolization per see does not cause loss of renal function, tubular vacuolization resolves spontaneously, more severe tubular damage may lead to the shedding of vacuolated cells into the urinary space. New cells replace the shed epithelial cells.
This
Vacuolization can be
Although the condition was named after the swelling of tubular epithelial cells, the reason for this swelling is not osmotic pressure but the formation of vacuoles [
Hydropic vacuolization is reversible [
Osmotic diuresis (e.g., induced by
Calcineurin inhibitors
CI-AKI vacuoles were located primarily in the proximal tubules and are lysosomes [
Despite high RCM doses, vacuoles were absent in healthy kidneys and no tubular necrosis or atrophy developed unless there was some concomitant or predisposing renal damage [
It is generally accepted that hypoxia plays an important role in the development of CI-AKI [
The contribution of reactive oxygen species (ROS) to CI-AKI pathology is widely accepted. ROS contribute to intrarenal vasoconstriction by scavenging NO. Endothelin also contributes to the vasoconstriction [
Tubular epithelial cells are the most sensitive to hypoxia. However, there are substantial regional differences in the severity of hypoxia. As detailed below, there is an inverse relationship between oxygen supply and need from outer cortex to inner medulla. Furthermore, with increasing distance from vasa recta oxygenation is decreasing. Due to these regional differences of oxygen supply and demand, histological changes are often focal or patchy and inhomogeneous in the postischemic or CI-AKI kidney. This
Renal tubular epithelial cells are the most sensitive to hypoxia due to their high metabolic demand. Furthermore, due to the countercurrent circulatory system of the kidney, the oxygen supply decreases towards the medulla as the oxygen demand increases. Thus, tubular epithelial cells are the first to suffer from hypoxic damage. Despite many papers describing hypoxia as an important contributor to CI-AKI,
Histopathology of CI-AKI models with RCM administration and hypoxia. (a–d) RCM + renal ischemia mouse model (unpublished own data) (PAS, 400x). (a, b) Hypoxic tubular damage in mice 24 h after Omnipaque 350 iv. alone but no additional ischemia: besides vacuolization and mild tubular cell injury, no necrosis can be observed. Proximal tubuli have an intact brush border. (c, d) Ischemic changes in mice after 22 min ischemia + Omnipaque 350 iv. + 24 h reperfusion: more severe tubular damage, flattening of tubular epithelial cells, loss of nuclei, dilation of tubular lumen, and cast formation demonstrate tubular necrosis. (e) RCM + indomethacin rat model. Necrotic tubular cells (arrow) and inflammatory cell infiltration
Direct tubular toxicity of RCM is considered to participate in the pathomechanism of CI-AKI [
Rodent CI-AKI models apply renal hypoxia to aggravate the kidney damage that is subclinical if RCM is given alone (Table
In these models a control group with renal ischemia/hypoxia but without RCM is necessary to differentiate the effects of RCM from clamping. The severity of ischemia/hypoxia has to be adjusted as too severe damage may prohibit the evaluation of RCM-induced pathology, whereas if the model is too mild, kidneys may remain unaffected.
A disadvantage is the fundamental difference between rodent renal ischemia-reperfusion injury and human hypoxic AKI. Important differences include the following: complete cessation of blood flow (anoxia) in rodent models versus reduced blood flow (hypoxia) in humans, and temperature during the anoxia/hypoxia is close to physiologic in rodent models, whereas it is often reduced in human AKI. Warm ischemia primarily affects the cortex and the outer stripe, whereas cold ischemia damages the inner stripe and the renal papilla [
In healthy (sham operated) kidneys tubules have narrow lumen in the cortex (Figure
A common pathomechanism in the nephrotoxicity of nonsteroidal anti-inflammatory drugs (
Prolonged (72 h) dehydration combined with RCM causes CI-AKI in mice [
Intramuscular glycerol injection-induced rhabdomyolysis is a model of acute renal failure. As the hydration status of the body during rhabdomyolysis significantly influences the development of renal failure, 24-hour water deprivation precedes glycerol injection in this model [
CKD is an important risk factor for CI-AKI. Thus, CKD rodent models plus iv. RCM injection is also used to model CI-AKI [
In summary, the most specific histopathological lesions in rodent CI-AKI models are vacuolization of tubular epithelial cells and medullary hypoxia. Necrosis is only present if other hypoxia triggers are also applied as part of the model. As histopathologic changes lack specificity it is a relevant marker but not sufficient enough. Thus, further functional parameters and molecular biomarkers should be included in CI-AKI animal studies for a comprehensive analysis of disease progression. As the injection of RCM alone does not cause overt AKI in rodents, multiple insults are necessary for inducing histopathological and functional decline. The difference in sensitivity between species and the correlation with renal concentrating ability suggests that high concentrating ability may protect from CI-AKI.
The authors have declared that no competing interests exist.
Support was provided to Péter Hamar from the Hungarian Research Fund: OTKA-ANN (FWF) 110810 and OTKA-SNN 114619. Péter Hamar acknowledges support from the Bolyai Research Scholarship of the Hungarian Academy of Sciences and the Merit Prize of the Semmelweis University.