Original Report: Laboratory Investigation American
Journal of
Nephrology
Received: November 20, 2013 Accepted: January 25, 2014 Published online: March 21, 2014
Am J Nephrol 2014;39:268–278 DOI: 10.1159/000360093
Assessment with Unenhanced MRI Techniques of Renal Morphology and Hemodynamic Changes during Acute Kidney Injury and Chronic Kidney Disease in Mice Zohar Milman a, b Jonathan H. Axelrod a Samuel N. Heyman c Nathalie Nachmansson a, b Rinat Abramovitch a, b
a The Goldyne Savad Institute of Gene Therapy, b MRI/MRS Lab HBRC and c Department of Internal Medicine, Hadassah Hebrew University Medical Center, Jerusalem, Israel
Key Words Acute kidney injury · Chronic kidney disease · Kidney imaging · Renal perfusion
Abstract Background/Aims: Changes in renal oxygenation and perfusion have been identified as common pathways to the development and progression of renal disease. Recently, the sensitivity of hemodynamic response imaging (HRI) was demonstrated; this is a functional magnetic resonance imaging (MRI) method combined with transient hypercapnia and hyperoxia for the evaluation of renal perfusion and vascular reactivity. The aim of this study was to utilize HRI for the noninvasive evaluation of changes in renal hemodynamics and morphology during acute, chronic and acute-on-chronic renal failures. Methods: Renal-HRI maps and true fast imaging with steady-state precession (True-FISP) images were used to evaluate renal perfusion, morphology and corticomedullary differentiation (CMD). MR images were acquired on two mouse models of kidney injury: adenine-induced chronic kidney disease (CKD) and rhabdomyolysis-induced acute
© 2014 S. Karger AG, Basel 0250–8095/14/0393–0268$39.50/0 E-Mail [email protected] www.karger.com/ajn
kidney injury (AKI). Serum urea was measured from these mice in order to determine renal function. Results: Renal-HRI maps revealed a blunted response to hypercapnia and hyperoxia with evolving kidney dysfunction in both models, reflecting hampered renal vascular reactivity and perfusion. True-FISP images showed a high sensitivity to renal morphological changes, with different patterns characterizing each model. Calculated data obtained from HRI and True-FISP during the evolution of renal failure and upon recovery, with and without protective intervention, closely correlated with the degree of renal impairment. Conclusions: This study suggests the potential combined usage of two noninvasive MRI methods, HRI and True-FISP, for the assessment of renal dysfunction without the potential risk associated with contrast-agents administration. HRI may also serve as a research tool in experimental settings, revealing the hemodynamic changes associated with kidney dysfunction. © 2014 S. Karger AG, Basel
The results presented in this paper have not been published previously, except in abstract format.
Rinat Abramovitch, PhD The Goldyne Savad Institute of Gene Therapy Hadassah Hebrew University Medical Center POB 12000, Jerusalem 91120 (Israel) E-Mail rinat @ hadassah.org.il
Downloaded by: Temple University 155.247.166.234 - 10/5/2016 2:38:51 PM
Acute kidney injury (AKI) is increasingly common among hospitalized patients, with a profound impact upon morbidity, mortality and expense [1, 2]. Chronic kidney disease (CKD) is an additional increasing medical problem as the life-expectancy of patients increases with the introduction of new treatment regimens. Since preexisting kidney disease is a major risk factor for AKI, acute-on-chronic renal failure is becoming highly prevalent among hospitalized patients [3]. Changes in renal oxygenation and perfusion have been identified as common pathways to the development and progression of renal disease [4, 5]. The radiologic assessment of changes in renal morphology and renal impairment is well established. The use of iodinated contrast media in computerized tomography (CT) and of gadolinium-based contrast-enhanced MRI (CE-MRI) greatly improves the imaging quality of renal morphology, and provides information regarding the perfusion and excretory capacity of the kidneys. Yet, the use of these imaging techniques is limited in AKI and CKD patients, since it harbors the risk of worsening renal function and the development of nephrogenic systemic fibrosis, respectively [6, 7]. In recent years, several noncontrast MRI methods for evaluating changes in renal oxygenation and perfusion have been developed and introduced to clinical practice. Non-contrast-enhanced MR angiography is gaining popularity [8], arterial spin labeling-based MR perfusion [9] can be used and phase-contrast measurements can provide flow profiles across renal arteries. Blood-oxygenated-level-dependent (BOLD) MRI provides additional important information regarding the renal oxygenation profile [10–13]. BOLD MRI uses the paramagnetic properties of deoxyhemoglobin to acquire images sensitive to local tissue oxygen concentration. As the deoxyhemoglobin concentration in blood increases, the T2 relaxation time of the protons decreases, resulting in signal intensity loss. A strong correlation has been demonstrated between renal BOLD MRI to tissue oxygen partial pressure (pO2). We have recently demonstrated the sensitivity of hemodynamic response imaging (HRI), a functional BOLD MRI, where transient alterations in inspired gases from normoxia to hypercapnia and subsequently to hyperoxia enable the evaluation of renal oxygenation, perfusion and vascular reactivity without the need for injected contrast agents [14]. This method, first applied in hepatic imaging [15], has been found particularly applicable in the kidney, where altered vascular reactivity may be both a hallmark Renal Diseases Assessment by Nonenhanced MRI
for and a causative parameter in the pathogenesis of renal impairment. A recent study showed that renal cortical thickness has a strong positive relationship with renal function [16]. Assessment of cortical thickness achieved with the help of corticomedullary differentiation (CMD) on MR images has been postulated to reflect differences in T1 values as a consequence of water content [16]. However, static MRI without the administration of contrast agent generally suffers from a small difference in signal intensity between the cortex and the medulla. In fact, diminished CMD on unenhanced MRI has been observed in patients with renal insufficiency [17]. True fast imaging with steady-state precession (True-FISP), an implementation of steadystate free precession, is an increasingly utilized MRI method that provides a high-resolution anatomical image in a short acquisition time [18]. True-FISP shows a strong contrast between tissues with different ratios of T2 and T1, providing excellent visualization of fluid compartments and vasculature. So far, it has been specifically utilized for the morphological imaging of the cardiovascular system [19]. Lately, steady-state free precession combined with time-spatial labeling inversion pulse was evaluated for CMD in the normal kidney [20]. The first aim of this study was to evaluate whether vascular reactivity is affected during the progression of AKI, CKD and in the development of acute-on-chronic renal failure, as evaluated by HRI. An additional aim was to establish the use of these nonenhanced MRI methods for the evaluation of changes in renal morphology, vascular reactivity and perfusion in the presence of acute or chronic renal impairment, without the need for administration of potentially harmful contrast agents. Materials and Methods Animal Models Animal experiments were performed in accordance with the guidelines and approval of the Animal Care and Use Institutional Committee, which holds NIH approval (OPRR-A01–5011), on CB6F1/OlaHsd mice (n = 60; Harlan, Jerusalem, Israel) that were 7–8 weeks old. Blood samples were obtained by tail-vein bleeding. Serum urea and creatine kinase levels were determined using the Reflotron system (Roche Diagnostics, Basel, Switzerland). AKI Model. Rhabdomyolysis-induced AKI (myoglobinuric AKI) was induced by glycerol 50% (8 ml/kg body weight) injected into the anterior thigh muscle of both hind limbs. MRI scans were executed repeatedly on days 1, 4, 8, 15 and 22 following glycerol injection (n = 15). In an additional set of experiments (n = 10), we assessed the proportion of altered MRI data in correlation with the extent of renal injury. For this, we used an interleukin-6 (IL-6)/soluble IL-6
Am J Nephrol 2014;39:268–278 DOI: 10.1159/000360093
269
Downloaded by: Temple University 155.247.166.234 - 10/5/2016 2:38:51 PM
Introduction
MRI MRI experiments were performed on a horizontal 4.7-tesla Biospec spectrometer (Bruker, Ettlingen, Germany) using a 3.5cm birdcage coil. Mice were anesthetized with intraperitoneal injection of pentobarbital (70 mg/kg; CTS Group, Hod-Hasharon, Israel) and placed in a supine position. Kidney morphology was assessed using both True-FISP [repetition time/echo time (TR/ TE) = 3/1.5 ms, field of view 3.4 cm, in-plane resolution 133 μm, slice thickness 1 mm and flip angle 60°; 16 averages] and T2weighted (T2W) fast spin echo (SE) images (RARE sequence; TR = 2,000 ms, TE = 37 ms, field of view 3.4 cm, in-plane resolution 133 μm and slice thickness 1 mm). The chosen TR/TE values may be suboptimal. It is possible that a longer TE would prove more suitable in future studies. Regions of interest (ROIs) were drawn manually by using Analyze-7.0 (BIR, Mayo Clinic, Rochester, Minn., USA) based on axial True-FISP images. On the True-FISP images, renal ROIs covered the cortex or outer medulla of each kidney (fig. 1b) whereas on HRI maps, the entire kidneys were assessed. Two additional ROIs covered the adjacent back muscle and the spleen for normalization of the renal values as shown previously [14]. Kidney area was calculated from the central axial slice of each kidney using Analyze-7.0 software. The HRI protocol was implemented as previously described [15], using axial T2W gradient echo images which are sensitive to the local tissue oxygenation (TR/TE = 147/10 ms, field-of-view 3.4 cm, in-plane resolution 133 μm, slice thickness 1 mm, 2 averages, 256 × 256 pixels, resulting in a temporal resolution of 37 s) during breathing of ambient air (normoxia), with hypercapnia (ΔSco2, 5% CO2) for 4 min, followed by hyperoxia (ΔSo2, 5% CO2 + 95% O2) for 4 min, administered through a face mask (4 liters/min). HRI maps were generated as reported previously [15] using a home-built IDL (Interactive Data Language of ITT Visual Information Solutions, Boulder, Colo., USA) program. Briefly, the percentage change in signal intensity (SI), induced by ΔSco2 and ΔSo2 was calculated from the average of 6 repetitions for each gas admixture, by including only pixels with a statistical threshold of p < 0.05 calculated with a Mann-Whitney U test. Mean renal HRI values were calculated for the entire kidney, from the central axial slice of each kidney. CE-MRI was applied as a reference method to assess renal perfusion, performed with T1W gradient echo sequence (TR/TE = 58.3/5.2 ms, field-of-view 3.4 cm, in-plane resolution 133 μm, slice thickness 1 mm, 1 average and 60 repetitions, resulting in a tem-
270
Am J Nephrol 2014;39:268–278 DOI: 10.1159/000360093
poral resolution of 7.4 s). Ten repetitions were acquired at baseline, followed by 50 repetitions after Gd-DTPA (0.5 M, 100 μl; Magnetol, Soreq Radiopharmaceuticals, Yavne, Israel) administration via a tail-vein catheter at a dose of 0.1 mmol/kg. Histological and Immunohistochemical Analysis Kidney sections, obtained parallel to MRI scans, were routinely stained with hematoxylin and eosin (HE) and Masson’s trichrome staining was performed to assess fibrosis in the CKD model. In addition, apoptosis was assessed by immunostaining using TUNEL (terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end-labeling), performed with an in situ cell death detection kit (Roche Diagnostics Corp. Mannheim, Germany). Bromodeoxyuridine (BrdU) immunostaining (GE Healthcare AmershamTM, UK) was performed to assess renal tubular cell proliferation, injected intraperitoneally (1 ml/100 g) 3 h before mice were sacrificed. Following target retrieval, sections were incubated with primary mouse monoclonal anti-BrdU antibody (Neomarker, Fremont, Calif., USA). Mapping of renal hypoxic regions (pO2
Journal of
Nephrology
Received: November 20, 2013 Accepted: January 25, 2014 Published online: March 21, 2014
Am J Nephrol 2014;39:268–278 DOI: 10.1159/000360093
Assessment with Unenhanced MRI Techniques of Renal Morphology and Hemodynamic Changes during Acute Kidney Injury and Chronic Kidney Disease in Mice Zohar Milman a, b Jonathan H. Axelrod a Samuel N. Heyman c Nathalie Nachmansson a, b Rinat Abramovitch a, b
a The Goldyne Savad Institute of Gene Therapy, b MRI/MRS Lab HBRC and c Department of Internal Medicine, Hadassah Hebrew University Medical Center, Jerusalem, Israel
Key Words Acute kidney injury · Chronic kidney disease · Kidney imaging · Renal perfusion
Abstract Background/Aims: Changes in renal oxygenation and perfusion have been identified as common pathways to the development and progression of renal disease. Recently, the sensitivity of hemodynamic response imaging (HRI) was demonstrated; this is a functional magnetic resonance imaging (MRI) method combined with transient hypercapnia and hyperoxia for the evaluation of renal perfusion and vascular reactivity. The aim of this study was to utilize HRI for the noninvasive evaluation of changes in renal hemodynamics and morphology during acute, chronic and acute-on-chronic renal failures. Methods: Renal-HRI maps and true fast imaging with steady-state precession (True-FISP) images were used to evaluate renal perfusion, morphology and corticomedullary differentiation (CMD). MR images were acquired on two mouse models of kidney injury: adenine-induced chronic kidney disease (CKD) and rhabdomyolysis-induced acute
© 2014 S. Karger AG, Basel 0250–8095/14/0393–0268$39.50/0 E-Mail [email protected] www.karger.com/ajn
kidney injury (AKI). Serum urea was measured from these mice in order to determine renal function. Results: Renal-HRI maps revealed a blunted response to hypercapnia and hyperoxia with evolving kidney dysfunction in both models, reflecting hampered renal vascular reactivity and perfusion. True-FISP images showed a high sensitivity to renal morphological changes, with different patterns characterizing each model. Calculated data obtained from HRI and True-FISP during the evolution of renal failure and upon recovery, with and without protective intervention, closely correlated with the degree of renal impairment. Conclusions: This study suggests the potential combined usage of two noninvasive MRI methods, HRI and True-FISP, for the assessment of renal dysfunction without the potential risk associated with contrast-agents administration. HRI may also serve as a research tool in experimental settings, revealing the hemodynamic changes associated with kidney dysfunction. © 2014 S. Karger AG, Basel
The results presented in this paper have not been published previously, except in abstract format.
Rinat Abramovitch, PhD The Goldyne Savad Institute of Gene Therapy Hadassah Hebrew University Medical Center POB 12000, Jerusalem 91120 (Israel) E-Mail rinat @ hadassah.org.il
Downloaded by: Temple University 155.247.166.234 - 10/5/2016 2:38:51 PM
Acute kidney injury (AKI) is increasingly common among hospitalized patients, with a profound impact upon morbidity, mortality and expense [1, 2]. Chronic kidney disease (CKD) is an additional increasing medical problem as the life-expectancy of patients increases with the introduction of new treatment regimens. Since preexisting kidney disease is a major risk factor for AKI, acute-on-chronic renal failure is becoming highly prevalent among hospitalized patients [3]. Changes in renal oxygenation and perfusion have been identified as common pathways to the development and progression of renal disease [4, 5]. The radiologic assessment of changes in renal morphology and renal impairment is well established. The use of iodinated contrast media in computerized tomography (CT) and of gadolinium-based contrast-enhanced MRI (CE-MRI) greatly improves the imaging quality of renal morphology, and provides information regarding the perfusion and excretory capacity of the kidneys. Yet, the use of these imaging techniques is limited in AKI and CKD patients, since it harbors the risk of worsening renal function and the development of nephrogenic systemic fibrosis, respectively [6, 7]. In recent years, several noncontrast MRI methods for evaluating changes in renal oxygenation and perfusion have been developed and introduced to clinical practice. Non-contrast-enhanced MR angiography is gaining popularity [8], arterial spin labeling-based MR perfusion [9] can be used and phase-contrast measurements can provide flow profiles across renal arteries. Blood-oxygenated-level-dependent (BOLD) MRI provides additional important information regarding the renal oxygenation profile [10–13]. BOLD MRI uses the paramagnetic properties of deoxyhemoglobin to acquire images sensitive to local tissue oxygen concentration. As the deoxyhemoglobin concentration in blood increases, the T2 relaxation time of the protons decreases, resulting in signal intensity loss. A strong correlation has been demonstrated between renal BOLD MRI to tissue oxygen partial pressure (pO2). We have recently demonstrated the sensitivity of hemodynamic response imaging (HRI), a functional BOLD MRI, where transient alterations in inspired gases from normoxia to hypercapnia and subsequently to hyperoxia enable the evaluation of renal oxygenation, perfusion and vascular reactivity without the need for injected contrast agents [14]. This method, first applied in hepatic imaging [15], has been found particularly applicable in the kidney, where altered vascular reactivity may be both a hallmark Renal Diseases Assessment by Nonenhanced MRI
for and a causative parameter in the pathogenesis of renal impairment. A recent study showed that renal cortical thickness has a strong positive relationship with renal function [16]. Assessment of cortical thickness achieved with the help of corticomedullary differentiation (CMD) on MR images has been postulated to reflect differences in T1 values as a consequence of water content [16]. However, static MRI without the administration of contrast agent generally suffers from a small difference in signal intensity between the cortex and the medulla. In fact, diminished CMD on unenhanced MRI has been observed in patients with renal insufficiency [17]. True fast imaging with steady-state precession (True-FISP), an implementation of steadystate free precession, is an increasingly utilized MRI method that provides a high-resolution anatomical image in a short acquisition time [18]. True-FISP shows a strong contrast between tissues with different ratios of T2 and T1, providing excellent visualization of fluid compartments and vasculature. So far, it has been specifically utilized for the morphological imaging of the cardiovascular system [19]. Lately, steady-state free precession combined with time-spatial labeling inversion pulse was evaluated for CMD in the normal kidney [20]. The first aim of this study was to evaluate whether vascular reactivity is affected during the progression of AKI, CKD and in the development of acute-on-chronic renal failure, as evaluated by HRI. An additional aim was to establish the use of these nonenhanced MRI methods for the evaluation of changes in renal morphology, vascular reactivity and perfusion in the presence of acute or chronic renal impairment, without the need for administration of potentially harmful contrast agents. Materials and Methods Animal Models Animal experiments were performed in accordance with the guidelines and approval of the Animal Care and Use Institutional Committee, which holds NIH approval (OPRR-A01–5011), on CB6F1/OlaHsd mice (n = 60; Harlan, Jerusalem, Israel) that were 7–8 weeks old. Blood samples were obtained by tail-vein bleeding. Serum urea and creatine kinase levels were determined using the Reflotron system (Roche Diagnostics, Basel, Switzerland). AKI Model. Rhabdomyolysis-induced AKI (myoglobinuric AKI) was induced by glycerol 50% (8 ml/kg body weight) injected into the anterior thigh muscle of both hind limbs. MRI scans were executed repeatedly on days 1, 4, 8, 15 and 22 following glycerol injection (n = 15). In an additional set of experiments (n = 10), we assessed the proportion of altered MRI data in correlation with the extent of renal injury. For this, we used an interleukin-6 (IL-6)/soluble IL-6
Am J Nephrol 2014;39:268–278 DOI: 10.1159/000360093
269
Downloaded by: Temple University 155.247.166.234 - 10/5/2016 2:38:51 PM
Introduction
MRI MRI experiments were performed on a horizontal 4.7-tesla Biospec spectrometer (Bruker, Ettlingen, Germany) using a 3.5cm birdcage coil. Mice were anesthetized with intraperitoneal injection of pentobarbital (70 mg/kg; CTS Group, Hod-Hasharon, Israel) and placed in a supine position. Kidney morphology was assessed using both True-FISP [repetition time/echo time (TR/ TE) = 3/1.5 ms, field of view 3.4 cm, in-plane resolution 133 μm, slice thickness 1 mm and flip angle 60°; 16 averages] and T2weighted (T2W) fast spin echo (SE) images (RARE sequence; TR = 2,000 ms, TE = 37 ms, field of view 3.4 cm, in-plane resolution 133 μm and slice thickness 1 mm). The chosen TR/TE values may be suboptimal. It is possible that a longer TE would prove more suitable in future studies. Regions of interest (ROIs) were drawn manually by using Analyze-7.0 (BIR, Mayo Clinic, Rochester, Minn., USA) based on axial True-FISP images. On the True-FISP images, renal ROIs covered the cortex or outer medulla of each kidney (fig. 1b) whereas on HRI maps, the entire kidneys were assessed. Two additional ROIs covered the adjacent back muscle and the spleen for normalization of the renal values as shown previously [14]. Kidney area was calculated from the central axial slice of each kidney using Analyze-7.0 software. The HRI protocol was implemented as previously described [15], using axial T2W gradient echo images which are sensitive to the local tissue oxygenation (TR/TE = 147/10 ms, field-of-view 3.4 cm, in-plane resolution 133 μm, slice thickness 1 mm, 2 averages, 256 × 256 pixels, resulting in a temporal resolution of 37 s) during breathing of ambient air (normoxia), with hypercapnia (ΔSco2, 5% CO2) for 4 min, followed by hyperoxia (ΔSo2, 5% CO2 + 95% O2) for 4 min, administered through a face mask (4 liters/min). HRI maps were generated as reported previously [15] using a home-built IDL (Interactive Data Language of ITT Visual Information Solutions, Boulder, Colo., USA) program. Briefly, the percentage change in signal intensity (SI), induced by ΔSco2 and ΔSo2 was calculated from the average of 6 repetitions for each gas admixture, by including only pixels with a statistical threshold of p < 0.05 calculated with a Mann-Whitney U test. Mean renal HRI values were calculated for the entire kidney, from the central axial slice of each kidney. CE-MRI was applied as a reference method to assess renal perfusion, performed with T1W gradient echo sequence (TR/TE = 58.3/5.2 ms, field-of-view 3.4 cm, in-plane resolution 133 μm, slice thickness 1 mm, 1 average and 60 repetitions, resulting in a tem-
270
Am J Nephrol 2014;39:268–278 DOI: 10.1159/000360093
poral resolution of 7.4 s). Ten repetitions were acquired at baseline, followed by 50 repetitions after Gd-DTPA (0.5 M, 100 μl; Magnetol, Soreq Radiopharmaceuticals, Yavne, Israel) administration via a tail-vein catheter at a dose of 0.1 mmol/kg. Histological and Immunohistochemical Analysis Kidney sections, obtained parallel to MRI scans, were routinely stained with hematoxylin and eosin (HE) and Masson’s trichrome staining was performed to assess fibrosis in the CKD model. In addition, apoptosis was assessed by immunostaining using TUNEL (terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end-labeling), performed with an in situ cell death detection kit (Roche Diagnostics Corp. Mannheim, Germany). Bromodeoxyuridine (BrdU) immunostaining (GE Healthcare AmershamTM, UK) was performed to assess renal tubular cell proliferation, injected intraperitoneally (1 ml/100 g) 3 h before mice were sacrificed. Following target retrieval, sections were incubated with primary mouse monoclonal anti-BrdU antibody (Neomarker, Fremont, Calif., USA). Mapping of renal hypoxic regions (pO2
